entry_point
stringlengths 1
65
| original_triton_code
stringlengths 4.5k
619k
| python_code
stringlengths 208
60.9k
| triton_code
stringlengths 1.15k
275k
| repo_name
stringlengths 7
115
| module_name
stringlengths 1
65
| synthetic
bool 1
class | uuid
int64 0
18.5k
| licenses
sequencelengths 1
6
| stars
int64 0
19.8k
| sha
stringlengths 40
40
| repo_link
stringlengths 72
180
| pytorch_code
stringlengths 200
4.05k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
NormalAttention_dot | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/zo/czobpmlyr5atbcpsuque6vcmk7nafmb3smtbzoqilz46drm7zbkm.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tl.store(in_out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/32/c32v7egt4mupqssam3gmac2qgv3ujprjybthsgweflmot256qqw7.py
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_2 => convolution_2
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/2x/c2xwbpm4ch2balahokzyrd5zmtndolk2v733exjzhzhai6vm7yxw.py
# Topologically Sorted Source Nodes: [energy_1, energy_2], Original ATen: [aten.elu, aten.div]
# Source node to ATen node mapping:
# energy_1 => expm1, gt, mul, mul_2, where
# energy_2 => div
# Graph fragment:
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%bmm, 0), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%bmm, 1.0), kwargs = {})
# %expm1 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1, 1.0), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %mul, %mul_2), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%where, 16), kwargs = {})
triton_poi_fused_div_elu_2 = async_compile.triton('triton_poi_fused_div_elu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_elu_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_elu_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = 0.0625
tmp9 = tmp7 * tmp8
tl.store(out_ptr0 + (x0), tmp9, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (1, ), (1, ))
assert_size_stride(primals_4, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (1, ), (1, ))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_9, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 1, 4, 4), (16, 16, 4, 1))
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(primals_1, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = reinterpret_tensor(buf0, (4, 1, 4, 4), (16, 1, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf2, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
buf3 = reinterpret_tensor(buf1, (4, 1, 4, 4), (16, 1, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf3, primals_5, 64, grid=grid(64), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (4, 16, 1), (16, 1, 0), 0), reinterpret_tensor(buf3, (4, 1, 16), (16, 0, 1), 0), out=buf4)
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf5 = extern_kernels.convolution(primals_1, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 4, 4), (64, 16, 4, 1))
buf6 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf6, primals_7, 256, grid=grid(256), stream=stream0)
del primals_7
buf7 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy_1, energy_2], Original ATen: [aten.elu, aten.div]
triton_poi_fused_div_elu_2.run(buf4, buf7, 1024, grid=grid(1024), stream=stream0)
buf8 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy_1, energy_2, bmm_1], Original ATen: [aten.elu, aten.div, aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf6, (4, 4, 16), (64, 16, 1), 0), buf7, out=buf8)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf9 = extern_kernels.convolution(reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0), primals_8, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 4, 4, 4), (64, 16, 4, 1))
buf10 = buf9; del buf9 # reuse
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf10, primals_9, 256, grid=grid(256), stream=stream0)
del primals_9
return (buf10, primals_1, primals_2, primals_4, primals_6, primals_8, buf4, reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf6, (4, 16, 4), (64, 1, 16), 0), reinterpret_tensor(buf7, (4, 16, 16), (256, 1, 16), 0), reinterpret_tensor(buf2, (4, 1, 16), (16, 16, 1), 0), reinterpret_tensor(buf3, (4, 16, 1), (16, 1, 16), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class NormalAttention_dot(nn.Module):
def __init__(self, input_channel_num, k=4):
super(NormalAttention_dot, self).__init__()
self.c_in = input_channel_num
self.query_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in // k, kernel_size=1)
self.key_conv = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in // k, kernel_size=1)
self.value_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in, kernel_size=1)
self.nonlin = nn.ELU()
self.gamma = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in, kernel_size=1)
def forward(self, x):
B, C, H, W = x.size()
proj_query = self.query_conv(x).view(B, -1, H * W).permute(0, 2, 1)
proj_key = self.key_conv(x).view(B, -1, H * W)
energy = torch.bmm(proj_query, proj_key)
energy = self.nonlin(energy)
energy = energy / (H * W)
proj_value = self.value_conv(x).view(B, -1, H * W)
out = torch.bmm(proj_value, energy).view(B, C, H, W)
out = self.gamma(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_channel_num': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tl.store(in_out_ptr0 + x0, tmp3, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_div_elu_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 1.0
tmp4 = tmp0 * tmp3
tmp5 = libdevice.expm1(tmp4)
tmp6 = tmp5 * tmp3
tmp7 = tl.where(tmp2, tmp4, tmp6)
tmp8 = 0.0625
tmp9 = tmp7 * tmp8
tl.store(out_ptr0 + x0, tmp9, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (1,), (1,))
assert_size_stride(primals_4, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (1,), (1,))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_9, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 1, 4, 4), (16, 16, 4, 1))
buf1 = extern_kernels.convolution(primals_1, primals_4, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = reinterpret_tensor(buf0, (4, 1, 4, 4), (16, 1, 4, 1), 0)
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(64)](buf2, primals_3, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_3
buf3 = reinterpret_tensor(buf1, (4, 1, 4, 4), (16, 1, 4, 1), 0)
del buf1
triton_poi_fused_convolution_0[grid(64)](buf3, primals_5, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf2, (4, 16, 1), (16, 1, 0),
0), reinterpret_tensor(buf3, (4, 1, 16), (16, 0, 1), 0), out=buf4)
buf5 = extern_kernels.convolution(primals_1, primals_6, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 4, 4), (64, 16, 4, 1))
buf6 = buf5
del buf5
triton_poi_fused_convolution_1[grid(256)](buf6, primals_7, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_7
buf7 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
triton_poi_fused_div_elu_2[grid(1024)](buf4, buf7, 1024, XBLOCK=256,
num_warps=4, num_stages=1)
buf8 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf6, (4, 4, 16), (64, 16, 1),
0), buf7, out=buf8)
buf9 = extern_kernels.convolution(reinterpret_tensor(buf8, (4, 4, 4,
4), (64, 16, 4, 1), 0), primals_8, stride=(1, 1), padding=(0, 0
), dilation=(1, 1), transposed=False, output_padding=(0, 0),
groups=1, bias=None)
assert_size_stride(buf9, (4, 4, 4, 4), (64, 16, 4, 1))
buf10 = buf9
del buf9
triton_poi_fused_convolution_1[grid(256)](buf10, primals_9, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_9
return (buf10, primals_1, primals_2, primals_4, primals_6, primals_8,
buf4, reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0),
reinterpret_tensor(buf6, (4, 16, 4), (64, 1, 16), 0),
reinterpret_tensor(buf7, (4, 16, 16), (256, 1, 16), 0),
reinterpret_tensor(buf2, (4, 1, 16), (16, 16, 1), 0),
reinterpret_tensor(buf3, (4, 16, 1), (16, 1, 16), 0))
class NormalAttention_dotNew(nn.Module):
def __init__(self, input_channel_num, k=4):
super(NormalAttention_dotNew, self).__init__()
self.c_in = input_channel_num
self.query_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in // k, kernel_size=1)
self.key_conv = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in // k, kernel_size=1)
self.value_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in, kernel_size=1)
self.nonlin = nn.ELU()
self.gamma = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in, kernel_size=1)
def forward(self, input_0):
primals_2 = self.query_conv.weight
primals_3 = self.query_conv.bias
primals_4 = self.key_conv.weight
primals_5 = self.key_conv.bias
primals_6 = self.value_conv.weight
primals_7 = self.value_conv.bias
primals_8 = self.gamma.weight
primals_9 = self.gamma.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0]
| Schwartz-Zha/My-invertible-resnet | NormalAttention_dot | false | 1,037 | [
"MIT"
] | 0 | 5415975bb0d640f3bf3ef4a7b986563e84109270 | https://github.com/Schwartz-Zha/My-invertible-resnet/tree/5415975bb0d640f3bf3ef4a7b986563e84109270 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, input_channel_num, k=4):
super().__init__()
self.c_in = input_channel_num
self.query_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in // k, kernel_size=1)
self.key_conv = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in // k, kernel_size=1)
self.value_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in, kernel_size=1)
self.nonlin = nn.ELU()
self.gamma = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in, kernel_size=1)
def forward(self, x):
B, C, H, W = x.size()
proj_query = self.query_conv(x).view(B, -1, H * W).permute(0, 2, 1)
proj_key = self.key_conv(x).view(B, -1, H * W)
energy = torch.bmm(proj_query, proj_key)
energy = self.nonlin(energy)
energy = energy / (H * W)
proj_value = self.value_conv(x).view(B, -1, H * W)
out = torch.bmm(proj_value, energy).view(B, C, H, W)
out = self.gamma(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
ActNorm2D | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/mk/cmkfq3mg2sl66skmhudljmlrjckeersi3ye3frqm2bl7felzsn2n.py
# Topologically Sorted Source Nodes: [sum_1, mul, logdet], Original ATen: [aten.sum, aten.mul]
# Source node to ATen node mapping:
# logdet => mul_1
# mul => mul
# sum_1 => sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%unsqueeze_2,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, 4), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, 4), kwargs = {})
triton_per_fused_mul_sum_0 = async_compile.triton('triton_per_fused_mul_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mul_sum_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mul_sum_0(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.sum(tmp1, 1)[:, None]
tmp4 = 4.0
tmp5 = tmp3 * tmp4
tmp6 = tmp5 * tmp4
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp6, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ok/cok2tne7527mk7p2rdtta5caje65jdloa27vp4xewaco2m24lfgp.py
# Topologically Sorted Source Nodes: [exp, mul_2, add], Original ATen: [aten.exp, aten.mul, aten.add]
# Source node to ATen node mapping:
# add => add
# exp => exp
# mul_2 => mul_2
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%unsqueeze_2,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %exp), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %unsqueeze_5), kwargs = {})
triton_poi_fused_add_exp_mul_1 = async_compile.triton('triton_poi_fused_add_exp_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_exp_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_exp_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tl_math.exp(tmp1)
tmp3 = tmp0 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + (x3), tmp5, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, ), (1, ))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [sum_1, mul, logdet], Original ATen: [aten.sum, aten.mul]
stream0 = get_raw_stream(0)
triton_per_fused_mul_sum_0.run(buf2, primals_1, 1, 4, grid=grid(1), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [exp, mul_2, add], Original ATen: [aten.exp, aten.mul, aten.add]
triton_poi_fused_add_exp_mul_1.run(primals_2, primals_1, primals_3, buf1, 256, grid=grid(256), stream=stream0)
del primals_3
return (buf1, buf2, primals_1, primals_2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
from torch.nn import Parameter
class ActNorm2D(nn.Module):
def __init__(self, num_channels, eps=1e-05):
super(ActNorm2D, self).__init__()
self.eps = eps
self.num_channels = num_channels
self._log_scale = Parameter(torch.Tensor(num_channels))
self._shift = Parameter(torch.Tensor(num_channels))
self._init = False
def log_scale(self):
return self._log_scale[None, :, None, None]
def shift(self):
return self._shift[None, :, None, None]
def forward(self, x):
if not self._init:
with torch.no_grad():
assert self.num_channels == x.size(1)
mean = torch.transpose(x, 0, 1).contiguous().view(self.
num_channels, -1).mean(dim=1)
zero_mean = x - mean[None, :, None, None]
var = torch.transpose(zero_mean ** 2, 0, 1).contiguous().view(
self.num_channels, -1).mean(dim=1)
std = (var + self.eps) ** 0.5
log_scale = torch.log(1.0 / std)
self._shift.data = -mean * torch.exp(log_scale)
self._log_scale.data = log_scale
self._init = True
log_scale = self.log_scale()
logdet = log_scale.sum() * x.size(2) * x.size(3)
return x * torch.exp(log_scale) + self.shift(), logdet
def inverse(self, x):
return (x - self.shift()) * torch.exp(-self.log_scale())
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_channels': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_mul_sum_0(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.sum(tmp1, 1)[:, None]
tmp4 = 4.0
tmp5 = tmp3 * tmp4
tmp6 = tmp5 * tmp4
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp6, None)
@triton.jit
def triton_poi_fused_add_exp_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp2 = tl_math.exp(tmp1)
tmp3 = tmp0 * tmp2
tmp5 = tmp3 + tmp4
tl.store(out_ptr0 + x3, tmp5, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4,), (1,))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf2 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_mul_sum_0[grid(1)](buf2, primals_1, 1, 4, XBLOCK=1,
num_warps=2, num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_exp_mul_1[grid(256)](primals_2, primals_1,
primals_3, buf1, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_3
return buf1, buf2, primals_1, primals_2
class ActNorm2DNew(nn.Module):
def __init__(self, num_channels, eps=1e-05):
super(ActNorm2DNew, self).__init__()
self.eps = eps
self.num_channels = num_channels
self._log_scale = Parameter(torch.Tensor(num_channels))
self._shift = Parameter(torch.Tensor(num_channels))
self._init = False
def log_scale(self):
return self._log_scale[None, :, None, None]
def shift(self):
return self._shift[None, :, None, None]
def inverse(self, x):
return (x - self.shift()) * torch.exp(-self.log_scale())
def forward(self, input_0):
primals_1 = self._log_scale
primals_3 = self._shift
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0], output[1]
| Schwartz-Zha/My-invertible-resnet | ActNorm2D | false | 1,038 | [
"MIT"
] | 0 | 5415975bb0d640f3bf3ef4a7b986563e84109270 | https://github.com/Schwartz-Zha/My-invertible-resnet/tree/5415975bb0d640f3bf3ef4a7b986563e84109270 | import torch
import torch.nn as nn
from torch.nn import Parameter
class Model(nn.Module):
def __init__(self, num_channels, eps=1e-05):
super().__init__()
self.eps = eps
self.num_channels = num_channels
self._log_scale = Parameter(torch.Tensor(num_channels))
self._shift = Parameter(torch.Tensor(num_channels))
self._init = False
def log_scale(self):
return self._log_scale[None, :, None, None]
def shift(self):
return self._shift[None, :, None, None]
def forward(self, x):
if not self._init:
with torch.no_grad():
assert self.num_channels == x.size(1)
mean = torch.transpose(x, 0, 1).contiguous().view(self.
num_channels, -1).mean(dim=1)
zero_mean = x - mean[None, :, None, None]
var = torch.transpose(zero_mean ** 2, 0, 1).contiguous().view(
self.num_channels, -1).mean(dim=1)
std = (var + self.eps) ** 0.5
log_scale = torch.log(1.0 / std)
self._shift.data = -mean * torch.exp(log_scale)
self._log_scale.data = log_scale
self._init = True
log_scale = self.log_scale()
logdet = log_scale.sum() * x.size(2) * x.size(3)
return x * torch.exp(log_scale) + self.shift(), logdet
def inverse(self, x):
return (x - self.shift()) * torch.exp(-self.log_scale())
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
NormalAttention_embedded_gaussian | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/zo/czobpmlyr5atbcpsuque6vcmk7nafmb3smtbzoqilz46drm7zbkm.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tl.store(in_out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/63/c63abizbceybowmxasyhbvyg6gmcs7m552tgg653wdhaqei77hz3.py
# Topologically Sorted Source Nodes: [energy_1, energy_sum, energy_2], Original ATen: [aten.exp, aten.sum, aten.div]
# Source node to ATen node mapping:
# energy_1 => exp
# energy_2 => div
# energy_sum => sum_1
# Graph fragment:
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%bmm,), kwargs = {})
# %sum_1 : [num_users=2] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [2], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_per_fused_div_exp_sum_1 = async_compile.triton('triton_per_fused_div_exp_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[64, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_exp_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_div_exp_sum_1(in_ptr0, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl_math.exp(tmp0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = tmp1 / tmp5
tl.store(out_ptr1 + (r1 + (16*x0)), tmp6, xmask)
tl.store(out_ptr0 + (x0), tmp5, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/w5/cw5gytijzzkwnfpq2a2axdsj4pfxgxmwiuzizuyd4bw5uwnanzw7.py
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_2 => convolution_2
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_1, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (1, ), (1, ))
assert_size_stride(primals_4, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (1, ), (1, ))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_9, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 1, 4, 4), (16, 16, 4, 1))
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(primals_1, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = reinterpret_tensor(buf0, (4, 1, 4, 4), (16, 1, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf2, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
buf3 = reinterpret_tensor(buf1, (4, 1, 4, 4), (16, 1, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf3, primals_5, 64, grid=grid(64), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (4, 16, 1), (16, 1, 0), 0), reinterpret_tensor(buf3, (4, 1, 16), (16, 0, 1), 0), out=buf4)
buf5 = empty_strided_cuda((4, 16, 1), (16, 1, 1), torch.float32)
buf6 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [energy_1, energy_sum, energy_2], Original ATen: [aten.exp, aten.sum, aten.div]
triton_per_fused_div_exp_sum_1.run(buf4, buf5, buf6, 64, 16, grid=grid(64), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(primals_1, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 4, 4, 4), (64, 16, 4, 1))
buf8 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf8, primals_7, 256, grid=grid(256), stream=stream0)
del primals_7
buf9 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [bmm_1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf8, (4, 4, 16), (64, 16, 1), 0), buf6, out=buf9)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf10 = extern_kernels.convolution(reinterpret_tensor(buf9, (4, 4, 4, 4), (64, 16, 4, 1), 0), primals_8, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf10, (4, 4, 4, 4), (64, 16, 4, 1))
buf11 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf11, primals_9, 256, grid=grid(256), stream=stream0)
del primals_9
return (buf11, primals_1, primals_2, primals_4, primals_6, primals_8, buf4, buf5, buf6, reinterpret_tensor(buf9, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf8, (4, 16, 4), (64, 1, 16), 0), reinterpret_tensor(buf2, (4, 1, 16), (16, 16, 1), 0), reinterpret_tensor(buf3, (4, 16, 1), (16, 1, 16), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class NormalAttention_embedded_gaussian(nn.Module):
def __init__(self, input_channel_num, k=4):
super(NormalAttention_embedded_gaussian, self).__init__()
self.c_in = input_channel_num
self.query_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in // k, kernel_size=1)
self.key_conv = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in // k, kernel_size=1)
self.value_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in, kernel_size=1)
self.gamma = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in, kernel_size=1)
def forward(self, x):
B, C, H, W = x.size()
proj_query = self.query_conv(x).view(B, -1, H * W).permute(0, 2, 1)
proj_key = self.key_conv(x).view(B, -1, H * W)
energy = torch.bmm(proj_query, proj_key)
energy = torch.exp(energy)
energy_sum = torch.sum(energy, dim=2, keepdim=True)
energy = energy / energy_sum
proj_value = self.value_conv(x).view(B, -1, H * W)
out = torch.bmm(proj_value, energy).view(B, C, H, W)
out = self.gamma(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_channel_num': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tl.store(in_out_ptr0 + x0, tmp3, xmask)
@triton.jit
def triton_per_fused_div_exp_sum_1(in_ptr0, out_ptr0, out_ptr1, xnumel,
rnumel, XBLOCK: tl.constexpr):
xnumel = 64
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl_math.exp(tmp0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = tmp1 / tmp5
tl.store(out_ptr1 + (r1 + 16 * x0), tmp6, xmask)
tl.store(out_ptr0 + x0, tmp5, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (1,), (1,))
assert_size_stride(primals_4, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_5, (1,), (1,))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_9, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 1, 4, 4), (16, 16, 4, 1))
buf1 = extern_kernels.convolution(primals_1, primals_4, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = reinterpret_tensor(buf0, (4, 1, 4, 4), (16, 1, 4, 1), 0)
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(64)](buf2, primals_3, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_3
buf3 = reinterpret_tensor(buf1, (4, 1, 4, 4), (16, 1, 4, 1), 0)
del buf1
triton_poi_fused_convolution_0[grid(64)](buf3, primals_5, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf2, (4, 16, 1), (16, 1, 0),
0), reinterpret_tensor(buf3, (4, 1, 16), (16, 0, 1), 0), out=buf4)
buf5 = empty_strided_cuda((4, 16, 1), (16, 1, 1), torch.float32)
buf6 = empty_strided_cuda((4, 16, 16), (256, 16, 1), torch.float32)
triton_per_fused_div_exp_sum_1[grid(64)](buf4, buf5, buf6, 64, 16,
XBLOCK=32, num_warps=4, num_stages=1)
buf7 = extern_kernels.convolution(primals_1, primals_6, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 4, 4, 4), (64, 16, 4, 1))
buf8 = buf7
del buf7
triton_poi_fused_convolution_2[grid(256)](buf8, primals_7, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_7
buf9 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf8, (4, 4, 16), (64, 16, 1),
0), buf6, out=buf9)
buf10 = extern_kernels.convolution(reinterpret_tensor(buf9, (4, 4,
4, 4), (64, 16, 4, 1), 0), primals_8, stride=(1, 1), padding=(0,
0), dilation=(1, 1), transposed=False, output_padding=(0, 0),
groups=1, bias=None)
assert_size_stride(buf10, (4, 4, 4, 4), (64, 16, 4, 1))
buf11 = buf10
del buf10
triton_poi_fused_convolution_2[grid(256)](buf11, primals_9, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_9
return (buf11, primals_1, primals_2, primals_4, primals_6, primals_8,
buf4, buf5, buf6, reinterpret_tensor(buf9, (4, 4, 4, 4), (64, 16, 4,
1), 0), reinterpret_tensor(buf8, (4, 16, 4), (64, 1, 16), 0),
reinterpret_tensor(buf2, (4, 1, 16), (16, 16, 1), 0),
reinterpret_tensor(buf3, (4, 16, 1), (16, 1, 16), 0))
class NormalAttention_embedded_gaussianNew(nn.Module):
def __init__(self, input_channel_num, k=4):
super(NormalAttention_embedded_gaussianNew, self).__init__()
self.c_in = input_channel_num
self.query_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in // k, kernel_size=1)
self.key_conv = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in // k, kernel_size=1)
self.value_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in, kernel_size=1)
self.gamma = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in, kernel_size=1)
def forward(self, input_0):
primals_2 = self.query_conv.weight
primals_3 = self.query_conv.bias
primals_4 = self.key_conv.weight
primals_5 = self.key_conv.bias
primals_6 = self.value_conv.weight
primals_7 = self.value_conv.bias
primals_8 = self.gamma.weight
primals_9 = self.gamma.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0]
| Schwartz-Zha/My-invertible-resnet | NormalAttention_embedded_gaussian | false | 1,039 | [
"MIT"
] | 0 | 5415975bb0d640f3bf3ef4a7b986563e84109270 | https://github.com/Schwartz-Zha/My-invertible-resnet/tree/5415975bb0d640f3bf3ef4a7b986563e84109270 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, input_channel_num, k=4):
super().__init__()
self.c_in = input_channel_num
self.query_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in // k, kernel_size=1)
self.key_conv = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in // k, kernel_size=1)
self.value_conv = nn.Conv2d(in_channels=self.c_in, out_channels=
self.c_in, kernel_size=1)
self.gamma = nn.Conv2d(in_channels=self.c_in, out_channels=self.
c_in, kernel_size=1)
def forward(self, x):
B, C, H, W = x.size()
proj_query = self.query_conv(x).view(B, -1, H * W).permute(0, 2, 1)
proj_key = self.key_conv(x).view(B, -1, H * W)
energy = torch.bmm(proj_query, proj_key)
energy = torch.exp(energy)
energy_sum = torch.sum(energy, dim=2, keepdim=True)
energy = energy / energy_sum
proj_value = self.value_conv(x).view(B, -1, H * W)
out = torch.bmm(proj_value, energy).view(B, C, H, W)
out = self.gamma(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
Net | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/hr/chrb4easzia4dw753qymwrkwdvp5554y3k2pc6zt3dtjbxgoihzj.py
# Topologically Sorted Source Nodes: [pi1], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# pi1 => tanh
# Graph fragment:
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%view_1,), kwargs = {})
triton_poi_fused_tanh_0 = async_compile.triton('triton_poi_fused_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_tanh_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_tanh_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 128), (128, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (128, 4), (4, 1))
assert_size_stride(primals_7, (128, ), (1, ))
assert_size_stride(primals_8, (1, 128), (128, 1))
assert_size_stride(primals_9, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [pi1], Original ATen: [aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_tanh_0.run(buf1, primals_2, 8192, grid=grid(8192), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [logits], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 128), (128, 1), 0), reinterpret_tensor(primals_4, (128, 4), (1, 128), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 128), (1, 4), 0), out=buf3)
del primals_6
buf4 = reinterpret_tensor(buf3, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [v1], Original ATen: [aten.tanh]
triton_poi_fused_tanh_0.run(buf4, primals_7, 8192, grid=grid(8192), stream=stream0)
del primals_7
buf6 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [values], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf4, (64, 128), (128, 1), 0), reinterpret_tensor(primals_8, (128, 1), (1, 128), 0), alpha=1, beta=1, out=buf6)
del primals_9
return (reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf6, (4, 4, 4, 1), (16, 4, 1, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf1, buf4, primals_8, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class Net(nn.Module):
def __init__(self, s_dim, a_dim):
super(Net, self).__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.pi1 = nn.Linear(s_dim, 128)
self.pi2 = nn.Linear(128, a_dim)
self.v1 = nn.Linear(s_dim, 128)
self.v2 = nn.Linear(128, 1)
set_init([self.pi1, self.pi2, self.v1, self.v2])
self.distribution = torch.distributions.Categorical
def forward(self, x):
pi1 = torch.tanh(self.pi1(x))
logits = self.pi2(pi1)
v1 = torch.tanh(self.v1(x))
values = self.v2(v1)
return logits, values
def choose_action(self, s):
self.eval()
logits, _ = self.forward(s)
prob = F.softmax(logits, dim=1).data
m = self.distribution(prob)
return m.sample().numpy()[0]
def loss_func(self, s, a, v_t):
self.train()
logits, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
probs = F.softmax(logits, dim=1)
m = self.distribution(probs)
exp_v = m.log_prob(a) * td.detach().squeeze()
a_loss = -exp_v
total_loss = (c_loss + a_loss).mean()
return total_loss
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'s_dim': 4, 'a_dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_tanh_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, None)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 128), (128, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (128, 4), (4, 1))
assert_size_stride(primals_7, (128,), (1,))
assert_size_stride(primals_8, (1, 128), (128, 1))
assert_size_stride(primals_9, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf0
get_raw_stream(0)
triton_poi_fused_tanh_0[grid(8192)](buf1, primals_2, 8192, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 128),
(128, 1), 0), reinterpret_tensor(primals_4, (128, 4), (1, 128),
0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 128), (1, 4), 0), out=buf3)
del primals_6
buf4 = reinterpret_tensor(buf3, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf3
triton_poi_fused_tanh_0[grid(8192)](buf4, primals_7, 8192, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_7
buf6 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_9, reinterpret_tensor(buf4, (64, 128),
(128, 1), 0), reinterpret_tensor(primals_8, (128, 1), (1, 128),
0), alpha=1, beta=1, out=buf6)
del primals_9
return reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(buf6, (4, 4, 4, 1), (16, 4, 1, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf1, buf4, primals_8, primals_4
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class NetNew(nn.Module):
def __init__(self, s_dim, a_dim):
super(NetNew, self).__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.pi1 = nn.Linear(s_dim, 128)
self.pi2 = nn.Linear(128, a_dim)
self.v1 = nn.Linear(s_dim, 128)
self.v2 = nn.Linear(128, 1)
set_init([self.pi1, self.pi2, self.v1, self.v2])
self.distribution = torch.distributions.Categorical
def choose_action(self, s):
self.eval()
logits, _ = self.forward(s)
prob = F.softmax(logits, dim=1).data
m = self.distribution(prob)
return m.sample().numpy()[0]
def loss_func(self, s, a, v_t):
self.train()
logits, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
probs = F.softmax(logits, dim=1)
m = self.distribution(probs)
exp_v = m.log_prob(a) * td.detach().squeeze()
a_loss = -exp_v
total_loss = (c_loss + a_loss).mean()
return total_loss
def forward(self, input_0):
primals_1 = self.pi1.weight
primals_2 = self.pi1.bias
primals_4 = self.pi2.weight
primals_5 = self.pi2.bias
primals_6 = self.v1.weight
primals_7 = self.v1.bias
primals_8 = self.v2.weight
primals_9 = self.v2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0], output[1]
| SeungyounShin/pytorch-A3C | Net | false | 1,040 | [
"MIT"
] | 0 | acb9c05a5e1a697c48a7d4c1a48b1c86326faf91 | https://github.com/SeungyounShin/pytorch-A3C/tree/acb9c05a5e1a697c48a7d4c1a48b1c86326faf91 | import torch
import torch.nn as nn
import torch.nn.functional as F
def set_init(layers):
for layer in layers:
nn.init.normal_(layer.weight, mean=0.0, std=0.1)
nn.init.constant_(layer.bias, 0.0)
class Model(nn.Module):
def __init__(self, s_dim, a_dim):
super().__init__()
self.s_dim = s_dim
self.a_dim = a_dim
self.pi1 = nn.Linear(s_dim, 128)
self.pi2 = nn.Linear(128, a_dim)
self.v1 = nn.Linear(s_dim, 128)
self.v2 = nn.Linear(128, 1)
set_init([self.pi1, self.pi2, self.v1, self.v2])
self.distribution = torch.distributions.Categorical
def forward(self, x):
pi1 = torch.tanh(self.pi1(x))
logits = self.pi2(pi1)
v1 = torch.tanh(self.v1(x))
values = self.v2(v1)
return logits, values
def choose_action(self, s):
self.eval()
logits, _ = self.forward(s)
prob = F.softmax(logits, dim=1).data
m = self.distribution(prob)
return m.sample().numpy()[0]
def loss_func(self, s, a, v_t):
self.train()
logits, values = self.forward(s)
td = v_t - values
c_loss = td.pow(2)
probs = F.softmax(logits, dim=1)
m = self.distribution(probs)
exp_v = m.log_prob(a) * td.detach().squeeze()
a_loss = -exp_v
total_loss = (c_loss + a_loss).mean()
return total_loss
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
DiceLoss | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/um/cum65j23qchrjf5dndblqgbw6zomhgwfj2obfidtgy7b5j3zwklm.py
# Topologically Sorted Source Nodes: [pred], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# pred => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_0 = async_compile.triton('triton_poi_fused__softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/wk/cwk2wao7opapqbjj7klnqrd6tgist3ts3nc5veryzhzstwpx7d4l.py
# Topologically Sorted Source Nodes: [pred], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# pred => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=4] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/uq/cuqmt6h2h2qrwlfougoxbc4cnmi6o42afntw3nwkdzak2kpz2xum.py
# Topologically Sorted Source Nodes: [mul, ne, valid_mask, valid_mask_1, mul_1, sum_1, mul_2, num, pow_1, pow_2, add_1, sum_2, den, truediv, loss, loss_1, total_loss, mul_3, valid_mask_2, mul_4, sum_3, mul_5, num_1, pow_3, pow_4, add_5, sum_4, den_1, truediv_1, loss_2, loss_3, total_loss_1, mul_6, valid_mask_3, mul_7, sum_5, mul_8, num_2, pow_5, pow_6, add_8, sum_6, den_2, truediv_2, loss_4, loss_5, total_loss_2, mul_9, valid_mask_4, mul_10, sum_7, mul_11, num_3, pow_7, pow_8, add_11, sum_8, den_3, truediv_3, loss_6, loss_7, total_loss_3, loss_8, loss_9, loss_10], Original ATen: [aten.mul, aten.ne, aten._to_copy, aten.view, aten.sum, aten.add, aten.pow, aten.div, aten.rsub, aten.mean]
# Source node to ATen node mapping:
# add_1 => add_1
# add_11 => add_13
# add_5 => add_5
# add_8 => add_9
# den => add_2
# den_1 => add_6
# den_2 => add_10
# den_3 => add_14
# loss => sub_1
# loss_1 => mean
# loss_10 => mul_12
# loss_2 => sub_2
# loss_3 => mean_1
# loss_4 => sub_3
# loss_5 => mean_2
# loss_6 => sub_4
# loss_7 => mean_3
# loss_8 => div_5
# loss_9 => mean_4
# mul => mul
# mul_1 => mul_1
# mul_10 => mul_10
# mul_11 => mul_11
# mul_2 => mul_2
# mul_3 => mul_3
# mul_4 => mul_4
# mul_5 => mul_5
# mul_6 => mul_6
# mul_7 => mul_7
# mul_8 => mul_8
# mul_9 => mul_9
# ne => ne
# num => add
# num_1 => add_4
# num_2 => add_8
# num_3 => add_12
# pow_1 => pow_1
# pow_2 => pow_2
# pow_3 => pow_3
# pow_4 => pow_4
# pow_5 => pow_5
# pow_6 => pow_6
# pow_7 => pow_7
# pow_8 => pow_8
# sum_1 => sum_2
# sum_2 => sum_3
# sum_3 => sum_4
# sum_4 => sum_5
# sum_5 => sum_6
# sum_6 => sum_7
# sum_7 => sum_8
# sum_8 => sum_9
# total_loss => add_3
# total_loss_1 => add_7
# total_loss_2 => add_11
# total_loss_3 => add_15
# truediv => div_1
# truediv_1 => div_2
# truediv_2 => div_3
# truediv_3 => div_4
# valid_mask => convert_element_type_2
# valid_mask_1 => view_2
# valid_mask_2 => view_5
# valid_mask_3 => view_8
# valid_mask_4 => view_11
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %select_1), kwargs = {})
# %ne : [num_users=1] = call_function[target=torch.ops.aten.ne.Scalar](args = (%arg1_1, 255), kwargs = {})
# %convert_element_type_2 : [num_users=4] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%ne, torch.int64), kwargs = {})
# %view_2 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %view_2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_1, [1]), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_2, 2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, 1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view, 2), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_1, 2), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_1, %pow_2), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_1, [1]), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_3, 1), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %add_2), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_1), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_1,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, 0), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, %select_3), kwargs = {})
# %view_5 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_3, %view_5), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_4, [1]), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_4, 2), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_5, 1), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_3, 2), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_3, 2), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_3, %pow_4), kwargs = {})
# %sum_5 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_5, [1]), kwargs = {})
# %add_6 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_5, 1), kwargs = {})
# %div_2 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_4, %add_6), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_2), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_2,), kwargs = {})
# %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_3, %mean_1), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_6, %select_5), kwargs = {})
# %view_8 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_6, %view_8), kwargs = {})
# %sum_6 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_7, [1]), kwargs = {})
# %mul_8 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_6, 2), kwargs = {})
# %add_8 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_8, 1), kwargs = {})
# %pow_5 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_6, 2), kwargs = {})
# %pow_6 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_5, 2), kwargs = {})
# %add_9 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_5, %pow_6), kwargs = {})
# %sum_7 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_9, [1]), kwargs = {})
# %add_10 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_7, 1), kwargs = {})
# %div_3 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_8, %add_10), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_3), kwargs = {})
# %mean_2 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_3,), kwargs = {})
# %add_11 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_7, %mean_2), kwargs = {})
# %mul_9 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_9, %select_7), kwargs = {})
# %view_11 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_10 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_9, %view_11), kwargs = {})
# %sum_8 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_10, [1]), kwargs = {})
# %mul_11 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_8, 2), kwargs = {})
# %add_12 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_11, 1), kwargs = {})
# %pow_7 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_9, 2), kwargs = {})
# %pow_8 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_7, 2), kwargs = {})
# %add_13 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_7, %pow_8), kwargs = {})
# %sum_9 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_13, [1]), kwargs = {})
# %add_14 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_9, 1), kwargs = {})
# %div_4 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_12, %add_14), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_4), kwargs = {})
# %mean_3 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_4,), kwargs = {})
# %add_15 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_11, %mean_3), kwargs = {})
# %div_5 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_15, 4), kwargs = {})
# %mean_4 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%div_5,), kwargs = {})
# %mul_12 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean_4, 1.0), kwargs = {})
triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2 = async_compile.triton('triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (4*r0), None, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*r0), None, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr1 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr1 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp42 = tl.load(in_ptr1 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp71 = tl.load(in_ptr0 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp112 = tl.load(in_ptr0 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp153 = tl.load(in_ptr0 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp2 = tmp1.to(tl.int64)
tmp3 = tl.full([1, 1], 0, tl.int64)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = tl.full([1, 1], 3, tl.int64)
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp6 == tmp3
tmp8 = tmp7.to(tl.int64)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp0 * tmp9
tmp11 = 255.0
tmp12 = tmp1 != tmp11
tmp13 = tmp12.to(tl.int64)
tmp14 = tmp13.to(tl.float32)
tmp15 = tmp10 * tmp14
tmp17 = tmp16.to(tl.int64)
tmp18 = triton_helpers.maximum(tmp17, tmp3)
tmp19 = triton_helpers.minimum(tmp18, tmp5)
tmp20 = tmp19 == tmp3
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp21.to(tl.float32)
tmp23 = tmp0 * tmp22
tmp24 = tmp16 != tmp11
tmp25 = tmp24.to(tl.int64)
tmp26 = tmp25.to(tl.float32)
tmp27 = tmp23 * tmp26
tmp28 = tmp15 + tmp27
tmp30 = tmp29.to(tl.int64)
tmp31 = triton_helpers.maximum(tmp30, tmp3)
tmp32 = triton_helpers.minimum(tmp31, tmp5)
tmp33 = tmp32 == tmp3
tmp34 = tmp33.to(tl.int64)
tmp35 = tmp34.to(tl.float32)
tmp36 = tmp0 * tmp35
tmp37 = tmp29 != tmp11
tmp38 = tmp37.to(tl.int64)
tmp39 = tmp38.to(tl.float32)
tmp40 = tmp36 * tmp39
tmp41 = tmp28 + tmp40
tmp43 = tmp42.to(tl.int64)
tmp44 = triton_helpers.maximum(tmp43, tmp3)
tmp45 = triton_helpers.minimum(tmp44, tmp5)
tmp46 = tmp45 == tmp3
tmp47 = tmp46.to(tl.int64)
tmp48 = tmp47.to(tl.float32)
tmp49 = tmp0 * tmp48
tmp50 = tmp42 != tmp11
tmp51 = tmp50.to(tl.int64)
tmp52 = tmp51.to(tl.float32)
tmp53 = tmp49 * tmp52
tmp54 = tmp41 + tmp53
tmp55 = tmp0 * tmp0
tmp56 = tmp8 * tmp8
tmp57 = tmp56.to(tl.float32)
tmp58 = tmp55 + tmp57
tmp59 = tmp21 * tmp21
tmp60 = tmp59.to(tl.float32)
tmp61 = tmp55 + tmp60
tmp62 = tmp58 + tmp61
tmp63 = tmp34 * tmp34
tmp64 = tmp63.to(tl.float32)
tmp65 = tmp55 + tmp64
tmp66 = tmp62 + tmp65
tmp67 = tmp47 * tmp47
tmp68 = tmp67.to(tl.float32)
tmp69 = tmp55 + tmp68
tmp70 = tmp66 + tmp69
tmp72 = tl.full([1, 1], 1, tl.int64)
tmp73 = tmp6 == tmp72
tmp74 = tmp73.to(tl.int64)
tmp75 = tmp74.to(tl.float32)
tmp76 = tmp71 * tmp75
tmp77 = tmp76 * tmp14
tmp78 = tmp19 == tmp72
tmp79 = tmp78.to(tl.int64)
tmp80 = tmp79.to(tl.float32)
tmp81 = tmp71 * tmp80
tmp82 = tmp81 * tmp26
tmp83 = tmp77 + tmp82
tmp84 = tmp32 == tmp72
tmp85 = tmp84.to(tl.int64)
tmp86 = tmp85.to(tl.float32)
tmp87 = tmp71 * tmp86
tmp88 = tmp87 * tmp39
tmp89 = tmp83 + tmp88
tmp90 = tmp45 == tmp72
tmp91 = tmp90.to(tl.int64)
tmp92 = tmp91.to(tl.float32)
tmp93 = tmp71 * tmp92
tmp94 = tmp93 * tmp52
tmp95 = tmp89 + tmp94
tmp96 = tmp71 * tmp71
tmp97 = tmp74 * tmp74
tmp98 = tmp97.to(tl.float32)
tmp99 = tmp96 + tmp98
tmp100 = tmp79 * tmp79
tmp101 = tmp100.to(tl.float32)
tmp102 = tmp96 + tmp101
tmp103 = tmp99 + tmp102
tmp104 = tmp85 * tmp85
tmp105 = tmp104.to(tl.float32)
tmp106 = tmp96 + tmp105
tmp107 = tmp103 + tmp106
tmp108 = tmp91 * tmp91
tmp109 = tmp108.to(tl.float32)
tmp110 = tmp96 + tmp109
tmp111 = tmp107 + tmp110
tmp113 = tl.full([1, 1], 2, tl.int64)
tmp114 = tmp6 == tmp113
tmp115 = tmp114.to(tl.int64)
tmp116 = tmp115.to(tl.float32)
tmp117 = tmp112 * tmp116
tmp118 = tmp117 * tmp14
tmp119 = tmp19 == tmp113
tmp120 = tmp119.to(tl.int64)
tmp121 = tmp120.to(tl.float32)
tmp122 = tmp112 * tmp121
tmp123 = tmp122 * tmp26
tmp124 = tmp118 + tmp123
tmp125 = tmp32 == tmp113
tmp126 = tmp125.to(tl.int64)
tmp127 = tmp126.to(tl.float32)
tmp128 = tmp112 * tmp127
tmp129 = tmp128 * tmp39
tmp130 = tmp124 + tmp129
tmp131 = tmp45 == tmp113
tmp132 = tmp131.to(tl.int64)
tmp133 = tmp132.to(tl.float32)
tmp134 = tmp112 * tmp133
tmp135 = tmp134 * tmp52
tmp136 = tmp130 + tmp135
tmp137 = tmp112 * tmp112
tmp138 = tmp115 * tmp115
tmp139 = tmp138.to(tl.float32)
tmp140 = tmp137 + tmp139
tmp141 = tmp120 * tmp120
tmp142 = tmp141.to(tl.float32)
tmp143 = tmp137 + tmp142
tmp144 = tmp140 + tmp143
tmp145 = tmp126 * tmp126
tmp146 = tmp145.to(tl.float32)
tmp147 = tmp137 + tmp146
tmp148 = tmp144 + tmp147
tmp149 = tmp132 * tmp132
tmp150 = tmp149.to(tl.float32)
tmp151 = tmp137 + tmp150
tmp152 = tmp148 + tmp151
tmp154 = tmp6 == tmp5
tmp155 = tmp154.to(tl.int64)
tmp156 = tmp155.to(tl.float32)
tmp157 = tmp153 * tmp156
tmp158 = tmp157 * tmp14
tmp159 = tmp19 == tmp5
tmp160 = tmp159.to(tl.int64)
tmp161 = tmp160.to(tl.float32)
tmp162 = tmp153 * tmp161
tmp163 = tmp162 * tmp26
tmp164 = tmp158 + tmp163
tmp165 = tmp32 == tmp5
tmp166 = tmp165.to(tl.int64)
tmp167 = tmp166.to(tl.float32)
tmp168 = tmp153 * tmp167
tmp169 = tmp168 * tmp39
tmp170 = tmp164 + tmp169
tmp171 = tmp45 == tmp5
tmp172 = tmp171.to(tl.int64)
tmp173 = tmp172.to(tl.float32)
tmp174 = tmp153 * tmp173
tmp175 = tmp174 * tmp52
tmp176 = tmp170 + tmp175
tmp177 = tmp153 * tmp153
tmp178 = tmp155 * tmp155
tmp179 = tmp178.to(tl.float32)
tmp180 = tmp177 + tmp179
tmp181 = tmp160 * tmp160
tmp182 = tmp181.to(tl.float32)
tmp183 = tmp177 + tmp182
tmp184 = tmp180 + tmp183
tmp185 = tmp166 * tmp166
tmp186 = tmp185.to(tl.float32)
tmp187 = tmp177 + tmp186
tmp188 = tmp184 + tmp187
tmp189 = tmp172 * tmp172
tmp190 = tmp189.to(tl.float32)
tmp191 = tmp177 + tmp190
tmp192 = tmp188 + tmp191
tmp193 = 2.0
tmp194 = tmp54 * tmp193
tmp195 = 1.0
tmp196 = tmp194 + tmp195
tmp197 = tmp70 + tmp195
tmp198 = tmp196 / tmp197
tmp199 = tmp195 - tmp198
tmp200 = tl.broadcast_to(tmp199, [XBLOCK, RBLOCK])
tmp202 = tl.sum(tmp200, 1)[:, None]
tmp203 = tmp95 * tmp193
tmp204 = tmp203 + tmp195
tmp205 = tmp111 + tmp195
tmp206 = tmp204 / tmp205
tmp207 = tmp195 - tmp206
tmp208 = tl.broadcast_to(tmp207, [XBLOCK, RBLOCK])
tmp210 = tl.sum(tmp208, 1)[:, None]
tmp211 = tmp136 * tmp193
tmp212 = tmp211 + tmp195
tmp213 = tmp152 + tmp195
tmp214 = tmp212 / tmp213
tmp215 = tmp195 - tmp214
tmp216 = tl.broadcast_to(tmp215, [XBLOCK, RBLOCK])
tmp218 = tl.sum(tmp216, 1)[:, None]
tmp219 = tmp176 * tmp193
tmp220 = tmp219 + tmp195
tmp221 = tmp192 + tmp195
tmp222 = tmp220 / tmp221
tmp223 = tmp195 - tmp222
tmp224 = tl.broadcast_to(tmp223, [XBLOCK, RBLOCK])
tmp226 = tl.sum(tmp224, 1)[:, None]
tmp227 = 4.0
tmp228 = tmp202 / tmp227
tmp229 = 0.0
tmp230 = tmp228 + tmp229
tmp231 = tmp210 / tmp227
tmp232 = tmp230 + tmp231
tmp233 = tmp218 / tmp227
tmp234 = tmp232 + tmp233
tmp235 = tmp226 / tmp227
tmp236 = tmp234 + tmp235
tmp237 = 0.25
tmp238 = tmp236 * tmp237
tmp239 = tmp238 / tmp195
tmp240 = tmp239 * tmp195
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp240, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4), (4, 1))
assert_size_stride(arg1_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pred], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(arg0_1, buf0, 16, grid=grid(16), stream=stream0)
del arg0_1
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pred], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf0, buf1, 16, grid=grid(16), stream=stream0)
del buf0
buf10 = empty_strided_cuda((), (), torch.float32)
buf14 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [mul, ne, valid_mask, valid_mask_1, mul_1, sum_1, mul_2, num, pow_1, pow_2, add_1, sum_2, den, truediv, loss, loss_1, total_loss, mul_3, valid_mask_2, mul_4, sum_3, mul_5, num_1, pow_3, pow_4, add_5, sum_4, den_1, truediv_1, loss_2, loss_3, total_loss_1, mul_6, valid_mask_3, mul_7, sum_5, mul_8, num_2, pow_5, pow_6, add_8, sum_6, den_2, truediv_2, loss_4, loss_5, total_loss_2, mul_9, valid_mask_4, mul_10, sum_7, mul_11, num_3, pow_7, pow_8, add_11, sum_8, den_3, truediv_3, loss_6, loss_7, total_loss_3, loss_8, loss_9, loss_10], Original ATen: [aten.mul, aten.ne, aten._to_copy, aten.view, aten.sum, aten.add, aten.pow, aten.div, aten.rsub, aten.mean]
triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2.run(buf14, buf1, arg1_1, 1, 4, grid=grid(1), stream=stream0)
del arg1_1
del buf1
return (buf14, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import functools
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
def reduce_loss(loss, reduction):
"""Reduce loss as specified.
Args:
loss (Tensor): Elementwise loss tensor.
reduction (str): Options are "none", "mean" and "sum".
Return:
Tensor: Reduced loss tensor.
"""
reduction_enum = F._Reduction.get_enum(reduction)
if reduction_enum == 0:
return loss
elif reduction_enum == 1:
return loss.mean()
elif reduction_enum == 2:
return loss.sum()
def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
"""Apply element-wise weight and reduce loss.
Args:
loss (Tensor): Element-wise loss.
weight (Tensor): Element-wise weights.
reduction (str): Same as built-in losses of PyTorch.
avg_factor (float): Avarage factor when computing the mean of losses.
Returns:
Tensor: Processed loss values.
"""
if weight is not None:
assert weight.dim() == loss.dim()
if weight.dim() > 1:
assert weight.size(1) == 1 or weight.size(1) == loss.size(1)
loss = loss * weight
if avg_factor is None:
loss = reduce_loss(loss, reduction)
elif reduction == 'mean':
loss = loss.sum() / avg_factor
elif reduction != 'none':
raise ValueError('avg_factor can not be used with reduction="sum"')
return loss
def weighted_loss(loss_func):
"""Create a weighted version of a given loss function.
To use this decorator, the loss function must have the signature like
`loss_func(pred, target, **kwargs)`. The function only needs to compute
element-wise loss without any reduction. This decorator will add weight
and reduction arguments to the function. The decorated function will have
the signature like `loss_func(pred, target, weight=None, reduction='mean',
avg_factor=None, **kwargs)`.
:Example:
>>> import torch
>>> @weighted_loss
>>> def l1_loss(pred, target):
>>> return (pred - target).abs()
>>> pred = torch.Tensor([0, 2, 3])
>>> target = torch.Tensor([1, 1, 1])
>>> weight = torch.Tensor([1, 0, 1])
>>> l1_loss(pred, target)
tensor(1.3333)
>>> l1_loss(pred, target, weight)
tensor(1.)
>>> l1_loss(pred, target, reduction='none')
tensor([1., 1., 2.])
>>> l1_loss(pred, target, weight, avg_factor=2)
tensor(1.5000)
"""
@functools.wraps(loss_func)
def wrapper(pred, target, weight=None, reduction='mean', avg_factor=
None, **kwargs):
loss = loss_func(pred, target, **kwargs)
loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
return loss
return wrapper
@weighted_loss
def binary_dice_loss(pred, target, valid_mask, smooth=1, exponent=2, **kwards):
assert pred.shape[0] == target.shape[0]
pred = pred.reshape(pred.shape[0], -1)
target = target.reshape(target.shape[0], -1)
valid_mask = valid_mask.reshape(valid_mask.shape[0], -1)
num = torch.sum(torch.mul(pred, target) * valid_mask, dim=1) * 2 + smooth
den = torch.sum(pred.pow(exponent) + target.pow(exponent), dim=1) + smooth
return 1 - num / den
@weighted_loss
def dice_loss(pred, target, valid_mask, smooth=1, exponent=2, class_weight=
None, ignore_index=255):
assert pred.shape[0] == target.shape[0]
total_loss = 0
num_classes = pred.shape[1]
for i in range(num_classes):
if i != ignore_index:
dice_loss = binary_dice_loss(pred[:, i], target[..., i],
valid_mask=valid_mask, smooth=smooth, exponent=exponent)
if class_weight is not None:
dice_loss *= class_weight[i]
total_loss += dice_loss
return total_loss / num_classes
def get_class_weight(class_weight):
"""Get class weight for loss function.
Args:
class_weight (list[float] | str | None): If class_weight is a str,
take it as a file name and read from it.
"""
if isinstance(class_weight, str):
if class_weight.endswith('.npy'):
class_weight = np.load(class_weight)
else:
class_weight = mmcv.load(class_weight)
return class_weight
class DiceLoss(nn.Module):
"""DiceLoss.
This loss is proposed in `V-Net: Fully Convolutional Neural Networks for
Volumetric Medical Image Segmentation <https://arxiv.org/abs/1606.04797>`_.
Args:
loss_type (str, optional): Binary or multi-class loss.
Default: 'multi_class'. Options are "binary" and "multi_class".
smooth (float): A float number to smooth loss, and avoid NaN error.
Default: 1
exponent (float): An float number to calculate denominator
value: \\sum{x^exponent} + \\sum{y^exponent}. Default: 2.
reduction (str, optional): The method used to reduce the loss. Options
are "none", "mean" and "sum". This parameter only works when
per_image is True. Default: 'mean'.
class_weight (list[float] | str, optional): Weight of each class. If in
str format, read them from a file. Defaults to None.
loss_weight (float, optional): Weight of the loss. Default to 1.0.
ignore_index (int | None): The label index to be ignored. Default: 255.
"""
def __init__(self, smooth=1, exponent=2, reduction='mean', class_weight
=None, loss_weight=1.0, ignore_index=255, **kwards):
super(DiceLoss, self).__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.class_weight = get_class_weight(class_weight)
self.loss_weight = loss_weight
self.ignore_index = ignore_index
def forward(self, pred, target, avg_factor=None, reduction_override=
None, **kwards):
assert reduction_override in (None, 'none', 'mean', 'sum')
reduction = (reduction_override if reduction_override else self.
reduction)
if self.class_weight is not None:
class_weight = pred.new_tensor(self.class_weight)
else:
class_weight = None
pred = F.softmax(pred, dim=1)
num_classes = pred.shape[1]
one_hot_target = F.one_hot(torch.clamp(target.long(), 0,
num_classes - 1), num_classes=num_classes)
valid_mask = (target != self.ignore_index).long()
loss = self.loss_weight * dice_loss(pred, one_hot_target,
valid_mask=valid_mask, reduction=reduction, avg_factor=
avg_factor, smooth=self.smooth, exponent=self.exponent,
class_weight=class_weight, ignore_index=self.ignore_index)
return loss
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import functools
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2(
in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + 4 * r0, None, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * r0, None, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr1 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr1 + (2 + 4 * r0), None, eviction_policy='evict_last')
tmp42 = tl.load(in_ptr1 + (3 + 4 * r0), None, eviction_policy='evict_last')
tmp71 = tl.load(in_ptr0 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp112 = tl.load(in_ptr0 + (2 + 4 * r0), None, eviction_policy='evict_last'
)
tmp153 = tl.load(in_ptr0 + (3 + 4 * r0), None, eviction_policy='evict_last'
)
tmp2 = tmp1.to(tl.int64)
tmp3 = tl.full([1, 1], 0, tl.int64)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = tl.full([1, 1], 3, tl.int64)
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp6 == tmp3
tmp8 = tmp7.to(tl.int64)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp0 * tmp9
tmp11 = 255.0
tmp12 = tmp1 != tmp11
tmp13 = tmp12.to(tl.int64)
tmp14 = tmp13.to(tl.float32)
tmp15 = tmp10 * tmp14
tmp17 = tmp16.to(tl.int64)
tmp18 = triton_helpers.maximum(tmp17, tmp3)
tmp19 = triton_helpers.minimum(tmp18, tmp5)
tmp20 = tmp19 == tmp3
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp21.to(tl.float32)
tmp23 = tmp0 * tmp22
tmp24 = tmp16 != tmp11
tmp25 = tmp24.to(tl.int64)
tmp26 = tmp25.to(tl.float32)
tmp27 = tmp23 * tmp26
tmp28 = tmp15 + tmp27
tmp30 = tmp29.to(tl.int64)
tmp31 = triton_helpers.maximum(tmp30, tmp3)
tmp32 = triton_helpers.minimum(tmp31, tmp5)
tmp33 = tmp32 == tmp3
tmp34 = tmp33.to(tl.int64)
tmp35 = tmp34.to(tl.float32)
tmp36 = tmp0 * tmp35
tmp37 = tmp29 != tmp11
tmp38 = tmp37.to(tl.int64)
tmp39 = tmp38.to(tl.float32)
tmp40 = tmp36 * tmp39
tmp41 = tmp28 + tmp40
tmp43 = tmp42.to(tl.int64)
tmp44 = triton_helpers.maximum(tmp43, tmp3)
tmp45 = triton_helpers.minimum(tmp44, tmp5)
tmp46 = tmp45 == tmp3
tmp47 = tmp46.to(tl.int64)
tmp48 = tmp47.to(tl.float32)
tmp49 = tmp0 * tmp48
tmp50 = tmp42 != tmp11
tmp51 = tmp50.to(tl.int64)
tmp52 = tmp51.to(tl.float32)
tmp53 = tmp49 * tmp52
tmp54 = tmp41 + tmp53
tmp55 = tmp0 * tmp0
tmp56 = tmp8 * tmp8
tmp57 = tmp56.to(tl.float32)
tmp58 = tmp55 + tmp57
tmp59 = tmp21 * tmp21
tmp60 = tmp59.to(tl.float32)
tmp61 = tmp55 + tmp60
tmp62 = tmp58 + tmp61
tmp63 = tmp34 * tmp34
tmp64 = tmp63.to(tl.float32)
tmp65 = tmp55 + tmp64
tmp66 = tmp62 + tmp65
tmp67 = tmp47 * tmp47
tmp68 = tmp67.to(tl.float32)
tmp69 = tmp55 + tmp68
tmp70 = tmp66 + tmp69
tmp72 = tl.full([1, 1], 1, tl.int64)
tmp73 = tmp6 == tmp72
tmp74 = tmp73.to(tl.int64)
tmp75 = tmp74.to(tl.float32)
tmp76 = tmp71 * tmp75
tmp77 = tmp76 * tmp14
tmp78 = tmp19 == tmp72
tmp79 = tmp78.to(tl.int64)
tmp80 = tmp79.to(tl.float32)
tmp81 = tmp71 * tmp80
tmp82 = tmp81 * tmp26
tmp83 = tmp77 + tmp82
tmp84 = tmp32 == tmp72
tmp85 = tmp84.to(tl.int64)
tmp86 = tmp85.to(tl.float32)
tmp87 = tmp71 * tmp86
tmp88 = tmp87 * tmp39
tmp89 = tmp83 + tmp88
tmp90 = tmp45 == tmp72
tmp91 = tmp90.to(tl.int64)
tmp92 = tmp91.to(tl.float32)
tmp93 = tmp71 * tmp92
tmp94 = tmp93 * tmp52
tmp95 = tmp89 + tmp94
tmp96 = tmp71 * tmp71
tmp97 = tmp74 * tmp74
tmp98 = tmp97.to(tl.float32)
tmp99 = tmp96 + tmp98
tmp100 = tmp79 * tmp79
tmp101 = tmp100.to(tl.float32)
tmp102 = tmp96 + tmp101
tmp103 = tmp99 + tmp102
tmp104 = tmp85 * tmp85
tmp105 = tmp104.to(tl.float32)
tmp106 = tmp96 + tmp105
tmp107 = tmp103 + tmp106
tmp108 = tmp91 * tmp91
tmp109 = tmp108.to(tl.float32)
tmp110 = tmp96 + tmp109
tmp111 = tmp107 + tmp110
tmp113 = tl.full([1, 1], 2, tl.int64)
tmp114 = tmp6 == tmp113
tmp115 = tmp114.to(tl.int64)
tmp116 = tmp115.to(tl.float32)
tmp117 = tmp112 * tmp116
tmp118 = tmp117 * tmp14
tmp119 = tmp19 == tmp113
tmp120 = tmp119.to(tl.int64)
tmp121 = tmp120.to(tl.float32)
tmp122 = tmp112 * tmp121
tmp123 = tmp122 * tmp26
tmp124 = tmp118 + tmp123
tmp125 = tmp32 == tmp113
tmp126 = tmp125.to(tl.int64)
tmp127 = tmp126.to(tl.float32)
tmp128 = tmp112 * tmp127
tmp129 = tmp128 * tmp39
tmp130 = tmp124 + tmp129
tmp131 = tmp45 == tmp113
tmp132 = tmp131.to(tl.int64)
tmp133 = tmp132.to(tl.float32)
tmp134 = tmp112 * tmp133
tmp135 = tmp134 * tmp52
tmp136 = tmp130 + tmp135
tmp137 = tmp112 * tmp112
tmp138 = tmp115 * tmp115
tmp139 = tmp138.to(tl.float32)
tmp140 = tmp137 + tmp139
tmp141 = tmp120 * tmp120
tmp142 = tmp141.to(tl.float32)
tmp143 = tmp137 + tmp142
tmp144 = tmp140 + tmp143
tmp145 = tmp126 * tmp126
tmp146 = tmp145.to(tl.float32)
tmp147 = tmp137 + tmp146
tmp148 = tmp144 + tmp147
tmp149 = tmp132 * tmp132
tmp150 = tmp149.to(tl.float32)
tmp151 = tmp137 + tmp150
tmp152 = tmp148 + tmp151
tmp154 = tmp6 == tmp5
tmp155 = tmp154.to(tl.int64)
tmp156 = tmp155.to(tl.float32)
tmp157 = tmp153 * tmp156
tmp158 = tmp157 * tmp14
tmp159 = tmp19 == tmp5
tmp160 = tmp159.to(tl.int64)
tmp161 = tmp160.to(tl.float32)
tmp162 = tmp153 * tmp161
tmp163 = tmp162 * tmp26
tmp164 = tmp158 + tmp163
tmp165 = tmp32 == tmp5
tmp166 = tmp165.to(tl.int64)
tmp167 = tmp166.to(tl.float32)
tmp168 = tmp153 * tmp167
tmp169 = tmp168 * tmp39
tmp170 = tmp164 + tmp169
tmp171 = tmp45 == tmp5
tmp172 = tmp171.to(tl.int64)
tmp173 = tmp172.to(tl.float32)
tmp174 = tmp153 * tmp173
tmp175 = tmp174 * tmp52
tmp176 = tmp170 + tmp175
tmp177 = tmp153 * tmp153
tmp178 = tmp155 * tmp155
tmp179 = tmp178.to(tl.float32)
tmp180 = tmp177 + tmp179
tmp181 = tmp160 * tmp160
tmp182 = tmp181.to(tl.float32)
tmp183 = tmp177 + tmp182
tmp184 = tmp180 + tmp183
tmp185 = tmp166 * tmp166
tmp186 = tmp185.to(tl.float32)
tmp187 = tmp177 + tmp186
tmp188 = tmp184 + tmp187
tmp189 = tmp172 * tmp172
tmp190 = tmp189.to(tl.float32)
tmp191 = tmp177 + tmp190
tmp192 = tmp188 + tmp191
tmp193 = 2.0
tmp194 = tmp54 * tmp193
tmp195 = 1.0
tmp196 = tmp194 + tmp195
tmp197 = tmp70 + tmp195
tmp198 = tmp196 / tmp197
tmp199 = tmp195 - tmp198
tmp200 = tl.broadcast_to(tmp199, [XBLOCK, RBLOCK])
tmp202 = tl.sum(tmp200, 1)[:, None]
tmp203 = tmp95 * tmp193
tmp204 = tmp203 + tmp195
tmp205 = tmp111 + tmp195
tmp206 = tmp204 / tmp205
tmp207 = tmp195 - tmp206
tmp208 = tl.broadcast_to(tmp207, [XBLOCK, RBLOCK])
tmp210 = tl.sum(tmp208, 1)[:, None]
tmp211 = tmp136 * tmp193
tmp212 = tmp211 + tmp195
tmp213 = tmp152 + tmp195
tmp214 = tmp212 / tmp213
tmp215 = tmp195 - tmp214
tmp216 = tl.broadcast_to(tmp215, [XBLOCK, RBLOCK])
tmp218 = tl.sum(tmp216, 1)[:, None]
tmp219 = tmp176 * tmp193
tmp220 = tmp219 + tmp195
tmp221 = tmp192 + tmp195
tmp222 = tmp220 / tmp221
tmp223 = tmp195 - tmp222
tmp224 = tl.broadcast_to(tmp223, [XBLOCK, RBLOCK])
tmp226 = tl.sum(tmp224, 1)[:, None]
tmp227 = 4.0
tmp228 = tmp202 / tmp227
tmp229 = 0.0
tmp230 = tmp228 + tmp229
tmp231 = tmp210 / tmp227
tmp232 = tmp230 + tmp231
tmp233 = tmp218 / tmp227
tmp234 = tmp232 + tmp233
tmp235 = tmp226 / tmp227
tmp236 = tmp234 + tmp235
tmp237 = 0.25
tmp238 = tmp236 * tmp237
tmp239 = tmp238 / tmp195
tmp240 = tmp239 * tmp195
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp240, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4), (4, 1))
assert_size_stride(arg1_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(16)](arg0_1, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del arg0_1
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(16)](buf0, buf1, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del buf0
buf10 = empty_strided_cuda((), (), torch.float32)
buf14 = buf10
del buf10
triton_per_fused__to_copy_add_div_mean_mul_ne_pow_rsub_sum_view_2[grid
(1)](buf14, buf1, arg1_1, 1, 4, XBLOCK=1, num_warps=2, num_stages=1
)
del arg1_1
del buf1
return buf14,
def reduce_loss(loss, reduction):
"""Reduce loss as specified.
Args:
loss (Tensor): Elementwise loss tensor.
reduction (str): Options are "none", "mean" and "sum".
Return:
Tensor: Reduced loss tensor.
"""
reduction_enum = F._Reduction.get_enum(reduction)
if reduction_enum == 0:
return loss
elif reduction_enum == 1:
return loss.mean()
elif reduction_enum == 2:
return loss.sum()
def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
"""Apply element-wise weight and reduce loss.
Args:
loss (Tensor): Element-wise loss.
weight (Tensor): Element-wise weights.
reduction (str): Same as built-in losses of PyTorch.
avg_factor (float): Avarage factor when computing the mean of losses.
Returns:
Tensor: Processed loss values.
"""
if weight is not None:
assert weight.dim() == loss.dim()
if weight.dim() > 1:
assert weight.size(1) == 1 or weight.size(1) == loss.size(1)
loss = loss * weight
if avg_factor is None:
loss = reduce_loss(loss, reduction)
elif reduction == 'mean':
loss = loss.sum() / avg_factor
elif reduction != 'none':
raise ValueError('avg_factor can not be used with reduction="sum"')
return loss
def weighted_loss(loss_func):
"""Create a weighted version of a given loss function.
To use this decorator, the loss function must have the signature like
`loss_func(pred, target, **kwargs)`. The function only needs to compute
element-wise loss without any reduction. This decorator will add weight
and reduction arguments to the function. The decorated function will have
the signature like `loss_func(pred, target, weight=None, reduction='mean',
avg_factor=None, **kwargs)`.
:Example:
>>> import torch
>>> @weighted_loss
>>> def l1_loss(pred, target):
>>> return (pred - target).abs()
>>> pred = torch.Tensor([0, 2, 3])
>>> target = torch.Tensor([1, 1, 1])
>>> weight = torch.Tensor([1, 0, 1])
>>> l1_loss(pred, target)
tensor(1.3333)
>>> l1_loss(pred, target, weight)
tensor(1.)
>>> l1_loss(pred, target, reduction='none')
tensor([1., 1., 2.])
>>> l1_loss(pred, target, weight, avg_factor=2)
tensor(1.5000)
"""
@functools.wraps(loss_func)
def wrapper(pred, target, weight=None, reduction='mean', avg_factor=
None, **kwargs):
loss = loss_func(pred, target, **kwargs)
loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
return loss
return wrapper
@weighted_loss
def binary_dice_loss(pred, target, valid_mask, smooth=1, exponent=2, **kwards):
assert pred.shape[0] == target.shape[0]
pred = pred.reshape(pred.shape[0], -1)
target = target.reshape(target.shape[0], -1)
valid_mask = valid_mask.reshape(valid_mask.shape[0], -1)
num = torch.sum(torch.mul(pred, target) * valid_mask, dim=1) * 2 + smooth
den = torch.sum(pred.pow(exponent) + target.pow(exponent), dim=1) + smooth
return 1 - num / den
@weighted_loss
def dice_loss(pred, target, valid_mask, smooth=1, exponent=2, class_weight=
None, ignore_index=255):
assert pred.shape[0] == target.shape[0]
total_loss = 0
num_classes = pred.shape[1]
for i in range(num_classes):
if i != ignore_index:
dice_loss = binary_dice_loss(pred[:, i], target[..., i],
valid_mask=valid_mask, smooth=smooth, exponent=exponent)
if class_weight is not None:
dice_loss *= class_weight[i]
total_loss += dice_loss
return total_loss / num_classes
def get_class_weight(class_weight):
"""Get class weight for loss function.
Args:
class_weight (list[float] | str | None): If class_weight is a str,
take it as a file name and read from it.
"""
if isinstance(class_weight, str):
if class_weight.endswith('.npy'):
class_weight = np.load(class_weight)
else:
class_weight = mmcv.load(class_weight)
return class_weight
class DiceLossNew(nn.Module):
"""DiceLoss.
This loss is proposed in `V-Net: Fully Convolutional Neural Networks for
Volumetric Medical Image Segmentation <https://arxiv.org/abs/1606.04797>`_.
Args:
loss_type (str, optional): Binary or multi-class loss.
Default: 'multi_class'. Options are "binary" and "multi_class".
smooth (float): A float number to smooth loss, and avoid NaN error.
Default: 1
exponent (float): An float number to calculate denominator
value: \\sum{x^exponent} + \\sum{y^exponent}. Default: 2.
reduction (str, optional): The method used to reduce the loss. Options
are "none", "mean" and "sum". This parameter only works when
per_image is True. Default: 'mean'.
class_weight (list[float] | str, optional): Weight of each class. If in
str format, read them from a file. Defaults to None.
loss_weight (float, optional): Weight of the loss. Default to 1.0.
ignore_index (int | None): The label index to be ignored. Default: 255.
"""
def __init__(self, smooth=1, exponent=2, reduction='mean', class_weight
=None, loss_weight=1.0, ignore_index=255, **kwards):
super(DiceLossNew, self).__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.class_weight = get_class_weight(class_weight)
self.loss_weight = loss_weight
self.ignore_index = ignore_index
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| SeHwanJoo/mmsegmentation_body | DiceLoss | false | 1,041 | [
"Apache-2.0"
] | 0 | 31c4bf27c3dc0a84bfbb06a0c017c5908c17f0ac | https://github.com/SeHwanJoo/mmsegmentation_body/tree/31c4bf27c3dc0a84bfbb06a0c017c5908c17f0ac | import functools
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
def reduce_loss(loss, reduction):
"""Reduce loss as specified.
Args:
loss (Tensor): Elementwise loss tensor.
reduction (str): Options are "none", "mean" and "sum".
Return:
Tensor: Reduced loss tensor.
"""
reduction_enum = F._Reduction.get_enum(reduction)
if reduction_enum == 0:
return loss
elif reduction_enum == 1:
return loss.mean()
elif reduction_enum == 2:
return loss.sum()
def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
"""Apply element-wise weight and reduce loss.
Args:
loss (Tensor): Element-wise loss.
weight (Tensor): Element-wise weights.
reduction (str): Same as built-in losses of PyTorch.
avg_factor (float): Avarage factor when computing the mean of losses.
Returns:
Tensor: Processed loss values.
"""
if weight is not None:
assert weight.dim() == loss.dim()
if weight.dim() > 1:
assert weight.size(1) == 1 or weight.size(1) == loss.size(1)
loss = loss * weight
if avg_factor is None:
loss = reduce_loss(loss, reduction)
elif reduction == 'mean':
loss = loss.sum() / avg_factor
elif reduction != 'none':
raise ValueError('avg_factor can not be used with reduction="sum"')
return loss
def weighted_loss(loss_func):
"""Create a weighted version of a given loss function.
To use this decorator, the loss function must have the signature like
`loss_func(pred, target, **kwargs)`. The function only needs to compute
element-wise loss without any reduction. This decorator will add weight
and reduction arguments to the function. The decorated function will have
the signature like `loss_func(pred, target, weight=None, reduction='mean',
avg_factor=None, **kwargs)`.
:Example:
>>> import torch
>>> @weighted_loss
>>> def l1_loss(pred, target):
>>> return (pred - target).abs()
>>> pred = torch.Tensor([0, 2, 3])
>>> target = torch.Tensor([1, 1, 1])
>>> weight = torch.Tensor([1, 0, 1])
>>> l1_loss(pred, target)
tensor(1.3333)
>>> l1_loss(pred, target, weight)
tensor(1.)
>>> l1_loss(pred, target, reduction='none')
tensor([1., 1., 2.])
>>> l1_loss(pred, target, weight, avg_factor=2)
tensor(1.5000)
"""
@functools.wraps(loss_func)
def wrapper(pred, target, weight=None, reduction='mean', avg_factor=
None, **kwargs):
loss = loss_func(pred, target, **kwargs)
loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
return loss
return wrapper
@weighted_loss
def binary_dice_loss(pred, target, valid_mask, smooth=1, exponent=2, **kwards):
assert pred.shape[0] == target.shape[0]
pred = pred.reshape(pred.shape[0], -1)
target = target.reshape(target.shape[0], -1)
valid_mask = valid_mask.reshape(valid_mask.shape[0], -1)
num = torch.sum(torch.mul(pred, target) * valid_mask, dim=1) * 2 + smooth
den = torch.sum(pred.pow(exponent) + target.pow(exponent), dim=1) + smooth
return 1 - num / den
@weighted_loss
def dice_loss(pred, target, valid_mask, smooth=1, exponent=2, class_weight=
None, ignore_index=255):
assert pred.shape[0] == target.shape[0]
total_loss = 0
num_classes = pred.shape[1]
for i in range(num_classes):
if i != ignore_index:
dice_loss = binary_dice_loss(pred[:, i], target[..., i],
valid_mask=valid_mask, smooth=smooth, exponent=exponent)
if class_weight is not None:
dice_loss *= class_weight[i]
total_loss += dice_loss
return total_loss / num_classes
def get_class_weight(class_weight):
"""Get class weight for loss function.
Args:
class_weight (list[float] | str
# ... truncated (>4000 chars) for memory efficiency |
VAE | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/3q/c3qwr2d2rrpjzvnddomnmdy6cwva4hjlvrn2y5epemk4ak3k2m6c.py
# Topologically Sorted Source Nodes: [h1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# h1 => relu
# Graph fragment:
# %add_tensor_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_2, %primals_3), kwargs = {})
# %relu : [num_users=3] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_2,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2048],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 400
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/a6/ca6vhpnsixur65k3b6hxegp4job3ylimpyatml46dzhhhkxeihd5.py
# Topologically Sorted Source Nodes: [mul, std, mul_1, z], Original ATen: [aten.mul, aten.exp, aten.add]
# Source node to ATen node mapping:
# mul => mul
# mul_1 => mul_1
# std => exp
# z => add
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%addmm_2, 0.5), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%mul,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%randn, %exp), kwargs = {})
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %addmm_1), kwargs = {})
triton_poi_fused_add_exp_mul_1 = async_compile.triton('triton_poi_fused_add_exp_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_exp_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_exp_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 80
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask)
tmp6 = tl.load(in_ptr2 + (x0), xmask)
tmp2 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp0 * tmp4
tmp7 = tmp5 + tmp6
tl.store(out_ptr0 + (x0), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/hb/chbjjrtszu6f3bhry7ireqcm3ie3twpz5s7g7owb3zuauqhiqcby.py
# Topologically Sorted Source Nodes: [sigmoid], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# sigmoid => sigmoid
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_11), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_sigmoid_2 = async_compile.triton('triton_poi_fused_sigmoid_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 3136
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 784
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 784), (784, 1))
assert_size_stride(primals_2, (400, 784), (784, 1))
assert_size_stride(primals_3, (400, ), (1, ))
assert_size_stride(primals_4, (20, 400), (400, 1))
assert_size_stride(primals_5, (20, ), (1, ))
assert_size_stride(primals_6, (20, 400), (400, 1))
assert_size_stride(primals_7, (20, ), (1, ))
assert_size_stride(primals_8, (400, 20), (20, 1))
assert_size_stride(primals_9, (400, ), (1, ))
assert_size_stride(primals_10, (784, 400), (400, 1))
assert_size_stride(primals_11, (784, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784, 400), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [h1], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_3, 1600, grid=grid(1600), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((4, 20), (20, 1), torch.float32)
# Topologically Sorted Source Nodes: [mu], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, buf1, reinterpret_tensor(primals_4, (400, 20), (1, 400), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 20), (20, 1), torch.float32)
# Topologically Sorted Source Nodes: [logvar], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf1, reinterpret_tensor(primals_6, (400, 20), (1, 400), 0), alpha=1, beta=1, out=buf3)
del primals_7
# Topologically Sorted Source Nodes: [eps], Original ATen: [aten.randn_like]
buf4 = torch.ops.aten.randn.default([4, 20], dtype=torch.float32, device=device(type='cuda', index=0), pin_memory=False)
buf5 = buf4
del buf4
buf6 = empty_strided_cuda((4, 20), (20, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, std, mul_1, z], Original ATen: [aten.mul, aten.exp, aten.add]
triton_poi_fused_add_exp_mul_1.run(buf5, buf3, buf2, buf6, 80, grid=grid(80), stream=stream0)
buf7 = empty_strided_cuda((4, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf6, reinterpret_tensor(primals_8, (20, 400), (1, 20), 0), out=buf7)
buf8 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [h3], Original ATen: [aten.relu]
triton_poi_fused_relu_0.run(buf8, primals_9, 1600, grid=grid(1600), stream=stream0)
del primals_9
buf9 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf8, reinterpret_tensor(primals_10, (400, 784), (1, 400), 0), out=buf9)
buf10 = buf9; del buf9 # reuse
# Topologically Sorted Source Nodes: [sigmoid], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_2.run(buf10, primals_11, 3136, grid=grid(3136), stream=stream0)
del primals_11
return (buf10, buf2, buf3, primals_1, buf1, buf3, buf5, buf6, buf8, buf10, primals_10, primals_8, primals_6, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((400, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((20, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((20, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((20, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((20, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((400, 20), (20, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((784, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((784, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.onnx
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.autograd
class VAE(nn.Module):
def __init__(self):
super(VAE, self).__init__()
self.fc1 = nn.Linear(784, 400)
self.fc21 = nn.Linear(400, 20)
self.fc22 = nn.Linear(400, 20)
self.fc3 = nn.Linear(20, 400)
self.fc4 = nn.Linear(400, 784)
def encode(self, x):
h1 = F.relu(self.fc1(x))
return self.fc21(h1), self.fc22(h1)
def reparameterize(self, mu, logvar):
std = torch.exp(0.5 * logvar)
eps = torch.randn_like(std)
return eps.mul(std).add_(mu)
def decode(self, z):
h3 = F.relu(self.fc3(z))
return torch.sigmoid(self.fc4(h3))
def forward(self, x):
mu, logvar = self.encode(x.view(-1, 784))
z = self.reparameterize(mu, logvar)
return self.decode(z), mu, logvar
def get_inputs():
return [torch.rand([4, 784])]
def get_init_inputs():
return [[], {}]
| import torch
from torch import device
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.nn.functional as F
import torch.onnx
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.autograd
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 1600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 400
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_exp_mul_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 80
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask)
tmp6 = tl.load(in_ptr2 + x0, xmask)
tmp2 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp0 * tmp4
tmp7 = tmp5 + tmp6
tl.store(out_ptr0 + x0, tmp7, xmask)
@triton.jit
def triton_poi_fused_sigmoid_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 3136
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 784
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 784), (784, 1))
assert_size_stride(primals_2, (400, 784), (784, 1))
assert_size_stride(primals_3, (400,), (1,))
assert_size_stride(primals_4, (20, 400), (400, 1))
assert_size_stride(primals_5, (20,), (1,))
assert_size_stride(primals_6, (20, 400), (400, 1))
assert_size_stride(primals_7, (20,), (1,))
assert_size_stride(primals_8, (400, 20), (20, 1))
assert_size_stride(primals_9, (400,), (1,))
assert_size_stride(primals_10, (784, 400), (400, 1))
assert_size_stride(primals_11, (784,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 400), (400, 1), torch.float32)
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784,
400), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(1600)](buf1, primals_3, 1600, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((4, 20), (20, 1), torch.float32)
extern_kernels.addmm(primals_5, buf1, reinterpret_tensor(primals_4,
(400, 20), (1, 400), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 20), (20, 1), torch.float32)
extern_kernels.addmm(primals_7, buf1, reinterpret_tensor(primals_6,
(400, 20), (1, 400), 0), alpha=1, beta=1, out=buf3)
del primals_7
buf4 = torch.ops.aten.randn.default([4, 20], dtype=torch.float32,
device=device(type='cuda', index=0), pin_memory=False)
buf5 = buf4
del buf4
buf6 = empty_strided_cuda((4, 20), (20, 1), torch.float32)
triton_poi_fused_add_exp_mul_1[grid(80)](buf5, buf3, buf2, buf6, 80,
XBLOCK=128, num_warps=4, num_stages=1)
buf7 = empty_strided_cuda((4, 400), (400, 1), torch.float32)
extern_kernels.mm(buf6, reinterpret_tensor(primals_8, (20, 400), (1,
20), 0), out=buf7)
buf8 = buf7
del buf7
triton_poi_fused_relu_0[grid(1600)](buf8, primals_9, 1600, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_9
buf9 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
extern_kernels.mm(buf8, reinterpret_tensor(primals_10, (400, 784),
(1, 400), 0), out=buf9)
buf10 = buf9
del buf9
triton_poi_fused_sigmoid_2[grid(3136)](buf10, primals_11, 3136,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_11
return (buf10, buf2, buf3, primals_1, buf1, buf3, buf5, buf6, buf8,
buf10, primals_10, primals_8, primals_6, primals_4)
class VAENew(nn.Module):
def __init__(self):
super(VAENew, self).__init__()
self.fc1 = nn.Linear(784, 400)
self.fc21 = nn.Linear(400, 20)
self.fc22 = nn.Linear(400, 20)
self.fc3 = nn.Linear(20, 400)
self.fc4 = nn.Linear(400, 784)
def encode(self, x):
h1 = F.relu(self.fc1(x))
return self.fc21(h1), self.fc22(h1)
def reparameterize(self, mu, logvar):
std = torch.exp(0.5 * logvar)
eps = torch.randn_like(std)
return eps.mul(std).add_(mu)
def decode(self, z):
h3 = F.relu(self.fc3(z))
return torch.sigmoid(self.fc4(h3))
def forward(self, input_0):
primals_2 = self.fc1.weight
primals_3 = self.fc1.bias
primals_4 = self.fc21.weight
primals_5 = self.fc21.bias
primals_6 = self.fc22.weight
primals_7 = self.fc22.bias
primals_8 = self.fc3.weight
primals_9 = self.fc3.bias
primals_10 = self.fc4.weight
primals_11 = self.fc4.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0], output[1], output[2]
| ScorpioDoctor/antares02 | VAE | false | 1,042 | [
"BSD-3-Clause"
] | 0 | 631b817d2e98f351d1173b620d15c4a5efed11da | https://github.com/ScorpioDoctor/antares02/tree/631b817d2e98f351d1173b620d15c4a5efed11da | import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.onnx
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torch.autograd
class Model(nn.Module):
def __init__(self):
super().__init__()
self.fc1 = nn.Linear(784, 400)
self.fc21 = nn.Linear(400, 20)
self.fc22 = nn.Linear(400, 20)
self.fc3 = nn.Linear(20, 400)
self.fc4 = nn.Linear(400, 784)
def encode(self, x):
h1 = F.relu(self.fc1(x))
return self.fc21(h1), self.fc22(h1)
def reparameterize(self, mu, logvar):
std = torch.exp(0.5 * logvar)
eps = torch.randn_like(std)
return eps.mul(std).add_(mu)
def decode(self, z):
h3 = F.relu(self.fc3(z))
return torch.sigmoid(self.fc4(h3))
def forward(self, x):
mu, logvar = self.encode(x.view(-1, 784))
z = self.reparameterize(mu, logvar)
return self.decode(z), mu, logvar
def get_inputs():
return [torch.rand([4, 784])]
def get_init_inputs():
return []
|
MultiHeadAttention | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/yj/cyjay5w6yqtzjhwrra5u36ugjzcr3epsjgaero2ugqpqx7xjycai.py
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention_score_2 => exp, sum_1
# Graph fragment:
# %mul_tensor : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_11, 1), kwargs = {})
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%mul_tensor, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_tensor, %amax_default), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_tensor, 1.0), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%div_tensor,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
triton_poi_fused__softmax_0 = async_compile.triton('triton_poi_fused__softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp6 = tmp0 * tmp5
tmp7 = tmp6 * tmp3
tmp8 = triton_helpers.maximum(tmp4, tmp7)
tmp10 = tmp0 * tmp9
tmp11 = tmp10 * tmp3
tmp12 = triton_helpers.maximum(tmp8, tmp11)
tmp14 = tmp0 * tmp13
tmp15 = tmp14 * tmp3
tmp16 = triton_helpers.maximum(tmp12, tmp15)
tmp17 = tmp4 - tmp16
tmp18 = tmp17 * tmp3
tmp19 = tl_math.exp(tmp18)
tmp20 = tmp7 - tmp16
tmp21 = tmp20 * tmp3
tmp22 = tl_math.exp(tmp21)
tmp23 = tmp19 + tmp22
tmp24 = tmp11 - tmp16
tmp25 = tmp24 * tmp3
tmp26 = tl_math.exp(tmp25)
tmp27 = tmp23 + tmp26
tmp28 = tmp15 - tmp16
tmp29 = tmp28 * tmp3
tmp30 = tl_math.exp(tmp29)
tmp31 = tmp27 + tmp30
tl.store(out_ptr0 + (x2), tmp16, xmask)
tl.store(out_ptr1 + (x2), tmp31, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/eq/ceqzjquocsdhf6v3jex4eilxuabityntyouuqz2e5hg6cpglxepm.py
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention_score_2 => div_1, exp
# Graph fragment:
# %mul_tensor : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_11, 1), kwargs = {})
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%mul_tensor, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_tensor, %amax_default), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_tensor, 1.0), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%div_tensor,), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = (xindex // 4)
x0 = xindex % 4
x2 = (xindex // 16)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x3), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr2 + (x3), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr3 + (x3), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp6 = tmp4 - tmp5
tmp7 = tmp6 * tmp3
tmp8 = tl_math.exp(tmp7)
tmp10 = tmp8 / tmp9
tl.store(out_ptr0 + (x4), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_query], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_key], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_6, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_value], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_8, reinterpret_tensor(primals_9, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_7
del primals_8
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(buf0, buf1, buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf0, buf1, buf3, buf4, buf5, 256, grid=grid(256), stream=stream0)
buf6 = reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [result], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf5, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0), out=buf6)
buf7 = reinterpret_tensor(buf3, (16, 4), (4, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [output_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_11, reinterpret_tensor(buf6, (16, 4), (4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf7)
del primals_11
return (reinterpret_tensor(buf7, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), buf0, reinterpret_tensor(primals_6, (16, 4), (4, 1), 0), buf1, reinterpret_tensor(primals_9, (16, 4), (4, 1), 0), buf5, reinterpret_tensor(buf6, (16, 4), (4, 1), 0), primals_10, reinterpret_tensor(buf2, (16, 1, 4), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super(MultiHeadAttention, self).__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [[], {'model_dim': 4, 'key_dim': 4, 'value_dim': 4, 'num_head': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp6 = tmp0 * tmp5
tmp7 = tmp6 * tmp3
tmp8 = triton_helpers.maximum(tmp4, tmp7)
tmp10 = tmp0 * tmp9
tmp11 = tmp10 * tmp3
tmp12 = triton_helpers.maximum(tmp8, tmp11)
tmp14 = tmp0 * tmp13
tmp15 = tmp14 * tmp3
tmp16 = triton_helpers.maximum(tmp12, tmp15)
tmp17 = tmp4 - tmp16
tmp18 = tmp17 * tmp3
tmp19 = tl_math.exp(tmp18)
tmp20 = tmp7 - tmp16
tmp21 = tmp20 * tmp3
tmp22 = tl_math.exp(tmp21)
tmp23 = tmp19 + tmp22
tmp24 = tmp11 - tmp16
tmp25 = tmp24 * tmp3
tmp26 = tl_math.exp(tmp25)
tmp27 = tmp23 + tmp26
tmp28 = tmp15 - tmp16
tmp29 = tmp28 * tmp3
tmp30 = tl_math.exp(tmp29)
tmp31 = tmp27 + tmp30
tl.store(out_ptr0 + x2, tmp16, xmask)
tl.store(out_ptr1 + x2, tmp31, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex // 4
x0 = xindex % 4
x2 = xindex // 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + x3, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp5 = tl.load(in_ptr2 + x3, xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr3 + x3, xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tmp6 = tmp4 - tmp5
tmp7 = tmp6 * tmp3
tmp8 = tl_math.exp(tmp7)
tmp10 = tmp8 / tmp9
tl.store(out_ptr0 + x4, tmp10, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (16,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_6, (16,
4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_8, reinterpret_tensor(primals_9, (16,
4), (4, 1), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf2)
del primals_7
del primals_8
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(64)](buf0, buf1, buf3, buf4, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf0, buf1, buf3, buf4, buf5,
256, XBLOCK=256, num_warps=4, num_stages=1)
buf6 = reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 1), 0)
del buf4
extern_kernels.bmm(reinterpret_tensor(buf5, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0), out=buf6)
buf7 = reinterpret_tensor(buf3, (16, 4), (4, 1), 0)
del buf3
extern_kernels.addmm(primals_11, reinterpret_tensor(buf6, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf7)
del primals_11
return reinterpret_tensor(buf7, (4, 4, 4), (16, 4, 1), 0
), reinterpret_tensor(primals_3, (16, 4), (4, 1), 0
), buf0, reinterpret_tensor(primals_6, (16, 4), (4, 1), 0
), buf1, reinterpret_tensor(primals_9, (16, 4), (4, 1), 0
), buf5, reinterpret_tensor(buf6, (16, 4), (4, 1), 0
), primals_10, reinterpret_tensor(buf2, (16, 1, 4), (4, 1, 1), 0)
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttentionNew(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super(MultiHeadAttentionNew, self).__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
def forward(self, input_0, input_1, input_2):
primals_1 = self.Wq.weight
primals_2 = self.Wq.bias
primals_4 = self.Wk.weight
primals_5 = self.Wk.bias
primals_7 = self.Wv.weight
primals_8 = self.Wv.bias
primals_10 = self.Wo.weight
primals_11 = self.Wo.bias
primals_3 = input_0
primals_6 = input_1
primals_9 = input_2
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
| SeungoneKim/Transformer_implementation | MultiHeadAttention | false | 1,043 | [
"Apache-2.0"
] | 0 | a52bf552eb645fc9bfb812cc26842fc147d6c008 | https://github.com/SeungoneKim/Transformer_implementation/tree/a52bf552eb645fc9bfb812cc26842fc147d6c008 | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super().__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class Model(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super().__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [4, 4, 4, 4]
|
Encoding | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/yx/cyx5u6kg47bcb2a4mvrlkw5ynd42h4mj76hk2j6tveptehbkmic4.py
# Topologically Sorted Source Nodes: [sub, pow_1, sum_1, scaled_l2_norm], Original ATen: [aten.sub, aten.pow, aten.sum, aten.mul]
# Source node to ATen node mapping:
# pow_1 => pow_1
# scaled_l2_norm => mul
# sub => sub
# sum_1 => sum_1
# Graph fragment:
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%expand, %view_2), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [3]), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %sum_1), kwargs = {})
triton_poi_fused_mul_pow_sub_sum_0 = async_compile.triton('triton_poi_fused_mul_pow_sub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_pow_sub_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_pow_sub_sum_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4) % 16
x2 = (xindex // 64)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x1 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (4*x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (16 + x1 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr2 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr1 + (32 + x1 + (64*x2)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr1 + (48 + x1 + (64*x2)), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr2 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 - tmp2
tmp4 = tmp3 * tmp3
tmp7 = tmp5 - tmp6
tmp8 = tmp7 * tmp7
tmp9 = tmp4 + tmp8
tmp12 = tmp10 - tmp11
tmp13 = tmp12 * tmp12
tmp14 = tmp9 + tmp13
tmp17 = tmp15 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp14 + tmp18
tmp20 = tmp0 * tmp19
tl.store(out_ptr0 + (x4), tmp20, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/hz/chz2sqsqk26mwhf2dxhgh44jfpu2er5yqjftwkzfav5ctqtx5e7f.py
# Topologically Sorted Source Nodes: [assignment_weights], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# assignment_weights => amax, exp, sub_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%mul, [2], True), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_1,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [assignment_weights], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# assignment_weights => div, sum_2
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [2], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_2), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/36/c36bunrpokrzs5svt4e6kwmfyyitmjh7nivu5rc6sidx55znumrf.py
# Topologically Sorted Source Nodes: [sub, mul_1, encoded_feat], Original ATen: [aten.sub, aten.mul, aten.sum]
# Source node to ATen node mapping:
# encoded_feat => sum_3
# mul_1 => mul_1
# sub => sub
# Graph fragment:
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%expand, %view_2), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%unsqueeze_2, %sub), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_1, [1]), kwargs = {})
triton_per_fused_mul_sub_sum_3 = async_compile.triton('triton_per_fused_mul_sub_sum_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[64, 16],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mul_sub_sum_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mul_sub_sum_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r3 = rindex
x1 = (xindex // 4) % 4
x2 = (xindex // 16)
x0 = xindex % 4
x4 = xindex % 16
x5 = xindex
tmp0 = tl.load(in_ptr0 + (x1 + (4*r3) + (64*x2)), xmask, eviction_policy='evict_last', other=0.0)
tmp1 = tl.load(in_ptr1 + (r3 + (16*x0) + (64*x2)), xmask, eviction_policy='evict_last', other=0.0)
tmp2 = tl.load(in_ptr2 + (x4), xmask, eviction_policy='evict_last')
tmp3 = tmp1 - tmp2
tmp4 = tmp0 * tmp3
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp7 = tl.where(xmask, tmp5, 0)
tmp8 = tl.sum(tmp7, 1)[:, None]
tl.store(out_ptr0 + (x5), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sub, pow_1, sum_1, scaled_l2_norm], Original ATen: [aten.sub, aten.pow, aten.sum, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_pow_sub_sum_0.run(primals_3, primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
buf1 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [assignment_weights], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [assignment_weights], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf1, buf2, 256, grid=grid(256), stream=stream0)
del buf1
buf3 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sub, mul_1, encoded_feat], Original ATen: [aten.sub, aten.mul, aten.sum]
triton_per_fused_mul_sub_sum_3.run(buf2, primals_1, primals_2, buf3, 64, 16, grid=grid(64), stream=stream0)
del buf2
return (buf3, primals_1, primals_2, primals_3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
class Encoding(nn.Module):
"""Encoding Layer: a learnable residual encoder.
Input is of shape (batch_size, channels, height, width).
Output is of shape (batch_size, num_codes, channels).
Args:
channels: dimension of the features or feature channels
num_codes: number of code words
"""
def __init__(self, channels, num_codes):
super(Encoding, self).__init__()
self.channels, self.num_codes = channels, num_codes
std = 1.0 / (num_codes * channels) ** 0.5
self.codewords = nn.Parameter(torch.empty(num_codes, channels,
dtype=torch.float).uniform_(-std, std), requires_grad=True)
self.scale = nn.Parameter(torch.empty(num_codes, dtype=torch.float)
.uniform_(-1, 0), requires_grad=True)
@staticmethod
def scaled_l2(x, codewords, scale):
num_codes, channels = codewords.size()
batch_size = x.size(0)
reshaped_scale = scale.view((1, 1, num_codes))
expanded_x = x.unsqueeze(2).expand((batch_size, x.size(1),
num_codes, channels))
reshaped_codewords = codewords.view((1, 1, num_codes, channels))
scaled_l2_norm = reshaped_scale * (expanded_x - reshaped_codewords
).pow(2).sum(dim=3)
return scaled_l2_norm
@staticmethod
def aggregate(assignment_weights, x, codewords):
num_codes, channels = codewords.size()
reshaped_codewords = codewords.view((1, 1, num_codes, channels))
batch_size = x.size(0)
expanded_x = x.unsqueeze(2).expand((batch_size, x.size(1),
num_codes, channels))
encoded_feat = (assignment_weights.unsqueeze(3) * (expanded_x -
reshaped_codewords)).sum(dim=1)
return encoded_feat
def forward(self, x):
assert x.dim() == 4 and x.size(1) == self.channels
batch_size = x.size(0)
x = x.view(batch_size, self.channels, -1).transpose(1, 2).contiguous()
assignment_weights = F.softmax(self.scaled_l2(x, self.codewords,
self.scale), dim=2)
encoded_feat = self.aggregate(assignment_weights, x, self.codewords)
return encoded_feat
def __repr__(self):
repr_str = self.__class__.__name__
repr_str += (
f'(Nx{self.channels}xHxW =>Nx{self.num_codes}x{self.channels})')
return repr_str
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'channels': 4, 'num_codes': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch._C
import torch.serialization
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_pow_sub_sum_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4 % 16
x2 = xindex // 64
x4 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x1 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr2 + 4 * x0, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (16 + x1 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr2 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr1 + (32 + x1 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp11 = tl.load(in_ptr2 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp15 = tl.load(in_ptr1 + (48 + x1 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp16 = tl.load(in_ptr2 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp3 = tmp1 - tmp2
tmp4 = tmp3 * tmp3
tmp7 = tmp5 - tmp6
tmp8 = tmp7 * tmp7
tmp9 = tmp4 + tmp8
tmp12 = tmp10 - tmp11
tmp13 = tmp12 * tmp12
tmp14 = tmp9 + tmp13
tmp17 = tmp15 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp14 + tmp18
tmp20 = tmp0 * tmp19
tl.store(out_ptr0 + x4, tmp20, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_per_fused_mul_sub_sum_3(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 64
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r3 = rindex
x1 = xindex // 4 % 4
x2 = xindex // 16
x0 = xindex % 4
x4 = xindex % 16
x5 = xindex
tmp0 = tl.load(in_ptr0 + (x1 + 4 * r3 + 64 * x2), xmask,
eviction_policy='evict_last', other=0.0)
tmp1 = tl.load(in_ptr1 + (r3 + 16 * x0 + 64 * x2), xmask,
eviction_policy='evict_last', other=0.0)
tmp2 = tl.load(in_ptr2 + x4, xmask, eviction_policy='evict_last')
tmp3 = tmp1 - tmp2
tmp4 = tmp0 * tmp3
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp7 = tl.where(xmask, tmp5, 0)
tmp8 = tl.sum(tmp7, 1)[:, None]
tl.store(out_ptr0 + x5, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_pow_sub_sum_0[grid(256)](primals_3, primals_1,
primals_2, buf0, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf1 = empty_strided_cuda((4, 16, 4), (64, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf0, buf1, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf2 = buf0
del buf0
triton_poi_fused__softmax_2[grid(256)](buf1, buf2, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf1
buf3 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_per_fused_mul_sub_sum_3[grid(64)](buf2, primals_1, primals_2,
buf3, 64, 16, XBLOCK=32, num_warps=4, num_stages=1)
del buf2
return buf3, primals_1, primals_2, primals_3
class EncodingNew(nn.Module):
"""Encoding Layer: a learnable residual encoder.
Input is of shape (batch_size, channels, height, width).
Output is of shape (batch_size, num_codes, channels).
Args:
channels: dimension of the features or feature channels
num_codes: number of code words
"""
def __init__(self, channels, num_codes):
super(EncodingNew, self).__init__()
self.channels, self.num_codes = channels, num_codes
std = 1.0 / (num_codes * channels) ** 0.5
self.codewords = nn.Parameter(torch.empty(num_codes, channels,
dtype=torch.float).uniform_(-std, std), requires_grad=True)
self.scale = nn.Parameter(torch.empty(num_codes, dtype=torch.float)
.uniform_(-1, 0), requires_grad=True)
@staticmethod
def scaled_l2(x, codewords, scale):
num_codes, channels = codewords.size()
batch_size = x.size(0)
reshaped_scale = scale.view((1, 1, num_codes))
expanded_x = x.unsqueeze(2).expand((batch_size, x.size(1),
num_codes, channels))
reshaped_codewords = codewords.view((1, 1, num_codes, channels))
scaled_l2_norm = reshaped_scale * (expanded_x - reshaped_codewords
).pow(2).sum(dim=3)
return scaled_l2_norm
@staticmethod
def aggregate(assignment_weights, x, codewords):
num_codes, channels = codewords.size()
reshaped_codewords = codewords.view((1, 1, num_codes, channels))
batch_size = x.size(0)
expanded_x = x.unsqueeze(2).expand((batch_size, x.size(1),
num_codes, channels))
encoded_feat = (assignment_weights.unsqueeze(3) * (expanded_x -
reshaped_codewords)).sum(dim=1)
return encoded_feat
def __repr__(self):
repr_str = self.__class__.__name__
repr_str += (
f'(Nx{self.channels}xHxW =>Nx{self.num_codes}x{self.channels})')
return repr_str
def forward(self, input_0):
primals_2 = self.codewords
primals_3 = self.scale
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| SeHwanJoo/mmsegmentation_body | Encoding | false | 1,044 | [
"Apache-2.0"
] | 0 | 31c4bf27c3dc0a84bfbb06a0c017c5908c17f0ac | https://github.com/SeHwanJoo/mmsegmentation_body/tree/31c4bf27c3dc0a84bfbb06a0c017c5908c17f0ac | import torch
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
class Model(nn.Module):
"""Encoding Layer: a learnable residual encoder.
Input is of shape (batch_size, channels, height, width).
Output is of shape (batch_size, num_codes, channels).
Args:
channels: dimension of the features or feature channels
num_codes: number of code words
"""
def __init__(self, channels, num_codes):
super().__init__()
self.channels, self.num_codes = channels, num_codes
std = 1.0 / (num_codes * channels) ** 0.5
self.codewords = nn.Parameter(torch.empty(num_codes, channels,
dtype=torch.float).uniform_(-std, std), requires_grad=True)
self.scale = nn.Parameter(torch.empty(num_codes, dtype=torch.float)
.uniform_(-1, 0), requires_grad=True)
@staticmethod
def scaled_l2(x, codewords, scale):
num_codes, channels = codewords.size()
batch_size = x.size(0)
reshaped_scale = scale.view((1, 1, num_codes))
expanded_x = x.unsqueeze(2).expand((batch_size, x.size(1),
num_codes, channels))
reshaped_codewords = codewords.view((1, 1, num_codes, channels))
scaled_l2_norm = reshaped_scale * (expanded_x - reshaped_codewords
).pow(2).sum(dim=3)
return scaled_l2_norm
@staticmethod
def aggregate(assignment_weights, x, codewords):
num_codes, channels = codewords.size()
reshaped_codewords = codewords.view((1, 1, num_codes, channels))
batch_size = x.size(0)
expanded_x = x.unsqueeze(2).expand((batch_size, x.size(1),
num_codes, channels))
encoded_feat = (assignment_weights.unsqueeze(3) * (expanded_x -
reshaped_codewords)).sum(dim=1)
return encoded_feat
def forward(self, x):
assert x.dim() == 4 and x.size(1) == self.channels
batch_size = x.size(0)
x = x.view(batch_size, self.channels, -1).transpose(1, 2).contiguous()
assignment_weights = F.softmax(self.scaled_l2(x, self.codewords,
self.scale), dim=2)
encoded_feat = self.aggregate(assignment_weights, x, self.codewords)
return encoded_feat
def __repr__(self):
repr_str = self.__class__.__name__
repr_str += (
f'(Nx{self.channels}xHxW =>Nx{self.num_codes}x{self.channels})')
return repr_str
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
CnnNet | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/t4/ct4qtnuxylso56lddzt74wqzxctat7bgtw43ghj6wwn25j2c7msg.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_0 = async_compile.triton('triton_poi_fused_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2048, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 2048
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 32
y1 = (yindex // 32)
tmp0 = tl.load(in_ptr0 + (x2 + (4*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (32*x2) + (128*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/gk/cgk5p6nacclp2ot6bn6nbnbresf5zlczkwvdxtsnsyi5hbqd3xsi.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_1 = async_compile.triton('triton_poi_fused_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 8192
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = (yindex // 64)
tmp0 = tl.load(in_ptr0 + (x2 + (4*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (64*x2) + (256*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/uc/cuconme7l5q53n3sxefuxolugwrfgw262laltqajqmscann6mybi.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x => convolution
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128, 16384], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 128
xnumel = 8836
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 32
y1 = (yindex // 32)
tmp0 = tl.load(in_ptr0 + (x2 + (8836*y3)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (y0 + (32*x2) + (282752*y1)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ld/cldjotsifmpevio42aswbawvkexbmusj2bxnqz7d4jaevrxn5vht.py
# Topologically Sorted Source Nodes: [max_pool2d, x_1], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
# Source node to ATen node mapping:
# max_pool2d => _low_memory_max_pool2d_with_offsets, getitem_1
# x_1 => relu
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%convolution, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%getitem,), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_relu_3 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_relu_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[524288],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_relu_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_3(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 282752
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 32
x1 = (xindex // 32) % 47
x2 = (xindex // 1504)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1) + (6016*x2)), xmask)
tmp1 = tl.load(in_ptr0 + (32 + x0 + (64*x1) + (6016*x2)), xmask)
tmp7 = tl.load(in_ptr0 + (3008 + x0 + (64*x1) + (6016*x2)), xmask)
tmp12 = tl.load(in_ptr0 + (3040 + x0 + (64*x1) + (6016*x2)), xmask)
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + (x3), tmp15, xmask)
tl.store(out_ptr1 + (x3), tmp18, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/dj/cdjdw64ewlwdxlw4dclphhticnt624zm6q6wavwuwwofbgbk4ply.py
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_2 => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_4 = async_compile.triton('triton_poi_fused_convolution_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1048576],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 541696
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/b6/cb6puipugnxx2nscy7ndskuler7ba6tm3lmyvm6pr62rqflvor2h.py
# Topologically Sorted Source Nodes: [max_pool2d_1, x_3], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
# Source node to ATen node mapping:
# max_pool2d_1 => _low_memory_max_pool2d_with_offsets_1, getitem_3
# x_3 => relu_1
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets_1 : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%convolution_1, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_3 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_1, 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%getitem_2,), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_relu_5 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_relu_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[262144],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_relu_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_5(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 135424
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 64
x1 = (xindex // 64) % 23
x2 = (xindex // 1472)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (128*x1) + (5888*x2)), xmask)
tmp1 = tl.load(in_ptr0 + (64 + x0 + (128*x1) + (5888*x2)), xmask)
tmp7 = tl.load(in_ptr0 + (2944 + x0 + (128*x1) + (5888*x2)), xmask)
tmp12 = tl.load(in_ptr0 + (3008 + x0 + (128*x1) + (5888*x2)), xmask)
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + (x3), tmp15, xmask)
tl.store(out_ptr1 + (x3), tmp18, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/vm/cvmtij7djyjxpxvgju7slqbaipji5f5nz3ocixx4tzb3zoqbrr3j.py
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_4 => convolution_2
# Graph fragment:
# %convolution_2 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_1, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_6 = async_compile.triton('triton_poi_fused_convolution_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[262144],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_6', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_6(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 247808
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x2), tmp2, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/a2/ca2xazqvkvyqidyiiiw4ixs4ax4ahwzigfwnnkafigdyruyulcac.py
# Topologically Sorted Source Nodes: [max_pool2d_2, x_5], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
# Source node to ATen node mapping:
# max_pool2d_2 => _low_memory_max_pool2d_with_offsets_2, getitem_5
# x_5 => relu_2
# Graph fragment:
# %_low_memory_max_pool2d_with_offsets_2 : [num_users=2] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%convolution_2, [2, 2], [2, 2], [0, 0], [1, 1], False), kwargs = {})
# %getitem_5 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_2, 1), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%getitem_4,), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_relu_7 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_relu_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512, 128], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i8', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_relu_7', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_7(in_ptr0, out_ptr0, out_ptr1, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 484
xnumel = 128
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 11
y1 = (yindex // 11)
y5 = yindex
y4 = (yindex // 121)
y6 = yindex % 121
tmp0 = tl.load(in_ptr0 + (x2 + (256*y0) + (5632*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (128 + x2 + (256*y0) + (5632*y1)), xmask & ymask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2816 + x2 + (256*y0) + (5632*y1)), xmask & ymask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2944 + x2 + (256*y0) + (5632*y1)), xmask & ymask, eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1, 1], 1, tl.int8)
tmp4 = tl.full([1, 1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1, 1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1, 1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1, 1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + (x2 + (128*y5)), tmp15, xmask & ymask)
tl.store(out_ptr1 + (y6 + (121*x2) + (15488*y4)), tmp18, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/bk/cbkmobcq4xugig5i4hltumomgc5bjkhuql4r3rz23mkiqfo33z4t.py
# Topologically Sorted Source Nodes: [x_7], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_7 => relu_3
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_9), kwargs = {})
# %relu_3 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
triton_poi_fused_relu_8 = async_compile.triton('triton_poi_fused_relu_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2048],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_8', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_8(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 2000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 500
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/y3/cy3sy4qwyqyz63qaftf2sjogpitcadukpnborr6vso5qvhgtdvue.py
# Topologically Sorted Source Nodes: [], Original ATen: [aten.threshold_backward]
# Source node to ATen node mapping:
# Graph fragment:
# %le_2 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_2, 0), kwargs = {})
triton_poi_fused_threshold_backward_9 = async_compile.triton('triton_poi_fused_threshold_backward_9', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512, 128], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i1', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_threshold_backward_9', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_threshold_backward_9(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 512
xnumel = 121
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = (yindex // 128)
tmp0 = tl.load(in_ptr0 + (x2 + (121*y3)), xmask & ymask, eviction_policy='evict_last')
tmp1 = 0.0
tmp2 = tmp0 <= tmp1
tl.store(out_ptr0 + (y0 + (128*x2) + (15488*y1)), tmp2, xmask & ymask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13 = args
args.clear()
assert_size_stride(primals_1, (32, 1, 3, 3), (9, 9, 3, 1))
assert_size_stride(primals_2, (32, ), (1, ))
assert_size_stride(primals_3, (4, 1, 96, 96), (9216, 9216, 96, 1))
assert_size_stride(primals_4, (64, 32, 2, 2), (128, 4, 2, 1))
assert_size_stride(primals_5, (64, ), (1, ))
assert_size_stride(primals_6, (128, 64, 2, 2), (256, 4, 2, 1))
assert_size_stride(primals_7, (128, ), (1, ))
assert_size_stride(primals_8, (500, 15488), (15488, 1))
assert_size_stride(primals_9, (500, ), (1, ))
assert_size_stride(primals_10, (500, 500), (500, 1))
assert_size_stride(primals_11, (500, ), (1, ))
assert_size_stride(primals_12, (30, 500), (500, 1))
assert_size_stride(primals_13, (30, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 32, 2, 2), (128, 1, 64, 32), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
stream0 = get_raw_stream(0)
triton_poi_fused_0.run(primals_4, buf0, 2048, 4, grid=grid(2048, 4), stream=stream0)
del primals_4
buf1 = empty_strided_cuda((128, 64, 2, 2), (256, 1, 128, 64), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_1.run(primals_6, buf1, 8192, 4, grid=grid(8192, 4), stream=stream0)
del primals_6
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 32, 94, 94), (282752, 8836, 94, 1))
buf3 = empty_strided_cuda((4, 32, 94, 94), (282752, 1, 3008, 32), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf2, primals_2, buf3, 128, 8836, grid=grid(128, 8836), stream=stream0)
del buf2
del primals_2
buf4 = empty_strided_cuda((4, 32, 47, 47), (70688, 1, 1504, 32), torch.int8)
buf5 = empty_strided_cuda((4, 32, 47, 47), (70688, 1, 1504, 32), torch.float32)
# Topologically Sorted Source Nodes: [max_pool2d, x_1], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
triton_poi_fused_max_pool2d_with_indices_relu_3.run(buf3, buf4, buf5, 282752, grid=grid(282752), stream=stream0)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, buf0, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 64, 46, 46), (135424, 1, 2944, 64))
buf7 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_4.run(buf7, primals_5, 541696, grid=grid(541696), stream=stream0)
del primals_5
buf8 = empty_strided_cuda((4, 64, 23, 23), (33856, 1, 1472, 64), torch.int8)
buf9 = empty_strided_cuda((4, 64, 23, 23), (33856, 1, 1472, 64), torch.float32)
# Topologically Sorted Source Nodes: [max_pool2d_1, x_3], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
triton_poi_fused_max_pool2d_with_indices_relu_5.run(buf7, buf8, buf9, 135424, grid=grid(135424), stream=stream0)
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.convolution]
buf10 = extern_kernels.convolution(buf9, buf1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf10, (4, 128, 22, 22), (61952, 1, 2816, 128))
buf11 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.convolution]
triton_poi_fused_convolution_6.run(buf11, primals_7, 247808, grid=grid(247808), stream=stream0)
del primals_7
buf12 = empty_strided_cuda((4, 128, 11, 11), (15488, 1, 1408, 128), torch.int8)
buf13 = empty_strided_cuda((4, 128, 11, 11), (15488, 121, 11, 1), torch.float32)
# Topologically Sorted Source Nodes: [max_pool2d_2, x_5], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
triton_poi_fused_max_pool2d_with_indices_relu_7.run(buf11, buf12, buf13, 484, 128, grid=grid(484, 128), stream=stream0)
buf14 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf13, (4, 15488), (15488, 1), 0), reinterpret_tensor(primals_8, (15488, 500), (1, 15488), 0), out=buf14)
buf15 = buf14; del buf14 # reuse
# Topologically Sorted Source Nodes: [x_7], Original ATen: [aten.relu]
triton_poi_fused_relu_8.run(buf15, primals_9, 2000, grid=grid(2000), stream=stream0)
del primals_9
buf16 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf15, reinterpret_tensor(primals_10, (500, 500), (1, 500), 0), out=buf16)
buf17 = buf16; del buf16 # reuse
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.relu]
triton_poi_fused_relu_8.run(buf17, primals_11, 2000, grid=grid(2000), stream=stream0)
del primals_11
buf18 = empty_strided_cuda((4, 30), (30, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_9], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_13, buf17, reinterpret_tensor(primals_12, (500, 30), (1, 500), 0), alpha=1, beta=1, out=buf18)
del primals_13
buf19 = empty_strided_cuda((4, 128, 11, 11), (15488, 1, 1408, 128), torch.bool)
# Topologically Sorted Source Nodes: [], Original ATen: [aten.threshold_backward]
triton_poi_fused_threshold_backward_9.run(buf13, buf19, 512, 121, grid=grid(512, 121), stream=stream0)
return (buf18, primals_1, primals_3, buf0, buf1, buf3, buf4, buf5, buf7, buf8, buf9, buf11, buf12, reinterpret_tensor(buf13, (4, 15488), (15488, 1), 0), buf15, buf17, primals_12, primals_10, primals_8, buf19, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((32, 1, 3, 3), (9, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((32, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 1, 96, 96), (9216, 9216, 96, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((64, 32, 2, 2), (128, 4, 2, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((128, 64, 2, 2), (256, 4, 2, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((500, 15488), (15488, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((500, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((500, 500), (500, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((500, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((30, 500), (500, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((30, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
class CnnNet(nn.Module):
def __init__(self):
super(CnnNet, self).__init__()
self.conv1 = nn.Conv2d(1, 32, 3)
self.pool1 = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(32, 64, 2)
self.pool2 = nn.MaxPool2d(2, 2)
self.conv3 = nn.Conv2d(64, 128, 2)
self.pool3 = nn.MaxPool2d(2, 2)
self.fc1 = nn.Linear(15488, 500)
self.fc2 = nn.Linear(500, 500)
self.fc3 = nn.Linear(500, 30)
def forward(self, x):
x = self.conv1(x)
x = F.relu(self.pool1(x))
x = self.conv2(x)
x = F.relu(self.pool2(x))
x = self.conv3(x)
x = F.relu(self.pool3(x))
x = x.view(-1, 15488)
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
def get_inputs():
return [torch.rand([4, 1, 96, 96])]
def get_init_inputs():
return [[], {}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 32
y1 = yindex // 32
tmp0 = tl.load(in_ptr0 + (x2 + 4 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 32 * x2 + 128 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = yindex // 64
tmp0 = tl.load(in_ptr0 + (x2 + 4 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 64 * x2 + 256 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_ptr0, in_ptr1, out_ptr0, ynumel,
xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 128
xnumel = 8836
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 32
y1 = yindex // 32
tmp0 = tl.load(in_ptr0 + (x2 + 8836 * y3), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (y0 + 32 * x2 + 282752 * y1), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_3(in_ptr0, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 282752
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 32
x1 = xindex // 32 % 47
x2 = xindex // 1504
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1 + 6016 * x2), xmask)
tmp1 = tl.load(in_ptr0 + (32 + x0 + 64 * x1 + 6016 * x2), xmask)
tmp7 = tl.load(in_ptr0 + (3008 + x0 + 64 * x1 + 6016 * x2), xmask)
tmp12 = tl.load(in_ptr0 + (3040 + x0 + 64 * x1 + 6016 * x2), xmask)
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + x3, tmp15, xmask)
tl.store(out_ptr1 + x3, tmp18, xmask)
@triton.jit
def triton_poi_fused_convolution_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 541696
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x2, tmp2, xmask)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_5(in_ptr0, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 135424
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 64
x1 = xindex // 64 % 23
x2 = xindex // 1472
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 128 * x1 + 5888 * x2), xmask)
tmp1 = tl.load(in_ptr0 + (64 + x0 + 128 * x1 + 5888 * x2), xmask)
tmp7 = tl.load(in_ptr0 + (2944 + x0 + 128 * x1 + 5888 * x2), xmask)
tmp12 = tl.load(in_ptr0 + (3008 + x0 + 128 * x1 + 5888 * x2), xmask)
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1], 1, tl.int8)
tmp4 = tl.full([1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + x3, tmp15, xmask)
tl.store(out_ptr1 + x3, tmp18, xmask)
@triton.jit
def triton_poi_fused_convolution_6(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x2, tmp2, None)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_relu_7(in_ptr0, out_ptr0,
out_ptr1, ynumel, xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 484
xnumel = 128
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 11
y1 = yindex // 11
y5 = yindex
y4 = yindex // 121
y6 = yindex % 121
tmp0 = tl.load(in_ptr0 + (x2 + 256 * y0 + 5632 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (128 + x2 + 256 * y0 + 5632 * y1), xmask &
ymask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2816 + x2 + 256 * y0 + 5632 * y1), xmask &
ymask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2944 + x2 + 256 * y0 + 5632 * y1), xmask &
ymask, eviction_policy='evict_last')
tmp2 = tmp1 > tmp0
tmp3 = tl.full([1, 1], 1, tl.int8)
tmp4 = tl.full([1, 1], 0, tl.int8)
tmp5 = tl.where(tmp2, tmp3, tmp4)
tmp6 = triton_helpers.maximum(tmp1, tmp0)
tmp8 = tmp7 > tmp6
tmp9 = tl.full([1, 1], 2, tl.int8)
tmp10 = tl.where(tmp8, tmp9, tmp5)
tmp11 = triton_helpers.maximum(tmp7, tmp6)
tmp13 = tmp12 > tmp11
tmp14 = tl.full([1, 1], 3, tl.int8)
tmp15 = tl.where(tmp13, tmp14, tmp10)
tmp16 = triton_helpers.maximum(tmp12, tmp11)
tmp17 = tl.full([1, 1], 0, tl.int32)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tl.store(out_ptr0 + (x2 + 128 * y5), tmp15, xmask & ymask)
tl.store(out_ptr1 + (y6 + 121 * x2 + 15488 * y4), tmp18, xmask & ymask)
@triton.jit
def triton_poi_fused_relu_8(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 2000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 500
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_threshold_backward_9(in_ptr0, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 512
xnumel = 121
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = yindex // 128
tmp0 = tl.load(in_ptr0 + (x2 + 121 * y3), xmask & ymask,
eviction_policy='evict_last')
tmp1 = 0.0
tmp2 = tmp0 <= tmp1
tl.store(out_ptr0 + (y0 + 128 * x2 + 15488 * y1), tmp2, xmask & ymask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13) = args
args.clear()
assert_size_stride(primals_1, (32, 1, 3, 3), (9, 9, 3, 1))
assert_size_stride(primals_2, (32,), (1,))
assert_size_stride(primals_3, (4, 1, 96, 96), (9216, 9216, 96, 1))
assert_size_stride(primals_4, (64, 32, 2, 2), (128, 4, 2, 1))
assert_size_stride(primals_5, (64,), (1,))
assert_size_stride(primals_6, (128, 64, 2, 2), (256, 4, 2, 1))
assert_size_stride(primals_7, (128,), (1,))
assert_size_stride(primals_8, (500, 15488), (15488, 1))
assert_size_stride(primals_9, (500,), (1,))
assert_size_stride(primals_10, (500, 500), (500, 1))
assert_size_stride(primals_11, (500,), (1,))
assert_size_stride(primals_12, (30, 500), (500, 1))
assert_size_stride(primals_13, (30,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 32, 2, 2), (128, 1, 64, 32), torch.
float32)
get_raw_stream(0)
triton_poi_fused_0[grid(2048, 4)](primals_4, buf0, 2048, 4, XBLOCK=
4, YBLOCK=256, num_warps=4, num_stages=1)
del primals_4
buf1 = empty_strided_cuda((128, 64, 2, 2), (256, 1, 128, 64), torch
.float32)
triton_poi_fused_1[grid(8192, 4)](primals_6, buf1, 8192, 4, XBLOCK=
4, YBLOCK=256, num_warps=4, num_stages=1)
del primals_6
buf2 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 32, 94, 94), (282752, 8836, 94, 1))
buf3 = empty_strided_cuda((4, 32, 94, 94), (282752, 1, 3008, 32),
torch.float32)
triton_poi_fused_convolution_2[grid(128, 8836)](buf2, primals_2,
buf3, 128, 8836, XBLOCK=32, YBLOCK=32, num_warps=4, num_stages=1)
del buf2
del primals_2
buf4 = empty_strided_cuda((4, 32, 47, 47), (70688, 1, 1504, 32),
torch.int8)
buf5 = empty_strided_cuda((4, 32, 47, 47), (70688, 1, 1504, 32),
torch.float32)
triton_poi_fused_max_pool2d_with_indices_relu_3[grid(282752)](buf3,
buf4, buf5, 282752, XBLOCK=512, num_warps=8, num_stages=1)
buf6 = extern_kernels.convolution(buf5, buf0, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 64, 46, 46), (135424, 1, 2944, 64))
buf7 = buf6
del buf6
triton_poi_fused_convolution_4[grid(541696)](buf7, primals_5,
541696, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_5
buf8 = empty_strided_cuda((4, 64, 23, 23), (33856, 1, 1472, 64),
torch.int8)
buf9 = empty_strided_cuda((4, 64, 23, 23), (33856, 1, 1472, 64),
torch.float32)
triton_poi_fused_max_pool2d_with_indices_relu_5[grid(135424)](buf7,
buf8, buf9, 135424, XBLOCK=1024, num_warps=4, num_stages=1)
buf10 = extern_kernels.convolution(buf9, buf1, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf10, (4, 128, 22, 22), (61952, 1, 2816, 128))
buf11 = buf10
del buf10
triton_poi_fused_convolution_6[grid(247808)](buf11, primals_7,
247808, XBLOCK=512, num_warps=8, num_stages=1)
del primals_7
buf12 = empty_strided_cuda((4, 128, 11, 11), (15488, 1, 1408, 128),
torch.int8)
buf13 = empty_strided_cuda((4, 128, 11, 11), (15488, 121, 11, 1),
torch.float32)
triton_poi_fused_max_pool2d_with_indices_relu_7[grid(484, 128)](buf11,
buf12, buf13, 484, 128, XBLOCK=128, YBLOCK=4, num_warps=4,
num_stages=1)
buf14 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf13, (4, 15488), (15488, 1),
0), reinterpret_tensor(primals_8, (15488, 500), (1, 15488), 0),
out=buf14)
buf15 = buf14
del buf14
triton_poi_fused_relu_8[grid(2000)](buf15, primals_9, 2000, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_9
buf16 = empty_strided_cuda((4, 500), (500, 1), torch.float32)
extern_kernels.mm(buf15, reinterpret_tensor(primals_10, (500, 500),
(1, 500), 0), out=buf16)
buf17 = buf16
del buf16
triton_poi_fused_relu_8[grid(2000)](buf17, primals_11, 2000, XBLOCK
=256, num_warps=4, num_stages=1)
del primals_11
buf18 = empty_strided_cuda((4, 30), (30, 1), torch.float32)
extern_kernels.addmm(primals_13, buf17, reinterpret_tensor(
primals_12, (500, 30), (1, 500), 0), alpha=1, beta=1, out=buf18)
del primals_13
buf19 = empty_strided_cuda((4, 128, 11, 11), (15488, 1, 1408, 128),
torch.bool)
triton_poi_fused_threshold_backward_9[grid(512, 121)](buf13, buf19,
512, 121, XBLOCK=32, YBLOCK=32, num_warps=4, num_stages=1)
return (buf18, primals_1, primals_3, buf0, buf1, buf3, buf4, buf5, buf7,
buf8, buf9, buf11, buf12, reinterpret_tensor(buf13, (4, 15488), (
15488, 1), 0), buf15, buf17, primals_12, primals_10, primals_8, buf19)
class CnnNetNew(nn.Module):
def __init__(self):
super(CnnNetNew, self).__init__()
self.conv1 = nn.Conv2d(1, 32, 3)
self.pool1 = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(32, 64, 2)
self.pool2 = nn.MaxPool2d(2, 2)
self.conv3 = nn.Conv2d(64, 128, 2)
self.pool3 = nn.MaxPool2d(2, 2)
self.fc1 = nn.Linear(15488, 500)
self.fc2 = nn.Linear(500, 500)
self.fc3 = nn.Linear(500, 30)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_6 = self.conv3.weight
primals_7 = self.conv3.bias
primals_8 = self.fc1.weight
primals_9 = self.fc1.bias
primals_10 = self.fc2.weight
primals_11 = self.fc2.bias
primals_12 = self.fc3.weight
primals_13 = self.fc3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13])
return output[0]
| RoyHirsch/DeepLearningCourse | CnnNet | false | 1,045 | [
"MIT"
] | 0 | 9036c0fdbb08b610524d7be991f8e4b490a82c6c | https://github.com/RoyHirsch/DeepLearningCourse/tree/9036c0fdbb08b610524d7be991f8e4b490a82c6c | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super().__init__()
self.conv1 = nn.Conv2d(1, 32, 3)
self.pool1 = nn.MaxPool2d(2, 2)
self.conv2 = nn.Conv2d(32, 64, 2)
self.pool2 = nn.MaxPool2d(2, 2)
self.conv3 = nn.Conv2d(64, 128, 2)
self.pool3 = nn.MaxPool2d(2, 2)
self.fc1 = nn.Linear(15488, 500)
self.fc2 = nn.Linear(500, 500)
self.fc3 = nn.Linear(500, 30)
def forward(self, x):
x = self.conv1(x)
x = F.relu(self.pool1(x))
x = self.conv2(x)
x = F.relu(self.pool2(x))
x = self.conv3(x)
x = F.relu(self.pool3(x))
x = x.view(-1, 15488)
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
def get_inputs():
return [torch.rand([4, 1, 96, 96])]
def get_init_inputs():
return []
|
BILM | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/xq/cxq2qmrgyky7dqwgqwftj6uhbj5vdl2rg7fomifs252vwbdicjhh.py
# Topologically Sorted Source Nodes: [pos_sig, pos_sig_1, neg_sig, neg_sig_1, sum_sig, x], Original ATen: [aten.sigmoid, aten.max_pool2d_with_indices, aten.mul, aten.add]
# Source node to ATen node mapping:
# neg_sig => mul
# neg_sig_1 => _low_memory_max_pool2d_with_offsets_1
# pos_sig => sigmoid
# pos_sig_1 => _low_memory_max_pool2d_with_offsets
# sum_sig => add
# x => mul_1
# Graph fragment:
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%arg0_1,), kwargs = {})
# %_low_memory_max_pool2d_with_offsets : [num_users=1] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%sigmoid, [3, 3], [1, 1], [1, 1], [1, 1], False), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, -1), kwargs = {})
# %_low_memory_max_pool2d_with_offsets_1 : [num_users=1] = call_function[target=torch.ops.prims._low_memory_max_pool2d_with_offsets.default](args = (%mul, [3, 3], [1, 1], [1, 1], [1, 1], False), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, %getitem_2), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %add), kwargs = {})
triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0 = async_compile.triton('triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4) % 4
x0 = xindex % 4
x3 = xindex
tmp115 = tl.load(in_ptr0 + (x3), xmask)
tmp0 = (-1) + x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tmp2 & tmp4
tmp6 = (-1) + x0
tmp7 = tmp6 >= tmp1
tmp8 = tmp6 < tmp3
tmp9 = tmp7 & tmp8
tmp10 = tmp5 & tmp9
tmp11 = tl.load(in_ptr0 + ((-5) + x3), tmp10 & xmask, other=0.0)
tmp12 = tl.sigmoid(tmp11)
tmp13 = tl.full(tmp12.shape, float("-inf"), tmp12.dtype)
tmp14 = tl.where(tmp10, tmp12, tmp13)
tmp15 = x0
tmp16 = tmp15 >= tmp1
tmp17 = tmp15 < tmp3
tmp18 = tmp16 & tmp17
tmp19 = tmp5 & tmp18
tmp20 = tl.load(in_ptr0 + ((-4) + x3), tmp19 & xmask, other=0.0)
tmp21 = tl.sigmoid(tmp20)
tmp22 = tl.full(tmp21.shape, float("-inf"), tmp21.dtype)
tmp23 = tl.where(tmp19, tmp21, tmp22)
tmp24 = triton_helpers.maximum(tmp23, tmp14)
tmp25 = 1 + x0
tmp26 = tmp25 >= tmp1
tmp27 = tmp25 < tmp3
tmp28 = tmp26 & tmp27
tmp29 = tmp5 & tmp28
tmp30 = tl.load(in_ptr0 + ((-3) + x3), tmp29 & xmask, other=0.0)
tmp31 = tl.sigmoid(tmp30)
tmp32 = tl.full(tmp31.shape, float("-inf"), tmp31.dtype)
tmp33 = tl.where(tmp29, tmp31, tmp32)
tmp34 = triton_helpers.maximum(tmp33, tmp24)
tmp35 = x1
tmp36 = tmp35 >= tmp1
tmp37 = tmp35 < tmp3
tmp38 = tmp36 & tmp37
tmp39 = tmp38 & tmp9
tmp40 = tl.load(in_ptr0 + ((-1) + x3), tmp39 & xmask, other=0.0)
tmp41 = tl.sigmoid(tmp40)
tmp42 = tl.full(tmp41.shape, float("-inf"), tmp41.dtype)
tmp43 = tl.where(tmp39, tmp41, tmp42)
tmp44 = triton_helpers.maximum(tmp43, tmp34)
tmp45 = tmp38 & tmp18
tmp46 = tl.load(in_ptr0 + (x3), tmp45 & xmask, other=0.0)
tmp47 = tl.sigmoid(tmp46)
tmp48 = tl.full(tmp47.shape, float("-inf"), tmp47.dtype)
tmp49 = tl.where(tmp45, tmp47, tmp48)
tmp50 = triton_helpers.maximum(tmp49, tmp44)
tmp51 = tmp38 & tmp28
tmp52 = tl.load(in_ptr0 + (1 + x3), tmp51 & xmask, other=0.0)
tmp53 = tl.sigmoid(tmp52)
tmp54 = tl.full(tmp53.shape, float("-inf"), tmp53.dtype)
tmp55 = tl.where(tmp51, tmp53, tmp54)
tmp56 = triton_helpers.maximum(tmp55, tmp50)
tmp57 = 1 + x1
tmp58 = tmp57 >= tmp1
tmp59 = tmp57 < tmp3
tmp60 = tmp58 & tmp59
tmp61 = tmp60 & tmp9
tmp62 = tl.load(in_ptr0 + (3 + x3), tmp61 & xmask, other=0.0)
tmp63 = tl.sigmoid(tmp62)
tmp64 = tl.full(tmp63.shape, float("-inf"), tmp63.dtype)
tmp65 = tl.where(tmp61, tmp63, tmp64)
tmp66 = triton_helpers.maximum(tmp65, tmp56)
tmp67 = tmp60 & tmp18
tmp68 = tl.load(in_ptr0 + (4 + x3), tmp67 & xmask, other=0.0)
tmp69 = tl.sigmoid(tmp68)
tmp70 = tl.full(tmp69.shape, float("-inf"), tmp69.dtype)
tmp71 = tl.where(tmp67, tmp69, tmp70)
tmp72 = triton_helpers.maximum(tmp71, tmp66)
tmp73 = tmp60 & tmp28
tmp74 = tl.load(in_ptr0 + (5 + x3), tmp73 & xmask, other=0.0)
tmp75 = tl.sigmoid(tmp74)
tmp76 = tl.full(tmp75.shape, float("-inf"), tmp75.dtype)
tmp77 = tl.where(tmp73, tmp75, tmp76)
tmp78 = triton_helpers.maximum(tmp77, tmp72)
tmp79 = -1.0
tmp80 = tmp12 * tmp79
tmp81 = tl.full(tmp80.shape, float("-inf"), tmp80.dtype)
tmp82 = tl.where(tmp10, tmp80, tmp81)
tmp83 = tmp21 * tmp79
tmp84 = tl.full(tmp83.shape, float("-inf"), tmp83.dtype)
tmp85 = tl.where(tmp19, tmp83, tmp84)
tmp86 = triton_helpers.maximum(tmp85, tmp82)
tmp87 = tmp31 * tmp79
tmp88 = tl.full(tmp87.shape, float("-inf"), tmp87.dtype)
tmp89 = tl.where(tmp29, tmp87, tmp88)
tmp90 = triton_helpers.maximum(tmp89, tmp86)
tmp91 = tmp41 * tmp79
tmp92 = tl.full(tmp91.shape, float("-inf"), tmp91.dtype)
tmp93 = tl.where(tmp39, tmp91, tmp92)
tmp94 = triton_helpers.maximum(tmp93, tmp90)
tmp95 = tmp47 * tmp79
tmp96 = tl.full(tmp95.shape, float("-inf"), tmp95.dtype)
tmp97 = tl.where(tmp45, tmp95, tmp96)
tmp98 = triton_helpers.maximum(tmp97, tmp94)
tmp99 = tmp53 * tmp79
tmp100 = tl.full(tmp99.shape, float("-inf"), tmp99.dtype)
tmp101 = tl.where(tmp51, tmp99, tmp100)
tmp102 = triton_helpers.maximum(tmp101, tmp98)
tmp103 = tmp63 * tmp79
tmp104 = tl.full(tmp103.shape, float("-inf"), tmp103.dtype)
tmp105 = tl.where(tmp61, tmp103, tmp104)
tmp106 = triton_helpers.maximum(tmp105, tmp102)
tmp107 = tmp69 * tmp79
tmp108 = tl.full(tmp107.shape, float("-inf"), tmp107.dtype)
tmp109 = tl.where(tmp67, tmp107, tmp108)
tmp110 = triton_helpers.maximum(tmp109, tmp106)
tmp111 = tmp75 * tmp79
tmp112 = tl.full(tmp111.shape, float("-inf"), tmp111.dtype)
tmp113 = tl.where(tmp73, tmp111, tmp112)
tmp114 = triton_helpers.maximum(tmp113, tmp110)
tmp116 = tmp78 + tmp114
tmp117 = tmp115 * tmp116
tl.store(in_out_ptr0 + (x3), tmp117, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [pos_sig, pos_sig_1, neg_sig, neg_sig_1, sum_sig, x], Original ATen: [aten.sigmoid, aten.max_pool2d_with_indices, aten.mul, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0.run(buf2, arg0_1, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class BILM(nn.Module):
def __init__(self):
super(BILM, self).__init__()
self.maxpool1 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
self.maxpool2 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
def forward(self, feat):
pos_sig = torch.sigmoid(feat)
neg_sig = -1 * pos_sig
pos_sig = self.maxpool1(pos_sig)
neg_sig = self.maxpool2(neg_sig)
sum_sig = pos_sig + neg_sig
x = feat * sum_sig
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0(in_out_ptr0,
in_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4 % 4
x0 = xindex % 4
x3 = xindex
tmp115 = tl.load(in_ptr0 + x3, xmask)
tmp0 = -1 + x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tmp2 & tmp4
tmp6 = -1 + x0
tmp7 = tmp6 >= tmp1
tmp8 = tmp6 < tmp3
tmp9 = tmp7 & tmp8
tmp10 = tmp5 & tmp9
tmp11 = tl.load(in_ptr0 + (-5 + x3), tmp10 & xmask, other=0.0)
tmp12 = tl.sigmoid(tmp11)
tmp13 = tl.full(tmp12.shape, float('-inf'), tmp12.dtype)
tmp14 = tl.where(tmp10, tmp12, tmp13)
tmp15 = x0
tmp16 = tmp15 >= tmp1
tmp17 = tmp15 < tmp3
tmp18 = tmp16 & tmp17
tmp19 = tmp5 & tmp18
tmp20 = tl.load(in_ptr0 + (-4 + x3), tmp19 & xmask, other=0.0)
tmp21 = tl.sigmoid(tmp20)
tmp22 = tl.full(tmp21.shape, float('-inf'), tmp21.dtype)
tmp23 = tl.where(tmp19, tmp21, tmp22)
tmp24 = triton_helpers.maximum(tmp23, tmp14)
tmp25 = 1 + x0
tmp26 = tmp25 >= tmp1
tmp27 = tmp25 < tmp3
tmp28 = tmp26 & tmp27
tmp29 = tmp5 & tmp28
tmp30 = tl.load(in_ptr0 + (-3 + x3), tmp29 & xmask, other=0.0)
tmp31 = tl.sigmoid(tmp30)
tmp32 = tl.full(tmp31.shape, float('-inf'), tmp31.dtype)
tmp33 = tl.where(tmp29, tmp31, tmp32)
tmp34 = triton_helpers.maximum(tmp33, tmp24)
tmp35 = x1
tmp36 = tmp35 >= tmp1
tmp37 = tmp35 < tmp3
tmp38 = tmp36 & tmp37
tmp39 = tmp38 & tmp9
tmp40 = tl.load(in_ptr0 + (-1 + x3), tmp39 & xmask, other=0.0)
tmp41 = tl.sigmoid(tmp40)
tmp42 = tl.full(tmp41.shape, float('-inf'), tmp41.dtype)
tmp43 = tl.where(tmp39, tmp41, tmp42)
tmp44 = triton_helpers.maximum(tmp43, tmp34)
tmp45 = tmp38 & tmp18
tmp46 = tl.load(in_ptr0 + x3, tmp45 & xmask, other=0.0)
tmp47 = tl.sigmoid(tmp46)
tmp48 = tl.full(tmp47.shape, float('-inf'), tmp47.dtype)
tmp49 = tl.where(tmp45, tmp47, tmp48)
tmp50 = triton_helpers.maximum(tmp49, tmp44)
tmp51 = tmp38 & tmp28
tmp52 = tl.load(in_ptr0 + (1 + x3), tmp51 & xmask, other=0.0)
tmp53 = tl.sigmoid(tmp52)
tmp54 = tl.full(tmp53.shape, float('-inf'), tmp53.dtype)
tmp55 = tl.where(tmp51, tmp53, tmp54)
tmp56 = triton_helpers.maximum(tmp55, tmp50)
tmp57 = 1 + x1
tmp58 = tmp57 >= tmp1
tmp59 = tmp57 < tmp3
tmp60 = tmp58 & tmp59
tmp61 = tmp60 & tmp9
tmp62 = tl.load(in_ptr0 + (3 + x3), tmp61 & xmask, other=0.0)
tmp63 = tl.sigmoid(tmp62)
tmp64 = tl.full(tmp63.shape, float('-inf'), tmp63.dtype)
tmp65 = tl.where(tmp61, tmp63, tmp64)
tmp66 = triton_helpers.maximum(tmp65, tmp56)
tmp67 = tmp60 & tmp18
tmp68 = tl.load(in_ptr0 + (4 + x3), tmp67 & xmask, other=0.0)
tmp69 = tl.sigmoid(tmp68)
tmp70 = tl.full(tmp69.shape, float('-inf'), tmp69.dtype)
tmp71 = tl.where(tmp67, tmp69, tmp70)
tmp72 = triton_helpers.maximum(tmp71, tmp66)
tmp73 = tmp60 & tmp28
tmp74 = tl.load(in_ptr0 + (5 + x3), tmp73 & xmask, other=0.0)
tmp75 = tl.sigmoid(tmp74)
tmp76 = tl.full(tmp75.shape, float('-inf'), tmp75.dtype)
tmp77 = tl.where(tmp73, tmp75, tmp76)
tmp78 = triton_helpers.maximum(tmp77, tmp72)
tmp79 = -1.0
tmp80 = tmp12 * tmp79
tmp81 = tl.full(tmp80.shape, float('-inf'), tmp80.dtype)
tmp82 = tl.where(tmp10, tmp80, tmp81)
tmp83 = tmp21 * tmp79
tmp84 = tl.full(tmp83.shape, float('-inf'), tmp83.dtype)
tmp85 = tl.where(tmp19, tmp83, tmp84)
tmp86 = triton_helpers.maximum(tmp85, tmp82)
tmp87 = tmp31 * tmp79
tmp88 = tl.full(tmp87.shape, float('-inf'), tmp87.dtype)
tmp89 = tl.where(tmp29, tmp87, tmp88)
tmp90 = triton_helpers.maximum(tmp89, tmp86)
tmp91 = tmp41 * tmp79
tmp92 = tl.full(tmp91.shape, float('-inf'), tmp91.dtype)
tmp93 = tl.where(tmp39, tmp91, tmp92)
tmp94 = triton_helpers.maximum(tmp93, tmp90)
tmp95 = tmp47 * tmp79
tmp96 = tl.full(tmp95.shape, float('-inf'), tmp95.dtype)
tmp97 = tl.where(tmp45, tmp95, tmp96)
tmp98 = triton_helpers.maximum(tmp97, tmp94)
tmp99 = tmp53 * tmp79
tmp100 = tl.full(tmp99.shape, float('-inf'), tmp99.dtype)
tmp101 = tl.where(tmp51, tmp99, tmp100)
tmp102 = triton_helpers.maximum(tmp101, tmp98)
tmp103 = tmp63 * tmp79
tmp104 = tl.full(tmp103.shape, float('-inf'), tmp103.dtype)
tmp105 = tl.where(tmp61, tmp103, tmp104)
tmp106 = triton_helpers.maximum(tmp105, tmp102)
tmp107 = tmp69 * tmp79
tmp108 = tl.full(tmp107.shape, float('-inf'), tmp107.dtype)
tmp109 = tl.where(tmp67, tmp107, tmp108)
tmp110 = triton_helpers.maximum(tmp109, tmp106)
tmp111 = tmp75 * tmp79
tmp112 = tl.full(tmp111.shape, float('-inf'), tmp111.dtype)
tmp113 = tl.where(tmp73, tmp111, tmp112)
tmp114 = triton_helpers.maximum(tmp113, tmp110)
tmp116 = tmp78 + tmp114
tmp117 = tmp115 * tmp116
tl.store(in_out_ptr0 + x3, tmp117, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_add_max_pool2d_with_indices_mul_sigmoid_0[grid(256)](
buf2, arg0_1, 256, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf2,
class BILMNew(nn.Module):
def __init__(self):
super(BILMNew, self).__init__()
self.maxpool1 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
self.maxpool2 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| SeunghwanByun/Real-Time-Road-Detection-Network | BILM | false | 1,046 | [
"MIT"
] | 0 | bc46615adef0e2b1a9a03dd4951559ca5849e6e1 | https://github.com/SeunghwanByun/Real-Time-Road-Detection-Network/tree/bc46615adef0e2b1a9a03dd4951559ca5849e6e1 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self):
super().__init__()
self.maxpool1 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
self.maxpool2 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
def forward(self, feat):
pos_sig = torch.sigmoid(feat)
neg_sig = -1 * pos_sig
pos_sig = self.maxpool1(pos_sig)
neg_sig = self.maxpool2(neg_sig)
sum_sig = pos_sig + neg_sig
x = feat * sum_sig
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
KLDLoss | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/5v/c5vtwldilipwi3r3wx5rsxypbvfmsav7s7xybl5bf72d3sckrknw.py
# Topologically Sorted Source Nodes: [add, pow_1, sub, exp, sub_1, sum_1, mul], Original ATen: [aten.add, aten.pow, aten.sub, aten.exp, aten.sum, aten.mul]
# Source node to ATen node mapping:
# add => add
# exp => exp
# mul => mul
# pow_1 => pow_1
# sub => sub
# sub_1 => sub_1
# sum_1 => sum_1
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, 1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg1_1, 2), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %pow_1), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%arg0_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %exp), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%sub_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, -0.5), kwargs = {})
triton_per_fused_add_exp_mul_pow_sub_sum_0 = async_compile.triton('triton_per_fused_add_exp_mul_pow_sub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_exp_mul_pow_sub_sum_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_add_exp_mul_pow_sub_sum_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp3 = tl.load(in_ptr1 + (r0), None)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 - tmp4
tmp6 = tl_math.exp(tmp0)
tmp7 = tmp5 - tmp6
tmp8 = tl.broadcast_to(tmp7, [RBLOCK])
tmp10 = triton_helpers.promote_to_tensor(tl.sum(tmp8, 0))
tmp11 = -0.5
tmp12 = tmp10 * tmp11
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp12, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [add, pow_1, sub, exp, sub_1, sum_1, mul], Original ATen: [aten.add, aten.pow, aten.sub, aten.exp, aten.sum, aten.mul]
stream0 = get_raw_stream(0)
triton_per_fused_add_exp_mul_pow_sub_sum_0.run(buf1, arg0_1, arg1_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
class KLDLoss(nn.Module):
def forward(self, mu, logvar):
return -0.5 * torch.sum(1 + logvar - mu.pow(2) - logvar.exp())
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_add_exp_mul_pow_sub_sum_0(in_out_ptr0, in_ptr0,
in_ptr1, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp3 = tl.load(in_ptr1 + r0, None)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 - tmp4
tmp6 = tl_math.exp(tmp0)
tmp7 = tmp5 - tmp6
tmp8 = tl.broadcast_to(tmp7, [RBLOCK])
tmp10 = triton_helpers.promote_to_tensor(tl.sum(tmp8, 0))
tmp11 = -0.5
tmp12 = tmp10 * tmp11
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp12, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_add_exp_mul_pow_sub_sum_0[grid(1)](buf1, arg0_1,
arg1_1, 1, 256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf1,
class KLDLossNew(nn.Module):
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| SebyakinAndrei/MichiGAN | KLDLoss | false | 1,047 | [
"MIT"
] | 0 | 6584c9a106b33096f38e8f5b11d0320f7065fd26 | https://github.com/SebyakinAndrei/MichiGAN/tree/6584c9a106b33096f38e8f5b11d0320f7065fd26 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def forward(self, mu, logvar):
return -0.5 * torch.sum(1 + logvar - mu.pow(2) - logvar.exp())
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
AddCoords | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/nz/cnzr3enannjni75kec3qorz6jm6lyd5whz6u5l3ih55bgihwnb2u.py
# Topologically Sorted Source Nodes: [ret], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# ret => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%arg0_1, %device_put, %device_put_1], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 16) % 6
x3 = (xindex // 96)
x4 = xindex % 16
x1 = (xindex // 4) % 4
x0 = xindex % 4
x5 = xindex
tmp0 = x2
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x4 + (16*x2) + (64*x3)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 5, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = x1
tmp11 = tmp10.to(tl.float32)
tmp12 = 0.3333333333333333
tmp13 = tmp11 * tmp12
tmp14 = 2.0
tmp15 = tmp13 * tmp14
tmp16 = 1.0
tmp17 = tmp15 - tmp16
tmp18 = tl.full(tmp17.shape, 0.0, tmp17.dtype)
tmp19 = tl.where(tmp9, tmp17, tmp18)
tmp20 = tmp0 >= tmp7
tmp21 = tl.full([1], 6, tl.int64)
tmp22 = tmp0 < tmp21
tmp23 = x0
tmp24 = tmp23.to(tl.float32)
tmp25 = tmp24 * tmp12
tmp26 = tmp25 * tmp14
tmp27 = tmp26 - tmp16
tmp28 = tl.full(tmp27.shape, 0.0, tmp27.dtype)
tmp29 = tl.where(tmp20, tmp27, tmp28)
tmp30 = tl.where(tmp9, tmp19, tmp29)
tmp31 = tl.where(tmp4, tmp5, tmp30)
tl.store(out_ptr0 + (x5), tmp31, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 6, 4, 4), (96, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [ret], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(arg0_1, buf0, 384, grid=grid(384), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class AddCoords(nn.Module):
def __init__(self, with_r=False):
super().__init__()
self.with_r = with_r
def forward(self, input_tensor):
"""
Args:
input_tensor: shape(batch, channel, x_dim, y_dim)
"""
batch_size, _, x_dim, y_dim = input_tensor.size()
xx_channel = torch.arange(x_dim).repeat(1, y_dim, 1)
yy_channel = torch.arange(y_dim).repeat(1, x_dim, 1).transpose(1, 2)
xx_channel = xx_channel.float() / (x_dim - 1)
yy_channel = yy_channel.float() / (y_dim - 1)
xx_channel = xx_channel * 2 - 1
yy_channel = yy_channel * 2 - 1
xx_channel = xx_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
yy_channel = yy_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
ret = torch.cat([input_tensor, xx_channel.type_as(input_tensor),
yy_channel.type_as(input_tensor)], dim=1)
if self.with_r:
rr = torch.sqrt(torch.pow(xx_channel.type_as(input_tensor) -
0.5, 2) + torch.pow(yy_channel.type_as(input_tensor) - 0.5, 2))
ret = torch.cat([ret, rr], dim=1)
return ret
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 16 % 6
x3 = xindex // 96
x4 = xindex % 16
x1 = xindex // 4 % 4
x0 = xindex % 4
x5 = xindex
tmp0 = x2
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x4 + 16 * x2 + 64 * x3), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 5, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = x1
tmp11 = tmp10.to(tl.float32)
tmp12 = 0.3333333333333333
tmp13 = tmp11 * tmp12
tmp14 = 2.0
tmp15 = tmp13 * tmp14
tmp16 = 1.0
tmp17 = tmp15 - tmp16
tmp18 = tl.full(tmp17.shape, 0.0, tmp17.dtype)
tmp19 = tl.where(tmp9, tmp17, tmp18)
tmp20 = tmp0 >= tmp7
tl.full([1], 6, tl.int64)
tmp23 = x0
tmp24 = tmp23.to(tl.float32)
tmp25 = tmp24 * tmp12
tmp26 = tmp25 * tmp14
tmp27 = tmp26 - tmp16
tmp28 = tl.full(tmp27.shape, 0.0, tmp27.dtype)
tmp29 = tl.where(tmp20, tmp27, tmp28)
tmp30 = tl.where(tmp9, tmp19, tmp29)
tmp31 = tl.where(tmp4, tmp5, tmp30)
tl.store(out_ptr0 + x5, tmp31, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 6, 4, 4), (96, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(384)](arg0_1, buf0, 384, XBLOCK=128,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class AddCoordsNew(nn.Module):
def __init__(self, with_r=False):
super().__init__()
self.with_r = with_r
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| SeunghwanByun/Real-Time-Road-Detection-Network | AddCoords | false | 1,048 | [
"MIT"
] | 0 | bc46615adef0e2b1a9a03dd4951559ca5849e6e1 | https://github.com/SeunghwanByun/Real-Time-Road-Detection-Network/tree/bc46615adef0e2b1a9a03dd4951559ca5849e6e1 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, with_r=False):
super().__init__()
self.with_r = with_r
def forward(self, input_tensor):
"""
Args:
input_tensor: shape(batch, channel, x_dim, y_dim)
"""
batch_size, _, x_dim, y_dim = input_tensor.size()
xx_channel = torch.arange(x_dim).repeat(1, y_dim, 1)
yy_channel = torch.arange(y_dim).repeat(1, x_dim, 1).transpose(1, 2)
xx_channel = xx_channel.float() / (x_dim - 1)
yy_channel = yy_channel.float() / (y_dim - 1)
xx_channel = xx_channel * 2 - 1
yy_channel = yy_channel * 2 - 1
xx_channel = xx_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
yy_channel = yy_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
ret = torch.cat([input_tensor, xx_channel.type_as(input_tensor),
yy_channel.type_as(input_tensor)], dim=1)
if self.with_r:
rr = torch.sqrt(torch.pow(xx_channel.type_as(input_tensor) -
0.5, 2) + torch.pow(yy_channel.type_as(input_tensor) - 0.5, 2))
ret = torch.cat([ret, rr], dim=1)
return ret
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
BCEDiceLoss | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/ij/cijqe3u6tb2m4eqcjcyn2yvugi5jt3lklpnhpcyrzltvdudroq22.py
# Topologically Sorted Source Nodes: [loss, pred], Original ATen: [aten._log_softmax, aten._softmax]
# Source node to ATen node mapping:
# loss => amax, sub
# pred => amax_1, exp_1, sub_2
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax), kwargs = {})
# %amax_1 : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [1], True), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax_1), kwargs = {})
# %exp_1 : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_2,), kwargs = {})
triton_poi_fused__log_softmax__softmax_0 = async_compile.triton('triton_poi_fused__log_softmax__softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax__softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax__softmax_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp8, xmask)
tl.store(out_ptr1 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/wk/cwk2wao7opapqbjj7klnqrd6tgist3ts3nc5veryzhzstwpx7d4l.py
# Topologically Sorted Source Nodes: [pred], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# pred => div, sum_3
# Graph fragment:
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_1, [1], True), kwargs = {})
# %div : [num_users=4] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_1, %sum_3), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/de/cdeyqtmcoa2bhcp6bumzbremn4d7odew3le3fnk2ptjpoeuhteru.py
# Topologically Sorted Source Nodes: [loss, loss_1, loss_bce, mul_14, mul_1, ne, valid_mask, valid_mask_1, mul_2, sum_1, mul_3, num, pow_1, pow_2, add_1, sum_2, den, truediv, loss_2, loss_3, total_loss, mul_4, valid_mask_2, mul_5, sum_3, mul_6, num_1, pow_3, pow_4, add_5, sum_4, den_1, truediv_1, loss_4, loss_5, total_loss_1, mul_7, valid_mask_3, mul_8, sum_5, mul_9, num_2, pow_5, pow_6, add_8, sum_6, den_2, truediv_2, loss_6, loss_7, total_loss_2, mul_10, valid_mask_4, mul_11, sum_7, mul_12, num_3, pow_7, pow_8, add_11, sum_8, den_3, truediv_3, loss_8, loss_9, total_loss_3, loss_10, loss_11, loss_dice, mul_15, add_13], Original ATen: [aten._log_softmax, aten.mul, aten.sum, aten.neg, aten.mean, aten.ne, aten._to_copy, aten.view, aten.add, aten.pow, aten.div, aten.rsub]
# Source node to ATen node mapping:
# add_1 => add_1
# add_11 => add_13
# add_13 => add_16
# add_5 => add_5
# add_8 => add_9
# den => add_2
# den_1 => add_6
# den_2 => add_10
# den_3 => add_14
# loss => exp, log, mul, neg, sub_1, sum_1, sum_2
# loss_1 => mean
# loss_10 => div_5
# loss_11 => mean_5
# loss_2 => sub_3
# loss_3 => mean_1
# loss_4 => sub_4
# loss_5 => mean_2
# loss_6 => sub_5
# loss_7 => mean_3
# loss_8 => sub_6
# loss_9 => mean_4
# loss_bce => mul_1
# loss_dice => mul_14
# mul_1 => mul_2
# mul_10 => mul_11
# mul_11 => mul_12
# mul_12 => mul_13
# mul_14 => mul_15
# mul_15 => mul_16
# mul_2 => mul_3
# mul_3 => mul_4
# mul_4 => mul_5
# mul_5 => mul_6
# mul_6 => mul_7
# mul_7 => mul_8
# mul_8 => mul_9
# mul_9 => mul_10
# ne => ne
# num => add
# num_1 => add_4
# num_2 => add_8
# num_3 => add_12
# pow_1 => pow_1
# pow_2 => pow_2
# pow_3 => pow_3
# pow_4 => pow_4
# pow_5 => pow_5
# pow_6 => pow_6
# pow_7 => pow_7
# pow_8 => pow_8
# sum_1 => sum_4
# sum_2 => sum_5
# sum_3 => sum_6
# sum_4 => sum_7
# sum_5 => sum_8
# sum_6 => sum_9
# sum_7 => sum_10
# sum_8 => sum_11
# total_loss => add_3
# total_loss_1 => add_7
# total_loss_2 => add_11
# total_loss_3 => add_15
# truediv => div_1
# truediv_1 => div_2
# truediv_2 => div_3
# truediv_3 => div_4
# valid_mask => convert_element_type_2
# valid_mask_1 => view_2
# valid_mask_2 => view_5
# valid_mask_3 => view_8
# valid_mask_4 => view_11
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %log), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %arg1_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%sum_2,), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%neg,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, 1.0), kwargs = {})
# %mul_15 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_1, 0.5), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %select_1), kwargs = {})
# %ne : [num_users=1] = call_function[target=torch.ops.aten.ne.Scalar](args = (%arg1_1, 255), kwargs = {})
# %convert_element_type_2 : [num_users=4] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%ne, torch.int64), kwargs = {})
# %view_2 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_2, %view_2), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_3, [1]), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_4, 2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_4, 1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view, 2), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_1, 2), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_1, %pow_2), kwargs = {})
# %sum_5 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_1, [1]), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_5, 1), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %add_2), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_1), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_3,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean_1, 0), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, %select_3), kwargs = {})
# %view_5 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_5, %view_5), kwargs = {})
# %sum_6 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_6, [1]), kwargs = {})
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_6, 2), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_7, 1), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_3, 2), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_3, 2), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_3, %pow_4), kwargs = {})
# %sum_7 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_5, [1]), kwargs = {})
# %add_6 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_7, 1), kwargs = {})
# %div_2 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_4, %add_6), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_2), kwargs = {})
# %mean_2 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_4,), kwargs = {})
# %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_3, %mean_2), kwargs = {})
# %mul_8 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_6, %select_5), kwargs = {})
# %view_8 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_9 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_8, %view_8), kwargs = {})
# %sum_8 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_9, [1]), kwargs = {})
# %mul_10 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_8, 2), kwargs = {})
# %add_8 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_10, 1), kwargs = {})
# %pow_5 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_6, 2), kwargs = {})
# %pow_6 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_5, 2), kwargs = {})
# %add_9 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_5, %pow_6), kwargs = {})
# %sum_9 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_9, [1]), kwargs = {})
# %add_10 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_9, 1), kwargs = {})
# %div_3 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_8, %add_10), kwargs = {})
# %sub_5 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_3), kwargs = {})
# %mean_3 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_5,), kwargs = {})
# %add_11 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_7, %mean_3), kwargs = {})
# %mul_11 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_9, %select_7), kwargs = {})
# %view_11 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%convert_element_type_2, [4, -1]), kwargs = {})
# %mul_12 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_11, %view_11), kwargs = {})
# %sum_10 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_12, [1]), kwargs = {})
# %mul_13 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_10, 2), kwargs = {})
# %add_12 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_13, 1), kwargs = {})
# %pow_7 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_9, 2), kwargs = {})
# %pow_8 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%select_7, 2), kwargs = {})
# %add_13 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_7, %pow_8), kwargs = {})
# %sum_11 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%add_13, [1]), kwargs = {})
# %add_14 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_11, 1), kwargs = {})
# %div_4 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_12, %add_14), kwargs = {})
# %sub_6 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div_4), kwargs = {})
# %mean_4 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_6,), kwargs = {})
# %add_15 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_11, %mean_4), kwargs = {})
# %div_5 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add_15, 4), kwargs = {})
# %mean_5 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%div_5,), kwargs = {})
# %mul_14 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean_5, 1.0), kwargs = {})
# %mul_16 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_14, 0.5), kwargs = {})
# %add_16 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_15, %mul_16), kwargs = {})
triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2 = async_compile.triton('triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {4: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=(4,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 5, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 4
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (4*r0), None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (4*r0), None, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr1 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr1 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr1 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp31 = tl.load(in_ptr2 + (4*r0), None, eviction_policy='evict_last')
tmp98 = tl.load(in_ptr2 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp139 = tl.load(in_ptr2 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp180 = tl.load(in_ptr2 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp1 = tl_math.exp(tmp0)
tmp3 = tl_math.exp(tmp2)
tmp4 = tmp1 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp14 = tmp12 * tmp13
tmp15 = tmp2 - tmp11
tmp17 = tmp15 * tmp16
tmp18 = tmp14 + tmp17
tmp19 = tmp5 - tmp11
tmp21 = tmp19 * tmp20
tmp22 = tmp18 + tmp21
tmp23 = tmp8 - tmp11
tmp25 = tmp23 * tmp24
tmp26 = tmp22 + tmp25
tmp27 = -tmp26
tmp28 = tl.broadcast_to(tmp27, [XBLOCK, RBLOCK])
tmp30 = tl.sum(tmp28, 1)[:, None]
tmp32 = tmp13.to(tl.int64)
tmp33 = tl.full([1, 1], 0, tl.int64)
tmp34 = triton_helpers.maximum(tmp32, tmp33)
tmp35 = tl.full([1, 1], 3, tl.int64)
tmp36 = triton_helpers.minimum(tmp34, tmp35)
tmp37 = tmp36 == tmp33
tmp38 = tmp37.to(tl.int64)
tmp39 = tmp38.to(tl.float32)
tmp40 = tmp31 * tmp39
tmp41 = 255.0
tmp42 = tmp13 != tmp41
tmp43 = tmp42.to(tl.int64)
tmp44 = tmp43.to(tl.float32)
tmp45 = tmp40 * tmp44
tmp46 = tmp16.to(tl.int64)
tmp47 = triton_helpers.maximum(tmp46, tmp33)
tmp48 = triton_helpers.minimum(tmp47, tmp35)
tmp49 = tmp48 == tmp33
tmp50 = tmp49.to(tl.int64)
tmp51 = tmp50.to(tl.float32)
tmp52 = tmp31 * tmp51
tmp53 = tmp16 != tmp41
tmp54 = tmp53.to(tl.int64)
tmp55 = tmp54.to(tl.float32)
tmp56 = tmp52 * tmp55
tmp57 = tmp45 + tmp56
tmp58 = tmp20.to(tl.int64)
tmp59 = triton_helpers.maximum(tmp58, tmp33)
tmp60 = triton_helpers.minimum(tmp59, tmp35)
tmp61 = tmp60 == tmp33
tmp62 = tmp61.to(tl.int64)
tmp63 = tmp62.to(tl.float32)
tmp64 = tmp31 * tmp63
tmp65 = tmp20 != tmp41
tmp66 = tmp65.to(tl.int64)
tmp67 = tmp66.to(tl.float32)
tmp68 = tmp64 * tmp67
tmp69 = tmp57 + tmp68
tmp70 = tmp24.to(tl.int64)
tmp71 = triton_helpers.maximum(tmp70, tmp33)
tmp72 = triton_helpers.minimum(tmp71, tmp35)
tmp73 = tmp72 == tmp33
tmp74 = tmp73.to(tl.int64)
tmp75 = tmp74.to(tl.float32)
tmp76 = tmp31 * tmp75
tmp77 = tmp24 != tmp41
tmp78 = tmp77.to(tl.int64)
tmp79 = tmp78.to(tl.float32)
tmp80 = tmp76 * tmp79
tmp81 = tmp69 + tmp80
tmp82 = tmp31 * tmp31
tmp83 = tmp38 * tmp38
tmp84 = tmp83.to(tl.float32)
tmp85 = tmp82 + tmp84
tmp86 = tmp50 * tmp50
tmp87 = tmp86.to(tl.float32)
tmp88 = tmp82 + tmp87
tmp89 = tmp85 + tmp88
tmp90 = tmp62 * tmp62
tmp91 = tmp90.to(tl.float32)
tmp92 = tmp82 + tmp91
tmp93 = tmp89 + tmp92
tmp94 = tmp74 * tmp74
tmp95 = tmp94.to(tl.float32)
tmp96 = tmp82 + tmp95
tmp97 = tmp93 + tmp96
tmp99 = tl.full([1, 1], 1, tl.int64)
tmp100 = tmp36 == tmp99
tmp101 = tmp100.to(tl.int64)
tmp102 = tmp101.to(tl.float32)
tmp103 = tmp98 * tmp102
tmp104 = tmp103 * tmp44
tmp105 = tmp48 == tmp99
tmp106 = tmp105.to(tl.int64)
tmp107 = tmp106.to(tl.float32)
tmp108 = tmp98 * tmp107
tmp109 = tmp108 * tmp55
tmp110 = tmp104 + tmp109
tmp111 = tmp60 == tmp99
tmp112 = tmp111.to(tl.int64)
tmp113 = tmp112.to(tl.float32)
tmp114 = tmp98 * tmp113
tmp115 = tmp114 * tmp67
tmp116 = tmp110 + tmp115
tmp117 = tmp72 == tmp99
tmp118 = tmp117.to(tl.int64)
tmp119 = tmp118.to(tl.float32)
tmp120 = tmp98 * tmp119
tmp121 = tmp120 * tmp79
tmp122 = tmp116 + tmp121
tmp123 = tmp98 * tmp98
tmp124 = tmp101 * tmp101
tmp125 = tmp124.to(tl.float32)
tmp126 = tmp123 + tmp125
tmp127 = tmp106 * tmp106
tmp128 = tmp127.to(tl.float32)
tmp129 = tmp123 + tmp128
tmp130 = tmp126 + tmp129
tmp131 = tmp112 * tmp112
tmp132 = tmp131.to(tl.float32)
tmp133 = tmp123 + tmp132
tmp134 = tmp130 + tmp133
tmp135 = tmp118 * tmp118
tmp136 = tmp135.to(tl.float32)
tmp137 = tmp123 + tmp136
tmp138 = tmp134 + tmp137
tmp140 = tl.full([1, 1], 2, tl.int64)
tmp141 = tmp36 == tmp140
tmp142 = tmp141.to(tl.int64)
tmp143 = tmp142.to(tl.float32)
tmp144 = tmp139 * tmp143
tmp145 = tmp144 * tmp44
tmp146 = tmp48 == tmp140
tmp147 = tmp146.to(tl.int64)
tmp148 = tmp147.to(tl.float32)
tmp149 = tmp139 * tmp148
tmp150 = tmp149 * tmp55
tmp151 = tmp145 + tmp150
tmp152 = tmp60 == tmp140
tmp153 = tmp152.to(tl.int64)
tmp154 = tmp153.to(tl.float32)
tmp155 = tmp139 * tmp154
tmp156 = tmp155 * tmp67
tmp157 = tmp151 + tmp156
tmp158 = tmp72 == tmp140
tmp159 = tmp158.to(tl.int64)
tmp160 = tmp159.to(tl.float32)
tmp161 = tmp139 * tmp160
tmp162 = tmp161 * tmp79
tmp163 = tmp157 + tmp162
tmp164 = tmp139 * tmp139
tmp165 = tmp142 * tmp142
tmp166 = tmp165.to(tl.float32)
tmp167 = tmp164 + tmp166
tmp168 = tmp147 * tmp147
tmp169 = tmp168.to(tl.float32)
tmp170 = tmp164 + tmp169
tmp171 = tmp167 + tmp170
tmp172 = tmp153 * tmp153
tmp173 = tmp172.to(tl.float32)
tmp174 = tmp164 + tmp173
tmp175 = tmp171 + tmp174
tmp176 = tmp159 * tmp159
tmp177 = tmp176.to(tl.float32)
tmp178 = tmp164 + tmp177
tmp179 = tmp175 + tmp178
tmp181 = tmp36 == tmp35
tmp182 = tmp181.to(tl.int64)
tmp183 = tmp182.to(tl.float32)
tmp184 = tmp180 * tmp183
tmp185 = tmp184 * tmp44
tmp186 = tmp48 == tmp35
tmp187 = tmp186.to(tl.int64)
tmp188 = tmp187.to(tl.float32)
tmp189 = tmp180 * tmp188
tmp190 = tmp189 * tmp55
tmp191 = tmp185 + tmp190
tmp192 = tmp60 == tmp35
tmp193 = tmp192.to(tl.int64)
tmp194 = tmp193.to(tl.float32)
tmp195 = tmp180 * tmp194
tmp196 = tmp195 * tmp67
tmp197 = tmp191 + tmp196
tmp198 = tmp72 == tmp35
tmp199 = tmp198.to(tl.int64)
tmp200 = tmp199.to(tl.float32)
tmp201 = tmp180 * tmp200
tmp202 = tmp201 * tmp79
tmp203 = tmp197 + tmp202
tmp204 = tmp180 * tmp180
tmp205 = tmp182 * tmp182
tmp206 = tmp205.to(tl.float32)
tmp207 = tmp204 + tmp206
tmp208 = tmp187 * tmp187
tmp209 = tmp208.to(tl.float32)
tmp210 = tmp204 + tmp209
tmp211 = tmp207 + tmp210
tmp212 = tmp193 * tmp193
tmp213 = tmp212.to(tl.float32)
tmp214 = tmp204 + tmp213
tmp215 = tmp211 + tmp214
tmp216 = tmp199 * tmp199
tmp217 = tmp216.to(tl.float32)
tmp218 = tmp204 + tmp217
tmp219 = tmp215 + tmp218
tmp220 = 2.0
tmp221 = tmp81 * tmp220
tmp222 = 1.0
tmp223 = tmp221 + tmp222
tmp224 = tmp97 + tmp222
tmp225 = tmp223 / tmp224
tmp226 = tmp222 - tmp225
tmp227 = tl.broadcast_to(tmp226, [XBLOCK, RBLOCK])
tmp229 = tl.sum(tmp227, 1)[:, None]
tmp230 = tmp122 * tmp220
tmp231 = tmp230 + tmp222
tmp232 = tmp138 + tmp222
tmp233 = tmp231 / tmp232
tmp234 = tmp222 - tmp233
tmp235 = tl.broadcast_to(tmp234, [XBLOCK, RBLOCK])
tmp237 = tl.sum(tmp235, 1)[:, None]
tmp238 = tmp163 * tmp220
tmp239 = tmp238 + tmp222
tmp240 = tmp179 + tmp222
tmp241 = tmp239 / tmp240
tmp242 = tmp222 - tmp241
tmp243 = tl.broadcast_to(tmp242, [XBLOCK, RBLOCK])
tmp245 = tl.sum(tmp243, 1)[:, None]
tmp246 = tmp203 * tmp220
tmp247 = tmp246 + tmp222
tmp248 = tmp219 + tmp222
tmp249 = tmp247 / tmp248
tmp250 = tmp222 - tmp249
tmp251 = tl.broadcast_to(tmp250, [XBLOCK, RBLOCK])
tmp253 = tl.sum(tmp251, 1)[:, None]
tmp254 = 4.0
tmp255 = tmp30 / tmp254
tmp256 = tmp255 * tmp222
tmp257 = 0.5
tmp258 = tmp256 * tmp257
tmp259 = tmp229 / tmp254
tmp260 = 0.0
tmp261 = tmp259 + tmp260
tmp262 = tmp237 / tmp254
tmp263 = tmp261 + tmp262
tmp264 = tmp245 / tmp254
tmp265 = tmp263 + tmp264
tmp266 = tmp253 / tmp254
tmp267 = tmp265 + tmp266
tmp268 = 0.25
tmp269 = tmp267 * tmp268
tmp270 = tmp269 / tmp222
tmp271 = tmp270 * tmp222
tmp272 = tmp271 * tmp257
tmp273 = tmp258 + tmp272
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp273, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4), (4, 1))
assert_size_stride(arg1_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [loss, pred], Original ATen: [aten._log_softmax, aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax__softmax_0.run(arg0_1, buf0, buf2, 16, grid=grid(16), stream=stream0)
del arg0_1
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pred], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf2, buf3, 16, grid=grid(16), stream=stream0)
del buf2
buf1 = empty_strided_cuda((), (), torch.float32)
buf16 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [loss, loss_1, loss_bce, mul_14, mul_1, ne, valid_mask, valid_mask_1, mul_2, sum_1, mul_3, num, pow_1, pow_2, add_1, sum_2, den, truediv, loss_2, loss_3, total_loss, mul_4, valid_mask_2, mul_5, sum_3, mul_6, num_1, pow_3, pow_4, add_5, sum_4, den_1, truediv_1, loss_4, loss_5, total_loss_1, mul_7, valid_mask_3, mul_8, sum_5, mul_9, num_2, pow_5, pow_6, add_8, sum_6, den_2, truediv_2, loss_6, loss_7, total_loss_2, mul_10, valid_mask_4, mul_11, sum_7, mul_12, num_3, pow_7, pow_8, add_11, sum_8, den_3, truediv_3, loss_8, loss_9, total_loss_3, loss_10, loss_11, loss_dice, mul_15, add_13], Original ATen: [aten._log_softmax, aten.mul, aten.sum, aten.neg, aten.mean, aten.ne, aten._to_copy, aten.view, aten.add, aten.pow, aten.div, aten.rsub]
triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2.run(buf16, buf0, arg1_1, buf3, 1, 4, grid=grid(1), stream=stream0)
del arg1_1
del buf0
del buf3
return (buf16, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import functools
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
def _expand_onehot_labels(labels, label_weights, target_shape, ignore_index):
"""Expand onehot labels to match the size of prediction."""
bin_labels = labels.new_zeros(target_shape)
valid_mask = (labels >= 0) & (labels != ignore_index)
inds = torch.nonzero(valid_mask, as_tuple=True)
if inds[0].numel() > 0:
if labels.dim() == 3:
bin_labels[inds[0], labels[valid_mask], inds[1], inds[2]] = 1
else:
bin_labels[inds[0], labels[valid_mask]] = 1
valid_mask = valid_mask.unsqueeze(1).expand(target_shape).float()
if label_weights is None:
bin_label_weights = valid_mask
else:
bin_label_weights = label_weights.unsqueeze(1).expand(target_shape)
bin_label_weights *= valid_mask
return bin_labels, bin_label_weights
def reduce_loss(loss, reduction):
"""Reduce loss as specified.
Args:
loss (Tensor): Elementwise loss tensor.
reduction (str): Options are "none", "mean" and "sum".
Return:
Tensor: Reduced loss tensor.
"""
reduction_enum = F._Reduction.get_enum(reduction)
if reduction_enum == 0:
return loss
elif reduction_enum == 1:
return loss.mean()
elif reduction_enum == 2:
return loss.sum()
def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
"""Apply element-wise weight and reduce loss.
Args:
loss (Tensor): Element-wise loss.
weight (Tensor): Element-wise weights.
reduction (str): Same as built-in losses of PyTorch.
avg_factor (float): Avarage factor when computing the mean of losses.
Returns:
Tensor: Processed loss values.
"""
if weight is not None:
assert weight.dim() == loss.dim()
if weight.dim() > 1:
assert weight.size(1) == 1 or weight.size(1) == loss.size(1)
loss = loss * weight
if avg_factor is None:
loss = reduce_loss(loss, reduction)
elif reduction == 'mean':
loss = loss.sum() / avg_factor
elif reduction != 'none':
raise ValueError('avg_factor can not be used with reduction="sum"')
return loss
def binary_cross_entropy(pred, label, weight=None, reduction='mean',
avg_factor=None, class_weight=None, ignore_index=255):
"""Calculate the binary CrossEntropy loss.
Args:
pred (torch.Tensor): The prediction with shape (N, 1).
label (torch.Tensor): The learning label of the prediction.
weight (torch.Tensor, optional): Sample-wise loss weight.
reduction (str, optional): The method used to reduce the loss.
Options are "none", "mean" and "sum".
avg_factor (int, optional): Average factor that is used to average
the loss. Defaults to None.
class_weight (list[float], optional): The weight for each class.
ignore_index (int | None): The label index to be ignored. Default: 255
Returns:
torch.Tensor: The calculated loss
"""
if pred.dim() != label.dim():
assert pred.dim() == 2 and label.dim() == 1 or pred.dim(
) == 4 and label.dim(
) == 3, 'Only pred shape [N, C], label shape [N] or pred shape [N, C, H, W], label shape [N, H, W] are supported'
label, weight = _expand_onehot_labels(label, weight, pred.shape,
ignore_index)
if weight is not None:
weight = weight.float()
loss = F.binary_cross_entropy_with_logits(pred, label.float(),
pos_weight=class_weight, reduction='none')
loss = weight_reduce_loss(loss, weight, reduction=reduction, avg_factor
=avg_factor)
return loss
def cross_entropy(pred, label, weight=None, class_weight=None, reduction=
'mean', avg_factor=None, ignore_index=-100):
"""The wrapper function for :func:`F.cross_entropy`"""
loss = F.cross_entropy(pred, label, weight=class_weight, reduction=
'none', ignore_index=ignore_index)
if weight is not None:
weight = weight.float()
loss = weight_reduce_loss(loss, weight=weight, reduction=reduction,
avg_factor=avg_factor)
return loss
def weighted_loss(loss_func):
"""Create a weighted version of a given loss function.
To use this decorator, the loss function must have the signature like
`loss_func(pred, target, **kwargs)`. The function only needs to compute
element-wise loss without any reduction. This decorator will add weight
and reduction arguments to the function. The decorated function will have
the signature like `loss_func(pred, target, weight=None, reduction='mean',
avg_factor=None, **kwargs)`.
:Example:
>>> import torch
>>> @weighted_loss
>>> def l1_loss(pred, target):
>>> return (pred - target).abs()
>>> pred = torch.Tensor([0, 2, 3])
>>> target = torch.Tensor([1, 1, 1])
>>> weight = torch.Tensor([1, 0, 1])
>>> l1_loss(pred, target)
tensor(1.3333)
>>> l1_loss(pred, target, weight)
tensor(1.)
>>> l1_loss(pred, target, reduction='none')
tensor([1., 1., 2.])
>>> l1_loss(pred, target, weight, avg_factor=2)
tensor(1.5000)
"""
@functools.wraps(loss_func)
def wrapper(pred, target, weight=None, reduction='mean', avg_factor=
None, **kwargs):
loss = loss_func(pred, target, **kwargs)
loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
return loss
return wrapper
@weighted_loss
def binary_dice_loss(pred, target, valid_mask, smooth=1, exponent=2, **kwards):
assert pred.shape[0] == target.shape[0]
pred = pred.reshape(pred.shape[0], -1)
target = target.reshape(target.shape[0], -1)
valid_mask = valid_mask.reshape(valid_mask.shape[0], -1)
num = torch.sum(torch.mul(pred, target) * valid_mask, dim=1) * 2 + smooth
den = torch.sum(pred.pow(exponent) + target.pow(exponent), dim=1) + smooth
return 1 - num / den
@weighted_loss
def dice_loss(pred, target, valid_mask, smooth=1, exponent=2, class_weight=
None, ignore_index=255):
assert pred.shape[0] == target.shape[0]
total_loss = 0
num_classes = pred.shape[1]
for i in range(num_classes):
if i != ignore_index:
dice_loss = binary_dice_loss(pred[:, i], target[..., i],
valid_mask=valid_mask, smooth=smooth, exponent=exponent)
if class_weight is not None:
dice_loss *= class_weight[i]
total_loss += dice_loss
return total_loss / num_classes
def get_class_weight(class_weight):
"""Get class weight for loss function.
Args:
class_weight (list[float] | str | None): If class_weight is a str,
take it as a file name and read from it.
"""
if isinstance(class_weight, str):
if class_weight.endswith('.npy'):
class_weight = np.load(class_weight)
else:
class_weight = mmcv.load(class_weight)
return class_weight
def mask_cross_entropy(pred, target, label, reduction='mean', avg_factor=
None, class_weight=None, ignore_index=None):
"""Calculate the CrossEntropy loss for masks.
Args:
pred (torch.Tensor): The prediction with shape (N, C), C is the number
of classes.
target (torch.Tensor): The learning label of the prediction.
label (torch.Tensor): ``label`` indicates the class label of the mask'
corresponding object. This will be used to select the mask in the
of the class which the object belongs to when the mask prediction
if not class-agnostic.
reduction (str, optional): The method used to reduce the loss.
Options are "none", "mean" and "sum".
avg_factor (int, optional): Average factor that is used to average
the loss. Defaults to None.
class_weight (list[float], optional): The weight for each class.
ignore_index (None): Placeholder, to be consistent with other loss.
Default: None.
Returns:
torch.Tensor: The calculated loss
"""
assert ignore_index is None, 'BCE loss does not support ignore_index'
assert reduction == 'mean' and avg_factor is None
num_rois = pred.size()[0]
inds = torch.arange(0, num_rois, dtype=torch.long, device=pred.device)
pred_slice = pred[inds, label].squeeze(1)
return F.binary_cross_entropy_with_logits(pred_slice, target, weight=
class_weight, reduction='mean')[None]
class BCEDiceLoss(nn.Module):
"""CrossEntropyLoss.
Args:
use_sigmoid (bool, optional): Whether the prediction uses sigmoid
of softmax. Defaults to False.
use_mask (bool, optional): Whether to use mask cross entropy loss.
Defaults to False.
reduction (str, optional): . Defaults to 'mean'.
Options are "none", "mean" and "sum".
class_weight (list[float] | str, optional): Weight of each class. If in
str format, read them from a file. Defaults to None.
loss_weight (float, optional): Weight of the loss. Defaults to 1.0.
"""
def __init__(self, use_sigmoid=False, use_mask=False, smooth=1,
exponent=2, reduction='mean', class_weight=None, loss_weight=1.0,
ignore_index=255, bce_weight=0.5, **kwards):
super(BCEDiceLoss, self).__init__()
assert use_sigmoid is False or use_mask is False
self.use_sigmoid = use_sigmoid
self.use_mask = use_mask
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.class_weight = get_class_weight(class_weight)
self.loss_weight = loss_weight
self.ignore_index = ignore_index
self.bce_weight = bce_weight
if self.use_sigmoid:
self.cls_criterion = binary_cross_entropy
elif self.use_mask:
self.cls_criterion = mask_cross_entropy
else:
self.cls_criterion = cross_entropy
def forward(self, pred, target, weight=None, avg_factor=None,
reduction_override=None, **kwards):
assert reduction_override in (None, 'none', 'mean', 'sum')
reduction = (reduction_override if reduction_override else self.
reduction)
if self.class_weight is not None:
class_weight = pred.new_tensor(self.class_weight)
else:
class_weight = None
loss_bce = self.loss_weight * self.cls_criterion(pred, target,
weight, class_weight=class_weight, reduction=reduction,
avg_factor=avg_factor, **kwards)
pred = F.softmax(pred, dim=1)
num_classes = pred.shape[1]
one_hot_target = F.one_hot(torch.clamp(target.long(), 0,
num_classes - 1), num_classes=num_classes)
valid_mask = (target != self.ignore_index).long()
loss_dice = self.loss_weight * dice_loss(pred, one_hot_target,
valid_mask=valid_mask, reduction=reduction, avg_factor=
avg_factor, smooth=self.smooth, exponent=self.exponent,
class_weight=class_weight, ignore_index=self.ignore_index)
return loss_bce * self.bce_weight + loss_dice * (1 - self.bce_weight)
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import functools
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__log_softmax__softmax_0(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp8, xmask)
tl.store(out_ptr1 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2(
in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, xnumel, rnumel, XBLOCK: tl.
constexpr):
RBLOCK: tl.constexpr = 4
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + 4 * r0, None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (2 + 4 * r0), None, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (3 + 4 * r0), None, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + 4 * r0, None, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr1 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr1 + (2 + 4 * r0), None, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr1 + (3 + 4 * r0), None, eviction_policy='evict_last')
tmp31 = tl.load(in_ptr2 + 4 * r0, None, eviction_policy='evict_last')
tmp98 = tl.load(in_ptr2 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp139 = tl.load(in_ptr2 + (2 + 4 * r0), None, eviction_policy='evict_last'
)
tmp180 = tl.load(in_ptr2 + (3 + 4 * r0), None, eviction_policy='evict_last'
)
tmp1 = tl_math.exp(tmp0)
tmp3 = tl_math.exp(tmp2)
tmp4 = tmp1 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp14 = tmp12 * tmp13
tmp15 = tmp2 - tmp11
tmp17 = tmp15 * tmp16
tmp18 = tmp14 + tmp17
tmp19 = tmp5 - tmp11
tmp21 = tmp19 * tmp20
tmp22 = tmp18 + tmp21
tmp23 = tmp8 - tmp11
tmp25 = tmp23 * tmp24
tmp26 = tmp22 + tmp25
tmp27 = -tmp26
tmp28 = tl.broadcast_to(tmp27, [XBLOCK, RBLOCK])
tmp30 = tl.sum(tmp28, 1)[:, None]
tmp32 = tmp13.to(tl.int64)
tmp33 = tl.full([1, 1], 0, tl.int64)
tmp34 = triton_helpers.maximum(tmp32, tmp33)
tmp35 = tl.full([1, 1], 3, tl.int64)
tmp36 = triton_helpers.minimum(tmp34, tmp35)
tmp37 = tmp36 == tmp33
tmp38 = tmp37.to(tl.int64)
tmp39 = tmp38.to(tl.float32)
tmp40 = tmp31 * tmp39
tmp41 = 255.0
tmp42 = tmp13 != tmp41
tmp43 = tmp42.to(tl.int64)
tmp44 = tmp43.to(tl.float32)
tmp45 = tmp40 * tmp44
tmp46 = tmp16.to(tl.int64)
tmp47 = triton_helpers.maximum(tmp46, tmp33)
tmp48 = triton_helpers.minimum(tmp47, tmp35)
tmp49 = tmp48 == tmp33
tmp50 = tmp49.to(tl.int64)
tmp51 = tmp50.to(tl.float32)
tmp52 = tmp31 * tmp51
tmp53 = tmp16 != tmp41
tmp54 = tmp53.to(tl.int64)
tmp55 = tmp54.to(tl.float32)
tmp56 = tmp52 * tmp55
tmp57 = tmp45 + tmp56
tmp58 = tmp20.to(tl.int64)
tmp59 = triton_helpers.maximum(tmp58, tmp33)
tmp60 = triton_helpers.minimum(tmp59, tmp35)
tmp61 = tmp60 == tmp33
tmp62 = tmp61.to(tl.int64)
tmp63 = tmp62.to(tl.float32)
tmp64 = tmp31 * tmp63
tmp65 = tmp20 != tmp41
tmp66 = tmp65.to(tl.int64)
tmp67 = tmp66.to(tl.float32)
tmp68 = tmp64 * tmp67
tmp69 = tmp57 + tmp68
tmp70 = tmp24.to(tl.int64)
tmp71 = triton_helpers.maximum(tmp70, tmp33)
tmp72 = triton_helpers.minimum(tmp71, tmp35)
tmp73 = tmp72 == tmp33
tmp74 = tmp73.to(tl.int64)
tmp75 = tmp74.to(tl.float32)
tmp76 = tmp31 * tmp75
tmp77 = tmp24 != tmp41
tmp78 = tmp77.to(tl.int64)
tmp79 = tmp78.to(tl.float32)
tmp80 = tmp76 * tmp79
tmp81 = tmp69 + tmp80
tmp82 = tmp31 * tmp31
tmp83 = tmp38 * tmp38
tmp84 = tmp83.to(tl.float32)
tmp85 = tmp82 + tmp84
tmp86 = tmp50 * tmp50
tmp87 = tmp86.to(tl.float32)
tmp88 = tmp82 + tmp87
tmp89 = tmp85 + tmp88
tmp90 = tmp62 * tmp62
tmp91 = tmp90.to(tl.float32)
tmp92 = tmp82 + tmp91
tmp93 = tmp89 + tmp92
tmp94 = tmp74 * tmp74
tmp95 = tmp94.to(tl.float32)
tmp96 = tmp82 + tmp95
tmp97 = tmp93 + tmp96
tmp99 = tl.full([1, 1], 1, tl.int64)
tmp100 = tmp36 == tmp99
tmp101 = tmp100.to(tl.int64)
tmp102 = tmp101.to(tl.float32)
tmp103 = tmp98 * tmp102
tmp104 = tmp103 * tmp44
tmp105 = tmp48 == tmp99
tmp106 = tmp105.to(tl.int64)
tmp107 = tmp106.to(tl.float32)
tmp108 = tmp98 * tmp107
tmp109 = tmp108 * tmp55
tmp110 = tmp104 + tmp109
tmp111 = tmp60 == tmp99
tmp112 = tmp111.to(tl.int64)
tmp113 = tmp112.to(tl.float32)
tmp114 = tmp98 * tmp113
tmp115 = tmp114 * tmp67
tmp116 = tmp110 + tmp115
tmp117 = tmp72 == tmp99
tmp118 = tmp117.to(tl.int64)
tmp119 = tmp118.to(tl.float32)
tmp120 = tmp98 * tmp119
tmp121 = tmp120 * tmp79
tmp122 = tmp116 + tmp121
tmp123 = tmp98 * tmp98
tmp124 = tmp101 * tmp101
tmp125 = tmp124.to(tl.float32)
tmp126 = tmp123 + tmp125
tmp127 = tmp106 * tmp106
tmp128 = tmp127.to(tl.float32)
tmp129 = tmp123 + tmp128
tmp130 = tmp126 + tmp129
tmp131 = tmp112 * tmp112
tmp132 = tmp131.to(tl.float32)
tmp133 = tmp123 + tmp132
tmp134 = tmp130 + tmp133
tmp135 = tmp118 * tmp118
tmp136 = tmp135.to(tl.float32)
tmp137 = tmp123 + tmp136
tmp138 = tmp134 + tmp137
tmp140 = tl.full([1, 1], 2, tl.int64)
tmp141 = tmp36 == tmp140
tmp142 = tmp141.to(tl.int64)
tmp143 = tmp142.to(tl.float32)
tmp144 = tmp139 * tmp143
tmp145 = tmp144 * tmp44
tmp146 = tmp48 == tmp140
tmp147 = tmp146.to(tl.int64)
tmp148 = tmp147.to(tl.float32)
tmp149 = tmp139 * tmp148
tmp150 = tmp149 * tmp55
tmp151 = tmp145 + tmp150
tmp152 = tmp60 == tmp140
tmp153 = tmp152.to(tl.int64)
tmp154 = tmp153.to(tl.float32)
tmp155 = tmp139 * tmp154
tmp156 = tmp155 * tmp67
tmp157 = tmp151 + tmp156
tmp158 = tmp72 == tmp140
tmp159 = tmp158.to(tl.int64)
tmp160 = tmp159.to(tl.float32)
tmp161 = tmp139 * tmp160
tmp162 = tmp161 * tmp79
tmp163 = tmp157 + tmp162
tmp164 = tmp139 * tmp139
tmp165 = tmp142 * tmp142
tmp166 = tmp165.to(tl.float32)
tmp167 = tmp164 + tmp166
tmp168 = tmp147 * tmp147
tmp169 = tmp168.to(tl.float32)
tmp170 = tmp164 + tmp169
tmp171 = tmp167 + tmp170
tmp172 = tmp153 * tmp153
tmp173 = tmp172.to(tl.float32)
tmp174 = tmp164 + tmp173
tmp175 = tmp171 + tmp174
tmp176 = tmp159 * tmp159
tmp177 = tmp176.to(tl.float32)
tmp178 = tmp164 + tmp177
tmp179 = tmp175 + tmp178
tmp181 = tmp36 == tmp35
tmp182 = tmp181.to(tl.int64)
tmp183 = tmp182.to(tl.float32)
tmp184 = tmp180 * tmp183
tmp185 = tmp184 * tmp44
tmp186 = tmp48 == tmp35
tmp187 = tmp186.to(tl.int64)
tmp188 = tmp187.to(tl.float32)
tmp189 = tmp180 * tmp188
tmp190 = tmp189 * tmp55
tmp191 = tmp185 + tmp190
tmp192 = tmp60 == tmp35
tmp193 = tmp192.to(tl.int64)
tmp194 = tmp193.to(tl.float32)
tmp195 = tmp180 * tmp194
tmp196 = tmp195 * tmp67
tmp197 = tmp191 + tmp196
tmp198 = tmp72 == tmp35
tmp199 = tmp198.to(tl.int64)
tmp200 = tmp199.to(tl.float32)
tmp201 = tmp180 * tmp200
tmp202 = tmp201 * tmp79
tmp203 = tmp197 + tmp202
tmp204 = tmp180 * tmp180
tmp205 = tmp182 * tmp182
tmp206 = tmp205.to(tl.float32)
tmp207 = tmp204 + tmp206
tmp208 = tmp187 * tmp187
tmp209 = tmp208.to(tl.float32)
tmp210 = tmp204 + tmp209
tmp211 = tmp207 + tmp210
tmp212 = tmp193 * tmp193
tmp213 = tmp212.to(tl.float32)
tmp214 = tmp204 + tmp213
tmp215 = tmp211 + tmp214
tmp216 = tmp199 * tmp199
tmp217 = tmp216.to(tl.float32)
tmp218 = tmp204 + tmp217
tmp219 = tmp215 + tmp218
tmp220 = 2.0
tmp221 = tmp81 * tmp220
tmp222 = 1.0
tmp223 = tmp221 + tmp222
tmp224 = tmp97 + tmp222
tmp225 = tmp223 / tmp224
tmp226 = tmp222 - tmp225
tmp227 = tl.broadcast_to(tmp226, [XBLOCK, RBLOCK])
tmp229 = tl.sum(tmp227, 1)[:, None]
tmp230 = tmp122 * tmp220
tmp231 = tmp230 + tmp222
tmp232 = tmp138 + tmp222
tmp233 = tmp231 / tmp232
tmp234 = tmp222 - tmp233
tmp235 = tl.broadcast_to(tmp234, [XBLOCK, RBLOCK])
tmp237 = tl.sum(tmp235, 1)[:, None]
tmp238 = tmp163 * tmp220
tmp239 = tmp238 + tmp222
tmp240 = tmp179 + tmp222
tmp241 = tmp239 / tmp240
tmp242 = tmp222 - tmp241
tmp243 = tl.broadcast_to(tmp242, [XBLOCK, RBLOCK])
tmp245 = tl.sum(tmp243, 1)[:, None]
tmp246 = tmp203 * tmp220
tmp247 = tmp246 + tmp222
tmp248 = tmp219 + tmp222
tmp249 = tmp247 / tmp248
tmp250 = tmp222 - tmp249
tmp251 = tl.broadcast_to(tmp250, [XBLOCK, RBLOCK])
tmp253 = tl.sum(tmp251, 1)[:, None]
tmp254 = 4.0
tmp255 = tmp30 / tmp254
tmp256 = tmp255 * tmp222
tmp257 = 0.5
tmp258 = tmp256 * tmp257
tmp259 = tmp229 / tmp254
tmp260 = 0.0
tmp261 = tmp259 + tmp260
tmp262 = tmp237 / tmp254
tmp263 = tmp261 + tmp262
tmp264 = tmp245 / tmp254
tmp265 = tmp263 + tmp264
tmp266 = tmp253 / tmp254
tmp267 = tmp265 + tmp266
tmp268 = 0.25
tmp269 = tmp267 * tmp268
tmp270 = tmp269 / tmp222
tmp271 = tmp270 * tmp222
tmp272 = tmp271 * tmp257
tmp273 = tmp258 + tmp272
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp273, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4), (4, 1))
assert_size_stride(arg1_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax__softmax_0[grid(16)](arg0_1, buf0,
buf2, 16, XBLOCK=16, num_warps=1, num_stages=1)
del arg0_1
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(16)](buf2, buf3, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del buf2
buf1 = empty_strided_cuda((), (), torch.float32)
buf16 = buf1
del buf1
triton_per_fused__log_softmax__to_copy_add_div_mean_mul_ne_neg_pow_rsub_sum_view_2[
grid(1)](buf16, buf0, arg1_1, buf3, 1, 4, XBLOCK=1, num_warps=2,
num_stages=1)
del arg1_1
del buf0
del buf3
return buf16,
def _expand_onehot_labels(labels, label_weights, target_shape, ignore_index):
"""Expand onehot labels to match the size of prediction."""
bin_labels = labels.new_zeros(target_shape)
valid_mask = (labels >= 0) & (labels != ignore_index)
inds = torch.nonzero(valid_mask, as_tuple=True)
if inds[0].numel() > 0:
if labels.dim() == 3:
bin_labels[inds[0], labels[valid_mask], inds[1], inds[2]] = 1
else:
bin_labels[inds[0], labels[valid_mask]] = 1
valid_mask = valid_mask.unsqueeze(1).expand(target_shape).float()
if label_weights is None:
bin_label_weights = valid_mask
else:
bin_label_weights = label_weights.unsqueeze(1).expand(target_shape)
bin_label_weights *= valid_mask
return bin_labels, bin_label_weights
def reduce_loss(loss, reduction):
"""Reduce loss as specified.
Args:
loss (Tensor): Elementwise loss tensor.
reduction (str): Options are "none", "mean" and "sum".
Return:
Tensor: Reduced loss tensor.
"""
reduction_enum = F._Reduction.get_enum(reduction)
if reduction_enum == 0:
return loss
elif reduction_enum == 1:
return loss.mean()
elif reduction_enum == 2:
return loss.sum()
def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
"""Apply element-wise weight and reduce loss.
Args:
loss (Tensor): Element-wise loss.
weight (Tensor): Element-wise weights.
reduction (str): Same as built-in losses of PyTorch.
avg_factor (float): Avarage factor when computing the mean of losses.
Returns:
Tensor: Processed loss values.
"""
if weight is not None:
assert weight.dim() == loss.dim()
if weight.dim() > 1:
assert weight.size(1) == 1 or weight.size(1) == loss.size(1)
loss = loss * weight
if avg_factor is None:
loss = reduce_loss(loss, reduction)
elif reduction == 'mean':
loss = loss.sum() / avg_factor
elif reduction != 'none':
raise ValueError('avg_factor can not be used with reduction="sum"')
return loss
def binary_cross_entropy(pred, label, weight=None, reduction='mean',
avg_factor=None, class_weight=None, ignore_index=255):
"""Calculate the binary CrossEntropy loss.
Args:
pred (torch.Tensor): The prediction with shape (N, 1).
label (torch.Tensor): The learning label of the prediction.
weight (torch.Tensor, optional): Sample-wise loss weight.
reduction (str, optional): The method used to reduce the loss.
Options are "none", "mean" and "sum".
avg_factor (int, optional): Average factor that is used to average
the loss. Defaults to None.
class_weight (list[float], optional): The weight for each class.
ignore_index (int | None): The label index to be ignored. Default: 255
Returns:
torch.Tensor: The calculated loss
"""
if pred.dim() != label.dim():
assert pred.dim() == 2 and label.dim() == 1 or pred.dim(
) == 4 and label.dim(
) == 3, 'Only pred shape [N, C], label shape [N] or pred shape [N, C, H, W], label shape [N, H, W] are supported'
label, weight = _expand_onehot_labels(label, weight, pred.shape,
ignore_index)
if weight is not None:
weight = weight.float()
loss = F.binary_cross_entropy_with_logits(pred, label.float(),
pos_weight=class_weight, reduction='none')
loss = weight_reduce_loss(loss, weight, reduction=reduction, avg_factor
=avg_factor)
return loss
def cross_entropy(pred, label, weight=None, class_weight=None, reduction=
'mean', avg_factor=None, ignore_index=-100):
"""The wrapper function for :func:`F.cross_entropy`"""
loss = F.cross_entropy(pred, label, weight=class_weight, reduction=
'none', ignore_index=ignore_index)
if weight is not None:
weight = weight.float()
loss = weight_reduce_loss(loss, weight=weight, reduction=reduction,
avg_factor=avg_factor)
return loss
def weighted_loss(loss_func):
"""Create a weighted version of a given loss function.
To use this decorator, the loss function must have the signature like
`loss_func(pred, target, **kwargs)`. The function only needs to compute
element-wise loss without any reduction. This decorator will add weight
and reduction arguments to the function. The decorated function will have
the signature like `loss_func(pred, target, weight=None, reduction='mean',
avg_factor=None, **kwargs)`.
:Example:
>>> import torch
>>> @weighted_loss
>>> def l1_loss(pred, target):
>>> return (pred - target).abs()
>>> pred = torch.Tensor([0, 2, 3])
>>> target = torch.Tensor([1, 1, 1])
>>> weight = torch.Tensor([1, 0, 1])
>>> l1_loss(pred, target)
tensor(1.3333)
>>> l1_loss(pred, target, weight)
tensor(1.)
>>> l1_loss(pred, target, reduction='none')
tensor([1., 1., 2.])
>>> l1_loss(pred, target, weight, avg_factor=2)
tensor(1.5000)
"""
@functools.wraps(loss_func)
def wrapper(pred, target, weight=None, reduction='mean', avg_factor=
None, **kwargs):
loss = loss_func(pred, target, **kwargs)
loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
return loss
return wrapper
@weighted_loss
def binary_dice_loss(pred, target, valid_mask, smooth=1, exponent=2, **kwards):
assert pred.shape[0] == target.shape[0]
pred = pred.reshape(pred.shape[0], -1)
target = target.reshape(target.shape[0], -1)
valid_mask = valid_mask.reshape(valid_mask.shape[0], -1)
num = torch.sum(torch.mul(pred, target) * valid_mask, dim=1) * 2 + smooth
den = torch.sum(pred.pow(exponent) + target.pow(exponent), dim=1) + smooth
return 1 - num / den
@weighted_loss
def dice_loss(pred, target, valid_mask, smooth=1, exponent=2, class_weight=
None, ignore_index=255):
assert pred.shape[0] == target.shape[0]
total_loss = 0
num_classes = pred.shape[1]
for i in range(num_classes):
if i != ignore_index:
dice_loss = binary_dice_loss(pred[:, i], target[..., i],
valid_mask=valid_mask, smooth=smooth, exponent=exponent)
if class_weight is not None:
dice_loss *= class_weight[i]
total_loss += dice_loss
return total_loss / num_classes
def get_class_weight(class_weight):
"""Get class weight for loss function.
Args:
class_weight (list[float] | str | None): If class_weight is a str,
take it as a file name and read from it.
"""
if isinstance(class_weight, str):
if class_weight.endswith('.npy'):
class_weight = np.load(class_weight)
else:
class_weight = mmcv.load(class_weight)
return class_weight
def mask_cross_entropy(pred, target, label, reduction='mean', avg_factor=
None, class_weight=None, ignore_index=None):
"""Calculate the CrossEntropy loss for masks.
Args:
pred (torch.Tensor): The prediction with shape (N, C), C is the number
of classes.
target (torch.Tensor): The learning label of the prediction.
label (torch.Tensor): ``label`` indicates the class label of the mask'
corresponding object. This will be used to select the mask in the
of the class which the object belongs to when the mask prediction
if not class-agnostic.
reduction (str, optional): The method used to reduce the loss.
Options are "none", "mean" and "sum".
avg_factor (int, optional): Average factor that is used to average
the loss. Defaults to None.
class_weight (list[float], optional): The weight for each class.
ignore_index (None): Placeholder, to be consistent with other loss.
Default: None.
Returns:
torch.Tensor: The calculated loss
"""
assert ignore_index is None, 'BCE loss does not support ignore_index'
assert reduction == 'mean' and avg_factor is None
num_rois = pred.size()[0]
inds = torch.arange(0, num_rois, dtype=torch.long, device=pred.device)
pred_slice = pred[inds, label].squeeze(1)
return F.binary_cross_entropy_with_logits(pred_slice, target, weight=
class_weight, reduction='mean')[None]
class BCEDiceLossNew(nn.Module):
"""CrossEntropyLoss.
Args:
use_sigmoid (bool, optional): Whether the prediction uses sigmoid
of softmax. Defaults to False.
use_mask (bool, optional): Whether to use mask cross entropy loss.
Defaults to False.
reduction (str, optional): . Defaults to 'mean'.
Options are "none", "mean" and "sum".
class_weight (list[float] | str, optional): Weight of each class. If in
str format, read them from a file. Defaults to None.
loss_weight (float, optional): Weight of the loss. Defaults to 1.0.
"""
def __init__(self, use_sigmoid=False, use_mask=False, smooth=1,
exponent=2, reduction='mean', class_weight=None, loss_weight=1.0,
ignore_index=255, bce_weight=0.5, **kwards):
super(BCEDiceLossNew, self).__init__()
assert use_sigmoid is False or use_mask is False
self.use_sigmoid = use_sigmoid
self.use_mask = use_mask
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.class_weight = get_class_weight(class_weight)
self.loss_weight = loss_weight
self.ignore_index = ignore_index
self.bce_weight = bce_weight
if self.use_sigmoid:
self.cls_criterion = binary_cross_entropy
elif self.use_mask:
self.cls_criterion = mask_cross_entropy
else:
self.cls_criterion = cross_entropy
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| SeHwanJoo/mmsegmentation_body | BCEDiceLoss | false | 1,049 | [
"Apache-2.0"
] | 0 | 31c4bf27c3dc0a84bfbb06a0c017c5908c17f0ac | https://github.com/SeHwanJoo/mmsegmentation_body/tree/31c4bf27c3dc0a84bfbb06a0c017c5908c17f0ac | import functools
import torch
import numpy as np
import torch.nn.functional as F
import torch.nn as nn
import torch._C
import torch.serialization
def _expand_onehot_labels(labels, label_weights, target_shape, ignore_index):
"""Expand onehot labels to match the size of prediction."""
bin_labels = labels.new_zeros(target_shape)
valid_mask = (labels >= 0) & (labels != ignore_index)
inds = torch.nonzero(valid_mask, as_tuple=True)
if inds[0].numel() > 0:
if labels.dim() == 3:
bin_labels[inds[0], labels[valid_mask], inds[1], inds[2]] = 1
else:
bin_labels[inds[0], labels[valid_mask]] = 1
valid_mask = valid_mask.unsqueeze(1).expand(target_shape).float()
if label_weights is None:
bin_label_weights = valid_mask
else:
bin_label_weights = label_weights.unsqueeze(1).expand(target_shape)
bin_label_weights *= valid_mask
return bin_labels, bin_label_weights
def reduce_loss(loss, reduction):
"""Reduce loss as specified.
Args:
loss (Tensor): Elementwise loss tensor.
reduction (str): Options are "none", "mean" and "sum".
Return:
Tensor: Reduced loss tensor.
"""
reduction_enum = F._Reduction.get_enum(reduction)
if reduction_enum == 0:
return loss
elif reduction_enum == 1:
return loss.mean()
elif reduction_enum == 2:
return loss.sum()
def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
"""Apply element-wise weight and reduce loss.
Args:
loss (Tensor): Element-wise loss.
weight (Tensor): Element-wise weights.
reduction (str): Same as built-in losses of PyTorch.
avg_factor (float): Avarage factor when computing the mean of losses.
Returns:
Tensor: Processed loss values.
"""
if weight is not None:
assert weight.dim() == loss.dim()
if weight.dim() > 1:
assert weight.size(1) == 1 or weight.size(1) == loss.size(1)
loss = loss * weight
if avg_factor is None:
loss = reduce_loss(loss, reduction)
elif reduction == 'mean':
loss = loss.sum() / avg_factor
elif reduction != 'none':
raise ValueError('avg_factor can not be used with reduction="sum"')
return loss
def binary_cross_entropy(pred, label, weight=None, reduction='mean',
avg_factor=None, class_weight=None, ignore_index=255):
"""Calculate the binary CrossEntropy loss.
Args:
pred (torch.Tensor): The prediction with shape (N, 1).
label (torch.Tensor): The learning label of the prediction.
weight (torch.Tensor, optional): Sample-wise loss weight.
reduction (str, optional): The method used to reduce the loss.
Options are "none", "mean" and "sum".
avg_factor (int, optional): Average factor that is used to average
the loss. Defaults to None.
class_weight (list[float], optional): The weight for each class.
ignore_index (int | None): The label index to be ignored. Default: 255
Returns:
torch.Tensor: The calculated loss
"""
if pred.dim() != label.dim():
assert pred.dim() == 2 and label.dim() == 1 or pred.dim(
) == 4 and label.dim(
) == 3, 'Only pred shape [N, C], label shape [N] or pred shape [N, C, H, W], label shape [N, H, W] are supported'
label, weight = _expand_onehot_labels(label, weight, pred.shape,
ignore_index)
if weight is not None:
weight = weight.float()
loss = F.binary_cross_entropy_with_logits(pred, label.float(),
pos_weight=class_weight, reduction='none')
loss = weight_reduce_loss(loss, weight, reduction=reduction, avg_factor
=avg_factor)
return loss
def cross_entropy(pred, label, weight=None, class_weight=None, reduction=
'mean', avg_factor=None, ignore_index=-100):
"""The wrapper function for :func:`F.cross_entropy`"""
loss = F.cross_
# ... truncated (>4000 chars) for memory efficiency |
BasicBlock | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/fv/cfvfxpmpur3qlmurffwz4u56tgvw75i4lbjvzd25ortunbobyxnh.py
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten._native_batch_norm_legit, aten.relu]
# Source node to ATen node mapping:
# out_1 => add, rsqrt, var_mean
# out_2 => relu
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view, [0, 2, 3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %relu : [num_users=1] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
triton_per_fused__native_batch_norm_legit_relu_0 = async_compile.triton('triton_per_fused__native_batch_norm_legit_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__native_batch_norm_legit_relu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__native_batch_norm_legit_relu_0(in_ptr0, out_ptr0, out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp24 = tl.full([1, 1], 0, tl.int32)
tmp25 = triton_helpers.maximum(tmp24, tmp23)
tl.store(out_ptr2 + (r1 + (16*x0)), tmp25, xmask)
tl.store(out_ptr3 + (x0), tmp22, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/lk/clk53igo2wowtd6pq5zu23svybsi67ef5dmegrb3qnpketcev22z.py
# Topologically Sorted Source Nodes: [out_4, out_6], Original ATen: [aten._native_batch_norm_legit, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# out_4 => add_1, rsqrt_1, var_mean_1
# out_6 => relu_1
# Graph fragment:
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_5, [0, 2, 3]), kwargs = {correction: 0, keepdim: True})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-05), kwargs = {})
# %rsqrt_1 : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_1,), kwargs = {})
# %relu_1 : [num_users=1] = call_function[target=torch.ops.aten.relu.default](args = (%view_8,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%view_16, 0), kwargs = {})
triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1 = async_compile.triton('triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: '*fp32', 6: 'i32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1(in_ptr0, in_ptr1, out_ptr0, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp25 = tmp23 + tmp24
tmp26 = tl.full([1, 1], 0, tl.int32)
tmp27 = triton_helpers.maximum(tmp26, tmp25)
tmp28 = 0.0
tmp29 = tmp27 <= tmp28
tl.store(out_ptr2 + (r1 + (16*x0)), tmp27, xmask)
tl.store(out_ptr3 + (r1 + (16*x0)), tmp29, xmask)
tl.store(out_ptr4 + (x0), tmp22, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4, 4, 3, 3), (36, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf4 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten._native_batch_norm_legit, aten.relu]
stream0 = get_raw_stream(0)
triton_per_fused__native_batch_norm_legit_relu_0.run(buf0, buf1, buf5, buf4, 16, 16, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 4, 4, 4), (64, 16, 4, 1))
buf7 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
buf11 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf12 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
buf10 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [out_4, out_6], Original ATen: [aten._native_batch_norm_legit, aten.relu, aten.threshold_backward]
triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1.run(buf6, primals_1, buf7, buf11, buf12, buf10, 16, 16, grid=grid(16), stream=stream0)
return (buf11, primals_1, primals_2, primals_3, buf0, reinterpret_tensor(buf4, (16, ), (1, ), 0), buf5, buf6, reinterpret_tensor(buf10, (16, ), (1, ), 0), buf12, reinterpret_tensor(buf7, (1, 16, 1, 1), (16, 1, 1, 1), 0), reinterpret_tensor(buf1, (1, 16, 1, 1), (16, 1, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
def conv1x1(in_planes, out_planes, stride=1):
"""1x1 convolution"""
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride,
bias=False)
def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
class BasicBlock(nn.Module):
expansion = 1
__constants__ = ['downsample']
def __init__(self, inplanes, planes, stride=1, downsample=None, groups=
1, base_width=64, dilation=1, norm_model='instance'):
super(BasicBlock, self).__init__()
if 'instance' == norm_model:
norm_layer = nn.InstanceNorm2d
else:
norm_layer = nn.BatchNorm2d
if groups != 1 or base_width != 64:
raise ValueError(
'BasicBlock only supports groups=1 and base_width=64')
if dilation > 1:
raise NotImplementedError(
'Dilation > 1 not supported in BasicBlock')
self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = norm_layer(planes)
self.relu = nn.ReLU(inplace=True)
self.conv2 = conv3x3(planes, planes)
self.bn2 = norm_layer(planes)
self.stride = stride
if stride != 1 or inplanes != planes:
self.downsample = nn.Sequential(conv1x1(inplanes, planes,
stride), norm_layer(planes))
else:
self.downsample = downsample
def forward(self, x):
identity = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
if self.downsample is not None:
identity = self.downsample(x)
out += identity
out = self.relu(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'inplanes': 4, 'planes': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused__native_batch_norm_legit_relu_0(in_ptr0, out_ptr0,
out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp24 = tl.full([1, 1], 0, tl.int32)
tmp25 = triton_helpers.maximum(tmp24, tmp23)
tl.store(out_ptr2 + (r1 + 16 * x0), tmp25, xmask)
tl.store(out_ptr3 + x0, tmp22, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
@triton.jit
def triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1(in_ptr0
, in_ptr1, out_ptr0, out_ptr2, out_ptr3, out_ptr4, xnumel, rnumel,
XBLOCK: tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = tmp0 - tmp10
tmp18 = 16.0
tmp19 = tmp16 / tmp18
tmp20 = 1e-05
tmp21 = tmp19 + tmp20
tmp22 = libdevice.rsqrt(tmp21)
tmp23 = tmp17 * tmp22
tmp25 = tmp23 + tmp24
tmp26 = tl.full([1, 1], 0, tl.int32)
tmp27 = triton_helpers.maximum(tmp26, tmp25)
tmp28 = 0.0
tmp29 = tmp27 <= tmp28
tl.store(out_ptr2 + (r1 + 16 * x0), tmp27, xmask)
tl.store(out_ptr3 + (r1 + 16 * x0), tmp29, xmask)
tl.store(out_ptr4 + x0, tmp22, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4, 4, 3, 3), (36, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_1, primals_2, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32
)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf4 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32
)
get_raw_stream(0)
triton_per_fused__native_batch_norm_legit_relu_0[grid(16)](buf0,
buf1, buf5, buf4, 16, 16, XBLOCK=8, num_warps=2, num_stages=1)
buf6 = extern_kernels.convolution(buf5, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 4, 4, 4), (64, 16, 4, 1))
buf7 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32
)
buf11 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf12 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
buf10 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.
float32)
triton_per_fused__native_batch_norm_legit_relu_threshold_backward_1[
grid(16)](buf6, primals_1, buf7, buf11, buf12, buf10, 16, 16,
XBLOCK=8, num_warps=2, num_stages=1)
return buf11, primals_1, primals_2, primals_3, buf0, reinterpret_tensor(
buf4, (16,), (1,), 0), buf5, buf6, reinterpret_tensor(buf10, (16,),
(1,), 0), buf12, reinterpret_tensor(buf7, (1, 16, 1, 1), (16, 1, 1,
1), 0), reinterpret_tensor(buf1, (1, 16, 1, 1), (16, 1, 1, 1), 0)
def conv1x1(in_planes, out_planes, stride=1):
"""1x1 convolution"""
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride,
bias=False)
def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
class BasicBlockNew(nn.Module):
expansion = 1
__constants__ = ['downsample']
def __init__(self, inplanes, planes, stride=1, downsample=None, groups=
1, base_width=64, dilation=1, norm_model='instance'):
super(BasicBlockNew, self).__init__()
if 'instance' == norm_model:
norm_layer = nn.InstanceNorm2d
else:
norm_layer = nn.BatchNorm2d
if groups != 1 or base_width != 64:
raise ValueError(
'BasicBlock only supports groups=1 and base_width=64')
if dilation > 1:
raise NotImplementedError(
'Dilation > 1 not supported in BasicBlock')
self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = norm_layer(planes)
self.relu = nn.ReLU(inplace=True)
self.conv2 = conv3x3(planes, planes)
self.bn2 = norm_layer(planes)
self.stride = stride
if stride != 1 or inplanes != planes:
self.downsample = nn.Sequential(conv1x1(inplanes, planes,
stride), norm_layer(planes))
else:
self.downsample = downsample
def forward(self, input_0):
primals_2 = self.conv1.weight
primals_3 = self.conv2.weight
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| SebyakinAndrei/MichiGAN | BasicBlock | false | 1,050 | [
"MIT"
] | 0 | 6584c9a106b33096f38e8f5b11d0320f7065fd26 | https://github.com/SebyakinAndrei/MichiGAN/tree/6584c9a106b33096f38e8f5b11d0320f7065fd26 | import torch
import torch.nn as nn
import torch.utils.data
def conv1x1(in_planes, out_planes, stride=1):
"""1x1 convolution"""
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride,
bias=False)
def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=dilation, groups=groups, bias=False, dilation=dilation)
class Model(nn.Module):
expansion = 1
__constants__ = ['downsample']
def __init__(self, inplanes, planes, stride=1, downsample=None, groups=
1, base_width=64, dilation=1, norm_model='instance'):
super().__init__()
if 'instance' == norm_model:
norm_layer = nn.InstanceNorm2d
else:
norm_layer = nn.BatchNorm2d
if groups != 1 or base_width != 64:
raise ValueError(
'BasicBlock only supports groups=1 and base_width=64')
if dilation > 1:
raise NotImplementedError(
'Dilation > 1 not supported in BasicBlock')
self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = norm_layer(planes)
self.relu = nn.ReLU(inplace=True)
self.conv2 = conv3x3(planes, planes)
self.bn2 = norm_layer(planes)
self.stride = stride
if stride != 1 or inplanes != planes:
self.downsample = nn.Sequential(conv1x1(inplanes, planes,
stride), norm_layer(planes))
else:
self.downsample = downsample
def forward(self, x):
identity = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.conv2(out)
out = self.bn2(out)
if self.downsample is not None:
identity = self.downsample(x)
out += identity
out = self.relu(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
BahdanauAttention | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/ao/caoovxtqrx42gvkmjirowqmmbh6kppvfh5ebrzzv4kzkgwm2umii.py
# Topologically Sorted Source Nodes: [processed_query], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# processed_query => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_1,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2) + (16*x1)), xmask)
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/bg/cbgmsaps4ljzc6rkbd4imsj3jo73tgvkd46dy7obklnnvintmaea.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mv]
# Source node to ATen node mapping:
# out => mul, sum_1
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_4, %primals_5), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
triton_poi_fused_mv_1 = async_compile.triton('triton_poi_fused_mv_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mv_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mv_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*(x0 // 4)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + ((4*(x0 % 4)) + (16*(x0 // 16))), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + (0))
tmp5 = tl.broadcast_to(tmp4, [XBLOCK])
tmp7 = tl.load(in_ptr0 + (1 + (4*(x0 // 4))), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (1 + (4*(x0 % 4)) + (16*(x0 // 16))), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + (1))
tmp12 = tl.broadcast_to(tmp11, [XBLOCK])
tmp15 = tl.load(in_ptr0 + (2 + (4*(x0 // 4))), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr1 + (2 + (4*(x0 % 4)) + (16*(x0 // 16))), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + (2))
tmp20 = tl.broadcast_to(tmp19, [XBLOCK])
tmp23 = tl.load(in_ptr0 + (3 + (4*(x0 // 4))), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr1 + (3 + (4*(x0 % 4)) + (16*(x0 // 16))), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + (3))
tmp28 = tl.broadcast_to(tmp27, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp3 = libdevice.tanh(tmp2)
tmp6 = tmp3 * tmp5
tmp9 = tmp7 + tmp8
tmp10 = libdevice.tanh(tmp9)
tmp13 = tmp10 * tmp12
tmp14 = tmp6 + tmp13
tmp17 = tmp15 + tmp16
tmp18 = libdevice.tanh(tmp17)
tmp21 = tmp18 * tmp20
tmp22 = tmp14 + tmp21
tmp25 = tmp23 + tmp24
tmp26 = libdevice.tanh(tmp25)
tmp29 = tmp26 * tmp28
tmp30 = tmp22 + tmp29
tl.store(out_ptr0 + (x0), tmp30, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/hg/chg3iq6bscxmmxv5f7tuzgwycb4mgrimwfhv2nauw5rj4tt5cmv2.py
# Topologically Sorted Source Nodes: [scores_normalized], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# scores_normalized => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_5, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_5, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/zu/czuvep3dmpmqmhiiliwubh4ghdt2qr27va67sszkua7trziinwov.py
# Topologically Sorted Source Nodes: [scores_normalized], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# scores_normalized => div, sum_2
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div : [num_users=3] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_2), kwargs = {})
triton_poi_fused__softmax_3 = async_compile.triton('triton_poi_fused__softmax_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [processed_query], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(primals_2, buf0, 64, grid=grid(64), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [processed_query], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
del primals_3
buf2 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [processed_key], Original ATen: [aten.clone]
triton_poi_fused_clone_0.run(primals_1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [processed_key], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf3)
del primals_4
buf4 = empty_strided_cuda((64, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mv]
triton_poi_fused_mv_1.run(buf1, buf3, primals_5, buf4, 64, grid=grid(64), stream=stream0)
buf5 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [scores_normalized], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf4, buf5, 64, grid=grid(64), stream=stream0)
buf6 = reinterpret_tensor(buf4, (4, 4, 4), (16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [scores_normalized], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf5, buf6, 64, grid=grid(64), stream=stream0)
buf7 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [context], Original ATen: [aten.bmm]
extern_kernels.bmm(buf6, reinterpret_tensor(primals_1, (4, 4, 4), (4, 16, 1), 0), out=buf7)
return (reinterpret_tensor(buf7, (4, 4, 4), (4, 16, 1), 0), reinterpret_tensor(buf6, (4, 4, 4), (4, 16, 1), 0), primals_5, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), buf1, reinterpret_tensor(buf2, (16, 4), (4, 1), 0), buf3, buf6, reinterpret_tensor(primals_1, (4, 4, 4), (4, 1, 16), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn.functional as F
from torch import nn
from torch.nn.parameter import Parameter
import torch.nn.parallel
import torch.utils.data
import torch.onnx
import torch.optim
import torch.utils.data.distributed
class BahdanauAttention(nn.Module):
"""
It should be very similar to tf.contrib.seq2seq.BahdanauAttention
"""
def __init__(self, query_size, key_size, num_units, normalize=False,
dropout=0, batch_first=False):
super(BahdanauAttention, self).__init__()
self.normalize = normalize
self.batch_first = batch_first
self.num_units = num_units
self.linear_q = nn.Linear(query_size, num_units, bias=False)
self.linear_k = nn.Linear(key_size, num_units, bias=False)
self.linear_att = Parameter(torch.Tensor(num_units))
self.dropout = nn.Dropout(dropout)
self.mask = None
if self.normalize:
self.normalize_scalar = Parameter(torch.Tensor(1))
self.normalize_bias = Parameter(torch.Tensor(num_units))
else:
self.register_parameter('normalize_scalar', None)
self.register_parameter('normalize_bias', None)
self.reset_parameters()
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.num_units)
self.linear_att.data.uniform_(-stdv, stdv)
if self.normalize:
self.normalize_scalar.data.fill_(stdv)
self.normalize_bias.data.zero_()
def set_mask(self, context_len, context):
"""
sets self.mask which is applied before softmax
ones for inactive context fields, zeros for active context fields
:param context_len: b
:param context: if batch_first: (b x t_k x n) else: (t_k x b x n)
self.mask: (b x t_k)
"""
if self.batch_first:
max_len = context.size(1)
else:
max_len = context.size(0)
indices = torch.arange(0, max_len, dtype=torch.int64, device=
context.device)
self.mask = indices >= context_len.unsqueeze(1)
def calc_score(self, att_query, att_keys):
"""
Calculate Bahdanau score
:param att_query: b x t_q x n
:param att_keys: b x t_k x n
return b x t_q x t_k scores
"""
b, t_k, n = att_keys.size()
t_q = att_query.size(1)
att_query = att_query.unsqueeze(2).expand(b, t_q, t_k, n)
att_keys = att_keys.unsqueeze(1).expand(b, t_q, t_k, n)
sum_qk = att_query + att_keys
if self.normalize:
sum_qk = sum_qk + self.normalize_bias
tmp = self.linear_att
linear_att = tmp / tmp.norm()
linear_att = linear_att
linear_att = linear_att * self.normalize_scalar
else:
linear_att = self.linear_att
out = F.tanh(sum_qk).matmul(linear_att)
return out
def forward(self, query, keys):
"""
:param query: if batch_first: (b x t_q x n) else: (t_q x b x n)
:param keys: if batch_first: (b x t_k x n) else (t_k x b x n)
:returns: (context, scores_normalized)
context: if batch_first: (b x t_q x n) else (t_q x b x n)
scores_normalized: if batch_first (b x t_q x t_k) else (t_q x b x t_k)
"""
if not self.batch_first:
keys = keys.transpose(0, 1)
if query.dim() == 3:
query = query.transpose(0, 1)
if query.dim() == 2:
single_query = True
query = query.unsqueeze(1)
else:
single_query = False
b = query.size(0)
t_k = keys.size(1)
t_q = query.size(1)
processed_query = self.linear_q(query)
processed_key = self.linear_k(keys)
scores = self.calc_score(processed_query, processed_key)
if self.mask is not None:
mask = self.mask.unsqueeze(1).expand(b, t_q, t_k)
scores.data.masked_fill_(mask, -65504.0)
scores_normalized = F.softmax(scores, dim=-1)
scores_normalized = self.dropout(scores_normalized)
context = torch.bmm(scores_normalized, keys)
if single_query:
context = context.squeeze(1)
scores_normalized = scores_normalized.squeeze(1)
elif not self.batch_first:
context = context.transpose(0, 1)
scores_normalized = scores_normalized.transpose(0, 1)
return context, scores_normalized
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'query_size': 4, 'key_size': 4, 'num_units': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import math
import torch.nn.functional as F
from torch import nn
from torch.nn.parameter import Parameter
import torch.nn.parallel
import torch.utils.data
import torch.onnx
import torch.optim
import torch.utils.data.distributed
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2 + 16 * x1), xmask)
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_mv_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * (x0 // 4), xmask, eviction_policy='evict_last'
)
tmp1 = tl.load(in_ptr1 + (4 * (x0 % 4) + 16 * (x0 // 16)), xmask,
eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + 0)
tmp5 = tl.broadcast_to(tmp4, [XBLOCK])
tmp7 = tl.load(in_ptr0 + (1 + 4 * (x0 // 4)), xmask, eviction_policy=
'evict_last')
tmp8 = tl.load(in_ptr1 + (1 + 4 * (x0 % 4) + 16 * (x0 // 16)), xmask,
eviction_policy='evict_last')
tmp11 = tl.load(in_ptr2 + 1)
tmp12 = tl.broadcast_to(tmp11, [XBLOCK])
tmp15 = tl.load(in_ptr0 + (2 + 4 * (x0 // 4)), xmask, eviction_policy=
'evict_last')
tmp16 = tl.load(in_ptr1 + (2 + 4 * (x0 % 4) + 16 * (x0 // 16)), xmask,
eviction_policy='evict_last')
tmp19 = tl.load(in_ptr2 + 2)
tmp20 = tl.broadcast_to(tmp19, [XBLOCK])
tmp23 = tl.load(in_ptr0 + (3 + 4 * (x0 // 4)), xmask, eviction_policy=
'evict_last')
tmp24 = tl.load(in_ptr1 + (3 + 4 * (x0 % 4) + 16 * (x0 // 16)), xmask,
eviction_policy='evict_last')
tmp27 = tl.load(in_ptr2 + 3)
tmp28 = tl.broadcast_to(tmp27, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp3 = libdevice.tanh(tmp2)
tmp6 = tmp3 * tmp5
tmp9 = tmp7 + tmp8
tmp10 = libdevice.tanh(tmp9)
tmp13 = tmp10 * tmp12
tmp14 = tmp6 + tmp13
tmp17 = tmp15 + tmp16
tmp18 = libdevice.tanh(tmp17)
tmp21 = tmp18 * tmp20
tmp22 = tmp14 + tmp21
tmp25 = tmp23 + tmp24
tmp26 = libdevice.tanh(tmp25)
tmp29 = tmp26 * tmp28
tmp30 = tmp22 + tmp29
tl.store(out_ptr0 + x0, tmp30, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(64)](primals_2, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
del primals_3
buf2 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_clone_0[grid(64)](primals_1, buf2, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf2, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf3)
del primals_4
buf4 = empty_strided_cuda((64,), (1,), torch.float32)
triton_poi_fused_mv_1[grid(64)](buf1, buf3, primals_5, buf4, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_2[grid(64)](buf4, buf5, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf6 = reinterpret_tensor(buf4, (4, 4, 4), (16, 4, 1), 0)
del buf4
triton_poi_fused__softmax_3[grid(64)](buf5, buf6, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf7 = buf5
del buf5
extern_kernels.bmm(buf6, reinterpret_tensor(primals_1, (4, 4, 4), (
4, 16, 1), 0), out=buf7)
return reinterpret_tensor(buf7, (4, 4, 4), (4, 16, 1), 0
), reinterpret_tensor(buf6, (4, 4, 4), (4, 16, 1), 0
), primals_5, reinterpret_tensor(buf0, (16, 4), (4, 1), 0
), buf1, reinterpret_tensor(buf2, (16, 4), (4, 1), 0
), buf3, buf6, reinterpret_tensor(primals_1, (4, 4, 4), (4, 1, 16), 0)
class BahdanauAttentionNew(nn.Module):
"""
It should be very similar to tf.contrib.seq2seq.BahdanauAttention
"""
def __init__(self, query_size, key_size, num_units, normalize=False,
dropout=0, batch_first=False):
super(BahdanauAttentionNew, self).__init__()
self.normalize = normalize
self.batch_first = batch_first
self.num_units = num_units
self.linear_q = nn.Linear(query_size, num_units, bias=False)
self.linear_k = nn.Linear(key_size, num_units, bias=False)
self.linear_att = Parameter(torch.Tensor(num_units))
self.dropout = nn.Dropout(dropout)
self.mask = None
if self.normalize:
self.normalize_scalar = Parameter(torch.Tensor(1))
self.normalize_bias = Parameter(torch.Tensor(num_units))
else:
self.register_parameter('normalize_scalar', None)
self.register_parameter('normalize_bias', None)
self.reset_parameters()
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.num_units)
self.linear_att.data.uniform_(-stdv, stdv)
if self.normalize:
self.normalize_scalar.data.fill_(stdv)
self.normalize_bias.data.zero_()
def set_mask(self, context_len, context):
"""
sets self.mask which is applied before softmax
ones for inactive context fields, zeros for active context fields
:param context_len: b
:param context: if batch_first: (b x t_k x n) else: (t_k x b x n)
self.mask: (b x t_k)
"""
if self.batch_first:
max_len = context.size(1)
else:
max_len = context.size(0)
indices = torch.arange(0, max_len, dtype=torch.int64, device=
context.device)
self.mask = indices >= context_len.unsqueeze(1)
def calc_score(self, att_query, att_keys):
"""
Calculate Bahdanau score
:param att_query: b x t_q x n
:param att_keys: b x t_k x n
return b x t_q x t_k scores
"""
b, t_k, n = att_keys.size()
t_q = att_query.size(1)
att_query = att_query.unsqueeze(2).expand(b, t_q, t_k, n)
att_keys = att_keys.unsqueeze(1).expand(b, t_q, t_k, n)
sum_qk = att_query + att_keys
if self.normalize:
sum_qk = sum_qk + self.normalize_bias
tmp = self.linear_att
linear_att = tmp / tmp.norm()
linear_att = linear_att
linear_att = linear_att * self.normalize_scalar
else:
linear_att = self.linear_att
out = F.tanh(sum_qk).matmul(linear_att)
return out
def forward(self, input_0, input_1):
primals_5 = self.linear_att
primals_3 = self.linear_q.weight
primals_4 = self.linear_k.weight
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0], output[1]
| SerailHydra/examples | BahdanauAttention | false | 1,051 | [
"BSD-3-Clause"
] | 0 | 547226ff28032d4dab1dbf26e0b5f8b8276d79ae | https://github.com/SerailHydra/examples/tree/547226ff28032d4dab1dbf26e0b5f8b8276d79ae | import math
import torch
import torch.nn.functional as F
from torch import nn
from torch.nn.parameter import Parameter
import torch.nn.parallel
import torch.utils.data
import torch.onnx
import torch.optim
import torch.utils.data.distributed
class Model(nn.Module):
"""
It should be very similar to tf.contrib.seq2seq.BahdanauAttention
"""
def __init__(self, query_size, key_size, num_units, normalize=False,
dropout=0, batch_first=False):
super().__init__()
self.normalize = normalize
self.batch_first = batch_first
self.num_units = num_units
self.linear_q = nn.Linear(query_size, num_units, bias=False)
self.linear_k = nn.Linear(key_size, num_units, bias=False)
self.linear_att = Parameter(torch.Tensor(num_units))
self.dropout = nn.Dropout(dropout)
self.mask = None
if self.normalize:
self.normalize_scalar = Parameter(torch.Tensor(1))
self.normalize_bias = Parameter(torch.Tensor(num_units))
else:
self.register_parameter('normalize_scalar', None)
self.register_parameter('normalize_bias', None)
self.reset_parameters()
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.num_units)
self.linear_att.data.uniform_(-stdv, stdv)
if self.normalize:
self.normalize_scalar.data.fill_(stdv)
self.normalize_bias.data.zero_()
def set_mask(self, context_len, context):
"""
sets self.mask which is applied before softmax
ones for inactive context fields, zeros for active context fields
:param context_len: b
:param context: if batch_first: (b x t_k x n) else: (t_k x b x n)
self.mask: (b x t_k)
"""
if self.batch_first:
max_len = context.size(1)
else:
max_len = context.size(0)
indices = torch.arange(0, max_len, dtype=torch.int64, device=
context.device)
self.mask = indices >= context_len.unsqueeze(1)
def calc_score(self, att_query, att_keys):
"""
Calculate Bahdanau score
:param att_query: b x t_q x n
:param att_keys: b x t_k x n
return b x t_q x t_k scores
"""
b, t_k, n = att_keys.size()
t_q = att_query.size(1)
att_query = att_query.unsqueeze(2).expand(b, t_q, t_k, n)
att_keys = att_keys.unsqueeze(1).expand(b, t_q, t_k, n)
sum_qk = att_query + att_keys
if self.normalize:
sum_qk = sum_qk + self.normalize_bias
tmp = self.linear_att
linear_att = tmp / tmp.norm()
linear_att = linear_att
linear_att = linear_att * self.normalize_scalar
else:
linear_att = self.linear_att
out = F.tanh(sum_qk).matmul(linear_att)
return out
def forward(self, query, keys):
"""
:param query: if batch_first: (b x t_q x n) else: (t_q x b x n)
:param keys: if batch_first: (b x t_k x n) else (t_k x b x n)
:returns: (context, scores_normalized)
context: if batch_first: (b x t_q x n) else (t_q x b x n)
scores_normalized: if batch_first (b x t_q x t_k) else (t_q x b x t_k)
"""
if not self.batch_first:
keys = keys.transpose(0, 1)
if query.dim() == 3:
query = query.transpose(0, 1)
if query.dim() == 2:
single_query = True
query = query.unsqueeze(1)
else:
single_query = False
b = query.size(0)
t_k = keys.size(1)
t_q = query.size(1)
processed_query = self.linear_q(query)
processed_key = self.linear_k(keys)
scores = self.calc_score(processed_query, processed_key)
if self.mask is not None:
mask = self.mask.unsqueeze(1).expand(b, t_q, t_k)
scores.data.masked_fill_(mask, -65504.0)
scores_normalized = F.softmax
# ... truncated (>4000 chars) for memory efficiency |
StateActionEmbedding | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/t4/ct4abpxorr7ywpffawu2dyuam53y6icpnxs5d4cvirqopww6wnuh.py
# Topologically Sorted Source Nodes: [concatinated_batch], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# concatinated_batch => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%arg0_1, %arg1_1], 2), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4) % 8
x0 = xindex % 4
x2 = (xindex // 32)
x3 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (4*x1) + (16*x2)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr1 + (x0 + (4*((-4) + x1)) + (16*x2)), tmp6 & xmask, other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + (x3), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 4), (128, 32, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [concatinated_batch], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(arg0_1, arg1_1, buf0, 512, grid=grid(512), stream=stream0)
del arg0_1
del arg1_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import numpy as np
from abc import ABC
from abc import abstractmethod
from abc import abstractproperty
from torch import nn
from enum import Enum
def tensor_to_numpy(tensor):
return tensor.detach().cpu().numpy()
class MLPParamHandler(ABC):
def __init__(self) ->None:
"""Interface for parameter handler. For Algorithms that require data on past model parameters this module handels this data such that is it is in the right format according to the specific algorithm.
"""
super().__init__()
@abstractmethod
def get_policy_replay_data(self, model: 'torch.nn.Module'):
"""Function to extract replay data from a module in a format that it can be saved to the replay buffer.
Args:
model (torch.nn.Module): Module to extract data from.
"""
...
@abstractmethod
def get_policy_critic_data(self, model: 'torch.nn.Module'):
"""Function to extract data from a policy that the critic requires to evaluate it.
Args:
model (torch.nn.Module): Module to extract data from.
"""
...
@abstractmethod
def format_replay_buffer_data(self, **kwargs):
"""Function to format data from the replay buffer such that it can be used as input to the critic afterwards.
"""
...
@abstractproperty
def replay_data_keys(self):
"""Keys used to save data to the replay buffer.
"""
...
@abstractproperty
def replay_data_info(self):
"""Description of the data needed to initialize the replay buffer.
"""
...
class FlatParamHandler(MLPParamHandler):
def __init__(self, example_policy: 'torch.nn.Module') ->None:
"""Parameter handler that simply takes all parameters flattens them and saves them in one vector.
Args:
example_policy (torch.nn.Module): Example policy network to acquire shape of data.
"""
super().__init__()
self._replay_data_keys = [constants.DATA_PARAMETERS]
self._replay_data_info = {self._replay_data_keys[0]: {'shape': self
.get_policy_critic_data(example_policy).shape}}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
return {self._replay_data_keys[0]: tensor_to_numpy(torch.nn.utils.
parameters_to_vector(model.parameters())).reshape(1, -1)}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
return torch.nn.utils.parameters_to_vector(model.parameters()).reshape(
1, -1)
def format_replay_buffer_data(self, **kwargs):
return kwargs[self._replay_data_keys[0]]
@property
def replay_data_keys(self):
return self._replay_data_keys
@property
def replay_data_info(self):
return self._replay_data_info
class NamedParamHandler(MLPParamHandler):
def __init__(self, example_policy: 'torch.nn.Module') ->None:
"""Parameter handler that saves parameters in a dictionary shape such that the parameters are saved in a similar format of how they are used in the actual module. Useful if the Parameters are later reused similarly to how the are used within the module they are extracted from.
Args:
example_policy (torch.nn.Module): Example policy network to acquire structure of module and according dictionary.
"""
super().__init__()
actor_parameter_dict = self.get_policy_critic_data(example_policy)
self._replay_data_keys = actor_parameter_dict.keys()
self._replay_data_info = {key: {'shape': actor_parameter_dict[key].
shape[1:]} for key in self._replay_data_keys}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
batched_param_dict = self.get_policy_critic_data(model)
return {key: tensor_to_numpy(value) for key, value in
batched_param_dict.items()}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
param_dict = dict(model.named_parameters())
return {key: torch.unsqueeze(tensor, dim=0) for key, tensor in
param_dict.items()}
def format_replay_buffer_data(self, **kwargs):
return {key: kwargs[key] for key in self._replay_data_keys}
@property
def replay_data_info(self):
return self._replay_data_info
@property
def replay_data_keys(self):
return self._replay_data_keys
class StateActionHandler(MLPParamHandler):
def __init__(self, num_state_action_pairs: 'int', episode_length: 'int',
rollout_handler: 'RolloutHandler') ->None:
"""Parameter handler that does not actually use parameters but rather state action pairs as representation of policies.
Args:
num_state_action_pairs (int): Number of state action pairs used as a representation for a policy.
episode_length (int): Maximal time steps of a episode used for representation purposes.
rollout_handler (RolloutHandler): Rollout handler used to execute rollouts when needed.
"""
super().__init__()
self.rollout_handler = rollout_handler
self.num_state_action_pairs = num_state_action_pairs
self.episode_length = episode_length
self._replay_data_keys = [constants.DATA_OBSERVATIONS, constants.
DATA_ACTIONS]
self._replay_data_info = {constants.DATA_OBSERVATIONS: {'shape': (
episode_length, *rollout_handler.environment_handler.
exploration_environment.observation_space.shape)}, constants.
DATA_ACTIONS: {'shape': (episode_length, *rollout_handler.
environment_handler.exploration_environment.action_space.shape)}}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
return {}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
rollout_data = self.rollout_handler.update_rollout(policy=model,
extraction_keys=[constants.DATA_OBSERVATIONS, constants.
DATA_ACTIONS])
sampled_states, sampeled_actions = self.format_replay_buffer_data(**
rollout_data)
return sampled_states, sampeled_actions
def format_replay_buffer_data(self, **kwargs):
states = kwargs[constants.DATA_OBSERVATIONS]
actions = kwargs[constants.DATA_ACTIONS]
sampled_states, sampeled_actions = self._sample_state_action_paris(
states, actions)
return sampled_states, sampeled_actions
def _sample_state_action_paris(self, states, actions):
"""To make sure the number of state actions paris is always the same, this function sub samples the desired amount from a state action batch. This also acts as a kind of data augmentation as the representation of a single policy will consist of different state action pairs if called multiple times.
Args:
states (np.ndarray): Batch of states to sub sample from.
actions (np.ndarray): Batch of actions (according to states) to sum sample from.
Returns:
sampled_states (np.ndarray): Sub sampled states
sampeled_actions (np.ndarray): Sub sampled actions
"""
sample_id = np.random.choice(range(self.episode_length), size=self.
num_state_action_pairs, replace=False)
sampled_states = states[:, sample_id]
sampeled_actions = actions[:, sample_id]
return sampled_states, sampeled_actions
@property
def replay_data_info(self):
return self._replay_data_info
@property
def replay_data_keys(self):
return self._replay_data_keys
class ParameterFormat(Enum):
FlatParameters = FlatParamHandler
NamedParameters = NamedParamHandler
StateAction = StateActionHandler
class MLPEmbeddingNetwork(nn.Module):
def __init__(self):
super(MLPEmbeddingNetwork, self).__init__()
@abstractproperty
def embedding_size(self) ->int:
...
@abstractproperty
def input_type(self) ->ParameterFormat:
...
class StateActionEmbedding(MLPEmbeddingNetwork):
def __init__(self, num_state_action_pairs: 'int', observation_space:
'gym.Space', action_space: 'gym.Space'):
super(StateActionEmbedding, self).__init__()
self.observation_shape = observation_space.shape
self.action_shape = action_space.shape
self.num_state_action_pairs = num_state_action_pairs
self._embedding_size = num_state_action_pairs * (math.prod(self.
observation_shape) + math.prod(self.action_shape))
def forward(self, states, actions):
concatinated_batch = torch.cat([states, actions], dim=2)
return concatinated_batch
@property
def embedding_size(self) ->int:
return self._embedding_size
@property
def input_type(self) ->ParameterFormat:
return ParameterFormat.StateAction
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_state_action_pairs': 4, 'observation_space': torch.
rand([4, 4]), 'action_space': torch.rand([4, 4])}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import math
import numpy as np
from abc import ABC
from abc import abstractmethod
from abc import abstractproperty
from torch import nn
from enum import Enum
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4 % 8
x0 = xindex % 4
x2 = xindex // 32
x3 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 4 * x1 + 16 * x2), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp9 = tl.load(in_ptr1 + (x0 + 4 * (-4 + x1) + 16 * x2), tmp6 & xmask,
other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + x3, tmp10, xmask)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 4), (128, 32, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(512)](arg0_1, arg1_1, buf0, 512, XBLOCK
=256, num_warps=4, num_stages=1)
del arg0_1
del arg1_1
return buf0,
def tensor_to_numpy(tensor):
return tensor.detach().cpu().numpy()
class MLPParamHandler(ABC):
def __init__(self) ->None:
"""Interface for parameter handler. For Algorithms that require data on past model parameters this module handels this data such that is it is in the right format according to the specific algorithm.
"""
super().__init__()
@abstractmethod
def get_policy_replay_data(self, model: 'torch.nn.Module'):
"""Function to extract replay data from a module in a format that it can be saved to the replay buffer.
Args:
model (torch.nn.Module): Module to extract data from.
"""
...
@abstractmethod
def get_policy_critic_data(self, model: 'torch.nn.Module'):
"""Function to extract data from a policy that the critic requires to evaluate it.
Args:
model (torch.nn.Module): Module to extract data from.
"""
...
@abstractmethod
def format_replay_buffer_data(self, **kwargs):
"""Function to format data from the replay buffer such that it can be used as input to the critic afterwards.
"""
...
@abstractproperty
def replay_data_keys(self):
"""Keys used to save data to the replay buffer.
"""
...
@abstractproperty
def replay_data_info(self):
"""Description of the data needed to initialize the replay buffer.
"""
...
class FlatParamHandler(MLPParamHandler):
def __init__(self, example_policy: 'torch.nn.Module') ->None:
"""Parameter handler that simply takes all parameters flattens them and saves them in one vector.
Args:
example_policy (torch.nn.Module): Example policy network to acquire shape of data.
"""
super().__init__()
self._replay_data_keys = [constants.DATA_PARAMETERS]
self._replay_data_info = {self._replay_data_keys[0]: {'shape': self
.get_policy_critic_data(example_policy).shape}}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
return {self._replay_data_keys[0]: tensor_to_numpy(torch.nn.utils.
parameters_to_vector(model.parameters())).reshape(1, -1)}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
return torch.nn.utils.parameters_to_vector(model.parameters()).reshape(
1, -1)
def format_replay_buffer_data(self, **kwargs):
return kwargs[self._replay_data_keys[0]]
@property
def replay_data_keys(self):
return self._replay_data_keys
@property
def replay_data_info(self):
return self._replay_data_info
class NamedParamHandler(MLPParamHandler):
def __init__(self, example_policy: 'torch.nn.Module') ->None:
"""Parameter handler that saves parameters in a dictionary shape such that the parameters are saved in a similar format of how they are used in the actual module. Useful if the Parameters are later reused similarly to how the are used within the module they are extracted from.
Args:
example_policy (torch.nn.Module): Example policy network to acquire structure of module and according dictionary.
"""
super().__init__()
actor_parameter_dict = self.get_policy_critic_data(example_policy)
self._replay_data_keys = actor_parameter_dict.keys()
self._replay_data_info = {key: {'shape': actor_parameter_dict[key].
shape[1:]} for key in self._replay_data_keys}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
batched_param_dict = self.get_policy_critic_data(model)
return {key: tensor_to_numpy(value) for key, value in
batched_param_dict.items()}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
param_dict = dict(model.named_parameters())
return {key: torch.unsqueeze(tensor, dim=0) for key, tensor in
param_dict.items()}
def format_replay_buffer_data(self, **kwargs):
return {key: kwargs[key] for key in self._replay_data_keys}
@property
def replay_data_info(self):
return self._replay_data_info
@property
def replay_data_keys(self):
return self._replay_data_keys
class StateActionHandler(MLPParamHandler):
def __init__(self, num_state_action_pairs: 'int', episode_length: 'int',
rollout_handler: 'RolloutHandler') ->None:
"""Parameter handler that does not actually use parameters but rather state action pairs as representation of policies.
Args:
num_state_action_pairs (int): Number of state action pairs used as a representation for a policy.
episode_length (int): Maximal time steps of a episode used for representation purposes.
rollout_handler (RolloutHandler): Rollout handler used to execute rollouts when needed.
"""
super().__init__()
self.rollout_handler = rollout_handler
self.num_state_action_pairs = num_state_action_pairs
self.episode_length = episode_length
self._replay_data_keys = [constants.DATA_OBSERVATIONS, constants.
DATA_ACTIONS]
self._replay_data_info = {constants.DATA_OBSERVATIONS: {'shape': (
episode_length, *rollout_handler.environment_handler.
exploration_environment.observation_space.shape)}, constants.
DATA_ACTIONS: {'shape': (episode_length, *rollout_handler.
environment_handler.exploration_environment.action_space.shape)}}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
return {}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
rollout_data = self.rollout_handler.update_rollout(policy=model,
extraction_keys=[constants.DATA_OBSERVATIONS, constants.
DATA_ACTIONS])
sampled_states, sampeled_actions = self.format_replay_buffer_data(**
rollout_data)
return sampled_states, sampeled_actions
def format_replay_buffer_data(self, **kwargs):
states = kwargs[constants.DATA_OBSERVATIONS]
actions = kwargs[constants.DATA_ACTIONS]
sampled_states, sampeled_actions = self._sample_state_action_paris(
states, actions)
return sampled_states, sampeled_actions
def _sample_state_action_paris(self, states, actions):
"""To make sure the number of state actions paris is always the same, this function sub samples the desired amount from a state action batch. This also acts as a kind of data augmentation as the representation of a single policy will consist of different state action pairs if called multiple times.
Args:
states (np.ndarray): Batch of states to sub sample from.
actions (np.ndarray): Batch of actions (according to states) to sum sample from.
Returns:
sampled_states (np.ndarray): Sub sampled states
sampeled_actions (np.ndarray): Sub sampled actions
"""
sample_id = np.random.choice(range(self.episode_length), size=self.
num_state_action_pairs, replace=False)
sampled_states = states[:, sample_id]
sampeled_actions = actions[:, sample_id]
return sampled_states, sampeled_actions
@property
def replay_data_info(self):
return self._replay_data_info
@property
def replay_data_keys(self):
return self._replay_data_keys
class ParameterFormat(Enum):
FlatParameters = FlatParamHandler
NamedParameters = NamedParamHandler
StateAction = StateActionHandler
class MLPEmbeddingNetwork(nn.Module):
def __init__(self):
super(MLPEmbeddingNetwork, self).__init__()
@abstractproperty
def embedding_size(self) ->int:
...
@abstractproperty
def input_type(self) ->ParameterFormat:
...
class StateActionEmbeddingNew(MLPEmbeddingNetwork):
def __init__(self, num_state_action_pairs: 'int', observation_space:
'gym.Space', action_space: 'gym.Space'):
super(StateActionEmbeddingNew, self).__init__()
self.observation_shape = observation_space.shape
self.action_shape = action_space.shape
self.num_state_action_pairs = num_state_action_pairs
self._embedding_size = num_state_action_pairs * (math.prod(self.
observation_shape) + math.prod(self.action_shape))
@property
def embedding_size(self) ->int:
return self._embedding_size
@property
def input_type(self) ->ParameterFormat:
return ParameterFormat.StateAction
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| Sebastian-Griesbach/Improving-Policy-Conditioned-Value-Functions | StateActionEmbedding | false | 1,052 | [
"MIT"
] | 0 | ec4125c5e056753e507df0406fcd60b6b6c3dc25 | https://github.com/Sebastian-Griesbach/Improving-Policy-Conditioned-Value-Functions/tree/ec4125c5e056753e507df0406fcd60b6b6c3dc25 | import math
import torch
import numpy as np
from abc import ABC
from abc import abstractmethod
from abc import abstractproperty
from torch import nn
from enum import Enum
def tensor_to_numpy(tensor):
return tensor.detach().cpu().numpy()
class MLPParamHandler(ABC):
def __init__(self) ->None:
"""Interface for parameter handler. For Algorithms that require data on past model parameters this module handels this data such that is it is in the right format according to the specific algorithm.
"""
super().__init__()
@abstractmethod
def get_policy_replay_data(self, model: 'torch.nn.Module'):
"""Function to extract replay data from a module in a format that it can be saved to the replay buffer.
Args:
model (torch.nn.Module): Module to extract data from.
"""
...
@abstractmethod
def get_policy_critic_data(self, model: 'torch.nn.Module'):
"""Function to extract data from a policy that the critic requires to evaluate it.
Args:
model (torch.nn.Module): Module to extract data from.
"""
...
@abstractmethod
def format_replay_buffer_data(self, **kwargs):
"""Function to format data from the replay buffer such that it can be used as input to the critic afterwards.
"""
...
@abstractproperty
def replay_data_keys(self):
"""Keys used to save data to the replay buffer.
"""
...
@abstractproperty
def replay_data_info(self):
"""Description of the data needed to initialize the replay buffer.
"""
...
class FlatParamHandler(MLPParamHandler):
def __init__(self, example_policy: 'torch.nn.Module') ->None:
"""Parameter handler that simply takes all parameters flattens them and saves them in one vector.
Args:
example_policy (torch.nn.Module): Example policy network to acquire shape of data.
"""
super().__init__()
self._replay_data_keys = [constants.DATA_PARAMETERS]
self._replay_data_info = {self._replay_data_keys[0]: {'shape': self
.get_policy_critic_data(example_policy).shape}}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
return {self._replay_data_keys[0]: tensor_to_numpy(torch.nn.utils.
parameters_to_vector(model.parameters())).reshape(1, -1)}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
return torch.nn.utils.parameters_to_vector(model.parameters()).reshape(
1, -1)
def format_replay_buffer_data(self, **kwargs):
return kwargs[self._replay_data_keys[0]]
@property
def replay_data_keys(self):
return self._replay_data_keys
@property
def replay_data_info(self):
return self._replay_data_info
class NamedParamHandler(MLPParamHandler):
def __init__(self, example_policy: 'torch.nn.Module') ->None:
"""Parameter handler that saves parameters in a dictionary shape such that the parameters are saved in a similar format of how they are used in the actual module. Useful if the Parameters are later reused similarly to how the are used within the module they are extracted from.
Args:
example_policy (torch.nn.Module): Example policy network to acquire structure of module and according dictionary.
"""
super().__init__()
actor_parameter_dict = self.get_policy_critic_data(example_policy)
self._replay_data_keys = actor_parameter_dict.keys()
self._replay_data_info = {key: {'shape': actor_parameter_dict[key].
shape[1:]} for key in self._replay_data_keys}
def get_policy_replay_data(self, model: 'torch.nn.Module'):
batched_param_dict = self.get_policy_critic_data(model)
return {key: tensor_to_numpy(value) for key, value in
batched_param_dict.items()}
def get_policy_critic_data(self, model: 'torch.nn.Module'):
para
# ... truncated (>4000 chars) for memory efficiency |
GatedConv2d | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/mi/cmi6pvfnxxh5g6vyzi42gzjlvkcallq3paz67ydqnujyhpolr3bm.py
# Topologically Sorted Source Nodes: [x, mask, sigmoid, x_1, x_2], Original ATen: [aten.convolution, aten.sigmoid, aten.mul, aten.elu]
# Source node to ATen node mapping:
# mask => convolution_1
# sigmoid => sigmoid
# x => convolution
# x_1 => mul
# x_2 => expm1, gt, mul_1, mul_3, where
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_4, %primals_5, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution_1,), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, %sigmoid), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%mul, 0), kwargs = {})
# %mul_1 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, 1.0), kwargs = {})
# %expm1 : [num_users=1] = call_function[target=torch.ops.aten.expm1.default](args = (%mul_1,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expm1, 1.0), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %mul_1, %mul_3), kwargs = {})
triton_poi_fused_convolution_elu_mul_sigmoid_0 = async_compile.triton('triton_poi_fused_convolution_elu_mul_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_elu_mul_sigmoid_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_elu_mul_sigmoid_0(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 144
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 9) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + (x3), xmask)
tmp4 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tl.sigmoid(tmp5)
tmp7 = tmp2 * tmp6
tmp8 = 0.0
tmp9 = tmp7 > tmp8
tmp10 = 1.0
tmp11 = tmp7 * tmp10
tmp12 = libdevice.expm1(tmp11)
tmp13 = tmp12 * tmp10
tmp14 = tl.where(tmp9, tmp11, tmp13)
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
tl.store(in_out_ptr1 + (x3), tmp5, xmask)
tl.store(out_ptr0 + (x3), tmp14, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 3, 3), (36, 9, 3, 1))
# Topologically Sorted Source Nodes: [mask], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_3, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 3, 3), (36, 9, 3, 1))
buf1 = buf0; del buf0 # reuse
buf3 = buf2; del buf2 # reuse
buf4 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.float32)
# Topologically Sorted Source Nodes: [x, mask, sigmoid, x_1, x_2], Original ATen: [aten.convolution, aten.sigmoid, aten.mul, aten.elu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_elu_mul_sigmoid_0.run(buf1, buf3, primals_2, primals_5, buf4, 144, grid=grid(144), stream=stream0)
del primals_2
del primals_5
return (buf4, primals_1, primals_3, primals_4, buf1, buf3, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class GatedConv2d(torch.nn.Module):
"""
Gated Convlution layer with activation (default activation:LeakyReLU)
Params: same as conv2d
Input: The feature from last layer "I"
Output:\\phi(f(I))*\\sigmoid(g(I))
"""
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
dilation=1, bias=True, activation=torch.nn.ELU(1.0, inplace=True),
dropout=0, gate_type='regular_conv'):
super(GatedConv2d, self).__init__()
self.stride = stride
padding = dilation * (kernel_size - 1) // 2
self.activation = activation
self.conv2d = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=kernel_size, stride=stride, padding=
padding, dilation=dilation, bias=bias)
if gate_type == 'regular_conv':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=kernel_size, stride=
stride, padding=padding, dilation=dilation, bias=bias)
elif gate_type == 'single_channel':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=1, kernel_size=kernel_size, stride=stride,
padding=padding, dilation=dilation, bias=bias)
elif gate_type == 'pixel_wise':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=1, stride=stride,
padding=0, dilation=dilation, bias=bias)
elif gate_type == 'depth_separable':
self.mask_conv2d = nn.Sequential(nn.Conv2d(in_channels=
in_channels, out_channels=in_channels, kernel_size=
kernel_size, stride=stride, padding=padding, dilation=
dilation, bias=bias, groups=in_channels), nn.Conv2d(
in_channels=in_channels, out_channels=out_channels,
kernel_size=1, padding=0, bias=bias))
self.sigmoid = nn.Sigmoid()
if dropout > 0:
self.dropout = nn.Dropout(dropout)
else:
self.dropout = None
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight)
def forward(self, input):
x = self.conv2d(input)
mask = self.mask_conv2d(input)
x = x * self.sigmoid(mask)
if self.activation is not None:
x = self.activation(x)
if self.dropout:
x = self.dropout(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_convolution_elu_mul_sigmoid_0(in_out_ptr0, in_out_ptr1,
in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 144
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 9 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_out_ptr1 + x3, xmask)
tmp4 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tl.sigmoid(tmp5)
tmp7 = tmp2 * tmp6
tmp8 = 0.0
tmp9 = tmp7 > tmp8
tmp10 = 1.0
tmp11 = tmp7 * tmp10
tmp12 = libdevice.expm1(tmp11)
tmp13 = tmp12 * tmp10
tmp14 = tl.where(tmp9, tmp11, tmp13)
tl.store(in_out_ptr0 + x3, tmp2, xmask)
tl.store(in_out_ptr1 + x3, tmp5, xmask)
tl.store(out_ptr0 + x3, tmp14, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 3, 3), (36, 9, 3, 1))
buf2 = extern_kernels.convolution(primals_3, primals_4, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 3, 3), (36, 9, 3, 1))
buf1 = buf0
del buf0
buf3 = buf2
del buf2
buf4 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_convolution_elu_mul_sigmoid_0[grid(144)](buf1,
buf3, primals_2, primals_5, buf4, 144, XBLOCK=128, num_warps=4,
num_stages=1)
del primals_2
del primals_5
return buf4, primals_1, primals_3, primals_4, buf1, buf3, buf4
class GatedConv2dNew(torch.nn.Module):
"""
Gated Convlution layer with activation (default activation:LeakyReLU)
Params: same as conv2d
Input: The feature from last layer "I"
Output:\\phi(f(I))*\\sigmoid(g(I))
"""
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
dilation=1, bias=True, activation=torch.nn.ELU(1.0, inplace=True),
dropout=0, gate_type='regular_conv'):
super(GatedConv2dNew, self).__init__()
self.stride = stride
padding = dilation * (kernel_size - 1) // 2
self.activation = activation
self.conv2d = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=kernel_size, stride=stride, padding=
padding, dilation=dilation, bias=bias)
if gate_type == 'regular_conv':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=kernel_size, stride=
stride, padding=padding, dilation=dilation, bias=bias)
elif gate_type == 'single_channel':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=1, kernel_size=kernel_size, stride=stride,
padding=padding, dilation=dilation, bias=bias)
elif gate_type == 'pixel_wise':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=1, stride=stride,
padding=0, dilation=dilation, bias=bias)
elif gate_type == 'depth_separable':
self.mask_conv2d = nn.Sequential(nn.Conv2d(in_channels=
in_channels, out_channels=in_channels, kernel_size=
kernel_size, stride=stride, padding=padding, dilation=
dilation, bias=bias, groups=in_channels), nn.Conv2d(
in_channels=in_channels, out_channels=out_channels,
kernel_size=1, padding=0, bias=bias))
self.sigmoid = nn.Sigmoid()
if dropout > 0:
self.dropout = nn.Dropout(dropout)
else:
self.dropout = None
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight)
def forward(self, input_0):
primals_1 = self.conv2d.weight
primals_2 = self.conv2d.bias
primals_3 = self.mask_conv2d.weight
primals_5 = self.mask_conv2d.bias
primals_4 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| ShiraLightricks/3d-photo-inpainting | GatedConv2d | false | 1,053 | [
"MIT"
] | 0 | c42ac41576690b765e50f5281ddbfb58439ff36d | https://github.com/ShiraLightricks/3d-photo-inpainting/tree/c42ac41576690b765e50f5281ddbfb58439ff36d | import torch
import torch.nn as nn
class Model(torch.nn.Module):
"""
Gated Convlution layer with activation (default activation:LeakyReLU)
Params: same as conv2d
Input: The feature from last layer "I"
Output:\\phi(f(I))*\\sigmoid(g(I))
"""
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
dilation=1, bias=True, activation=torch.nn.ELU(1.0, inplace=True),
dropout=0, gate_type='regular_conv'):
super().__init__()
self.stride = stride
padding = dilation * (kernel_size - 1) // 2
self.activation = activation
self.conv2d = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=kernel_size, stride=stride, padding=
padding, dilation=dilation, bias=bias)
if gate_type == 'regular_conv':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=kernel_size, stride=
stride, padding=padding, dilation=dilation, bias=bias)
elif gate_type == 'single_channel':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=1, kernel_size=kernel_size, stride=stride,
padding=padding, dilation=dilation, bias=bias)
elif gate_type == 'pixel_wise':
self.mask_conv2d = nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=1, stride=stride,
padding=0, dilation=dilation, bias=bias)
elif gate_type == 'depth_separable':
self.mask_conv2d = nn.Sequential(nn.Conv2d(in_channels=
in_channels, out_channels=in_channels, kernel_size=
kernel_size, stride=stride, padding=padding, dilation=
dilation, bias=bias, groups=in_channels), nn.Conv2d(
in_channels=in_channels, out_channels=out_channels,
kernel_size=1, padding=0, bias=bias))
self.sigmoid = nn.Sigmoid()
if dropout > 0:
self.dropout = nn.Dropout(dropout)
else:
self.dropout = None
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight)
def forward(self, input):
x = self.conv2d(input)
mask = self.mask_conv2d(input)
x = x * self.sigmoid(mask)
if self.activation is not None:
x = self.activation(x)
if self.dropout:
x = self.dropout(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
CoordConv | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/nz/cnzr3enannjni75kec3qorz6jm6lyd5whz6u5l3ih55bgihwnb2u.py
# Topologically Sorted Source Nodes: [ret], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# ret => cat
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_1, %device_put, %device_put_1], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 16) % 6
x3 = (xindex // 96)
x4 = xindex % 16
x1 = (xindex // 4) % 4
x0 = xindex % 4
x5 = xindex
tmp0 = x2
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x4 + (16*x2) + (64*x3)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 5, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = x1
tmp11 = tmp10.to(tl.float32)
tmp12 = 0.3333333333333333
tmp13 = tmp11 * tmp12
tmp14 = 2.0
tmp15 = tmp13 * tmp14
tmp16 = 1.0
tmp17 = tmp15 - tmp16
tmp18 = tl.full(tmp17.shape, 0.0, tmp17.dtype)
tmp19 = tl.where(tmp9, tmp17, tmp18)
tmp20 = tmp0 >= tmp7
tmp21 = tl.full([1], 6, tl.int64)
tmp22 = tmp0 < tmp21
tmp23 = x0
tmp24 = tmp23.to(tl.float32)
tmp25 = tmp24 * tmp12
tmp26 = tmp25 * tmp14
tmp27 = tmp26 - tmp16
tmp28 = tl.full(tmp27.shape, 0.0, tmp27.dtype)
tmp29 = tl.where(tmp20, tmp27, tmp28)
tmp30 = tl.where(tmp9, tmp19, tmp29)
tmp31 = tl.where(tmp4, tmp5, tmp30)
tl.store(out_ptr0 + (x5), tmp31, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/32/c32v7egt4mupqssam3gmac2qgv3ujprjybthsgweflmot256qqw7.py
# Topologically Sorted Source Nodes: [ret_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# ret_1 => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%cat, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 6, 1, 1), (6, 1, 1, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 6, 4, 4), (96, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [ret], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_1, buf0, 384, grid=grid(384), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [ret_1], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [ret_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf2, primals_3, 256, grid=grid(256), stream=stream0)
del primals_3
return (buf2, primals_2, buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 6, 1, 1), (6, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class AddCoords(nn.Module):
def __init__(self, with_r=False):
super().__init__()
self.with_r = with_r
def forward(self, input_tensor):
"""
Args:
input_tensor: shape(batch, channel, x_dim, y_dim)
"""
batch_size, _, x_dim, y_dim = input_tensor.size()
xx_channel = torch.arange(x_dim).repeat(1, y_dim, 1)
yy_channel = torch.arange(y_dim).repeat(1, x_dim, 1).transpose(1, 2)
xx_channel = xx_channel.float() / (x_dim - 1)
yy_channel = yy_channel.float() / (y_dim - 1)
xx_channel = xx_channel * 2 - 1
yy_channel = yy_channel * 2 - 1
xx_channel = xx_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
yy_channel = yy_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
ret = torch.cat([input_tensor, xx_channel.type_as(input_tensor),
yy_channel.type_as(input_tensor)], dim=1)
if self.with_r:
rr = torch.sqrt(torch.pow(xx_channel.type_as(input_tensor) -
0.5, 2) + torch.pow(yy_channel.type_as(input_tensor) - 0.5, 2))
ret = torch.cat([ret, rr], dim=1)
return ret
class CoordConv(nn.Module):
def __init__(self, in_channels, out_channels, with_r=False, **kwargs):
super().__init__()
self.addcoords = AddCoords(with_r=with_r)
in_size = in_channels + 2
if with_r:
in_size += 1
self.conv = nn.Conv2d(in_size, out_channels, kernel_size=1, stride=
1, padding=0, dilation=1, groups=1, bias=True)
def forward(self, x):
ret = self.addcoords(x)
ret = self.conv(ret)
return ret
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 16 % 6
x3 = xindex // 96
x4 = xindex % 16
x1 = xindex // 4 % 4
x0 = xindex % 4
x5 = xindex
tmp0 = x2
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x4 + 16 * x2 + 64 * x3), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 5, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = x1
tmp11 = tmp10.to(tl.float32)
tmp12 = 0.3333333333333333
tmp13 = tmp11 * tmp12
tmp14 = 2.0
tmp15 = tmp13 * tmp14
tmp16 = 1.0
tmp17 = tmp15 - tmp16
tmp18 = tl.full(tmp17.shape, 0.0, tmp17.dtype)
tmp19 = tl.where(tmp9, tmp17, tmp18)
tmp20 = tmp0 >= tmp7
tl.full([1], 6, tl.int64)
tmp23 = x0
tmp24 = tmp23.to(tl.float32)
tmp25 = tmp24 * tmp12
tmp26 = tmp25 * tmp14
tmp27 = tmp26 - tmp16
tmp28 = tl.full(tmp27.shape, 0.0, tmp27.dtype)
tmp29 = tl.where(tmp20, tmp27, tmp28)
tmp30 = tl.where(tmp9, tmp19, tmp29)
tmp31 = tl.where(tmp4, tmp5, tmp30)
tl.store(out_ptr0 + x5, tmp31, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 6, 1, 1), (6, 1, 1, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 6, 4, 4), (96, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(384)](primals_1, buf0, 384, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_1[grid(256)](buf2, primals_3, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
return buf2, primals_2, buf0
class AddCoords(nn.Module):
def __init__(self, with_r=False):
super().__init__()
self.with_r = with_r
def forward(self, input_tensor):
"""
Args:
input_tensor: shape(batch, channel, x_dim, y_dim)
"""
batch_size, _, x_dim, y_dim = input_tensor.size()
xx_channel = torch.arange(x_dim).repeat(1, y_dim, 1)
yy_channel = torch.arange(y_dim).repeat(1, x_dim, 1).transpose(1, 2)
xx_channel = xx_channel.float() / (x_dim - 1)
yy_channel = yy_channel.float() / (y_dim - 1)
xx_channel = xx_channel * 2 - 1
yy_channel = yy_channel * 2 - 1
xx_channel = xx_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
yy_channel = yy_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
ret = torch.cat([input_tensor, xx_channel.type_as(input_tensor),
yy_channel.type_as(input_tensor)], dim=1)
if self.with_r:
rr = torch.sqrt(torch.pow(xx_channel.type_as(input_tensor) -
0.5, 2) + torch.pow(yy_channel.type_as(input_tensor) - 0.5, 2))
ret = torch.cat([ret, rr], dim=1)
return ret
class CoordConvNew(nn.Module):
def __init__(self, in_channels, out_channels, with_r=False, **kwargs):
super().__init__()
self.addcoords = AddCoords(with_r=with_r)
in_size = in_channels + 2
if with_r:
in_size += 1
self.conv = nn.Conv2d(in_size, out_channels, kernel_size=1, stride=
1, padding=0, dilation=1, groups=1, bias=True)
def forward(self, input_0):
primals_2 = self.conv.weight
primals_3 = self.conv.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| SeunghwanByun/Real-Time-Road-Detection-Network | CoordConv | false | 1,054 | [
"MIT"
] | 0 | bc46615adef0e2b1a9a03dd4951559ca5849e6e1 | https://github.com/SeunghwanByun/Real-Time-Road-Detection-Network/tree/bc46615adef0e2b1a9a03dd4951559ca5849e6e1 | import torch
import torch.nn as nn
class AddCoords(nn.Module):
def __init__(self, with_r=False):
super().__init__()
self.with_r = with_r
def forward(self, input_tensor):
"""
Args:
input_tensor: shape(batch, channel, x_dim, y_dim)
"""
batch_size, _, x_dim, y_dim = input_tensor.size()
xx_channel = torch.arange(x_dim).repeat(1, y_dim, 1)
yy_channel = torch.arange(y_dim).repeat(1, x_dim, 1).transpose(1, 2)
xx_channel = xx_channel.float() / (x_dim - 1)
yy_channel = yy_channel.float() / (y_dim - 1)
xx_channel = xx_channel * 2 - 1
yy_channel = yy_channel * 2 - 1
xx_channel = xx_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
yy_channel = yy_channel.repeat(batch_size, 1, 1, 1).transpose(2, 3)
ret = torch.cat([input_tensor, xx_channel.type_as(input_tensor),
yy_channel.type_as(input_tensor)], dim=1)
if self.with_r:
rr = torch.sqrt(torch.pow(xx_channel.type_as(input_tensor) -
0.5, 2) + torch.pow(yy_channel.type_as(input_tensor) - 0.5, 2))
ret = torch.cat([ret, rr], dim=1)
return ret
class Model(nn.Module):
def __init__(self, in_channels, out_channels, with_r=False, **kwargs):
super().__init__()
self.addcoords = AddCoords(with_r=with_r)
in_size = in_channels + 2
if with_r:
in_size += 1
self.conv = nn.Conv2d(in_size, out_channels, kernel_size=1, stride=
1, padding=0, dilation=1, groups=1, bias=True)
def forward(self, x):
ret = self.addcoords(x)
ret = self.conv(ret)
return ret
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ScaledDotProductAttention | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/nu/cnuc7ivckuuly7yn2763pwt3sw72jd6vuwpeeu4sfespm5iz7fq4.py
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention_score_2 => exp
# Graph fragment:
# %mul_tensor : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_5, 1), kwargs = {})
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%mul_tensor, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_tensor, %amax_default), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_tensor, 2.0), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%div_tensor,), kwargs = {})
triton_poi_fused__softmax_0 = async_compile.triton('triton_poi_fused__softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = 0.5
tmp16 = tmp14 * tmp15
tmp17 = tl_math.exp(tmp16)
tl.store(out_ptr0 + (x2), tmp17, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/fj/cfjl47pvhwbpfbvh6rfehwy5ijxc5p3zgkld2lwf3mw5bl6pbkak.py
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attention_score_2 => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention_score], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(arg0_1, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(arg1_1, (16, 4, 4), (16, 4, 1), 0), out=buf0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [attention_score_2], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf1, buf2, 256, grid=grid(256), stream=stream0)
buf3 = reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [result], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(arg2_1, (16, 4, 4), (16, 4, 1), 0), out=buf3)
del arg2_1
return (reinterpret_tensor(buf3, (4, 4, 4, 4), (64, 16, 4, 1), 0), buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused__softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = 0.5
tmp16 = tmp14 * tmp15
tmp17 = tl_math.exp(tmp16)
tl.store(out_ptr0 + x2, tmp17, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(arg0_1, (16, 4, 4), (16, 4, 1
), 0), reinterpret_tensor(arg1_1, (16, 4, 4), (16, 4, 1), 0),
out=buf0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(256)](buf0, buf1, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf2 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
triton_poi_fused__softmax_1[grid(256)](buf1, buf2, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf3 = reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0)
del buf1
extern_kernels.bmm(reinterpret_tensor(buf2, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(arg2_1, (16, 4, 4), (16, 4, 1), 0), out=buf3
)
del arg2_1
return reinterpret_tensor(buf3, (4, 4, 4, 4), (64, 16, 4, 1), 0), buf2
class ScaledDotProductAttentionNew(nn.Module):
def __init__(self):
super(ScaledDotProductAttentionNew, self).__init__()
def forward(self, input_0, input_1, input_2):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
output = call([arg0_1, arg1_1, arg2_1])
return output[0], output[1]
| SeungoneKim/Transformer_implementation | ScaledDotProductAttention | false | 1,055 | [
"Apache-2.0"
] | 0 | a52bf552eb645fc9bfb812cc26842fc147d6c008 | https://github.com/SeungoneKim/Transformer_implementation/tree/a52bf552eb645fc9bfb812cc26842fc147d6c008 | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super().__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return []
|
Swish | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/lc/clcsmfjtwsnzu2llmpsmvwrt5ojf76ozdv5ttluhi2gtpojjo6lv.py
# Topologically Sorted Source Nodes: [sigmoid, mul_], Original ATen: [aten.sigmoid, aten.mul]
# Source node to ATen node mapping:
# mul_ => mul
# sigmoid => sigmoid
# Graph fragment:
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%arg0_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %sigmoid), kwargs = {})
# %copy_ : [num_users=1] = call_function[target=torch.ops.aten.copy_.default](args = (%arg0_1, %mul), kwargs = {})
triton_poi_fused_mul_sigmoid_0 = async_compile.triton('triton_poi_fused_mul_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_sigmoid_0', 'mutated_arg_names': ['in_ptr0', 'out_ptr1'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_sigmoid_0(in_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tmp0 * tmp1
tl.store(out_ptr1 + (x0), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [sigmoid, mul_], Original ATen: [aten.sigmoid, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_sigmoid_0.run(arg0_1, arg0_1, 256, grid=grid(256), stream=stream0)
return (arg0_1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class Swish(nn.Module):
def __init__(self, inplace=True):
super(Swish, self).__init__()
self.inplace = inplace
def forward(self, x):
return x.mul_(x.sigmoid()) if self.inplace else x.mul(x.sigmoid())
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
@triton.jit
def triton_poi_fused_mul_sigmoid_0(in_ptr0, out_ptr1, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tmp0 * tmp1
tl.store(out_ptr1 + x0, tmp2, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
get_raw_stream(0)
triton_poi_fused_mul_sigmoid_0[grid(256)](arg0_1, arg0_1, 256,
XBLOCK=128, num_warps=4, num_stages=1)
return arg0_1,
class SwishNew(nn.Module):
def __init__(self, inplace=True):
super(SwishNew, self).__init__()
self.inplace = inplace
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| ShowLo/Networks | Swish | false | 1,056 | [
"MIT"
] | 0 | 48f8545783966c383b6c3b600fbe37a15ea8ae3c | https://github.com/ShowLo/Networks/tree/48f8545783966c383b6c3b600fbe37a15ea8ae3c | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, inplace=True):
super().__init__()
self.inplace = inplace
def forward(self, x):
return x.mul_(x.sigmoid()) if self.inplace else x.mul(x.sigmoid())
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Bicubic | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/j5/cj5es5bvythkhvipgt3xmi2elrhehimxnb5sn3noat3qbbqagqx7.py
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.floor, aten.arange, aten._to_copy, aten.add, aten.mul, aten.sub, aten._unsafe_index, aten.clamp, aten.rsub]
# Source node to ATen node mapping:
# interpolate => _unsafe_index, _unsafe_index_1, _unsafe_index_10, _unsafe_index_11, _unsafe_index_12, _unsafe_index_13, _unsafe_index_14, _unsafe_index_15, _unsafe_index_2, _unsafe_index_3, _unsafe_index_4, _unsafe_index_5, _unsafe_index_6, _unsafe_index_7, _unsafe_index_8, _unsafe_index_9, add, add_10, add_11, add_12, add_13, add_14, add_15, add_16, add_17, add_18, add_19, add_20, add_21, add_22, add_23, add_24, add_25, add_26, add_27, add_28, add_29, add_30, add_6, add_7, add_8, add_9, clamp_max, clamp_max_1, clamp_min, clamp_min_1, convert_element_type_1, floor, floor_1, iota_1, mul, mul_10, mul_11, mul_12, mul_13, mul_14, mul_15, mul_16, mul_17, mul_18, mul_19, mul_2, mul_20, mul_21, mul_22, mul_23, mul_24, mul_25, mul_26, mul_27, mul_28, mul_29, mul_3, mul_30, mul_31, mul_32, mul_33, mul_34, mul_35, mul_36, mul_37, mul_38, mul_39, mul_4, mul_40, mul_41, mul_42, mul_43, mul_44, mul_45, mul_5, mul_6, mul_7, mul_8, mul_9, sub, sub_10, sub_11, sub_12, sub_13, sub_14, sub_15, sub_16, sub_17, sub_18, sub_19, sub_2, sub_20, sub_21, sub_3, sub_6, sub_7, sub_8, sub_9
# Graph fragment:
# %floor_1 : [num_users=2] = call_function[target=torch.ops.aten.floor.default](args = (%unsqueeze,), kwargs = {})
# %iota_1 : [num_users=1] = call_function[target=torch.ops.prims.iota.default](args = (4,), kwargs = {start: 0, step: 1, dtype: torch.int64, device: cuda:0, requires_grad: False})
# %convert_element_type_1 : [num_users=1] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%iota_1, torch.float32), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convert_element_type_1, 0.5), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add, 1.0), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul, 0.5), kwargs = {})
# %floor : [num_users=2] = call_function[target=torch.ops.aten.floor.default](args = (%sub,), kwargs = {})
# %_unsafe_index : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_2, %clamp_max_3]), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %floor), kwargs = {})
# %clamp_min_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sub_3, 0.0), kwargs = {})
# %clamp_max_1 : [num_users=6] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min_1, 1.0), kwargs = {})
# %add_6 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%clamp_max_1, 1.0), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_6, -0.75), kwargs = {})
# %sub_6 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_2, -3.75), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_6, %add_6), kwargs = {})
# %add_7 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_3, -6.0), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_7, %add_6), kwargs = {})
# %sub_7 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_4, -3.0), kwargs = {})
# %mul_26 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index, %sub_7), kwargs = {})
# %_unsafe_index_1 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_4, %clamp_max_5]), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%clamp_max_1, 1.25), kwargs = {})
# %sub_8 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_5, 2.25), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_8, %clamp_max_1), kwargs = {})
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_6, %clamp_max_1), kwargs = {})
# %add_8 : [num_users=4] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_7, 1), kwargs = {})
# %mul_27 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_1, %add_8), kwargs = {})
# %add_16 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_26, %mul_27), kwargs = {})
# %_unsafe_index_2 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_6, %clamp_max_7]), kwargs = {})
# %sub_9 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %clamp_max_1), kwargs = {})
# %mul_8 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_9, 1.25), kwargs = {})
# %sub_10 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_8, 2.25), kwargs = {})
# %mul_9 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_10, %sub_9), kwargs = {})
# %mul_10 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_9, %sub_9), kwargs = {})
# %add_9 : [num_users=4] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_10, 1), kwargs = {})
# %mul_28 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_2, %add_9), kwargs = {})
# %add_17 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_16, %mul_28), kwargs = {})
# %_unsafe_index_3 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_8, %clamp_max_9]), kwargs = {})
# %sub_11 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (2.0, %clamp_max_1), kwargs = {})
# %mul_11 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_11, -0.75), kwargs = {})
# %sub_12 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_11, -3.75), kwargs = {})
# %mul_12 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_12, %sub_11), kwargs = {})
# %add_10 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_12, -6.0), kwargs = {})
# %mul_13 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_10, %sub_11), kwargs = {})
# %sub_13 : [num_users=4] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_13, -3.0), kwargs = {})
# %mul_29 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_3, %sub_13), kwargs = {})
# %add_18 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_17, %mul_29), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%unsqueeze, %floor_1), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sub_2, 0.0), kwargs = {})
# %clamp_max : [num_users=6] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 1.0), kwargs = {})
# %add_11 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%clamp_max, 1.0), kwargs = {})
# %mul_14 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_11, -0.75), kwargs = {})
# %sub_14 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_14, -3.75), kwargs = {})
# %mul_15 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_14, %add_11), kwargs = {})
# %add_12 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_15, -6.0), kwargs = {})
# %mul_16 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_12, %add_11), kwargs = {})
# %sub_15 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_16, -3.0), kwargs = {})
# %mul_42 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_18, %sub_15), kwargs = {})
# %_unsafe_index_4 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_10, %clamp_max_11]), kwargs = {})
# %mul_30 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_4, %sub_7), kwargs = {})
# %_unsafe_index_5 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_12, %clamp_max_13]), kwargs = {})
# %mul_31 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_5, %add_8), kwargs = {})
# %add_19 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_30, %mul_31), kwargs = {})
# %_unsafe_index_6 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_14, %clamp_max_15]), kwargs = {})
# %mul_32 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_6, %add_9), kwargs = {})
# %add_20 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_19, %mul_32), kwargs = {})
# %_unsafe_index_7 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_16, %clamp_max_17]), kwargs = {})
# %mul_33 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_7, %sub_13), kwargs = {})
# %add_21 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_20, %mul_33), kwargs = {})
# %mul_17 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%clamp_max, 1.25), kwargs = {})
# %sub_16 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_17, 2.25), kwargs = {})
# %mul_18 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_16, %clamp_max), kwargs = {})
# %mul_19 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_18, %clamp_max), kwargs = {})
# %add_13 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_19, 1), kwargs = {})
# %mul_43 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_21, %add_13), kwargs = {})
# %add_28 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_42, %mul_43), kwargs = {})
# %_unsafe_index_8 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_18, %clamp_max_19]), kwargs = {})
# %mul_34 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_8, %sub_7), kwargs = {})
# %_unsafe_index_9 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_20, %clamp_max_21]), kwargs = {})
# %mul_35 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_9, %add_8), kwargs = {})
# %add_22 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_34, %mul_35), kwargs = {})
# %_unsafe_index_10 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_22, %clamp_max_23]), kwargs = {})
# %mul_36 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_10, %add_9), kwargs = {})
# %add_23 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_22, %mul_36), kwargs = {})
# %_unsafe_index_11 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_24, %clamp_max_25]), kwargs = {})
# %mul_37 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_11, %sub_13), kwargs = {})
# %add_24 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_23, %mul_37), kwargs = {})
# %sub_17 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %clamp_max), kwargs = {})
# %mul_20 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_17, 1.25), kwargs = {})
# %sub_18 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_20, 2.25), kwargs = {})
# %mul_21 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_18, %sub_17), kwargs = {})
# %mul_22 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_21, %sub_17), kwargs = {})
# %add_14 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_22, 1), kwargs = {})
# %mul_44 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_24, %add_14), kwargs = {})
# %add_29 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_28, %mul_44), kwargs = {})
# %_unsafe_index_12 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_26, %clamp_max_27]), kwargs = {})
# %mul_38 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_12, %sub_7), kwargs = {})
# %_unsafe_index_13 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_28, %clamp_max_29]), kwargs = {})
# %mul_39 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_13, %add_8), kwargs = {})
# %add_25 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_38, %mul_39), kwargs = {})
# %_unsafe_index_14 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_30, %clamp_max_31]), kwargs = {})
# %mul_40 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_14, %add_9), kwargs = {})
# %add_26 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_25, %mul_40), kwargs = {})
# %_unsafe_index_15 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%arg0_1, [None, None, %clamp_max_32, %clamp_max_33]), kwargs = {})
# %mul_41 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%_unsafe_index_15, %sub_13), kwargs = {})
# %add_27 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_26, %mul_41), kwargs = {})
# %sub_19 : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (2.0, %clamp_max), kwargs = {})
# %mul_23 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_19, -0.75), kwargs = {})
# %sub_20 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_23, -3.75), kwargs = {})
# %mul_24 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_20, %sub_19), kwargs = {})
# %add_15 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_24, -6.0), kwargs = {})
# %mul_25 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_15, %sub_19), kwargs = {})
# %sub_21 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_25, -3.0), kwargs = {})
# %mul_45 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_27, %sub_21), kwargs = {})
# %add_30 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_29, %mul_45), kwargs = {})
triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0 = async_compile.triton('triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0', 'mutated_arg_names': ['in_out_ptr1'], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0(in_out_ptr1, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4) % 4
x0 = xindex % 4
x2 = (xindex // 16)
x3 = xindex
tmp0 = x1
tmp1 = tmp0.to(tl.float32)
tmp2 = 0.5
tmp3 = tmp1 + tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = tmp5 - tmp2
tmp7 = libdevice.floor(tmp6)
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.full([1], 1, tl.int64)
tmp10 = tmp8 - tmp9
tmp11 = tl.full([1], 0, tl.int64)
tmp12 = triton_helpers.maximum(tmp10, tmp11)
tmp13 = tl.full([1], 3, tl.int64)
tmp14 = triton_helpers.minimum(tmp12, tmp13)
tmp15 = x0
tmp16 = tmp15.to(tl.float32)
tmp17 = tmp16 + tmp2
tmp18 = tmp17 * tmp4
tmp19 = tmp18 - tmp2
tmp20 = libdevice.floor(tmp19)
tmp21 = tmp20.to(tl.int32)
tmp22 = tmp21 - tmp9
tmp23 = triton_helpers.maximum(tmp22, tmp11)
tmp24 = triton_helpers.minimum(tmp23, tmp13)
tmp25 = tl.load(in_ptr0 + (tmp24 + (4*tmp14) + (16*x2)), xmask, eviction_policy='evict_last')
tmp26 = tmp19 - tmp20
tmp27 = 0.0
tmp28 = triton_helpers.maximum(tmp26, tmp27)
tmp29 = triton_helpers.minimum(tmp28, tmp4)
tmp30 = tmp29 + tmp4
tmp31 = -0.75
tmp32 = tmp30 * tmp31
tmp33 = -3.75
tmp34 = tmp32 - tmp33
tmp35 = tmp34 * tmp30
tmp36 = -6.0
tmp37 = tmp35 + tmp36
tmp38 = tmp37 * tmp30
tmp39 = -3.0
tmp40 = tmp38 - tmp39
tmp41 = tmp25 * tmp40
tmp42 = triton_helpers.maximum(tmp21, tmp11)
tmp43 = triton_helpers.minimum(tmp42, tmp13)
tmp44 = tl.load(in_ptr0 + (tmp43 + (4*tmp14) + (16*x2)), xmask, eviction_policy='evict_last')
tmp45 = 1.25
tmp46 = tmp29 * tmp45
tmp47 = 2.25
tmp48 = tmp46 - tmp47
tmp49 = tmp48 * tmp29
tmp50 = tmp49 * tmp29
tmp51 = tmp50 + tmp4
tmp52 = tmp44 * tmp51
tmp53 = tmp21 + tmp9
tmp54 = triton_helpers.maximum(tmp53, tmp11)
tmp55 = triton_helpers.minimum(tmp54, tmp13)
tmp56 = tl.load(in_ptr0 + (tmp55 + (4*tmp14) + (16*x2)), xmask, eviction_policy='evict_last')
tmp57 = tmp4 - tmp29
tmp58 = tmp57 * tmp45
tmp59 = tmp58 - tmp47
tmp60 = tmp59 * tmp57
tmp61 = tmp60 * tmp57
tmp62 = tmp61 + tmp4
tmp63 = tmp56 * tmp62
tmp64 = tl.full([1], 2, tl.int64)
tmp65 = tmp21 + tmp64
tmp66 = triton_helpers.maximum(tmp65, tmp11)
tmp67 = triton_helpers.minimum(tmp66, tmp13)
tmp68 = tl.load(in_ptr0 + (tmp67 + (4*tmp14) + (16*x2)), xmask, eviction_policy='evict_last')
tmp69 = 2.0
tmp70 = tmp69 - tmp29
tmp71 = tmp70 * tmp31
tmp72 = tmp71 - tmp33
tmp73 = tmp72 * tmp70
tmp74 = tmp73 + tmp36
tmp75 = tmp74 * tmp70
tmp76 = tmp75 - tmp39
tmp77 = tmp68 * tmp76
tmp78 = tmp41 + tmp52
tmp79 = tmp78 + tmp63
tmp80 = tmp79 + tmp77
tmp81 = tmp6 - tmp7
tmp82 = triton_helpers.maximum(tmp81, tmp27)
tmp83 = triton_helpers.minimum(tmp82, tmp4)
tmp84 = tmp83 + tmp4
tmp85 = tmp84 * tmp31
tmp86 = tmp85 - tmp33
tmp87 = tmp86 * tmp84
tmp88 = tmp87 + tmp36
tmp89 = tmp88 * tmp84
tmp90 = tmp89 - tmp39
tmp91 = tmp80 * tmp90
tmp92 = triton_helpers.maximum(tmp8, tmp11)
tmp93 = triton_helpers.minimum(tmp92, tmp13)
tmp94 = tl.load(in_ptr0 + (tmp24 + (4*tmp93) + (16*x2)), xmask, eviction_policy='evict_last')
tmp95 = tmp94 * tmp40
tmp96 = tl.load(in_ptr0 + (tmp43 + (4*tmp93) + (16*x2)), xmask, eviction_policy='evict_last')
tmp97 = tmp96 * tmp51
tmp98 = tl.load(in_ptr0 + (tmp55 + (4*tmp93) + (16*x2)), xmask, eviction_policy='evict_last')
tmp99 = tmp98 * tmp62
tmp100 = tl.load(in_ptr0 + (tmp67 + (4*tmp93) + (16*x2)), xmask, eviction_policy='evict_last')
tmp101 = tmp100 * tmp76
tmp102 = tmp8 + tmp9
tmp103 = triton_helpers.maximum(tmp102, tmp11)
tmp104 = triton_helpers.minimum(tmp103, tmp13)
tmp105 = tl.load(in_ptr0 + (tmp24 + (4*tmp104) + (16*x2)), xmask, eviction_policy='evict_last')
tmp106 = tmp105 * tmp40
tmp107 = tl.load(in_ptr0 + (tmp43 + (4*tmp104) + (16*x2)), xmask, eviction_policy='evict_last')
tmp108 = tmp107 * tmp51
tmp109 = tl.load(in_ptr0 + (tmp55 + (4*tmp104) + (16*x2)), xmask, eviction_policy='evict_last')
tmp110 = tmp109 * tmp62
tmp111 = tl.load(in_ptr0 + (tmp67 + (4*tmp104) + (16*x2)), xmask, eviction_policy='evict_last')
tmp112 = tmp111 * tmp76
tmp113 = tmp95 + tmp97
tmp114 = tmp113 + tmp99
tmp115 = tmp114 + tmp101
tmp116 = tmp83 * tmp45
tmp117 = tmp116 - tmp47
tmp118 = tmp117 * tmp83
tmp119 = tmp118 * tmp83
tmp120 = tmp119 + tmp4
tmp121 = tmp115 * tmp120
tmp122 = tmp91 + tmp121
tmp123 = tmp106 + tmp108
tmp124 = tmp123 + tmp110
tmp125 = tmp124 + tmp112
tmp126 = tmp4 - tmp83
tmp127 = tmp126 * tmp45
tmp128 = tmp127 - tmp47
tmp129 = tmp128 * tmp126
tmp130 = tmp129 * tmp126
tmp131 = tmp130 + tmp4
tmp132 = tmp125 * tmp131
tmp133 = tmp122 + tmp132
tmp134 = tmp8 + tmp64
tmp135 = triton_helpers.maximum(tmp134, tmp11)
tmp136 = triton_helpers.minimum(tmp135, tmp13)
tmp137 = tl.load(in_ptr0 + (tmp24 + (4*tmp136) + (16*x2)), xmask, eviction_policy='evict_last')
tmp138 = tmp137 * tmp40
tmp139 = tl.load(in_ptr0 + (tmp43 + (4*tmp136) + (16*x2)), xmask, eviction_policy='evict_last')
tmp140 = tmp139 * tmp51
tmp141 = tl.load(in_ptr0 + (tmp55 + (4*tmp136) + (16*x2)), xmask, eviction_policy='evict_last')
tmp142 = tmp141 * tmp62
tmp143 = tl.load(in_ptr0 + (tmp67 + (4*tmp136) + (16*x2)), xmask, eviction_policy='evict_last')
tmp144 = tmp143 * tmp76
tmp145 = tmp138 + tmp140
tmp146 = tmp145 + tmp142
tmp147 = tmp146 + tmp144
tmp148 = tmp69 - tmp83
tmp149 = tmp148 * tmp31
tmp150 = tmp149 - tmp33
tmp151 = tmp150 * tmp148
tmp152 = tmp151 + tmp36
tmp153 = tmp152 * tmp148
tmp154 = tmp153 - tmp39
tmp155 = tmp147 * tmp154
tmp156 = tmp133 + tmp155
tl.store(in_out_ptr1 + (x3), tmp156, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf10 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf13 = buf10; del buf10 # reuse
buf19 = buf13; del buf13 # reuse
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.floor, aten.arange, aten._to_copy, aten.add, aten.mul, aten.sub, aten._unsafe_index, aten.clamp, aten.rsub]
stream0 = get_raw_stream(0)
triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0.run(buf19, arg0_1, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf19, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| from torch.nn import Module
import torch
import torch.nn.functional as F
class Bicubic(Module):
def __init__(self, scale_factor):
super().__init__()
self.scale_factor = scale_factor
def forward(self, x):
return F.interpolate(x, scale_factor=self.scale_factor, mode='bicubic')
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'scale_factor': 1.0}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
from torch.nn import Module
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0(
in_out_ptr1, in_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4 % 4
x0 = xindex % 4
x2 = xindex // 16
x3 = xindex
tmp0 = x1
tmp1 = tmp0.to(tl.float32)
tmp2 = 0.5
tmp3 = tmp1 + tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = tmp5 - tmp2
tmp7 = libdevice.floor(tmp6)
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.full([1], 1, tl.int64)
tmp10 = tmp8 - tmp9
tmp11 = tl.full([1], 0, tl.int64)
tmp12 = triton_helpers.maximum(tmp10, tmp11)
tmp13 = tl.full([1], 3, tl.int64)
tmp14 = triton_helpers.minimum(tmp12, tmp13)
tmp15 = x0
tmp16 = tmp15.to(tl.float32)
tmp17 = tmp16 + tmp2
tmp18 = tmp17 * tmp4
tmp19 = tmp18 - tmp2
tmp20 = libdevice.floor(tmp19)
tmp21 = tmp20.to(tl.int32)
tmp22 = tmp21 - tmp9
tmp23 = triton_helpers.maximum(tmp22, tmp11)
tmp24 = triton_helpers.minimum(tmp23, tmp13)
tmp25 = tl.load(in_ptr0 + (tmp24 + 4 * tmp14 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp26 = tmp19 - tmp20
tmp27 = 0.0
tmp28 = triton_helpers.maximum(tmp26, tmp27)
tmp29 = triton_helpers.minimum(tmp28, tmp4)
tmp30 = tmp29 + tmp4
tmp31 = -0.75
tmp32 = tmp30 * tmp31
tmp33 = -3.75
tmp34 = tmp32 - tmp33
tmp35 = tmp34 * tmp30
tmp36 = -6.0
tmp37 = tmp35 + tmp36
tmp38 = tmp37 * tmp30
tmp39 = -3.0
tmp40 = tmp38 - tmp39
tmp41 = tmp25 * tmp40
tmp42 = triton_helpers.maximum(tmp21, tmp11)
tmp43 = triton_helpers.minimum(tmp42, tmp13)
tmp44 = tl.load(in_ptr0 + (tmp43 + 4 * tmp14 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp45 = 1.25
tmp46 = tmp29 * tmp45
tmp47 = 2.25
tmp48 = tmp46 - tmp47
tmp49 = tmp48 * tmp29
tmp50 = tmp49 * tmp29
tmp51 = tmp50 + tmp4
tmp52 = tmp44 * tmp51
tmp53 = tmp21 + tmp9
tmp54 = triton_helpers.maximum(tmp53, tmp11)
tmp55 = triton_helpers.minimum(tmp54, tmp13)
tmp56 = tl.load(in_ptr0 + (tmp55 + 4 * tmp14 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp57 = tmp4 - tmp29
tmp58 = tmp57 * tmp45
tmp59 = tmp58 - tmp47
tmp60 = tmp59 * tmp57
tmp61 = tmp60 * tmp57
tmp62 = tmp61 + tmp4
tmp63 = tmp56 * tmp62
tmp64 = tl.full([1], 2, tl.int64)
tmp65 = tmp21 + tmp64
tmp66 = triton_helpers.maximum(tmp65, tmp11)
tmp67 = triton_helpers.minimum(tmp66, tmp13)
tmp68 = tl.load(in_ptr0 + (tmp67 + 4 * tmp14 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp69 = 2.0
tmp70 = tmp69 - tmp29
tmp71 = tmp70 * tmp31
tmp72 = tmp71 - tmp33
tmp73 = tmp72 * tmp70
tmp74 = tmp73 + tmp36
tmp75 = tmp74 * tmp70
tmp76 = tmp75 - tmp39
tmp77 = tmp68 * tmp76
tmp78 = tmp41 + tmp52
tmp79 = tmp78 + tmp63
tmp80 = tmp79 + tmp77
tmp81 = tmp6 - tmp7
tmp82 = triton_helpers.maximum(tmp81, tmp27)
tmp83 = triton_helpers.minimum(tmp82, tmp4)
tmp84 = tmp83 + tmp4
tmp85 = tmp84 * tmp31
tmp86 = tmp85 - tmp33
tmp87 = tmp86 * tmp84
tmp88 = tmp87 + tmp36
tmp89 = tmp88 * tmp84
tmp90 = tmp89 - tmp39
tmp91 = tmp80 * tmp90
tmp92 = triton_helpers.maximum(tmp8, tmp11)
tmp93 = triton_helpers.minimum(tmp92, tmp13)
tmp94 = tl.load(in_ptr0 + (tmp24 + 4 * tmp93 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp95 = tmp94 * tmp40
tmp96 = tl.load(in_ptr0 + (tmp43 + 4 * tmp93 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp97 = tmp96 * tmp51
tmp98 = tl.load(in_ptr0 + (tmp55 + 4 * tmp93 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp99 = tmp98 * tmp62
tmp100 = tl.load(in_ptr0 + (tmp67 + 4 * tmp93 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp101 = tmp100 * tmp76
tmp102 = tmp8 + tmp9
tmp103 = triton_helpers.maximum(tmp102, tmp11)
tmp104 = triton_helpers.minimum(tmp103, tmp13)
tmp105 = tl.load(in_ptr0 + (tmp24 + 4 * tmp104 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp106 = tmp105 * tmp40
tmp107 = tl.load(in_ptr0 + (tmp43 + 4 * tmp104 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp108 = tmp107 * tmp51
tmp109 = tl.load(in_ptr0 + (tmp55 + 4 * tmp104 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp110 = tmp109 * tmp62
tmp111 = tl.load(in_ptr0 + (tmp67 + 4 * tmp104 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp112 = tmp111 * tmp76
tmp113 = tmp95 + tmp97
tmp114 = tmp113 + tmp99
tmp115 = tmp114 + tmp101
tmp116 = tmp83 * tmp45
tmp117 = tmp116 - tmp47
tmp118 = tmp117 * tmp83
tmp119 = tmp118 * tmp83
tmp120 = tmp119 + tmp4
tmp121 = tmp115 * tmp120
tmp122 = tmp91 + tmp121
tmp123 = tmp106 + tmp108
tmp124 = tmp123 + tmp110
tmp125 = tmp124 + tmp112
tmp126 = tmp4 - tmp83
tmp127 = tmp126 * tmp45
tmp128 = tmp127 - tmp47
tmp129 = tmp128 * tmp126
tmp130 = tmp129 * tmp126
tmp131 = tmp130 + tmp4
tmp132 = tmp125 * tmp131
tmp133 = tmp122 + tmp132
tmp134 = tmp8 + tmp64
tmp135 = triton_helpers.maximum(tmp134, tmp11)
tmp136 = triton_helpers.minimum(tmp135, tmp13)
tmp137 = tl.load(in_ptr0 + (tmp24 + 4 * tmp136 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp138 = tmp137 * tmp40
tmp139 = tl.load(in_ptr0 + (tmp43 + 4 * tmp136 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp140 = tmp139 * tmp51
tmp141 = tl.load(in_ptr0 + (tmp55 + 4 * tmp136 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp142 = tmp141 * tmp62
tmp143 = tl.load(in_ptr0 + (tmp67 + 4 * tmp136 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp144 = tmp143 * tmp76
tmp145 = tmp138 + tmp140
tmp146 = tmp145 + tmp142
tmp147 = tmp146 + tmp144
tmp148 = tmp69 - tmp83
tmp149 = tmp148 * tmp31
tmp150 = tmp149 - tmp33
tmp151 = tmp150 * tmp148
tmp152 = tmp151 + tmp36
tmp153 = tmp152 * tmp148
tmp154 = tmp153 - tmp39
tmp155 = tmp147 * tmp154
tmp156 = tmp133 + tmp155
tl.store(in_out_ptr1 + x3, tmp156, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf10 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf13 = buf10
del buf10
buf19 = buf13
del buf13
get_raw_stream(0)
triton_poi_fused__to_copy__unsafe_index_add_arange_clamp_floor_mul_rsub_sub_0[
grid(256)](buf19, arg0_1, 256, XBLOCK=128, num_warps=4,
num_stages=1)
del arg0_1
return buf19,
class BicubicNew(Module):
def __init__(self, scale_factor):
super().__init__()
self.scale_factor = scale_factor
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| ShivanshuPurohit/Diffusion | Bicubic | false | 1,057 | [
"MIT"
] | 0 | 9a190d9aa4ed9767cf223e4ef57d0c31690f92cc | https://github.com/ShivanshuPurohit/Diffusion/tree/9a190d9aa4ed9767cf223e4ef57d0c31690f92cc | from torch.nn import Module
import torch
import torch.nn.functional as F
class Model(Module):
def __init__(self, scale_factor):
super().__init__()
self.scale_factor = scale_factor
def forward(self, x):
return F.interpolate(x, scale_factor=self.scale_factor, mode='bicubic')
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [1.0]
|
adder2d | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/w7/cw7nwm4phyff5u2mxvg4wffuk3qkl3kcrshhf3qydzhb7zd4zl75.py
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous_1 => clone_1
# Graph fragment:
# %clone_1 : [num_users=3] = call_function[target=torch.ops.aten.clone.default](args = (%permute_2,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/fh/cfhxlck7uzhxtofknhjghf2xokzgxovbt22nsyc7lfq6lggpmsc6.py
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.neg]
# Source node to ATen node mapping:
# out_2 => neg
# Graph fragment:
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%_cdist_forward,), kwargs = {})
triton_poi_fused_neg_1 = async_compile.triton('triton_poi_fused_neg_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_neg_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_neg_1(in_out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = -tmp0
tl.store(in_out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/xb/cxb4lxq2tbmyzyffeyfgkeb34canjnomidn7ybre2qjgrajgfrs5.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.neg]
# Source node to ATen node mapping:
# out => neg_1
# Graph fragment:
# %neg_1 : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%_cdist_forward_1,), kwargs = {})
triton_poi_fused_neg_2 = async_compile.triton('triton_poi_fused_neg_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_neg_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_neg_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = -tmp0
tl.store(out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 64), (64, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous_1], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(primals_2, buf0, 256, grid=grid(256), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [cdist], Original ATen: [aten._cdist_forward]
buf1 = torch.ops.aten._cdist_forward.default(reinterpret_tensor(primals_1, (4, 64), (64, 1), 0), buf0, 1.0, None)
buf2 = buf1
del buf1
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.neg]
triton_poi_fused_neg_1.run(buf3, 16, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [cdist_1], Original ATen: [aten._cdist_forward]
buf4 = torch.ops.aten._cdist_forward.default(reinterpret_tensor(primals_1, (4, 64), (64, 1), 0), buf0, 2.0, None)
buf5 = buf4
del buf4
buf6 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.neg]
triton_poi_fused_neg_2.run(buf5, buf6, 16, grid=grid(16), stream=stream0)
return (buf3, buf6, reinterpret_tensor(primals_1, (4, 64), (64, 1), 0), buf0, buf5, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
def adder2d_function(X, W, stride=1, padding=0, groups=1):
n_filters, _d_filter, h_filter, w_filter = W.size()
n_x, _d_x, h_x, w_x = X.size()
h_out = (h_x - h_filter + 2 * padding) / stride + 1
w_out = (w_x - w_filter + 2 * padding) / stride + 1
h_out, w_out = int(h_out), int(w_out)
if groups == 1:
X_col = torch.nn.functional.unfold(X.view(1, -1, h_x, w_x),
h_filter, dilation=1, padding=padding, stride=stride).view(n_x,
-1, h_out * w_out)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W.view(n_filters, -1)
out_2 = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 1)
out_2 = out_2.detach()
out = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 2)
out.data = out_2.data
out = out.view(n_filters, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
return out
else:
torch.Tensor(n_x, n_filters, h_out, w_out)
outs = []
for g in range(groups):
X_part = X[:, g, ...].unsqueeze(1)
W_part = W[g, ...].unsqueeze(0)
X_col = torch.nn.functional.unfold(X_part.view(1, -1, h_x, w_x),
h_filter, dilation=1, padding=padding, stride=stride).view(n_x,
-1, h_out * w_out)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W_part.view(1, -1)
out_2 = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 1)
out_2 = out_2.detach()
out = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 2)
out.data = out_2.data
out = out.view(1, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
outs.append(out)
return torch.cat(outs, dim=1)
class adder2d(nn.Module):
def __init__(self, input_channel, output_channel, kernel_size, stride=1,
padding=0, groups=1, bias=False):
super(adder2d, self).__init__()
self.stride = stride
self.padding = padding
self.input_channel = input_channel
self.output_channel = output_channel
self.kernel_size = kernel_size
self.groups = groups
self.adder = torch.nn.Parameter(nn.init.kaiming_normal_(torch.randn
(output_channel, input_channel // groups, kernel_size,
kernel_size)))
self.bias = bias
if bias:
self.b = torch.nn.Parameter(nn.init.kaiming_uniform_(torch.
zeros(output_channel)))
def forward(self, x):
output = adder2d_function(x, self.adder, self.stride, self.padding,
self.groups)
if self.bias:
output += self.b.unsqueeze(0).unsqueeze(2).unsqueeze(3)
return output
def extra_repr(self):
return (
f"{'v2'.upper()}, {self.input_channel}, {self.output_channel}, " +
f'kenrel_size={self.kernel_size}, stride={self.stride}, padding={self.padding}, groups={self.groups}, bias={self.bias}'
)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_channel': 4, 'output_channel': 4, 'kernel_size': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
tmp0 = tl.load(in_ptr0 + x2, xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
@triton.jit
def triton_poi_fused_neg_1(in_out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = -tmp0
tl.store(in_out_ptr0 + x0, tmp1, xmask)
@triton.jit
def triton_poi_fused_neg_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = -tmp0
tl.store(out_ptr0 + x0, tmp1, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 64), (64, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(256)](primals_2, buf0, 256, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_2
buf1 = torch.ops.aten._cdist_forward.default(reinterpret_tensor(
primals_1, (4, 64), (64, 1), 0), buf0, 1.0, None)
buf2 = buf1
del buf1
buf3 = buf2
del buf2
triton_poi_fused_neg_1[grid(16)](buf3, 16, XBLOCK=16, num_warps=1,
num_stages=1)
buf4 = torch.ops.aten._cdist_forward.default(reinterpret_tensor(
primals_1, (4, 64), (64, 1), 0), buf0, 2.0, None)
buf5 = buf4
del buf4
buf6 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_neg_2[grid(16)](buf5, buf6, 16, XBLOCK=16,
num_warps=1, num_stages=1)
return buf3, buf6, reinterpret_tensor(primals_1, (4, 64), (64, 1), 0
), buf0, buf5
def adder2d_function(X, W, stride=1, padding=0, groups=1):
n_filters, _d_filter, h_filter, w_filter = W.size()
n_x, _d_x, h_x, w_x = X.size()
h_out = (h_x - h_filter + 2 * padding) / stride + 1
w_out = (w_x - w_filter + 2 * padding) / stride + 1
h_out, w_out = int(h_out), int(w_out)
if groups == 1:
X_col = torch.nn.functional.unfold(X.view(1, -1, h_x, w_x),
h_filter, dilation=1, padding=padding, stride=stride).view(n_x,
-1, h_out * w_out)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W.view(n_filters, -1)
out_2 = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 1)
out_2 = out_2.detach()
out = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 2)
out.data = out_2.data
out = out.view(n_filters, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
return out
else:
torch.Tensor(n_x, n_filters, h_out, w_out)
outs = []
for g in range(groups):
X_part = X[:, g, ...].unsqueeze(1)
W_part = W[g, ...].unsqueeze(0)
X_col = torch.nn.functional.unfold(X_part.view(1, -1, h_x, w_x),
h_filter, dilation=1, padding=padding, stride=stride).view(n_x,
-1, h_out * w_out)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W_part.view(1, -1)
out_2 = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 1)
out_2 = out_2.detach()
out = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 2)
out.data = out_2.data
out = out.view(1, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
outs.append(out)
return torch.cat(outs, dim=1)
class adder2dNew(nn.Module):
def __init__(self, input_channel, output_channel, kernel_size, stride=1,
padding=0, groups=1, bias=False):
super(adder2dNew, self).__init__()
self.stride = stride
self.padding = padding
self.input_channel = input_channel
self.output_channel = output_channel
self.kernel_size = kernel_size
self.groups = groups
self.adder = torch.nn.Parameter(nn.init.kaiming_normal_(torch.randn
(output_channel, input_channel // groups, kernel_size,
kernel_size)))
self.bias = bias
if bias:
self.b = torch.nn.Parameter(nn.init.kaiming_uniform_(torch.
zeros(output_channel)))
def extra_repr(self):
return (
f"{'v2'.upper()}, {self.input_channel}, {self.output_channel}, " +
f'kenrel_size={self.kernel_size}, stride={self.stride}, padding={self.padding}, groups={self.groups}, bias={self.bias}'
)
def forward(self, input_0):
primals_1 = self.adder
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| ShangyinGao/pytorch-cifar | adder2d | false | 1,058 | [
"MIT"
] | 0 | 480e19825bb155e3d0fafae3545faa3a4165bd77 | https://github.com/ShangyinGao/pytorch-cifar/tree/480e19825bb155e3d0fafae3545faa3a4165bd77 | import torch
import torch.nn as nn
def adder2d_function(X, W, stride=1, padding=0, groups=1):
n_filters, _d_filter, h_filter, w_filter = W.size()
n_x, _d_x, h_x, w_x = X.size()
h_out = (h_x - h_filter + 2 * padding) / stride + 1
w_out = (w_x - w_filter + 2 * padding) / stride + 1
h_out, w_out = int(h_out), int(w_out)
if groups == 1:
X_col = torch.nn.functional.unfold(X.view(1, -1, h_x, w_x),
h_filter, dilation=1, padding=padding, stride=stride).view(n_x,
-1, h_out * w_out)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W.view(n_filters, -1)
out_2 = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 1)
out_2 = out_2.detach()
out = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 2)
out.data = out_2.data
out = out.view(n_filters, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
return out
else:
torch.Tensor(n_x, n_filters, h_out, w_out)
outs = []
for g in range(groups):
X_part = X[:, g, ...].unsqueeze(1)
W_part = W[g, ...].unsqueeze(0)
X_col = torch.nn.functional.unfold(X_part.view(1, -1, h_x, w_x),
h_filter, dilation=1, padding=padding, stride=stride).view(n_x,
-1, h_out * w_out)
X_col = X_col.permute(1, 2, 0).contiguous().view(X_col.size(1), -1)
W_col = W_part.view(1, -1)
out_2 = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 1)
out_2 = out_2.detach()
out = -torch.cdist(W_col, X_col.transpose(0, 1).contiguous(), 2)
out.data = out_2.data
out = out.view(1, h_out, w_out, n_x)
out = out.permute(3, 0, 1, 2).contiguous()
outs.append(out)
return torch.cat(outs, dim=1)
class Model(nn.Module):
def __init__(self, input_channel, output_channel, kernel_size, stride=1,
padding=0, groups=1, bias=False):
super().__init__()
self.stride = stride
self.padding = padding
self.input_channel = input_channel
self.output_channel = output_channel
self.kernel_size = kernel_size
self.groups = groups
self.adder = torch.nn.Parameter(nn.init.kaiming_normal_(torch.randn
(output_channel, input_channel // groups, kernel_size,
kernel_size)))
self.bias = bias
if bias:
self.b = torch.nn.Parameter(nn.init.kaiming_uniform_(torch.
zeros(output_channel)))
def forward(self, x):
output = adder2d_function(x, self.adder, self.stride, self.padding,
self.groups)
if self.bias:
output += self.b.unsqueeze(0).unsqueeze(2).unsqueeze(3)
return output
def extra_repr(self):
return (
f"{'v2'.upper()}, {self.input_channel}, {self.output_channel}, " +
f'kenrel_size={self.kernel_size}, stride={self.stride}, padding={self.padding}, groups={self.groups}, bias={self.bias}'
)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
FFN | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/r3/cr3febcwm3t44fuoitsx3ou2p6xg4sk4f7unagmmrvffasxf47te.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ji/cji7mw45fbdoanjc5e6qu3e2bf5d6jnnjabskl6onjlk7uv7oqud.py
# Topologically Sorted Source Nodes: [tgt, tgt_1], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# tgt => add
# tgt_1 => var_mean
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_3, %view_3), kwargs = {})
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add, [3]), kwargs = {correction: 0, keepdim: True})
triton_poi_fused_add_native_layer_norm_1 = async_compile.triton('triton_poi_fused_add_native_layer_norm_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_1(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + (x0), tmp16, xmask)
tl.store(out_ptr1 + (x0), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/xy/cxyvzp6lij7d3yqq2ut3vi6guk7xnzb7qwqb66dthlly44r65vfk.py
# Topologically Sorted Source Nodes: [tgt, tgt_1], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# tgt => add
# tgt_1 => add_1, add_2, mul, mul_1, rsqrt, sub
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_3, %view_3), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %getitem_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %rsqrt), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %primals_6), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_7), kwargs = {})
triton_poi_fused_add_native_layer_norm_2 = async_compile.triton('triton_poi_fused_add_native_layer_norm_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + (x2), tmp13, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf6, 256, grid=grid(256), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [tgt2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [tgt, tgt_1], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_1.run(primals_3, buf2, buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [tgt, tgt_1], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_2.run(primals_3, buf2, buf3, buf4, primals_6, primals_7, buf5, 256, grid=grid(256), stream=stream0)
del buf3
del buf4
del primals_7
return (buf5, primals_3, primals_6, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), buf2, primals_4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.utils.data
import torchvision.transforms.functional as F
import torch.nn as nn
import torch.nn.functional as F
class FFN(nn.Module):
def __init__(self, d_model, d_ffn, dropout=0):
super().__init__()
self.linear1 = nn.Linear(d_model, d_ffn)
self.activation = F.relu
self.dropout1 = nn.Dropout(dropout)
self.linear2 = nn.Linear(d_ffn, d_model)
self.dropout2 = nn.Dropout(dropout)
self.norm = nn.LayerNorm(d_model)
def forward(self, tgt):
tgt2 = self.linear2(self.dropout1(self.activation(self.linear1(tgt))))
tgt = tgt + self.dropout2(tgt2)
tgt = self.norm(tgt)
return tgt
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'d_model': 4, 'd_ffn': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.utils.data
import torchvision.transforms.functional as F
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_1(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp12 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + x0, tmp16, xmask)
tl.store(out_ptr1 + x0, tmp28, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_2(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + x2, tmp13, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(256)](buf1,
primals_2, buf6, 256, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf2)
del primals_5
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused_add_native_layer_norm_1[grid(64)](primals_3, buf2,
buf3, buf4, 64, XBLOCK=64, num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_native_layer_norm_2[grid(256)](primals_3, buf2,
buf3, buf4, primals_6, primals_7, buf5, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf3
del buf4
del primals_7
return buf5, primals_3, primals_6, reinterpret_tensor(buf1, (64, 4), (4,
1), 0), buf2, primals_4, buf6
class FFNNew(nn.Module):
def __init__(self, d_model, d_ffn, dropout=0):
super().__init__()
self.linear1 = nn.Linear(d_model, d_ffn)
self.activation = F.relu
self.dropout1 = nn.Dropout(dropout)
self.linear2 = nn.Linear(d_ffn, d_model)
self.dropout2 = nn.Dropout(dropout)
self.norm = nn.LayerNorm(d_model)
def forward(self, input_0):
primals_1 = self.linear1.weight
primals_2 = self.linear1.bias
primals_4 = self.linear2.weight
primals_5 = self.linear2.bias
primals_6 = self.norm.weight
primals_7 = self.norm.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| SelvamArul/MOTR | FFN | false | 1,059 | [
"MIT"
] | 0 | 2a0b70288feaca665d460096159100d5077e9312 | https://github.com/SelvamArul/MOTR/tree/2a0b70288feaca665d460096159100d5077e9312 | import torch
import torch.utils.data
import torchvision.transforms.functional as F
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, d_model, d_ffn, dropout=0):
super().__init__()
self.linear1 = nn.Linear(d_model, d_ffn)
self.activation = F.relu
self.dropout1 = nn.Dropout(dropout)
self.linear2 = nn.Linear(d_ffn, d_model)
self.dropout2 = nn.Dropout(dropout)
self.norm = nn.LayerNorm(d_model)
def forward(self, tgt):
tgt2 = self.linear2(self.dropout1(self.activation(self.linear1(tgt))))
tgt = tgt + self.dropout2(tgt2)
tgt = self.norm(tgt)
return tgt
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
BinaryReg | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/dx/cdxzp5g4colwzu5hffvbsmu53nbfzo5dpcn2duoxzlf2thhrusqc.py
# Topologically Sorted Source Nodes: [diff, abs_1, diff_1, truediv, loss, mul], Original ATen: [aten.sub, aten.abs, aten.clamp, aten.reciprocal, aten.mul, aten.mean]
# Source node to ATen node mapping:
# abs_1 => abs_1
# diff => sub
# diff_1 => clamp_min
# loss => mean
# mul => mul_1
# truediv => mul, reciprocal
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, 0.5), kwargs = {})
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%sub,), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%abs_1, 0.01), kwargs = {})
# %reciprocal : [num_users=1] = call_function[target=torch.ops.aten.reciprocal.default](args = (%clamp_min,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%reciprocal, 1.0), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%mul,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, 0.1), kwargs = {})
triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0 = async_compile.triton('triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0(in_out_ptr0, in_ptr0, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = 0.5
tmp2 = tmp0 - tmp1
tmp3 = tl_math.abs(tmp2)
tmp4 = 0.01
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tl.full([1], 1, tl.int32)
tmp7 = tmp6 / tmp5
tmp8 = 1.0
tmp9 = tmp7 * tmp8
tmp10 = tl.broadcast_to(tmp9, [RBLOCK])
tmp12 = triton_helpers.promote_to_tensor(tl.sum(tmp10, 0))
tmp13 = 256.0
tmp14 = tmp12 / tmp13
tmp15 = 0.1
tmp16 = tmp14 * tmp15
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp16, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [diff, abs_1, diff_1, truediv, loss, mul], Original ATen: [aten.sub, aten.abs, aten.clamp, aten.reciprocal, aten.mul, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0.run(buf1, arg0_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
class BinaryReg(nn.Module):
"""Regularization for encouraging the outputs to be binary.
"""
def __init__(self, alpha=0.1):
super().__init__()
self.alpha = alpha
def forward(self, pred):
diff = pred - 0.5
diff = torch.clamp(torch.abs(diff), min=0.01)
loss = (1.0 / diff).mean()
return self.alpha * loss
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0(in_out_ptr0,
in_ptr0, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = 0.5
tmp2 = tmp0 - tmp1
tmp3 = tl_math.abs(tmp2)
tmp4 = 0.01
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tl.full([1], 1, tl.int32)
tmp7 = tmp6 / tmp5
tmp8 = 1.0
tmp9 = tmp7 * tmp8
tmp10 = tl.broadcast_to(tmp9, [RBLOCK])
tmp12 = triton_helpers.promote_to_tensor(tl.sum(tmp10, 0))
tmp13 = 256.0
tmp14 = tmp12 / tmp13
tmp15 = 0.1
tmp16 = tmp14 * tmp15
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp16, None)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_abs_clamp_mean_mul_reciprocal_sub_0[grid(1)](buf1,
arg0_1, 1, 256, num_warps=2, num_stages=1)
del arg0_1
return buf1,
class BinaryRegNew(nn.Module):
"""Regularization for encouraging the outputs to be binary.
"""
def __init__(self, alpha=0.1):
super().__init__()
self.alpha = alpha
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Shray64/pytorch_connectomics | BinaryReg | false | 1,060 | [
"MIT"
] | 0 | d6c814f11ac2f8418ede5ae220a93016f50214fc | https://github.com/Shray64/pytorch_connectomics/tree/d6c814f11ac2f8418ede5ae220a93016f50214fc | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
"""Regularization for encouraging the outputs to be binary.
"""
def __init__(self, alpha=0.1):
super().__init__()
self.alpha = alpha
def forward(self, pred):
diff = pred - 0.5
diff = torch.clamp(torch.abs(diff), min=0.01)
loss = (1.0 / diff).mean()
return self.alpha * loss
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
MessageNormalizer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/5p/c5pq5ihkqog6yst24r6r2rrf5qe3nsxkwwpixhsiqjqc6rcatvet.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, %primals_2), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, ), (1, ))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
return (buf0, primals_2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class MessageNormalizer(nn.Module):
def __init__(self, in_features, init_mean=1.0, init_stddev=0.01):
super(MessageNormalizer, self).__init__()
self.in_features = in_features
self.out_features = in_features
self.weight = torch.nn.Parameter(torch.Tensor(in_features))
self.init_mean = init_mean
self.init_stddev = init_stddev
self.reset_parameters()
def reset_parameters(self):
self.weight.data.normal_(mean=self.init_mean, std=self.init_stddev)
def forward(self, message):
return self.weight * message
def __repr__(self):
return self.__class__.__name__ + '(' + 'in_features=' + str(self.
in_features) + 'out_features=' + str(self.out_features) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_features': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x2, tmp2, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4,), (1,))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(256)](primals_1, primals_2, buf0, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_1
return buf0, primals_2
class MessageNormalizerNew(nn.Module):
def __init__(self, in_features, init_mean=1.0, init_stddev=0.01):
super(MessageNormalizerNew, self).__init__()
self.in_features = in_features
self.out_features = in_features
self.weight = torch.nn.Parameter(torch.Tensor(in_features))
self.init_mean = init_mean
self.init_stddev = init_stddev
self.reset_parameters()
def reset_parameters(self):
self.weight.data.normal_(mean=self.init_mean, std=self.init_stddev)
def __repr__(self):
return self.__class__.__name__ + '(' + 'in_features=' + str(self.
in_features) + 'out_features=' + str(self.out_features) + ')'
def forward(self, input_0):
primals_1 = self.weight
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| ShinyaFUKUMOTO/LeMPA | MessageNormalizer | false | 1,061 | [
"BSD-2-Clause"
] | 0 | 23b8c9f60fc13cf28d4485757d2ae0b3465b3e92 | https://github.com/ShinyaFUKUMOTO/LeMPA/tree/23b8c9f60fc13cf28d4485757d2ae0b3465b3e92 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_features, init_mean=1.0, init_stddev=0.01):
super().__init__()
self.in_features = in_features
self.out_features = in_features
self.weight = torch.nn.Parameter(torch.Tensor(in_features))
self.init_mean = init_mean
self.init_stddev = init_stddev
self.reset_parameters()
def reset_parameters(self):
self.weight.data.normal_(mean=self.init_mean, std=self.init_stddev)
def forward(self, message):
return self.weight * message
def __repr__(self):
return self.__class__.__name__ + '(' + 'in_features=' + str(self.
in_features) + 'out_features=' + str(self.out_features) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
MaxPoolStride1 | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/rz/crzg7kxbg2534vzhc3kxhbostzh2mlj5xgy25tqm37zxwoqenf5z.py
# Topologically Sorted Source Nodes: [pooled_x], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# pooled_x => getitem
# Graph fragment:
# %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 0), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_0 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 2
x1 = (xindex // 2) % 2
x2 = (xindex // 4)
x4 = xindex
tmp0 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (3*x1)) + (3*x1) * ((3*x1) < (3)))) + (16*x2) + ((3) * ((3) <= (3*x0)) + (3*x0) * ((3*x0) < (3)))), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (3*x1)) + (3*x1) * ((3*x1) < (3)))) + (16*x2) + ((3) * ((3) <= (1 + (3*x0))) + (1 + (3*x0)) * ((1 + (3*x0)) < (3)))), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (3*x1)) + (3*x1) * ((3*x1) < (3)))) + (16*x2) + ((3) * ((3) <= (2 + (3*x0))) + (2 + (3*x0)) * ((2 + (3*x0)) < (3)))), xmask)
tmp5 = tl.load(in_ptr0 + (3 + (4*((3) * ((3) <= (3*x1)) + (3*x1) * ((3*x1) < (3)))) + (16*x2)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (1 + (3*x1))) + (1 + (3*x1)) * ((1 + (3*x1)) < (3)))) + (16*x2) + ((3) * ((3) <= (3*x0)) + (3*x0) * ((3*x0) < (3)))), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (1 + (3*x1))) + (1 + (3*x1)) * ((1 + (3*x1)) < (3)))) + (16*x2) + ((3) * ((3) <= (1 + (3*x0))) + (1 + (3*x0)) * ((1 + (3*x0)) < (3)))), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (1 + (3*x1))) + (1 + (3*x1)) * ((1 + (3*x1)) < (3)))) + (16*x2) + ((3) * ((3) <= (2 + (3*x0))) + (2 + (3*x0)) * ((2 + (3*x0)) < (3)))), xmask)
tmp13 = tl.load(in_ptr0 + (3 + (4*((3) * ((3) <= (1 + (3*x1))) + (1 + (3*x1)) * ((1 + (3*x1)) < (3)))) + (16*x2)), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (2 + (3*x1))) + (2 + (3*x1)) * ((2 + (3*x1)) < (3)))) + (16*x2) + ((3) * ((3) <= (3*x0)) + (3*x0) * ((3*x0) < (3)))), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (2 + (3*x1))) + (2 + (3*x1)) * ((2 + (3*x1)) < (3)))) + (16*x2) + ((3) * ((3) <= (1 + (3*x0))) + (1 + (3*x0)) * ((1 + (3*x0)) < (3)))), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr0 + ((4*((3) * ((3) <= (2 + (3*x1))) + (2 + (3*x1)) * ((2 + (3*x1)) < (3)))) + (16*x2) + ((3) * ((3) <= (2 + (3*x0))) + (2 + (3*x0)) * ((2 + (3*x0)) < (3)))), xmask)
tmp21 = tl.load(in_ptr0 + (3 + (4*((3) * ((3) <= (2 + (3*x1))) + (2 + (3*x1)) * ((2 + (3*x1)) < (3)))) + (16*x2)), xmask, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr0 + (12 + (16*x2) + ((3) * ((3) <= (3*x0)) + (3*x0) * ((3*x0) < (3)))), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr0 + (12 + (16*x2) + ((3) * ((3) <= (1 + (3*x0))) + (1 + (3*x0)) * ((1 + (3*x0)) < (3)))), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr0 + (12 + (16*x2) + ((3) * ((3) <= (2 + (3*x0))) + (2 + (3*x0)) * ((2 + (3*x0)) < (3)))), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr0 + (15 + (16*x2)), xmask, eviction_policy='evict_last')
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp8 = triton_helpers.maximum(tmp7, tmp6)
tmp10 = triton_helpers.maximum(tmp9, tmp8)
tmp12 = triton_helpers.maximum(tmp11, tmp10)
tmp14 = triton_helpers.maximum(tmp13, tmp12)
tmp16 = triton_helpers.maximum(tmp15, tmp14)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp20 = triton_helpers.maximum(tmp19, tmp18)
tmp22 = triton_helpers.maximum(tmp21, tmp20)
tmp24 = triton_helpers.maximum(tmp23, tmp22)
tmp26 = triton_helpers.maximum(tmp25, tmp24)
tmp28 = triton_helpers.maximum(tmp27, tmp26)
tmp30 = triton_helpers.maximum(tmp29, tmp28)
tl.store(out_ptr0 + (x4), tmp30, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [pooled_x], Original ATen: [aten.max_pool2d_with_indices]
stream0 = get_raw_stream(0)
triton_poi_fused_max_pool2d_with_indices_0.run(arg0_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch.nn.functional as F
import torch._utils
class MaxPoolStride1(nn.Module):
def __init__(self, kernel_size):
super(MaxPoolStride1, self).__init__()
self.kernel_size = kernel_size
self.pad = kernel_size - 1
def forward(self, x):
padded_x = F.pad(x, (0, self.pad, 0, self.pad), mode='replicate')
pooled_x = nn.MaxPool2d(self.kernel_size, self.pad)(padded_x)
return pooled_x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'kernel_size': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch._utils
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_0(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 2
x1 = xindex // 2 % 2
x2 = xindex // 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (4 * (3 * (3 <= 3 * x1) + 3 * x1 * (3 * x1 < 3
)) + 16 * x2 + (3 * (3 <= 3 * x0) + 3 * x0 * (3 * x0 < 3))), xmask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (4 * (3 * (3 <= 3 * x1) + 3 * x1 * (3 * x1 < 3
)) + 16 * x2 + (3 * (3 <= 1 + 3 * x0) + (1 + 3 * x0) * (1 + 3 * x0 <
3))), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (4 * (3 * (3 <= 3 * x1) + 3 * x1 * (3 * x1 < 3
)) + 16 * x2 + (3 * (3 <= 2 + 3 * x0) + (2 + 3 * x0) * (2 + 3 * x0 <
3))), xmask)
tmp5 = tl.load(in_ptr0 + (3 + 4 * (3 * (3 <= 3 * x1) + 3 * x1 * (3 * x1 <
3)) + 16 * x2), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (4 * (3 * (3 <= 1 + 3 * x1) + (1 + 3 * x1) * (
1 + 3 * x1 < 3)) + 16 * x2 + (3 * (3 <= 3 * x0) + 3 * x0 * (3 * x0 <
3))), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (4 * (3 * (3 <= 1 + 3 * x1) + (1 + 3 * x1) * (
1 + 3 * x1 < 3)) + 16 * x2 + (3 * (3 <= 1 + 3 * x0) + (1 + 3 * x0) *
(1 + 3 * x0 < 3))), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (4 * (3 * (3 <= 1 + 3 * x1) + (1 + 3 * x1) *
(1 + 3 * x1 < 3)) + 16 * x2 + (3 * (3 <= 2 + 3 * x0) + (2 + 3 * x0) *
(2 + 3 * x0 < 3))), xmask)
tmp13 = tl.load(in_ptr0 + (3 + 4 * (3 * (3 <= 1 + 3 * x1) + (1 + 3 * x1
) * (1 + 3 * x1 < 3)) + 16 * x2), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr0 + (4 * (3 * (3 <= 2 + 3 * x1) + (2 + 3 * x1) *
(2 + 3 * x1 < 3)) + 16 * x2 + (3 * (3 <= 3 * x0) + 3 * x0 * (3 * x0 <
3))), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (4 * (3 * (3 <= 2 + 3 * x1) + (2 + 3 * x1) *
(2 + 3 * x1 < 3)) + 16 * x2 + (3 * (3 <= 1 + 3 * x0) + (1 + 3 * x0) *
(1 + 3 * x0 < 3))), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr0 + (4 * (3 * (3 <= 2 + 3 * x1) + (2 + 3 * x1) *
(2 + 3 * x1 < 3)) + 16 * x2 + (3 * (3 <= 2 + 3 * x0) + (2 + 3 * x0) *
(2 + 3 * x0 < 3))), xmask)
tmp21 = tl.load(in_ptr0 + (3 + 4 * (3 * (3 <= 2 + 3 * x1) + (2 + 3 * x1
) * (2 + 3 * x1 < 3)) + 16 * x2), xmask, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr0 + (12 + 16 * x2 + (3 * (3 <= 3 * x0) + 3 * x0 *
(3 * x0 < 3))), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr0 + (12 + 16 * x2 + (3 * (3 <= 1 + 3 * x0) + (1 +
3 * x0) * (1 + 3 * x0 < 3))), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr0 + (12 + 16 * x2 + (3 * (3 <= 2 + 3 * x0) + (2 +
3 * x0) * (2 + 3 * x0 < 3))), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr0 + (15 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp8 = triton_helpers.maximum(tmp7, tmp6)
tmp10 = triton_helpers.maximum(tmp9, tmp8)
tmp12 = triton_helpers.maximum(tmp11, tmp10)
tmp14 = triton_helpers.maximum(tmp13, tmp12)
tmp16 = triton_helpers.maximum(tmp15, tmp14)
tmp18 = triton_helpers.maximum(tmp17, tmp16)
tmp20 = triton_helpers.maximum(tmp19, tmp18)
tmp22 = triton_helpers.maximum(tmp21, tmp20)
tmp24 = triton_helpers.maximum(tmp23, tmp22)
tmp26 = triton_helpers.maximum(tmp25, tmp24)
tmp28 = triton_helpers.maximum(tmp27, tmp26)
tmp30 = triton_helpers.maximum(tmp29, tmp28)
tl.store(out_ptr0 + x4, tmp30, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 2, 2), (16, 4, 2, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_max_pool2d_with_indices_0[grid(64)](arg0_1, buf0,
64, XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
return buf0,
class MaxPoolStride1New(nn.Module):
def __init__(self, kernel_size):
super(MaxPoolStride1New, self).__init__()
self.kernel_size = kernel_size
self.pad = kernel_size - 1
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Sarathismg/Pose-Estimator-Old-Version | MaxPoolStride1 | false | 1,062 | [
"Apache-2.0"
] | 0 | ecaa03769323b94a4d7222e2d3606d1ce92a2fae | https://github.com/Sarathismg/Pose-Estimator-Old-Version/tree/ecaa03769323b94a4d7222e2d3606d1ce92a2fae | import torch
import torch.nn as nn
import torch.utils.data
import torch.utils.data.distributed
import torch.nn.functional as F
import torch._utils
class Model(nn.Module):
def __init__(self, kernel_size):
super().__init__()
self.kernel_size = kernel_size
self.pad = kernel_size - 1
def forward(self, x):
padded_x = F.pad(x, (0, self.pad, 0, self.pad), mode='replicate')
pooled_x = nn.MaxPool2d(self.kernel_size, self.pad)(padded_x)
return pooled_x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
GroupNorm32 | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/zy/czyl6lkgkemly2rorp7dytnq5bsikpucv7yhz2var2hddv5xnkj4.py
# Topologically Sorted Source Nodes: [y, mul, sigmoid, y_1], Original ATen: [aten.native_group_norm, aten.mul, aten.sigmoid]
# Source node to ATen node mapping:
# mul => mul_2
# sigmoid => sigmoid
# y => add, add_1, mul_1, rsqrt, var_mean
# y_1 => mul_3
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %unsqueeze_5), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %unsqueeze_2), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_1, 4.0), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%mul_2,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_1, %sigmoid), kwargs = {})
triton_per_fused_mul_native_group_norm_sigmoid_0 = async_compile.triton('triton_per_fused_mul_native_group_norm_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mul_native_group_norm_sigmoid_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mul_native_group_norm_sigmoid_0(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
r3 = (rindex // 16)
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (r3), None, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + (r3), None, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = 64.0
tmp18 = tmp16 / tmp17
tmp19 = 1e-05
tmp20 = tmp18 + tmp19
tmp21 = libdevice.rsqrt(tmp20)
tmp22 = tmp0 - tmp10
tmp23 = tmp22 * tmp21
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp28 = 4.0
tmp29 = tmp27 * tmp28
tmp30 = tl.sigmoid(tmp29)
tmp31 = tmp27 * tmp30
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp21, xmask)
tl.store(in_out_ptr1 + (r1 + (64*x0)), tmp31, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 1), (1, 1, 1, 1), torch.float32)
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf3 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf1 # reuse
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [y, mul, sigmoid, y_1], Original ATen: [aten.native_group_norm, aten.mul, aten.sigmoid]
stream0 = get_raw_stream(0)
triton_per_fused_mul_native_group_norm_sigmoid_0.run(buf3, buf5, primals_1, primals_2, primals_3, buf0, 4, 64, grid=grid(4), stream=stream0)
return (buf5, primals_1, primals_2, primals_3, buf0, buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
class GroupNorm32(nn.GroupNorm):
def __init__(self, num_groups, num_channels, swish, eps=1e-05):
super().__init__(num_groups=num_groups, num_channels=num_channels,
eps=eps)
self.swish = swish
def forward(self, x):
y = super().forward(x.float())
if self.swish == 1.0:
y = F.silu(y)
elif self.swish:
y = y * F.sigmoid(y * float(self.swish))
return y
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_groups': 1, 'num_channels': 4, 'swish': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_mul_native_group_norm_sigmoid_0(in_out_ptr0,
in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, rnumel,
XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
r3 = rindex // 16
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + r3, None, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + r3, None, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tl.where(xmask, tmp1, 0)
tmp4 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp6 = tl.where(xmask, tmp4, 0)
tmp7 = tl.sum(tmp6, 1)[:, None]
tmp8 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tmp17 = 64.0
tmp18 = tmp16 / tmp17
tmp19 = 1e-05
tmp20 = tmp18 + tmp19
tmp21 = libdevice.rsqrt(tmp20)
tmp22 = tmp0 - tmp10
tmp23 = tmp22 * tmp21
tmp25 = tmp23 * tmp24
tmp27 = tmp25 + tmp26
tmp28 = 4.0
tmp29 = tmp27 * tmp28
tmp30 = tl.sigmoid(tmp29)
tmp31 = tmp27 * tmp30
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp21, xmask)
tl.store(in_out_ptr1 + (r1 + 64 * x0), tmp31, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 1), (1, 1, 1, 1), torch.float32)
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf3 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf1
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = buf4
del buf4
get_raw_stream(0)
triton_per_fused_mul_native_group_norm_sigmoid_0[grid(4)](buf3,
buf5, primals_1, primals_2, primals_3, buf0, 4, 64, XBLOCK=1,
num_warps=2, num_stages=1)
return buf5, primals_1, primals_2, primals_3, buf0, buf3
class GroupNorm32New(nn.GroupNorm):
def __init__(self, num_groups, num_channels, swish, eps=1e-05):
super().__init__(num_groups=num_groups, num_channels=num_channels,
eps=eps)
self.swish = swish
def forward(self, input_0):
primals_2 = self.weight
primals_3 = self.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| ShivanshuPurohit/Diffusion | GroupNorm32 | false | 1,063 | [
"MIT"
] | 0 | 9a190d9aa4ed9767cf223e4ef57d0c31690f92cc | https://github.com/ShivanshuPurohit/Diffusion/tree/9a190d9aa4ed9767cf223e4ef57d0c31690f92cc | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.GroupNorm):
def __init__(self, num_groups, num_channels, swish, eps=1e-05):
super().__init__(num_groups=num_groups, num_channels=num_channels,
eps=eps)
self.swish = swish
def forward(self, x):
y = super().forward(x.float())
if self.swish == 1.0:
y = F.silu(y)
elif self.swish:
y = y * F.sigmoid(y * float(self.swish))
return y
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [1, 4, 4]
|
HardSigmoid | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/n6/cn6mtix3xxpl4wrz6ijcstysgsoj75bwqjpwztauquzjojdcfz37.py
# Topologically Sorted Source Nodes: [mul, x, x_1, neg, result, neg_1, result_1], Original ATen: [aten.mul, aten.add, aten.clamp, aten.neg, aten.threshold]
# Source node to ATen node mapping:
# mul => mul
# neg => neg
# neg_1 => neg_1
# result => full_default, le, where
# result_1 => full_default_1, le_1, where_1
# x => add
# x_1 => clamp_max, clamp_min
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 0.2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 0.5), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%add, 0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 1), kwargs = {})
# %neg : [num_users=2] = call_function[target=torch.ops.aten.neg.default](args = (%clamp_max,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%neg, -1), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -1.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%le, %full_default, %neg), kwargs = {})
# %neg_1 : [num_users=2] = call_function[target=torch.ops.aten.neg.default](args = (%where,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%neg_1, 0), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where_1 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%le_1, %full_default_1, %neg_1), kwargs = {})
triton_poi_fused_add_clamp_mul_neg_threshold_0 = async_compile.triton('triton_poi_fused_add_clamp_mul_neg_threshold_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_clamp_mul_neg_threshold_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_clamp_mul_neg_threshold_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.2
tmp2 = tmp0 * tmp1
tmp3 = 0.5
tmp4 = tmp2 + tmp3
tmp5 = 0.0
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tmp7 = 1.0
tmp8 = triton_helpers.minimum(tmp6, tmp7)
tmp9 = -tmp8
tmp10 = -1.0
tmp11 = tmp9 <= tmp10
tmp12 = tl.where(tmp11, tmp10, tmp9)
tmp13 = -tmp12
tmp14 = tmp13 <= tmp5
tmp15 = tl.where(tmp14, tmp5, tmp13)
tl.store(out_ptr0 + (x0), tmp15, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, x, x_1, neg, result, neg_1, result_1], Original ATen: [aten.mul, aten.add, aten.clamp, aten.neg, aten.threshold]
stream0 = get_raw_stream(0)
triton_poi_fused_add_clamp_mul_neg_threshold_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
class HardSigmoid(torch.nn.Module):
"""
Pytorch implementation of the hard sigmoid activation function
"""
def __init__(self):
super(HardSigmoid, self).__init__()
def forward(self, input):
x = 0.2 * input + 0.5
x = torch.clamp(x, 0, 1)
x = F.threshold(-x, -1, -1)
x = F.threshold(-x, 0, 0)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_clamp_mul_neg_threshold_0(in_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.2
tmp2 = tmp0 * tmp1
tmp3 = 0.5
tmp4 = tmp2 + tmp3
tmp5 = 0.0
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tmp7 = 1.0
tmp8 = triton_helpers.minimum(tmp6, tmp7)
tmp9 = -tmp8
tmp10 = -1.0
tmp11 = tmp9 <= tmp10
tmp12 = tl.where(tmp11, tmp10, tmp9)
tmp13 = -tmp12
tmp14 = tmp13 <= tmp5
tmp15 = tl.where(tmp14, tmp5, tmp13)
tl.store(out_ptr0 + x0, tmp15, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_clamp_mul_neg_threshold_0[grid(256)](arg0_1,
buf0, 256, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class HardSigmoidNew(torch.nn.Module):
"""
Pytorch implementation of the hard sigmoid activation function
"""
def __init__(self):
super(HardSigmoidNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| ShiraLightricks/3d-photo-inpainting | HardSigmoid | false | 1,064 | [
"MIT"
] | 0 | c42ac41576690b765e50f5281ddbfb58439ff36d | https://github.com/ShiraLightricks/3d-photo-inpainting/tree/c42ac41576690b765e50f5281ddbfb58439ff36d | import torch
import torch.nn.functional as F
class Model(torch.nn.Module):
"""
Pytorch implementation of the hard sigmoid activation function
"""
def __init__(self):
super().__init__()
def forward(self, input):
x = 0.2 * input + 0.5
x = torch.clamp(x, 0, 1)
x = F.threshold(-x, -1, -1)
x = F.threshold(-x, 0, 0)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Classifier | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/l3/cl35tzbhrd24dhunkbb6gjs54aklpyr46oikqhoylcgmkcmhujil.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# out => mean
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [-1, -2], True), kwargs = {})
triton_per_fused_mean_0 = async_compile.triton('triton_per_fused_mean_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_0(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp5 = 16.0
tmp6 = tmp4 / tmp5
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_mean_0.run(buf1, primals_1, 16, 16, grid=grid(16), stream=stream0)
del primals_1
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(buf1, (4, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_2
del primals_3
return (buf2, reinterpret_tensor(buf1, (4, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
from abc import *
class Classifier(nn.Module):
def __init__(self, in_channels, num_classes):
super(Classifier, self).__init__()
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.fc = nn.Linear(in_channels, num_classes)
def forward(self, x):
out = self.avgpool(x)
out = torch.flatten(out, 1)
out = self.fc(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'num_classes': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
from abc import *
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_mean_0(in_out_ptr0, in_ptr0, xnumel, rnumel, XBLOCK:
tl.constexpr):
xnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp5 = 16.0
tmp6 = tmp4 / tmp5
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp6, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf1 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_mean_0[grid(16)](buf1, primals_1, 16, 16, XBLOCK=1,
num_warps=2, num_stages=1)
del primals_1
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(buf1, (4, 4), (4,
1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha
=1, beta=1, out=buf2)
del primals_2
del primals_3
return buf2, reinterpret_tensor(buf1, (4, 4), (4, 1), 0)
class ClassifierNew(nn.Module):
def __init__(self, in_channels, num_classes):
super(ClassifierNew, self).__init__()
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.fc = nn.Linear(in_channels, num_classes)
def forward(self, input_0):
primals_2 = self.fc.weight
primals_3 = self.fc.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| Slime0519/simple-faster-rcnn-pytorch | Classifier | false | 1,065 | [
"MIT"
] | 0 | 0503e9b4d07a24ae0bc1789a61ed937709f5304c | https://github.com/Slime0519/simple-faster-rcnn-pytorch/tree/0503e9b4d07a24ae0bc1789a61ed937709f5304c | import torch
import torch.nn as nn
from abc import *
class Model(nn.Module):
def __init__(self, in_channels, num_classes):
super().__init__()
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.fc = nn.Linear(in_channels, num_classes)
def forward(self, x):
out = self.avgpool(x)
out = torch.flatten(out, 1)
out = self.fc(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
TemporalAttention | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/7j/c7jy63he3fhmqd62giamt227766vsitw2rh2xrr6mmcphr7netrz.py
# Topologically Sorted Source Nodes: [add, add_1, tanh], Original ATen: [aten.add, aten.tanh]
# Source node to ATen node mapping:
# add => add
# add_1 => add_1
# tanh => tanh
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%expand, %view_3), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %primals_5), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add_1,), kwargs = {})
triton_poi_fused_add_tanh_0 = async_compile.triton('triton_poi_fused_add_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_tanh_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_tanh_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = (xindex // 64)
x4 = xindex % 16
x5 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x4 + (16*x3)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_out_ptr0 + (x5), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = libdevice.tanh(tmp4)
tl.store(in_out_ptr0 + (x5), tmp5, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/dm/cdmkcxuzpnailvibeivaikqdr4zvashgzwju7qijhq5aizlo3aor.py
# Topologically Sorted Source Nodes: [weights], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# weights => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_5, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_5, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 4
x2 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (4 + x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (8 + x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (12 + x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x3), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/kt/cktghousutx6xui2sl2rvevzmb7gkacvfhntjq5n2xzeu7v57oz6.py
# Topologically Sorted Source Nodes: [weights], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# weights => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 4
x2 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (4 + x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (8 + x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (12 + x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/de/cde5wp4wyi2tqgfwsvoxmxb77ndgieetn2kkhp3yursmtfolgix6.py
# Topologically Sorted Source Nodes: [weighted_feats, attn_feats], Original ATen: [aten.mul, aten.sum]
# Source node to ATen node mapping:
# attn_feats => sum_2
# weighted_feats => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_4, %div), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
triton_poi_fused_mul_sum_3 = async_compile.triton('triton_poi_fused_mul_sum_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_sum_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_sum_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 16)
x3 = xindex % 16
x1 = (xindex // 4) % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x3 + (64*x2)), xmask)
tmp1 = tl.load(in_ptr1 + (x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x3 + (64*x2)), xmask)
tmp4 = tl.load(in_ptr1 + (4 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (32 + x3 + (64*x2)), xmask)
tmp8 = tl.load(in_ptr1 + (8 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (48 + x3 + (64*x2)), xmask)
tmp12 = tl.load(in_ptr1 + (12 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 * tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 * tmp12
tmp14 = tmp10 + tmp13
tl.store(out_ptr0 + (x4), tmp14, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [Wh], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [Uv], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [add, add_1, tanh], Original ATen: [aten.add, aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_add_tanh_0.run(buf2, buf0, primals_5, 256, grid=grid(256), stream=stream0)
del primals_5
buf3 = reinterpret_tensor(buf0, (64, 1), (1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [energies], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf2, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 1), (1, 4), 0), out=buf3)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [weights], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = reinterpret_tensor(buf3, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [weights], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf4, buf5, 64, grid=grid(64), stream=stream0)
buf6 = reinterpret_tensor(buf4, (4, 4, 4), (16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [weighted_feats, attn_feats], Original ATen: [aten.mul, aten.sum]
triton_poi_fused_mul_sum_3.run(primals_4, buf5, buf6, 64, grid=grid(64), stream=stream0)
return (buf6, buf5, primals_4, reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), buf2, buf5, primals_6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
class TemporalAttention(nn.Module):
def __init__(self, hidden_size, feat_size, bottleneck_size):
super(TemporalAttention, self).__init__()
self.hidden_size = hidden_size
self.feat_size = feat_size
self.bottleneck_size = bottleneck_size
self.W = nn.Linear(self.hidden_size, self.bottleneck_size, bias=False)
self.U = nn.Linear(self.feat_size, self.bottleneck_size, bias=False)
self.b = nn.Parameter(torch.ones(self.bottleneck_size),
requires_grad=True)
self.w = nn.Linear(self.bottleneck_size, 1, bias=False)
def forward(self, hidden, feats, masks=None):
Wh = self.W(hidden)
Uv = self.U(feats)
Wh = Wh.unsqueeze(1).expand_as(Uv)
energies = self.w(torch.tanh(Wh + Uv + self.b))
if masks is not None:
energies = energies.squeeze(2)
energies[~masks] = -float('inf')
energies = energies.unsqueeze(2)
weights = F.softmax(energies, dim=1)
weighted_feats = feats * weights
attn_feats = weighted_feats.sum(dim=1)
return attn_feats, weights
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'hidden_size': 4, 'feat_size': 4, 'bottleneck_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_tanh_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex // 64
x4 = xindex % 16
x5 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x4 + 16 * x3), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_out_ptr0 + x5, xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = libdevice.tanh(tmp4)
tl.store(in_out_ptr0 + x5, tmp5, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 4
x2 = xindex // 16
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (4 + x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (8 + x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (12 + x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x3, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 4
x2 = xindex // 16
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (4 + x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (8 + x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (12 + x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_poi_fused_mul_sum_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 16
x3 = xindex % 16
x1 = xindex // 4 % 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x3 + 64 * x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x3 + 64 * x2), xmask)
tmp4 = tl.load(in_ptr1 + (4 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp7 = tl.load(in_ptr0 + (32 + x3 + 64 * x2), xmask)
tmp8 = tl.load(in_ptr1 + (8 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp11 = tl.load(in_ptr0 + (48 + x3 + 64 * x2), xmask)
tmp12 = tl.load(in_ptr1 + (12 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 * tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 * tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 * tmp12
tmp14 = tmp10 + tmp13
tl.store(out_ptr0 + x4, tmp14, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf1
get_raw_stream(0)
triton_poi_fused_add_tanh_0[grid(256)](buf2, buf0, primals_5, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf3 = reinterpret_tensor(buf0, (64, 1), (1, 1), 0)
del buf0
extern_kernels.mm(reinterpret_tensor(buf2, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 1), (1, 4), 0), out=buf3)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_1[grid(64)](buf3, buf4, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf5 = reinterpret_tensor(buf3, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf3
triton_poi_fused__softmax_2[grid(64)](buf4, buf5, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf6 = reinterpret_tensor(buf4, (4, 4, 4), (16, 4, 1), 0)
del buf4
triton_poi_fused_mul_sum_3[grid(64)](primals_4, buf5, buf6, 64,
XBLOCK=64, num_warps=1, num_stages=1)
return buf6, buf5, primals_4, reinterpret_tensor(primals_2, (16, 4), (4,
1), 0), buf2, buf5, primals_6
class TemporalAttentionNew(nn.Module):
def __init__(self, hidden_size, feat_size, bottleneck_size):
super(TemporalAttentionNew, self).__init__()
self.hidden_size = hidden_size
self.feat_size = feat_size
self.bottleneck_size = bottleneck_size
self.W = nn.Linear(self.hidden_size, self.bottleneck_size, bias=False)
self.U = nn.Linear(self.feat_size, self.bottleneck_size, bias=False)
self.b = nn.Parameter(torch.ones(self.bottleneck_size),
requires_grad=True)
self.w = nn.Linear(self.bottleneck_size, 1, bias=False)
def forward(self, input_0, input_1):
primals_5 = self.b
primals_1 = self.W.weight
primals_3 = self.U.weight
primals_6 = self.w.weight
primals_2 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6])
return output[0], output[1]
| Shashwat07gupta/MSVD | TemporalAttention | false | 1,066 | [
"MIT"
] | 0 | 8026557ef7681a504b5140560ec4aaad9944de2d | https://github.com/Shashwat07gupta/MSVD/tree/8026557ef7681a504b5140560ec4aaad9944de2d | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, hidden_size, feat_size, bottleneck_size):
super().__init__()
self.hidden_size = hidden_size
self.feat_size = feat_size
self.bottleneck_size = bottleneck_size
self.W = nn.Linear(self.hidden_size, self.bottleneck_size, bias=False)
self.U = nn.Linear(self.feat_size, self.bottleneck_size, bias=False)
self.b = nn.Parameter(torch.ones(self.bottleneck_size),
requires_grad=True)
self.w = nn.Linear(self.bottleneck_size, 1, bias=False)
def forward(self, hidden, feats, masks=None):
Wh = self.W(hidden)
Uv = self.U(feats)
Wh = Wh.unsqueeze(1).expand_as(Uv)
energies = self.w(torch.tanh(Wh + Uv + self.b))
if masks is not None:
energies = energies.squeeze(2)
energies[~masks] = -float('inf')
energies = energies.unsqueeze(2)
weights = F.softmax(energies, dim=1)
weighted_feats = feats * weights
attn_feats = weighted_feats.sum(dim=1)
return attn_feats, weights
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
FastRNNCell | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/tb/ctbc62iwsfjylunbfj67j7mqt6piotcv6s56drmljjg2lp7vfyhv.py
# Topologically Sorted Source Nodes: [pre_comp, add_1, c, sigmoid, mul, sigmoid_1, mul_1, new_h], Original ATen: [aten.add, aten.tanh, aten.sigmoid, aten.mul]
# Source node to ATen node mapping:
# add_1 => add_1
# c => tanh
# mul => mul
# mul_1 => mul_1
# new_h => add_2
# pre_comp => add
# sigmoid => sigmoid
# sigmoid_1 => sigmoid_1
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %view_3), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %primals_5), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add_1,), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_6,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_4), kwargs = {})
# %sigmoid_1 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_7,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %tanh), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
triton_poi_fused_add_mul_sigmoid_tanh_0 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_tanh_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_tanh_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr2 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp9 = tl.load(in_ptr3 + (x2), xmask)
tmp11 = tl.load(in_ptr4 + (0))
tmp12 = tl.broadcast_to(tmp11, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = libdevice.tanh(tmp4)
tmp8 = tl.sigmoid(tmp7)
tmp10 = tmp8 * tmp9
tmp13 = tl.sigmoid(tmp12)
tmp14 = tmp13 * tmp5
tmp15 = tmp10 + tmp14
tl.store(in_out_ptr0 + (x2), tmp5, xmask)
tl.store(out_ptr0 + (x2), tmp15, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (1, 4), (4, 1))
assert_size_stride(primals_6, (1, 1), (1, 1))
assert_size_stride(primals_7, (1, 1), (1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [uComp], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0), primals_3, out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pre_comp, add_1, c, sigmoid, mul, sigmoid_1, mul_1, new_h], Original ATen: [aten.add, aten.tanh, aten.sigmoid, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_sigmoid_tanh_0.run(buf2, buf1, primals_5, primals_6, primals_4, primals_7, buf3, 256, grid=grid(256), stream=stream0)
del buf1
del primals_5
return (buf3, primals_4, primals_6, primals_7, buf2, reinterpret_tensor(primals_2, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, 1), (1, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((1, 1), (1, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class FastRNNCell(RNNCell):
"""
FastRNN Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix (creates two matrices if not None)
uRank = rank of U matrix (creates two matrices if not None)
wSparsity = intended sparsity of W matrix(ces)
uSparsity = intended sparsity of U matrix(ces)
Warning:
The Cell will not automatically sparsify.
The user must invoke .sparsify to hard threshold.
alphaInit = init for alpha, the update scalar
betaInit = init for beta, the weight for previous state
FastRNN architecture and compression techniques are found in
FastGRNN(LINK) paper
Basic architecture is like:
h_t^ = update_nl(Wx_t + Uh_{t-1} + B_h)
h_t = sigmoid(beta)*h_{t-1} + sigmoid(alpha)*h_t^
W and U can further parameterised into low rank version by
W = matmul(W_1, W_2) and U = matmul(U_1, U_2)
"""
def __init__(self, input_size, hidden_size, update_nonlinearity='tanh',
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0, alphaInit=-
3.0, betaInit=3.0, name='FastRNN'):
super(FastRNNCell, self).__init__(input_size, hidden_size, None,
update_nonlinearity, 1, 1, 1, wRank, uRank, wSparsity, uSparsity)
self._alphaInit = alphaInit
self._betaInit = betaInit
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W = nn.Parameter(0.1 * torch.randn([input_size, hidden_size]))
else:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, hidden_size])
)
else:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self.alpha = nn.Parameter(self._alphaInit * torch.ones([1, 1]))
self.beta = nn.Parameter(self._betaInit * torch.ones([1, 1]))
@property
def name(self):
return self._name
@property
def cellType(self):
return 'FastRNN'
def forward(self, input, state):
if self._wRank is None:
wComp = torch.matmul(input, self.W)
else:
wComp = torch.matmul(torch.matmul(input, self.W1), self.W2)
if self._uRank is None:
uComp = torch.matmul(state, self.U)
else:
uComp = torch.matmul(torch.matmul(state, self.U1), self.U2)
pre_comp = wComp + uComp
c = gen_nonlinearity(pre_comp + self.bias_update, self.
_update_nonlinearity)
new_h = torch.sigmoid(self.beta) * state + torch.sigmoid(self.alpha
) * c
return new_h
def getVars(self):
Vars = []
if self._num_W_matrices == 1:
Vars.append(self.W)
else:
Vars.extend([self.W1, self.W2])
if self._num_U_matrices == 1:
Vars.append(self.U)
else:
Vars.extend([self.U1, self.U2])
Vars.extend([self.bias_update])
Vars.extend([self.alpha, self.beta])
return Vars
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.onnx
from itertools import product as product
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_mul_sigmoid_tanh_0(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr2 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp9 = tl.load(in_ptr3 + x2, xmask)
tmp11 = tl.load(in_ptr4 + 0)
tmp12 = tl.broadcast_to(tmp11, [XBLOCK])
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = libdevice.tanh(tmp4)
tmp8 = tl.sigmoid(tmp7)
tmp10 = tmp8 * tmp9
tmp13 = tl.sigmoid(tmp12)
tmp14 = tmp13 * tmp5
tmp15 = tmp10 + tmp14
tl.store(in_out_ptr0 + x2, tmp5, xmask)
tl.store(out_ptr0 + x2, tmp15, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (1, 4), (4, 1))
assert_size_stride(primals_6, (1, 1), (1, 1))
assert_size_stride(primals_7, (1, 1), (1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0),
primals_3, out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_sigmoid_tanh_0[grid(256)](buf2, buf1,
primals_5, primals_6, primals_4, primals_7, buf3, 256, XBLOCK=
128, num_warps=4, num_stages=1)
del buf1
del primals_5
return buf3, primals_4, primals_6, primals_7, buf2, reinterpret_tensor(
primals_2, (4, 64), (1, 4), 0)
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class FastRNNCellNew(RNNCell):
"""
FastRNN Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix (creates two matrices if not None)
uRank = rank of U matrix (creates two matrices if not None)
wSparsity = intended sparsity of W matrix(ces)
uSparsity = intended sparsity of U matrix(ces)
Warning:
The Cell will not automatically sparsify.
The user must invoke .sparsify to hard threshold.
alphaInit = init for alpha, the update scalar
betaInit = init for beta, the weight for previous state
FastRNN architecture and compression techniques are found in
FastGRNN(LINK) paper
Basic architecture is like:
h_t^ = update_nl(Wx_t + Uh_{t-1} + B_h)
h_t = sigmoid(beta)*h_{t-1} + sigmoid(alpha)*h_t^
W and U can further parameterised into low rank version by
W = matmul(W_1, W_2) and U = matmul(U_1, U_2)
"""
def __init__(self, input_size, hidden_size, update_nonlinearity='tanh',
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0, alphaInit=-
3.0, betaInit=3.0, name='FastRNN'):
super(FastRNNCellNew, self).__init__(input_size, hidden_size, None,
update_nonlinearity, 1, 1, 1, wRank, uRank, wSparsity, uSparsity)
self._alphaInit = alphaInit
self._betaInit = betaInit
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W = nn.Parameter(0.1 * torch.randn([input_size, hidden_size]))
else:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, hidden_size])
)
else:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self.alpha = nn.Parameter(self._alphaInit * torch.ones([1, 1]))
self.beta = nn.Parameter(self._betaInit * torch.ones([1, 1]))
@property
def name(self):
return self._name
@property
def cellType(self):
return 'FastRNN'
def getVars(self):
Vars = []
if self._num_W_matrices == 1:
Vars.append(self.W)
else:
Vars.extend([self.W1, self.W2])
if self._num_U_matrices == 1:
Vars.append(self.U)
else:
Vars.extend([self.U1, self.U2])
Vars.extend([self.bias_update])
Vars.extend([self.alpha, self.beta])
return Vars
def forward(self, input_0, input_1):
primals_1 = self.W
primals_3 = self.U
primals_5 = self.bias_update
primals_6 = self.alpha
primals_7 = self.beta
primals_2 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| ShishirPatil/EdgeML-1 | FastRNNCell | false | 1,067 | [
"MIT"
] | 0 | cbba9f8b989e545788427c004eb8450e7e4c1a21 | https://github.com/ShishirPatil/EdgeML-1/tree/cbba9f8b989e545788427c004eb8450e7e4c1a21 | import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super().__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(to
# ... truncated (>4000 chars) for memory efficiency |
Downsample | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/cu/ccutvo2v4333pq6xhrg2zryqqwthm7dmmuqprvva2xdwiodpz5jn.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [2, 2], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 4) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(2, 2), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 2, 2), (16, 4, 2, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf1, primals_2, 64, grid=grid(64), stream=stream0)
del primals_2
return (buf1, primals_1, primals_3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import numpy as np
class BaseModule(torch.nn.Module):
def __init__(self):
super(BaseModule, self).__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class Downsample(BaseModule):
def __init__(self, dim):
super(Downsample, self).__init__()
self.conv = torch.nn.Conv2d(dim, dim, 3, 2, 1)
def forward(self, x):
return self.conv(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import numpy as np
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 4 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(2,
2), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 2, 2), (16, 4, 2, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(64)](buf1, primals_2, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_2
return buf1, primals_1, primals_3
class BaseModule(torch.nn.Module):
def __init__(self):
super(BaseModule, self).__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class DownsampleNew(BaseModule):
def __init__(self, dim):
super(DownsampleNew, self).__init__()
self.conv = torch.nn.Conv2d(dim, dim, 3, 2, 1)
def forward(self, input_0):
primals_1 = self.conv.weight
primals_2 = self.conv.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| Sobsz/uberduck-ml-dev | Downsample | false | 1,068 | [
"Apache-2.0"
] | 0 | f099238f6f2e3f600d72d89dea3c883c59d91387 | https://github.com/Sobsz/uberduck-ml-dev/tree/f099238f6f2e3f600d72d89dea3c883c59d91387 | import torch
import numpy as np
class BaseModule(torch.nn.Module):
def __init__(self):
super().__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class Model(BaseModule):
def __init__(self, dim):
super().__init__()
self.conv = torch.nn.Conv2d(dim, dim, 3, 2, 1)
def forward(self, x):
return self.conv(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
AddFunction | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/zi/czioyfiql36jvbru3amu3iggyuvnn5c4pypwuaiss36muc2jqtqb.py
# Topologically Sorted Source Nodes: [add], Original ATen: [aten.add]
# Source node to ATen node mapping:
# add => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
triton_poi_fused_add_0 = async_compile.triton('triton_poi_fused_add_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add], Original ATen: [aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_0.run(arg0_1, arg1_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
del arg1_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
class AddFunction(nn.Module):
def __init__(self):
super(AddFunction, self).__init__()
def forward(self, x, y):
return x + y
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask)
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_0[grid(256)](arg0_1, arg1_1, buf0, 256, XBLOCK
=128, num_warps=4, num_stages=1)
del arg0_1
del arg1_1
return buf0,
class AddFunctionNew(nn.Module):
def __init__(self):
super(AddFunctionNew, self).__init__()
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| ShounoLab/res-net-interpretation-open | AddFunction | false | 1,069 | [
"MIT"
] | 0 | 282dc0ae261467ee1866996416149959db216c02 | https://github.com/ShounoLab/res-net-interpretation-open/tree/282dc0ae261467ee1866996416149959db216c02 | import torch
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
class Model(nn.Module):
def __init__(self):
super().__init__()
def forward(self, x, y):
return x + y
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
WeightedCE | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/td/ctdj5kazgiki6gdaadhqtp2x7tq2ee5ey5hqqdcoqmp54jyhf74f.py
# Topologically Sorted Source Nodes: [loss], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# loss => amax, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg1_1, [1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %amax), kwargs = {})
triton_poi_fused__log_softmax_0 = async_compile.triton('triton_poi_fused__log_softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/hl/chlfdbgpszjgvc5lbbjy2patr43syrsegknoj2ftqczeybmrnw76.py
# Topologically Sorted Source Nodes: [loss, mean], Original ATen: [aten._log_softmax, aten.mul, aten.sum, aten.neg, aten.mean]
# Source node to ATen node mapping:
# loss => exp, log, mul, neg, sub_1, sum_1, sum_2
# mean => mean
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %log), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %arg0_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%sum_2,), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%neg,), kwargs = {})
triton_per_fused__log_softmax_mean_mul_neg_sum_1 = async_compile.triton('triton_per_fused__log_softmax_mean_mul_neg_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__log_softmax_mean_mul_neg_sum_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__log_softmax_mean_mul_neg_sum_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex % 16
r1 = (rindex // 16)
tmp0 = tl.load(in_ptr0 + (r0 + (64*r1)), None)
tmp2 = tl.load(in_ptr0 + (16 + r0 + (64*r1)), None)
tmp5 = tl.load(in_ptr0 + (32 + r0 + (64*r1)), None)
tmp8 = tl.load(in_ptr0 + (48 + r0 + (64*r1)), None)
tmp13 = tl.load(in_ptr1 + (r0 + (64*r1)), None)
tmp16 = tl.load(in_ptr1 + (16 + r0 + (64*r1)), None)
tmp20 = tl.load(in_ptr1 + (32 + r0 + (64*r1)), None)
tmp24 = tl.load(in_ptr1 + (48 + r0 + (64*r1)), None)
tmp1 = tl_math.exp(tmp0)
tmp3 = tl_math.exp(tmp2)
tmp4 = tmp1 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp14 = tmp12 * tmp13
tmp15 = tmp2 - tmp11
tmp17 = tmp15 * tmp16
tmp18 = tmp14 + tmp17
tmp19 = tmp5 - tmp11
tmp21 = tmp19 * tmp20
tmp22 = tmp18 + tmp21
tmp23 = tmp8 - tmp11
tmp25 = tmp23 * tmp24
tmp26 = tmp22 + tmp25
tmp27 = -tmp26
tmp28 = tl.broadcast_to(tmp27, [XBLOCK, RBLOCK])
tmp30 = tl.sum(tmp28, 1)[:, None]
tmp31 = 64.0
tmp32 = tmp30 / tmp31
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp32, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [loss], Original ATen: [aten._log_softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax_0.run(arg1_1, buf0, 256, grid=grid(256), stream=stream0)
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [loss, mean], Original ATen: [aten._log_softmax, aten.mul, aten.sum, aten.neg, aten.mean]
triton_per_fused__log_softmax_mean_mul_neg_sum_1.run(buf2, buf0, arg0_1, 1, 64, grid=grid(1), stream=stream0)
del arg0_1
del buf0
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
import torch.nn.functional as F
class WeightedCE(nn.Module):
"""Mask weighted multi-class cross-entropy (CE) loss.
"""
def __init__(self):
super().__init__()
def forward(self, pred, target, weight_mask=None):
loss = F.cross_entropy(pred, target, reduction='none')
if weight_mask is not None:
loss = loss * weight_mask
return loss.mean()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_per_fused__log_softmax_mean_mul_neg_sum_1(in_out_ptr0, in_ptr0,
in_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex % 16
r1 = rindex // 16
tmp0 = tl.load(in_ptr0 + (r0 + 64 * r1), None)
tmp2 = tl.load(in_ptr0 + (16 + r0 + 64 * r1), None)
tmp5 = tl.load(in_ptr0 + (32 + r0 + 64 * r1), None)
tmp8 = tl.load(in_ptr0 + (48 + r0 + 64 * r1), None)
tmp13 = tl.load(in_ptr1 + (r0 + 64 * r1), None)
tmp16 = tl.load(in_ptr1 + (16 + r0 + 64 * r1), None)
tmp20 = tl.load(in_ptr1 + (32 + r0 + 64 * r1), None)
tmp24 = tl.load(in_ptr1 + (48 + r0 + 64 * r1), None)
tmp1 = tl_math.exp(tmp0)
tmp3 = tl_math.exp(tmp2)
tmp4 = tmp1 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp14 = tmp12 * tmp13
tmp15 = tmp2 - tmp11
tmp17 = tmp15 * tmp16
tmp18 = tmp14 + tmp17
tmp19 = tmp5 - tmp11
tmp21 = tmp19 * tmp20
tmp22 = tmp18 + tmp21
tmp23 = tmp8 - tmp11
tmp25 = tmp23 * tmp24
tmp26 = tmp22 + tmp25
tmp27 = -tmp26
tmp28 = tl.broadcast_to(tmp27, [XBLOCK, RBLOCK])
tmp30 = tl.sum(tmp28, 1)[:, None]
tmp31 = 64.0
tmp32 = tmp30 / tmp31
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp32, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax_0[grid(256)](arg1_1, buf0, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
triton_per_fused__log_softmax_mean_mul_neg_sum_1[grid(1)](buf2,
buf0, arg0_1, 1, 64, XBLOCK=1, num_warps=2, num_stages=1)
del arg0_1
del buf0
return buf2,
class WeightedCENew(nn.Module):
"""Mask weighted multi-class cross-entropy (CE) loss.
"""
def __init__(self):
super().__init__()
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| Shray64/pytorch_connectomics | WeightedCE | false | 1,070 | [
"MIT"
] | 0 | d6c814f11ac2f8418ede5ae220a93016f50214fc | https://github.com/Shray64/pytorch_connectomics/tree/d6c814f11ac2f8418ede5ae220a93016f50214fc | import torch
import torch.nn as nn
import torch.utils.data
import torch.nn.functional as F
class Model(nn.Module):
"""Mask weighted multi-class cross-entropy (CE) loss.
"""
def __init__(self):
super().__init__()
def forward(self, pred, target, weight_mask=None):
loss = F.cross_entropy(pred, target, reduction='none')
if weight_mask is not None:
loss = loss * weight_mask
return loss.mean()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
PartialConv | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/p3/cp3qleddjiuuytozrtebx5pzf2ycpwtw4mkq2jsx7qqswymv2bm6.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, %primals_2), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask)
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/fj/cfjzvvqpec3bkltw3iqoicgjjjbjt5sjrhrkdgaltmhp7rqzs5eb.py
# Topologically Sorted Source Nodes: [output, no_update_holes, mask_sum, sub, mul_1, truediv, output_pre, output_1, new_mask, new_mask_1], Original ATen: [aten.convolution, aten.eq, aten.masked_fill, aten.sub, aten.mul, aten.div, aten.add, aten.ones_like]
# Source node to ATen node mapping:
# mask_sum => full_default, where
# mul_1 => mul_1
# new_mask => full_default_2
# new_mask_1 => where_2
# no_update_holes => eq
# output => convolution
# output_1 => full_default_1, where_1
# output_pre => add
# sub => sub
# truediv => div
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%mul, %primals_3, %primals_4, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %eq : [num_users=3] = call_function[target=torch.ops.aten.eq.Scalar](args = (%convolution_1, 0), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 1.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %convolution_1), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%convolution, %expand), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, 64), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_1, %where), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div, %expand), kwargs = {})
# %full_default_1 : [num_users=2] = call_function[target=torch.ops.aten.full.default](args = ([], 0.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where_1 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default_1, %add), kwargs = {})
# %full_default_2 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 1, 1], 1), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where_2 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default_1, %full_default_2), kwargs = {})
triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1 = async_compile.triton('triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_out_ptr0 + (x2), xmask)
tmp4 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp1 = 0.0
tmp2 = tmp0 == tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp5 - tmp4
tmp7 = 64.0
tmp8 = tmp6 * tmp7
tmp9 = 1.0
tmp10 = tl.where(tmp2, tmp9, tmp0)
tmp11 = tmp8 / tmp10
tmp12 = tmp11 + tmp4
tmp13 = tl.where(tmp2, tmp1, tmp12)
tmp14 = tl.where(tmp2, tmp1, tmp9)
tl.store(in_out_ptr0 + (x2), tmp13, xmask)
tl.store(out_ptr0 + (x2), tmp14, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_3, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 1, 1), (4, 1, 1, 1))
# Topologically Sorted Source Nodes: [output_mask], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_2, primals_5, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 1, 1), (4, 1, 1, 1))
del primals_2
del primals_5
buf3 = buf1; del buf1 # reuse
buf4 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [output, no_update_holes, mask_sum, sub, mul_1, truediv, output_pre, output_1, new_mask, new_mask_1], Original ATen: [aten.convolution, aten.eq, aten.masked_fill, aten.sub, aten.mul, aten.div, aten.add, aten.ones_like]
triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1.run(buf3, buf2, primals_4, buf4, 16, grid=grid(16), stream=stream0)
del primals_4
return (buf3, buf4, primals_3, buf0, buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn as nn
def weights_init(init_type='gaussian'):
def init_fun(m):
classname = m.__class__.__name__
if (classname.find('Conv') == 0 or classname.find('Linear') == 0
) and hasattr(m, 'weight'):
if init_type == 'gaussian':
nn.init.normal_(m.weight, 0.0, 0.02)
elif init_type == 'xavier':
nn.init.xavier_normal_(m.weight, gain=math.sqrt(2))
elif init_type == 'kaiming':
nn.init.kaiming_normal_(m.weight, a=0, mode='fan_in')
elif init_type == 'orthogonal':
nn.init.orthogonal_(m.weight, gain=math.sqrt(2))
elif init_type == 'default':
pass
else:
assert 0, 'Unsupported initialization: {}'.format(init_type)
if hasattr(m, 'bias') and m.bias is not None:
nn.init.constant_(m.bias, 0.0)
return init_fun
class PartialConv(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, groups=1, bias=True):
super().__init__()
self.input_conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, bias)
self.mask_conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, False)
self.input_conv.apply(weights_init('kaiming'))
self.slide_winsize = in_channels * kernel_size * kernel_size
torch.nn.init.constant_(self.mask_conv.weight, 1.0)
for param in self.mask_conv.parameters():
param.requires_grad = False
def forward(self, input, mask):
output = self.input_conv(input * mask)
if self.input_conv.bias is not None:
output_bias = self.input_conv.bias.view(1, -1, 1, 1).expand_as(
output)
else:
output_bias = torch.zeros_like(output)
with torch.no_grad():
output_mask = self.mask_conv(mask)
no_update_holes = output_mask == 0
mask_sum = output_mask.masked_fill_(no_update_holes, 1.0)
output_pre = (output - output_bias
) * self.slide_winsize / mask_sum + output_bias
output = output_pre.masked_fill_(no_update_holes, 0.0)
new_mask = torch.ones_like(output)
new_mask = new_mask.masked_fill_(no_update_holes, 0.0)
return output, new_mask
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask)
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1(
in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_out_ptr0 + x2, xmask)
tmp4 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp1 = 0.0
tmp2 = tmp0 == tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp5 - tmp4
tmp7 = 64.0
tmp8 = tmp6 * tmp7
tmp9 = 1.0
tmp10 = tl.where(tmp2, tmp9, tmp0)
tmp11 = tmp8 / tmp10
tmp12 = tmp11 + tmp4
tmp13 = tl.where(tmp2, tmp1, tmp12)
tmp14 = tl.where(tmp2, tmp1, tmp9)
tl.store(in_out_ptr0 + x2, tmp13, xmask)
tl.store(out_ptr0 + x2, tmp14, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(256)](primals_1, primals_2, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_3, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 1, 1), (4, 1, 1, 1))
buf2 = extern_kernels.convolution(primals_2, primals_5, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 1, 1), (4, 1, 1, 1))
del primals_2
del primals_5
buf3 = buf1
del buf1
buf4 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
triton_poi_fused_add_convolution_div_eq_masked_fill_mul_ones_like_sub_1[
grid(16)](buf3, buf2, primals_4, buf4, 16, XBLOCK=16, num_warps
=1, num_stages=1)
del primals_4
return buf3, buf4, primals_3, buf0, buf2
def weights_init(init_type='gaussian'):
def init_fun(m):
classname = m.__class__.__name__
if (classname.find('Conv') == 0 or classname.find('Linear') == 0
) and hasattr(m, 'weight'):
if init_type == 'gaussian':
nn.init.normal_(m.weight, 0.0, 0.02)
elif init_type == 'xavier':
nn.init.xavier_normal_(m.weight, gain=math.sqrt(2))
elif init_type == 'kaiming':
nn.init.kaiming_normal_(m.weight, a=0, mode='fan_in')
elif init_type == 'orthogonal':
nn.init.orthogonal_(m.weight, gain=math.sqrt(2))
elif init_type == 'default':
pass
else:
assert 0, 'Unsupported initialization: {}'.format(init_type)
if hasattr(m, 'bias') and m.bias is not None:
nn.init.constant_(m.bias, 0.0)
return init_fun
class PartialConvNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, groups=1, bias=True):
super().__init__()
self.input_conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, bias)
self.mask_conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, False)
self.input_conv.apply(weights_init('kaiming'))
self.slide_winsize = in_channels * kernel_size * kernel_size
torch.nn.init.constant_(self.mask_conv.weight, 1.0)
for param in self.mask_conv.parameters():
param.requires_grad = False
def forward(self, input_0, input_1):
primals_1 = self.input_conv.weight
primals_4 = self.input_conv.bias
primals_2 = self.mask_conv.weight
primals_3 = input_0
primals_5 = input_1
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0], output[1]
| ShiraLightricks/3d-photo-inpainting | PartialConv | false | 1,071 | [
"MIT"
] | 0 | c42ac41576690b765e50f5281ddbfb58439ff36d | https://github.com/ShiraLightricks/3d-photo-inpainting/tree/c42ac41576690b765e50f5281ddbfb58439ff36d | import math
import torch
import torch.nn as nn
def weights_init(init_type='gaussian'):
def init_fun(m):
classname = m.__class__.__name__
if (classname.find('Conv') == 0 or classname.find('Linear') == 0
) and hasattr(m, 'weight'):
if init_type == 'gaussian':
nn.init.normal_(m.weight, 0.0, 0.02)
elif init_type == 'xavier':
nn.init.xavier_normal_(m.weight, gain=math.sqrt(2))
elif init_type == 'kaiming':
nn.init.kaiming_normal_(m.weight, a=0, mode='fan_in')
elif init_type == 'orthogonal':
nn.init.orthogonal_(m.weight, gain=math.sqrt(2))
elif init_type == 'default':
pass
else:
assert 0, 'Unsupported initialization: {}'.format(init_type)
if hasattr(m, 'bias') and m.bias is not None:
nn.init.constant_(m.bias, 0.0)
return init_fun
class Model(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, groups=1, bias=True):
super().__init__()
self.input_conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, bias)
self.mask_conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, False)
self.input_conv.apply(weights_init('kaiming'))
self.slide_winsize = in_channels * kernel_size * kernel_size
torch.nn.init.constant_(self.mask_conv.weight, 1.0)
for param in self.mask_conv.parameters():
param.requires_grad = False
def forward(self, input, mask):
output = self.input_conv(input * mask)
if self.input_conv.bias is not None:
output_bias = self.input_conv.bias.view(1, -1, 1, 1).expand_as(
output)
else:
output_bias = torch.zeros_like(output)
with torch.no_grad():
output_mask = self.mask_conv(mask)
no_update_holes = output_mask == 0
mask_sum = output_mask.masked_fill_(no_update_holes, 1.0)
output_pre = (output - output_bias
) * self.slide_winsize / mask_sum + output_bias
output = output_pre.masked_fill_(no_update_holes, 0.0)
new_mask = torch.ones_like(output)
new_mask = new_mask.masked_fill_(no_update_holes, 0.0)
return output, new_mask
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
ProtoNN | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/dj/cdjhwfxa4rmj7sf4p4wvsuo75o54dwfcioekxuix47trjvtclo2h.py
# Topologically Sorted Source Nodes: [l2sim, l2sim_1, l2sim_2], Original ATen: [aten.sub, aten.pow, aten.sum]
# Source node to ATen node mapping:
# l2sim => sub
# l2sim_1 => pow_1
# l2sim_2 => sum_1
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %view_2), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %sum_1 : [num_users=2] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1], True), kwargs = {})
triton_poi_fused_pow_sub_sum_0 = async_compile.triton('triton_poi_fused_pow_sub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_pow_sub_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_pow_sub_sum_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (4 + x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (8 + x0), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (12 + x0), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp6 = tmp4 - tmp5
tmp7 = tmp6 * tmp6
tmp8 = tmp3 + tmp7
tmp11 = tmp9 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tmp8 + tmp12
tmp16 = tmp14 - tmp15
tmp17 = tmp16 * tmp16
tmp18 = tmp13 + tmp17
tl.store(out_ptr0 + (x2), tmp18, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/vq/cvqpmuazxnwnsptaxlt2b7ipl2pprxtdxzwiesxonokjn253gipy.py
# Topologically Sorted Source Nodes: [gammal2sim, M, y, y_1], Original ATen: [aten.mul, aten.exp, aten.sum]
# Source node to ATen node mapping:
# M => exp
# gammal2sim => mul
# y => mul_1
# y_1 => sum_2
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, -16), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%mul,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_4, %exp), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_1, [2]), kwargs = {})
triton_poi_fused_exp_mul_sum_1 = async_compile.triton('triton_poi_fused_exp_mul_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_exp_mul_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_exp_mul_sum_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp2 = -16.0
tmp3 = tmp1 * tmp2
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp0 * tmp4
tmp8 = tmp7 * tmp2
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp6 * tmp9
tmp11 = tmp5 + tmp10
tmp14 = tmp13 * tmp2
tmp15 = tl_math.exp(tmp14)
tmp16 = tmp12 * tmp15
tmp17 = tmp11 + tmp16
tmp20 = tmp19 * tmp2
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp18 * tmp21
tmp23 = tmp17 + tmp22
tl.store(out_ptr0 + (x2), tmp23, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [WX], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 1, 4), (4, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [l2sim, l2sim_1, l2sim_2], Original ATen: [aten.sub, aten.pow, aten.sum]
stream0 = get_raw_stream(0)
triton_poi_fused_pow_sub_sum_0.run(primals_2, buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [gammal2sim, M, y, y_1], Original ATen: [aten.mul, aten.exp, aten.sum]
triton_poi_fused_exp_mul_sum_1.run(primals_3, buf1, buf2, 256, grid=grid(256), stream=stream0)
return (buf2, buf1, primals_2, primals_3, buf0, buf1, reinterpret_tensor(primals_4, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import numpy as np
import torch.nn as nn
import torch.onnx
from itertools import product as product
class ProtoNN(nn.Module):
def __init__(self, inputDimension, projectionDimension, numPrototypes,
numOutputLabels, gamma, W=None, B=None, Z=None):
"""
Forward computation graph for ProtoNN.
inputDimension: Input data dimension or feature dimension.
projectionDimension: hyperparameter
numPrototypes: hyperparameter
numOutputLabels: The number of output labels or classes
W, B, Z: Numpy matrices that can be used to initialize
projection matrix(W), prototype matrix (B) and prototype labels
matrix (B).
Expected Dimensions:
W inputDimension (d) x projectionDimension (d_cap)
B projectionDimension (d_cap) x numPrototypes (m)
Z numOutputLabels (L) x numPrototypes (m)
"""
super(ProtoNN, self).__init__()
self.__d = inputDimension
self.__d_cap = projectionDimension
self.__m = numPrototypes
self.__L = numOutputLabels
self.W, self.B, self.Z = None, None, None
self.gamma = gamma
self.__validInit = False
self.__initWBZ(W, B, Z)
self.__validateInit()
def __validateInit(self):
self.__validinit = False
errmsg = 'Dimensions mismatch! Should be W[d, d_cap]'
errmsg += ', B[d_cap, m] and Z[L, m]'
d, d_cap, m, L, _ = self.getHyperParams()
assert self.W.shape[0] == d, errmsg
assert self.W.shape[1] == d_cap, errmsg
assert self.B.shape[0] == d_cap, errmsg
assert self.B.shape[1] == m, errmsg
assert self.Z.shape[0] == L, errmsg
assert self.Z.shape[1] == m, errmsg
self.__validInit = True
def __initWBZ(self, inW, inB, inZ):
if inW is None:
self.W = torch.randn([self.__d, self.__d_cap])
self.W = nn.Parameter(self.W)
else:
self.W = nn.Parameter(torch.from_numpy(inW.astype(np.float32)))
if inB is None:
self.B = torch.randn([self.__d_cap, self.__m])
self.B = nn.Parameter(self.B)
else:
self.B = nn.Parameter(torch.from_numpy(inB.astype(np.float32)))
if inZ is None:
self.Z = torch.randn([self.__L, self.__m])
self.Z = nn.Parameter(self.Z)
else:
self.Z = nn.Parameter(torch.from_numpy(inZ.astype(np.float32)))
def getHyperParams(self):
"""
Returns the model hyperparameters:
[inputDimension, projectionDimension, numPrototypes,
numOutputLabels, gamma]
"""
d = self.__d
dcap = self.__d_cap
m = self.__m
L = self.__L
return d, dcap, m, L, self.gamma
def getModelMatrices(self):
"""
Returns model matrices, which can then be evaluated to obtain
corresponding numpy arrays. These can then be exported as part of
other implementations of ProtonNN, for instance a C++ implementation or
pure python implementation.
Returns
[ProjectionMatrix (W), prototypeMatrix (B),
prototypeLabelsMatrix (Z), gamma]
"""
return self.W, self.B, self.Z, self.gamma
def forward(self, X):
"""
This method is responsible for construction of the forward computation
graph. The end point of the computation graph, or in other words the
output operator for the forward computation is returned.
X: Input of shape [-1, inputDimension]
returns: The forward computation outputs, self.protoNNOut
"""
assert self.__validInit is True, 'Initialization failed!'
W, B, Z, gamma = self.W, self.B, self.Z, self.gamma
WX = torch.matmul(X, W)
dim = [-1, WX.shape[1], 1]
WX = torch.reshape(WX, dim)
dim = [1, B.shape[0], -1]
B_ = torch.reshape(B, dim)
l2sim = B_ - WX
l2sim = torch.pow(l2sim, 2)
l2sim = torch.sum(l2sim, dim=1, keepdim=True)
self.l2sim = l2sim
gammal2sim = -1 * gamma * gamma * l2sim
M = torch.exp(gammal2sim)
dim = [1] + list(Z.shape)
Z_ = torch.reshape(Z, dim)
y = Z_ * M
y = torch.sum(y, dim=2)
return y
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'inputDimension': 4, 'projectionDimension': 4,
'numPrototypes': 4, 'numOutputLabels': 4, 'gamma': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
import numpy as np
import torch.nn as nn
import torch.onnx
from itertools import product as product
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_pow_sub_sum_0(in_ptr0, in_ptr1, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (4 + x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (8 + x0), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp14 = tl.load(in_ptr0 + (12 + x0), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp6 = tmp4 - tmp5
tmp7 = tmp6 * tmp6
tmp8 = tmp3 + tmp7
tmp11 = tmp9 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tmp8 + tmp12
tmp16 = tmp14 - tmp15
tmp17 = tmp16 * tmp16
tmp18 = tmp13 + tmp17
tl.store(out_ptr0 + x2, tmp18, xmask)
@triton.jit
def triton_poi_fused_exp_mul_sum_1(in_ptr0, in_ptr1, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp13 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp18 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp19 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp2 = -16.0
tmp3 = tmp1 * tmp2
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp0 * tmp4
tmp8 = tmp7 * tmp2
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp6 * tmp9
tmp11 = tmp5 + tmp10
tmp14 = tmp13 * tmp2
tmp15 = tl_math.exp(tmp14)
tmp16 = tmp12 * tmp15
tmp17 = tmp11 + tmp16
tmp20 = tmp19 * tmp2
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp18 * tmp21
tmp23 = tmp17 + tmp22
tl.store(out_ptr0 + x2, tmp23, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 1, 4), (4, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_pow_sub_sum_0[grid(256)](primals_2, buf0, buf1,
256, XBLOCK=256, num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
triton_poi_fused_exp_mul_sum_1[grid(256)](primals_3, buf1, buf2,
256, XBLOCK=128, num_warps=4, num_stages=1)
return buf2, buf1, primals_2, primals_3, buf0, buf1, reinterpret_tensor(
primals_4, (4, 64), (1, 4), 0)
class ProtoNNNew(nn.Module):
def __init__(self, inputDimension, projectionDimension, numPrototypes,
numOutputLabels, gamma, W=None, B=None, Z=None):
"""
Forward computation graph for ProtoNN.
inputDimension: Input data dimension or feature dimension.
projectionDimension: hyperparameter
numPrototypes: hyperparameter
numOutputLabels: The number of output labels or classes
W, B, Z: Numpy matrices that can be used to initialize
projection matrix(W), prototype matrix (B) and prototype labels
matrix (B).
Expected Dimensions:
W inputDimension (d) x projectionDimension (d_cap)
B projectionDimension (d_cap) x numPrototypes (m)
Z numOutputLabels (L) x numPrototypes (m)
"""
super(ProtoNNNew, self).__init__()
self.__d = inputDimension
self.__d_cap = projectionDimension
self.__m = numPrototypes
self.__L = numOutputLabels
self.W, self.B, self.Z = None, None, None
self.gamma = gamma
self.__validInit = False
self.__initWBZ(W, B, Z)
self.__validateInit()
def __validateInit(self):
self.__validinit = False
errmsg = 'Dimensions mismatch! Should be W[d, d_cap]'
errmsg += ', B[d_cap, m] and Z[L, m]'
d, d_cap, m, L, _ = self.getHyperParams()
assert self.W.shape[0] == d, errmsg
assert self.W.shape[1] == d_cap, errmsg
assert self.B.shape[0] == d_cap, errmsg
assert self.B.shape[1] == m, errmsg
assert self.Z.shape[0] == L, errmsg
assert self.Z.shape[1] == m, errmsg
self.__validInit = True
def __initWBZ(self, inW, inB, inZ):
if inW is None:
self.W = torch.randn([self.__d, self.__d_cap])
self.W = nn.Parameter(self.W)
else:
self.W = nn.Parameter(torch.from_numpy(inW.astype(np.float32)))
if inB is None:
self.B = torch.randn([self.__d_cap, self.__m])
self.B = nn.Parameter(self.B)
else:
self.B = nn.Parameter(torch.from_numpy(inB.astype(np.float32)))
if inZ is None:
self.Z = torch.randn([self.__L, self.__m])
self.Z = nn.Parameter(self.Z)
else:
self.Z = nn.Parameter(torch.from_numpy(inZ.astype(np.float32)))
def getHyperParams(self):
"""
Returns the model hyperparameters:
[inputDimension, projectionDimension, numPrototypes,
numOutputLabels, gamma]
"""
d = self.__d
dcap = self.__d_cap
m = self.__m
L = self.__L
return d, dcap, m, L, self.gamma
def getModelMatrices(self):
"""
Returns model matrices, which can then be evaluated to obtain
corresponding numpy arrays. These can then be exported as part of
other implementations of ProtonNN, for instance a C++ implementation or
pure python implementation.
Returns
[ProjectionMatrix (W), prototypeMatrix (B),
prototypeLabelsMatrix (Z), gamma]
"""
return self.W, self.B, self.Z, self.gamma
def forward(self, input_0):
primals_1 = self.W
primals_2 = self.B
primals_3 = self.Z
primals_4 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| ShishirPatil/EdgeML-1 | ProtoNN | false | 1,072 | [
"MIT"
] | 0 | cbba9f8b989e545788427c004eb8450e7e4c1a21 | https://github.com/ShishirPatil/EdgeML-1/tree/cbba9f8b989e545788427c004eb8450e7e4c1a21 | import torch
import numpy as np
import torch.nn as nn
import torch.onnx
from itertools import product as product
class Model(nn.Module):
def __init__(self, inputDimension, projectionDimension, numPrototypes,
numOutputLabels, gamma, W=None, B=None, Z=None):
"""
Forward computation graph for ProtoNN.
inputDimension: Input data dimension or feature dimension.
projectionDimension: hyperparameter
numPrototypes: hyperparameter
numOutputLabels: The number of output labels or classes
W, B, Z: Numpy matrices that can be used to initialize
projection matrix(W), prototype matrix (B) and prototype labels
matrix (B).
Expected Dimensions:
W inputDimension (d) x projectionDimension (d_cap)
B projectionDimension (d_cap) x numPrototypes (m)
Z numOutputLabels (L) x numPrototypes (m)
"""
super().__init__()
self.__d = inputDimension
self.__d_cap = projectionDimension
self.__m = numPrototypes
self.__L = numOutputLabels
self.W, self.B, self.Z = None, None, None
self.gamma = gamma
self.__validInit = False
self.__initWBZ(W, B, Z)
self.__validateInit()
def __validateInit(self):
self.__validinit = False
errmsg = 'Dimensions mismatch! Should be W[d, d_cap]'
errmsg += ', B[d_cap, m] and Z[L, m]'
d, d_cap, m, L, _ = self.getHyperParams()
assert self.W.shape[0] == d, errmsg
assert self.W.shape[1] == d_cap, errmsg
assert self.B.shape[0] == d_cap, errmsg
assert self.B.shape[1] == m, errmsg
assert self.Z.shape[0] == L, errmsg
assert self.Z.shape[1] == m, errmsg
self.__validInit = True
def __initWBZ(self, inW, inB, inZ):
if inW is None:
self.W = torch.randn([self.__d, self.__d_cap])
self.W = nn.Parameter(self.W)
else:
self.W = nn.Parameter(torch.from_numpy(inW.astype(np.float32)))
if inB is None:
self.B = torch.randn([self.__d_cap, self.__m])
self.B = nn.Parameter(self.B)
else:
self.B = nn.Parameter(torch.from_numpy(inB.astype(np.float32)))
if inZ is None:
self.Z = torch.randn([self.__L, self.__m])
self.Z = nn.Parameter(self.Z)
else:
self.Z = nn.Parameter(torch.from_numpy(inZ.astype(np.float32)))
def getHyperParams(self):
"""
Returns the model hyperparameters:
[inputDimension, projectionDimension, numPrototypes,
numOutputLabels, gamma]
"""
d = self.__d
dcap = self.__d_cap
m = self.__m
L = self.__L
return d, dcap, m, L, self.gamma
def getModelMatrices(self):
"""
Returns model matrices, which can then be evaluated to obtain
corresponding numpy arrays. These can then be exported as part of
other implementations of ProtonNN, for instance a C++ implementation or
pure python implementation.
Returns
[ProjectionMatrix (W), prototypeMatrix (B),
prototypeLabelsMatrix (Z), gamma]
"""
return self.W, self.B, self.Z, self.gamma
def forward(self, X):
"""
This method is responsible for construction of the forward computation
graph. The end point of the computation graph, or in other words the
output operator for the forward computation is returned.
X: Input of shape [-1, inputDimension]
returns: The forward computation outputs, self.protoNNOut
"""
assert self.__validInit is True, 'Initialization failed!'
W, B, Z, gamma = self.W, self.B, self.Z, self.gamma
WX = torch.matmul(X, W)
dim = [-1, WX.shape[1], 1]
WX = torch.reshape(WX, dim)
dim = [1, B.shape[0], -1]
B_ = torch.reshape(B, d
# ... truncated (>4000 chars) for memory efficiency |
GRULRCell | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/3l/c3lvfnzdm4j54qmpqlq4wtibe7wlwzukm4xfgt3bw5bdnfre2jng.py
# Topologically Sorted Source Nodes: [pre_comp1, add_2, r, mul], Original ATen: [aten.add, aten.sigmoid, aten.mul, aten.sigmoid_backward]
# Source node to ATen node mapping:
# add_2 => add_2
# mul => mul
# pre_comp1 => add
# r => sigmoid
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %view_7), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %primals_8), kwargs = {})
# %sigmoid : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_2,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_6), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid), kwargs = {})
# %mul_11 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %sub_3), kwargs = {})
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp3 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp7 = tmp5 * tmp6
tmp8 = 1.0
tmp9 = tmp8 - tmp5
tmp10 = tmp5 * tmp9
tl.store(out_ptr0 + (x2), tmp7, xmask)
tl.store(out_ptr1 + (x2), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/lz/clzjpvamptc4sf4uhsz3jl6umhb36eb2okmo444rgajkmwalgs3q.py
# Topologically Sorted Source Nodes: [pre_comp2, add_3, z, pre_comp3, add_5, c, mul_1, sub, mul_2, new_h], Original ATen: [aten.add, aten.sigmoid, aten.tanh, aten.mul, aten.rsub]
# Source node to ATen node mapping:
# add_3 => add_3
# add_5 => add_5
# c => tanh
# mul_1 => mul_1
# mul_2 => mul_2
# new_h => add_6
# pre_comp2 => add_1
# pre_comp3 => add_4
# sub => sub
# z => sigmoid_1
# Graph fragment:
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, %view_9), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, %primals_9), kwargs = {})
# %sigmoid_1 : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_3,), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_5, %view_11), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_4, %primals_11), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add_5,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %primals_6), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %sigmoid_1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %tanh), kwargs = {})
# %add_6 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %mul_2), kwargs = {})
triton_poi_fused_add_mul_rsub_sigmoid_tanh_1 = async_compile.triton('triton_poi_fused_add_mul_rsub_sigmoid_tanh_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_rsub_sigmoid_tanh_1', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_tanh_1(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_out_ptr1 + (x2), xmask)
tmp7 = tl.load(in_ptr2 + (x2), xmask)
tmp9 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr4 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = libdevice.tanh(tmp10)
tmp13 = tmp5 * tmp12
tmp14 = 1.0
tmp15 = tmp14 - tmp5
tmp16 = tmp15 * tmp11
tmp17 = tmp13 + tmp16
tl.store(in_out_ptr0 + (x2), tmp5, xmask)
tl.store(in_out_ptr1 + (x2), tmp11, xmask)
tl.store(out_ptr0 + (x2), tmp17, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (1, 4), (4, 1))
assert_size_stride(primals_9, (1, 4), (4, 1))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp2], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_3, out=buf1)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp3], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_4, out=buf2)
del primals_4
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [uComp1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_6, (64, 4), (4, 1), 0), primals_5, out=buf3)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [uComp2], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_6, (64, 4), (4, 1), 0), primals_7, out=buf4)
del primals_7
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf10 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pre_comp1, add_2, r, mul], Original ATen: [aten.add, aten.sigmoid, aten.mul, aten.sigmoid_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0.run(buf0, buf3, primals_8, primals_6, buf6, buf10, 256, grid=grid(256), stream=stream0)
del primals_8
buf7 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [matmul_5], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf6, (64, 4), (4, 1), 0), primals_10, out=buf7)
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf1 # reuse
buf8 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf2 # reuse
buf9 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [pre_comp2, add_3, z, pre_comp3, add_5, c, mul_1, sub, mul_2, new_h], Original ATen: [aten.add, aten.sigmoid, aten.tanh, aten.mul, aten.rsub]
triton_poi_fused_add_mul_rsub_sigmoid_tanh_1.run(buf5, buf8, buf4, primals_9, buf7, primals_11, primals_6, buf9, 256, grid=grid(256), stream=stream0)
del buf4
del buf7
del primals_11
del primals_9
return (buf9, primals_6, buf5, buf8, reinterpret_tensor(buf6, (4, 64), (1, 4), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), buf10, reinterpret_tensor(primals_2, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class GRULRCell(RNNCell):
"""
GRU LR Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
gate_nonlinearity = nonlinearity for the gate can be chosen from
[tanh, sigmoid, relu, quantTanh, quantSigm]
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix
(creates 4 matrices if not None else creates 3 matrices)
uRank = rank of U matrix
(creates 4 matrices if not None else creates 3 matrices)
GRU architecture and compression techniques are found in
GRU(LINK) paper
Basic architecture is like:
r_t = gate_nl(W1x_t + U1h_{t-1} + B_r)
z_t = gate_nl(W2x_t + U2h_{t-1} + B_g)
h_t^ = update_nl(W3x_t + r_t*U3(h_{t-1}) + B_h)
h_t = z_t*h_{t-1} + (1-z_t)*h_t^
Wi and Ui can further parameterised into low rank version by
Wi = matmul(W, W_i) and Ui = matmul(U, U_i)
"""
def __init__(self, input_size, hidden_size, gate_nonlinearity='sigmoid',
update_nonlinearity='tanh', wRank=None, uRank=None, wSparsity=1.0,
uSparsity=1.0, name='GRULR'):
super(GRULRCell, self).__init__(input_size, hidden_size,
gate_nonlinearity, update_nonlinearity, 3, 3, 3, wRank, uRank,
wSparsity, uSparsity)
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
self.W2 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
self.W3 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
else:
self.W = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W1 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
self.W3 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
self.U3 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
else:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U1 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.U3 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_r = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_gate = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self._device = self.bias_update.device
@property
def name(self):
return self._name
@property
def cellType(self):
return 'GRULR'
def forward(self, input, state):
if self._wRank is None:
wComp1 = torch.matmul(input, self.W1)
wComp2 = torch.matmul(input, self.W2)
wComp3 = torch.matmul(input, self.W3)
else:
wComp1 = torch.matmul(torch.matmul(input, self.W), self.W1)
wComp2 = torch.matmul(torch.matmul(input, self.W), self.W2)
wComp3 = torch.matmul(torch.matmul(input, self.W), self.W3)
if self._uRank is None:
uComp1 = torch.matmul(state, self.U1)
uComp2 = torch.matmul(state, self.U2)
else:
uComp1 = torch.matmul(torch.matmul(state, self.U), self.U1)
uComp2 = torch.matmul(torch.matmul(state, self.U), self.U2)
pre_comp1 = wComp1 + uComp1
pre_comp2 = wComp2 + uComp2
r = gen_nonlinearity(pre_comp1 + self.bias_r, self._gate_nonlinearity)
z = gen_nonlinearity(pre_comp2 + self.bias_gate, self.
_gate_nonlinearity)
if self._uRank is None:
pre_comp3 = wComp3 + torch.matmul(r * state, self.U3)
else:
pre_comp3 = wComp3 + torch.matmul(torch.matmul(r * state, self.
U), self.U3)
c = gen_nonlinearity(pre_comp3 + self.bias_update, self.
_update_nonlinearity)
new_h = z * state + (1.0 - z) * c
return new_h
def getVars(self):
Vars = []
if self._num_W_matrices == 3:
Vars.extend([self.W1, self.W2, self.W3])
else:
Vars.extend([self.W, self.W1, self.W2, self.W3])
if self._num_U_matrices == 3:
Vars.extend([self.U1, self.U2, self.U3])
else:
Vars.extend([self.U, self.U1, self.U2, self.U3])
Vars.extend([self.bias_r, self.bias_gate, self.bias_update])
return Vars
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.onnx
from itertools import product as product
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0(in_ptr0, in_ptr1,
in_ptr2, in_ptr3, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp3 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp7 = tmp5 * tmp6
tmp8 = 1.0
tmp9 = tmp8 - tmp5
tmp10 = tmp5 * tmp9
tl.store(out_ptr0 + x2, tmp7, xmask)
tl.store(out_ptr1 + x2, tmp10, xmask)
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_tanh_1(in_out_ptr0, in_out_ptr1,
in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_out_ptr1 + x2, xmask)
tmp7 = tl.load(in_ptr2 + x2, xmask)
tmp9 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr4 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = libdevice.tanh(tmp10)
tmp13 = tmp5 * tmp12
tmp14 = 1.0
tmp15 = tmp14 - tmp5
tmp16 = tmp15 * tmp11
tmp17 = tmp13 + tmp16
tl.store(in_out_ptr0 + x2, tmp5, xmask)
tl.store(in_out_ptr1 + x2, tmp11, xmask)
tl.store(out_ptr0 + x2, tmp17, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (1, 4), (4, 1))
assert_size_stride(primals_9, (1, 4), (4, 1))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_3, out=buf1)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_4, out=buf2)
del primals_4
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_6, (64, 4), (4, 1), 0),
primals_5, out=buf3)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_6, (64, 4), (4, 1), 0),
primals_7, out=buf4)
del primals_7
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf10 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_sigmoid_sigmoid_backward_0[grid(256)](buf0,
buf3, primals_8, primals_6, buf6, buf10, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_8
buf7 = buf3
del buf3
extern_kernels.mm(reinterpret_tensor(buf6, (64, 4), (4, 1), 0),
primals_10, out=buf7)
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf1
buf8 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf2
buf9 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
triton_poi_fused_add_mul_rsub_sigmoid_tanh_1[grid(256)](buf5, buf8,
buf4, primals_9, buf7, primals_11, primals_6, buf9, 256, XBLOCK
=256, num_warps=4, num_stages=1)
del buf4
del buf7
del primals_11
del primals_9
return buf9, primals_6, buf5, buf8, reinterpret_tensor(buf6, (4, 64), (
1, 4), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0
), buf10, reinterpret_tensor(primals_2, (4, 64), (1, 4), 0)
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class GRULRCellNew(RNNCell):
"""
GRU LR Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
gate_nonlinearity = nonlinearity for the gate can be chosen from
[tanh, sigmoid, relu, quantTanh, quantSigm]
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix
(creates 4 matrices if not None else creates 3 matrices)
uRank = rank of U matrix
(creates 4 matrices if not None else creates 3 matrices)
GRU architecture and compression techniques are found in
GRU(LINK) paper
Basic architecture is like:
r_t = gate_nl(W1x_t + U1h_{t-1} + B_r)
z_t = gate_nl(W2x_t + U2h_{t-1} + B_g)
h_t^ = update_nl(W3x_t + r_t*U3(h_{t-1}) + B_h)
h_t = z_t*h_{t-1} + (1-z_t)*h_t^
Wi and Ui can further parameterised into low rank version by
Wi = matmul(W, W_i) and Ui = matmul(U, U_i)
"""
def __init__(self, input_size, hidden_size, gate_nonlinearity='sigmoid',
update_nonlinearity='tanh', wRank=None, uRank=None, wSparsity=1.0,
uSparsity=1.0, name='GRULR'):
super(GRULRCellNew, self).__init__(input_size, hidden_size,
gate_nonlinearity, update_nonlinearity, 3, 3, 3, wRank, uRank,
wSparsity, uSparsity)
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
self.W2 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
self.W3 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
else:
self.W = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W1 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
self.W3 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
self.U3 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
else:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U1 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.U3 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_r = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_gate = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self._device = self.bias_update.device
@property
def name(self):
return self._name
@property
def cellType(self):
return 'GRULR'
def getVars(self):
Vars = []
if self._num_W_matrices == 3:
Vars.extend([self.W1, self.W2, self.W3])
else:
Vars.extend([self.W, self.W1, self.W2, self.W3])
if self._num_U_matrices == 3:
Vars.extend([self.U1, self.U2, self.U3])
else:
Vars.extend([self.U, self.U1, self.U2, self.U3])
Vars.extend([self.bias_r, self.bias_gate, self.bias_update])
return Vars
def forward(self, input_0, input_1):
primals_1 = self.W1
primals_3 = self.W2
primals_4 = self.W3
primals_5 = self.U1
primals_7 = self.U2
primals_10 = self.U3
primals_8 = self.bias_r
primals_9 = self.bias_gate
primals_11 = self.bias_update
primals_2 = input_0
primals_6 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11])
return output[0]
| ShishirPatil/EdgeML-1 | GRULRCell | false | 1,073 | [
"MIT"
] | 0 | cbba9f8b989e545788427c004eb8450e7e4c1a21 | https://github.com/ShishirPatil/EdgeML-1/tree/cbba9f8b989e545788427c004eb8450e7e4c1a21 | import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super().__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(to
# ... truncated (>4000 chars) for memory efficiency |
Connect2Model | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/v7/cv7zazascu4rpkkwoxbiwk6c2le2e6wshdhae73bmaoapelvwguv.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/xk/cxkugsynlmnyrjhah42fewrhwovuvurnuv2qimo2qhxq27wjmq7q.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_5, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_5, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x3), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/jf/cjfzp64ny4hf7wdw5wptah3hqv5fcsh5rrw4brz7uxcy6ad57n7h.py
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# softmax => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [1], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/mq/cmqqbqcbxlzvej3i4srpbdfgxirnooh2wgnwtdxbgyt3e2d762nv.py
# Topologically Sorted Source Nodes: [tanh], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# tanh => tanh
# Graph fragment:
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%view_7,), kwargs = {})
triton_poi_fused_tanh_3 = async_compile.triton('triton_poi_fused_tanh_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_tanh_3', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_tanh_3(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = libdevice.tanh(tmp3)
tl.store(in_out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (16, 4), (4, 1))
assert_size_stride(primals_2, (16, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (16, 16), (16, 1))
assert_size_stride(primals_5, (16, ), (1, ))
assert_size_stride(primals_6, (4, 16), (16, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (1, 16), (16, 1))
assert_size_stride(primals_9, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 16), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 16), (256, 64, 16, 1), 0); del buf0 # reuse
buf10 = empty_strided_cuda((4, 4, 4, 16), (256, 64, 16, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf10, 1024, grid=grid(1024), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 16), (16, 1), 0), reinterpret_tensor(primals_4, (16, 16), (1, 16), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 16), (256, 64, 16, 1), 0); del buf2 # reuse
buf9 = empty_strided_cuda((4, 4, 4, 16), (256, 64, 16, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf9, 1024, grid=grid(1024), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [action_logits], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 16), (16, 1), 0), reinterpret_tensor(primals_6, (16, 4), (1, 16), 0), alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf3, (64, 16), (16, 1), 0), reinterpret_tensor(primals_8, (16, 1), (1, 16), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf4, buf6, 256, grid=grid(256), stream=stream0)
buf7 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
buf8 = reinterpret_tensor(buf5, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [tanh], Original ATen: [aten.tanh]
triton_poi_fused_tanh_3.run(buf8, primals_9, 64, grid=grid(64), stream=stream0)
del primals_9
return (buf7, buf8, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 16), (16, 1), 0), reinterpret_tensor(buf3, (64, 16), (16, 1), 0), buf7, buf8, primals_8, primals_6, buf9, primals_4, buf10, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((16, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((16, 16), (16, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 16), (16, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 16), (16, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
class Connect2Model(nn.Module):
def __init__(self, board_size, action_size, device):
super(Connect2Model, self).__init__()
self.device = device
self.size = board_size
self.action_size = action_size
self.fc1 = nn.Linear(in_features=self.size, out_features=16)
self.fc2 = nn.Linear(in_features=16, out_features=16)
self.action_head = nn.Linear(in_features=16, out_features=self.
action_size)
self.value_head = nn.Linear(in_features=16, out_features=1)
self
def forward(self, x):
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
action_logits = self.action_head(x)
value_logit = self.value_head(x)
return F.softmax(action_logits, dim=1), torch.tanh(value_logit)
def predict(self, board):
board = torch.FloatTensor(board.astype(np.float32))
board = board.view(1, self.size)
self.eval()
with torch.no_grad():
pi, v = self.forward(board)
return pi.data.cpu().numpy()[0], v.data.cpu().numpy()[0]
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'board_size': 4, 'action_size': 4, 'device': 0}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import numpy as np
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x3, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_poi_fused_tanh_3(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = libdevice.tanh(tmp3)
tl.store(in_out_ptr0 + x0, tmp4, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (16, 4), (4, 1))
assert_size_stride(primals_2, (16,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (16, 16), (16, 1))
assert_size_stride(primals_5, (16,), (1,))
assert_size_stride(primals_6, (4, 16), (16, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (1, 16), (16, 1))
assert_size_stride(primals_9, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 16), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 16), (256, 64, 16, 1), 0)
del buf0
buf10 = empty_strided_cuda((4, 4, 4, 16), (256, 64, 16, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(1024)](buf1,
primals_2, buf10, 1024, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 16), (16, 1), 0),
reinterpret_tensor(primals_4, (16, 16), (1, 16), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 16), (256, 64, 16, 1), 0)
del buf2
buf9 = empty_strided_cuda((4, 4, 4, 16), (256, 64, 16, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(1024)](buf3,
primals_5, buf9, 1024, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 16),
(16, 1), 0), reinterpret_tensor(primals_6, (16, 4), (1, 16), 0),
alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf3, (64, 16), (16, 1), 0),
reinterpret_tensor(primals_8, (16, 1), (1, 16), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf4, buf6, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf7 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf4
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
buf8 = reinterpret_tensor(buf5, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf5
triton_poi_fused_tanh_3[grid(64)](buf8, primals_9, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_9
return buf7, buf8, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 16), (16, 1), 0), reinterpret_tensor(
buf3, (64, 16), (16, 1), 0
), buf7, buf8, primals_8, primals_6, buf9, primals_4, buf10
class Connect2ModelNew(nn.Module):
def __init__(self, board_size, action_size, device):
super(Connect2ModelNew, self).__init__()
self.device = device
self.size = board_size
self.action_size = action_size
self.fc1 = nn.Linear(in_features=self.size, out_features=16)
self.fc2 = nn.Linear(in_features=16, out_features=16)
self.action_head = nn.Linear(in_features=16, out_features=self.
action_size)
self.value_head = nn.Linear(in_features=16, out_features=1)
self
def predict(self, board):
board = torch.FloatTensor(board.astype(np.float32))
board = board.view(1, self.size)
self.eval()
with torch.no_grad():
pi, v = self.forward(board)
return pi.data.cpu().numpy()[0], v.data.cpu().numpy()[0]
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.action_head.weight
primals_7 = self.action_head.bias
primals_8 = self.value_head.weight
primals_9 = self.value_head.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0], output[1]
| ShokuninSan/AlphaZeroSimple | Connect2Model | false | 1,074 | [
"MIT"
] | 0 | e32e6a28f872a046705a3f68882139688d5a43c3 | https://github.com/ShokuninSan/AlphaZeroSimple/tree/e32e6a28f872a046705a3f68882139688d5a43c3 | import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, board_size, action_size, device):
super().__init__()
self.device = device
self.size = board_size
self.action_size = action_size
self.fc1 = nn.Linear(in_features=self.size, out_features=16)
self.fc2 = nn.Linear(in_features=16, out_features=16)
self.action_head = nn.Linear(in_features=16, out_features=self.
action_size)
self.value_head = nn.Linear(in_features=16, out_features=1)
self
def forward(self, x):
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
action_logits = self.action_head(x)
value_logit = self.value_head(x)
return F.softmax(action_logits, dim=1), torch.tanh(value_logit)
def predict(self, board):
board = torch.FloatTensor(board.astype(np.float32))
board = board.view(1, self.size)
self.eval()
with torch.no_grad():
pi, v = self.forward(board)
return pi.data.cpu().numpy()[0], v.data.cpu().numpy()[0]
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 0]
|
CausalConv2d | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/2a/c2aucoxj5ek5nl3pl5n67fq4vdl55x2rw66tvv2aq23fsb3jvf5q.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.constant_pad_nd]
# Source node to ATen node mapping:
# out => constant_pad_nd
# Graph fragment:
# %constant_pad_nd : [num_users=2] = call_function[target=torch.ops.aten.constant_pad_nd.default](args = (%primals_1, [3, 0, 3, 0], 0.0), kwargs = {})
triton_poi_fused_constant_pad_nd_0 = async_compile.triton('triton_poi_fused_constant_pad_nd_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_constant_pad_nd_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_constant_pad_nd_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 784
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 7) % 7
x0 = xindex % 7
x2 = (xindex // 49)
x4 = xindex
tmp0 = (-3) + x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = (-3) + x0
tmp4 = tmp3 >= tmp1
tmp5 = tmp2 & tmp4
tmp6 = tl.load(in_ptr0 + ((-15) + x0 + (4*x1) + (16*x2)), tmp5 & xmask, other=0.0)
tl.store(out_ptr0 + (x4), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/6r/c6rjd3rxthhw2ub6d2gtgjazrjeoyhalmj36ujwzgwexsis27e73.py
# Topologically Sorted Source Nodes: [_weight_norm], Original ATen: [aten._weight_norm_interface]
# Source node to ATen node mapping:
# _weight_norm => div, mul, pow_1, pow_2, sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_3, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1, 2, 3], True), kwargs = {})
# %pow_2 : [num_users=2] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_2, %pow_2), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_3, %div), kwargs = {})
triton_per_fused__weight_norm_interface_1 = async_compile.triton('triton_per_fused__weight_norm_interface_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__weight_norm_interface_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__weight_norm_interface_1(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp7 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp8 = tmp7 / tmp6
tmp9 = tmp0 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr0 + (r1 + (64*x0)), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/w5/cw5gytijzzkwnfpq2a2axdsj4pfxgxmwiuzizuyd4bw5uwnanzw7.py
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# out_1 => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%constant_pad_nd, %mul, %primals_4, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 1, 1, 1), (1, 1, 1, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 7, 7), (196, 49, 7, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.constant_pad_nd]
stream0 = get_raw_stream(0)
triton_poi_fused_constant_pad_nd_0.run(primals_1, buf0, 784, grid=grid(784), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf2 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf1 # reuse
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [_weight_norm], Original ATen: [aten._weight_norm_interface]
triton_per_fused__weight_norm_interface_1.run(buf2, primals_3, primals_2, buf3, 4, 64, grid=grid(4), stream=stream0)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf0, buf3, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 4, 4, 4), (64, 16, 4, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf5, primals_4, 256, grid=grid(256), stream=stream0)
del primals_4
return (buf5, buf3, primals_2, primals_3, buf0, buf2, buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 1, 1, 1), (1, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
import torch.utils.data
class WNConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding=0, bias=True, activation=None):
super().__init__()
self.conv = nn.utils.weight_norm(nn.Conv2d(in_channel, out_channel,
kernel_size, stride=stride, padding=padding, bias=bias))
self.out_channel = out_channel
if isinstance(kernel_size, int):
kernel_size = [kernel_size, kernel_size]
self.kernel_size = kernel_size
self.activation = activation
def forward(self, input):
out = self.conv(input)
if self.activation is not None:
out = self.activation(out)
return out
class CausalConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding='downright', activation=None):
super().__init__()
if isinstance(kernel_size, int):
kernel_size = [kernel_size] * 2
self.kernel_size = kernel_size
if padding == 'downright':
pad = [kernel_size[1] - 1, 0, kernel_size[0] - 1, 0]
elif padding == 'down' or padding == 'causal':
pad = kernel_size[1] // 2
pad = [pad, pad, kernel_size[0] - 1, 0]
self.causal = 0
if padding == 'causal':
self.causal = kernel_size[1] // 2
self.pad = nn.ZeroPad2d(pad)
self.conv = WNConv2d(in_channel, out_channel, kernel_size, stride=
stride, padding=0, activation=activation)
def forward(self, input):
out = self.pad(input)
if self.causal > 0:
self.conv.conv.weight_v.data[:, :, -1, self.causal:].zero_()
out = self.conv(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channel': 4, 'out_channel': 4, 'kernel_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_constant_pad_nd_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 784
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 7 % 7
x0 = xindex % 7
x2 = xindex // 49
x4 = xindex
tmp0 = -3 + x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = -3 + x0
tmp4 = tmp3 >= tmp1
tmp5 = tmp2 & tmp4
tmp6 = tl.load(in_ptr0 + (-15 + x0 + 4 * x1 + 16 * x2), tmp5 & xmask,
other=0.0)
tl.store(out_ptr0 + x4, tmp6, xmask)
@triton.jit
def triton_per_fused__weight_norm_interface_1(in_out_ptr0, in_ptr0, in_ptr1,
out_ptr0, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp7 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp8 = tmp7 / tmp6
tmp9 = tmp0 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr0 + (r1 + 64 * x0), tmp9, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 1, 1, 1), (1, 1, 1, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 7, 7), (196, 49, 7, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_constant_pad_nd_0[grid(784)](primals_1, buf0, 784,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf2 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf1
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_per_fused__weight_norm_interface_1[grid(4)](buf2, primals_3,
primals_2, buf3, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
buf4 = extern_kernels.convolution(buf0, buf3, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 4, 4, 4), (64, 16, 4, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_2[grid(256)](buf5, primals_4, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_4
return buf5, buf3, primals_2, primals_3, buf0, buf2, buf3
class WNConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding=0, bias=True, activation=None):
super().__init__()
self.conv = nn.utils.weight_norm(nn.Conv2d(in_channel, out_channel,
kernel_size, stride=stride, padding=padding, bias=bias))
self.out_channel = out_channel
if isinstance(kernel_size, int):
kernel_size = [kernel_size, kernel_size]
self.kernel_size = kernel_size
self.activation = activation
def forward(self, input):
out = self.conv(input)
if self.activation is not None:
out = self.activation(out)
return out
class CausalConv2dNew(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding='downright', activation=None):
super().__init__()
if isinstance(kernel_size, int):
kernel_size = [kernel_size] * 2
self.kernel_size = kernel_size
if padding == 'downright':
pad = [kernel_size[1] - 1, 0, kernel_size[0] - 1, 0]
elif padding == 'down' or padding == 'causal':
pad = kernel_size[1] // 2
pad = [pad, pad, kernel_size[0] - 1, 0]
self.causal = 0
if padding == 'causal':
self.causal = kernel_size[1] // 2
self.pad = nn.ZeroPad2d(pad)
self.conv = WNConv2d(in_channel, out_channel, kernel_size, stride=
stride, padding=0, activation=activation)
def forward(self, input_0):
primals_4 = self.conv.conv.bias
primals_2 = self.conv.conv.weight_g
primals_1 = self.conv.conv.weight_v
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| Shivanshu-Gupta/KaoKore-VQ-VAE2 | CausalConv2d | false | 1,075 | [
"MIT"
] | 0 | 38a88ba312dee3c0e2c1aaf02e1c1754ba19ac0c | https://github.com/Shivanshu-Gupta/KaoKore-VQ-VAE2/tree/38a88ba312dee3c0e2c1aaf02e1c1754ba19ac0c | import torch
from torch import nn
import torch.utils.data
class WNConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding=0, bias=True, activation=None):
super().__init__()
self.conv = nn.utils.weight_norm(nn.Conv2d(in_channel, out_channel,
kernel_size, stride=stride, padding=padding, bias=bias))
self.out_channel = out_channel
if isinstance(kernel_size, int):
kernel_size = [kernel_size, kernel_size]
self.kernel_size = kernel_size
self.activation = activation
def forward(self, input):
out = self.conv(input)
if self.activation is not None:
out = self.activation(out)
return out
class Model(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding='downright', activation=None):
super().__init__()
if isinstance(kernel_size, int):
kernel_size = [kernel_size] * 2
self.kernel_size = kernel_size
if padding == 'downright':
pad = [kernel_size[1] - 1, 0, kernel_size[0] - 1, 0]
elif padding == 'down' or padding == 'causal':
pad = kernel_size[1] // 2
pad = [pad, pad, kernel_size[0] - 1, 0]
self.causal = 0
if padding == 'causal':
self.causal = kernel_size[1] // 2
self.pad = nn.ZeroPad2d(pad)
self.conv = WNConv2d(in_channel, out_channel, kernel_size, stride=
stride, padding=0, activation=activation)
def forward(self, input):
out = self.pad(input)
if self.causal > 0:
self.conv.conv.weight_v.data[:, :, -1, self.causal:].zero_()
out = self.conv(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
FastGRNNCell | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/w4/cw4ae3o32nppo7llvm7ul7p5le255pvoa5wqcrceacbu2xvt2ydy.py
# Topologically Sorted Source Nodes: [pre_comp, add_1, z, add_2, c, mul, sigmoid_1, sub, mul_1, sigmoid_2, add_3, mul_2, new_h], Original ATen: [aten.add, aten.sigmoid, aten.tanh, aten.mul, aten.rsub]
# Source node to ATen node mapping:
# add_1 => add_1
# add_2 => add_2
# add_3 => add_3
# c => tanh
# mul => mul
# mul_1 => mul_1
# mul_2 => mul_2
# new_h => add_4
# pre_comp => add
# sigmoid_1 => sigmoid_1
# sigmoid_2 => sigmoid_2
# sub => sub
# z => sigmoid
# Graph fragment:
# %add : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %view_3), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %primals_5), kwargs = {})
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_1,), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %primals_6), kwargs = {})
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_2,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_4), kwargs = {})
# %sigmoid_1 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_7,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %sigmoid), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, %sub), kwargs = {})
# %sigmoid_2 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%primals_8,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %sigmoid_2), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_3, %tanh), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_2), kwargs = {})
triton_poi_fused_add_mul_rsub_sigmoid_tanh_0 = async_compile.triton('triton_poi_fused_add_mul_rsub_sigmoid_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_rsub_sigmoid_tanh_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_tanh_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
x1 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask)
tmp3 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr2 + (x0), xmask)
tmp8 = tl.load(in_ptr3 + (0))
tmp9 = tl.broadcast_to(tmp8, [XBLOCK])
tmp14 = tl.load(in_ptr4 + (0))
tmp15 = tl.broadcast_to(tmp14, [XBLOCK])
tmp18 = tl.load(in_ptr5 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp7 = tmp5 * tmp6
tmp10 = tl.sigmoid(tmp9)
tmp11 = 1.0
tmp12 = tmp11 - tmp5
tmp13 = tmp10 * tmp12
tmp16 = tl.sigmoid(tmp15)
tmp17 = tmp13 + tmp16
tmp19 = tmp2 + tmp18
tmp20 = libdevice.tanh(tmp19)
tmp21 = tmp17 * tmp20
tmp22 = tmp7 + tmp21
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
tl.store(out_ptr0 + (x0), tmp22, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (1, 4), (4, 1))
assert_size_stride(primals_6, (1, 4), (4, 1))
assert_size_stride(primals_7, (1, 1), (1, 1))
assert_size_stride(primals_8, (1, 1), (1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [uComp], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0), primals_3, out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pre_comp, add_1, z, add_2, c, mul, sigmoid_1, sub, mul_1, sigmoid_2, add_3, mul_2, new_h], Original ATen: [aten.add, aten.sigmoid, aten.tanh, aten.mul, aten.rsub]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_rsub_sigmoid_tanh_0.run(buf2, buf1, primals_5, primals_4, primals_7, primals_8, primals_6, buf3, 256, grid=grid(256), stream=stream0)
del buf1
return (buf3, primals_4, primals_5, primals_6, primals_7, primals_8, buf2, reinterpret_tensor(primals_2, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((1, 1), (1, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 1), (1, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class FastGRNNCell(RNNCell):
"""
FastGRNN Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
gate_nonlinearity = nonlinearity for the gate can be chosen from
[tanh, sigmoid, relu, quantTanh, quantSigm]
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix (creates two matrices if not None)
uRank = rank of U matrix (creates two matrices if not None)
wSparsity = intended sparsity of W matrix(ces)
uSparsity = intended sparsity of U matrix(ces)
Warning:
The Cell will not automatically sparsify.
The user must invoke .sparsify to hard threshold.
zetaInit = init for zeta, the scale param
nuInit = init for nu, the translation param
FastGRNN architecture and compression techniques are found in
FastGRNN(LINK) paper
Basic architecture is like:
z_t = gate_nl(Wx_t + Uh_{t-1} + B_g)
h_t^ = update_nl(Wx_t + Uh_{t-1} + B_h)
h_t = z_t*h_{t-1} + (sigmoid(zeta)(1-z_t) + sigmoid(nu))*h_t^
W and U can further parameterised into low rank version by
W = matmul(W_1, W_2) and U = matmul(U_1, U_2)
"""
def __init__(self, input_size, hidden_size, gate_nonlinearity='sigmoid',
update_nonlinearity='tanh', wRank=None, uRank=None, wSparsity=1.0,
uSparsity=1.0, zetaInit=1.0, nuInit=-4.0, name='FastGRNN'):
super(FastGRNNCell, self).__init__(input_size, hidden_size,
gate_nonlinearity, update_nonlinearity, 1, 1, 2, wRank, uRank,
wSparsity, uSparsity)
self._zetaInit = zetaInit
self._nuInit = nuInit
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W = nn.Parameter(0.1 * torch.randn([input_size, hidden_size]))
else:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, hidden_size])
)
else:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_gate = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self.zeta = nn.Parameter(self._zetaInit * torch.ones([1, 1]))
self.nu = nn.Parameter(self._nuInit * torch.ones([1, 1]))
@property
def name(self):
return self._name
@property
def cellType(self):
return 'FastGRNN'
def forward(self, input, state):
if self._wRank is None:
wComp = torch.matmul(input, self.W)
else:
wComp = torch.matmul(torch.matmul(input, self.W1), self.W2)
if self._uRank is None:
uComp = torch.matmul(state, self.U)
else:
uComp = torch.matmul(torch.matmul(state, self.U1), self.U2)
pre_comp = wComp + uComp
z = gen_nonlinearity(pre_comp + self.bias_gate, self._gate_nonlinearity
)
c = gen_nonlinearity(pre_comp + self.bias_update, self.
_update_nonlinearity)
new_h = z * state + (torch.sigmoid(self.zeta) * (1.0 - z) + torch.
sigmoid(self.nu)) * c
return new_h
def getVars(self):
Vars = []
if self._num_W_matrices == 1:
Vars.append(self.W)
else:
Vars.extend([self.W1, self.W2])
if self._num_U_matrices == 1:
Vars.append(self.U)
else:
Vars.extend([self.U1, self.U2])
Vars.extend([self.bias_gate, self.bias_update])
Vars.extend([self.zeta, self.nu])
return Vars
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.onnx
from itertools import product as product
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_tanh_0(in_out_ptr0, in_ptr0,
in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
x1 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask)
tmp3 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr2 + x0, xmask)
tmp8 = tl.load(in_ptr3 + 0)
tmp9 = tl.broadcast_to(tmp8, [XBLOCK])
tmp14 = tl.load(in_ptr4 + 0)
tmp15 = tl.broadcast_to(tmp14, [XBLOCK])
tmp18 = tl.load(in_ptr5 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp7 = tmp5 * tmp6
tmp10 = tl.sigmoid(tmp9)
tmp11 = 1.0
tmp12 = tmp11 - tmp5
tmp13 = tmp10 * tmp12
tmp16 = tl.sigmoid(tmp15)
tmp17 = tmp13 + tmp16
tmp19 = tmp2 + tmp18
tmp20 = libdevice.tanh(tmp19)
tmp21 = tmp17 * tmp20
tmp22 = tmp7 + tmp21
tl.store(in_out_ptr0 + x0, tmp2, xmask)
tl.store(out_ptr0 + x0, tmp22, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_5, (1, 4), (4, 1))
assert_size_stride(primals_6, (1, 4), (4, 1))
assert_size_stride(primals_7, (1, 1), (1, 1))
assert_size_stride(primals_8, (1, 1), (1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_4, (64, 4), (4, 1), 0),
primals_3, out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_rsub_sigmoid_tanh_0[grid(256)](buf2, buf1,
primals_5, primals_4, primals_7, primals_8, primals_6, buf3,
256, XBLOCK=128, num_warps=4, num_stages=1)
del buf1
return (buf3, primals_4, primals_5, primals_6, primals_7, primals_8,
buf2, reinterpret_tensor(primals_2, (4, 64), (1, 4), 0))
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class FastGRNNCellNew(RNNCell):
"""
FastGRNN Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
gate_nonlinearity = nonlinearity for the gate can be chosen from
[tanh, sigmoid, relu, quantTanh, quantSigm]
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix (creates two matrices if not None)
uRank = rank of U matrix (creates two matrices if not None)
wSparsity = intended sparsity of W matrix(ces)
uSparsity = intended sparsity of U matrix(ces)
Warning:
The Cell will not automatically sparsify.
The user must invoke .sparsify to hard threshold.
zetaInit = init for zeta, the scale param
nuInit = init for nu, the translation param
FastGRNN architecture and compression techniques are found in
FastGRNN(LINK) paper
Basic architecture is like:
z_t = gate_nl(Wx_t + Uh_{t-1} + B_g)
h_t^ = update_nl(Wx_t + Uh_{t-1} + B_h)
h_t = z_t*h_{t-1} + (sigmoid(zeta)(1-z_t) + sigmoid(nu))*h_t^
W and U can further parameterised into low rank version by
W = matmul(W_1, W_2) and U = matmul(U_1, U_2)
"""
def __init__(self, input_size, hidden_size, gate_nonlinearity='sigmoid',
update_nonlinearity='tanh', wRank=None, uRank=None, wSparsity=1.0,
uSparsity=1.0, zetaInit=1.0, nuInit=-4.0, name='FastGRNN'):
super(FastGRNNCellNew, self).__init__(input_size, hidden_size,
gate_nonlinearity, update_nonlinearity, 1, 1, 2, wRank, uRank,
wSparsity, uSparsity)
self._zetaInit = zetaInit
self._nuInit = nuInit
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W = nn.Parameter(0.1 * torch.randn([input_size, hidden_size]))
else:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, hidden_size])
)
else:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_gate = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self.zeta = nn.Parameter(self._zetaInit * torch.ones([1, 1]))
self.nu = nn.Parameter(self._nuInit * torch.ones([1, 1]))
@property
def name(self):
return self._name
@property
def cellType(self):
return 'FastGRNN'
def getVars(self):
Vars = []
if self._num_W_matrices == 1:
Vars.append(self.W)
else:
Vars.extend([self.W1, self.W2])
if self._num_U_matrices == 1:
Vars.append(self.U)
else:
Vars.extend([self.U1, self.U2])
Vars.extend([self.bias_gate, self.bias_update])
Vars.extend([self.zeta, self.nu])
return Vars
def forward(self, input_0, input_1):
primals_1 = self.W
primals_3 = self.U
primals_5 = self.bias_gate
primals_6 = self.bias_update
primals_7 = self.zeta
primals_8 = self.nu
primals_2 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8])
return output[0]
| ShishirPatil/EdgeML-1 | FastGRNNCell | false | 1,076 | [
"MIT"
] | 0 | cbba9f8b989e545788427c004eb8450e7e4c1a21 | https://github.com/ShishirPatil/EdgeML-1/tree/cbba9f8b989e545788427c004eb8450e7e4c1a21 | import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super().__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(to
# ... truncated (>4000 chars) for memory efficiency |
ResidualConvUnit | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/6q/c6q46q7lsepa4jw5qgcgbc5kiud5wm57hubk6vfo4gk47vl2tprk.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# out => relu
# Graph fragment:
# %relu : [num_users=4] = call_function[target=torch.ops.aten.relu.default](args = (%primals_1,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/4e/c4efs56ymyev6yow4ruutakn3po5nni7rvtifmzxqreckdzecoje.py
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# out_1 => convolution
# out_2 => relu_1
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_2, %primals_3, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_1 = async_compile.triton('triton_poi_fused_convolution_relu_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/hc/chcx47xohlbah6m2k2mpiyct2oijgy7vuqj7z6iime4huasayvby.py
# Topologically Sorted Source Nodes: [add], Original ATen: [aten.add]
# Source node to ATen node mapping:
# add => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convolution_1, %relu), kwargs = {})
# %copy_ : [num_users=0] = call_function[target=torch.ops.aten.copy_.default](args = (%primals_1, %relu), kwargs = {})
triton_poi_fused_add_2 = async_compile.triton('triton_poi_fused_add_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_2', 'mutated_arg_names': ['in_out_ptr0', 'out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_2(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask)
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
tl.store(out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_1.run(buf2, primals_3, 256, grid=grid(256), stream=stream0)
del primals_3
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf2, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 4, 4), (64, 16, 4, 1))
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [add], Original ATen: [aten.add]
triton_poi_fused_add_2.run(buf4, buf0, primals_1, 256, grid=grid(256), stream=stream0)
del primals_1
return (buf4, primals_2, primals_4, buf0, buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class ResidualConvUnit(nn.Module):
"""Residual convolution module.
"""
def __init__(self, features):
"""Init.
Args:
features (int): number of features
"""
super().__init__()
self.conv1 = nn.Conv2d(features, features, kernel_size=3, stride=1,
padding=1, bias=True)
self.conv2 = nn.Conv2d(features, features, kernel_size=3, stride=1,
padding=1, bias=False)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
"""Forward pass.
Args:
x (tensor): input
Returns:
tensor: output
"""
out = self.relu(x)
out = self.conv1(out)
out = self.relu(out)
out = self.conv2(out)
return out + x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'features': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_1(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_2(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask)
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x0, tmp2, xmask)
tl.store(out_ptr0 + x0, tmp1, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_relu_0[grid(256)](primals_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_relu_1[grid(256)](buf2, primals_3, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
buf3 = extern_kernels.convolution(buf2, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 4, 4), (64, 16, 4, 1))
buf4 = buf3
del buf3
triton_poi_fused_add_2[grid(256)](buf4, buf0, primals_1, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_1
return buf4, primals_2, primals_4, buf0, buf2
class ResidualConvUnitNew(nn.Module):
"""Residual convolution module.
"""
def __init__(self, features):
"""Init.
Args:
features (int): number of features
"""
super().__init__()
self.conv1 = nn.Conv2d(features, features, kernel_size=3, stride=1,
padding=1, bias=True)
self.conv2 = nn.Conv2d(features, features, kernel_size=3, stride=1,
padding=1, bias=False)
self.relu = nn.ReLU(inplace=True)
def forward(self, input_0):
primals_2 = self.conv1.weight
primals_3 = self.conv1.bias
primals_4 = self.conv2.weight
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| ShiraLightricks/3d-photo-inpainting | ResidualConvUnit | false | 1,077 | [
"MIT"
] | 0 | c42ac41576690b765e50f5281ddbfb58439ff36d | https://github.com/ShiraLightricks/3d-photo-inpainting/tree/c42ac41576690b765e50f5281ddbfb58439ff36d | import torch
import torch.nn as nn
class Model(nn.Module):
"""Residual convolution module.
"""
def __init__(self, features):
"""Init.
Args:
features (int): number of features
"""
super().__init__()
self.conv1 = nn.Conv2d(features, features, kernel_size=3, stride=1,
padding=1, bias=True)
self.conv2 = nn.Conv2d(features, features, kernel_size=3, stride=1,
padding=1, bias=False)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
"""Forward pass.
Args:
x (tensor): input
Returns:
tensor: output
"""
out = self.relu(x)
out = self.conv1(out)
out = self.relu(out)
out = self.conv2(out)
return out + x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
LayerNorm | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/zf/czfnaeipqg4a3qzttb2l6zy5ng44vshk3lfmp25jc2er665hxsmw.py
# Topologically Sorted Source Nodes: [mean, sub], Original ATen: [aten.mean, aten.sub]
# Source node to ATen node mapping:
# mean => mean
# sub => sub
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %mean), kwargs = {})
triton_poi_fused_mean_sub_0 = async_compile.triton('triton_poi_fused_mean_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mean_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mean_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 4.0
tmp9 = tmp7 / tmp8
tmp10 = tmp0 - tmp9
tl.store(out_ptr0 + (x3), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ix/cix3agfx5ttdzebqllvt7xsyf7hguiv4c5ya7rpk6x57inkbm4xh.py
# Topologically Sorted Source Nodes: [pow_1, variance, add, rsqrt, x, mul_1, x_1], Original ATen: [aten.pow, aten.mean, aten.add, aten.rsqrt, aten.mul]
# Source node to ATen node mapping:
# add => add
# mul_1 => mul_1
# pow_1 => pow_1
# rsqrt => rsqrt
# variance => mean_1
# x => mul
# x_1 => add_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_1, [1], True), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean_1, 0.0001), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %rsqrt), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %view), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %view_1), kwargs = {})
triton_poi_fused_add_mean_mul_pow_rsqrt_1 = async_compile.triton('triton_poi_fused_add_mean_mul_pow_rsqrt_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mean_mul_pow_rsqrt_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mean_mul_pow_rsqrt_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = 0.0001
tmp15 = tmp13 + tmp14
tmp16 = libdevice.rsqrt(tmp15)
tmp17 = tmp0 * tmp16
tmp19 = tmp17 * tmp18
tmp21 = tmp19 + tmp20
tl.store(out_ptr0 + (x3), tmp21, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean, sub], Original ATen: [aten.mean, aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_mean_sub_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pow_1, variance, add, rsqrt, x, mul_1, x_1], Original ATen: [aten.pow, aten.mean, aten.add, aten.rsqrt, aten.mul]
triton_poi_fused_add_mean_mul_pow_rsqrt_1.run(buf0, primals_2, primals_3, buf1, 256, grid=grid(256), stream=stream0)
del buf0
del primals_2
del primals_3
return (buf1, primals_1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import numpy as np
class BaseModule(torch.nn.Module):
def __init__(self):
super(BaseModule, self).__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class LayerNorm(BaseModule):
def __init__(self, channels, eps=0.0001):
super(LayerNorm, self).__init__()
self.channels = channels
self.eps = eps
self.gamma = torch.nn.Parameter(torch.ones(channels))
self.beta = torch.nn.Parameter(torch.zeros(channels))
def forward(self, x):
n_dims = len(x.shape)
mean = torch.mean(x, 1, keepdim=True)
variance = torch.mean((x - mean) ** 2, 1, keepdim=True)
x = (x - mean) * torch.rsqrt(variance + self.eps)
shape = [1, -1] + [1] * (n_dims - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'channels': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import numpy as np
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mean_sub_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 4.0
tmp9 = tmp7 / tmp8
tmp10 = tmp0 - tmp9
tl.store(out_ptr0 + x3, tmp10, xmask)
@triton.jit
def triton_poi_fused_add_mean_mul_pow_rsqrt_1(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
x1 = xindex // 16 % 4
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp18 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = 0.0001
tmp15 = tmp13 + tmp14
tmp16 = libdevice.rsqrt(tmp15)
tmp17 = tmp0 * tmp16
tmp19 = tmp17 * tmp18
tmp21 = tmp19 + tmp20
tl.store(out_ptr0 + x3, tmp21, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mean_sub_0[grid(256)](primals_1, buf0, 256, XBLOCK
=128, num_warps=4, num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mean_mul_pow_rsqrt_1[grid(256)](buf0,
primals_2, primals_3, buf1, 256, XBLOCK=128, num_warps=4,
num_stages=1)
del buf0
del primals_2
del primals_3
return buf1, primals_1
class BaseModule(torch.nn.Module):
def __init__(self):
super(BaseModule, self).__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class LayerNormNew(BaseModule):
def __init__(self, channels, eps=0.0001):
super(LayerNormNew, self).__init__()
self.channels = channels
self.eps = eps
self.gamma = torch.nn.Parameter(torch.ones(channels))
self.beta = torch.nn.Parameter(torch.zeros(channels))
def forward(self, input_0):
primals_2 = self.gamma
primals_3 = self.beta
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| Sobsz/uberduck-ml-dev | LayerNorm | false | 1,078 | [
"Apache-2.0"
] | 0 | f099238f6f2e3f600d72d89dea3c883c59d91387 | https://github.com/Sobsz/uberduck-ml-dev/tree/f099238f6f2e3f600d72d89dea3c883c59d91387 | import torch
import numpy as np
class BaseModule(torch.nn.Module):
def __init__(self):
super().__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class Model(BaseModule):
def __init__(self, channels, eps=0.0001):
super().__init__()
self.channels = channels
self.eps = eps
self.gamma = torch.nn.Parameter(torch.ones(channels))
self.beta = torch.nn.Parameter(torch.zeros(channels))
def forward(self, x):
n_dims = len(x.shape)
mean = torch.mean(x, 1, keepdim=True)
variance = torch.mean((x - mean) ** 2, 1, keepdim=True)
x = (x - mean) * torch.rsqrt(variance + self.eps)
shape = [1, -1] + [1] * (n_dims - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
Loss | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/rg/crgspi4c5nlqn7h3ahzly6mzubmsww42rdge5ypcptnsedezem2z.py
# Topologically Sorted Source Nodes: [add, log, mul], Original ATen: [aten.add, aten.log, aten.mul]
# Source node to ATen node mapping:
# add => add
# log => log
# mul => mul
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, 1), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%add,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 4), kwargs = {})
triton_poi_fused_add_log_mul_0 = async_compile.triton('triton_poi_fused_add_log_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_log_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_log_mul_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp3 = tl_math.log(tmp2)
tmp4 = 4.0
tmp5 = tmp0 * tmp4
tl.store(out_ptr0 + (x0), tmp3, xmask)
tl.store(out_ptr1 + (x0), tmp5, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/li/clivgwtftzyd2kjyxscelobmyhmudzefuhzuzxrxyxyfbvqhgazb.py
# Topologically Sorted Source Nodes: [sub, mean, loss, neg], Original ATen: [aten.sub, aten.mean, aten.mul, aten.neg]
# Source node to ATen node mapping:
# loss => mul_1
# mean => mean
# neg => neg
# sub => sub
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_2, %view_5), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, 0.25), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%mul_1,), kwargs = {})
triton_per_fused_mean_mul_neg_sub_1 = async_compile.triton('triton_per_fused_mean_mul_neg_sub_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_mul_neg_sub_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_mul_neg_sub_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.load(in_ptr1 + (r0), None)
tmp2 = tmp0 - tmp1
tmp3 = tl.broadcast_to(tmp2, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(tl.sum(tmp3, 0))
tmp6 = 256.0
tmp7 = tmp5 / tmp6
tmp8 = 0.25
tmp9 = tmp7 * tmp8
tmp10 = -tmp9
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp10, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, log, mul], Original ATen: [aten.add, aten.log, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_add_log_mul_0.run(arg0_1, buf0, buf2, 256, grid=grid(256), stream=stream0)
del arg0_1
buf1 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(arg1_1, (16, 4, 4), (16, 4, 1), 0), out=buf1)
del arg1_1
buf3 = reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(arg2_1, (16, 4, 4), (16, 4, 1), 0), out=buf3)
del arg2_1
del buf2
buf4 = empty_strided_cuda((), (), torch.float32)
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [sub, mean, loss, neg], Original ATen: [aten.sub, aten.mean, aten.mul, aten.neg]
triton_per_fused_mean_mul_neg_sub_1.run(buf5, buf1, buf3, 1, 256, grid=grid(1), stream=stream0)
del buf1
del buf3
return (buf5, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch as t
import torch.nn as nn
def indicator(K):
"""
@K: number of users
"""
return t.eye(5 * K)
class Loss(nn.Module):
def __init__(self, K, Nt, Vartheta):
super(Loss, self).__init__()
self.K = K
self.Nt = Nt
self.Delta = indicator(self.K)
self.alpha = 1 / self.K
self.Vartheta = Vartheta
self.batchsize = 10
def forward(self, x, ind1, ind2):
"""
@x: output of the last layer, its dimmension is (batchsize, 2*K*K+3*K)
"""
loss = self.alpha * t.mean(t.matmul(t.log(1 + x), ind1) - t.matmul(
self.Vartheta * x, ind2))
return -loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [[], {'K': 4, 'Nt': 4, 'Vartheta': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch as t
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_log_mul_0(in_ptr0, out_ptr0, out_ptr1, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp3 = tl_math.log(tmp2)
tmp4 = 4.0
tmp5 = tmp0 * tmp4
tl.store(out_ptr0 + x0, tmp3, xmask)
tl.store(out_ptr1 + x0, tmp5, xmask)
@triton.jit
def triton_per_fused_mean_mul_neg_sub_1(in_out_ptr0, in_ptr0, in_ptr1,
xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.load(in_ptr1 + r0, None)
tmp2 = tmp0 - tmp1
tmp3 = tl.broadcast_to(tmp2, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(tl.sum(tmp3, 0))
tmp6 = 256.0
tmp7 = tmp5 / tmp6
tmp8 = 0.25
tmp9 = tmp7 * tmp8
tmp10 = -tmp9
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp10, None)
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_log_mul_0[grid(256)](arg0_1, buf0, buf2, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
buf1 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(arg1_1, (16, 4, 4), (16, 4, 1), 0), out=buf1
)
del arg1_1
buf3 = reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0)
del buf0
extern_kernels.bmm(reinterpret_tensor(buf2, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(arg2_1, (16, 4, 4), (16, 4, 1), 0), out=buf3
)
del arg2_1
del buf2
buf4 = empty_strided_cuda((), (), torch.float32)
buf5 = buf4
del buf4
triton_per_fused_mean_mul_neg_sub_1[grid(1)](buf5, buf1, buf3, 1,
256, num_warps=2, num_stages=1)
del buf1
del buf3
return buf5,
def indicator(K):
"""
@K: number of users
"""
return t.eye(5 * K)
class LossNew(nn.Module):
def __init__(self, K, Nt, Vartheta):
super(LossNew, self).__init__()
self.K = K
self.Nt = Nt
self.Delta = indicator(self.K)
self.alpha = 1 / self.K
self.Vartheta = Vartheta
self.batchsize = 10
def forward(self, input_0, input_1, input_2):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
output = call([arg0_1, arg1_1, arg2_1])
return output[0]
| SoulVen/USRMNet-HWGCN | Loss | false | 1,079 | [
"Apache-2.0"
] | 0 | 2f99f53150335be26270bd408ce59dc51c8435cc | https://github.com/SoulVen/USRMNet-HWGCN/tree/2f99f53150335be26270bd408ce59dc51c8435cc | import torch
import torch as t
import torch.nn as nn
def indicator(K):
"""
@K: number of users
"""
return t.eye(5 * K)
class Model(nn.Module):
def __init__(self, K, Nt, Vartheta):
super().__init__()
self.K = K
self.Nt = Nt
self.Delta = indicator(self.K)
self.alpha = 1 / self.K
self.Vartheta = Vartheta
self.batchsize = 10
def forward(self, x, ind1, ind2):
"""
@x: output of the last layer, its dimmension is (batchsize, 2*K*K+3*K)
"""
loss = self.alpha * t.mean(t.matmul(t.log(1 + x), ind1) - t.matmul(
self.Vartheta * x, ind2))
return -loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
AttNet | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/dq/cdq267yqqozcarovnsrpdv2jgticcu6trrauhb2nqucmjjuwswn5.py
# Topologically Sorted Source Nodes: [concat_feat], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# concat_feat => cat
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_1, %primals_2],), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 64)
x0 = xindex % 64
x2 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (64*x1)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 4, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr1 + (x0 + (64*((-2) + x1))), tmp6 & xmask, other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + (x2), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/d3/cd3wh33oi23pfdaubjmlj5web62n47mpc3cvroacr4ja7m4eyxes.py
# Topologically Sorted Source Nodes: [conv2d, weights], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d => convolution
# weights => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%cat, %primals_3, %primals_4, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_1 = async_compile.triton('triton_poi_fused_convolution_relu_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4096
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 16) % 64
tmp0 = tl.load(in_out_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/k3/ck3ww4lbm3ccb4nqtjzcx44tla623nm2tpo6tnos4pcpcrtdkfly.py
# Topologically Sorted Source Nodes: [conv2d_1, weights_1], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# weights_1 => relu_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_5, %primals_6, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_1,), kwargs = {})
triton_poi_fused_convolution_relu_2 = async_compile.triton('triton_poi_fused_convolution_relu_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 16
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/7d/c7dn7w5djo7urmcn4tiv47ezted4wgg5pnolpuxzmc3glaqrtwgn.py
# Topologically Sorted Source Nodes: [conv2d_2, weights_2], Original ATen: [aten.convolution, aten._softmax]
# Source node to ATen node mapping:
# conv2d_2 => convolution_2
# weights_2 => amax, exp, sub, sum_1
# Graph fragment:
# %convolution_2 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_1, %primals_7, %primals_8, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%convolution_2, [0], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%convolution_2, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [0], True), kwargs = {})
triton_poi_fused__softmax_convolution_3 = async_compile.triton('triton_poi_fused__softmax_convolution_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_convolution_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_convolution_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp4 = tl.load(in_ptr0 + (16 + x0), xmask)
tmp7 = tl.load(in_ptr0 + (32 + x0), xmask)
tmp10 = tl.load(in_ptr0 + (48 + x0), xmask)
tmp3 = tmp0 + tmp2
tmp5 = tmp4 + tmp2
tmp6 = triton_helpers.maximum(tmp3, tmp5)
tmp8 = tmp7 + tmp2
tmp9 = triton_helpers.maximum(tmp6, tmp8)
tmp11 = tmp10 + tmp2
tmp12 = triton_helpers.maximum(tmp9, tmp11)
tmp13 = tmp3 - tmp12
tmp14 = tl_math.exp(tmp13)
tmp15 = tmp5 - tmp12
tmp16 = tl_math.exp(tmp15)
tmp17 = tmp14 + tmp16
tmp18 = tmp8 - tmp12
tmp19 = tl_math.exp(tmp18)
tmp20 = tmp17 + tmp19
tmp21 = tmp11 - tmp12
tmp22 = tl_math.exp(tmp21)
tmp23 = tmp20 + tmp22
tl.store(out_ptr0 + (x0), tmp12, xmask)
tl.store(out_ptr1 + (x0), tmp23, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/vl/cvl4cyhfl76tcxqwtza2w75eqxw7wmlfnbja3yuh3nzvv64atxf4.py
# Topologically Sorted Source Nodes: [conv2d_2, weights_2], Original ATen: [aten.convolution, aten._softmax]
# Source node to ATen node mapping:
# conv2d_2 => convolution_2
# weights_2 => amax, div, exp, sub, sum_1
# Graph fragment:
# %convolution_2 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_1, %primals_7, %primals_8, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%convolution_2, [0], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%convolution_2, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [0], True), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_convolution_4 = async_compile.triton('triton_poi_fused__softmax_convolution_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_convolution_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_convolution_4(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp4 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp0 + tmp2
tmp5 = tmp3 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp8 = tmp6 / tmp7
tl.store(in_out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/uy/cuycvtk2spkqnzbhucr6bdz4m47aquwkozh5wxorhrxoikbox2fk.py
# Topologically Sorted Source Nodes: [weight1, mul, mul_1, out_feat], Original ATen: [aten.repeat, aten.mul, aten.add]
# Source node to ATen node mapping:
# mul => mul
# mul_1 => mul_1
# out_feat => add
# weight1 => repeat
# Graph fragment:
# %repeat : [num_users=2] = call_function[target=torch.ops.aten.repeat.default](args = (%getitem, [1, 4, 1, 1]), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%repeat, %primals_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%repeat, %primals_2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
triton_poi_fused_add_mul_repeat_5 = async_compile.triton('triton_poi_fused_add_mul_repeat_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_repeat_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_repeat_5(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = (xindex // 64)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x3), xmask)
tmp3 = tl.load(in_ptr2 + (x3), xmask)
tmp2 = tmp0 * tmp1
tmp4 = tmp0 * tmp3
tmp5 = tmp2 + tmp4
tl.store(out_ptr0 + (x3), tmp5, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8 = args
args.clear()
assert_size_stride(primals_1, (2, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (64, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_4, (64, ), (1, ))
assert_size_stride(primals_5, (16, 64, 1, 1), (64, 1, 1, 1))
assert_size_stride(primals_6, (16, ), (1, ))
assert_size_stride(primals_7, (1, 16, 1, 1), (16, 1, 1, 1))
assert_size_stride(primals_8, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [concat_feat], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 64, 4, 4), (1024, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [conv2d, weights], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_1.run(buf2, primals_4, 4096, grid=grid(4096), stream=stream0)
del primals_4
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf2, primals_5, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 16, 4, 4), (256, 16, 4, 1))
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [conv2d_1, weights_1], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_2.run(buf4, primals_6, 1024, grid=grid(1024), stream=stream0)
del primals_6
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf5 = extern_kernels.convolution(buf4, primals_7, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 1, 4, 4), (16, 16, 4, 1))
buf6 = empty_strided_cuda((1, 1, 4, 4), (16, 16, 4, 1), torch.float32)
buf7 = empty_strided_cuda((1, 1, 4, 4), (16, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [conv2d_2, weights_2], Original ATen: [aten.convolution, aten._softmax]
triton_poi_fused__softmax_convolution_3.run(buf5, primals_8, buf6, buf7, 16, grid=grid(16), stream=stream0)
buf8 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [conv2d_2, weights_2], Original ATen: [aten.convolution, aten._softmax]
triton_poi_fused__softmax_convolution_4.run(buf8, primals_8, buf6, buf7, 64, grid=grid(64), stream=stream0)
del buf6
del buf7
del primals_8
buf9 = empty_strided_cuda((2, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [weight1, mul, mul_1, out_feat], Original ATen: [aten.repeat, aten.mul, aten.add]
triton_poi_fused_add_mul_repeat_5.run(buf8, primals_1, primals_2, buf9, 128, grid=grid(128), stream=stream0)
return (buf9, primals_1, primals_2, primals_3, primals_5, primals_7, buf0, buf2, buf4, buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((2, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((2, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((64, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((16, 64, 1, 1), (64, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((1, 16, 1, 1), (16, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
class AttNet(nn.Module):
def __init__(self, num_input_ch):
super(AttNet, self).__init__()
self.num_input_ch = num_input_ch
self.conv1 = nn.Conv2d(self.num_input_ch, 64, 3, padding=1, bias=True)
self.conv2 = nn.Conv2d(64, 16, 1, bias=True)
self.conv3 = nn.Conv2d(16, 1, 1, bias=True)
def forward(self, warp_feat, conv_feat):
concat_feat = torch.cat([warp_feat, conv_feat], dim=0)
weights = F.relu(self.conv1(concat_feat))
weights = F.relu(self.conv2(weights))
weights = F.softmax(self.conv3(weights), dim=0)
weights = torch.split(weights, 2, dim=0)
weight1 = torch.tile(weights[0], (1, self.num_input_ch, 1, 1))
weight2 = torch.tile(weights[0], (1, self.num_input_ch, 1, 1))
out_feat = weight1 * warp_feat + weight2 * conv_feat
return out_feat
def get_inputs():
return [torch.rand([2, 4, 4, 4]), torch.rand([2, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_input_ch': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 64
x0 = xindex % 64
x2 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 64 * x1), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 4, tl.int64)
tmp9 = tl.load(in_ptr1 + (x0 + 64 * (-2 + x1)), tmp6 & xmask, other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + x2, tmp10, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_1(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = xindex // 16 % 64
tmp0 = tl.load(in_out_ptr0 + x3, None)
tmp1 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, None)
@triton.jit
def triton_poi_fused_convolution_relu_2(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 16
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused__softmax_convolution_3(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp4 = tl.load(in_ptr0 + (16 + x0), xmask)
tmp7 = tl.load(in_ptr0 + (32 + x0), xmask)
tmp10 = tl.load(in_ptr0 + (48 + x0), xmask)
tmp3 = tmp0 + tmp2
tmp5 = tmp4 + tmp2
tmp6 = triton_helpers.maximum(tmp3, tmp5)
tmp8 = tmp7 + tmp2
tmp9 = triton_helpers.maximum(tmp6, tmp8)
tmp11 = tmp10 + tmp2
tmp12 = triton_helpers.maximum(tmp9, tmp11)
tmp13 = tmp3 - tmp12
tmp14 = tl_math.exp(tmp13)
tmp15 = tmp5 - tmp12
tmp16 = tl_math.exp(tmp15)
tmp17 = tmp14 + tmp16
tmp18 = tmp8 - tmp12
tmp19 = tl_math.exp(tmp18)
tmp20 = tmp17 + tmp19
tmp21 = tmp11 - tmp12
tmp22 = tl_math.exp(tmp21)
tmp23 = tmp20 + tmp22
tl.store(out_ptr0 + x0, tmp12, xmask)
tl.store(out_ptr1 + x0, tmp23, xmask)
@triton.jit
def triton_poi_fused__softmax_convolution_4(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp4 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp0 + tmp2
tmp5 = tmp3 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp8 = tmp6 / tmp7
tl.store(in_out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_add_mul_repeat_5(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex // 64
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + x3, xmask)
tmp3 = tl.load(in_ptr2 + x3, xmask)
tmp2 = tmp0 * tmp1
tmp4 = tmp0 * tmp3
tmp5 = tmp2 + tmp4
tl.store(out_ptr0 + x3, tmp5, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8) = args
args.clear()
assert_size_stride(primals_1, (2, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (2, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (64, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_4, (64,), (1,))
assert_size_stride(primals_5, (16, 64, 1, 1), (64, 1, 1, 1))
assert_size_stride(primals_6, (16,), (1,))
assert_size_stride(primals_7, (1, 16, 1, 1), (16, 1, 1, 1))
assert_size_stride(primals_8, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(256)](primals_1, primals_2, buf0, 256,
XBLOCK=256, num_warps=4, num_stages=1)
buf1 = extern_kernels.convolution(buf0, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 64, 4, 4), (1024, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_relu_1[grid(4096)](buf2, primals_4,
4096, XBLOCK=256, num_warps=4, num_stages=1)
del primals_4
buf3 = extern_kernels.convolution(buf2, primals_5, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 16, 4, 4), (256, 16, 4, 1))
buf4 = buf3
del buf3
triton_poi_fused_convolution_relu_2[grid(1024)](buf4, primals_6,
1024, XBLOCK=256, num_warps=4, num_stages=1)
del primals_6
buf5 = extern_kernels.convolution(buf4, primals_7, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 1, 4, 4), (16, 16, 4, 1))
buf6 = empty_strided_cuda((1, 1, 4, 4), (16, 16, 4, 1), torch.float32)
buf7 = empty_strided_cuda((1, 1, 4, 4), (16, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_convolution_3[grid(16)](buf5, primals_8,
buf6, buf7, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf8 = buf5
del buf5
triton_poi_fused__softmax_convolution_4[grid(64)](buf8, primals_8,
buf6, buf7, 64, XBLOCK=64, num_warps=1, num_stages=1)
del buf6
del buf7
del primals_8
buf9 = empty_strided_cuda((2, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_mul_repeat_5[grid(128)](buf8, primals_1,
primals_2, buf9, 128, XBLOCK=128, num_warps=4, num_stages=1)
return (buf9, primals_1, primals_2, primals_3, primals_5, primals_7,
buf0, buf2, buf4, buf8)
class AttNetNew(nn.Module):
def __init__(self, num_input_ch):
super(AttNetNew, self).__init__()
self.num_input_ch = num_input_ch
self.conv1 = nn.Conv2d(self.num_input_ch, 64, 3, padding=1, bias=True)
self.conv2 = nn.Conv2d(64, 16, 1, bias=True)
self.conv3 = nn.Conv2d(16, 1, 1, bias=True)
def forward(self, input_0, input_1):
primals_3 = self.conv1.weight
primals_4 = self.conv1.bias
primals_5 = self.conv2.weight
primals_6 = self.conv2.bias
primals_7 = self.conv3.weight
primals_8 = self.conv3.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8])
return output[0]
| SionHu/LP-MOT | AttNet | false | 1,080 | [
"MIT"
] | 0 | 90e6a1d51ebe1a948ac5c018a5ee560654e824f1 | https://github.com/SionHu/LP-MOT/tree/90e6a1d51ebe1a948ac5c018a5ee560654e824f1 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, num_input_ch):
super().__init__()
self.num_input_ch = num_input_ch
self.conv1 = nn.Conv2d(self.num_input_ch, 64, 3, padding=1, bias=True)
self.conv2 = nn.Conv2d(64, 16, 1, bias=True)
self.conv3 = nn.Conv2d(16, 1, 1, bias=True)
def forward(self, warp_feat, conv_feat):
concat_feat = torch.cat([warp_feat, conv_feat], dim=0)
weights = F.relu(self.conv1(concat_feat))
weights = F.relu(self.conv2(weights))
weights = F.softmax(self.conv3(weights), dim=0)
weights = torch.split(weights, 2, dim=0)
weight1 = torch.tile(weights[0], (1, self.num_input_ch, 1, 1))
weight2 = torch.tile(weights[0], (1, self.num_input_ch, 1, 1))
out_feat = weight1 * warp_feat + weight2 * conv_feat
return out_feat
def get_inputs():
return [torch.rand([2, 4, 4, 4]), torch.rand([2, 4, 4, 4])]
def get_init_inputs():
return [4]
|
Net | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/es/cesrwuwkeluuf67vbj367seoctnhxwytz3gxjpe3dmncvzr46phg.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# out => cat
# Graph fragment:
# %cat : [num_users=3] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_2, %mm], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tl.store(out_ptr0 + (x0 + (8*x1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/d2/cd2iuy4xijl6opmrfcyh5n4ktkq3vxdew7ukyggi4oe4nmubofvx.py
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# out_1 => cat_1
# Graph fragment:
# %cat_1 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%cat, %mm_1], 1), kwargs = {})
triton_poi_fused_cat_1 = async_compile.triton('triton_poi_fused_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 8
x1 = (xindex // 8)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tl.store(out_ptr0 + (x0 + (12*x1)), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 8), (8, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf2 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
buf0 = reinterpret_tensor(buf2, (4, 4), (8, 1), 4) # alias
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.mm]
extern_kernels.mm(primals_2, reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf2, (4, 4), (8, 1), 0) # alias
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_2, buf1, 16, grid=grid(16), stream=stream0)
buf5 = empty_strided_cuda((4, 12), (12, 1), torch.float32)
buf3 = reinterpret_tensor(buf5, (4, 4), (12, 1), 8) # alias
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.mm]
extern_kernels.mm(buf2, reinterpret_tensor(primals_3, (8, 4), (1, 8), 0), out=buf3)
buf4 = reinterpret_tensor(buf5, (4, 8), (12, 1), 0) # alias
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(buf2, buf4, 32, grid=grid(32), stream=stream0)
return (buf5, primals_2, buf2, primals_3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 8), (8, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class FcCat(nn.Module):
def __init__(self, nIn, nOut):
super(FcCat, self).__init__()
self.fc = nn.Linear(nIn, nOut, bias=False)
def forward(self, x):
out = torch.cat((x, self.fc(x)), 1)
return out
class Net(nn.Module):
def __init__(self, nFeatures, nHidden1, nHidden2):
super(Net, self).__init__()
self.l1 = FcCat(nFeatures, nHidden1)
self.l2 = FcCat(nFeatures + nHidden1, nHidden2)
def forward(self, x):
out = self.l1(x)
out = self.l2(out)
return out
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'nFeatures': 4, 'nHidden1': 4, 'nHidden2': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tl.store(out_ptr0 + (x0 + 8 * x1), tmp0, xmask)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 8
x1 = xindex // 8
tmp0 = tl.load(in_ptr0 + x2, xmask)
tl.store(out_ptr0 + (x0 + 12 * x1), tmp0, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, 8), (8, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf2 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
buf0 = reinterpret_tensor(buf2, (4, 4), (8, 1), 4)
extern_kernels.mm(primals_2, reinterpret_tensor(primals_1, (4, 4),
(1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf2, (4, 4), (8, 1), 0)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(16)](primals_2, buf1, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 12), (12, 1), torch.float32)
buf3 = reinterpret_tensor(buf5, (4, 4), (12, 1), 8)
extern_kernels.mm(buf2, reinterpret_tensor(primals_3, (8, 4), (1, 8
), 0), out=buf3)
buf4 = reinterpret_tensor(buf5, (4, 8), (12, 1), 0)
triton_poi_fused_cat_1[grid(32)](buf2, buf4, 32, XBLOCK=32,
num_warps=1, num_stages=1)
return buf5, primals_2, buf2, primals_3
class FcCat(nn.Module):
def __init__(self, nIn, nOut):
super(FcCat, self).__init__()
self.fc = nn.Linear(nIn, nOut, bias=False)
def forward(self, x):
out = torch.cat((x, self.fc(x)), 1)
return out
class NetNew(nn.Module):
def __init__(self, nFeatures, nHidden1, nHidden2):
super(NetNew, self).__init__()
self.l1 = FcCat(nFeatures, nHidden1)
self.l2 = FcCat(nFeatures + nHidden1, nHidden2)
def forward(self, input_0):
primals_1 = self.l1.fc.weight
primals_3 = self.l2.fc.weight
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| Sreehari-S/Tiramisu_DigestPath | Net | false | 1,081 | [
"Apache-2.0"
] | 0 | a884ee911bc60ce997996e0ec2e6036600ffcffa | https://github.com/Sreehari-S/Tiramisu_DigestPath/tree/a884ee911bc60ce997996e0ec2e6036600ffcffa | import torch
import torch.nn as nn
class FcCat(nn.Module):
def __init__(self, nIn, nOut):
super().__init__()
self.fc = nn.Linear(nIn, nOut, bias=False)
def forward(self, x):
out = torch.cat((x, self.fc(x)), 1)
return out
class Model(nn.Module):
def __init__(self, nFeatures, nHidden1, nHidden2):
super().__init__()
self.l1 = FcCat(nFeatures, nHidden1)
self.l2 = FcCat(nFeatures + nHidden1, nHidden2)
def forward(self, x):
out = self.l1(x)
out = self.l2(out)
return out
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
DecoderLayer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/fg/cfg742icmosiwp5ugziye26din5ueqx3v7ntptkkpyackudldrxs.py
# Topologically Sorted Source Nodes: [eq], Original ATen: [aten.eq]
# Source node to ATen node mapping:
# eq => eq
# Graph fragment:
# %eq : [num_users=2] = call_function[target=torch.ops.aten.eq.Scalar](args = (%primals_8, 0), kwargs = {})
triton_poi_fused_eq_0 = async_compile.triton('triton_poi_fused_eq_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i1', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_eq_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_eq_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.0
tmp2 = tmp0 == tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/g7/cg74ay746yb3ivb67q6pz3e2xibwamviqoisgmyzopmnr47hrosh.py
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
# Source node to ATen node mapping:
# attention_score_1 => div
# attention_score_2 => full_default, where
# attention_score_3 => amax, exp, sub, sum_1
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_11, 1.0), kwargs = {})
# %full_default : [num_users=2] = call_function[target=torch.ops.aten.full.default](args = ([], -1000000000.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %div), kwargs = {})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%where, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
triton_poi_fused__softmax_div_masked_fill_1 = async_compile.triton('triton_poi_fused__softmax_div_masked_fill_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*i1', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_div_masked_fill_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (4*x2), xmask, eviction_policy='evict_last').to(tl.int1)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (4*x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (1 + (4*x2)), xmask, eviction_policy='evict_last').to(tl.int1)
tmp9 = tl.load(in_ptr2 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (2 + (4*x2)), xmask, eviction_policy='evict_last').to(tl.int1)
tmp15 = tl.load(in_ptr2 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr0 + (3 + (4*x2)), xmask, eviction_policy='evict_last').to(tl.int1)
tmp21 = tl.load(in_ptr2 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp10 = tmp1 * tmp9
tmp11 = tmp10 * tmp4
tmp12 = tl.where(tmp8, tmp6, tmp11)
tmp13 = triton_helpers.maximum(tmp7, tmp12)
tmp16 = tmp1 * tmp15
tmp17 = tmp16 * tmp4
tmp18 = tl.where(tmp14, tmp6, tmp17)
tmp19 = triton_helpers.maximum(tmp13, tmp18)
tmp22 = tmp1 * tmp21
tmp23 = tmp22 * tmp4
tmp24 = tl.where(tmp20, tmp6, tmp23)
tmp25 = triton_helpers.maximum(tmp19, tmp24)
tmp26 = tmp7 - tmp25
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp12 - tmp25
tmp29 = tl_math.exp(tmp28)
tmp30 = tmp27 + tmp29
tmp31 = tmp18 - tmp25
tmp32 = tl_math.exp(tmp31)
tmp33 = tmp30 + tmp32
tmp34 = tmp24 - tmp25
tmp35 = tl_math.exp(tmp34)
tmp36 = tmp33 + tmp35
tl.store(out_ptr0 + (x2), tmp25, xmask)
tl.store(out_ptr1 + (x2), tmp36, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/na/cna5revtaqihb3t6bnp24kdg6eqircycsi7slr6uzmhkwpelq6x7.py
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
# Source node to ATen node mapping:
# attention_score_1 => div
# attention_score_2 => full_default, where
# attention_score_3 => div_1, exp, sub
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_11, 1.0), kwargs = {})
# %full_default : [num_users=2] = call_function[target=torch.ops.aten.full.default](args = ([], -1000000000.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %div), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_div_masked_fill_2 = async_compile.triton('triton_poi_fused__softmax_div_masked_fill_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*i1', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_div_masked_fill_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x4 = (xindex // 4)
x0 = xindex % 4
x2 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x3), xmask).to(tl.int1)
tmp1 = tl.load(in_ptr1 + (x4), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr3 + (x4), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr4 + (x4), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp9 = tmp7 - tmp8
tmp10 = tl_math.exp(tmp9)
tmp12 = tmp10 / tmp11
tl.store(out_ptr0 + (x3), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/tc/ctcugqu2nbqlxcf2thnspnnypxifbalbzmclmutd5vaxdes2oyyk.py
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# layer_norm => var_mean
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %primals_1), kwargs = {})
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add, [2]), kwargs = {correction: 0, keepdim: True})
triton_poi_fused_add_native_layer_norm_3 = async_compile.triton('triton_poi_fused_add_native_layer_norm_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + (x0), tmp16, xmask)
tl.store(out_ptr1 + (x0), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ua/cuaquah4oaz43nhi25wixnpzlhvf2zfzdiezhmmwkuy5wfhtw6z4.py
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# layer_norm => add_1, add_2, mul_1, mul_2, rsqrt, sub_1
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %primals_1), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %getitem_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %rsqrt), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_1, %primals_11), kwargs = {})
# %add_2 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %primals_12), kwargs = {})
triton_poi_fused_add_native_layer_norm_4 = async_compile.triton('triton_poi_fused_add_native_layer_norm_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + (x2), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/56/c566ex7xddxc2fqpwqlmymdyd23nesbsyghxftm7cy73ebnuo3ke.py
# Topologically Sorted Source Nodes: [add_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# add_1 => add_3
# Graph fragment:
# %add_3 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_35, %add_2), kwargs = {})
triton_poi_fused_add_5 = async_compile.triton('triton_poi_fused_add_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_5(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/7b/c7b2yb3hgvixcalmjh5bpsuibu3izis7v3a6ccjg3fsa77ovsm2d.py
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# layer_norm_1 => add_4, rsqrt_1, var_mean_1
# Graph fragment:
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add_3, [2]), kwargs = {correction: 0, keepdim: True})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-05), kwargs = {})
# %rsqrt_1 : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_4,), kwargs = {})
triton_poi_fused_native_layer_norm_6 = async_compile.triton('triton_poi_fused_native_layer_norm_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_6(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + (x0), tmp8, xmask)
tl.store(out_ptr1 + (x0), tmp23, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/z6/cz657lehnuecvtrckws4hxctt6heob44wmjdfvlblgm7yc4swnal.py
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# layer_norm_1 => add_4, add_5, mul_4, mul_5, rsqrt_1, sub_3, var_mean_1
# Graph fragment:
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add_3, [2]), kwargs = {correction: 0, keepdim: True})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-05), kwargs = {})
# %rsqrt_1 : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_4,), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_3, %getitem_3), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_3, %rsqrt_1), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_4, %primals_23), kwargs = {})
# %add_5 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_5, %primals_24), kwargs = {})
triton_poi_fused_native_layer_norm_7 = async_compile.triton('triton_poi_fused_native_layer_norm_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_7', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_7(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/23/c23j2vk753qpctrm5kblwdo7f2zh4pnjylmlrcdhyl7syfjonudr.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_37,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_8 = async_compile.triton('triton_poi_fused_relu_threshold_backward_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_8', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_8(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25, primals_26, primals_27, primals_28, primals_29, primals_30 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, ), (1, ))
assert_size_stride(primals_12, (4, ), (1, ))
assert_size_stride(primals_13, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_14, (4, 4), (4, 1))
assert_size_stride(primals_15, (4, ), (1, ))
assert_size_stride(primals_16, (4, 4), (4, 1))
assert_size_stride(primals_17, (4, ), (1, ))
assert_size_stride(primals_18, (4, 4), (4, 1))
assert_size_stride(primals_19, (4, ), (1, ))
assert_size_stride(primals_20, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_21, (4, 4), (4, 1))
assert_size_stride(primals_22, (4, ), (1, ))
assert_size_stride(primals_23, (4, ), (1, ))
assert_size_stride(primals_24, (4, ), (1, ))
assert_size_stride(primals_25, (4, 4), (4, 1))
assert_size_stride(primals_26, (4, ), (1, ))
assert_size_stride(primals_27, (4, 4), (4, 1))
assert_size_stride(primals_28, (4, ), (1, ))
assert_size_stride(primals_29, (4, ), (1, ))
assert_size_stride(primals_30, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_query], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_key], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_value], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_6
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [eq], Original ATen: [aten.eq]
stream0 = get_raw_stream(0)
triton_poi_fused_eq_0.run(primals_8, buf3, 256, grid=grid(256), stream=stream0)
del primals_8
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
triton_poi_fused__softmax_div_masked_fill_1.run(buf3, buf0, buf1, buf4, buf5, 64, grid=grid(64), stream=stream0)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
triton_poi_fused__softmax_div_masked_fill_2.run(buf3, buf0, buf1, buf4, buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [result], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf6, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0), out=buf7)
buf8 = reinterpret_tensor(buf4, (16, 4), (4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [output_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_10, reinterpret_tensor(buf7, (16, 4), (4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf8)
del primals_10
buf9 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
buf10 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_3.run(buf8, primals_1, buf9, buf10, 16, grid=grid(16), stream=stream0)
buf11 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_4.run(buf8, primals_1, buf9, buf10, primals_11, primals_12, buf11, 64, grid=grid(64), stream=stream0)
del primals_12
buf12 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_query_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_15, reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(primals_14, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf12)
del primals_15
buf13 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_key_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_17, reinterpret_tensor(primals_13, (16, 4), (4, 1), 0), reinterpret_tensor(primals_16, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf13)
del primals_16
del primals_17
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_value_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_19, reinterpret_tensor(primals_13, (16, 4), (4, 1), 0), reinterpret_tensor(primals_18, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf14)
del primals_18
del primals_19
buf15 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [eq_1], Original ATen: [aten.eq]
triton_poi_fused_eq_0.run(primals_20, buf15, 256, grid=grid(256), stream=stream0)
del primals_20
buf16 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf17 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_2, attention_score_5, attention_score_6, attention_score_7], Original ATen: [aten.masked_fill, aten.div, aten._softmax]
triton_poi_fused__softmax_div_masked_fill_1.run(buf15, buf12, buf13, buf16, buf17, 64, grid=grid(64), stream=stream0)
buf18 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_2, attention_score_5, attention_score_6, attention_score_7], Original ATen: [aten.masked_fill, aten.div, aten._softmax]
triton_poi_fused__softmax_div_masked_fill_2.run(buf15, buf12, buf13, buf16, buf17, buf18, 256, grid=grid(256), stream=stream0)
buf19 = reinterpret_tensor(buf17, (16, 4, 1), (4, 1, 1), 0); del buf17 # reuse
# Topologically Sorted Source Nodes: [result_1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf18, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf14, (16, 4, 1), (4, 1, 1), 0), out=buf19)
buf20 = reinterpret_tensor(buf16, (16, 4), (4, 1), 0); del buf16 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf19, (16, 4), (4, 1), 0), reinterpret_tensor(primals_21, (4, 4), (1, 4), 0), out=buf20)
buf21 = reinterpret_tensor(buf20, (4, 4, 4), (16, 4, 1), 0); del buf20 # reuse
# Topologically Sorted Source Nodes: [add_1], Original ATen: [aten.add]
triton_poi_fused_add_5.run(buf21, primals_22, buf11, 64, grid=grid(64), stream=stream0)
del primals_22
buf22 = buf9; del buf9 # reuse
buf23 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_6.run(buf21, buf22, buf23, 16, grid=grid(16), stream=stream0)
buf24 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_7.run(buf21, buf22, buf23, primals_23, primals_24, buf24, 64, grid=grid(64), stream=stream0)
del primals_24
buf25 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf24, (16, 4), (4, 1), 0), reinterpret_tensor(primals_25, (4, 4), (1, 4), 0), out=buf25)
buf26 = reinterpret_tensor(buf25, (4, 4, 4), (16, 4, 1), 0); del buf25 # reuse
buf32 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_8.run(buf26, primals_26, buf32, 64, grid=grid(64), stream=stream0)
del primals_26
buf27 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf26, (16, 4), (4, 1), 0), reinterpret_tensor(primals_27, (4, 4), (1, 4), 0), out=buf27)
buf28 = reinterpret_tensor(buf27, (4, 4, 4), (16, 4, 1), 0); del buf27 # reuse
# Topologically Sorted Source Nodes: [add_2], Original ATen: [aten.add]
triton_poi_fused_add_5.run(buf28, primals_28, buf24, 64, grid=grid(64), stream=stream0)
del primals_28
buf29 = buf23; del buf23 # reuse
buf30 = buf22; del buf22 # reuse
# Topologically Sorted Source Nodes: [layer_norm_2], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_6.run(buf28, buf29, buf30, 16, grid=grid(16), stream=stream0)
buf31 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [layer_norm_2], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_7.run(buf28, buf29, buf30, primals_29, primals_30, buf31, 64, grid=grid(64), stream=stream0)
del buf29
del buf30
del primals_30
return (buf31, primals_1, primals_11, primals_23, primals_29, buf0, buf1, buf3, buf6, reinterpret_tensor(buf7, (16, 4), (4, 1), 0), buf8, reinterpret_tensor(buf11, (16, 4), (4, 1), 0), buf12, reinterpret_tensor(primals_13, (16, 4), (4, 1), 0), buf13, buf15, buf18, reinterpret_tensor(buf19, (16, 4), (4, 1), 0), buf21, reinterpret_tensor(buf24, (16, 4), (4, 1), 0), reinterpret_tensor(buf26, (16, 4), (4, 1), 0), buf28, primals_27, buf32, primals_25, primals_21, reinterpret_tensor(buf14, (16, 1, 4), (4, 1, 1), 0), primals_14, primals_9, reinterpret_tensor(buf2, (16, 1, 4), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_19 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_20 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_21 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_22 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_23 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_24 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_25 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_26 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_27 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_28 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_29 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_30 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25, primals_26, primals_27, primals_28, primals_29, primals_30])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super(MultiHeadAttention, self).__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
class FeedForward(nn.Module):
def __init__(self, model_dim, hidden_dim, drop_prob):
super(FeedForward, self).__init__()
self.model_dim = model_dim
self.hidden_dim = hidden_dim
self.drop_prob = drop_prob
self.linearlayer1 = nn.Linear(model_dim, hidden_dim)
self.linearlayer2 = nn.Linear(hidden_dim, model_dim)
self.relu = nn.ReLU()
self.dropout = nn.Dropout(drop_prob)
def forward(self, tensor):
tensor = self.dropout(self.relu(self.linearlayer1(tensor)))
return self.linearlayer2(tensor)
class DecoderLayer(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, hidden_dim, num_head,
drop_prob):
super(DecoderLayer, self).__init__()
self.self_attention = MultiHeadAttention(model_dim, key_dim,
value_dim, num_head)
self.normalization1 = nn.LayerNorm(model_dim)
self.dropout1 = nn.Dropout(drop_prob)
self.enc_dec_attention = MultiHeadAttention(model_dim, key_dim,
value_dim, num_head)
self.normalization2 = nn.LayerNorm(model_dim)
self.dropout2 = nn.Dropout(drop_prob)
self.ffn = FeedForward(model_dim, hidden_dim, drop_prob)
self.normalization3 = nn.LayerNorm(model_dim)
self.dropout3 = nn.Dropout(drop_prob)
def forward(self, dec_tensor, enc_tensor, source_mask, target_mask):
residual = dec_tensor
tensor = self.self_attention(query=dec_tensor, key=dec_tensor,
value=dec_tensor, mask=target_mask)
tensor = self.dropout1(self.normalization1(tensor + residual))
if enc_tensor is not None:
residual = tensor
tensor = self.enc_dec_attention(query=tensor, key=enc_tensor,
value=enc_tensor, mask=source_mask)
tensor = self.dropout2(self.normalization2(tensor + residual))
residual = tensor
tensor = self.ffn(tensor)
tensor = self.dropout3(self.normalization3(tensor + residual))
return tensor
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4,
4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'model_dim': 4, 'key_dim': 4, 'value_dim': 4, 'hidden_dim':
4, 'num_head': 4, 'drop_prob': 0.5}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import math
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_eq_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.0
tmp2 = tmp0 == tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_1(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + 4 * x2, xmask, eviction_policy='evict_last').to(tl
.int1)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + 4 * x1, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (1 + 4 * x2), xmask, eviction_policy='evict_last'
).to(tl.int1)
tmp9 = tl.load(in_ptr2 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (2 + 4 * x2), xmask, eviction_policy='evict_last'
).to(tl.int1)
tmp15 = tl.load(in_ptr2 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp20 = tl.load(in_ptr0 + (3 + 4 * x2), xmask, eviction_policy='evict_last'
).to(tl.int1)
tmp21 = tl.load(in_ptr2 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp10 = tmp1 * tmp9
tmp11 = tmp10 * tmp4
tmp12 = tl.where(tmp8, tmp6, tmp11)
tmp13 = triton_helpers.maximum(tmp7, tmp12)
tmp16 = tmp1 * tmp15
tmp17 = tmp16 * tmp4
tmp18 = tl.where(tmp14, tmp6, tmp17)
tmp19 = triton_helpers.maximum(tmp13, tmp18)
tmp22 = tmp1 * tmp21
tmp23 = tmp22 * tmp4
tmp24 = tl.where(tmp20, tmp6, tmp23)
tmp25 = triton_helpers.maximum(tmp19, tmp24)
tmp26 = tmp7 - tmp25
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp12 - tmp25
tmp29 = tl_math.exp(tmp28)
tmp30 = tmp27 + tmp29
tmp31 = tmp18 - tmp25
tmp32 = tl_math.exp(tmp31)
tmp33 = tmp30 + tmp32
tmp34 = tmp24 - tmp25
tmp35 = tl_math.exp(tmp34)
tmp36 = tmp33 + tmp35
tl.store(out_ptr0 + x2, tmp25, xmask)
tl.store(out_ptr1 + x2, tmp36, xmask)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_2(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x4 = xindex // 4
x0 = xindex % 4
x2 = xindex // 16
tmp0 = tl.load(in_ptr0 + x3, xmask).to(tl.int1)
tmp1 = tl.load(in_ptr1 + x4, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp8 = tl.load(in_ptr3 + x4, xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr4 + x4, xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp9 = tmp7 - tmp8
tmp10 = tl_math.exp(tmp9)
tmp12 = tmp10 / tmp11
tl.store(out_ptr0 + x3, tmp12, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_3(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp12 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + x0, tmp16, xmask)
tl.store(out_ptr1 + x0, tmp28, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_4(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + x2, tmp13, xmask)
@triton.jit
def triton_poi_fused_add_5(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_native_layer_norm_6(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + x0, tmp8, xmask)
tl.store(out_ptr1 + x0, tmp23, xmask)
@triton.jit
def triton_poi_fused_native_layer_norm_7(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_8(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17,
primals_18, primals_19, primals_20, primals_21, primals_22,
primals_23, primals_24, primals_25, primals_26, primals_27,
primals_28, primals_29, primals_30) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4,), (1,))
assert_size_stride(primals_12, (4,), (1,))
assert_size_stride(primals_13, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_14, (4, 4), (4, 1))
assert_size_stride(primals_15, (4,), (1,))
assert_size_stride(primals_16, (4, 4), (4, 1))
assert_size_stride(primals_17, (4,), (1,))
assert_size_stride(primals_18, (4, 4), (4, 1))
assert_size_stride(primals_19, (4,), (1,))
assert_size_stride(primals_20, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_21, (4, 4), (4, 1))
assert_size_stride(primals_22, (4,), (1,))
assert_size_stride(primals_23, (4,), (1,))
assert_size_stride(primals_24, (4,), (1,))
assert_size_stride(primals_25, (4, 4), (4, 1))
assert_size_stride(primals_26, (4,), (1,))
assert_size_stride(primals_27, (4, 4), (4, 1))
assert_size_stride(primals_28, (4,), (1,))
assert_size_stride(primals_29, (4,), (1,))
assert_size_stride(primals_30, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(primals_1, (16,
4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_1, (16,
4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(primals_1, (16,
4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf2)
del primals_6
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_eq_0[grid(256)](primals_8, buf3, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_8
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_div_masked_fill_1[grid(64)](buf3, buf0,
buf1, buf4, buf5, 64, XBLOCK=64, num_warps=1, num_stages=1)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_div_masked_fill_2[grid(256)](buf3, buf0,
buf1, buf4, buf5, buf6, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf7 = reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0)
del buf5
extern_kernels.bmm(reinterpret_tensor(buf6, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0), out=buf7)
buf8 = reinterpret_tensor(buf4, (16, 4), (4, 1), 0)
del buf4
extern_kernels.addmm(primals_10, reinterpret_tensor(buf7, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf8)
del primals_10
buf9 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
buf10 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
triton_poi_fused_add_native_layer_norm_3[grid(16)](buf8, primals_1,
buf9, buf10, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf11 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_native_layer_norm_4[grid(64)](buf8, primals_1,
buf9, buf10, primals_11, primals_12, buf11, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_12
buf12 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_15, reinterpret_tensor(buf11, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_14, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf12)
del primals_15
buf13 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_17, reinterpret_tensor(primals_13, (16,
4), (4, 1), 0), reinterpret_tensor(primals_16, (4, 4), (1, 4),
0), alpha=1, beta=1, out=buf13)
del primals_16
del primals_17
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_19, reinterpret_tensor(primals_13, (16,
4), (4, 1), 0), reinterpret_tensor(primals_18, (4, 4), (1, 4),
0), alpha=1, beta=1, out=buf14)
del primals_18
del primals_19
buf15 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_eq_0[grid(256)](primals_20, buf15, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_20
buf16 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf17 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_div_masked_fill_1[grid(64)](buf15, buf12,
buf13, buf16, buf17, 64, XBLOCK=64, num_warps=1, num_stages=1)
buf18 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_div_masked_fill_2[grid(256)](buf15, buf12,
buf13, buf16, buf17, buf18, 256, XBLOCK=128, num_warps=4,
num_stages=1)
buf19 = reinterpret_tensor(buf17, (16, 4, 1), (4, 1, 1), 0)
del buf17
extern_kernels.bmm(reinterpret_tensor(buf18, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf14, (16, 4, 1), (4, 1, 1), 0), out=buf19)
buf20 = reinterpret_tensor(buf16, (16, 4), (4, 1), 0)
del buf16
extern_kernels.mm(reinterpret_tensor(buf19, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_21, (4, 4), (1, 4), 0), out=buf20)
buf21 = reinterpret_tensor(buf20, (4, 4, 4), (16, 4, 1), 0)
del buf20
triton_poi_fused_add_5[grid(64)](buf21, primals_22, buf11, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_22
buf22 = buf9
del buf9
buf23 = buf10
del buf10
triton_poi_fused_native_layer_norm_6[grid(16)](buf21, buf22, buf23,
16, XBLOCK=16, num_warps=1, num_stages=1)
buf24 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_native_layer_norm_7[grid(64)](buf21, buf22, buf23,
primals_23, primals_24, buf24, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del primals_24
buf25 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf24, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_25, (4, 4), (1, 4), 0), out=buf25)
buf26 = reinterpret_tensor(buf25, (4, 4, 4), (16, 4, 1), 0)
del buf25
buf32 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_8[grid(64)](buf26,
primals_26, buf32, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_26
buf27 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf26, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_27, (4, 4), (1, 4), 0), out=buf27)
buf28 = reinterpret_tensor(buf27, (4, 4, 4), (16, 4, 1), 0)
del buf27
triton_poi_fused_add_5[grid(64)](buf28, primals_28, buf24, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_28
buf29 = buf23
del buf23
buf30 = buf22
del buf22
triton_poi_fused_native_layer_norm_6[grid(16)](buf28, buf29, buf30,
16, XBLOCK=16, num_warps=1, num_stages=1)
buf31 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_native_layer_norm_7[grid(64)](buf28, buf29, buf30,
primals_29, primals_30, buf31, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf29
del buf30
del primals_30
return (buf31, primals_1, primals_11, primals_23, primals_29, buf0,
buf1, buf3, buf6, reinterpret_tensor(buf7, (16, 4), (4, 1), 0),
buf8, reinterpret_tensor(buf11, (16, 4), (4, 1), 0), buf12,
reinterpret_tensor(primals_13, (16, 4), (4, 1), 0), buf13, buf15,
buf18, reinterpret_tensor(buf19, (16, 4), (4, 1), 0), buf21,
reinterpret_tensor(buf24, (16, 4), (4, 1), 0), reinterpret_tensor(
buf26, (16, 4), (4, 1), 0), buf28, primals_27, buf32, primals_25,
primals_21, reinterpret_tensor(buf14, (16, 1, 4), (4, 1, 1), 0),
primals_14, primals_9, reinterpret_tensor(buf2, (16, 1, 4), (4, 1,
1), 0))
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super(MultiHeadAttention, self).__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
class FeedForward(nn.Module):
def __init__(self, model_dim, hidden_dim, drop_prob):
super(FeedForward, self).__init__()
self.model_dim = model_dim
self.hidden_dim = hidden_dim
self.drop_prob = drop_prob
self.linearlayer1 = nn.Linear(model_dim, hidden_dim)
self.linearlayer2 = nn.Linear(hidden_dim, model_dim)
self.relu = nn.ReLU()
self.dropout = nn.Dropout(drop_prob)
def forward(self, tensor):
tensor = self.dropout(self.relu(self.linearlayer1(tensor)))
return self.linearlayer2(tensor)
class DecoderLayerNew(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, hidden_dim, num_head,
drop_prob):
super(DecoderLayerNew, self).__init__()
self.self_attention = MultiHeadAttention(model_dim, key_dim,
value_dim, num_head)
self.normalization1 = nn.LayerNorm(model_dim)
self.dropout1 = nn.Dropout(drop_prob)
self.enc_dec_attention = MultiHeadAttention(model_dim, key_dim,
value_dim, num_head)
self.normalization2 = nn.LayerNorm(model_dim)
self.dropout2 = nn.Dropout(drop_prob)
self.ffn = FeedForward(model_dim, hidden_dim, drop_prob)
self.normalization3 = nn.LayerNorm(model_dim)
self.dropout3 = nn.Dropout(drop_prob)
def forward(self, input_0, input_1, input_2, input_3):
primals_2 = self.self_attention.Wq.weight
primals_3 = self.self_attention.Wq.bias
primals_4 = self.self_attention.Wk.weight
primals_5 = self.self_attention.Wk.bias
primals_6 = self.self_attention.Wv.weight
primals_7 = self.self_attention.Wv.bias
primals_9 = self.self_attention.Wo.weight
primals_10 = self.self_attention.Wo.bias
primals_11 = self.normalization1.weight
primals_12 = self.normalization1.bias
primals_14 = self.enc_dec_attention.Wq.weight
primals_15 = self.enc_dec_attention.Wq.bias
primals_16 = self.enc_dec_attention.Wk.weight
primals_17 = self.enc_dec_attention.Wk.bias
primals_18 = self.enc_dec_attention.Wv.weight
primals_19 = self.enc_dec_attention.Wv.bias
primals_21 = self.enc_dec_attention.Wo.weight
primals_22 = self.enc_dec_attention.Wo.bias
primals_23 = self.normalization2.weight
primals_24 = self.normalization2.bias
primals_25 = self.ffn.linearlayer1.weight
primals_26 = self.ffn.linearlayer1.bias
primals_27 = self.ffn.linearlayer2.weight
primals_28 = self.ffn.linearlayer2.bias
primals_29 = self.normalization3.weight
primals_30 = self.normalization3.bias
primals_1 = input_0
primals_13 = input_1
primals_8 = input_2
primals_20 = input_3
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18, primals_19,
primals_20, primals_21, primals_22, primals_23, primals_24,
primals_25, primals_26, primals_27, primals_28, primals_29,
primals_30])
return output[0]
| SeungoneKim/Transformer_implementation | DecoderLayer | false | 1,082 | [
"Apache-2.0"
] | 0 | a52bf552eb645fc9bfb812cc26842fc147d6c008 | https://github.com/SeungoneKim/Transformer_implementation/tree/a52bf552eb645fc9bfb812cc26842fc147d6c008 | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super().__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super().__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
class FeedForward(nn.Module):
def __init__(self, model_dim, hidden_dim, drop_prob):
super().__init__()
self.model_dim = model_dim
self.hidden_dim = hidden_dim
self.drop_prob = drop_prob
self.linearlayer1 = nn.Linear(model_dim, hidden_dim)
self.linearlayer2 = nn.Linear(hidden_dim, model_dim)
self.relu = nn.ReLU()
self.dropout = nn.Dropout(drop_prob)
def forward(self, tensor):
tensor = self.dropout(self.relu(self.linearlayer1(tensor)))
return self.linearlayer2(tensor)
class Model(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, hidden_dim, num_head,
drop_prob):
super().__init__()
self.self_attention = MultiHeadAttention(model_dim, key_dim,
value_dim, num_head)
self.normalization1 = nn.LayerNorm(model_dim)
self.dropout1 = nn.Dropout(drop_prob)
self.enc_dec_attention = MultiHeadAttention(model_dim, key_dim,
value_dim, num_head)
self.normalization2 = nn.LayerNorm(model_dim)
self.dropout2 = nn.Dropout(drop_prob)
self.ffn = FeedForward(model_dim, hidden_dim, drop_prob)
self.normalization3 = nn.LayerNorm(model_dim)
self.dropout3 = nn.Dropout(drop_prob)
def forward(self, dec_tensor, enc_tensor, source_mask, target_mask):
residual = dec_tensor
tensor = self.self_attention(query=dec_tensor, key=dec_te
# ... truncated (>4000 chars) for memory efficiency |
TransitionUp | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/re/creleqpucbzvzrso3whbekvyjzfafblr33ygekztpuugjz5zfqbd.py
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# out_2 => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%slice_4, %primals_4], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 16) % 8
x0 = xindex % 4
x1 = (xindex // 4) % 4
x3 = (xindex // 128)
x4 = xindex % 16
x5 = xindex
tmp0 = x2
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (20 + x0 + (9*x1) + (81*x2) + (324*x3)), tmp4 & xmask, other=0.0)
tmp6 = tl.load(in_ptr1 + (x2), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tmp11 = tl.full([1], 8, tl.int64)
tmp12 = tmp0 < tmp11
tmp13 = tl.load(in_ptr2 + (x4 + (16*((-4) + x2)) + (64*x3)), tmp10 & xmask, other=0.0)
tmp14 = tl.where(tmp4, tmp9, tmp13)
tl.store(out_ptr0 + (x5), tmp14, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(2, 2), padding=(0, 0), dilation=(1, 1), transposed=True, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 9, 9), (324, 81, 9, 1))
buf1 = empty_strided_cuda((4, 8, 4, 4), (128, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(buf0, primals_2, primals_4, buf1, 512, grid=grid(512), stream=stream0)
del buf0
del primals_2
del primals_4
return (buf1, primals_1, primals_3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
def center_crop(layer, max_height, max_width):
_, _, h, w = layer.size()
xy1 = (w - max_width) // 2
xy2 = (h - max_height) // 2
return layer[:, :, xy2:xy2 + max_height, xy1:xy1 + max_width]
class TransitionUp(nn.Module):
def __init__(self, in_channels, out_channels):
super().__init__()
self.convTrans = nn.ConvTranspose2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=3, stride=2, padding=0,
bias=True)
def forward(self, x, skip):
out = self.convTrans(x)
out = center_crop(out, skip.size(2), skip.size(3))
out = torch.cat([out, skip], 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 16 % 8
x0 = xindex % 4
x1 = xindex // 4 % 4
x3 = xindex // 128
x4 = xindex % 16
x5 = xindex
tmp0 = x2
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (20 + x0 + 9 * x1 + 81 * x2 + 324 * x3), tmp4 &
xmask, other=0.0)
tmp6 = tl.load(in_ptr1 + x2, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp13 = tl.load(in_ptr2 + (x4 + 16 * (-4 + x2) + 64 * x3), tmp10 &
xmask, other=0.0)
tmp14 = tl.where(tmp4, tmp9, tmp13)
tl.store(out_ptr0 + x5, tmp14, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(2,
2), padding=(0, 0), dilation=(1, 1), transposed=True,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 9, 9), (324, 81, 9, 1))
buf1 = empty_strided_cuda((4, 8, 4, 4), (128, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(512)](buf0, primals_2, primals_4, buf1,
512, XBLOCK=256, num_warps=4, num_stages=1)
del buf0
del primals_2
del primals_4
return buf1, primals_1, primals_3
def center_crop(layer, max_height, max_width):
_, _, h, w = layer.size()
xy1 = (w - max_width) // 2
xy2 = (h - max_height) // 2
return layer[:, :, xy2:xy2 + max_height, xy1:xy1 + max_width]
class TransitionUpNew(nn.Module):
def __init__(self, in_channels, out_channels):
super().__init__()
self.convTrans = nn.ConvTranspose2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=3, stride=2, padding=0,
bias=True)
def forward(self, input_0, input_1):
primals_1 = self.convTrans.weight
primals_2 = self.convTrans.bias
primals_3 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| Sreehari-S/Tiramisu_DigestPath | TransitionUp | false | 1,083 | [
"Apache-2.0"
] | 0 | a884ee911bc60ce997996e0ec2e6036600ffcffa | https://github.com/Sreehari-S/Tiramisu_DigestPath/tree/a884ee911bc60ce997996e0ec2e6036600ffcffa | import torch
import torch.nn as nn
def center_crop(layer, max_height, max_width):
_, _, h, w = layer.size()
xy1 = (w - max_width) // 2
xy2 = (h - max_height) // 2
return layer[:, :, xy2:xy2 + max_height, xy1:xy1 + max_width]
class Model(nn.Module):
def __init__(self, in_channels, out_channels):
super().__init__()
self.convTrans = nn.ConvTranspose2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=3, stride=2, padding=0,
bias=True)
def forward(self, x, skip):
out = self.convTrans(x)
out = center_crop(out, skip.size(2), skip.size(3))
out = torch.cat([out, skip], 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ConvBlock | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/of/cofwklo5nuhcnwyox2iqrwvifqkquaz5nm5gngbpzzb6xu4usmqd.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, 0.23570226039551584), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.23570226039551584
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/32/c32v7egt4mupqssam3gmac2qgv3ujprjybthsgweflmot256qqw7.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=4] = call_function[target=torch.ops.aten.convolution.default](args = (%mul, %primals_2, %primals_3, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/4b/c4biasj4wzw4qmo6u767lcafgh6e6kct7c433onys3pjjjwfxxqj.py
# Topologically Sorted Source Nodes: [x, pow_1, mean, add, sqrt, x_1, mul_1], Original ATen: [aten.leaky_relu, aten.pow, aten.mean, aten.add, aten.sqrt, aten.div, aten.mul]
# Source node to ATen node mapping:
# add => add
# mean => mean
# mul_1 => mul_2
# pow_1 => pow_1
# sqrt => sqrt
# x => gt, mul_1, where
# x_1 => div
# Graph fragment:
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 0.2), kwargs = {})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%gt, %convolution, %mul_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%where, 2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_1, [1], True), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, 1e-08), kwargs = {})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%add,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%where, %sqrt), kwargs = {})
# %mul_2 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, 0.23570226039551584), kwargs = {})
triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2 = async_compile.triton('triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp6 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 0.2
tmp4 = tmp0 * tmp3
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp7 = tmp6 > tmp1
tmp8 = tmp6 * tmp3
tmp9 = tl.where(tmp7, tmp6, tmp8)
tmp10 = tmp9 * tmp9
tmp12 = tmp11 > tmp1
tmp13 = tmp11 * tmp3
tmp14 = tl.where(tmp12, tmp11, tmp13)
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp18 = tmp17 > tmp1
tmp19 = tmp17 * tmp3
tmp20 = tl.where(tmp18, tmp17, tmp19)
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp24 = tmp23 > tmp1
tmp25 = tmp23 * tmp3
tmp26 = tl.where(tmp24, tmp23, tmp25)
tmp27 = tmp26 * tmp26
tmp28 = tmp22 + tmp27
tmp29 = 4.0
tmp30 = tmp28 / tmp29
tmp31 = 1e-08
tmp32 = tmp30 + tmp31
tmp33 = libdevice.sqrt(tmp32)
tmp34 = tmp5 / tmp33
tmp35 = 0.23570226039551584
tmp36 = tmp34 * tmp35
tl.store(in_out_ptr0 + (x3), tmp36, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/7v/c7vm65ktpe4n23v24el4vlbfc7rafxngxnpi7ekdsaph6dnzyef4.py
# Topologically Sorted Source Nodes: [x_2, pow_2, mean_1, add_1, sqrt_1, x_3], Original ATen: [aten.leaky_relu, aten.pow, aten.mean, aten.add, aten.sqrt, aten.div]
# Source node to ATen node mapping:
# add_1 => add_1
# mean_1 => mean_1
# pow_2 => pow_2
# sqrt_1 => sqrt_1
# x_2 => gt_1, mul_3, where_1
# x_3 => div_1
# Graph fragment:
# %gt_1 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution_1, 0), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution_1, 0.2), kwargs = {})
# %where_1 : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%gt_1, %convolution_1, %mul_3), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%where_1, 2), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_2, [1], True), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean_1, 1e-08), kwargs = {})
# %sqrt_1 : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%add_1,), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%where_1, %sqrt_1), kwargs = {})
triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3 = async_compile.triton('triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp6 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 0.2
tmp4 = tmp0 * tmp3
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp7 = tmp6 > tmp1
tmp8 = tmp6 * tmp3
tmp9 = tl.where(tmp7, tmp6, tmp8)
tmp10 = tmp9 * tmp9
tmp12 = tmp11 > tmp1
tmp13 = tmp11 * tmp3
tmp14 = tl.where(tmp12, tmp11, tmp13)
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp18 = tmp17 > tmp1
tmp19 = tmp17 * tmp3
tmp20 = tl.where(tmp18, tmp17, tmp19)
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp24 = tmp23 > tmp1
tmp25 = tmp23 * tmp3
tmp26 = tl.where(tmp24, tmp23, tmp25)
tmp27 = tmp26 * tmp26
tmp28 = tmp22 + tmp27
tmp29 = 4.0
tmp30 = tmp28 / tmp29
tmp31 = 1e-08
tmp32 = tmp30 + tmp31
tmp33 = libdevice.sqrt(tmp32)
tmp34 = tmp5 / tmp33
tl.store(out_ptr0 + (x3), tmp34, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf2, primals_3, 256, grid=grid(256), stream=stream0)
del primals_3
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [x, pow_1, mean, add, sqrt, x_1, mul_1], Original ATen: [aten.leaky_relu, aten.pow, aten.mean, aten.add, aten.sqrt, aten.div, aten.mul]
triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2.run(buf4, buf2, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf5 = extern_kernels.convolution(buf4, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 4, 4), (64, 16, 4, 1))
buf6 = buf5; del buf5 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf6, primals_5, 256, grid=grid(256), stream=stream0)
del primals_5
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2, pow_2, mean_1, add_1, sqrt_1, x_3], Original ATen: [aten.leaky_relu, aten.pow, aten.mean, aten.add, aten.sqrt, aten.div]
triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
return (buf7, primals_2, primals_4, buf0, buf2, buf4, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
class WSConv2d(nn.Module):
"""
Weight scaled Conv2d (Equalized Learning Rate)
Note that input is multiplied rather than changing weights
this will have the same result.
Inspired by:
https://github.com/nvnbny/progressive_growing_of_gans/blob/master/modelUtils.py
"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1,
padding=1, gain=2):
super(WSConv2d, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding)
self.scale = (gain / self.conv.weight[0].numel()) ** 0.5
nn.init.normal_(self.conv.weight)
nn.init.zeros_(self.conv.bias)
def forward(self, x):
return self.conv(x * self.scale)
class PixelNorm(nn.Module):
def __init__(self):
super(PixelNorm, self).__init__()
self.epsilon = 1e-08
def forward(self, x):
return x / torch.sqrt(torch.mean(x ** 2, dim=1, keepdim=True) +
self.epsilon)
class ConvBlock(nn.Module):
def __init__(self, in_channels, out_channels, use_pixelnorm=True):
super(ConvBlock, self).__init__()
self.use_pn = use_pixelnorm
self.conv1 = WSConv2d(in_channels, out_channels)
self.conv2 = WSConv2d(out_channels, out_channels)
self.leaky = nn.LeakyReLU(0.2)
self.pn = PixelNorm()
def forward(self, x):
x = self.leaky(self.conv1(x))
x = self.pn(x) if self.use_pn else x
x = self.leaky(self.conv2(x))
x = self.pn(x) if self.use_pn else x
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.23570226039551584
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2(in_out_ptr0,
in_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp6 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp11 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp17 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp23 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 0.2
tmp4 = tmp0 * tmp3
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp7 = tmp6 > tmp1
tmp8 = tmp6 * tmp3
tmp9 = tl.where(tmp7, tmp6, tmp8)
tmp10 = tmp9 * tmp9
tmp12 = tmp11 > tmp1
tmp13 = tmp11 * tmp3
tmp14 = tl.where(tmp12, tmp11, tmp13)
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp18 = tmp17 > tmp1
tmp19 = tmp17 * tmp3
tmp20 = tl.where(tmp18, tmp17, tmp19)
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp24 = tmp23 > tmp1
tmp25 = tmp23 * tmp3
tmp26 = tl.where(tmp24, tmp23, tmp25)
tmp27 = tmp26 * tmp26
tmp28 = tmp22 + tmp27
tmp29 = 4.0
tmp30 = tmp28 / tmp29
tmp31 = 1e-08
tmp32 = tmp30 + tmp31
tmp33 = libdevice.sqrt(tmp32)
tmp34 = tmp5 / tmp33
tmp35 = 0.23570226039551584
tmp36 = tmp34 * tmp35
tl.store(in_out_ptr0 + x3, tmp36, xmask)
@triton.jit
def triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3(in_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp6 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp11 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp17 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp23 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 0.2
tmp4 = tmp0 * tmp3
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp7 = tmp6 > tmp1
tmp8 = tmp6 * tmp3
tmp9 = tl.where(tmp7, tmp6, tmp8)
tmp10 = tmp9 * tmp9
tmp12 = tmp11 > tmp1
tmp13 = tmp11 * tmp3
tmp14 = tl.where(tmp12, tmp11, tmp13)
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp18 = tmp17 > tmp1
tmp19 = tmp17 * tmp3
tmp20 = tl.where(tmp18, tmp17, tmp19)
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp24 = tmp23 > tmp1
tmp25 = tmp23 * tmp3
tmp26 = tl.where(tmp24, tmp23, tmp25)
tmp27 = tmp26 * tmp26
tmp28 = tmp22 + tmp27
tmp29 = 4.0
tmp30 = tmp28 / tmp29
tmp31 = 1e-08
tmp32 = tmp30 + tmp31
tmp33 = libdevice.sqrt(tmp32)
tmp34 = tmp5 / tmp33
tl.store(out_ptr0 + x3, tmp34, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(256)](primals_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_1[grid(256)](buf2, primals_3, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf4 = buf3
del buf3
triton_poi_fused_add_div_leaky_relu_mean_mul_pow_sqrt_2[grid(256)](buf4
, buf2, 256, XBLOCK=256, num_warps=4, num_stages=1)
buf5 = extern_kernels.convolution(buf4, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf5, (4, 4, 4, 4), (64, 16, 4, 1))
buf6 = buf5
del buf5
triton_poi_fused_convolution_1[grid(256)](buf6, primals_5, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_div_leaky_relu_mean_pow_sqrt_3[grid(256)](buf6,
buf7, 256, XBLOCK=128, num_warps=4, num_stages=1)
return buf7, primals_2, primals_4, buf0, buf2, buf4, buf6
class WSConv2d(nn.Module):
"""
Weight scaled Conv2d (Equalized Learning Rate)
Note that input is multiplied rather than changing weights
this will have the same result.
Inspired by:
https://github.com/nvnbny/progressive_growing_of_gans/blob/master/modelUtils.py
"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1,
padding=1, gain=2):
super(WSConv2d, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding)
self.scale = (gain / self.conv.weight[0].numel()) ** 0.5
nn.init.normal_(self.conv.weight)
nn.init.zeros_(self.conv.bias)
def forward(self, x):
return self.conv(x * self.scale)
class PixelNorm(nn.Module):
def __init__(self):
super(PixelNorm, self).__init__()
self.epsilon = 1e-08
def forward(self, x):
return x / torch.sqrt(torch.mean(x ** 2, dim=1, keepdim=True) +
self.epsilon)
class ConvBlockNew(nn.Module):
def __init__(self, in_channels, out_channels, use_pixelnorm=True):
super(ConvBlockNew, self).__init__()
self.use_pn = use_pixelnorm
self.conv1 = WSConv2d(in_channels, out_channels)
self.conv2 = WSConv2d(out_channels, out_channels)
self.leaky = nn.LeakyReLU(0.2)
self.pn = PixelNorm()
def forward(self, input_0):
primals_2 = self.conv1.conv.weight
primals_3 = self.conv1.conv.bias
primals_4 = self.conv2.conv.weight
primals_5 = self.conv2.conv.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| SongsLearning/Machine-Learning-Collection | ConvBlock | false | 1,084 | [
"MIT"
] | 0 | a8dff83969f67d37f70a89db06b851057d2da539 | https://github.com/SongsLearning/Machine-Learning-Collection/tree/a8dff83969f67d37f70a89db06b851057d2da539 | import torch
import torch.nn as nn
import torch.utils.data
class WSConv2d(nn.Module):
"""
Weight scaled Conv2d (Equalized Learning Rate)
Note that input is multiplied rather than changing weights
this will have the same result.
Inspired by:
https://github.com/nvnbny/progressive_growing_of_gans/blob/master/modelUtils.py
"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1,
padding=1, gain=2):
super().__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding)
self.scale = (gain / self.conv.weight[0].numel()) ** 0.5
nn.init.normal_(self.conv.weight)
nn.init.zeros_(self.conv.bias)
def forward(self, x):
return self.conv(x * self.scale)
class PixelNorm(nn.Module):
def __init__(self):
super().__init__()
self.epsilon = 1e-08
def forward(self, x):
return x / torch.sqrt(torch.mean(x ** 2, dim=1, keepdim=True) +
self.epsilon)
class Model(nn.Module):
def __init__(self, in_channels, out_channels, use_pixelnorm=True):
super().__init__()
self.use_pn = use_pixelnorm
self.conv1 = WSConv2d(in_channels, out_channels)
self.conv2 = WSConv2d(out_channels, out_channels)
self.leaky = nn.LeakyReLU(0.2)
self.pn = PixelNorm()
def forward(self, x):
x = self.leaky(self.conv1(x))
x = self.pn(x) if self.use_pn else x
x = self.leaky(self.conv2(x))
x = self.pn(x) if self.use_pn else x
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
FcCat | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/ie/ciettq2a3562jfpgfe75iig4ki2hbm6pmbwujlvp6mw26i2odufm.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# out => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_2, %view_1], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 16) % 8
x0 = xindex % 16
x2 = (xindex // 128)
x3 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (16*x1) + (64*x2)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr1 + (x0 + (16*((-4) + x1)) + (64*x2)), tmp6 & xmask, other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + (x3), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 8, 4, 4), (128, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_2, buf0, buf1, 512, grid=grid(512), stream=stream0)
del buf0
return (buf1, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class FcCat(nn.Module):
def __init__(self, nIn, nOut):
super(FcCat, self).__init__()
self.fc = nn.Linear(nIn, nOut, bias=False)
def forward(self, x):
out = torch.cat((x, self.fc(x)), 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'nIn': 4, 'nOut': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 16 % 8
x0 = xindex % 16
x2 = xindex // 128
x3 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 16 * x1 + 64 * x2), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp9 = tl.load(in_ptr1 + (x0 + 16 * (-4 + x1) + 64 * x2), tmp6 & xmask,
other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tl.store(out_ptr0 + x3, tmp10, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 8, 4, 4), (128, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(512)](primals_2, buf0, buf1, 512,
XBLOCK=256, num_warps=4, num_stages=1)
del buf0
return buf1, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0)
class FcCatNew(nn.Module):
def __init__(self, nIn, nOut):
super(FcCatNew, self).__init__()
self.fc = nn.Linear(nIn, nOut, bias=False)
def forward(self, input_0):
primals_1 = self.fc.weight
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| Sreehari-S/Tiramisu_DigestPath | FcCat | false | 1,086 | [
"Apache-2.0"
] | 0 | a884ee911bc60ce997996e0ec2e6036600ffcffa | https://github.com/Sreehari-S/Tiramisu_DigestPath/tree/a884ee911bc60ce997996e0ec2e6036600ffcffa | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, nIn, nOut):
super().__init__()
self.fc = nn.Linear(nIn, nOut, bias=False)
def forward(self, x):
out = torch.cat((x, self.fc(x)), 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
WSConv2d | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/of/cofwklo5nuhcnwyox2iqrwvifqkquaz5nm5gngbpzzb6xu4usmqd.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, 0.23570226039551584), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.23570226039551584
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/32/c32v7egt4mupqssam3gmac2qgv3ujprjybthsgweflmot256qqw7.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%mul, %primals_2, %primals_3, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf2, primals_3, 256, grid=grid(256), stream=stream0)
del primals_3
return (buf2, primals_2, buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
class WSConv2d(nn.Module):
"""
Weight scaled Conv2d (Equalized Learning Rate)
Note that input is multiplied rather than changing weights
this will have the same result.
Inspired by:
https://github.com/nvnbny/progressive_growing_of_gans/blob/master/modelUtils.py
"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1,
padding=1, gain=2):
super(WSConv2d, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding)
self.scale = (gain / self.conv.weight[0].numel()) ** 0.5
nn.init.normal_(self.conv.weight)
nn.init.zeros_(self.conv.bias)
def forward(self, x):
return self.conv(x * self.scale)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.23570226039551584
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(256)](primals_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4, 4), (64, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_1[grid(256)](buf2, primals_3, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
return buf2, primals_2, buf0
class WSConv2dNew(nn.Module):
"""
Weight scaled Conv2d (Equalized Learning Rate)
Note that input is multiplied rather than changing weights
this will have the same result.
Inspired by:
https://github.com/nvnbny/progressive_growing_of_gans/blob/master/modelUtils.py
"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1,
padding=1, gain=2):
super(WSConv2dNew, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding)
self.scale = (gain / self.conv.weight[0].numel()) ** 0.5
nn.init.normal_(self.conv.weight)
nn.init.zeros_(self.conv.bias)
def forward(self, input_0):
primals_2 = self.conv.weight
primals_3 = self.conv.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| SongsLearning/Machine-Learning-Collection | WSConv2d | false | 1,087 | [
"MIT"
] | 0 | a8dff83969f67d37f70a89db06b851057d2da539 | https://github.com/SongsLearning/Machine-Learning-Collection/tree/a8dff83969f67d37f70a89db06b851057d2da539 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
"""
Weight scaled Conv2d (Equalized Learning Rate)
Note that input is multiplied rather than changing weights
this will have the same result.
Inspired by:
https://github.com/nvnbny/progressive_growing_of_gans/blob/master/modelUtils.py
"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1,
padding=1, gain=2):
super().__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding)
self.scale = (gain / self.conv.weight[0].numel()) ** 0.5
nn.init.normal_(self.conv.weight)
nn.init.zeros_(self.conv.bias)
def forward(self, x):
return self.conv(x * self.scale)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Standardscaler | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/ft/cft4ssfca465zguug7muc7b6b3q6agysmujv6njryw7rnc62phca.py
# Topologically Sorted Source Nodes: [std_mean, sub, total], Original ATen: [aten.std_mean, aten.sub, aten.div]
# Source node to ATen node mapping:
# std_mean => sqrt, var_mean
# sub => sub
# total => div
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%arg0_1,), kwargs = {correction: 0.0})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %getitem_1), kwargs = {})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%getitem,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %sqrt), kwargs = {})
triton_per_fused_div_std_mean_sub_0 = async_compile.triton('triton_per_fused_div_std_mean_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_std_mean_sub_0', 'mutated_arg_names': [], 'no_x_dim': True, 'num_load': 1, 'num_reduction': 4, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_div_std_mean_sub_0(in_ptr0, out_ptr2, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.broadcast_to(tmp0, [RBLOCK])
tmp3 = tl.broadcast_to(tmp1, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(tl.sum(tmp3, 0))
tmp6 = tl.full([1], 256, tl.int32)
tmp7 = tmp6.to(tl.float32)
tmp8 = tmp5 / tmp7
tmp9 = tmp1 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tl.broadcast_to(tmp10, [RBLOCK])
tmp13 = triton_helpers.promote_to_tensor(tl.sum(tmp11, 0))
tmp14 = tmp0 - tmp8
tmp15 = 256.0
tmp16 = tmp13 / tmp15
tmp17 = libdevice.sqrt(tmp16)
tmp18 = tmp14 / tmp17
tl.store(out_ptr2 + (tl.broadcast_to(r0, [RBLOCK])), tmp18, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [std_mean, sub, total], Original ATen: [aten.std_mean, aten.sub, aten.div]
stream0 = get_raw_stream(0)
triton_per_fused_div_std_mean_sub_0.run(arg0_1, buf3, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
return (buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
class Standardscaler(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, input_batch):
std, mean = torch.std_mean(input_batch.type(torch.float32),
unbiased=False)
total = (input_batch - mean) / std
return total
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_div_std_mean_sub_0(in_ptr0, out_ptr2, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.broadcast_to(tmp0, [RBLOCK])
tmp3 = tl.broadcast_to(tmp1, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(tl.sum(tmp3, 0))
tmp6 = tl.full([1], 256, tl.int32)
tmp7 = tmp6.to(tl.float32)
tmp8 = tmp5 / tmp7
tmp9 = tmp1 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tl.broadcast_to(tmp10, [RBLOCK])
tmp13 = triton_helpers.promote_to_tensor(tl.sum(tmp11, 0))
tmp14 = tmp0 - tmp8
tmp15 = 256.0
tmp16 = tmp13 / tmp15
tmp17 = libdevice.sqrt(tmp16)
tmp18 = tmp14 / tmp17
tl.store(out_ptr2 + tl.broadcast_to(r0, [RBLOCK]), tmp18, None)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused_div_std_mean_sub_0[grid(1)](arg0_1, buf3, 1, 256,
num_warps=2, num_stages=1)
del arg0_1
return buf3,
class StandardscalerNew(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Stuksus/StandardScaler_for_pytorch | Standardscaler | false | 1,088 | [
"MIT"
] | 0 | 27da9afd111007f20a615bee9a5a7ac272adb241 | https://github.com/Stuksus/StandardScaler_for_pytorch/tree/27da9afd111007f20a615bee9a5a7ac272adb241 | import torch
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, input_batch):
std, mean = torch.std_mean(input_batch.type(torch.float32),
unbiased=False)
total = (input_batch - mean) / std
return total
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
FeatureResizer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/66/c666qvr725wwogti7syalhhjsndtv2n5sxzt6zi4wlesyjxocpic.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# x_1 => add, rsqrt, var_mean
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_1, [3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-12), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
triton_poi_fused_native_layer_norm_0 = async_compile.triton('triton_poi_fused_native_layer_norm_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-12
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + (x0), tmp8, xmask)
tl.store(out_ptr1 + (x0), tmp23, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/lh/clhh73owbiuj4adasmetdqsot2nlmw2ljupnw2q4yt3du76mikww.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# x_1 => add, add_1, mul, mul_1, rsqrt, sub, var_mean
# Graph fragment:
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_1, [3]), kwargs = {correction: 0, keepdim: True})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-12), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_1, %getitem_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %rsqrt), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %primals_4), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_5), kwargs = {})
triton_poi_fused_native_layer_norm_1 = async_compile.triton('triton_poi_fused_native_layer_norm_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
stream0 = get_raw_stream(0)
triton_poi_fused_native_layer_norm_0.run(buf0, buf1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_1.run(buf0, buf1, buf2, primals_4, primals_5, buf3, 256, grid=grid(256), stream=stream0)
del buf1
del buf2
del primals_5
return (buf3, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.utils.data
import torch
from torch import nn
class FeatureResizer(nn.Module):
"""
This class takes as input a set of embeddings of dimension C1 and outputs a set of
embedding of dimension C2, after a linear transformation, dropout and normalization (LN).
"""
def __init__(self, input_feat_size, output_feat_size, dropout, do_ln=True):
super().__init__()
self.do_ln = do_ln
self.fc = nn.Linear(input_feat_size, output_feat_size, bias=True)
self.layer_norm = nn.LayerNorm(output_feat_size, eps=1e-12)
self.dropout = nn.Dropout(dropout)
def forward(self, encoder_features):
x = self.fc(encoder_features)
if self.do_ln:
x = self.layer_norm(x)
output = self.dropout(x)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_feat_size': 4, 'output_feat_size': 4, 'dropout': 0.5}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.utils.data
import torch
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_native_layer_norm_0(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-12
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + x0, tmp8, xmask)
tl.store(out_ptr1 + x0, tmp23, xmask)
@triton.jit
def triton_poi_fused_native_layer_norm_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
get_raw_stream(0)
triton_poi_fused_native_layer_norm_0[grid(64)](buf0, buf1, buf2, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_native_layer_norm_1[grid(256)](buf0, buf1, buf2,
primals_4, primals_5, buf3, 256, XBLOCK=128, num_warps=4,
num_stages=1)
del buf1
del buf2
del primals_5
return buf3, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf0
class FeatureResizerNew(nn.Module):
"""
This class takes as input a set of embeddings of dimension C1 and outputs a set of
embedding of dimension C2, after a linear transformation, dropout and normalization (LN).
"""
def __init__(self, input_feat_size, output_feat_size, dropout, do_ln=True):
super().__init__()
self.do_ln = do_ln
self.fc = nn.Linear(input_feat_size, output_feat_size, bias=True)
self.layer_norm = nn.LayerNorm(output_feat_size, eps=1e-12)
self.dropout = nn.Dropout(dropout)
def forward(self, input_0):
primals_1 = self.fc.weight
primals_2 = self.fc.bias
primals_4 = self.layer_norm.weight
primals_5 = self.layer_norm.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| Sudhir11292rt/DefVisTR | FeatureResizer | false | 1,089 | [
"Apache-2.0"
] | 0 | d52b2d88c10c6239de1c1ff851a743c58b708b75 | https://github.com/Sudhir11292rt/DefVisTR/tree/d52b2d88c10c6239de1c1ff851a743c58b708b75 | import torch
import torch.utils.data
import torch
from torch import nn
class Model(nn.Module):
"""
This class takes as input a set of embeddings of dimension C1 and outputs a set of
embedding of dimension C2, after a linear transformation, dropout and normalization (LN).
"""
def __init__(self, input_feat_size, output_feat_size, dropout, do_ln=True):
super().__init__()
self.do_ln = do_ln
self.fc = nn.Linear(input_feat_size, output_feat_size, bias=True)
self.layer_norm = nn.LayerNorm(output_feat_size, eps=1e-12)
self.dropout = nn.Dropout(dropout)
def forward(self, encoder_features):
x = self.fc(encoder_features)
if self.do_ln:
x = self.layer_norm(x)
output = self.dropout(x)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 0.5]
|
UGRNNLRCell | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/id/cidxuzakshtvnrldbbqkyga2amymtgyw2cvpqv7d2kjchobiepep.py
# Topologically Sorted Source Nodes: [pre_comp1, pre_comp2, add_2, z, add_3, c, mul, sub, mul_1, new_h], Original ATen: [aten.add, aten.sigmoid, aten.tanh, aten.mul, aten.rsub]
# Source node to ATen node mapping:
# add_2 => add_2
# add_3 => add_3
# c => tanh
# mul => mul
# mul_1 => mul_1
# new_h => add_4
# pre_comp1 => add
# pre_comp2 => add_1
# sub => sub
# z => sigmoid
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %view_5), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, %view_7), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %primals_7), kwargs = {})
# %sigmoid : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_2,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, %primals_8), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add_3,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_5), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %sigmoid), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %tanh), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
triton_poi_fused_add_mul_rsub_sigmoid_tanh_0 = async_compile.triton('triton_poi_fused_add_mul_rsub_sigmoid_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_rsub_sigmoid_tanh_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 7, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_tanh_0(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_out_ptr1 + (x2), xmask)
tmp7 = tl.load(in_ptr2 + (x2), xmask)
tmp9 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr4 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = libdevice.tanh(tmp10)
tmp13 = tmp5 * tmp12
tmp14 = 1.0
tmp15 = tmp14 - tmp5
tmp16 = tmp15 * tmp11
tmp17 = tmp13 + tmp16
tl.store(in_out_ptr0 + (x2), tmp5, xmask)
tl.store(in_out_ptr1 + (x2), tmp11, xmask)
tl.store(out_ptr0 + (x2), tmp17, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (1, 4), (4, 1))
assert_size_stride(primals_8, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [wComp2], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_3, out=buf1)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [uComp1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_5, (64, 4), (4, 1), 0), primals_4, out=buf2)
del primals_4
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [uComp2], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_5, (64, 4), (4, 1), 0), primals_6, out=buf3)
del primals_6
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf1 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pre_comp1, pre_comp2, add_2, z, add_3, c, mul, sub, mul_1, new_h], Original ATen: [aten.add, aten.sigmoid, aten.tanh, aten.mul, aten.rsub]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_rsub_sigmoid_tanh_0.run(buf4, buf5, buf2, primals_7, buf3, primals_8, primals_5, buf6, 256, grid=grid(256), stream=stream0)
del buf2
del buf3
del primals_7
del primals_8
return (buf6, primals_5, buf4, buf5, reinterpret_tensor(primals_2, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class UGRNNLRCell(RNNCell):
"""
UGRNN LR Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
gate_nonlinearity = nonlinearity for the gate can be chosen from
[tanh, sigmoid, relu, quantTanh, quantSigm]
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix
(creates 3 matrices if not None else creates 2 matrices)
uRank = rank of U matrix
(creates 3 matrices if not None else creates 2 matrices)
UGRNN architecture and compression techniques are found in
UGRNN(LINK) paper
Basic architecture is like:
z_t = gate_nl(W1x_t + U1h_{t-1} + B_g)
h_t^ = update_nl(W1x_t + U1h_{t-1} + B_h)
h_t = z_t*h_{t-1} + (1-z_t)*h_t^
Wi and Ui can further parameterised into low rank version by
Wi = matmul(W, W_i) and Ui = matmul(U, U_i)
"""
def __init__(self, input_size, hidden_size, gate_nonlinearity='sigmoid',
update_nonlinearity='tanh', wRank=None, uRank=None, wSparsity=1.0,
uSparsity=1.0, name='UGRNNLR'):
super(UGRNNLRCell, self).__init__(input_size, hidden_size,
gate_nonlinearity, update_nonlinearity, 2, 2, 2, wRank, uRank,
wSparsity, uSparsity)
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
self.W2 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
else:
self.W = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W1 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
else:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U1 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_gate = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self._device = self.bias_update.device
@property
def name(self):
return self._name
@property
def cellType(self):
return 'UGRNNLR'
def forward(self, input, state):
if self._wRank is None:
wComp1 = torch.matmul(input, self.W1)
wComp2 = torch.matmul(input, self.W2)
else:
wComp1 = torch.matmul(torch.matmul(input, self.W), self.W1)
wComp2 = torch.matmul(torch.matmul(input, self.W), self.W2)
if self._uRank is None:
uComp1 = torch.matmul(state, self.U1)
uComp2 = torch.matmul(state, self.U2)
else:
uComp1 = torch.matmul(torch.matmul(state, self.U), self.U1)
uComp2 = torch.matmul(torch.matmul(state, self.U), self.U2)
pre_comp1 = wComp1 + uComp1
pre_comp2 = wComp2 + uComp2
z = gen_nonlinearity(pre_comp1 + self.bias_gate, self.
_gate_nonlinearity)
c = gen_nonlinearity(pre_comp2 + self.bias_update, self.
_update_nonlinearity)
new_h = z * state + (1.0 - z) * c
return new_h
def getVars(self):
Vars = []
if self._num_W_matrices == 2:
Vars.extend([self.W1, self.W2])
else:
Vars.extend([self.W, self.W1, self.W2])
if self._num_U_matrices == 2:
Vars.extend([self.U1, self.U2])
else:
Vars.extend([self.U, self.U1, self.U2])
Vars.extend([self.bias_gate, self.bias_update])
return Vars
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.onnx
from itertools import product as product
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_tanh_0(in_out_ptr0, in_out_ptr1,
in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_out_ptr1 + x2, xmask)
tmp7 = tl.load(in_ptr2 + x2, xmask)
tmp9 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr4 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.sigmoid(tmp4)
tmp8 = tmp6 + tmp7
tmp10 = tmp8 + tmp9
tmp11 = libdevice.tanh(tmp10)
tmp13 = tmp5 * tmp12
tmp14 = 1.0
tmp15 = tmp14 - tmp5
tmp16 = tmp15 * tmp11
tmp17 = tmp13 + tmp16
tl.store(in_out_ptr0 + x2, tmp5, xmask)
tl.store(in_out_ptr1 + x2, tmp11, xmask)
tl.store(out_ptr0 + x2, tmp17, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8) = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (1, 4), (4, 1))
assert_size_stride(primals_8, (1, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_3, out=buf1)
del primals_3
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_5, (64, 4), (4, 1), 0),
primals_4, out=buf2)
del primals_4
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_5, (64, 4), (4, 1), 0),
primals_6, out=buf3)
del primals_6
buf4 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
buf5 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf1
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_rsub_sigmoid_tanh_0[grid(256)](buf4, buf5,
buf2, primals_7, buf3, primals_8, primals_5, buf6, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del buf2
del buf3
del primals_7
del primals_8
return buf6, primals_5, buf4, buf5, reinterpret_tensor(primals_2, (4,
64), (1, 4), 0)
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super(RNNCell, self).__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(torch.FloatTensor())
for i in range(num_mats):
self.oldmats[i] = torch.FloatTensor(mats[i].detach().clone())
def sparsify(self):
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
for i in range(0, endW):
mats[i] = utils.hardThreshold(mats[i], self._wSparsity)
for i in range(endW, endU):
mats[i] = utils.hardThreshold(mats[i], self._uSparsity)
self.copy_previous_UW()
def sparsifyWithSupport(self):
mats = self.getVars()
endU = self._num_W_matrices + self._num_U_matrices
for i in range(0, endU):
mats[i] = utils.supportBasedThreshold(mats[i], self.oldmats[i])
class UGRNNLRCellNew(RNNCell):
"""
UGRNN LR Cell with Both Full Rank and Low Rank Formulations
Has multiple activation functions for the gates
hidden_size = # hidden units
gate_nonlinearity = nonlinearity for the gate can be chosen from
[tanh, sigmoid, relu, quantTanh, quantSigm]
update_nonlinearity = nonlinearity for final rnn update
can be chosen from [tanh, sigmoid, relu, quantTanh, quantSigm]
wRank = rank of W matrix
(creates 3 matrices if not None else creates 2 matrices)
uRank = rank of U matrix
(creates 3 matrices if not None else creates 2 matrices)
UGRNN architecture and compression techniques are found in
UGRNN(LINK) paper
Basic architecture is like:
z_t = gate_nl(W1x_t + U1h_{t-1} + B_g)
h_t^ = update_nl(W1x_t + U1h_{t-1} + B_h)
h_t = z_t*h_{t-1} + (1-z_t)*h_t^
Wi and Ui can further parameterised into low rank version by
Wi = matmul(W, W_i) and Ui = matmul(U, U_i)
"""
def __init__(self, input_size, hidden_size, gate_nonlinearity='sigmoid',
update_nonlinearity='tanh', wRank=None, uRank=None, wSparsity=1.0,
uSparsity=1.0, name='UGRNNLR'):
super(UGRNNLRCellNew, self).__init__(input_size, hidden_size,
gate_nonlinearity, update_nonlinearity, 2, 2, 2, wRank, uRank,
wSparsity, uSparsity)
if wRank is not None:
self._num_W_matrices += 1
self._num_weight_matrices[0] = self._num_W_matrices
if uRank is not None:
self._num_U_matrices += 1
self._num_weight_matrices[1] = self._num_U_matrices
self._name = name
if wRank is None:
self.W1 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
self.W2 = nn.Parameter(0.1 * torch.randn([input_size, hidden_size])
)
else:
self.W = nn.Parameter(0.1 * torch.randn([input_size, wRank]))
self.W1 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
self.W2 = nn.Parameter(0.1 * torch.randn([wRank, hidden_size]))
if uRank is None:
self.U1 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([hidden_size,
hidden_size]))
else:
self.U = nn.Parameter(0.1 * torch.randn([hidden_size, uRank]))
self.U1 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.U2 = nn.Parameter(0.1 * torch.randn([uRank, hidden_size]))
self.bias_gate = nn.Parameter(torch.ones([1, hidden_size]))
self.bias_update = nn.Parameter(torch.ones([1, hidden_size]))
self._device = self.bias_update.device
@property
def name(self):
return self._name
@property
def cellType(self):
return 'UGRNNLR'
def getVars(self):
Vars = []
if self._num_W_matrices == 2:
Vars.extend([self.W1, self.W2])
else:
Vars.extend([self.W, self.W1, self.W2])
if self._num_U_matrices == 2:
Vars.extend([self.U1, self.U2])
else:
Vars.extend([self.U, self.U1, self.U2])
Vars.extend([self.bias_gate, self.bias_update])
return Vars
def forward(self, input_0, input_1):
primals_1 = self.W1
primals_3 = self.W2
primals_4 = self.U1
primals_6 = self.U2
primals_7 = self.bias_gate
primals_8 = self.bias_update
primals_2 = input_0
primals_5 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8])
return output[0]
| ShishirPatil/EdgeML-1 | UGRNNLRCell | false | 1,090 | [
"MIT"
] | 0 | cbba9f8b989e545788427c004eb8450e7e4c1a21 | https://github.com/ShishirPatil/EdgeML-1/tree/cbba9f8b989e545788427c004eb8450e7e4c1a21 | import torch
import torch.nn as nn
import torch.onnx
from itertools import product as product
def gen_nonlinearity(A, nonlinearity):
"""
Returns required activation for a tensor based on the inputs
nonlinearity is either a callable or a value in
['tanh', 'sigmoid', 'relu', 'quantTanh', 'quantSigm', 'quantSigm4']
"""
if nonlinearity == 'tanh':
return torch.tanh(A)
elif nonlinearity == 'sigmoid':
return torch.sigmoid(A)
elif nonlinearity == 'relu':
return torch.relu(A, 0.0)
elif nonlinearity == 'quantTanh':
return torch.max(torch.min(A, torch.ones_like(A)), -1.0 * torch.
ones_like(A))
elif nonlinearity == 'quantSigm':
A = (A + 1.0) / 2.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
elif nonlinearity == 'quantSigm4':
A = (A + 2.0) / 4.0
return torch.max(torch.min(A, torch.ones_like(A)), torch.zeros_like(A))
else:
if not callable(nonlinearity):
raise ValueError(
'nonlinearity is either a callable or a value ' +
"['tanh', 'sigmoid', 'relu', 'quantTanh', " + "'quantSigm'")
return nonlinearity(A)
class RNNCell(nn.Module):
def __init__(self, input_size, hidden_size, gate_nonlinearity,
update_nonlinearity, num_W_matrices, num_U_matrices, num_biases,
wRank=None, uRank=None, wSparsity=1.0, uSparsity=1.0):
super().__init__()
self._input_size = input_size
self._hidden_size = hidden_size
self._gate_nonlinearity = gate_nonlinearity
self._update_nonlinearity = update_nonlinearity
self._num_W_matrices = num_W_matrices
self._num_U_matrices = num_U_matrices
self._num_biases = num_biases
self._num_weight_matrices = [self._num_W_matrices, self.
_num_U_matrices, self._num_biases]
self._wRank = wRank
self._uRank = uRank
self._wSparsity = wSparsity
self._uSparsity = uSparsity
self.oldmats = []
@property
def state_size(self):
return self._hidden_size
@property
def input_size(self):
return self._input_size
@property
def output_size(self):
return self._hidden_size
@property
def gate_nonlinearity(self):
return self._gate_nonlinearity
@property
def update_nonlinearity(self):
return self._update_nonlinearity
@property
def wRank(self):
return self._wRank
@property
def uRank(self):
return self._uRank
@property
def num_W_matrices(self):
return self._num_W_matrices
@property
def num_U_matrices(self):
return self._num_U_matrices
@property
def num_weight_matrices(self):
return self._num_weight_matrices
@property
def name(self):
raise NotImplementedError()
def forward(self, input, state):
raise NotImplementedError()
def getVars(self):
raise NotImplementedError()
def get_model_size(self):
"""
Function to get aimed model size
"""
mats = self.getVars()
endW = self._num_W_matrices
endU = endW + self._num_U_matrices
totalnnz = 2
for i in range(0, endW):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._wSparsity)
mats[i]
for i in range(endW, endU):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), self._uSparsity)
mats[i]
for i in range(endU, len(mats)):
mats[i].device
totalnnz += utils.countNNZ(mats[i].cpu(), False)
mats[i]
return totalnnz * 4
def copy_previous_UW(self):
mats = self.getVars()
num_mats = self._num_W_matrices + self._num_U_matrices
if len(self.oldmats) != num_mats:
for i in range(num_mats):
self.oldmats.append(to
# ... truncated (>4000 chars) for memory efficiency |
FCLayer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/dn/cdnhr6ixjduuhci57kobqjnehjrl22mcyjqzuuhvtxxshy437diy.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# x_1 => tanh
# Graph fragment:
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%primals_1,), kwargs = {})
triton_poi_fused_tanh_0 = async_compile.triton('triton_poi_fused_tanh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_tanh_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_tanh_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = libdevice.tanh(tmp0)
tl.store(out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_tanh_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(buf0, (64, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_2
del primals_3
return (reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf0, (64, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class FCLayer(nn.Module):
def __init__(self, input_dim, output_dim, dropout_rate=0.0,
use_activation=True):
super(FCLayer, self).__init__()
self.use_activation = use_activation
self.dropout = nn.Dropout(dropout_rate)
self.linear = nn.Linear(input_dim, output_dim)
self.tanh = nn.Tanh()
def forward(self, x):
x = self.dropout(x)
if self.use_activation:
x = self.tanh(x)
return self.linear(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'output_dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_tanh_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = libdevice.tanh(tmp0)
tl.store(out_ptr0 + x0, tmp1, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_tanh_0[grid(256)](primals_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(buf0, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf1)
del primals_2
del primals_3
return reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(buf0, (64, 4), (4, 1), 0)
class FCLayerNew(nn.Module):
def __init__(self, input_dim, output_dim, dropout_rate=0.0,
use_activation=True):
super(FCLayerNew, self).__init__()
self.use_activation = use_activation
self.dropout = nn.Dropout(dropout_rate)
self.linear = nn.Linear(input_dim, output_dim)
self.tanh = nn.Tanh()
def forward(self, input_0):
primals_2 = self.linear.weight
primals_3 = self.linear.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| StevenChaoo/R-BERT-DDI | FCLayer | false | 1,091 | [
"MIT"
] | 0 | 6d9666e0bc61397ca942ffad53653690c1e8a899 | https://github.com/StevenChaoo/R-BERT-DDI/tree/6d9666e0bc61397ca942ffad53653690c1e8a899 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, input_dim, output_dim, dropout_rate=0.0,
use_activation=True):
super().__init__()
self.use_activation = use_activation
self.dropout = nn.Dropout(dropout_rate)
self.linear = nn.Linear(input_dim, output_dim)
self.tanh = nn.Tanh()
def forward(self, x):
x = self.dropout(x)
if self.use_activation:
x = self.tanh(x)
return self.linear(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
MultiHeadAttention | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/cu/ccutvo2v4333pq6xhrg2zryqqwthm7dmmuqprvva2xdwiodpz5jn.py
# Topologically Sorted Source Nodes: [q], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# q => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1], [0], [1], False, [0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 4) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/s2/cs2rk3o3kmhydx4oijp6rsdb5atcrq5axy4adadrpl7gkt7scies.py
# Topologically Sorted Source Nodes: [p_attn], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# p_attn => exp
# Graph fragment:
# %mul_tensor : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_5, 1), kwargs = {})
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%mul_tensor, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_tensor, %amax_default), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub_tensor, 1.0), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%div_tensor,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = tmp14 * tmp1
tmp16 = tl_math.exp(tmp15)
tl.store(out_ptr0 + (x2), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [p_attn], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# p_attn => div_1, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_7, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_10, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [q], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4), (16, 4, 1))
# Topologically Sorted Source Nodes: [k], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(primals_6, primals_4, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4), (16, 4, 1))
# Topologically Sorted Source Nodes: [v], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_6, primals_7, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4), (16, 4, 1))
buf3 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [q], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf3, primals_2, 64, grid=grid(64), stream=stream0)
del primals_2
buf4 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [k], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf4, primals_5, 64, grid=grid(64), stream=stream0)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0), 0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [p_attn], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = reinterpret_tensor(buf5, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [p_attn], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
buf8 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [v], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf8, primals_8, 64, grid=grid(64), stream=stream0)
del primals_8
buf9 = empty_strided_cuda((16, 4, 1), (4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf7, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.convolution]
buf10 = extern_kernels.convolution(reinterpret_tensor(buf9, (4, 4, 4), (16, 4, 1), 0), primals_9, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf10, (4, 4, 4), (16, 4, 1))
buf11 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf11, primals_10, 64, grid=grid(64), stream=stream0)
del primals_10
return (buf11, buf7, primals_1, primals_3, primals_4, primals_6, primals_7, primals_9, buf7, reinterpret_tensor(buf9, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf8, (16, 1, 4), (4, 4, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 4, 1), 0), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4, 1), (4, 1, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import numpy as np
def convert_pad_shape(pad_shape):
"""Reverse, then flatten a list of lists."""
l = pad_shape[::-1]
pad_shape = [item for sublist in l for item in sublist]
return pad_shape
class BaseModule(torch.nn.Module):
def __init__(self):
super(BaseModule, self).__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class MultiHeadAttention(BaseModule):
def __init__(self, channels, out_channels, n_heads, window_size=None,
heads_share=True, p_dropout=0.0, proximal_bias=False, proximal_init
=False):
super(MultiHeadAttention, self).__init__()
assert channels % n_heads == 0
self.channels = channels
self.out_channels = out_channels
self.n_heads = n_heads
self.window_size = window_size
self.heads_share = heads_share
self.proximal_bias = proximal_bias
self.p_dropout = p_dropout
self.attn = None
self.k_channels = channels // n_heads
self.conv_q = torch.nn.Conv1d(channels, channels, 1)
self.conv_k = torch.nn.Conv1d(channels, channels, 1)
self.conv_v = torch.nn.Conv1d(channels, channels, 1)
if window_size is not None:
n_heads_rel = 1 if heads_share else n_heads
rel_stddev = self.k_channels ** -0.5
self.emb_rel_k = torch.nn.Parameter(torch.randn(n_heads_rel,
window_size * 2 + 1, self.k_channels) * rel_stddev)
self.emb_rel_v = torch.nn.Parameter(torch.randn(n_heads_rel,
window_size * 2 + 1, self.k_channels) * rel_stddev)
self.conv_o = torch.nn.Conv1d(channels, out_channels, 1)
self.drop = torch.nn.Dropout(p_dropout)
torch.nn.init.xavier_uniform_(self.conv_q.weight)
torch.nn.init.xavier_uniform_(self.conv_k.weight)
if proximal_init:
self.conv_k.weight.data.copy_(self.conv_q.weight.data)
self.conv_k.bias.data.copy_(self.conv_q.bias.data)
torch.nn.init.xavier_uniform_(self.conv_v.weight)
def forward(self, x, c, attn_mask=None):
q = self.conv_q(x)
k = self.conv_k(c)
v = self.conv_v(c)
x, self.attn = self.attention(q, k, v, mask=attn_mask)
x = self.conv_o(x)
return x
def attention(self, query, key, value, mask=None):
b, d, t_s, t_t = *key.size(), query.size(2)
query = query.view(b, self.n_heads, self.k_channels, t_t).transpose(
2, 3)
key = key.view(b, self.n_heads, self.k_channels, t_s).transpose(2, 3)
value = value.view(b, self.n_heads, self.k_channels, t_s).transpose(
2, 3)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(self
.k_channels)
if self.window_size is not None:
assert t_s == t_t, 'Relative attention is only available for self-attention.'
key_relative_embeddings = self._get_relative_embeddings(self.
emb_rel_k, t_s)
rel_logits = self._matmul_with_relative_keys(query,
key_relative_embeddings)
rel_logits = self._relative_position_to_absolute_position(
rel_logits)
scores_local = rel_logits / math.sqrt(self.k_channels)
scores = scores + scores_local
if self.proximal_bias:
assert t_s == t_t, 'Proximal bias is only available for self-attention.'
scores = scores + self._attention_bias_proximal(t_s)
if mask is not None:
scores = scores.masked_fill(mask == 0, -10000.0)
p_attn = torch.nn.functional.softmax(scores, dim=-1)
p_attn = self.drop(p_attn)
output = torch.matmul(p_attn, value)
if self.window_size is not None:
relative_weights = self._absolute_position_to_relative_position(
p_attn)
value_relative_embeddings = self._get_relative_embeddings(self.
emb_rel_v, t_s)
output = output + self._matmul_with_relative_values(
relative_weights, value_relative_embeddings)
output = output.transpose(2, 3).contiguous().view(b, d, t_t)
return output, p_attn
def _matmul_with_relative_values(self, x, y):
ret = torch.matmul(x, y.unsqueeze(0))
return ret
def _matmul_with_relative_keys(self, x, y):
ret = torch.matmul(x, y.unsqueeze(0).transpose(-2, -1))
return ret
def _get_relative_embeddings(self, relative_embeddings, length):
pad_length = max(length - (self.window_size + 1), 0)
slice_start_position = max(self.window_size + 1 - length, 0)
slice_end_position = slice_start_position + 2 * length - 1
if pad_length > 0:
padded_relative_embeddings = torch.nn.functional.pad(
relative_embeddings, convert_pad_shape([[0, 0], [pad_length,
pad_length], [0, 0]]))
else:
padded_relative_embeddings = relative_embeddings
used_relative_embeddings = padded_relative_embeddings[:,
slice_start_position:slice_end_position]
return used_relative_embeddings
def _relative_position_to_absolute_position(self, x):
batch, heads, length, _ = x.size()
x = torch.nn.functional.pad(x, convert_pad_shape([[0, 0], [0, 0], [
0, 0], [0, 1]]))
x_flat = x.view([batch, heads, length * 2 * length])
x_flat = torch.nn.functional.pad(x_flat, convert_pad_shape([[0, 0],
[0, 0], [0, length - 1]]))
x_final = x_flat.view([batch, heads, length + 1, 2 * length - 1])[:,
:, :length, length - 1:]
return x_final
def _absolute_position_to_relative_position(self, x):
batch, heads, length, _ = x.size()
x = torch.nn.functional.pad(x, convert_pad_shape([[0, 0], [0, 0], [
0, 0], [0, length - 1]]))
x_flat = x.view([batch, heads, length ** 2 + length * (length - 1)])
x_flat = torch.nn.functional.pad(x_flat, convert_pad_shape([[0, 0],
[0, 0], [length, 0]]))
x_final = x_flat.view([batch, heads, length, 2 * length])[:, :, :, 1:]
return x_final
def _attention_bias_proximal(self, length):
r = torch.arange(length, dtype=torch.float32)
diff = torch.unsqueeze(r, 0) - torch.unsqueeze(r, 1)
return torch.unsqueeze(torch.unsqueeze(-torch.log1p(torch.abs(diff)
), 0), 0)
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'channels': 4, 'out_channels': 4, 'n_heads': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import math
import numpy as np
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 4 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = tmp14 * tmp1
tmp16 = tl_math.exp(tmp15)
tl.store(out_ptr0 + x2, tmp16, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_4, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_7, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 4, 1), (4, 1, 1))
assert_size_stride(primals_10, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4), (16, 4, 1))
buf1 = extern_kernels.convolution(primals_6, primals_4, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 4), (16, 4, 1))
buf2 = extern_kernels.convolution(primals_6, primals_7, stride=(1,),
padding=(0,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4), (16, 4, 1))
buf3 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(64)](buf3, primals_2, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_2
buf4 = buf1
del buf1
triton_poi_fused_convolution_0[grid(64)](buf4, primals_5, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf3, (16, 4, 1), (4, 1, 0),
0), reinterpret_tensor(buf4, (16, 1, 4), (4, 0, 1), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf5, buf6, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf7 = reinterpret_tensor(buf5, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf5
triton_poi_fused__softmax_2[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
buf8 = buf2
del buf2
triton_poi_fused_convolution_0[grid(64)](buf8, primals_8, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_8
buf9 = empty_strided_cuda((16, 4, 1), (4, 1, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf7, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf8, (16, 4, 1), (4, 1, 0), 0), out=buf9)
buf10 = extern_kernels.convolution(reinterpret_tensor(buf9, (4, 4,
4), (16, 4, 1), 0), primals_9, stride=(1,), padding=(0,),
dilation=(1,), transposed=False, output_padding=(0,), groups=1,
bias=None)
assert_size_stride(buf10, (4, 4, 4), (16, 4, 1))
buf11 = buf10
del buf10
triton_poi_fused_convolution_0[grid(64)](buf11, primals_10, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_10
return (buf11, buf7, primals_1, primals_3, primals_4, primals_6,
primals_7, primals_9, buf7, reinterpret_tensor(buf9, (4, 4, 4), (16,
4, 1), 0), reinterpret_tensor(buf8, (16, 1, 4), (4, 4, 1), 0),
reinterpret_tensor(buf3, (16, 1, 4), (4, 4, 1), 0),
reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0))
def convert_pad_shape(pad_shape):
"""Reverse, then flatten a list of lists."""
l = pad_shape[::-1]
pad_shape = [item for sublist in l for item in sublist]
return pad_shape
class BaseModule(torch.nn.Module):
def __init__(self):
super(BaseModule, self).__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class MultiHeadAttentionNew(BaseModule):
def __init__(self, channels, out_channels, n_heads, window_size=None,
heads_share=True, p_dropout=0.0, proximal_bias=False, proximal_init
=False):
super(MultiHeadAttentionNew, self).__init__()
assert channels % n_heads == 0
self.channels = channels
self.out_channels = out_channels
self.n_heads = n_heads
self.window_size = window_size
self.heads_share = heads_share
self.proximal_bias = proximal_bias
self.p_dropout = p_dropout
self.attn = None
self.k_channels = channels // n_heads
self.conv_q = torch.nn.Conv1d(channels, channels, 1)
self.conv_k = torch.nn.Conv1d(channels, channels, 1)
self.conv_v = torch.nn.Conv1d(channels, channels, 1)
if window_size is not None:
n_heads_rel = 1 if heads_share else n_heads
rel_stddev = self.k_channels ** -0.5
self.emb_rel_k = torch.nn.Parameter(torch.randn(n_heads_rel,
window_size * 2 + 1, self.k_channels) * rel_stddev)
self.emb_rel_v = torch.nn.Parameter(torch.randn(n_heads_rel,
window_size * 2 + 1, self.k_channels) * rel_stddev)
self.conv_o = torch.nn.Conv1d(channels, out_channels, 1)
self.drop = torch.nn.Dropout(p_dropout)
torch.nn.init.xavier_uniform_(self.conv_q.weight)
torch.nn.init.xavier_uniform_(self.conv_k.weight)
if proximal_init:
self.conv_k.weight.data.copy_(self.conv_q.weight.data)
self.conv_k.bias.data.copy_(self.conv_q.bias.data)
torch.nn.init.xavier_uniform_(self.conv_v.weight)
def attention(self, query, key, value, mask=None):
b, d, t_s, t_t = *key.size(), query.size(2)
query = query.view(b, self.n_heads, self.k_channels, t_t).transpose(
2, 3)
key = key.view(b, self.n_heads, self.k_channels, t_s).transpose(2, 3)
value = value.view(b, self.n_heads, self.k_channels, t_s).transpose(
2, 3)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(self
.k_channels)
if self.window_size is not None:
assert t_s == t_t, 'Relative attention is only available for self-attention.'
key_relative_embeddings = self._get_relative_embeddings(self.
emb_rel_k, t_s)
rel_logits = self._matmul_with_relative_keys(query,
key_relative_embeddings)
rel_logits = self._relative_position_to_absolute_position(
rel_logits)
scores_local = rel_logits / math.sqrt(self.k_channels)
scores = scores + scores_local
if self.proximal_bias:
assert t_s == t_t, 'Proximal bias is only available for self-attention.'
scores = scores + self._attention_bias_proximal(t_s)
if mask is not None:
scores = scores.masked_fill(mask == 0, -10000.0)
p_attn = torch.nn.functional.softmax(scores, dim=-1)
p_attn = self.drop(p_attn)
output = torch.matmul(p_attn, value)
if self.window_size is not None:
relative_weights = self._absolute_position_to_relative_position(
p_attn)
value_relative_embeddings = self._get_relative_embeddings(self.
emb_rel_v, t_s)
output = output + self._matmul_with_relative_values(
relative_weights, value_relative_embeddings)
output = output.transpose(2, 3).contiguous().view(b, d, t_t)
return output, p_attn
def _matmul_with_relative_values(self, x, y):
ret = torch.matmul(x, y.unsqueeze(0))
return ret
def _matmul_with_relative_keys(self, x, y):
ret = torch.matmul(x, y.unsqueeze(0).transpose(-2, -1))
return ret
def _get_relative_embeddings(self, relative_embeddings, length):
pad_length = max(length - (self.window_size + 1), 0)
slice_start_position = max(self.window_size + 1 - length, 0)
slice_end_position = slice_start_position + 2 * length - 1
if pad_length > 0:
padded_relative_embeddings = torch.nn.functional.pad(
relative_embeddings, convert_pad_shape([[0, 0], [pad_length,
pad_length], [0, 0]]))
else:
padded_relative_embeddings = relative_embeddings
used_relative_embeddings = padded_relative_embeddings[:,
slice_start_position:slice_end_position]
return used_relative_embeddings
def _relative_position_to_absolute_position(self, x):
batch, heads, length, _ = x.size()
x = torch.nn.functional.pad(x, convert_pad_shape([[0, 0], [0, 0], [
0, 0], [0, 1]]))
x_flat = x.view([batch, heads, length * 2 * length])
x_flat = torch.nn.functional.pad(x_flat, convert_pad_shape([[0, 0],
[0, 0], [0, length - 1]]))
x_final = x_flat.view([batch, heads, length + 1, 2 * length - 1])[:,
:, :length, length - 1:]
return x_final
def _absolute_position_to_relative_position(self, x):
batch, heads, length, _ = x.size()
x = torch.nn.functional.pad(x, convert_pad_shape([[0, 0], [0, 0], [
0, 0], [0, length - 1]]))
x_flat = x.view([batch, heads, length ** 2 + length * (length - 1)])
x_flat = torch.nn.functional.pad(x_flat, convert_pad_shape([[0, 0],
[0, 0], [length, 0]]))
x_final = x_flat.view([batch, heads, length, 2 * length])[:, :, :, 1:]
return x_final
def _attention_bias_proximal(self, length):
r = torch.arange(length, dtype=torch.float32)
diff = torch.unsqueeze(r, 0) - torch.unsqueeze(r, 1)
return torch.unsqueeze(torch.unsqueeze(-torch.log1p(torch.abs(diff)
), 0), 0)
def forward(self, input_0, input_1):
primals_1 = self.conv_q.weight
primals_2 = self.conv_q.bias
primals_4 = self.conv_k.weight
primals_5 = self.conv_k.bias
primals_7 = self.conv_v.weight
primals_8 = self.conv_v.bias
primals_9 = self.conv_o.weight
primals_10 = self.conv_o.bias
primals_3 = input_0
primals_6 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9, primals_10])
return output[0]
| Sobsz/uberduck-ml-dev | MultiHeadAttention | false | 1,092 | [
"Apache-2.0"
] | 0 | f099238f6f2e3f600d72d89dea3c883c59d91387 | https://github.com/Sobsz/uberduck-ml-dev/tree/f099238f6f2e3f600d72d89dea3c883c59d91387 | import math
import torch
import numpy as np
def convert_pad_shape(pad_shape):
"""Reverse, then flatten a list of lists."""
l = pad_shape[::-1]
pad_shape = [item for sublist in l for item in sublist]
return pad_shape
class BaseModule(torch.nn.Module):
def __init__(self):
super().__init__()
@property
def nparams(self):
"""
Returns number of trainable parameters of the module.
"""
num_params = 0
for name, param in self.named_parameters():
if param.requires_grad:
num_params += np.prod(param.detach().cpu().numpy().shape)
return num_params
def relocate_input(self, x: 'list'):
"""
Relocates provided tensors to the same device set for the module.
"""
device = next(self.parameters()).device
for i in range(len(x)):
if isinstance(x[i], torch.Tensor) and x[i].device != device:
x[i] = x[i]
return x
class Model(BaseModule):
def __init__(self, channels, out_channels, n_heads, window_size=None,
heads_share=True, p_dropout=0.0, proximal_bias=False, proximal_init
=False):
super().__init__()
assert channels % n_heads == 0
self.channels = channels
self.out_channels = out_channels
self.n_heads = n_heads
self.window_size = window_size
self.heads_share = heads_share
self.proximal_bias = proximal_bias
self.p_dropout = p_dropout
self.attn = None
self.k_channels = channels // n_heads
self.conv_q = torch.nn.Conv1d(channels, channels, 1)
self.conv_k = torch.nn.Conv1d(channels, channels, 1)
self.conv_v = torch.nn.Conv1d(channels, channels, 1)
if window_size is not None:
n_heads_rel = 1 if heads_share else n_heads
rel_stddev = self.k_channels ** -0.5
self.emb_rel_k = torch.nn.Parameter(torch.randn(n_heads_rel,
window_size * 2 + 1, self.k_channels) * rel_stddev)
self.emb_rel_v = torch.nn.Parameter(torch.randn(n_heads_rel,
window_size * 2 + 1, self.k_channels) * rel_stddev)
self.conv_o = torch.nn.Conv1d(channels, out_channels, 1)
self.drop = torch.nn.Dropout(p_dropout)
torch.nn.init.xavier_uniform_(self.conv_q.weight)
torch.nn.init.xavier_uniform_(self.conv_k.weight)
if proximal_init:
self.conv_k.weight.data.copy_(self.conv_q.weight.data)
self.conv_k.bias.data.copy_(self.conv_q.bias.data)
torch.nn.init.xavier_uniform_(self.conv_v.weight)
def forward(self, x, c, attn_mask=None):
q = self.conv_q(x)
k = self.conv_k(c)
v = self.conv_v(c)
x, self.attn = self.attention(q, k, v, mask=attn_mask)
x = self.conv_o(x)
return x
def attention(self, query, key, value, mask=None):
b, d, t_s, t_t = *key.size(), query.size(2)
query = query.view(b, self.n_heads, self.k_channels, t_t).transpose(
2, 3)
key = key.view(b, self.n_heads, self.k_channels, t_s).transpose(2, 3)
value = value.view(b, self.n_heads, self.k_channels, t_s).transpose(
2, 3)
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(self
.k_channels)
if self.window_size is not None:
assert t_s == t_t, 'Relative attention is only available for self-attention.'
key_relative_embeddings = self._get_relative_embeddings(self.
emb_rel_k, t_s)
rel_logits = self._matmul_with_relative_keys(query,
key_relative_embeddings)
rel_logits = self._relative_position_to_absolute_position(
rel_logits)
scores_local = rel_logits / math.sqrt(self.k_channels)
scores = scores + scores_local
if self.proximal_bias:
assert t_s == t_t, 'Proximal bias is only available fo
# ... truncated (>4000 chars) for memory efficiency |
WNConv2d | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/4m/c4munax4bhqei64mhriszwqd42q6bjyh4nv3jazxhymu3f7wtucw.py
# Topologically Sorted Source Nodes: [_weight_norm], Original ATen: [aten._weight_norm_interface]
# Source node to ATen node mapping:
# _weight_norm => div, mul, pow_1, pow_2, sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_2, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1, 2, 3], True), kwargs = {})
# %pow_2 : [num_users=2] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_1, %pow_2), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %div), kwargs = {})
triton_per_fused__weight_norm_interface_0 = async_compile.triton('triton_per_fused__weight_norm_interface_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__weight_norm_interface_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__weight_norm_interface_0(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp7 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp8 = tmp7 / tmp6
tmp9 = tmp0 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr0 + (r1 + (64*x0)), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/tc/ctcagp37ljugm52zu6ckorigrppqo67voefe2f2odg5r6hyllhyu.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# out => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_4, %mul, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 1, 1, 1), (1, 1, 1, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf0 # reuse
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [_weight_norm], Original ATen: [aten._weight_norm_interface]
stream0 = get_raw_stream(0)
triton_per_fused__weight_norm_interface_0.run(buf1, primals_2, primals_1, buf2, 4, 64, grid=grid(4), stream=stream0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(primals_4, buf2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 1, 1), (4, 1, 1, 1))
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf4, primals_3, 16, grid=grid(16), stream=stream0)
del primals_3
return (buf4, buf2, primals_1, primals_2, primals_4, buf1, buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 1, 1, 1), (1, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
import torch.utils.data
class WNConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding=0, bias=True, activation=None):
super().__init__()
self.conv = nn.utils.weight_norm(nn.Conv2d(in_channel, out_channel,
kernel_size, stride=stride, padding=padding, bias=bias))
self.out_channel = out_channel
if isinstance(kernel_size, int):
kernel_size = [kernel_size, kernel_size]
self.kernel_size = kernel_size
self.activation = activation
def forward(self, input):
out = self.conv(input)
if self.activation is not None:
out = self.activation(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channel': 4, 'out_channel': 4, 'kernel_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused__weight_norm_interface_0(in_out_ptr0, in_ptr0, in_ptr1,
out_ptr0, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp7 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp8 = tmp7 / tmp6
tmp9 = tmp0 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr0 + (r1 + 64 * x0), tmp9, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x2, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 1, 1, 1), (1, 1, 1, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused__weight_norm_interface_0[grid(4)](buf1, primals_2,
primals_1, buf2, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
buf3 = extern_kernels.convolution(primals_4, buf2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 1, 1), (4, 1, 1, 1))
buf4 = buf3
del buf3
triton_poi_fused_convolution_1[grid(16)](buf4, primals_3, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_3
return buf4, buf2, primals_1, primals_2, primals_4, buf1, buf2
class WNConv2dNew(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding=0, bias=True, activation=None):
super().__init__()
self.conv = nn.utils.weight_norm(nn.Conv2d(in_channel, out_channel,
kernel_size, stride=stride, padding=padding, bias=bias))
self.out_channel = out_channel
if isinstance(kernel_size, int):
kernel_size = [kernel_size, kernel_size]
self.kernel_size = kernel_size
self.activation = activation
def forward(self, input_0):
primals_3 = self.conv.bias
primals_1 = self.conv.weight_g
primals_2 = self.conv.weight_v
primals_4 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| Shivanshu-Gupta/KaoKore-VQ-VAE2 | WNConv2d | false | 1,093 | [
"MIT"
] | 0 | 38a88ba312dee3c0e2c1aaf02e1c1754ba19ac0c | https://github.com/Shivanshu-Gupta/KaoKore-VQ-VAE2/tree/38a88ba312dee3c0e2c1aaf02e1c1754ba19ac0c | import torch
from torch import nn
import torch.utils.data
class Model(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, stride=1,
padding=0, bias=True, activation=None):
super().__init__()
self.conv = nn.utils.weight_norm(nn.Conv2d(in_channel, out_channel,
kernel_size, stride=stride, padding=padding, bias=bias))
self.out_channel = out_channel
if isinstance(kernel_size, int):
kernel_size = [kernel_size, kernel_size]
self.kernel_size = kernel_size
self.activation = activation
def forward(self, input):
out = self.conv(input)
if self.activation is not None:
out = self.activation(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
DisparityRegression | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/mt/cmtrzhde3xvkbbrbxbhlpwzr6zziv6z7o2vl6xhkqbvvxtyrdmk5.py
# Topologically Sorted Source Nodes: [max_d, index, setitem, setitem_1, prob, index_1, setitem_2, setitem_3, prob_1, index_2, setitem_4, setitem_5, prob_2, index_3, setitem_6, setitem_7, prob_3, index_4, setitem_8, setitem_9, prob_4, index_5, setitem_10, setitem_11, prob_5, index_6, setitem_12, setitem_13, prob_6, index_7, setitem_14, setitem_15, prob_7, index_8, setitem_16, setitem_17, prob_8], Original ATen: [aten.argmax, aten.add, aten.lift_fresh, aten.index_put, aten.gather]
# Source node to ATen node mapping:
# index => add
# index_1 => add_1
# index_2 => add_2
# index_3 => add_3
# index_4 => add_4
# index_5 => add_5
# index_6 => add_6
# index_7 => add_7
# index_8 => add_8
# max_d => argmax
# prob => gather
# prob_1 => gather_1
# prob_2 => gather_2
# prob_3 => gather_3
# prob_4 => gather_4
# prob_5 => gather_5
# prob_6 => gather_6
# prob_7 => gather_7
# prob_8 => gather_8
# setitem => full_default, index_put
# setitem_1 => full_default_1, index_put_1
# setitem_10 => full_default_10, index_put_10
# setitem_11 => full_default_11, index_put_11
# setitem_12 => full_default_12, index_put_12
# setitem_13 => full_default_13, index_put_13
# setitem_14 => full_default_14, index_put_14
# setitem_15 => full_default_15, index_put_15
# setitem_16 => full_default_16, index_put_16
# setitem_17 => full_default_17, index_put_17
# setitem_2 => full_default_2, index_put_2
# setitem_3 => full_default_3, index_put_3
# setitem_4 => full_default_4, index_put_4
# setitem_5 => full_default_5, index_put_5
# setitem_6 => full_default_6, index_put_6
# setitem_7 => full_default_7, index_put_7
# setitem_8 => full_default_8, index_put_8
# setitem_9 => full_default_9, index_put_9
# Graph fragment:
# %argmax : [num_users=9] = call_function[target=torch.ops.aten.argmax.default](args = (%arg0_1, 1, True), kwargs = {})
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, -4), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add, [%lt], %full_default), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_1 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put, [%gt], %full_default_1), kwargs = {})
# %gather : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_1), kwargs = {})
# %add_1 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, -3), kwargs = {})
# %full_default_2 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_2 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_1, [%lt_1], %full_default_2), kwargs = {})
# %full_default_3 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_3 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_2, [%gt_1], %full_default_3), kwargs = {})
# %gather_1 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_3), kwargs = {})
# %add_2 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, -2), kwargs = {})
# %full_default_4 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_4 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_2, [%lt_2], %full_default_4), kwargs = {})
# %full_default_5 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_5 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_4, [%gt_2], %full_default_5), kwargs = {})
# %gather_2 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_5), kwargs = {})
# %add_3 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, -1), kwargs = {})
# %full_default_6 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_6 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_3, [%lt_3], %full_default_6), kwargs = {})
# %full_default_7 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_7 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_6, [%gt_3], %full_default_7), kwargs = {})
# %gather_3 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_7), kwargs = {})
# %add_4 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, 0), kwargs = {})
# %full_default_8 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_8 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_4, [%lt_4], %full_default_8), kwargs = {})
# %full_default_9 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_9 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_8, [%gt_4], %full_default_9), kwargs = {})
# %gather_4 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_9), kwargs = {})
# %add_5 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, 1), kwargs = {})
# %full_default_10 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_10 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_5, [%lt_5], %full_default_10), kwargs = {})
# %full_default_11 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_11 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_10, [%gt_5], %full_default_11), kwargs = {})
# %gather_5 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_11), kwargs = {})
# %add_6 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, 2), kwargs = {})
# %full_default_12 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_12 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_6, [%lt_6], %full_default_12), kwargs = {})
# %full_default_13 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_13 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_12, [%gt_6], %full_default_13), kwargs = {})
# %gather_6 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_13), kwargs = {})
# %add_7 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, 3), kwargs = {})
# %full_default_14 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_14 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_7, [%lt_7], %full_default_14), kwargs = {})
# %full_default_15 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_15 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_14, [%gt_7], %full_default_15), kwargs = {})
# %gather_7 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_15), kwargs = {})
# %add_8 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%argmax, 4), kwargs = {})
# %full_default_16 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_16 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%add_8, [%lt_8], %full_default_16), kwargs = {})
# %full_default_17 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 3), kwargs = {dtype: torch.int64, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put_17 : [num_users=2] = call_function[target=torch.ops.aten.index_put_.default](args = (%index_put_16, [%gt_8], %full_default_17), kwargs = {})
# %gather_8 : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%arg0_1, 1, %index_put_17), kwargs = {})
triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0 = async_compile.triton('triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i64', 2: '*i64', 3: '*i64', 4: '*i64', 5: '*i64', 6: '*i64', 7: '*i64', 8: '*i64', 9: '*i64', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: '*fp32', 15: '*fp32', 16: '*fp32', 17: '*fp32', 18: '*fp32', 19: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 19), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0(in_ptr0, out_ptr2, out_ptr4, out_ptr6, out_ptr8, out_ptr10, out_ptr12, out_ptr14, out_ptr16, out_ptr18, out_ptr19, out_ptr20, out_ptr21, out_ptr22, out_ptr23, out_ptr24, out_ptr25, out_ptr26, out_ptr27, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp17 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp32 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp2 = tmp0 > tmp1
tmp3 = tmp0 == tmp1
tmp4 = tmp0 != tmp0
tmp5 = tmp1 != tmp1
tmp6 = tmp4 > tmp5
tmp7 = tmp2 | tmp6
tmp8 = tmp4 & tmp5
tmp9 = tmp3 | tmp8
tmp10 = tl.full([1], 0, tl.int64)
tmp11 = tl.full([1], 1, tl.int64)
tmp12 = tmp10 < tmp11
tmp13 = tmp9 & tmp12
tmp14 = tmp7 | tmp13
tmp15 = tl.where(tmp14, tmp0, tmp1)
tmp16 = tl.where(tmp14, tmp10, tmp11)
tmp18 = tmp15 > tmp17
tmp19 = tmp15 == tmp17
tmp20 = tmp15 != tmp15
tmp21 = tmp17 != tmp17
tmp22 = tmp20 > tmp21
tmp23 = tmp18 | tmp22
tmp24 = tmp20 & tmp21
tmp25 = tmp19 | tmp24
tmp26 = tl.full([1], 2, tl.int64)
tmp27 = tmp16 < tmp26
tmp28 = tmp25 & tmp27
tmp29 = tmp23 | tmp28
tmp30 = tl.where(tmp29, tmp15, tmp17)
tmp31 = tl.where(tmp29, tmp16, tmp26)
tmp33 = tmp30 > tmp32
tmp34 = tmp30 == tmp32
tmp35 = tmp30 != tmp30
tmp36 = tmp32 != tmp32
tmp37 = tmp35 > tmp36
tmp38 = tmp33 | tmp37
tmp39 = tmp35 & tmp36
tmp40 = tmp34 | tmp39
tmp41 = tl.full([1], 3, tl.int64)
tmp42 = tmp31 < tmp41
tmp43 = tmp40 & tmp42
tmp44 = tmp38 | tmp43
tmp45 = tl.where(tmp44, tmp30, tmp32)
tmp46 = tl.where(tmp44, tmp31, tmp41)
tmp47 = tl.full([1], -4, tl.int64)
tmp48 = tmp46 + tmp47
tmp49 = tmp48 < tmp10
tmp50 = tl.where(tmp49, tmp10, tmp48)
tmp51 = tmp50 > tmp41
tmp52 = tl.where(tmp51, tmp41, tmp50)
tmp53 = tl.full([1], -3, tl.int64)
tmp54 = tmp46 + tmp53
tmp55 = tmp54 < tmp10
tmp56 = tl.where(tmp55, tmp10, tmp54)
tmp57 = tmp56 > tmp41
tmp58 = tl.where(tmp57, tmp41, tmp56)
tmp59 = tl.full([1], -2, tl.int64)
tmp60 = tmp46 + tmp59
tmp61 = tmp60 < tmp10
tmp62 = tl.where(tmp61, tmp10, tmp60)
tmp63 = tmp62 > tmp41
tmp64 = tl.where(tmp63, tmp41, tmp62)
tmp65 = tl.full([1], -1, tl.int64)
tmp66 = tmp46 + tmp65
tmp67 = tmp66 < tmp10
tmp68 = tl.where(tmp67, tmp10, tmp66)
tmp69 = tmp68 > tmp41
tmp70 = tl.where(tmp69, tmp41, tmp68)
tmp71 = tmp46 + tmp10
tmp72 = tmp71 < tmp10
tmp73 = tl.where(tmp72, tmp10, tmp71)
tmp74 = tmp73 > tmp41
tmp75 = tl.where(tmp74, tmp41, tmp73)
tmp76 = tmp46 + tmp11
tmp77 = tmp76 < tmp10
tmp78 = tl.where(tmp77, tmp10, tmp76)
tmp79 = tmp78 > tmp41
tmp80 = tl.where(tmp79, tmp41, tmp78)
tmp81 = tmp46 + tmp26
tmp82 = tmp81 < tmp10
tmp83 = tl.where(tmp82, tmp10, tmp81)
tmp84 = tmp83 > tmp41
tmp85 = tl.where(tmp84, tmp41, tmp83)
tmp86 = tmp46 + tmp41
tmp87 = tmp86 < tmp10
tmp88 = tl.where(tmp87, tmp10, tmp86)
tmp89 = tmp88 > tmp41
tmp90 = tl.where(tmp89, tmp41, tmp88)
tmp91 = tl.full([1], 4, tl.int64)
tmp92 = tmp46 + tmp91
tmp93 = tmp92 < tmp10
tmp94 = tl.where(tmp93, tmp10, tmp92)
tmp95 = tmp94 > tmp41
tmp96 = tl.where(tmp95, tmp41, tmp94)
tmp97 = tl.full([XBLOCK], 4, tl.int32)
tmp98 = tmp52 + tmp97
tmp99 = tmp52 < 0
tmp100 = tl.where(tmp99, tmp98, tmp52)
tl.device_assert(((0 <= tmp100) & (tmp100 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp100 < 4")
tmp102 = tl.load(in_ptr0 + (x0 + (16*tmp100) + (64*x1)), xmask)
tmp103 = tmp58 + tmp97
tmp104 = tmp58 < 0
tmp105 = tl.where(tmp104, tmp103, tmp58)
tl.device_assert(((0 <= tmp105) & (tmp105 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp105 < 4")
tmp107 = tl.load(in_ptr0 + (x0 + (16*tmp105) + (64*x1)), xmask)
tmp108 = tmp64 + tmp97
tmp109 = tmp64 < 0
tmp110 = tl.where(tmp109, tmp108, tmp64)
tl.device_assert(((0 <= tmp110) & (tmp110 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp110 < 4")
tmp112 = tl.load(in_ptr0 + (x0 + (16*tmp110) + (64*x1)), xmask)
tmp113 = tmp70 + tmp97
tmp114 = tmp70 < 0
tmp115 = tl.where(tmp114, tmp113, tmp70)
tl.device_assert(((0 <= tmp115) & (tmp115 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp115 < 4")
tmp117 = tl.load(in_ptr0 + (x0 + (16*tmp115) + (64*x1)), xmask)
tmp118 = tmp75 + tmp97
tmp119 = tmp75 < 0
tmp120 = tl.where(tmp119, tmp118, tmp75)
tl.device_assert(((0 <= tmp120) & (tmp120 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp120 < 4")
tmp122 = tl.load(in_ptr0 + (x0 + (16*tmp120) + (64*x1)), xmask)
tmp123 = tmp80 + tmp97
tmp124 = tmp80 < 0
tmp125 = tl.where(tmp124, tmp123, tmp80)
tl.device_assert(((0 <= tmp125) & (tmp125 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp125 < 4")
tmp127 = tl.load(in_ptr0 + (x0 + (16*tmp125) + (64*x1)), xmask)
tmp128 = tmp85 + tmp97
tmp129 = tmp85 < 0
tmp130 = tl.where(tmp129, tmp128, tmp85)
tl.device_assert(((0 <= tmp130) & (tmp130 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp130 < 4")
tmp132 = tl.load(in_ptr0 + (x0 + (16*tmp130) + (64*x1)), xmask)
tmp133 = tmp90 + tmp97
tmp134 = tmp90 < 0
tmp135 = tl.where(tmp134, tmp133, tmp90)
tl.device_assert(((0 <= tmp135) & (tmp135 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp135 < 4")
tmp137 = tl.load(in_ptr0 + (x0 + (16*tmp135) + (64*x1)), xmask)
tmp138 = tmp96 + tmp97
tmp139 = tmp96 < 0
tmp140 = tl.where(tmp139, tmp138, tmp96)
tl.device_assert(((0 <= tmp140) & (tmp140 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp140 < 4")
tmp142 = tl.load(in_ptr0 + (x0 + (16*tmp140) + (64*x1)), xmask)
tl.store(out_ptr2 + (x0 + (144*x1)), tmp52, xmask)
tl.store(out_ptr4 + (x0 + (144*x1)), tmp58, xmask)
tl.store(out_ptr6 + (x0 + (144*x1)), tmp64, xmask)
tl.store(out_ptr8 + (x0 + (144*x1)), tmp70, xmask)
tl.store(out_ptr10 + (x0 + (144*x1)), tmp75, xmask)
tl.store(out_ptr12 + (x0 + (144*x1)), tmp80, xmask)
tl.store(out_ptr14 + (x0 + (144*x1)), tmp85, xmask)
tl.store(out_ptr16 + (x0 + (144*x1)), tmp90, xmask)
tl.store(out_ptr18 + (x0 + (144*x1)), tmp96, xmask)
tl.store(out_ptr19 + (9*x2), tmp102, xmask)
tl.store(out_ptr20 + (9*x2), tmp107, xmask)
tl.store(out_ptr21 + (9*x2), tmp112, xmask)
tl.store(out_ptr22 + (9*x2), tmp117, xmask)
tl.store(out_ptr23 + (9*x2), tmp122, xmask)
tl.store(out_ptr24 + (9*x2), tmp127, xmask)
tl.store(out_ptr25 + (9*x2), tmp132, xmask)
tl.store(out_ptr26 + (9*x2), tmp137, xmask)
tl.store(out_ptr27 + (9*x2), tmp142, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/hc/chcrv4z3xuibekjnrpmxzsnizu3g4ubg6rgohy7cmfzimhemq5jp.py
# Topologically Sorted Source Nodes: [sum_1, add_9, part_x_1, part_d, mul, out], Original ATen: [aten.sum, aten.add, aten.div, aten._to_copy, aten.mul]
# Source node to ATen node mapping:
# add_9 => add_9
# mul => mul
# out => sum_2
# part_d => convert_element_type_1
# part_x_1 => div
# sum_1 => sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%cat, [1], True), kwargs = {})
# %add_9 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, 1e-08), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%cat, %add_9), kwargs = {})
# %convert_element_type_1 : [num_users=1] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%cat_1, torch.float32), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %convert_element_type_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
triton_per_fused__to_copy_add_div_mul_sum_1 = async_compile.triton('triton_per_fused__to_copy_add_div_mul_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[64, 16],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i64', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__to_copy_add_div_mul_sum_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 2, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__to_copy_add_div_mul_sum_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 9
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = rindex < rnumel
r1 = rindex
x0 = xindex
x2 = xindex % 16
x3 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (r1 + (9*x0)), rmask & xmask, other=0.0)
tmp8 = tl.load(in_ptr1 + (x2 + (16*r1) + (144*x3)), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp5 = 1e-08
tmp6 = tmp4 + tmp5
tmp7 = tmp0 / tmp6
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 * tmp9
tmp11 = tl.broadcast_to(tmp10, [XBLOCK, RBLOCK])
tmp13 = tl.where(rmask & xmask, tmp11, 0)
tmp14 = tl.sum(tmp13, 1)[:, None]
tl.store(in_out_ptr0 + (x0), tmp14, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf30 = empty_strided_cuda((4, 9, 4, 4), (144, 16, 4, 1), torch.int64)
buf2 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 0) # alias
buf4 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 16) # alias
buf6 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 32) # alias
buf8 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 48) # alias
buf10 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 64) # alias
buf12 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 80) # alias
buf14 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 96) # alias
buf16 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 112) # alias
buf18 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 128) # alias
buf28 = empty_strided_cuda((4, 9, 4, 4), (144, 1, 36, 9), torch.float32)
buf19 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 0) # alias
buf20 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 1) # alias
buf21 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 2) # alias
buf22 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 3) # alias
buf23 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 4) # alias
buf24 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 5) # alias
buf25 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 6) # alias
buf26 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 7) # alias
buf27 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 8) # alias
# Topologically Sorted Source Nodes: [max_d, index, setitem, setitem_1, prob, index_1, setitem_2, setitem_3, prob_1, index_2, setitem_4, setitem_5, prob_2, index_3, setitem_6, setitem_7, prob_3, index_4, setitem_8, setitem_9, prob_4, index_5, setitem_10, setitem_11, prob_5, index_6, setitem_12, setitem_13, prob_6, index_7, setitem_14, setitem_15, prob_7, index_8, setitem_16, setitem_17, prob_8], Original ATen: [aten.argmax, aten.add, aten.lift_fresh, aten.index_put, aten.gather]
stream0 = get_raw_stream(0)
triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0.run(arg0_1, buf2, buf4, buf6, buf8, buf10, buf12, buf14, buf16, buf18, buf19, buf20, buf21, buf22, buf23, buf24, buf25, buf26, buf27, 64, grid=grid(64), stream=stream0)
del arg0_1
buf29 = empty_strided_cuda((4, 1, 4, 4), (16, 64, 4, 1), torch.float32)
buf31 = reinterpret_tensor(buf29, (4, 4, 4), (16, 4, 1), 0); del buf29 # reuse
# Topologically Sorted Source Nodes: [sum_1, add_9, part_x_1, part_d, mul, out], Original ATen: [aten.sum, aten.add, aten.div, aten._to_copy, aten.mul]
triton_per_fused__to_copy_add_div_mul_sum_1.run(buf31, buf28, buf30, 64, 9, grid=grid(64), stream=stream0)
del buf10
del buf12
del buf14
del buf16
del buf18
del buf19
del buf2
del buf20
del buf21
del buf22
del buf23
del buf24
del buf25
del buf26
del buf27
del buf28
del buf30
del buf4
del buf6
del buf8
return (buf31, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
class DisparityRegression(nn.Module):
def __init__(self, maxdisp, win_size):
super(DisparityRegression, self).__init__()
self.max_disp = maxdisp
self.win_size = win_size
def forward(self, x):
disp = torch.arange(0, self.max_disp).view(1, -1, 1, 1).float()
if self.win_size > 0:
max_d = torch.argmax(x, dim=1, keepdim=True)
d_value = []
prob_value = []
for d in range(-self.win_size, self.win_size + 1):
index = max_d + d
index[index < 0] = 0
index[index > x.shape[1] - 1] = x.shape[1] - 1
d_value.append(index)
prob = torch.gather(x, dim=1, index=index)
prob_value.append(prob)
part_x = torch.cat(prob_value, dim=1)
part_x = part_x / (torch.sum(part_x, dim=1, keepdim=True) + 1e-08)
part_d = torch.cat(d_value, dim=1).float()
out = torch.sum(part_x * part_d, dim=1)
else:
out = torch.sum(x * disp, 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'maxdisp': 4, 'win_size': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0(in_ptr0,
out_ptr2, out_ptr4, out_ptr6, out_ptr8, out_ptr10, out_ptr12, out_ptr14,
out_ptr16, out_ptr18, out_ptr19, out_ptr20, out_ptr21, out_ptr22,
out_ptr23, out_ptr24, out_ptr25, out_ptr26, out_ptr27, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp17 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp32 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp2 = tmp0 > tmp1
tmp3 = tmp0 == tmp1
tmp4 = tmp0 != tmp0
tmp5 = tmp1 != tmp1
tmp6 = tmp4 > tmp5
tmp7 = tmp2 | tmp6
tmp8 = tmp4 & tmp5
tmp9 = tmp3 | tmp8
tmp10 = tl.full([1], 0, tl.int64)
tmp11 = tl.full([1], 1, tl.int64)
tmp12 = tmp10 < tmp11
tmp13 = tmp9 & tmp12
tmp14 = tmp7 | tmp13
tmp15 = tl.where(tmp14, tmp0, tmp1)
tmp16 = tl.where(tmp14, tmp10, tmp11)
tmp18 = tmp15 > tmp17
tmp19 = tmp15 == tmp17
tmp20 = tmp15 != tmp15
tmp21 = tmp17 != tmp17
tmp22 = tmp20 > tmp21
tmp23 = tmp18 | tmp22
tmp24 = tmp20 & tmp21
tmp25 = tmp19 | tmp24
tmp26 = tl.full([1], 2, tl.int64)
tmp27 = tmp16 < tmp26
tmp28 = tmp25 & tmp27
tmp29 = tmp23 | tmp28
tmp30 = tl.where(tmp29, tmp15, tmp17)
tmp31 = tl.where(tmp29, tmp16, tmp26)
tmp33 = tmp30 > tmp32
tmp34 = tmp30 == tmp32
tmp35 = tmp30 != tmp30
tmp36 = tmp32 != tmp32
tmp37 = tmp35 > tmp36
tmp38 = tmp33 | tmp37
tmp39 = tmp35 & tmp36
tmp40 = tmp34 | tmp39
tmp41 = tl.full([1], 3, tl.int64)
tmp42 = tmp31 < tmp41
tmp43 = tmp40 & tmp42
tmp44 = tmp38 | tmp43
tl.where(tmp44, tmp30, tmp32)
tmp46 = tl.where(tmp44, tmp31, tmp41)
tmp47 = tl.full([1], -4, tl.int64)
tmp48 = tmp46 + tmp47
tmp49 = tmp48 < tmp10
tmp50 = tl.where(tmp49, tmp10, tmp48)
tmp51 = tmp50 > tmp41
tmp52 = tl.where(tmp51, tmp41, tmp50)
tmp53 = tl.full([1], -3, tl.int64)
tmp54 = tmp46 + tmp53
tmp55 = tmp54 < tmp10
tmp56 = tl.where(tmp55, tmp10, tmp54)
tmp57 = tmp56 > tmp41
tmp58 = tl.where(tmp57, tmp41, tmp56)
tmp59 = tl.full([1], -2, tl.int64)
tmp60 = tmp46 + tmp59
tmp61 = tmp60 < tmp10
tmp62 = tl.where(tmp61, tmp10, tmp60)
tmp63 = tmp62 > tmp41
tmp64 = tl.where(tmp63, tmp41, tmp62)
tmp65 = tl.full([1], -1, tl.int64)
tmp66 = tmp46 + tmp65
tmp67 = tmp66 < tmp10
tmp68 = tl.where(tmp67, tmp10, tmp66)
tmp69 = tmp68 > tmp41
tmp70 = tl.where(tmp69, tmp41, tmp68)
tmp71 = tmp46 + tmp10
tmp72 = tmp71 < tmp10
tmp73 = tl.where(tmp72, tmp10, tmp71)
tmp74 = tmp73 > tmp41
tmp75 = tl.where(tmp74, tmp41, tmp73)
tmp76 = tmp46 + tmp11
tmp77 = tmp76 < tmp10
tmp78 = tl.where(tmp77, tmp10, tmp76)
tmp79 = tmp78 > tmp41
tmp80 = tl.where(tmp79, tmp41, tmp78)
tmp81 = tmp46 + tmp26
tmp82 = tmp81 < tmp10
tmp83 = tl.where(tmp82, tmp10, tmp81)
tmp84 = tmp83 > tmp41
tmp85 = tl.where(tmp84, tmp41, tmp83)
tmp86 = tmp46 + tmp41
tmp87 = tmp86 < tmp10
tmp88 = tl.where(tmp87, tmp10, tmp86)
tmp89 = tmp88 > tmp41
tmp90 = tl.where(tmp89, tmp41, tmp88)
tmp91 = tl.full([1], 4, tl.int64)
tmp92 = tmp46 + tmp91
tmp93 = tmp92 < tmp10
tmp94 = tl.where(tmp93, tmp10, tmp92)
tmp95 = tmp94 > tmp41
tmp96 = tl.where(tmp95, tmp41, tmp94)
tmp97 = tl.full([XBLOCK], 4, tl.int32)
tmp98 = tmp52 + tmp97
tmp99 = tmp52 < 0
tmp100 = tl.where(tmp99, tmp98, tmp52)
tl.device_assert((0 <= tmp100) & (tmp100 < 4) | ~xmask,
'index out of bounds: 0 <= tmp100 < 4')
tmp102 = tl.load(in_ptr0 + (x0 + 16 * tmp100 + 64 * x1), xmask)
tmp103 = tmp58 + tmp97
tmp104 = tmp58 < 0
tmp105 = tl.where(tmp104, tmp103, tmp58)
tl.device_assert((0 <= tmp105) & (tmp105 < 4) | ~xmask,
'index out of bounds: 0 <= tmp105 < 4')
tmp107 = tl.load(in_ptr0 + (x0 + 16 * tmp105 + 64 * x1), xmask)
tmp108 = tmp64 + tmp97
tmp109 = tmp64 < 0
tmp110 = tl.where(tmp109, tmp108, tmp64)
tl.device_assert((0 <= tmp110) & (tmp110 < 4) | ~xmask,
'index out of bounds: 0 <= tmp110 < 4')
tmp112 = tl.load(in_ptr0 + (x0 + 16 * tmp110 + 64 * x1), xmask)
tmp113 = tmp70 + tmp97
tmp114 = tmp70 < 0
tmp115 = tl.where(tmp114, tmp113, tmp70)
tl.device_assert((0 <= tmp115) & (tmp115 < 4) | ~xmask,
'index out of bounds: 0 <= tmp115 < 4')
tmp117 = tl.load(in_ptr0 + (x0 + 16 * tmp115 + 64 * x1), xmask)
tmp118 = tmp75 + tmp97
tmp119 = tmp75 < 0
tmp120 = tl.where(tmp119, tmp118, tmp75)
tl.device_assert((0 <= tmp120) & (tmp120 < 4) | ~xmask,
'index out of bounds: 0 <= tmp120 < 4')
tmp122 = tl.load(in_ptr0 + (x0 + 16 * tmp120 + 64 * x1), xmask)
tmp123 = tmp80 + tmp97
tmp124 = tmp80 < 0
tmp125 = tl.where(tmp124, tmp123, tmp80)
tl.device_assert((0 <= tmp125) & (tmp125 < 4) | ~xmask,
'index out of bounds: 0 <= tmp125 < 4')
tmp127 = tl.load(in_ptr0 + (x0 + 16 * tmp125 + 64 * x1), xmask)
tmp128 = tmp85 + tmp97
tmp129 = tmp85 < 0
tmp130 = tl.where(tmp129, tmp128, tmp85)
tl.device_assert((0 <= tmp130) & (tmp130 < 4) | ~xmask,
'index out of bounds: 0 <= tmp130 < 4')
tmp132 = tl.load(in_ptr0 + (x0 + 16 * tmp130 + 64 * x1), xmask)
tmp133 = tmp90 + tmp97
tmp134 = tmp90 < 0
tmp135 = tl.where(tmp134, tmp133, tmp90)
tl.device_assert((0 <= tmp135) & (tmp135 < 4) | ~xmask,
'index out of bounds: 0 <= tmp135 < 4')
tmp137 = tl.load(in_ptr0 + (x0 + 16 * tmp135 + 64 * x1), xmask)
tmp138 = tmp96 + tmp97
tmp139 = tmp96 < 0
tmp140 = tl.where(tmp139, tmp138, tmp96)
tl.device_assert((0 <= tmp140) & (tmp140 < 4) | ~xmask,
'index out of bounds: 0 <= tmp140 < 4')
tmp142 = tl.load(in_ptr0 + (x0 + 16 * tmp140 + 64 * x1), xmask)
tl.store(out_ptr2 + (x0 + 144 * x1), tmp52, xmask)
tl.store(out_ptr4 + (x0 + 144 * x1), tmp58, xmask)
tl.store(out_ptr6 + (x0 + 144 * x1), tmp64, xmask)
tl.store(out_ptr8 + (x0 + 144 * x1), tmp70, xmask)
tl.store(out_ptr10 + (x0 + 144 * x1), tmp75, xmask)
tl.store(out_ptr12 + (x0 + 144 * x1), tmp80, xmask)
tl.store(out_ptr14 + (x0 + 144 * x1), tmp85, xmask)
tl.store(out_ptr16 + (x0 + 144 * x1), tmp90, xmask)
tl.store(out_ptr18 + (x0 + 144 * x1), tmp96, xmask)
tl.store(out_ptr19 + 9 * x2, tmp102, xmask)
tl.store(out_ptr20 + 9 * x2, tmp107, xmask)
tl.store(out_ptr21 + 9 * x2, tmp112, xmask)
tl.store(out_ptr22 + 9 * x2, tmp117, xmask)
tl.store(out_ptr23 + 9 * x2, tmp122, xmask)
tl.store(out_ptr24 + 9 * x2, tmp127, xmask)
tl.store(out_ptr25 + 9 * x2, tmp132, xmask)
tl.store(out_ptr26 + 9 * x2, tmp137, xmask)
tl.store(out_ptr27 + 9 * x2, tmp142, xmask)
@triton.jit
def triton_per_fused__to_copy_add_div_mul_sum_1(in_out_ptr0, in_ptr0,
in_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 64
rnumel = 9
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
rmask = rindex < rnumel
r1 = rindex
x0 = xindex
x2 = xindex % 16
x3 = xindex // 16
tmp0 = tl.load(in_ptr0 + (r1 + 9 * x0), rmask & xmask, other=0.0)
tmp8 = tl.load(in_ptr1 + (x2 + 16 * r1 + 144 * x3), rmask & xmask,
other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp5 = 1e-08
tmp6 = tmp4 + tmp5
tmp7 = tmp0 / tmp6
tmp9 = tmp8.to(tl.float32)
tmp10 = tmp7 * tmp9
tmp11 = tl.broadcast_to(tmp10, [XBLOCK, RBLOCK])
tmp13 = tl.where(rmask & xmask, tmp11, 0)
tmp14 = tl.sum(tmp13, 1)[:, None]
tl.store(in_out_ptr0 + x0, tmp14, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf30 = empty_strided_cuda((4, 9, 4, 4), (144, 16, 4, 1), torch.int64)
buf2 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 0)
buf4 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 16)
buf6 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 32)
buf8 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 48)
buf10 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 64)
buf12 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 80)
buf14 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 96)
buf16 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 112)
buf18 = reinterpret_tensor(buf30, (4, 1, 4, 4), (144, 16, 4, 1), 128)
buf28 = empty_strided_cuda((4, 9, 4, 4), (144, 1, 36, 9), torch.float32
)
buf19 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 0)
buf20 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 1)
buf21 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 2)
buf22 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 3)
buf23 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 4)
buf24 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 5)
buf25 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 6)
buf26 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 7)
buf27 = reinterpret_tensor(buf28, (4, 1, 4, 4), (144, 1, 36, 9), 8)
get_raw_stream(0)
triton_poi_fused_add_argmax_gather_index_put_lift_fresh_0[grid(64)](
arg0_1, buf2, buf4, buf6, buf8, buf10, buf12, buf14, buf16,
buf18, buf19, buf20, buf21, buf22, buf23, buf24, buf25, buf26,
buf27, 64, XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
buf29 = empty_strided_cuda((4, 1, 4, 4), (16, 64, 4, 1), torch.float32)
buf31 = reinterpret_tensor(buf29, (4, 4, 4), (16, 4, 1), 0)
del buf29
triton_per_fused__to_copy_add_div_mul_sum_1[grid(64)](buf31, buf28,
buf30, 64, 9, XBLOCK=8, num_warps=2, num_stages=1)
del buf10
del buf12
del buf14
del buf16
del buf18
del buf19
del buf2
del buf20
del buf21
del buf22
del buf23
del buf24
del buf25
del buf26
del buf27
del buf28
del buf30
del buf4
del buf6
del buf8
return buf31,
class DisparityRegressionNew(nn.Module):
def __init__(self, maxdisp, win_size):
super(DisparityRegressionNew, self).__init__()
self.max_disp = maxdisp
self.win_size = win_size
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| SpadeLiu/Graft-PSMNet | DisparityRegression | false | 1,094 | [
"MIT"
] | 0 | 1f2950d5afd85237f8d3604caab20dd47a8c9889 | https://github.com/SpadeLiu/Graft-PSMNet/tree/1f2950d5afd85237f8d3604caab20dd47a8c9889 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def __init__(self, maxdisp, win_size):
super().__init__()
self.max_disp = maxdisp
self.win_size = win_size
def forward(self, x):
disp = torch.arange(0, self.max_disp).view(1, -1, 1, 1).float()
if self.win_size > 0:
max_d = torch.argmax(x, dim=1, keepdim=True)
d_value = []
prob_value = []
for d in range(-self.win_size, self.win_size + 1):
index = max_d + d
index[index < 0] = 0
index[index > x.shape[1] - 1] = x.shape[1] - 1
d_value.append(index)
prob = torch.gather(x, dim=1, index=index)
prob_value.append(prob)
part_x = torch.cat(prob_value, dim=1)
part_x = part_x / (torch.sum(part_x, dim=1, keepdim=True) + 1e-08)
part_d = torch.cat(d_value, dim=1).float()
out = torch.sum(part_x * part_d, dim=1)
else:
out = torch.sum(x * disp, 1)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Message_Passing_Unit_v1 | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/oc/coczp7rkil43vyoaywnm6ib3ifhjmr5ddfinfrw2suxqd75btpch.py
# Topologically Sorted Source Nodes: [gate, gate_1], Original ATen: [aten.cat, aten.relu]
# Source node to ATen node mapping:
# gate => cat
# gate_1 => relu
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_1, %primals_2], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%cat,), kwargs = {})
triton_poi_fused_cat_relu_0 = async_compile.triton('triton_poi_fused_cat_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_relu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_relu_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = (xindex // 8)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((4*x1) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tl.load(in_ptr1 + ((4*x1) + ((-4) + x0)), tmp6 & xmask, eviction_policy='evict_last', other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tmp11 = tl.full([1], 0, tl.int32)
tmp12 = triton_helpers.maximum(tmp11, tmp10)
tl.store(out_ptr0 + (x2), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/rg/crgu7axt3pkodcuj44cx7l5wvaq2wktx2n3wt4l46yox7manytzy.py
# Topologically Sorted Source Nodes: [sigmoid, gate_2], Original ATen: [aten.sigmoid, aten.mean]
# Source node to ATen node mapping:
# gate_2 => mean
# sigmoid => sigmoid
# Graph fragment:
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%addmm,), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sigmoid, [1]), kwargs = {})
triton_per_fused_mean_sigmoid_1 = async_compile.triton('triton_per_fused_mean_sigmoid_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 128],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_sigmoid_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_sigmoid_1(in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 128
RBLOCK: tl.constexpr = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (128*x0)), xmask, other=0.0)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp5, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/o4/co4bcog5oirylpafkr3bzmpf7tpkw52rybp2isvs75jt6osdlvhc.py
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# output => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %expand), kwargs = {})
triton_poi_fused_mul_2 = async_compile.triton('triton_poi_fused_mul_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = 128.0
tmp3 = tmp1 / tmp2
tmp4 = tmp0 * tmp3
tl.store(out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (128, 8), (8, 1))
assert_size_stride(primals_4, (128, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [gate, gate_1], Original ATen: [aten.cat, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_relu_0.run(primals_1, primals_2, buf0, 32, grid=grid(32), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, buf0, reinterpret_tensor(primals_3, (8, 128), (1, 8), 0), alpha=1, beta=1, out=buf1)
del primals_3
del primals_4
buf2 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [sigmoid, gate_2], Original ATen: [aten.sigmoid, aten.mean]
triton_per_fused_mean_sigmoid_1.run(buf1, buf2, 4, 128, grid=grid(4), stream=stream0)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.mul]
triton_poi_fused_mul_2.run(primals_2, buf2, buf3, 16, grid=grid(16), stream=stream0)
del buf2
return (buf3, primals_2, buf0, buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((128, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torchvision.transforms import functional as F
import torch.utils.data
from torch import nn
import torch.nn.functional as F
class Message_Passing_Unit_v1(nn.Module):
def __init__(self, fea_size, filter_size=128):
super(Message_Passing_Unit_v1, self).__init__()
self.w = nn.Linear(fea_size * 2, filter_size, bias=True)
self.fea_size = fea_size
self.filter_size = filter_size
def forward(self, unary_term, pair_term):
if unary_term.size()[0] == 1 and pair_term.size()[0] > 1:
unary_term = unary_term.expand(pair_term.size()[0], unary_term.
size()[1])
if unary_term.size()[0] > 1 and pair_term.size()[0] == 1:
pair_term = pair_term.expand(unary_term.size()[0], pair_term.
size()[1])
gate = torch.cat([unary_term, pair_term], 1)
gate = F.relu(gate)
gate = torch.sigmoid(self.w(gate)).mean(1)
output = pair_term * gate.view(-1, 1).expand(gate.size()[0],
pair_term.size()[1])
return output
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'fea_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.utils.data
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_cat_relu_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 32
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = xindex // 8
x2 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 * x1 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp9 = tl.load(in_ptr1 + (4 * x1 + (-4 + x0)), tmp6 & xmask,
eviction_policy='evict_last', other=0.0)
tmp10 = tl.where(tmp4, tmp5, tmp9)
tmp11 = tl.full([1], 0, tl.int32)
tmp12 = triton_helpers.maximum(tmp11, tmp10)
tl.store(out_ptr0 + x2, tmp12, xmask)
@triton.jit
def triton_per_fused_mean_sigmoid_1(in_ptr0, out_ptr0, xnumel, rnumel,
XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 128 * x0), xmask, other=0.0)
tmp1 = tl.sigmoid(tmp0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tl.store(out_ptr0 + x0, tmp5, xmask)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = 128.0
tmp3 = tmp1 / tmp2
tmp4 = tmp0 * tmp3
tl.store(out_ptr0 + x2, tmp4, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (128, 8), (8, 1))
assert_size_stride(primals_4, (128,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_relu_0[grid(32)](primals_1, primals_2, buf0,
32, XBLOCK=32, num_warps=1, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
extern_kernels.addmm(primals_4, buf0, reinterpret_tensor(primals_3,
(8, 128), (1, 8), 0), alpha=1, beta=1, out=buf1)
del primals_3
del primals_4
buf2 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_mean_sigmoid_1[grid(4)](buf1, buf2, 4, 128, XBLOCK
=1, num_warps=2, num_stages=1)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_mul_2[grid(16)](primals_2, buf2, buf3, 16, XBLOCK=
16, num_warps=1, num_stages=1)
del buf2
return buf3, primals_2, buf0, buf1
class Message_Passing_Unit_v1New(nn.Module):
def __init__(self, fea_size, filter_size=128):
super(Message_Passing_Unit_v1New, self).__init__()
self.w = nn.Linear(fea_size * 2, filter_size, bias=True)
self.fea_size = fea_size
self.filter_size = filter_size
def forward(self, input_0, input_1):
primals_3 = self.w.weight
primals_4 = self.w.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| SpartaG117/scene_graph_benchmark | Message_Passing_Unit_v1 | false | 1,095 | [
"MIT"
] | 0 | e2e49940dd2f752b1faf9ae26707435ba3441bcb | https://github.com/SpartaG117/scene_graph_benchmark/tree/e2e49940dd2f752b1faf9ae26707435ba3441bcb | import torch
from torchvision.transforms import functional as F
import torch.utils.data
from torch import nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, fea_size, filter_size=128):
super().__init__()
self.w = nn.Linear(fea_size * 2, filter_size, bias=True)
self.fea_size = fea_size
self.filter_size = filter_size
def forward(self, unary_term, pair_term):
if unary_term.size()[0] == 1 and pair_term.size()[0] > 1:
unary_term = unary_term.expand(pair_term.size()[0], unary_term.
size()[1])
if unary_term.size()[0] > 1 and pair_term.size()[0] == 1:
pair_term = pair_term.expand(unary_term.size()[0], pair_term.
size()[1])
gate = torch.cat([unary_term, pair_term], 1)
gate = F.relu(gate)
gate = torch.sigmoid(self.w(gate)).mean(1)
output = pair_term * gate.view(-1, 1).expand(gate.size()[0],
pair_term.size()[1])
return output
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [4]
|
ExpModule | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/55/c554z2kehw4y75kmgw5gvu2rebcutcxsad4kke4bln6vvbihauog.py
# Topologically Sorted Source Nodes: [exp], Original ATen: [aten.exp]
# Source node to ATen node mapping:
# exp => exp
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%arg0_1,), kwargs = {})
triton_poi_fused_exp_0 = async_compile.triton('triton_poi_fused_exp_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_exp_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_exp_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl_math.exp(tmp0)
tl.store(out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [exp], Original ATen: [aten.exp]
stream0 = get_raw_stream(0)
triton_poi_fused_exp_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class ExpModule(nn.Module):
def __init__(self):
super(ExpModule, self).__init__()
def forward(self, x):
return torch.exp(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_exp_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl_math.exp(tmp0)
tl.store(out_ptr0 + x0, tmp1, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_exp_0[grid(256)](arg0_1, buf0, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class ExpModuleNew(nn.Module):
def __init__(self):
super(ExpModuleNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| SimonTreu/sdvae | ExpModule | false | 1,096 | [
"MIT"
] | 0 | e0270b9b2acf2d66eec93870f1c5633c8f04d9ab | https://github.com/SimonTreu/sdvae/tree/e0270b9b2acf2d66eec93870f1c5633c8f04d9ab | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self):
super().__init__()
def forward(self, x):
return torch.exp(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
EncoderLayer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/fg/cfg742icmosiwp5ugziye26din5ueqx3v7ntptkkpyackudldrxs.py
# Topologically Sorted Source Nodes: [eq], Original ATen: [aten.eq]
# Source node to ATen node mapping:
# eq => eq
# Graph fragment:
# %eq : [num_users=2] = call_function[target=torch.ops.aten.eq.Scalar](args = (%primals_8, 0), kwargs = {})
triton_poi_fused_eq_0 = async_compile.triton('triton_poi_fused_eq_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i1', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_eq_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_eq_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.0
tmp2 = tmp0 == tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/g7/cg74ay746yb3ivb67q6pz3e2xibwamviqoisgmyzopmnr47hrosh.py
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
# Source node to ATen node mapping:
# attention_score_1 => div
# attention_score_2 => full_default, where
# attention_score_3 => amax, exp, sub, sum_1
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_11, 1.0), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -1000000000.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %div), kwargs = {})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%where, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
triton_poi_fused__softmax_div_masked_fill_1 = async_compile.triton('triton_poi_fused__softmax_div_masked_fill_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*i1', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_div_masked_fill_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (4*x2), xmask, eviction_policy='evict_last').to(tl.int1)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr2 + (4*x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (1 + (4*x2)), xmask, eviction_policy='evict_last').to(tl.int1)
tmp9 = tl.load(in_ptr2 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (2 + (4*x2)), xmask, eviction_policy='evict_last').to(tl.int1)
tmp15 = tl.load(in_ptr2 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr0 + (3 + (4*x2)), xmask, eviction_policy='evict_last').to(tl.int1)
tmp21 = tl.load(in_ptr2 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp10 = tmp1 * tmp9
tmp11 = tmp10 * tmp4
tmp12 = tl.where(tmp8, tmp6, tmp11)
tmp13 = triton_helpers.maximum(tmp7, tmp12)
tmp16 = tmp1 * tmp15
tmp17 = tmp16 * tmp4
tmp18 = tl.where(tmp14, tmp6, tmp17)
tmp19 = triton_helpers.maximum(tmp13, tmp18)
tmp22 = tmp1 * tmp21
tmp23 = tmp22 * tmp4
tmp24 = tl.where(tmp20, tmp6, tmp23)
tmp25 = triton_helpers.maximum(tmp19, tmp24)
tmp26 = tmp7 - tmp25
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp12 - tmp25
tmp29 = tl_math.exp(tmp28)
tmp30 = tmp27 + tmp29
tmp31 = tmp18 - tmp25
tmp32 = tl_math.exp(tmp31)
tmp33 = tmp30 + tmp32
tmp34 = tmp24 - tmp25
tmp35 = tl_math.exp(tmp34)
tmp36 = tmp33 + tmp35
tl.store(out_ptr0 + (x2), tmp25, xmask)
tl.store(out_ptr1 + (x2), tmp36, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/na/cna5revtaqihb3t6bnp24kdg6eqircycsi7slr6uzmhkwpelq6x7.py
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
# Source node to ATen node mapping:
# attention_score_1 => div
# attention_score_2 => full_default, where
# attention_score_3 => div_1, exp, sub
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_11, 1.0), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -1000000000.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %div), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %div_1 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_div_masked_fill_2 = async_compile.triton('triton_poi_fused__softmax_div_masked_fill_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*i1', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_div_masked_fill_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x4 = (xindex // 4)
x0 = xindex % 4
x2 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x3), xmask).to(tl.int1)
tmp1 = tl.load(in_ptr1 + (x4), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr3 + (x4), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr4 + (x4), xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp9 = tmp7 - tmp8
tmp10 = tl_math.exp(tmp9)
tmp12 = tmp10 / tmp11
tl.store(out_ptr0 + (x3), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/tc/ctcugqu2nbqlxcf2thnspnnypxifbalbzmclmutd5vaxdes2oyyk.py
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# layer_norm => var_mean
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %primals_1), kwargs = {})
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add, [2]), kwargs = {correction: 0, keepdim: True})
triton_poi_fused_add_native_layer_norm_3 = async_compile.triton('triton_poi_fused_add_native_layer_norm_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + (x0), tmp16, xmask)
tl.store(out_ptr1 + (x0), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ua/cuaquah4oaz43nhi25wixnpzlhvf2zfzdiezhmmwkuy5wfhtw6z4.py
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# layer_norm => add_1, add_2, mul_1, mul_2, rsqrt, sub_1
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %primals_1), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1e-05), kwargs = {})
# %rsqrt : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %getitem_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %rsqrt), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_1, %primals_11), kwargs = {})
# %add_2 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %primals_12), kwargs = {})
triton_poi_fused_add_native_layer_norm_4 = async_compile.triton('triton_poi_fused_add_native_layer_norm_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_native_layer_norm_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 6, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_native_layer_norm_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + (x2), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/va/cvayouropyisaprtjrhemadbdvsels72axdjsrgmbayknhu335yc.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_19,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_5 = async_compile.triton('triton_poi_fused_relu_threshold_backward_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_5(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/c3/cc3wjckmv52z5p6lagnrhsfwt53rzdfhvzlxkm5tgkwbs3kuzwax.py
# Topologically Sorted Source Nodes: [add_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# add_1 => add_3
# Graph fragment:
# %add_3 : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_21, %add_2), kwargs = {})
triton_poi_fused_add_6 = async_compile.triton('triton_poi_fused_add_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_6', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_6(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ly/clyngaczex3aq23lq4vvbajq6ig7vbayovzffexoltgvkl4mnmwv.py
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# layer_norm_1 => add_4, rsqrt_1, var_mean_1
# Graph fragment:
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add_3, [2]), kwargs = {correction: 0, keepdim: True})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-05), kwargs = {})
# %rsqrt_1 : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_4,), kwargs = {})
triton_poi_fused_native_layer_norm_7 = async_compile.triton('triton_poi_fused_native_layer_norm_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_7', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_7(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + (x0), tmp8, xmask)
tl.store(out_ptr1 + (x0), tmp23, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ev/cevwu4ha52ebsczlu4ai4awtkxr6r5mjdoyil2av5baz2emhshoo.py
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
# Source node to ATen node mapping:
# layer_norm_1 => add_4, add_5, mul_3, mul_4, rsqrt_1, sub_2, var_mean_1
# Graph fragment:
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%add_3, [2]), kwargs = {correction: 0, keepdim: True})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_2, 1e-05), kwargs = {})
# %rsqrt_1 : [num_users=1] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_4,), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_3, %getitem_3), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_2, %rsqrt_1), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_3, %primals_17), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_4, %primals_18), kwargs = {})
triton_poi_fused_native_layer_norm_8 = async_compile.triton('triton_poi_fused_native_layer_norm_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_native_layer_norm_8', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_native_layer_norm_8(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, ), (1, ))
assert_size_stride(primals_12, (4, ), (1, ))
assert_size_stride(primals_13, (4, 4), (4, 1))
assert_size_stride(primals_14, (4, ), (1, ))
assert_size_stride(primals_15, (4, 4), (4, 1))
assert_size_stride(primals_16, (4, ), (1, ))
assert_size_stride(primals_17, (4, ), (1, ))
assert_size_stride(primals_18, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_query], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_key], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prj_value], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_6
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [eq], Original ATen: [aten.eq]
stream0 = get_raw_stream(0)
triton_poi_fused_eq_0.run(primals_8, buf3, 256, grid=grid(256), stream=stream0)
del primals_8
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
triton_poi_fused__softmax_div_masked_fill_1.run(buf3, buf0, buf1, buf4, buf5, 64, grid=grid(64), stream=stream0)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attention_score_1, attention_score_2, attention_score_3], Original ATen: [aten.div, aten.masked_fill, aten._softmax]
triton_poi_fused__softmax_div_masked_fill_2.run(buf3, buf0, buf1, buf4, buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [result], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf6, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0), out=buf7)
buf8 = reinterpret_tensor(buf4, (16, 4), (4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [output_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_10, reinterpret_tensor(buf7, (16, 4), (4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf8)
del primals_10
buf9 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
buf10 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_3.run(buf8, primals_1, buf9, buf10, 16, grid=grid(16), stream=stream0)
buf11 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, layer_norm], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_4.run(buf8, primals_1, buf9, buf10, primals_11, primals_12, buf11, 64, grid=grid(64), stream=stream0)
del primals_12
buf12 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(primals_13, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0); del buf12 # reuse
buf19 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_5.run(buf13, primals_14, buf19, 64, grid=grid(64), stream=stream0)
del primals_14
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf13, (16, 4), (4, 1), 0), reinterpret_tensor(primals_15, (4, 4), (1, 4), 0), out=buf14)
buf15 = reinterpret_tensor(buf14, (4, 4, 4), (16, 4, 1), 0); del buf14 # reuse
# Topologically Sorted Source Nodes: [add_1], Original ATen: [aten.add]
triton_poi_fused_add_6.run(buf15, primals_16, buf11, 64, grid=grid(64), stream=stream0)
del primals_16
buf16 = buf9; del buf9 # reuse
buf17 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_7.run(buf15, buf16, buf17, 16, grid=grid(16), stream=stream0)
buf18 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [layer_norm_1], Original ATen: [aten.native_layer_norm]
triton_poi_fused_native_layer_norm_8.run(buf15, buf16, buf17, primals_17, primals_18, buf18, 64, grid=grid(64), stream=stream0)
del buf16
del buf17
del primals_18
return (buf18, primals_1, primals_11, primals_17, buf0, buf1, buf3, buf6, reinterpret_tensor(buf7, (16, 4), (4, 1), 0), buf8, reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(buf13, (16, 4), (4, 1), 0), buf15, primals_15, buf19, primals_13, primals_9, reinterpret_tensor(buf2, (16, 1, 4), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super(MultiHeadAttention, self).__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
class FeedForward(nn.Module):
def __init__(self, model_dim, hidden_dim, drop_prob):
super(FeedForward, self).__init__()
self.model_dim = model_dim
self.hidden_dim = hidden_dim
self.drop_prob = drop_prob
self.linearlayer1 = nn.Linear(model_dim, hidden_dim)
self.linearlayer2 = nn.Linear(hidden_dim, model_dim)
self.relu = nn.ReLU()
self.dropout = nn.Dropout(drop_prob)
def forward(self, tensor):
tensor = self.dropout(self.relu(self.linearlayer1(tensor)))
return self.linearlayer2(tensor)
class EncoderLayer(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, hidden_dim, num_head,
drop_prob):
super(EncoderLayer, self).__init__()
self.attention = MultiHeadAttention(model_dim, key_dim, value_dim,
num_head)
self.normalization1 = nn.LayerNorm(model_dim)
self.dropout1 = nn.Dropout(drop_prob)
self.ffn = FeedForward(model_dim, hidden_dim, drop_prob)
self.normalization2 = nn.LayerNorm(model_dim)
self.dropout2 = nn.Dropout(drop_prob)
def forward(self, tensor, source_mask):
residual = tensor
tensor = self.attention(query=tensor, key=tensor, value=tensor,
mask=source_mask)
tensor = self.dropout1(self.normalization1(tensor + residual))
residual = tensor
tensor = self.ffn(tensor)
tensor = self.dropout2(self.normalization2(tensor + residual))
return tensor
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'model_dim': 4, 'key_dim': 4, 'value_dim': 4, 'hidden_dim':
4, 'num_head': 4, 'drop_prob': 0.5}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import math
import torch.nn as nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_eq_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.0
tmp2 = tmp0 == tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_1(in_ptr0, in_ptr1, in_ptr2,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + 4 * x2, xmask, eviction_policy='evict_last').to(tl
.int1)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr2 + 4 * x1, xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (1 + 4 * x2), xmask, eviction_policy='evict_last'
).to(tl.int1)
tmp9 = tl.load(in_ptr2 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (2 + 4 * x2), xmask, eviction_policy='evict_last'
).to(tl.int1)
tmp15 = tl.load(in_ptr2 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp20 = tl.load(in_ptr0 + (3 + 4 * x2), xmask, eviction_policy='evict_last'
).to(tl.int1)
tmp21 = tl.load(in_ptr2 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp10 = tmp1 * tmp9
tmp11 = tmp10 * tmp4
tmp12 = tl.where(tmp8, tmp6, tmp11)
tmp13 = triton_helpers.maximum(tmp7, tmp12)
tmp16 = tmp1 * tmp15
tmp17 = tmp16 * tmp4
tmp18 = tl.where(tmp14, tmp6, tmp17)
tmp19 = triton_helpers.maximum(tmp13, tmp18)
tmp22 = tmp1 * tmp21
tmp23 = tmp22 * tmp4
tmp24 = tl.where(tmp20, tmp6, tmp23)
tmp25 = triton_helpers.maximum(tmp19, tmp24)
tmp26 = tmp7 - tmp25
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp12 - tmp25
tmp29 = tl_math.exp(tmp28)
tmp30 = tmp27 + tmp29
tmp31 = tmp18 - tmp25
tmp32 = tl_math.exp(tmp31)
tmp33 = tmp30 + tmp32
tmp34 = tmp24 - tmp25
tmp35 = tl_math.exp(tmp34)
tmp36 = tmp33 + tmp35
tl.store(out_ptr0 + x2, tmp25, xmask)
tl.store(out_ptr1 + x2, tmp36, xmask)
@triton.jit
def triton_poi_fused__softmax_div_masked_fill_2(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x4 = xindex // 4
x0 = xindex % 4
x2 = xindex // 16
tmp0 = tl.load(in_ptr0 + x3, xmask).to(tl.int1)
tmp1 = tl.load(in_ptr1 + x4, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp8 = tl.load(in_ptr3 + x4, xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr4 + x4, xmask, eviction_policy='evict_last')
tmp3 = tmp1 * tmp2
tmp4 = 1.0
tmp5 = tmp3 * tmp4
tmp6 = -1000000000.0
tmp7 = tl.where(tmp0, tmp6, tmp5)
tmp9 = tmp7 - tmp8
tmp10 = tl_math.exp(tmp9)
tmp12 = tmp10 / tmp11
tl.store(out_ptr0 + x3, tmp12, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_3(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp12 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp9 = tmp7 + tmp8
tmp10 = tmp6 + tmp9
tmp13 = tmp11 + tmp12
tmp14 = tmp10 + tmp13
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp2 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp5 - tmp16
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = tmp9 - tmp16
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp25 = tmp13 - tmp16
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = tmp27 / tmp15
tl.store(out_ptr0 + x0, tmp16, xmask)
tl.store(out_ptr1 + x0, tmp28, xmask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_4(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, in_ptr5, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr5 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.rsqrt(tmp7)
tmp9 = tmp4 * tmp8
tmp11 = tmp9 * tmp10
tmp13 = tmp11 + tmp12
tl.store(out_ptr0 + x2, tmp13, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_5(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
@triton.jit
def triton_poi_fused_add_6(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_native_layer_norm_7(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = tmp19 / tmp7
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tl.store(out_ptr0 + x0, tmp8, xmask)
tl.store(out_ptr1 + x0, tmp23, xmask)
@triton.jit
def triton_poi_fused_native_layer_norm_8(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
in_ptr4, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp6 = tmp4 * tmp5
tmp8 = tmp6 + tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17, primals_18
) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_9, (4, 4), (4, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4,), (1,))
assert_size_stride(primals_12, (4,), (1,))
assert_size_stride(primals_13, (4, 4), (4, 1))
assert_size_stride(primals_14, (4,), (1,))
assert_size_stride(primals_15, (4, 4), (4, 1))
assert_size_stride(primals_16, (4,), (1,))
assert_size_stride(primals_17, (4,), (1,))
assert_size_stride(primals_18, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(primals_1, (16,
4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(primals_1, (16,
4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf1)
del primals_4
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(primals_1, (16,
4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf2)
del primals_6
del primals_7
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_eq_0[grid(256)](primals_8, buf3, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_8
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_div_masked_fill_1[grid(64)](buf3, buf0,
buf1, buf4, buf5, 64, XBLOCK=64, num_warps=1, num_stages=1)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_div_masked_fill_2[grid(256)](buf3, buf0,
buf1, buf4, buf5, buf6, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf7 = reinterpret_tensor(buf5, (16, 4, 1), (4, 1, 1), 0)
del buf5
extern_kernels.bmm(reinterpret_tensor(buf6, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0), out=buf7)
buf8 = reinterpret_tensor(buf4, (16, 4), (4, 1), 0)
del buf4
extern_kernels.addmm(primals_10, reinterpret_tensor(buf7, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_9, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf8)
del primals_10
buf9 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
buf10 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
triton_poi_fused_add_native_layer_norm_3[grid(16)](buf8, primals_1,
buf9, buf10, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf11 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_native_layer_norm_4[grid(64)](buf8, primals_1,
buf9, buf10, primals_11, primals_12, buf11, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_12
buf12 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf11, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_13, (4, 4), (1, 4), 0), out=buf12)
buf13 = reinterpret_tensor(buf12, (4, 4, 4), (16, 4, 1), 0)
del buf12
buf19 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_5[grid(64)](buf13,
primals_14, buf19, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_14
buf14 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf13, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_15, (4, 4), (1, 4), 0), out=buf14)
buf15 = reinterpret_tensor(buf14, (4, 4, 4), (16, 4, 1), 0)
del buf14
triton_poi_fused_add_6[grid(64)](buf15, primals_16, buf11, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del primals_16
buf16 = buf9
del buf9
buf17 = buf10
del buf10
triton_poi_fused_native_layer_norm_7[grid(16)](buf15, buf16, buf17,
16, XBLOCK=16, num_warps=1, num_stages=1)
buf18 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_native_layer_norm_8[grid(64)](buf15, buf16, buf17,
primals_17, primals_18, buf18, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf16
del buf17
del primals_18
return (buf18, primals_1, primals_11, primals_17, buf0, buf1, buf3,
buf6, reinterpret_tensor(buf7, (16, 4), (4, 1), 0), buf8,
reinterpret_tensor(buf11, (16, 4), (4, 1), 0), reinterpret_tensor(
buf13, (16, 4), (4, 1), 0), buf15, primals_15, buf19, primals_13,
primals_9, reinterpret_tensor(buf2, (16, 1, 4), (4, 1, 1), 0))
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super(ScaledDotProductAttention, self).__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super(MultiHeadAttention, self).__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
class FeedForward(nn.Module):
def __init__(self, model_dim, hidden_dim, drop_prob):
super(FeedForward, self).__init__()
self.model_dim = model_dim
self.hidden_dim = hidden_dim
self.drop_prob = drop_prob
self.linearlayer1 = nn.Linear(model_dim, hidden_dim)
self.linearlayer2 = nn.Linear(hidden_dim, model_dim)
self.relu = nn.ReLU()
self.dropout = nn.Dropout(drop_prob)
def forward(self, tensor):
tensor = self.dropout(self.relu(self.linearlayer1(tensor)))
return self.linearlayer2(tensor)
class EncoderLayerNew(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, hidden_dim, num_head,
drop_prob):
super(EncoderLayerNew, self).__init__()
self.attention = MultiHeadAttention(model_dim, key_dim, value_dim,
num_head)
self.normalization1 = nn.LayerNorm(model_dim)
self.dropout1 = nn.Dropout(drop_prob)
self.ffn = FeedForward(model_dim, hidden_dim, drop_prob)
self.normalization2 = nn.LayerNorm(model_dim)
self.dropout2 = nn.Dropout(drop_prob)
def forward(self, input_0, input_1):
primals_2 = self.attention.Wq.weight
primals_3 = self.attention.Wq.bias
primals_4 = self.attention.Wk.weight
primals_5 = self.attention.Wk.bias
primals_6 = self.attention.Wv.weight
primals_7 = self.attention.Wv.bias
primals_9 = self.attention.Wo.weight
primals_10 = self.attention.Wo.bias
primals_11 = self.normalization1.weight
primals_12 = self.normalization1.bias
primals_13 = self.ffn.linearlayer1.weight
primals_14 = self.ffn.linearlayer1.bias
primals_15 = self.ffn.linearlayer2.weight
primals_16 = self.ffn.linearlayer2.bias
primals_17 = self.normalization2.weight
primals_18 = self.normalization2.bias
primals_1 = input_0
primals_8 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18])
return output[0]
| SeungoneKim/Transformer_implementation | EncoderLayer | false | 1,097 | [
"Apache-2.0"
] | 0 | a52bf552eb645fc9bfb812cc26842fc147d6c008 | https://github.com/SeungoneKim/Transformer_implementation/tree/a52bf552eb645fc9bfb812cc26842fc147d6c008 | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class ScaledDotProductAttention(nn.Module):
def __init__(self):
super().__init__()
def forward(self, query, key, value, mask=None):
_1, _2, query_sequence_length, _3 = query.size()
batch_size, num_head, key_sequence_length, size_per_head = key.size()
query = query.view(batch_size, num_head, query_sequence_length,
size_per_head)
key = key.view(batch_size, num_head, size_per_head, key_sequence_length
)
attention_score = torch.einsum('abcd, abde -> abce', query, key)
attention_score = attention_score / math.sqrt(size_per_head)
if mask is not None:
attention_score = attention_score.masked_fill(mask == 0, -
1000000000.0)
attention_score = F.softmax(attention_score, dim=-1)
result = attention_score @ value
return result, attention_score
class MultiHeadAttention(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, num_head):
super().__init__()
self.model_dim = model_dim
self.key_dim = key_dim
self.value_dim = value_dim
self.num_head = num_head
self.Wq = nn.Linear(model_dim, key_dim)
self.Wk = nn.Linear(model_dim, key_dim)
self.Wv = nn.Linear(model_dim, value_dim)
self.attention = ScaledDotProductAttention()
self.Wo = nn.Linear(value_dim, model_dim)
def forward(self, query, key, value, mask=None):
prj_query = self.Wq(query)
prj_key = self.Wk(key)
prj_value = self.Wv(value)
multihead_query = self.multihead_split(prj_query)
multihead_key = self.multihead_split(prj_key)
multihead_value = self.multihead_split(prj_value)
attention_output, _attention_score = self.attention(multihead_query,
multihead_key, multihead_value, mask=mask)
output = self.multihead_concat(attention_output)
output = self.Wo(output)
return output
def multihead_split(self, tensor):
batch_size, sequence_length, hidden_size = tensor.size()
size_per_head = hidden_size // self.num_head
return tensor.view(batch_size, self.num_head, sequence_length,
size_per_head)
def multihead_concat(self, tensor):
batch_size, num_head, sequence_length, size_per_head = tensor.size()
hidden_size = num_head * size_per_head
return tensor.view(batch_size, sequence_length, hidden_size)
class FeedForward(nn.Module):
def __init__(self, model_dim, hidden_dim, drop_prob):
super().__init__()
self.model_dim = model_dim
self.hidden_dim = hidden_dim
self.drop_prob = drop_prob
self.linearlayer1 = nn.Linear(model_dim, hidden_dim)
self.linearlayer2 = nn.Linear(hidden_dim, model_dim)
self.relu = nn.ReLU()
self.dropout = nn.Dropout(drop_prob)
def forward(self, tensor):
tensor = self.dropout(self.relu(self.linearlayer1(tensor)))
return self.linearlayer2(tensor)
class Model(nn.Module):
def __init__(self, model_dim, key_dim, value_dim, hidden_dim, num_head,
drop_prob):
super().__init__()
self.attention = MultiHeadAttention(model_dim, key_dim, value_dim,
num_head)
self.normalization1 = nn.LayerNorm(model_dim)
self.dropout1 = nn.Dropout(drop_prob)
self.ffn = FeedForward(model_dim, hidden_dim, drop_prob)
self.normalization2 = nn.LayerNorm(model_dim)
self.dropout2 = nn.Dropout(drop_prob)
def forward(self, tensor, source_mask):
residual = tensor
tensor = self.attention(query=tensor, key=tensor, value=tensor,
mask=source_mask)
tensor = self.dropout1(self.normalization1(tensor + residual))
residual = tensor
tensor = self.ffn(tensor)
tensor = self.dropout2(self.normalization2(tensor + residual))
# ... truncated (>4000 chars) for memory efficiency |
Residual_Covolution | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/td/ctdybbibnws4d7ukbk3fpn35zkgapxylowdhzwx7vgsllncbdrxa.py
# Topologically Sorted Source Nodes: [dow1, dow1_1], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# dow1 => convolution
# dow1_1 => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [12, 12], [12, 12], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=3] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_0 = async_compile.triton('triton_poi_fused_convolution_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/32/c32v7egt4mupqssam3gmac2qgv3ujprjybthsgweflmot256qqw7.py
# Topologically Sorted Source Nodes: [seg], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# seg => convolution_1
# Graph fragment:
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_4, %primals_5, [1, 1], [12, 12], [12, 12], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/tr/ctr6ssoy3ohhjpgpoxiqv2ojm7yzgqk5hgm4lemyjegkqr4ohtne.py
# Topologically Sorted Source Nodes: [inc1, relu_1, add1], Original ATen: [aten.convolution, aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# add1 => add
# inc1 => convolution_2
# relu_1 => relu_1
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%convolution_1, %primals_6, %primals_7, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_2,), kwargs = {})
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%relu, %relu_1), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_add_convolution_relu_threshold_backward_2 = async_compile.triton('triton_poi_fused_add_convolution_relu_threshold_backward_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_convolution_relu_threshold_backward_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_convolution_relu_threshold_backward_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr1 + (x3), xmask)
tmp2 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + (x3), tmp6, xmask)
tl.store(out_ptr1 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_5, (4, ), (1, ))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_9, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [dow1], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(12, 12), dilation=(12, 12), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [dow1, dow1_1], Original ATen: [aten.convolution, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_0.run(buf1, primals_2, 256, grid=grid(256), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [seg], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, primals_4, stride=(1, 1), padding=(12, 12), dilation=(12, 12), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [seg], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf3, primals_5, 256, grid=grid(256), stream=stream0)
del primals_5
# Topologically Sorted Source Nodes: [inc1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 4, 4, 4), (64, 16, 4, 1))
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [inc1, relu_1, add1], Original ATen: [aten.convolution, aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_convolution_relu_threshold_backward_2.run(buf1, buf4, primals_7, buf5, buf9, 256, grid=grid(256), stream=stream0)
del primals_7
# Topologically Sorted Source Nodes: [inc2], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 4, 4, 4), (64, 16, 4, 1))
buf7 = buf4; del buf4 # reuse
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [inc2, relu_2, out], Original ATen: [aten.convolution, aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_convolution_relu_threshold_backward_2.run(primals_3, buf6, primals_9, buf7, buf8, 256, grid=grid(256), stream=stream0)
del buf6
del primals_9
return (buf7, buf3, primals_1, primals_3, primals_4, primals_6, primals_8, buf1, buf3, buf5, buf8, buf9, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class Residual_Covolution(nn.Module):
def __init__(self, icol, ocol, num_classes):
super(Residual_Covolution, self).__init__()
self.conv1 = nn.Conv2d(icol, ocol, kernel_size=3, stride=1, padding
=12, dilation=12, bias=True)
self.conv2 = nn.Conv2d(ocol, num_classes, kernel_size=3, stride=1,
padding=12, dilation=12, bias=True)
self.conv3 = nn.Conv2d(num_classes, ocol, kernel_size=1, stride=1,
padding=0, dilation=1, bias=True)
self.conv4 = nn.Conv2d(ocol, icol, kernel_size=1, stride=1, padding
=0, dilation=1, bias=True)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
dow1 = self.conv1(x)
dow1 = self.relu(dow1)
seg = self.conv2(dow1)
inc1 = self.conv3(seg)
add1 = dow1 + self.relu(inc1)
inc2 = self.conv4(add1)
out = x + self.relu(inc2)
return out, seg
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'icol': 4, 'ocol': 4, 'num_classes': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_convolution_relu_threshold_backward_2(in_ptr0,
in_ptr1, in_ptr2, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr1 + x3, xmask)
tmp2 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp4 = tl.full([1], 0, tl.int32)
tmp5 = triton_helpers.maximum(tmp4, tmp3)
tmp6 = tmp0 + tmp5
tmp7 = 0.0
tmp8 = tmp5 <= tmp7
tl.store(out_ptr0 + x3, tmp6, xmask)
tl.store(out_ptr1 + x3, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_5, (4,), (1,))
assert_size_stride(primals_6, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_9, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(12, 12), dilation=(12, 12), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_relu_0[grid(256)](buf1, primals_2, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = extern_kernels.convolution(buf1, primals_4, stride=(1, 1),
padding=(12, 12), dilation=(12, 12), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf2
del buf2
triton_poi_fused_convolution_1[grid(256)](buf3, primals_5, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 4, 4, 4), (64, 16, 4, 1))
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_add_convolution_relu_threshold_backward_2[grid(256)](
buf1, buf4, primals_7, buf5, buf9, 256, XBLOCK=256, num_warps=4,
num_stages=1)
del primals_7
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 4, 4, 4), (64, 16, 4, 1))
buf7 = buf4
del buf4
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_add_convolution_relu_threshold_backward_2[grid(256)](
primals_3, buf6, primals_9, buf7, buf8, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf6
del primals_9
return (buf7, buf3, primals_1, primals_3, primals_4, primals_6,
primals_8, buf1, buf3, buf5, buf8, buf9)
class Residual_CovolutionNew(nn.Module):
def __init__(self, icol, ocol, num_classes):
super(Residual_CovolutionNew, self).__init__()
self.conv1 = nn.Conv2d(icol, ocol, kernel_size=3, stride=1, padding
=12, dilation=12, bias=True)
self.conv2 = nn.Conv2d(ocol, num_classes, kernel_size=3, stride=1,
padding=12, dilation=12, bias=True)
self.conv3 = nn.Conv2d(num_classes, ocol, kernel_size=1, stride=1,
padding=0, dilation=1, bias=True)
self.conv4 = nn.Conv2d(ocol, icol, kernel_size=1, stride=1, padding
=0, dilation=1, bias=True)
self.relu = nn.ReLU(inplace=True)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_6 = self.conv3.weight
primals_7 = self.conv3.bias
primals_8 = self.conv4.weight
primals_9 = self.conv4.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9])
return output[0], output[1]
| SultanAbuGhazal/CGNet | Residual_Covolution | false | 1,098 | [
"MIT"
] | 0 | f10b976b984ba09be26b902ed4da97cd1311cf17 | https://github.com/SultanAbuGhazal/CGNet/tree/f10b976b984ba09be26b902ed4da97cd1311cf17 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, icol, ocol, num_classes):
super().__init__()
self.conv1 = nn.Conv2d(icol, ocol, kernel_size=3, stride=1, padding
=12, dilation=12, bias=True)
self.conv2 = nn.Conv2d(ocol, num_classes, kernel_size=3, stride=1,
padding=12, dilation=12, bias=True)
self.conv3 = nn.Conv2d(num_classes, ocol, kernel_size=1, stride=1,
padding=0, dilation=1, bias=True)
self.conv4 = nn.Conv2d(ocol, icol, kernel_size=1, stride=1, padding
=0, dilation=1, bias=True)
self.relu = nn.ReLU(inplace=True)
def forward(self, x):
dow1 = self.conv1(x)
dow1 = self.relu(dow1)
seg = self.conv2(dow1)
inc1 = self.conv3(seg)
add1 = dow1 + self.relu(inc1)
inc2 = self.conv4(add1)
out = x + self.relu(inc2)
return out, seg
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
ResidualBlockNoBN | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/td/ctdybbibnws4d7ukbk3fpn35zkgapxylowdhzwx7vgsllncbdrxa.py
# Topologically Sorted Source Nodes: [conv2d, out], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d => convolution
# out => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_0 = async_compile.triton('triton_poi_fused_convolution_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x3), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/l3/cl3kktfjbfxvoqsgvjon5fk5ycnqjp7n2a3dk3gk4gw4n2jfe25m.py
# Topologically Sorted Source Nodes: [out_1, out_2, out_3], Original ATen: [aten.convolution, aten.add, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# out_1 => convolution_1
# out_2 => add
# out_3 => relu_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_4, %primals_5, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convolution_1, %primals_3), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_add_convolution_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_add_convolution_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_convolution_relu_threshold_backward_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_convolution_relu_threshold_backward_1(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x3), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.full([1], 0, tl.int32)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = 0.0
tmp8 = tmp6 <= tmp7
tl.store(in_out_ptr0 + (x3), tmp6, xmask)
tl.store(out_ptr0 + (x3), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [conv2d, out], Original ATen: [aten.convolution, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_0.run(buf1, primals_2, 256, grid=grid(256), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf2; del buf2 # reuse
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [out_1, out_2, out_3], Original ATen: [aten.convolution, aten.add, aten.relu, aten.threshold_backward]
triton_poi_fused_add_convolution_relu_threshold_backward_1.run(buf3, primals_5, primals_3, buf4, 256, grid=grid(256), stream=stream0)
del primals_5
return (buf3, primals_1, primals_3, primals_4, buf1, buf4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
class ResidualBlockNoBN(nn.Module):
def __init__(self, in_channels, out_channels, stride=1):
super(ResidualBlockNoBN, self).__init__()
self.conv1 = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=(3, 3), stride=stride, padding=1,
bias=True)
self.conv2 = nn.Conv2d(in_channels=out_channels, out_channels=
out_channels, kernel_size=(3, 3), stride=1, padding=1, bias=True)
self.shortcut = nn.Sequential()
if stride != 1 or in_channels != out_channels:
self.shortcut = nn.Sequential(nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=(1, 1), stride=
stride, bias=False))
def forward(self, x):
out = nn.ReLU()(self.conv1(x))
out = self.conv2(out)
out += self.shortcut(x)
out = nn.ReLU()(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_convolution_relu_0(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x3, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_convolution_relu_threshold_backward_1(in_out_ptr0,
in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x3, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.full([1], 0, tl.int32)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = 0.0
tmp8 = tmp6 <= tmp7
tl.store(in_out_ptr0 + x3, tmp6, xmask)
tl.store(out_ptr0 + x3, tmp8, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_relu_0[grid(256)](buf1, primals_2, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = extern_kernels.convolution(buf1, primals_4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf2
del buf2
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_add_convolution_relu_threshold_backward_1[grid(256)](
buf3, primals_5, primals_3, buf4, 256, XBLOCK=256, num_warps=4,
num_stages=1)
del primals_5
return buf3, primals_1, primals_3, primals_4, buf1, buf4
class ResidualBlockNoBNNew(nn.Module):
def __init__(self, in_channels, out_channels, stride=1):
super(ResidualBlockNoBNNew, self).__init__()
self.conv1 = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=(3, 3), stride=stride, padding=1,
bias=True)
self.conv2 = nn.Conv2d(in_channels=out_channels, out_channels=
out_channels, kernel_size=(3, 3), stride=1, padding=1, bias=True)
self.shortcut = nn.Sequential()
if stride != 1 or in_channels != out_channels:
self.shortcut = nn.Sequential(nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=(1, 1), stride=
stride, bias=False))
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.conv2.weight
primals_5 = self.conv2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| Suvapna/ArtificialLaughter | ResidualBlockNoBN | false | 1,100 | [
"MIT"
] | 0 | a7114134b698f829e05e74cac30052e18b260f85 | https://github.com/Suvapna/ArtificialLaughter/tree/a7114134b698f829e05e74cac30052e18b260f85 | import torch
from torch import nn
class Model(nn.Module):
def __init__(self, in_channels, out_channels, stride=1):
super().__init__()
self.conv1 = nn.Conv2d(in_channels=in_channels, out_channels=
out_channels, kernel_size=(3, 3), stride=stride, padding=1,
bias=True)
self.conv2 = nn.Conv2d(in_channels=out_channels, out_channels=
out_channels, kernel_size=(3, 3), stride=1, padding=1, bias=True)
self.shortcut = nn.Sequential()
if stride != 1 or in_channels != out_channels:
self.shortcut = nn.Sequential(nn.Conv2d(in_channels=in_channels,
out_channels=out_channels, kernel_size=(1, 1), stride=
stride, bias=False))
def forward(self, x):
out = nn.ReLU()(self.conv1(x))
out = self.conv2(out)
out += self.shortcut(x)
out = nn.ReLU()(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
SpatialAttention | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/46/c46mg7rvdztu6n5oosf5c4if7ziag6obrxhwbn43lcdfibfuom7w.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# x => cat
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%mean, %getitem], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[128],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 16) % 2
x0 = xindex % 16
x2 = (xindex // 32)
x3 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (64*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp9 = tmp7 + tmp8
tmp10 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp11 = tmp9 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp4, tmp13, tmp14)
tmp16 = tmp0 >= tmp3
tmp17 = tl.full([1], 2, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tl.load(in_ptr0 + (x0 + (64*x2)), tmp16 & xmask, eviction_policy='evict_last', other=0.0)
tmp20 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), tmp16 & xmask, eviction_policy='evict_last', other=0.0)
tmp21 = triton_helpers.maximum(tmp19, tmp20)
tmp22 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), tmp16 & xmask, eviction_policy='evict_last', other=0.0)
tmp23 = triton_helpers.maximum(tmp21, tmp22)
tmp24 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), tmp16 & xmask, eviction_policy='evict_last', other=0.0)
tmp25 = triton_helpers.maximum(tmp23, tmp24)
tmp26 = tl.full(tmp25.shape, 0.0, tmp25.dtype)
tmp27 = tl.where(tmp16, tmp25, tmp26)
tmp28 = tl.where(tmp4, tmp15, tmp27)
tl.store(out_ptr0 + (x3), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/go/cgofqcgduqrtcjakfd7uk3wkcrpwsqxispluihwsstry6ekodk2u.py
# Topologically Sorted Source Nodes: [x_1, sigmoid], Original ATen: [aten.convolution, aten.sigmoid]
# Source node to ATen node mapping:
# sigmoid => sigmoid
# x_1 => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%cat, %primals_2, %primals_3, [1, 1], [3, 3], [1, 1], False, [0, 0], 1), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_sigmoid_1 = async_compile.triton('triton_poi_fused_convolution_sigmoid_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_sigmoid_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_sigmoid_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = tl.sigmoid(tmp3)
tl.store(in_out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, 2, 7, 7), (98, 49, 7, 1))
assert_size_stride(primals_3, (1, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 2, 4, 4), (32, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_1, buf0, 128, grid=grid(128), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(3, 3), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [x_1, sigmoid], Original ATen: [aten.convolution, aten.sigmoid]
triton_poi_fused_convolution_sigmoid_1.run(buf2, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
return (buf2, primals_2, buf0, buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, 2, 7, 7), (98, 49, 7, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class SpatialAttention(nn.Module):
def __init__(self, kernel_size=7, bias=True):
super(SpatialAttention, self).__init__()
assert kernel_size in (3, 7), 'kernel size must be 3 or 7'
padding = 3 if kernel_size == 7 else 1
self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=bias)
self.sigmoid = nn.Sigmoid()
def init_weighs(self):
normal_init(self.spatial_layer.conv1, std=0.01)
def forward(self, x):
avg_out = torch.mean(x, dim=1, keepdim=True)
max_out, _ = torch.max(x, dim=1, keepdim=True)
x = torch.cat([avg_out, max_out], dim=1)
x = self.conv1(x)
return self.sigmoid(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 16 % 2
x0 = xindex % 16
x2 = xindex // 32
x3 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 1, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 64 * x2), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), tmp4 & xmask,
eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), tmp4 & xmask,
eviction_policy='evict_last', other=0.0)
tmp9 = tmp7 + tmp8
tmp10 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), tmp4 & xmask,
eviction_policy='evict_last', other=0.0)
tmp11 = tmp9 + tmp10
tmp12 = 4.0
tmp13 = tmp11 / tmp12
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp4, tmp13, tmp14)
tmp16 = tmp0 >= tmp3
tl.full([1], 2, tl.int64)
tmp19 = tl.load(in_ptr0 + (x0 + 64 * x2), tmp16 & xmask,
eviction_policy='evict_last', other=0.0)
tmp20 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), tmp16 & xmask,
eviction_policy='evict_last', other=0.0)
tmp21 = triton_helpers.maximum(tmp19, tmp20)
tmp22 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), tmp16 & xmask,
eviction_policy='evict_last', other=0.0)
tmp23 = triton_helpers.maximum(tmp21, tmp22)
tmp24 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), tmp16 & xmask,
eviction_policy='evict_last', other=0.0)
tmp25 = triton_helpers.maximum(tmp23, tmp24)
tmp26 = tl.full(tmp25.shape, 0.0, tmp25.dtype)
tmp27 = tl.where(tmp16, tmp25, tmp26)
tmp28 = tl.where(tmp4, tmp15, tmp27)
tl.store(out_ptr0 + x3, tmp28, xmask)
@triton.jit
def triton_poi_fused_convolution_sigmoid_1(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp3 = tmp0 + tmp2
tmp4 = tl.sigmoid(tmp3)
tl.store(in_out_ptr0 + x0, tmp4, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (1, 2, 7, 7), (98, 49, 7, 1))
assert_size_stride(primals_3, (1,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 2, 4, 4), (32, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(128)](primals_1, buf0, 128, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(3, 3), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 1, 4, 4), (16, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_sigmoid_1[grid(64)](buf2, primals_3,
64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_3
return buf2, primals_2, buf0, buf2
class SpatialAttentionNew(nn.Module):
def __init__(self, kernel_size=7, bias=True):
super(SpatialAttentionNew, self).__init__()
assert kernel_size in (3, 7), 'kernel size must be 3 or 7'
padding = 3 if kernel_size == 7 else 1
self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=bias)
self.sigmoid = nn.Sigmoid()
def init_weighs(self):
normal_init(self.spatial_layer.conv1, std=0.01)
def forward(self, input_0):
primals_2 = self.conv1.weight
primals_3 = self.conv1.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| SuzaKrish/mmdetection | SpatialAttention | false | 1,101 | [
"Apache-2.0"
] | 0 | 31c16891d7493252262e738bcbf05326dba866b2 | https://github.com/SuzaKrish/mmdetection/tree/31c16891d7493252262e738bcbf05326dba866b2 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, kernel_size=7, bias=True):
super().__init__()
assert kernel_size in (3, 7), 'kernel size must be 3 or 7'
padding = 3 if kernel_size == 7 else 1
self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=bias)
self.sigmoid = nn.Sigmoid()
def init_weighs(self):
normal_init(self.spatial_layer.conv1, std=0.01)
def forward(self, x):
avg_out = torch.mean(x, dim=1, keepdim=True)
max_out, _ = torch.max(x, dim=1, keepdim=True)
x = torch.cat([avg_out, max_out], dim=1)
x = self.conv1(x)
return self.sigmoid(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Message_Passing_Unit_v2 | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/wt/cwtjehqlpgsottzm6ypzegm4s3jeoie6y6l4z5tnv4psexwmszst.py
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%primals_1,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/7e/c7ewtx3sknj4widkgiiygn7cetpykui7eqvnbrietgcfvo46wzhb.py
# Topologically Sorted Source Nodes: [gate, sum_1, gate_1], Original ATen: [aten.mul, aten.sum, aten.sigmoid, aten.sigmoid_backward]
# Source node to ATen node mapping:
# gate => mul
# gate_1 => sigmoid
# sum_1 => sum_1
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%addmm, %addmm_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
# %sigmoid : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%sum_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %sub), kwargs = {})
triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1 = async_compile.triton('triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 128],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 128
RBLOCK: tl.constexpr = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (128*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (r1 + (128*x0)), xmask, other=0.0)
tmp2 = tmp0 * tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(xmask, tmp3, 0)
tmp6 = tl.sum(tmp5, 1)[:, None]
tmp7 = tl.sigmoid(tmp6)
tmp8 = 1.0
tmp9 = tmp8 - tmp7
tmp10 = tmp7 * tmp9
tl.store(out_ptr1 + (x0), tmp10, xmask)
tl.store(out_ptr0 + (x0), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ts/ctsemlty4h3qjpwtckapy4evmw7oic6xfmvqzokqs4ma7l3irspn.py
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# output => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %expand), kwargs = {})
triton_poi_fused_mul_2 = async_compile.triton('triton_poi_fused_mul_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.sigmoid(tmp1)
tmp3 = tmp0 * tmp2
tl.store(out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (128, 4), (4, 1))
assert_size_stride(primals_4, (128, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(primals_1, buf0, 16, grid=grid(16), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, buf0, reinterpret_tensor(primals_3, (4, 128), (1, 4), 0), alpha=1, beta=1, out=buf1)
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [relu_1], Original ATen: [aten.relu]
triton_poi_fused_relu_0.run(primals_2, buf2, 16, grid=grid(16), stream=stream0)
buf3 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, buf2, reinterpret_tensor(primals_3, (4, 128), (1, 4), 0), alpha=1, beta=1, out=buf3)
del primals_3
del primals_4
buf4 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf6 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [gate, sum_1, gate_1], Original ATen: [aten.mul, aten.sum, aten.sigmoid, aten.sigmoid_backward]
triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1.run(buf1, buf3, buf4, buf6, 4, 128, grid=grid(4), stream=stream0)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.mul]
triton_poi_fused_mul_2.run(primals_2, buf4, buf5, 16, grid=grid(16), stream=stream0)
del buf4
return (buf5, primals_2, buf0, buf1, buf2, buf3, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torchvision.transforms import functional as F
import torch.utils.data
from torch import nn
import torch.nn.functional as F
class Message_Passing_Unit_v2(nn.Module):
def __init__(self, fea_size, filter_size=128):
super(Message_Passing_Unit_v2, self).__init__()
self.w = nn.Linear(fea_size, filter_size, bias=True)
self.fea_size = fea_size
self.filter_size = filter_size
def forward(self, unary_term, pair_term):
if unary_term.size()[0] == 1 and pair_term.size()[0] > 1:
unary_term = unary_term.expand(pair_term.size()[0], unary_term.
size()[1])
if unary_term.size()[0] > 1 and pair_term.size()[0] == 1:
pair_term = pair_term.expand(unary_term.size()[0], pair_term.
size()[1])
gate = self.w(F.relu(unary_term)) * self.w(F.relu(pair_term))
gate = torch.sigmoid(gate.sum(1))
output = pair_term * gate.expand(gate.size()[0], pair_term.size()[1])
return output
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'fea_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.utils.data
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 128 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (r1 + 128 * x0), xmask, other=0.0)
tmp2 = tmp0 * tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(xmask, tmp3, 0)
tmp6 = tl.sum(tmp5, 1)[:, None]
tmp7 = tl.sigmoid(tmp6)
tmp8 = 1.0
tmp9 = tmp8 - tmp7
tmp10 = tmp7 * tmp9
tl.store(out_ptr1 + x0, tmp10, xmask)
tl.store(out_ptr0 + x0, tmp6, xmask)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.sigmoid(tmp1)
tmp3 = tmp0 * tmp2
tl.store(out_ptr0 + x2, tmp3, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (128, 4), (4, 1))
assert_size_stride(primals_4, (128,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_relu_0[grid(16)](primals_1, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
extern_kernels.addmm(primals_4, buf0, reinterpret_tensor(primals_3,
(4, 128), (1, 4), 0), alpha=1, beta=1, out=buf1)
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_relu_0[grid(16)](primals_2, buf2, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
extern_kernels.addmm(primals_4, buf2, reinterpret_tensor(primals_3,
(4, 128), (1, 4), 0), alpha=1, beta=1, out=buf3)
del primals_3
del primals_4
buf4 = empty_strided_cuda((4,), (1,), torch.float32)
buf6 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_mul_sigmoid_sigmoid_backward_sum_1[grid(4)](buf1,
buf3, buf4, buf6, 4, 128, XBLOCK=1, num_warps=2, num_stages=1)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_mul_2[grid(16)](primals_2, buf4, buf5, 16, XBLOCK=
16, num_warps=1, num_stages=1)
del buf4
return buf5, primals_2, buf0, buf1, buf2, buf3, buf6
class Message_Passing_Unit_v2New(nn.Module):
def __init__(self, fea_size, filter_size=128):
super(Message_Passing_Unit_v2New, self).__init__()
self.w = nn.Linear(fea_size, filter_size, bias=True)
self.fea_size = fea_size
self.filter_size = filter_size
def forward(self, input_0, input_1):
primals_3 = self.w.weight
primals_4 = self.w.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| SpartaG117/scene_graph_benchmark | Message_Passing_Unit_v2 | false | 1,102 | [
"MIT"
] | 0 | e2e49940dd2f752b1faf9ae26707435ba3441bcb | https://github.com/SpartaG117/scene_graph_benchmark/tree/e2e49940dd2f752b1faf9ae26707435ba3441bcb | import torch
from torchvision.transforms import functional as F
import torch.utils.data
from torch import nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, fea_size, filter_size=128):
super().__init__()
self.w = nn.Linear(fea_size, filter_size, bias=True)
self.fea_size = fea_size
self.filter_size = filter_size
def forward(self, unary_term, pair_term):
if unary_term.size()[0] == 1 and pair_term.size()[0] > 1:
unary_term = unary_term.expand(pair_term.size()[0], unary_term.
size()[1])
if unary_term.size()[0] > 1 and pair_term.size()[0] == 1:
pair_term = pair_term.expand(unary_term.size()[0], pair_term.
size()[1])
gate = self.w(F.relu(unary_term)) * self.w(F.relu(pair_term))
gate = torch.sigmoid(gate.sum(1))
output = pair_term * gate.expand(gate.size()[0], pair_term.size()[1])
return output
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [4]
|
PositionalEmbedding | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/xc/cxchazmi4sbqmo6kndy2optflcbyzztvdu4wnn2b25nwssqinnbj.py
# Topologically Sorted Source Nodes: [signal], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# signal => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%sin, %cos], 1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tmp0.to(tl.float32)
tmp6 = -9.210340371976184
tmp7 = tmp5 * tmp6
tmp8 = tl_math.exp(tmp7)
tmp9 = x1
tmp10 = tmp9.to(tl.float32)
tmp11 = tmp10 * tmp8
tmp12 = tl_math.sin(tmp11)
tmp13 = tl.full(tmp12.shape, 0.0, tmp12.dtype)
tmp14 = tl.where(tmp4, tmp12, tmp13)
tmp15 = tmp0 >= tmp3
tmp16 = tl.full([1], 4, tl.int64)
tmp17 = tmp0 < tmp16
tmp18 = (-2) + x0
tmp19 = tmp18.to(tl.float32)
tmp20 = tmp19 * tmp6
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp10 * tmp21
tmp23 = tl_math.cos(tmp22)
tmp24 = tl.full(tmp23.shape, 0.0, tmp23.dtype)
tmp25 = tl.where(tmp15, tmp23, tmp24)
tmp26 = tl.where(tmp4, tmp14, tmp25)
tl.store(out_ptr0 + (x2), tmp26, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/li/clijl6e37kmnvjjduuogi6e6ltrhmuxbtlkhmiy5ijbkqmjtltpl.py
# Topologically Sorted Source Nodes: [add], Original ATen: [aten.add]
# Source node to ATen node mapping:
# add => add_2
# Graph fragment:
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, %view), kwargs = {})
triton_poi_fused_add_1 = async_compile.triton('triton_poi_fused_add_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [signal], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(buf0, 16, grid=grid(16), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add], Original ATen: [aten.add]
triton_poi_fused_add_1.run(arg0_1, buf0, buf1, 64, grid=grid(64), stream=stream0)
del arg0_1
del buf0
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
class PositionalEmbedding(torch.nn.Module):
def __init__(self):
super(PositionalEmbedding, self).__init__()
def forward(self, inputs):
if inputs.dim() != 3:
raise ValueError('The rank of input must be 3.')
length = inputs.shape[1]
channels = inputs.shape[2]
half_dim = channels // 2
positions = torch.arange(length, dtype=inputs.dtype, device=inputs.
device)
dimensions = torch.arange(half_dim, dtype=inputs.dtype, device=
inputs.device)
scale = math.log(10000.0) / float(half_dim - 1)
dimensions.mul_(-scale).exp_()
scaled_time = positions.unsqueeze(1) * dimensions.unsqueeze(0)
signal = torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)],
dim=1)
if channels % 2 == 1:
pad = torch.zeros([signal.shape[0], 1], dtype=inputs.dtype,
device=inputs.device)
signal = torch.cat([signal, pad], axis=1)
return inputs + torch.reshape(signal, [1, -1, channels])
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_cat_0(out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tmp0.to(tl.float32)
tmp6 = -9.210340371976184
tmp7 = tmp5 * tmp6
tmp8 = tl_math.exp(tmp7)
tmp9 = x1
tmp10 = tmp9.to(tl.float32)
tmp11 = tmp10 * tmp8
tmp12 = tl_math.sin(tmp11)
tmp13 = tl.full(tmp12.shape, 0.0, tmp12.dtype)
tmp14 = tl.where(tmp4, tmp12, tmp13)
tmp15 = tmp0 >= tmp3
tl.full([1], 4, tl.int64)
tmp18 = -2 + x0
tmp19 = tmp18.to(tl.float32)
tmp20 = tmp19 * tmp6
tmp21 = tl_math.exp(tmp20)
tmp22 = tmp10 * tmp21
tmp23 = tl_math.cos(tmp22)
tmp24 = tl.full(tmp23.shape, 0.0, tmp23.dtype)
tmp25 = tl.where(tmp15, tmp23, tmp24)
tmp26 = tl.where(tmp4, tmp14, tmp25)
tl.store(out_ptr0 + x2, tmp26, xmask)
@triton.jit
def triton_poi_fused_add_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 16
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + x2, tmp2, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(16)](buf0, 16, XBLOCK=16, num_warps=1,
num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_1[grid(64)](arg0_1, buf0, buf1, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del arg0_1
del buf0
return buf1,
class PositionalEmbeddingNew(torch.nn.Module):
def __init__(self):
super(PositionalEmbeddingNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| THUNLP-MT/PLM4MT | PositionalEmbedding | false | 1,103 | [
"BSD-3-Clause"
] | 0 | 85bd2ee9d96b07ac827e14d4b3e5b0d0924c3401 | https://github.com/THUNLP-MT/PLM4MT/tree/85bd2ee9d96b07ac827e14d4b3e5b0d0924c3401 | import math
import torch
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, inputs):
if inputs.dim() != 3:
raise ValueError('The rank of input must be 3.')
length = inputs.shape[1]
channels = inputs.shape[2]
half_dim = channels // 2
positions = torch.arange(length, dtype=inputs.dtype, device=inputs.
device)
dimensions = torch.arange(half_dim, dtype=inputs.dtype, device=
inputs.device)
scale = math.log(10000.0) / float(half_dim - 1)
dimensions.mul_(-scale).exp_()
scaled_time = positions.unsqueeze(1) * dimensions.unsqueeze(0)
signal = torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)],
dim=1)
if channels % 2 == 1:
pad = torch.zeros([signal.shape[0], 1], dtype=inputs.dtype,
device=inputs.device)
signal = torch.cat([signal, pad], axis=1)
return inputs + torch.reshape(signal, [1, -1, channels])
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return []
|
MaxPool | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/2l/c2lm5wvy5varadxpp77k6lvi6yjwzernwi4uqg6gmabg2nygeeur.py
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
# Source node to ATen node mapping:
# max_1 => getitem
# Graph fragment:
# %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%max_1, 0), kwargs = {})
triton_poi_fused_max_0 = async_compile.triton('triton_poi_fused_max_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
stream0 = get_raw_stream(0)
triton_poi_fused_max_0.run(arg0_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class MaxPool(nn.Module):
def __init__(self, dim=1):
super(MaxPool, self).__init__()
self.dim = dim
def forward(self, input):
return torch.max(input, self.dim)[0]
def __repr__(self):
return self.__class__.__name__ + ' (' + 'dim=' + str(self.dim) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_max_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp6 = triton_helpers.maximum(tmp4, tmp5)
tl.store(out_ptr0 + x2, tmp6, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_max_0[grid(64)](arg0_1, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del arg0_1
return buf0,
class MaxPoolNew(nn.Module):
def __init__(self, dim=1):
super(MaxPoolNew, self).__init__()
self.dim = dim
def __repr__(self):
return self.__class__.__name__ + ' (' + 'dim=' + str(self.dim) + ')'
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| SwaggyZhang/Geometry-aware | MaxPool | false | 1,104 | [
"Apache-2.0"
] | 0 | a750c00aa2f0bda5160dfdeee2eef5230fd9d993 | https://github.com/SwaggyZhang/Geometry-aware/tree/a750c00aa2f0bda5160dfdeee2eef5230fd9d993 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, dim=1):
super().__init__()
self.dim = dim
def forward(self, input):
return torch.max(input, self.dim)[0]
def __repr__(self):
return self.__class__.__name__ + ' (' + 'dim=' + str(self.dim) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Transpose | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/64/c64ahxnpt5ixqrlolbug3qf6y4u2zqmcjekif2yu4ba4hcze2fom.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16) % 4
x2 = (xindex // 64)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x2) + (64*x1)), xmask)
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class Transpose(nn.Module):
def __init__(self, dim1=0, dim2=1):
super(Transpose, self).__init__()
self.dim1 = dim1
self.dim2 = dim2
def forward(self, input):
return input.transpose(self.dim1, self.dim2).contiguous()
def __repr__(self):
return self.__class__.__name__ + ' (' + 'between=' + str(self.dim1
) + ',' + str(self.dim2) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16 % 4
x2 = xindex // 64
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x2 + 64 * x1), xmask)
tl.store(out_ptr0 + x3, tmp0, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(256)](arg0_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class TransposeNew(nn.Module):
def __init__(self, dim1=0, dim2=1):
super(TransposeNew, self).__init__()
self.dim1 = dim1
self.dim2 = dim2
def __repr__(self):
return self.__class__.__name__ + ' (' + 'between=' + str(self.dim1
) + ',' + str(self.dim2) + ')'
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| SwaggyZhang/Geometry-aware | Transpose | false | 1,105 | [
"Apache-2.0"
] | 0 | a750c00aa2f0bda5160dfdeee2eef5230fd9d993 | https://github.com/SwaggyZhang/Geometry-aware/tree/a750c00aa2f0bda5160dfdeee2eef5230fd9d993 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, dim1=0, dim2=1):
super().__init__()
self.dim1 = dim1
self.dim2 = dim2
def forward(self, input):
return input.transpose(self.dim1, self.dim2).contiguous()
def __repr__(self):
return self.__class__.__name__ + ' (' + 'between=' + str(self.dim1
) + ',' + str(self.dim2) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
GraphConv | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/rl/crljeqnoa6ykpfmvk4fgvc6cahtaak6ilhiaoxyzgcd7ynbpfnj2.py
# Topologically Sorted Source Nodes: [ones], Original ATen: [aten.ones]
# Source node to ATen node mapping:
# ones => full_default
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 1], 1), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
triton_poi_fused_ones_0 = async_compile.triton('triton_poi_fused_ones_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0,), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_ones_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_ones_0(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = 1.0
tl.store(out_ptr0 + (x0), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/dn/cdnkxcepcwle4u3a2zcw72axoqvmyd4bn2kvjp72wctqmnh3vrpp.py
# Topologically Sorted Source Nodes: [result, result_1], Original ATen: [aten.div, aten.add]
# Source node to ATen node mapping:
# result => div
# result_1 => add
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_6, %view_1), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div, %view_8), kwargs = {})
triton_poi_fused_add_div_1 = async_compile.triton('triton_poi_fused_add_div_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_1(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp4 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 / tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tl.store(in_out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [ones], Original ATen: [aten.ones]
stream0 = get_raw_stream(0)
triton_poi_fused_ones_0.run(buf0, 4, grid=grid(4), stream=stream0)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [ones, norm], Original ATen: [aten.ones, aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), buf0, out=buf1)
del buf0
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_3
del primals_4
buf3 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(primals_1, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf2, (16, 4, 4), (16, 4, 1), 0), out=buf3)
buf4 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf4)
del primals_5
buf5 = reinterpret_tensor(buf3, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [result, result_1], Original ATen: [aten.div, aten.add]
triton_poi_fused_add_div_1.run(buf5, buf1, buf4, primals_6, 256, grid=grid(256), stream=stream0)
del buf4
del primals_6
return (buf5, buf1, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (16, 4, 4), (16, 1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
import torch.nn
import torch.autograd
def sparse_bmm(sparse_matrix, dense_matrix_batch):
"""
Perform torch.bmm on an unbatched sparse matrix and a batched dense matrix.
Args:
sparse_matrix (torch.sparse.FloatTensor): Shape = (m, n)
dense_matrix_batch (torch.FloatTensor): Shape = (b, n, p)
Returns:
(torch.FloatTensor):
Result of the batched matrix multiplication. Shape = (b, n, p)
"""
m = sparse_matrix.shape[0]
b, n, p = dense_matrix_batch.shape
dense_matrix = dense_matrix_batch.transpose(0, 1).reshape(n, b * p)
result = torch.sparse.mm(sparse_matrix, dense_matrix)
return result.reshape(m, b, p).transpose(0, 1)
class GraphConv(nn.Module):
"""
A simple graph convolution layer, similar to the one defined in Kipf et al.
https://arxiv.org/abs/1609.02907
This operation with self_layer=False is equivalent to :math:`(A H W)` where:
- :math:`H` is the node features with shape (batch_size, num_nodes, input_dim)
- :math:`W` is a weight matrix of shape (input_dim, output_dim)
- :math:`A` is the adjacency matrix of shape (num_nodes, num_nodes).
It can include self-loop.
With normalize_adj=True, it is equivalent to :math:`(D^{-1} A H W)`, where:
- :math:`D` is a diagonal matrix with :math:`D_{ii}` = the sum of the i-th row of A.
In other words, :math:`D` is the incoming degree of each node.
With self_layer=True, it is equivalent to the above plus :math:`(H W_{\\text{self}})`, where:
- :math:`W_{\\text{self}}` is a separate weight matrix to filter each node's self features.
Note that when self_layer is True, A should not include self-loop.
Example:
>>> node_feat = torch.rand(1, 3, 5)
>>> i = torch.LongTensor(
... [[0, 1, 1, 2, 2, 0], [1, 0, 2, 1, 0, 2]])
>>> v = torch.FloatTensor([1, 1, 1, 1, 1, 1])
>>> adj = torch.sparse.FloatTensor(i, v, torch.Size([3, 3]))
>>> model = GraphConv(5, 10)
>>> output = model(node_feat, adj)
>>> # pre-normalize adj
>>> adj = normalize_adj(adj)
>>> output = model(node_feat, adj, normalize_adj=False)
If you use this code, please cite the original paper in addition to Kaolin.
.. code-block::
@article{kipf2016semi,
title={Semi-Supervised Classification with Graph Convolutional Networks},
author={Kipf, Thomas N and Welling, Max},
journal={arXiv preprint arXiv:1609.02907},
year={2016}
}
Args:
input_dim (int): The number of features in each input node.
output_dim (int): The number of features in each output node.
bias (bool): Whether to add bias after the node-wise linear layer.
"""
def __init__(self, input_dim, output_dim, self_layer=True, bias=True):
super(GraphConv, self).__init__()
self.self_layer = self_layer
self.linear = nn.Linear(input_dim, output_dim, bias=bias)
if self_layer:
self.linear_self = nn.Linear(input_dim, output_dim, bias=bias)
else:
self.linear_self = None
self.initialize()
def initialize(self):
nn.init.xavier_uniform_(self.linear.weight.data)
if self.linear.bias is not None:
self.linear.bias.data.uniform_(-1.0, 1.0)
if self.self_layer:
nn.init.xavier_uniform_(self.linear_self.weight.data)
if self.linear_self.bias is not None:
self.linear_self.bias.data.uniform_(-1.0, 1.0)
def forward(self, node_feat, adj, normalize_adj=True):
"""
Args:
node_feat (torch.FloatTensor):
Shape = (batch_size, num_nodes, input_dim)
The input features of each node.
adj (torch.sparse.FloatTensor or torch.FloatTensor):
Shape = (num_nodes, num_nodes)
The adjacency matrix. adj[i, j] is non-zero if there's an
incoming edge from j to i. Should not include self-loop if
self_layer is True.
normalize_adj (bool):
Set this to true to apply normalization to adjacency; that is,
each output feature will be divided by the number of incoming
neighbors. If normalization is not desired, or if the adjacency
matrix is pre-normalized, set this to False to improve
performance.
Returns:
(torch.FloatTensor):
The output features of each node.
Shape = (batch_size, num_nodes, output_dim)
"""
if adj.type().endswith('sparse.FloatTensor'):
if normalize_adj:
norm = torch.sparse.mm(adj, torch.ones((adj.shape[0], 1),
device=node_feat.device))
result = sparse_bmm(adj, self.linear(node_feat)) / norm
else:
result = sparse_bmm(adj, self.linear(node_feat))
elif normalize_adj:
norm = torch.matmul(adj, torch.ones((adj.shape[0], 1), device=
node_feat.device))
result = torch.matmul(adj, self.linear(node_feat)) / norm
else:
result = torch.matmul(adj, self.linear(node_feat))
if self.self_layer:
result += self.linear_self(node_feat)
return result
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'output_dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch import nn
import torch.nn
import torch.autograd
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_ones_0(out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = 1.0
tl.store(out_ptr0 + x0, tmp0, xmask)
@triton.jit
def triton_poi_fused_add_div_1(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2,
xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp4 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 / tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tl.store(in_out_ptr0 + x2, tmp6, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_ones_0[grid(4)](buf0, 4, XBLOCK=4, num_warps=1,
num_stages=1)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0),
buf0, out=buf1)
del buf0
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_4, reinterpret_tensor(primals_2, (64,
4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf2)
del primals_3
del primals_4
buf3 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(primals_1, (16, 4, 4), (16, 4,
1), 0), reinterpret_tensor(buf2, (16, 4, 4), (16, 4, 1), 0),
out=buf3)
buf4 = buf2
del buf2
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf4)
del primals_5
buf5 = reinterpret_tensor(buf3, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf3
triton_poi_fused_add_div_1[grid(256)](buf5, buf1, buf4, primals_6,
256, XBLOCK=128, num_warps=4, num_stages=1)
del buf4
del primals_6
return buf5, buf1, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0
), reinterpret_tensor(primals_1, (16, 4, 4), (16, 1, 4), 0)
def sparse_bmm(sparse_matrix, dense_matrix_batch):
"""
Perform torch.bmm on an unbatched sparse matrix and a batched dense matrix.
Args:
sparse_matrix (torch.sparse.FloatTensor): Shape = (m, n)
dense_matrix_batch (torch.FloatTensor): Shape = (b, n, p)
Returns:
(torch.FloatTensor):
Result of the batched matrix multiplication. Shape = (b, n, p)
"""
m = sparse_matrix.shape[0]
b, n, p = dense_matrix_batch.shape
dense_matrix = dense_matrix_batch.transpose(0, 1).reshape(n, b * p)
result = torch.sparse.mm(sparse_matrix, dense_matrix)
return result.reshape(m, b, p).transpose(0, 1)
class GraphConvNew(nn.Module):
"""
A simple graph convolution layer, similar to the one defined in Kipf et al.
https://arxiv.org/abs/1609.02907
This operation with self_layer=False is equivalent to :math:`(A H W)` where:
- :math:`H` is the node features with shape (batch_size, num_nodes, input_dim)
- :math:`W` is a weight matrix of shape (input_dim, output_dim)
- :math:`A` is the adjacency matrix of shape (num_nodes, num_nodes).
It can include self-loop.
With normalize_adj=True, it is equivalent to :math:`(D^{-1} A H W)`, where:
- :math:`D` is a diagonal matrix with :math:`D_{ii}` = the sum of the i-th row of A.
In other words, :math:`D` is the incoming degree of each node.
With self_layer=True, it is equivalent to the above plus :math:`(H W_{\\text{self}})`, where:
- :math:`W_{\\text{self}}` is a separate weight matrix to filter each node's self features.
Note that when self_layer is True, A should not include self-loop.
Example:
>>> node_feat = torch.rand(1, 3, 5)
>>> i = torch.LongTensor(
... [[0, 1, 1, 2, 2, 0], [1, 0, 2, 1, 0, 2]])
>>> v = torch.FloatTensor([1, 1, 1, 1, 1, 1])
>>> adj = torch.sparse.FloatTensor(i, v, torch.Size([3, 3]))
>>> model = GraphConv(5, 10)
>>> output = model(node_feat, adj)
>>> # pre-normalize adj
>>> adj = normalize_adj(adj)
>>> output = model(node_feat, adj, normalize_adj=False)
If you use this code, please cite the original paper in addition to Kaolin.
.. code-block::
@article{kipf2016semi,
title={Semi-Supervised Classification with Graph Convolutional Networks},
author={Kipf, Thomas N and Welling, Max},
journal={arXiv preprint arXiv:1609.02907},
year={2016}
}
Args:
input_dim (int): The number of features in each input node.
output_dim (int): The number of features in each output node.
bias (bool): Whether to add bias after the node-wise linear layer.
"""
def __init__(self, input_dim, output_dim, self_layer=True, bias=True):
super(GraphConvNew, self).__init__()
self.self_layer = self_layer
self.linear = nn.Linear(input_dim, output_dim, bias=bias)
if self_layer:
self.linear_self = nn.Linear(input_dim, output_dim, bias=bias)
else:
self.linear_self = None
self.initialize()
def initialize(self):
nn.init.xavier_uniform_(self.linear.weight.data)
if self.linear.bias is not None:
self.linear.bias.data.uniform_(-1.0, 1.0)
if self.self_layer:
nn.init.xavier_uniform_(self.linear_self.weight.data)
if self.linear_self.bias is not None:
self.linear_self.bias.data.uniform_(-1.0, 1.0)
def forward(self, input_0, input_1):
primals_3 = self.linear.weight
primals_4 = self.linear.bias
primals_5 = self.linear_self.weight
primals_6 = self.linear_self.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6])
return output[0]
| T0mt0mp/kaolin | GraphConv | false | 1,106 | [
"ECL-2.0",
"Apache-2.0"
] | 0 | 57d1e1478eec8df49dc7cc492f25637cec40399f | https://github.com/T0mt0mp/kaolin/tree/57d1e1478eec8df49dc7cc492f25637cec40399f | import torch
from torch import nn
import torch.nn
import torch.autograd
def sparse_bmm(sparse_matrix, dense_matrix_batch):
"""
Perform torch.bmm on an unbatched sparse matrix and a batched dense matrix.
Args:
sparse_matrix (torch.sparse.FloatTensor): Shape = (m, n)
dense_matrix_batch (torch.FloatTensor): Shape = (b, n, p)
Returns:
(torch.FloatTensor):
Result of the batched matrix multiplication. Shape = (b, n, p)
"""
m = sparse_matrix.shape[0]
b, n, p = dense_matrix_batch.shape
dense_matrix = dense_matrix_batch.transpose(0, 1).reshape(n, b * p)
result = torch.sparse.mm(sparse_matrix, dense_matrix)
return result.reshape(m, b, p).transpose(0, 1)
class Model(nn.Module):
"""
A simple graph convolution layer, similar to the one defined in Kipf et al.
https://arxiv.org/abs/1609.02907
This operation with self_layer=False is equivalent to :math:`(A H W)` where:
- :math:`H` is the node features with shape (batch_size, num_nodes, input_dim)
- :math:`W` is a weight matrix of shape (input_dim, output_dim)
- :math:`A` is the adjacency matrix of shape (num_nodes, num_nodes).
It can include self-loop.
With normalize_adj=True, it is equivalent to :math:`(D^{-1} A H W)`, where:
- :math:`D` is a diagonal matrix with :math:`D_{ii}` = the sum of the i-th row of A.
In other words, :math:`D` is the incoming degree of each node.
With self_layer=True, it is equivalent to the above plus :math:`(H W_{\\text{self}})`, where:
- :math:`W_{\\text{self}}` is a separate weight matrix to filter each node's self features.
Note that when self_layer is True, A should not include self-loop.
Example:
>>> node_feat = torch.rand(1, 3, 5)
>>> i = torch.LongTensor(
... [[0, 1, 1, 2, 2, 0], [1, 0, 2, 1, 0, 2]])
>>> v = torch.FloatTensor([1, 1, 1, 1, 1, 1])
>>> adj = torch.sparse.FloatTensor(i, v, torch.Size([3, 3]))
>>> model = GraphConv(5, 10)
>>> output = model(node_feat, adj)
>>> # pre-normalize adj
>>> adj = normalize_adj(adj)
>>> output = model(node_feat, adj, normalize_adj=False)
If you use this code, please cite the original paper in addition to Kaolin.
.. code-block::
@article{kipf2016semi,
title={Semi-Supervised Classification with Graph Convolutional Networks},
author={Kipf, Thomas N and Welling, Max},
journal={arXiv preprint arXiv:1609.02907},
year={2016}
}
Args:
input_dim (int): The number of features in each input node.
output_dim (int): The number of features in each output node.
bias (bool): Whether to add bias after the node-wise linear layer.
"""
def __init__(self, input_dim, output_dim, self_layer=True, bias=True):
super().__init__()
self.self_layer = self_layer
self.linear = nn.Linear(input_dim, output_dim, bias=bias)
if self_layer:
self.linear_self = nn.Linear(input_dim, output_dim, bias=bias)
else:
self.linear_self = None
self.initialize()
def initialize(self):
nn.init.xavier_uniform_(self.linear.weight.data)
if self.linear.bias is not None:
self.linear.bias.data.uniform_(-1.0, 1.0)
if self.self_layer:
nn.init.xavier_uniform_(self.linear_self.weight.data)
if self.linear_self.bias is not None:
self.linear_self.bias.data.uniform_(-1.0, 1.0)
def forward(self, node_feat, adj, normalize_adj=True):
"""
Args:
node_feat (torch.FloatTensor):
Shape = (batch_size, num_nodes, input_dim)
The input features of each node.
adj (torch.sparse.FloatTensor or torch.FloatTensor):
Shape = (num_nodes, num_nodes)
The adjacency matrix. adj[i, j] is non-zero if there's an
i
# ... truncated (>4000 chars) for memory efficiency |
Align | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/7l/c7lrnzz7u6zsph7qtb6own22vqc34am7brqewva4qc33ddgzxza7.py
# Topologically Sorted Source Nodes: [sub, norm, pred], Original ATen: [aten.sub, aten.linalg_vector_norm, aten.neg]
# Source node to ATen node mapping:
# norm => pow_1, pow_2, sum_1
# pred => neg
# sub => sub
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 4), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1]), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.25), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%pow_2,), kwargs = {})
triton_poi_fused_linalg_vector_norm_neg_sub_0 = async_compile.triton('triton_poi_fused_linalg_vector_norm_neg_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_linalg_vector_norm_neg_sub_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_linalg_vector_norm_neg_sub_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp6 = tl.load(in_ptr1 + (16 + x0 + (64*x1)), xmask)
tmp11 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp12 = tl.load(in_ptr1 + (32 + x0 + (64*x1)), xmask)
tmp17 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp18 = tl.load(in_ptr1 + (48 + x0 + (64*x1)), xmask)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp4 = tmp3 * tmp3
tmp7 = tmp5 - tmp6
tmp8 = tmp7 * tmp7
tmp9 = tmp8 * tmp8
tmp10 = tmp4 + tmp9
tmp13 = tmp11 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp19 = tmp17 - tmp18
tmp20 = tmp19 * tmp19
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp23 = 0.25
tmp24 = libdevice.pow(tmp22, tmp23)
tmp25 = -tmp24
tl.store(out_ptr0 + (x2), tmp25, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sub, norm, pred], Original ATen: [aten.sub, aten.linalg_vector_norm, aten.neg]
stream0 = get_raw_stream(0)
triton_poi_fused_linalg_vector_norm_neg_sub_0.run(arg0_1, arg1_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
del arg1_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
class Align(torch.nn.Module):
def __init__(self, p):
super(Align, self).__init__()
self.p = p
def forward(self, e1, e2):
pred = -torch.norm(e1 - e2, p=self.p, dim=1)
return pred
def only_pos_loss(self, e1, r, e2):
return -F.logsigmoid(-torch.sum(torch.pow(e1 + r - e2, 2), 1)).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'p': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_linalg_vector_norm_neg_sub_0(in_ptr0, in_ptr1,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp6 = tl.load(in_ptr1 + (16 + x0 + 64 * x1), xmask)
tmp11 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp12 = tl.load(in_ptr1 + (32 + x0 + 64 * x1), xmask)
tmp17 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp18 = tl.load(in_ptr1 + (48 + x0 + 64 * x1), xmask)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp4 = tmp3 * tmp3
tmp7 = tmp5 - tmp6
tmp8 = tmp7 * tmp7
tmp9 = tmp8 * tmp8
tmp10 = tmp4 + tmp9
tmp13 = tmp11 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp19 = tmp17 - tmp18
tmp20 = tmp19 * tmp19
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp23 = 0.25
tmp24 = libdevice.pow(tmp22, tmp23)
tmp25 = -tmp24
tl.store(out_ptr0 + x2, tmp25, xmask)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_linalg_vector_norm_neg_sub_0[grid(64)](arg0_1,
arg1_1, buf0, 64, XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
del arg1_1
return buf0,
class AlignNew(torch.nn.Module):
def __init__(self, p):
super(AlignNew, self).__init__()
self.p = p
def only_pos_loss(self, e1, r, e2):
return -F.logsigmoid(-torch.sum(torch.pow(e1 + r - e2, 2), 1)).sum()
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| TMUITLab/EAFR | Align | false | 1,108 | [
"MIT"
] | 0 | dadb6485d48711ccb8aa2f03760aeb437645f1ff | https://github.com/TMUITLab/EAFR/tree/dadb6485d48711ccb8aa2f03760aeb437645f1ff | import torch
import torch.nn.functional as F
class Model(torch.nn.Module):
def __init__(self, p):
super().__init__()
self.p = p
def forward(self, e1, e2):
pred = -torch.norm(e1 - e2, p=self.p, dim=1)
return pred
def only_pos_loss(self, e1, r, e2):
return -F.logsigmoid(-torch.sum(torch.pow(e1 + r - e2, 2), 1)).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
MNISTGenerator | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/hy/chyn4ucj7uhqavrcrhxk2c5izzfdiw63bn3glmpyn3tpx5bpigdc.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_2), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
triton_poi_fused_relu_0 = async_compile.triton('triton_poi_fused_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/mk/cmkwbx5yy6hhpvjwjmlmi7fvwp45sgwxrqj3vqfqnafxubycnrag.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.sigmoid, aten.sigmoid_backward]
# Source node to ATen node mapping:
# x_1 => sigmoid
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_5), kwargs = {})
# %sigmoid : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add_tensor,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %sigmoid), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %sub), kwargs = {})
triton_poi_fused_sigmoid_sigmoid_backward_1 = async_compile.triton('triton_poi_fused_sigmoid_sigmoid_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_sigmoid_sigmoid_backward_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_sigmoid_sigmoid_backward_1(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 3136
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 784
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tmp4 = 1.0
tmp5 = tmp4 - tmp3
tmp6 = tmp3 * tmp5
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
tl.store(out_ptr0 + (x2), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128, ), (1, ))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (784, 128), (128, 1))
assert_size_stride(primals_5, (784, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 128), (1, 4), 0), out=buf0)
del primals_1
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_2, 512, grid=grid(512), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (128, 784), (1, 128), 0), out=buf2)
buf3 = buf2; del buf2 # reuse
buf4 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.sigmoid, aten.sigmoid_backward]
triton_poi_fused_sigmoid_sigmoid_backward_1.run(buf3, primals_5, buf4, 3136, grid=grid(3136), stream=stream0)
del primals_5
return (reinterpret_tensor(buf3, (4, 1, 28, 28), (784, 784, 28, 1), 0), primals_3, buf1, buf4, primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((784, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((784, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn as nn
from torch import optim as optim
from torchvision import transforms as transforms
class MNISTGenerator(nn.Module):
def __init__(self, latent_dim):
super(MNISTGenerator, self).__init__()
self.image_shape = 1, 28, 28
self.latent_dim = latent_dim
self.dense1 = nn.Linear(self.latent_dim, 128, True)
self.dense2 = nn.Linear(128, 784, True)
def forward(self, x):
x = nn.functional.relu(self.dense1(x))
x = nn.functional.sigmoid(self.dense2(x))
return x.view(x.shape[0], *self.image_shape)
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'latent_dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn as nn
from torch import optim as optim
from torchvision import transforms as transforms
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, xmask)
@triton.jit
def triton_poi_fused_sigmoid_sigmoid_backward_1(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 3136
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 784
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tmp4 = 1.0
tmp5 = tmp4 - tmp3
tmp6 = tmp3 * tmp5
tl.store(in_out_ptr0 + x2, tmp3, xmask)
tl.store(out_ptr0 + x2, tmp6, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (128, 4), (4, 1))
assert_size_stride(primals_2, (128,), (1,))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (784, 128), (128, 1))
assert_size_stride(primals_5, (784,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 128), (128, 1), torch.float32)
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 128),
(1, 4), 0), out=buf0)
del primals_1
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(512)](buf1, primals_2, 512, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (128, 784), (
1, 128), 0), out=buf2)
buf3 = buf2
del buf2
buf4 = empty_strided_cuda((4, 784), (784, 1), torch.float32)
triton_poi_fused_sigmoid_sigmoid_backward_1[grid(3136)](buf3,
primals_5, buf4, 3136, XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
return reinterpret_tensor(buf3, (4, 1, 28, 28), (784, 784, 28, 1), 0
), primals_3, buf1, buf4, primals_4
class MNISTGeneratorNew(nn.Module):
def __init__(self, latent_dim):
super(MNISTGeneratorNew, self).__init__()
self.image_shape = 1, 28, 28
self.latent_dim = latent_dim
self.dense1 = nn.Linear(self.latent_dim, 128, True)
self.dense2 = nn.Linear(128, 784, True)
def forward(self, input_0):
primals_1 = self.dense1.weight
primals_2 = self.dense1.bias
primals_4 = self.dense2.weight
primals_5 = self.dense2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| RobinMaas95/GTSRB_Visualization | MNISTGenerator | false | 1,109 | [
"MIT"
] | 0 | fa837ff94e089a936ef4f4418970d262b35f70b6 | https://github.com/RobinMaas95/GTSRB_Visualization/tree/fa837ff94e089a936ef4f4418970d262b35f70b6 | import torch
from torch import nn as nn
from torch import optim as optim
from torchvision import transforms as transforms
class Model(nn.Module):
def __init__(self, latent_dim):
super().__init__()
self.image_shape = 1, 28, 28
self.latent_dim = latent_dim
self.dense1 = nn.Linear(self.latent_dim, 128, True)
self.dense2 = nn.Linear(128, 784, True)
def forward(self, x):
x = nn.functional.relu(self.dense1(x))
x = nn.functional.sigmoid(self.dense2(x))
return x.view(x.shape[0], *self.image_shape)
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [4]
|
Conv2d | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/yj/cyjqxrdr34zdlpnaqepj4py4tvwh2ebdslxkfeu7skxqjn4syiak.py
# Topologically Sorted Source Nodes: [mean, mean_1], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# mean => mean
# mean_1 => mean_1
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [1], True), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mean, [2], True), kwargs = {})
triton_poi_fused_mean_0 = async_compile.triton('triton_poi_fused_mean_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mean_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mean_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp9 = tl.load(in_ptr0 + (4 + x0 + (64*x1)), xmask)
tmp10 = tl.load(in_ptr0 + (20 + x0 + (64*x1)), xmask)
tmp12 = tl.load(in_ptr0 + (36 + x0 + (64*x1)), xmask)
tmp14 = tl.load(in_ptr0 + (52 + x0 + (64*x1)), xmask)
tmp18 = tl.load(in_ptr0 + (8 + x0 + (64*x1)), xmask)
tmp19 = tl.load(in_ptr0 + (24 + x0 + (64*x1)), xmask)
tmp21 = tl.load(in_ptr0 + (40 + x0 + (64*x1)), xmask)
tmp23 = tl.load(in_ptr0 + (56 + x0 + (64*x1)), xmask)
tmp27 = tl.load(in_ptr0 + (12 + x0 + (64*x1)), xmask)
tmp28 = tl.load(in_ptr0 + (28 + x0 + (64*x1)), xmask)
tmp30 = tl.load(in_ptr0 + (44 + x0 + (64*x1)), xmask)
tmp32 = tl.load(in_ptr0 + (60 + x0 + (64*x1)), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp11 = tmp9 + tmp10
tmp13 = tmp11 + tmp12
tmp15 = tmp13 + tmp14
tmp16 = tmp15 / tmp7
tmp17 = tmp8 + tmp16
tmp20 = tmp18 + tmp19
tmp22 = tmp20 + tmp21
tmp24 = tmp22 + tmp23
tmp25 = tmp24 / tmp7
tmp26 = tmp17 + tmp25
tmp29 = tmp27 + tmp28
tmp31 = tmp29 + tmp30
tmp33 = tmp31 + tmp32
tmp34 = tmp33 / tmp7
tmp35 = tmp26 + tmp34
tmp36 = tmp35 / tmp7
tl.store(out_ptr0 + (x2), tmp36, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/nn/cnnlxy27st37ghyasby2wfzvdhe7fuiziq2zan6tepfxf2oe5jld.py
# Topologically Sorted Source Nodes: [weight_mean, weight, std, weight_1], Original ATen: [aten.mean, aten.sub, aten.std, aten.div]
# Source node to ATen node mapping:
# std => sqrt, var
# weight => sub
# weight_1 => div
# weight_mean => mean_2
# Graph fragment:
# %mean_2 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mean_1, [3], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %mean_2), kwargs = {})
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%view, [1]), kwargs = {correction: 1.0})
# %sqrt : [num_users=2] = call_function[target=torch.ops.aten.sqrt.default](args = (%var,), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %expand), kwargs = {})
triton_per_fused_div_mean_std_sub_1 = async_compile.triton('triton_per_fused_div_mean_std_sub_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[4, 64],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32', 6: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 6), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_mean_std_sub_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 3, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_div_mean_std_sub_1(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (64*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 4.0
tmp9 = tmp7 / tmp8
tmp10 = tmp0 - tmp9
tmp11 = tl.broadcast_to(tmp10, [XBLOCK, RBLOCK])
tmp13 = tl.where(xmask, tmp11, 0)
tmp14 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp16 = tl.where(xmask, tmp14, 0)
tmp17 = tl.sum(tmp16, 1)[:, None]
tmp18 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp19 = tmp18.to(tl.float32)
tmp20 = tmp17 / tmp19
tmp21 = tmp11 - tmp20
tmp22 = tmp21 * tmp21
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.where(xmask, tmp23, 0)
tmp26 = tl.sum(tmp25, 1)[:, None]
tmp27 = 63.0
tmp28 = tmp26 / tmp27
tmp29 = libdevice.sqrt(tmp28)
tmp30 = 1e-05
tmp31 = tmp29 + tmp30
tmp32 = tmp10 / tmp31
tl.store(out_ptr0 + (r1 + (64*x0)), tmp10, xmask)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp29, xmask)
tl.store(out_ptr1 + (r1 + (64*x0)), tmp32, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/k2/ck2mamkqpmuzem4n3p4ij6fmfpy2bcbblg6sx6wwslgqwuqq5ifh.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %div, %primals_2, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_2 = async_compile.triton('triton_poi_fused_convolution_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_2', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 4), (4, 16, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean, mean_1], Original ATen: [aten.mean]
stream0 = get_raw_stream(0)
triton_poi_fused_mean_0.run(primals_1, buf0, 16, grid=grid(16), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf3 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf5 = buf3; del buf3 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [weight_mean, weight, std, weight_1], Original ATen: [aten.mean, aten.sub, aten.std, aten.div]
triton_per_fused_div_mean_std_sub_1.run(buf5, primals_1, buf0, buf1, buf6, 4, 64, grid=grid(4), stream=stream0)
del buf0
del buf1
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(primals_3, buf6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 4, 1, 1), (4, 1, 1, 1))
buf8 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf8, primals_2, 16, grid=grid(16), stream=stream0)
del primals_2
return (buf8, primals_1, primals_3, buf5, buf6, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.parallel
import torch.nn.functional as F
class Conv2d(nn.Conv2d):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, groups=1, bias=True):
super(Conv2d, self).__init__(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, bias)
def forward(self, x):
weight = self.weight
weight_mean = weight.mean(dim=1, keepdim=True).mean(dim=2, keepdim=True
).mean(dim=3, keepdim=True)
weight = weight - weight_mean
std = weight.view(weight.size(0), -1).std(dim=1).view(-1, 1, 1, 1
) + 1e-05
weight = weight / std.expand_as(weight)
return F.conv2d(x, weight, self.bias, self.stride, self.padding,
self.dilation, self.groups)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.nn.parallel
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_mean_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp9 = tl.load(in_ptr0 + (4 + x0 + 64 * x1), xmask)
tmp10 = tl.load(in_ptr0 + (20 + x0 + 64 * x1), xmask)
tmp12 = tl.load(in_ptr0 + (36 + x0 + 64 * x1), xmask)
tmp14 = tl.load(in_ptr0 + (52 + x0 + 64 * x1), xmask)
tmp18 = tl.load(in_ptr0 + (8 + x0 + 64 * x1), xmask)
tmp19 = tl.load(in_ptr0 + (24 + x0 + 64 * x1), xmask)
tmp21 = tl.load(in_ptr0 + (40 + x0 + 64 * x1), xmask)
tmp23 = tl.load(in_ptr0 + (56 + x0 + 64 * x1), xmask)
tmp27 = tl.load(in_ptr0 + (12 + x0 + 64 * x1), xmask)
tmp28 = tl.load(in_ptr0 + (28 + x0 + 64 * x1), xmask)
tmp30 = tl.load(in_ptr0 + (44 + x0 + 64 * x1), xmask)
tmp32 = tl.load(in_ptr0 + (60 + x0 + 64 * x1), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp11 = tmp9 + tmp10
tmp13 = tmp11 + tmp12
tmp15 = tmp13 + tmp14
tmp16 = tmp15 / tmp7
tmp17 = tmp8 + tmp16
tmp20 = tmp18 + tmp19
tmp22 = tmp20 + tmp21
tmp24 = tmp22 + tmp23
tmp25 = tmp24 / tmp7
tmp26 = tmp17 + tmp25
tmp29 = tmp27 + tmp28
tmp31 = tmp29 + tmp30
tmp33 = tmp31 + tmp32
tmp34 = tmp33 / tmp7
tmp35 = tmp26 + tmp34
tmp36 = tmp35 / tmp7
tl.store(out_ptr0 + x2, tmp36, xmask)
@triton.jit
def triton_per_fused_div_mean_std_sub_1(in_out_ptr0, in_ptr0, in_ptr1,
out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 64 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 4.0
tmp9 = tmp7 / tmp8
tmp10 = tmp0 - tmp9
tmp11 = tl.broadcast_to(tmp10, [XBLOCK, RBLOCK])
tl.where(xmask, tmp11, 0)
tmp14 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp16 = tl.where(xmask, tmp14, 0)
tmp17 = tl.sum(tmp16, 1)[:, None]
tmp18 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp19 = tmp18.to(tl.float32)
tmp20 = tmp17 / tmp19
tmp21 = tmp11 - tmp20
tmp22 = tmp21 * tmp21
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.where(xmask, tmp23, 0)
tmp26 = tl.sum(tmp25, 1)[:, None]
tmp27 = 63.0
tmp28 = tmp26 / tmp27
tmp29 = libdevice.sqrt(tmp28)
tmp30 = 1e-05
tmp31 = tmp29 + tmp30
tmp32 = tmp10 / tmp31
tl.store(out_ptr0 + (r1 + 64 * x0), tmp10, xmask)
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp29, xmask)
tl.store(out_ptr1 + (r1 + 64 * x0), tmp32, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x2, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1, 4), (4, 16, 16, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mean_0[grid(16)](primals_1, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf3 = empty_strided_cuda((4,), (1,), torch.float32)
buf5 = buf3
del buf3
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_per_fused_div_mean_std_sub_1[grid(4)](buf5, primals_1, buf0,
buf1, buf6, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
del buf0
del buf1
buf7 = extern_kernels.convolution(primals_3, buf6, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf7, (4, 4, 1, 1), (4, 1, 1, 1))
buf8 = buf7
del buf7
triton_poi_fused_convolution_2[grid(16)](buf8, primals_2, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_2
return buf8, primals_1, primals_3, buf5, buf6
class Conv2dNew(nn.Conv2d):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, groups=1, bias=True):
super(Conv2dNew, self).__init__(in_channels, out_channels,
kernel_size, stride, padding, dilation, groups, bias)
def forward(self, input_0):
primals_1 = self.weight
primals_2 = self.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| T1anZhenYu/pytorch-classification | Conv2d | false | 1,110 | [
"MIT"
] | 0 | ad68e09f20a98541bcb437a7df8e7d14e8c21636 | https://github.com/T1anZhenYu/pytorch-classification/tree/ad68e09f20a98541bcb437a7df8e7d14e8c21636 | import torch
import torch.nn as nn
import torch.nn.parallel
import torch.nn.functional as F
class Model(nn.Conv2d):
def __init__(self, in_channels, out_channels, kernel_size, stride=1,
padding=0, dilation=1, groups=1, bias=True):
super().__init__(in_channels, out_channels, kernel_size,
stride, padding, dilation, groups, bias)
def forward(self, x):
weight = self.weight
weight_mean = weight.mean(dim=1, keepdim=True).mean(dim=2, keepdim=True
).mean(dim=3, keepdim=True)
weight = weight - weight_mean
std = weight.view(weight.size(0), -1).std(dim=1).view(-1, 1, 1, 1
) + 1e-05
weight = weight / std.expand_as(weight)
return F.conv2d(x, weight, self.bias, self.stride, self.padding,
self.dilation, self.groups)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
lovasz_hinge | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/qm/cqmiebj3omt27mbkqpnnkw5z77l3s4xgfv6gqogfvmadhw2rd3m7.py
# Topologically Sorted Source Nodes: [valid], Original ATen: [aten.ne]
# Source node to ATen node mapping:
# valid => ne
# Graph fragment:
# %ne : [num_users=1] = call_function[target=torch.ops.aten.ne.Scalar](args = (%view_1, 255), kwargs = {})
triton_poi_fused_ne_0 = async_compile.triton('triton_poi_fused_ne_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i1', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_ne_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_ne_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 255.0
tmp2 = tmp0 != tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((256, ), (1, ), torch.bool)
# Topologically Sorted Source Nodes: [valid], Original ATen: [aten.ne]
stream0 = get_raw_stream(0)
triton_poi_fused_ne_0.run(arg1_1, buf0, 256, grid=grid(256), stream=stream0)
return (reinterpret_tensor(arg0_1, (256, ), (1, ), 0), buf0, reinterpret_tensor(arg1_1, (256, ), (1, ), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.parallel
import torch.utils.data
from torchvision.transforms import functional as F
import torch.nn.functional as F
from torch.autograd import Variable
def flatten_binary_scores(scores, labels, ignore=255):
"""
Flattens predictions in the batch (binary case)
Remove labels equal to 'ignore'
"""
scores = scores.view(-1)
labels = labels.view(-1)
if ignore is None:
return scores, labels
valid = labels != ignore
vscores = scores[valid]
vlabels = labels[valid]
return vscores, vlabels
def lovasz_grad(gt_sorted):
"""
Computes gradient of the Lovasz extension w.r.t sorted errors
See Alg. 1 in paper
"""
p = len(gt_sorted)
gts = gt_sorted.sum()
intersection = gts.float() - gt_sorted.float().cumsum(0)
union = gts.float() + (1 - gt_sorted).float().cumsum(0)
jaccard = 1.0 - intersection / union
if p > 1:
jaccard[1:p] = jaccard[1:p] - jaccard[0:-1]
return jaccard
def isnan(x):
return x != x
def mean(l, ignore_nan=False, empty=0):
"""
nanmean compatible with generators.
"""
l = iter(l)
if ignore_nan:
l = ifilterfalse(isnan, l)
try:
n = 1
acc = next(l)
except StopIteration:
if empty == 'raise':
raise ValueError('Empty mean')
return empty
for n, v in enumerate(l, 2):
acc += v
if n == 1:
return acc
return acc / n
class lovasz_hinge(torch.nn.Module):
def __init__(self, per_img=False, ignore=255):
"""
:param weight: 1D weight vector to deal with the class-imbalance
"""
super().__init__()
self.per_image = per_img
self.ignore = ignore
def lovasz_hinge_flat(self, logits, labels):
"""
Binary Lovasz hinge loss
logits: [P] Variable, logits at each prediction (between -\\infty and +\\infty)
labels: [P] Tensor, binary ground truth labels (0 or 1)
ignore: label to ignore
"""
if len(labels) == 0:
return logits.sum() * 0.0
signs = 2.0 * labels.float() - 1.0
errors = 1.0 - logits * Variable(signs)
errors_sorted, perm = torch.sort(errors, dim=0, descending=True)
perm = perm.data
gt_sorted = labels[perm]
grad = lovasz_grad(gt_sorted)
loss = torch.dot(F.relu(errors_sorted), Variable(grad))
return loss
def forward(self, logits, labels):
"""
Binary Lovasz hinge loss
logits: [B, H, W] Variable, logits at each pixel (between -\\infty and +\\infty)
labels: [B, H, W] Tensor, binary ground truth masks (0 or 1)
per_image: compute the loss per image instead of per batch
ignore: void class id
"""
if self.per_image:
loss = mean(self.lovasz_hinge_flat(*flatten_binary_scores(log.
unsqueeze(0), lab.unsqueeze(0), self.ignore)) for log, lab in
zip(logits, labels))
else:
loss = self.lovasz_hinge_flat(*flatten_binary_scores(logits,
labels, self.ignore))
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn.parallel
import torch.utils.data
from torchvision.transforms import functional as F
import torch.nn.functional as F
from torch.autograd import Variable
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_ne_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 255.0
tmp2 = tmp0 != tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((256,), (1,), torch.bool)
get_raw_stream(0)
triton_poi_fused_ne_0[grid(256)](arg1_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
return reinterpret_tensor(arg0_1, (256,), (1,), 0
), buf0, reinterpret_tensor(arg1_1, (256,), (1,), 0)
def flatten_binary_scores(scores, labels, ignore=255):
"""
Flattens predictions in the batch (binary case)
Remove labels equal to 'ignore'
"""
scores = scores.view(-1)
labels = labels.view(-1)
if ignore is None:
return scores, labels
valid = labels != ignore
vscores = scores[valid]
vlabels = labels[valid]
return vscores, vlabels
def lovasz_grad(gt_sorted):
"""
Computes gradient of the Lovasz extension w.r.t sorted errors
See Alg. 1 in paper
"""
p = len(gt_sorted)
gts = gt_sorted.sum()
intersection = gts.float() - gt_sorted.float().cumsum(0)
union = gts.float() + (1 - gt_sorted).float().cumsum(0)
jaccard = 1.0 - intersection / union
if p > 1:
jaccard[1:p] = jaccard[1:p] - jaccard[0:-1]
return jaccard
def isnan(x):
return x != x
def mean(l, ignore_nan=False, empty=0):
"""
nanmean compatible with generators.
"""
l = iter(l)
if ignore_nan:
l = ifilterfalse(isnan, l)
try:
n = 1
acc = next(l)
except StopIteration:
if empty == 'raise':
raise ValueError('Empty mean')
return empty
for n, v in enumerate(l, 2):
acc += v
if n == 1:
return acc
return acc / n
class lovasz_hingeNew(torch.nn.Module):
def __init__(self, per_img=False, ignore=255):
"""
:param weight: 1D weight vector to deal with the class-imbalance
"""
super().__init__()
self.per_image = per_img
self.ignore = ignore
def lovasz_hinge_flat(self, logits, labels):
"""
Binary Lovasz hinge loss
logits: [P] Variable, logits at each prediction (between -\\infty and +\\infty)
labels: [P] Tensor, binary ground truth labels (0 or 1)
ignore: label to ignore
"""
if len(labels) == 0:
return logits.sum() * 0.0
signs = 2.0 * labels.float() - 1.0
errors = 1.0 - logits * Variable(signs)
errors_sorted, perm = torch.sort(errors, dim=0, descending=True)
perm = perm.data
gt_sorted = labels[perm]
grad = lovasz_grad(gt_sorted)
loss = torch.dot(F.relu(errors_sorted), Variable(grad))
return loss
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| PhillipHuang2017/ext_portrait_segmentation | lovasz_hinge | false | 1,111 | [
"MIT"
] | 0 | 6d0cec0a953dacbc94a01ea8b719feb687b7c029 | https://github.com/PhillipHuang2017/ext_portrait_segmentation/tree/6d0cec0a953dacbc94a01ea8b719feb687b7c029 | import torch
import torch.nn.parallel
import torch.utils.data
from torchvision.transforms import functional as F
import torch.nn.functional as F
from torch.autograd import Variable
def flatten_binary_scores(scores, labels, ignore=255):
"""
Flattens predictions in the batch (binary case)
Remove labels equal to 'ignore'
"""
scores = scores.view(-1)
labels = labels.view(-1)
if ignore is None:
return scores, labels
valid = labels != ignore
vscores = scores[valid]
vlabels = labels[valid]
return vscores, vlabels
def lovasz_grad(gt_sorted):
"""
Computes gradient of the Lovasz extension w.r.t sorted errors
See Alg. 1 in paper
"""
p = len(gt_sorted)
gts = gt_sorted.sum()
intersection = gts.float() - gt_sorted.float().cumsum(0)
union = gts.float() + (1 - gt_sorted).float().cumsum(0)
jaccard = 1.0 - intersection / union
if p > 1:
jaccard[1:p] = jaccard[1:p] - jaccard[0:-1]
return jaccard
def isnan(x):
return x != x
def mean(l, ignore_nan=False, empty=0):
"""
nanmean compatible with generators.
"""
l = iter(l)
if ignore_nan:
l = ifilterfalse(isnan, l)
try:
n = 1
acc = next(l)
except StopIteration:
if empty == 'raise':
raise ValueError('Empty mean')
return empty
for n, v in enumerate(l, 2):
acc += v
if n == 1:
return acc
return acc / n
class Model(torch.nn.Module):
def __init__(self, per_img=False, ignore=255):
"""
:param weight: 1D weight vector to deal with the class-imbalance
"""
super().__init__()
self.per_image = per_img
self.ignore = ignore
def lovasz_hinge_flat(self, logits, labels):
"""
Binary Lovasz hinge loss
logits: [P] Variable, logits at each prediction (between -\\infty and +\\infty)
labels: [P] Tensor, binary ground truth labels (0 or 1)
ignore: label to ignore
"""
if len(labels) == 0:
return logits.sum() * 0.0
signs = 2.0 * labels.float() - 1.0
errors = 1.0 - logits * Variable(signs)
errors_sorted, perm = torch.sort(errors, dim=0, descending=True)
perm = perm.data
gt_sorted = labels[perm]
grad = lovasz_grad(gt_sorted)
loss = torch.dot(F.relu(errors_sorted), Variable(grad))
return loss
def forward(self, logits, labels):
"""
Binary Lovasz hinge loss
logits: [B, H, W] Variable, logits at each pixel (between -\\infty and +\\infty)
labels: [B, H, W] Tensor, binary ground truth masks (0 or 1)
per_image: compute the loss per image instead of per batch
ignore: void class id
"""
if self.per_image:
loss = mean(self.lovasz_hinge_flat(*flatten_binary_scores(log.
unsqueeze(0), lab.unsqueeze(0), self.ignore)) for log, lab in
zip(logits, labels))
else:
loss = self.lovasz_hinge_flat(*flatten_binary_scores(logits,
labels, self.ignore))
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
AlignEA | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/az/cazasdcnfd4vdvnvqp76gldss6jpfbd546uzzb2vc5bkonvjunrj.py
# Topologically Sorted Source Nodes: [add, sub, pow_1, sum_1], Original ATen: [aten.add, aten.sub, aten.pow, aten.sum]
# Source node to ATen node mapping:
# add => add
# pow_1 => pow_1
# sub => sub
# sum_1 => sum_1
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %arg2_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1]), kwargs = {})
triton_poi_fused_add_pow_sub_sum_0 = async_compile.triton('triton_poi_fused_add_pow_sub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_pow_sub_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_pow_sub_sum_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr2 + (x0 + (64*x1)), xmask)
tmp6 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp7 = tl.load(in_ptr1 + (16 + x0 + (64*x1)), xmask)
tmp9 = tl.load(in_ptr2 + (16 + x0 + (64*x1)), xmask)
tmp13 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp14 = tl.load(in_ptr1 + (32 + x0 + (64*x1)), xmask)
tmp16 = tl.load(in_ptr2 + (32 + x0 + (64*x1)), xmask)
tmp20 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp21 = tl.load(in_ptr1 + (48 + x0 + (64*x1)), xmask)
tmp23 = tl.load(in_ptr2 + (48 + x0 + (64*x1)), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp5 = tmp4 * tmp4
tmp8 = tmp6 + tmp7
tmp10 = tmp8 - tmp9
tmp11 = tmp10 * tmp10
tmp12 = tmp5 + tmp11
tmp15 = tmp13 + tmp14
tmp17 = tmp15 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp12 + tmp18
tmp22 = tmp20 + tmp21
tmp24 = tmp22 - tmp23
tmp25 = tmp24 * tmp24
tmp26 = tmp19 + tmp25
tl.store(out_ptr0 + (x2), tmp26, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, sub, pow_1, sum_1], Original ATen: [aten.add, aten.sub, aten.pow, aten.sum]
stream0 = get_raw_stream(0)
triton_poi_fused_add_pow_sub_sum_0.run(arg0_1, arg1_1, arg2_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
class AlignEA(torch.nn.Module):
def __init__(self, p, feat_drop, params):
super(AlignEA, self).__init__()
self.params = params
def forward(self, e1, r, e2):
return torch.sum(torch.pow(e1 + r - e2, 2), 1)
def only_pos_loss(self, e1, r, e2):
return -F.logsigmoid(-torch.sum(torch.pow(e1 + r - e2, 2), 1)).sum()
def loss(self, pos_score, neg_score, target):
return F.relu(pos_score - self.params[0]).sum() + self.params[1
] * F.relu(self.params[2] - neg_score).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [[], {'p': 4, 'feat_drop': 4, 'params': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_pow_sub_sum_0(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr2 + (x0 + 64 * x1), xmask)
tmp6 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp7 = tl.load(in_ptr1 + (16 + x0 + 64 * x1), xmask)
tmp9 = tl.load(in_ptr2 + (16 + x0 + 64 * x1), xmask)
tmp13 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp14 = tl.load(in_ptr1 + (32 + x0 + 64 * x1), xmask)
tmp16 = tl.load(in_ptr2 + (32 + x0 + 64 * x1), xmask)
tmp20 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp21 = tl.load(in_ptr1 + (48 + x0 + 64 * x1), xmask)
tmp23 = tl.load(in_ptr2 + (48 + x0 + 64 * x1), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp5 = tmp4 * tmp4
tmp8 = tmp6 + tmp7
tmp10 = tmp8 - tmp9
tmp11 = tmp10 * tmp10
tmp12 = tmp5 + tmp11
tmp15 = tmp13 + tmp14
tmp17 = tmp15 - tmp16
tmp18 = tmp17 * tmp17
tmp19 = tmp12 + tmp18
tmp22 = tmp20 + tmp21
tmp24 = tmp22 - tmp23
tmp25 = tmp24 * tmp24
tmp26 = tmp19 + tmp25
tl.store(out_ptr0 + x2, tmp26, xmask)
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_pow_sub_sum_0[grid(64)](arg0_1, arg1_1, arg2_1,
buf0, 64, XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
del arg1_1
del arg2_1
return buf0,
class AlignEANew(torch.nn.Module):
def __init__(self, p, feat_drop, params):
super(AlignEANew, self).__init__()
self.params = params
def only_pos_loss(self, e1, r, e2):
return -F.logsigmoid(-torch.sum(torch.pow(e1 + r - e2, 2), 1)).sum()
def loss(self, pos_score, neg_score, target):
return F.relu(pos_score - self.params[0]).sum() + self.params[1
] * F.relu(self.params[2] - neg_score).sum()
def forward(self, input_0, input_1, input_2):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
output = call([arg0_1, arg1_1, arg2_1])
return output[0]
| TMUITLab/EAFR | AlignEA | false | 1,112 | [
"MIT"
] | 0 | dadb6485d48711ccb8aa2f03760aeb437645f1ff | https://github.com/TMUITLab/EAFR/tree/dadb6485d48711ccb8aa2f03760aeb437645f1ff | import torch
import torch.nn.functional as F
class Model(torch.nn.Module):
def __init__(self, p, feat_drop, params):
super().__init__()
self.params = params
def forward(self, e1, r, e2):
return torch.sum(torch.pow(e1 + r - e2, 2), 1)
def only_pos_loss(self, e1, r, e2):
return -F.logsigmoid(-torch.sum(torch.pow(e1 + r - e2, 2), 1)).sum()
def loss(self, pos_score, neg_score, target):
return F.relu(pos_score - self.params[0]).sum() + self.params[1
] * F.relu(self.params[2] - neg_score).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
fpn_module | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/w5/cw5skfgpv5nrk42leelcp4yltyqwex7mnbyce3n3bj4ebm2z7azu.py
# Topologically Sorted Source Nodes: [p5], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# p5 => convolution
# Graph fragment:
# %convolution : [num_users=6] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_0 = async_compile.triton('triton_poi_fused_convolution_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4194304],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4194304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 4096) % 256
tmp0 = tl.load(in_out_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ib/cib2eq5hg22wfnjfpqps64l2xqhjhgautrnryu7xii6hu5r6ufgg.py
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten._to_copy]
# Source node to ATen node mapping:
# interpolate => convert_element_type_1
# Graph fragment:
# %convert_element_type_1 : [num_users=15] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%view, torch.int64), kwargs = {})
triton_poi_fused__to_copy_1 = async_compile.triton('triton_poi_fused__to_copy_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*i64', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__to_copy_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__to_copy_1(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp5.to(tl.int32)
tl.store(out_ptr0 + (x0), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/wr/cwrn4qfpnl5h5jxhhbi2ktzjxlgmnqsqet77wksszluqa7tvrvn2.py
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.add, aten.clamp]
# Source node to ATen node mapping:
# interpolate => add, clamp_max
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convert_element_type_1, 1), kwargs = {})
# %clamp_max : [num_users=13] = call_function[target=torch.ops.aten.clamp_max.default](args = (%add, 63), kwargs = {})
triton_poi_fused_add_clamp_2 = async_compile.triton('triton_poi_fused_add_clamp_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*i64', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_clamp_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_clamp_2(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp5.to(tl.int32)
tmp7 = tl.full([1], 1, tl.int64)
tmp8 = tmp6 + tmp7
tmp9 = tl.full([1], 63, tl.int64)
tmp10 = triton_helpers.minimum(tmp8, tmp9)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/rk/crkhx22brl3wt2ooij4yqrdzmhxwkdsxlxgfpso5oebqhikf32hy.py
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.arange, aten._to_copy, aten.mul, aten.clamp, aten.sub]
# Source node to ATen node mapping:
# interpolate => clamp_max_2, clamp_min, clamp_min_2, convert_element_type, iota, mul, sub
# Graph fragment:
# %iota : [num_users=1] = call_function[target=torch.ops.prims.iota.default](args = (64,), kwargs = {start: 0, step: 1, dtype: torch.int64, device: cuda:0, requires_grad: False})
# %convert_element_type : [num_users=1] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%iota, torch.float32), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convert_element_type, 1.0), kwargs = {})
# %clamp_min : [num_users=3] = call_function[target=torch.ops.aten.clamp_min.default](args = (%mul, 0.0), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%clamp_min, %convert_element_type_3), kwargs = {})
# %clamp_min_2 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sub, 0.0), kwargs = {})
# %clamp_max_2 : [num_users=13] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min_2, 1.0), kwargs = {})
triton_poi_fused__to_copy_arange_clamp_mul_sub_3 = async_compile.triton('triton_poi_fused__to_copy_arange_clamp_mul_sub_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__to_copy_arange_clamp_mul_sub_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__to_copy_arange_clamp_mul_sub_3(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp5.to(tl.int32)
tmp7 = tmp6.to(tl.float32)
tmp8 = tmp5 - tmp7
tmp9 = triton_helpers.maximum(tmp8, tmp4)
tmp10 = triton_helpers.minimum(tmp9, tmp2)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/dn/cdno47sb4iuwlzp3u5bnrx5nfsdaxpfpneo66rlgu43bh4opcxpe.py
# Topologically Sorted Source Nodes: [conv2d_1, interpolate, p4], Original ATen: [aten.convolution, aten._unsafe_index, aten.sub, aten.mul, aten.add]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# interpolate => _unsafe_index, _unsafe_index_1, _unsafe_index_2, _unsafe_index_3, add_2, add_3, add_4, mul_2, mul_3, mul_4, sub_1, sub_2, sub_4
# p4 => add_5
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_6, %primals_4, %primals_5, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %_unsafe_index : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution, [None, None, %convert_element_type_1, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_1 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution, [None, None, %convert_element_type_1, %clamp_max_1]), kwargs = {})
# %_unsafe_index_2 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution, [None, None, %clamp_max, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_3 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution, [None, None, %clamp_max, %clamp_max_1]), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_1, %_unsafe_index), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %clamp_max_2), kwargs = {})
# %add_2 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index, %mul_2), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_3, %_unsafe_index_2), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_2, %clamp_max_2), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_2, %mul_3), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_3, %add_2), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_4, %clamp_max_3), kwargs = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_2, %mul_4), kwargs = {})
# %add_5 : [num_users=6] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_4, %convolution_1), kwargs = {})
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4 = async_compile.triton('triton_poi_fused__unsafe_index_add_convolution_mul_sub_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4194304],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i64', 2: '*i64', 3: '*fp32', 4: '*i64', 5: '*fp32', 6: '*i64', 7: '*fp32', 8: '*fp32', 9: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_index_add_convolution_mul_sub_4', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__unsafe_index_add_convolution_mul_sub_4(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, xnumel, XBLOCK : tl.constexpr):
xnumel = 4194304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x1 = (xindex // 64) % 64
x0 = xindex % 64
x2 = (xindex // 4096)
x6 = xindex
x3 = (xindex // 4096) % 256
tmp0 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (x0), None, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr3 + (x0), None, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr4 + (x0), None, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr5 + (x1), None, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr6 + (x1), None, eviction_policy='evict_last')
tmp32 = tl.load(in_out_ptr0 + (x6), None)
tmp33 = tl.load(in_ptr7 + (x3), None, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 64, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tmp6 = tmp5 + tmp1
tmp7 = tmp5 < 0
tmp8 = tl.where(tmp7, tmp6, tmp5)
tmp9 = tl.load(in_ptr2 + (tmp8 + (64*tmp4) + (4096*x2)), None, eviction_policy='evict_last')
tmp11 = tmp10 + tmp1
tmp12 = tmp10 < 0
tmp13 = tl.where(tmp12, tmp11, tmp10)
tmp14 = tl.load(in_ptr2 + (tmp13 + (64*tmp4) + (4096*x2)), None, eviction_policy='evict_last')
tmp15 = tmp14 - tmp9
tmp17 = tmp15 * tmp16
tmp18 = tmp9 + tmp17
tmp20 = tmp19 + tmp1
tmp21 = tmp19 < 0
tmp22 = tl.where(tmp21, tmp20, tmp19)
tmp23 = tl.load(in_ptr2 + (tmp8 + (64*tmp22) + (4096*x2)), None, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr2 + (tmp13 + (64*tmp22) + (4096*x2)), None, eviction_policy='evict_last')
tmp25 = tmp24 - tmp23
tmp26 = tmp25 * tmp16
tmp27 = tmp23 + tmp26
tmp28 = tmp27 - tmp18
tmp30 = tmp28 * tmp29
tmp31 = tmp18 + tmp30
tmp34 = tmp32 + tmp33
tmp35 = tmp31 + tmp34
tl.store(in_out_ptr0 + (x6), tmp35, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/he/che3l4rv6rafbjk6jebrweq77ki4pey5sowezfrn6tmor5h7vaw6.py
# Topologically Sorted Source Nodes: [p5_1, p4_1, p3_1, p5_2, p4_2, p3_2], Original ATen: [aten.convolution, aten._unsafe_index, aten.sub, aten.mul, aten.add]
# Source node to ATen node mapping:
# p3_1 => convolution_9
# p3_2 => _unsafe_index_20, _unsafe_index_21, _unsafe_index_22, _unsafe_index_23, add_30, add_31, mul_27, mul_28, mul_29, sub_26, sub_27, sub_29
# p4_1 => convolution_7
# p4_2 => _unsafe_index_16, _unsafe_index_17, _unsafe_index_18, _unsafe_index_19, add_25, add_26, mul_22, mul_23, mul_24, sub_21, sub_22, sub_24
# p5_1 => convolution_5
# p5_2 => _unsafe_index_12, _unsafe_index_13, _unsafe_index_14, _unsafe_index_15, add_20, add_21, mul_17, mul_18, mul_19, sub_16, sub_17, sub_19
# Graph fragment:
# %convolution_5 : [num_users=4] = call_function[target=torch.ops.aten.convolution.default](args = (%convolution_4, %primals_15, %primals_16, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_7 : [num_users=4] = call_function[target=torch.ops.aten.convolution.default](args = (%convolution_6, %primals_19, %primals_20, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %convolution_9 : [num_users=4] = call_function[target=torch.ops.aten.convolution.default](args = (%convolution_8, %primals_23, %primals_24, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %_unsafe_index_12 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_5, [None, None, %convert_element_type_1, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_13 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_5, [None, None, %convert_element_type_1, %clamp_max_1]), kwargs = {})
# %_unsafe_index_14 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_5, [None, None, %clamp_max, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_15 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_5, [None, None, %clamp_max, %clamp_max_1]), kwargs = {})
# %sub_16 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_13, %_unsafe_index_12), kwargs = {})
# %mul_17 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_16, %clamp_max_2), kwargs = {})
# %add_20 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_12, %mul_17), kwargs = {})
# %sub_17 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_15, %_unsafe_index_14), kwargs = {})
# %mul_18 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_17, %clamp_max_2), kwargs = {})
# %add_21 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_14, %mul_18), kwargs = {})
# %sub_19 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_21, %add_20), kwargs = {})
# %mul_19 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_19, %clamp_max_3), kwargs = {})
# %_unsafe_index_16 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_7, [None, None, %convert_element_type_1, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_17 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_7, [None, None, %convert_element_type_1, %clamp_max_1]), kwargs = {})
# %_unsafe_index_18 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_7, [None, None, %clamp_max, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_19 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_7, [None, None, %clamp_max, %clamp_max_1]), kwargs = {})
# %sub_21 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_17, %_unsafe_index_16), kwargs = {})
# %mul_22 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_21, %clamp_max_2), kwargs = {})
# %add_25 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_16, %mul_22), kwargs = {})
# %sub_22 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_19, %_unsafe_index_18), kwargs = {})
# %mul_23 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_22, %clamp_max_2), kwargs = {})
# %add_26 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_18, %mul_23), kwargs = {})
# %sub_24 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_26, %add_25), kwargs = {})
# %mul_24 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_24, %clamp_max_3), kwargs = {})
# %_unsafe_index_20 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_9, [None, None, %convert_element_type_1, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_21 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_9, [None, None, %convert_element_type_1, %clamp_max_1]), kwargs = {})
# %_unsafe_index_22 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_9, [None, None, %clamp_max, %convert_element_type_3]), kwargs = {})
# %_unsafe_index_23 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%convolution_9, [None, None, %clamp_max, %clamp_max_1]), kwargs = {})
# %sub_26 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_21, %_unsafe_index_20), kwargs = {})
# %mul_27 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_26, %clamp_max_2), kwargs = {})
# %add_30 : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_20, %mul_27), kwargs = {})
# %sub_27 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%_unsafe_index_23, %_unsafe_index_22), kwargs = {})
# %mul_28 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_27, %clamp_max_2), kwargs = {})
# %add_31 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%_unsafe_index_22, %mul_28), kwargs = {})
# %sub_29 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add_31, %add_30), kwargs = {})
# %mul_29 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_29, %clamp_max_3), kwargs = {})
triton_poi_fused__unsafe_index_add_convolution_mul_sub_5 = async_compile.triton('triton_poi_fused__unsafe_index_add_convolution_mul_sub_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[2097152],
filename=__file__,
triton_meta={'signature': {0: '*i64', 1: '*i64', 2: '*fp32', 3: '*fp32', 4: '*i64', 5: '*fp32', 6: '*i64', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: '*fp32', 11: '*fp32', 12: '*fp32', 13: '*fp32', 14: '*fp32', 15: '*fp32', 16: '*fp32', 17: '*fp32', 18: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_index_add_convolution_mul_sub_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__unsafe_index_add_convolution_mul_sub_5(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8, in_ptr9, in_ptr10, in_ptr11, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, xnumel, XBLOCK : tl.constexpr):
xnumel = 2097152
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x1 = (xindex // 64) % 64
x0 = xindex % 64
x5 = (xindex // 4096)
x2 = (xindex // 4096) % 128
x6 = xindex
tmp0 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (x0), None, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr3 + (x2), None, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr4 + (x0), None, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr5 + (x0), None, eviction_policy='evict_last')
tmp22 = tl.load(in_ptr6 + (x1), None, eviction_policy='evict_last')
tmp34 = tl.load(in_ptr7 + (x1), None, eviction_policy='evict_last')
tmp37 = tl.load(in_ptr9 + (x2), None, eviction_policy='evict_last')
tmp54 = tl.load(in_ptr11 + (x2), None, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 64, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tmp6 = tmp5 + tmp1
tmp7 = tmp5 < 0
tmp8 = tl.where(tmp7, tmp6, tmp5)
tmp9 = tl.load(in_ptr2 + (tmp8 + (64*tmp4) + (4096*x5)), None, eviction_policy='evict_last')
tmp11 = tmp9 + tmp10
tmp13 = tmp12 + tmp1
tmp14 = tmp12 < 0
tmp15 = tl.where(tmp14, tmp13, tmp12)
tmp16 = tl.load(in_ptr2 + (tmp15 + (64*tmp4) + (4096*x5)), None, eviction_policy='evict_last')
tmp17 = tmp16 + tmp10
tmp18 = tmp17 - tmp11
tmp20 = tmp18 * tmp19
tmp21 = tmp11 + tmp20
tmp23 = tmp22 + tmp1
tmp24 = tmp22 < 0
tmp25 = tl.where(tmp24, tmp23, tmp22)
tmp26 = tl.load(in_ptr2 + (tmp8 + (64*tmp25) + (4096*x5)), None, eviction_policy='evict_last')
tmp27 = tmp26 + tmp10
tmp28 = tl.load(in_ptr2 + (tmp15 + (64*tmp25) + (4096*x5)), None, eviction_policy='evict_last')
tmp29 = tmp28 + tmp10
tmp30 = tmp29 - tmp27
tmp31 = tmp30 * tmp19
tmp32 = tmp27 + tmp31
tmp33 = tmp32 - tmp21
tmp35 = tmp33 * tmp34
tmp36 = tl.load(in_ptr8 + (tmp8 + (64*tmp4) + (4096*x5)), None, eviction_policy='evict_last')
tmp38 = tmp36 + tmp37
tmp39 = tl.load(in_ptr8 + (tmp15 + (64*tmp4) + (4096*x5)), None, eviction_policy='evict_last')
tmp40 = tmp39 + tmp37
tmp41 = tmp40 - tmp38
tmp42 = tmp41 * tmp19
tmp43 = tmp38 + tmp42
tmp44 = tl.load(in_ptr8 + (tmp8 + (64*tmp25) + (4096*x5)), None, eviction_policy='evict_last')
tmp45 = tmp44 + tmp37
tmp46 = tl.load(in_ptr8 + (tmp15 + (64*tmp25) + (4096*x5)), None, eviction_policy='evict_last')
tmp47 = tmp46 + tmp37
tmp48 = tmp47 - tmp45
tmp49 = tmp48 * tmp19
tmp50 = tmp45 + tmp49
tmp51 = tmp50 - tmp43
tmp52 = tmp51 * tmp34
tmp53 = tl.load(in_ptr10 + (tmp8 + (64*tmp4) + (4096*x5)), None, eviction_policy='evict_last')
tmp55 = tmp53 + tmp54
tmp56 = tl.load(in_ptr10 + (tmp15 + (64*tmp4) + (4096*x5)), None, eviction_policy='evict_last')
tmp57 = tmp56 + tmp54
tmp58 = tmp57 - tmp55
tmp59 = tmp58 * tmp19
tmp60 = tmp55 + tmp59
tmp61 = tl.load(in_ptr10 + (tmp8 + (64*tmp25) + (4096*x5)), None, eviction_policy='evict_last')
tmp62 = tmp61 + tmp54
tmp63 = tl.load(in_ptr10 + (tmp15 + (64*tmp25) + (4096*x5)), None, eviction_policy='evict_last')
tmp64 = tmp63 + tmp54
tmp65 = tmp64 - tmp62
tmp66 = tmp65 * tmp19
tmp67 = tmp62 + tmp66
tmp68 = tmp67 - tmp60
tmp69 = tmp68 * tmp34
tl.store(out_ptr0 + (x6), tmp21, None)
tl.store(out_ptr1 + (x6), tmp35, None)
tl.store(out_ptr2 + (x6), tmp43, None)
tl.store(out_ptr3 + (x6), tmp52, None)
tl.store(out_ptr4 + (x6), tmp60, None)
tl.store(out_ptr5 + (x6), tmp69, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/qq/cqqy6dwgchmh6da5wvexvll4djghft4xnm4crpatgerre2lzchix.py
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# cat => cat
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%add_22, %add_27, %add_32, %convolution_11], 1), kwargs = {})
triton_poi_fused_cat_6 = async_compile.triton('triton_poi_fused_cat_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8388608],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8, 9), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_6(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8388608
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x1 = (xindex // 4096) % 512
x0 = xindex % 4096
x2 = (xindex // 2097152)
x3 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 128, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + (4096*x1) + (524288*x2)), tmp4, other=0.0)
tmp6 = tl.load(in_ptr1 + (x0 + (4096*x1) + (524288*x2)), tmp4, other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tmp11 = tl.full([1], 256, tl.int64)
tmp12 = tmp0 < tmp11
tmp13 = tmp10 & tmp12
tmp14 = tl.load(in_ptr2 + (x0 + (4096*((-128) + x1)) + (524288*x2)), tmp13, other=0.0)
tmp15 = tl.load(in_ptr3 + (x0 + (4096*((-128) + x1)) + (524288*x2)), tmp13, other=0.0)
tmp16 = tmp14 + tmp15
tmp17 = tl.full(tmp16.shape, 0.0, tmp16.dtype)
tmp18 = tl.where(tmp13, tmp16, tmp17)
tmp19 = tmp0 >= tmp11
tmp20 = tl.full([1], 384, tl.int64)
tmp21 = tmp0 < tmp20
tmp22 = tmp19 & tmp21
tmp23 = tl.load(in_ptr4 + (x0 + (4096*((-256) + x1)) + (524288*x2)), tmp22, other=0.0)
tmp24 = tl.load(in_ptr5 + (x0 + (4096*((-256) + x1)) + (524288*x2)), tmp22, other=0.0)
tmp25 = tmp23 + tmp24
tmp26 = tl.full(tmp25.shape, 0.0, tmp25.dtype)
tmp27 = tl.where(tmp22, tmp25, tmp26)
tmp28 = tmp0 >= tmp20
tmp29 = tl.full([1], 512, tl.int64)
tmp30 = tmp0 < tmp29
tmp31 = tl.load(in_ptr6 + (x0 + (4096*((-384) + x1)) + (524288*x2)), tmp28, other=0.0)
tmp32 = tl.load(in_ptr7 + ((-384) + x1), tmp28, eviction_policy='evict_last', other=0.0)
tmp33 = tmp31 + tmp32
tmp34 = tl.full(tmp33.shape, 0.0, tmp33.dtype)
tmp35 = tl.where(tmp28, tmp33, tmp34)
tmp36 = tl.where(tmp22, tmp27, tmp35)
tmp37 = tl.where(tmp13, tmp18, tmp36)
tmp38 = tl.where(tmp4, tmp9, tmp37)
tl.store(out_ptr0 + (x3), tmp38, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/aa/caad7kyjzosottzpivtkindhaxpfkjxgqczfry2hxfky2fvdwwvu.py
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# output => convolution_12
# Graph fragment:
# %convolution_12 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%cat, %primals_29, %primals_30, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_7 = async_compile.triton('triton_poi_fused_convolution_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_7', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_7(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 65536
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 4096) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr0 + (x1), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25, primals_26, primals_27, primals_28, primals_29, primals_30 = args
args.clear()
assert_size_stride(primals_1, (256, 2048, 1, 1), (2048, 1, 1, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 2048, 64, 64), (8388608, 4096, 64, 1))
assert_size_stride(primals_4, (256, 1024, 1, 1), (1024, 1, 1, 1))
assert_size_stride(primals_5, (256, ), (1, ))
assert_size_stride(primals_6, (4, 1024, 64, 64), (4194304, 4096, 64, 1))
assert_size_stride(primals_7, (256, 512, 1, 1), (512, 1, 1, 1))
assert_size_stride(primals_8, (256, ), (1, ))
assert_size_stride(primals_9, (4, 512, 64, 64), (2097152, 4096, 64, 1))
assert_size_stride(primals_10, (256, 256, 1, 1), (256, 1, 1, 1))
assert_size_stride(primals_11, (256, ), (1, ))
assert_size_stride(primals_12, (4, 256, 64, 64), (1048576, 4096, 64, 1))
assert_size_stride(primals_13, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_14, (256, ), (1, ))
assert_size_stride(primals_15, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_16, (128, ), (1, ))
assert_size_stride(primals_17, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_18, (256, ), (1, ))
assert_size_stride(primals_19, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_20, (128, ), (1, ))
assert_size_stride(primals_21, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_22, (256, ), (1, ))
assert_size_stride(primals_23, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_24, (128, ), (1, ))
assert_size_stride(primals_25, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_26, (256, ), (1, ))
assert_size_stride(primals_27, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_28, (128, ), (1, ))
assert_size_stride(primals_29, (4, 512, 3, 3), (4608, 9, 3, 1))
assert_size_stride(primals_30, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [p5], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [p5], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf1, primals_2, 4194304, grid=grid(4194304), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(primals_6, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf3 = empty_strided_cuda((64, 1), (1, 1), torch.int64)
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten._to_copy]
triton_poi_fused__to_copy_1.run(buf3, 64, grid=grid(64), stream=stream0)
buf4 = empty_strided_cuda((64, 1), (1, 1), torch.int64)
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.add, aten.clamp]
triton_poi_fused_add_clamp_2.run(buf4, 64, grid=grid(64), stream=stream0)
buf5 = empty_strided_cuda((64, ), (1, ), torch.int64)
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.arange, aten._to_copy, aten.mul, aten.clamp]
triton_poi_fused__to_copy_1.run(buf5, 64, grid=grid(64), stream=stream0)
buf6 = empty_strided_cuda((64, ), (1, ), torch.int64)
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.add, aten.clamp]
triton_poi_fused_add_clamp_2.run(buf6, 64, grid=grid(64), stream=stream0)
buf7 = empty_strided_cuda((64, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.arange, aten._to_copy, aten.mul, aten.clamp, aten.sub]
triton_poi_fused__to_copy_arange_clamp_mul_sub_3.run(buf7, 64, grid=grid(64), stream=stream0)
buf9 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [interpolate], Original ATen: [aten.sub, aten.clamp]
triton_poi_fused__to_copy_arange_clamp_mul_sub_3.run(buf9, 64, grid=grid(64), stream=stream0)
buf10 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [conv2d_1, interpolate, p4], Original ATen: [aten.convolution, aten._unsafe_index, aten.sub, aten.mul, aten.add]
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4.run(buf10, buf3, buf5, buf1, buf6, buf7, buf4, buf9, primals_5, 4194304, grid=grid(4194304), stream=stream0)
del primals_5
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf11 = extern_kernels.convolution(primals_9, primals_7, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf11, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf13 = buf11; del buf11 # reuse
# Topologically Sorted Source Nodes: [conv2d_2, interpolate_1, p3], Original ATen: [aten.convolution, aten._unsafe_index, aten.sub, aten.mul, aten.add]
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4.run(buf13, buf3, buf5, buf10, buf6, buf7, buf4, buf9, primals_8, 4194304, grid=grid(4194304), stream=stream0)
del primals_8
# Topologically Sorted Source Nodes: [conv2d_3], Original ATen: [aten.convolution]
buf14 = extern_kernels.convolution(primals_12, primals_10, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf14, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf16 = buf14; del buf14 # reuse
# Topologically Sorted Source Nodes: [conv2d_3, interpolate_2, p2], Original ATen: [aten.convolution, aten._unsafe_index, aten.sub, aten.mul, aten.add]
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4.run(buf16, buf3, buf5, buf13, buf6, buf7, buf4, buf9, primals_11, 4194304, grid=grid(4194304), stream=stream0)
del primals_11
# Topologically Sorted Source Nodes: [conv2d_4], Original ATen: [aten.convolution]
buf17 = extern_kernels.convolution(buf1, primals_13, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf17, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf18 = buf17; del buf17 # reuse
# Topologically Sorted Source Nodes: [conv2d_4], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf18, primals_14, 4194304, grid=grid(4194304), stream=stream0)
del primals_14
# Topologically Sorted Source Nodes: [p5_1], Original ATen: [aten.convolution]
buf19 = extern_kernels.convolution(buf18, primals_15, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf19, (4, 128, 64, 64), (524288, 4096, 64, 1))
# Topologically Sorted Source Nodes: [conv2d_6], Original ATen: [aten.convolution]
buf20 = extern_kernels.convolution(buf10, primals_17, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf20, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf21 = buf20; del buf20 # reuse
# Topologically Sorted Source Nodes: [conv2d_6], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf21, primals_18, 4194304, grid=grid(4194304), stream=stream0)
del primals_18
# Topologically Sorted Source Nodes: [p4_1], Original ATen: [aten.convolution]
buf22 = extern_kernels.convolution(buf21, primals_19, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf22, (4, 128, 64, 64), (524288, 4096, 64, 1))
# Topologically Sorted Source Nodes: [conv2d_8], Original ATen: [aten.convolution]
buf23 = extern_kernels.convolution(buf13, primals_21, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf23, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf24 = buf23; del buf23 # reuse
# Topologically Sorted Source Nodes: [conv2d_8], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf24, primals_22, 4194304, grid=grid(4194304), stream=stream0)
del primals_22
# Topologically Sorted Source Nodes: [p3_1], Original ATen: [aten.convolution]
buf25 = extern_kernels.convolution(buf24, primals_23, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf25, (4, 128, 64, 64), (524288, 4096, 64, 1))
# Topologically Sorted Source Nodes: [conv2d_10], Original ATen: [aten.convolution]
buf26 = extern_kernels.convolution(buf16, primals_25, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf26, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf27 = buf26; del buf26 # reuse
# Topologically Sorted Source Nodes: [conv2d_10], Original ATen: [aten.convolution]
triton_poi_fused_convolution_0.run(buf27, primals_26, 4194304, grid=grid(4194304), stream=stream0)
del primals_26
# Topologically Sorted Source Nodes: [p2_1], Original ATen: [aten.convolution]
buf28 = extern_kernels.convolution(buf27, primals_27, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf28, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf29 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1), torch.float32)
buf30 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1), torch.float32)
buf31 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1), torch.float32)
buf32 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1), torch.float32)
buf33 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1), torch.float32)
buf34 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1), torch.float32)
# Topologically Sorted Source Nodes: [p5_1, p4_1, p3_1, p5_2, p4_2, p3_2], Original ATen: [aten.convolution, aten._unsafe_index, aten.sub, aten.mul, aten.add]
triton_poi_fused__unsafe_index_add_convolution_mul_sub_5.run(buf3, buf5, buf19, primals_16, buf6, buf7, buf4, buf9, buf22, primals_20, buf25, primals_24, buf29, buf30, buf31, buf32, buf33, buf34, 2097152, grid=grid(2097152), stream=stream0)
del buf19
del buf22
del buf25
del primals_16
del primals_20
del primals_24
buf35 = empty_strided_cuda((4, 512, 64, 64), (2097152, 4096, 64, 1), torch.float32)
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
triton_poi_fused_cat_6.run(buf29, buf30, buf31, buf32, buf33, buf34, buf28, primals_28, buf35, 8388608, grid=grid(8388608), stream=stream0)
del buf28
del buf29
del buf30
del buf31
del buf32
del buf33
del buf34
del primals_28
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.convolution]
buf36 = extern_kernels.convolution(buf35, primals_29, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf36, (4, 4, 64, 64), (16384, 4096, 64, 1))
buf37 = buf36; del buf36 # reuse
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.convolution]
triton_poi_fused_convolution_7.run(buf37, primals_30, 65536, grid=grid(65536), stream=stream0)
del primals_30
return (buf37, primals_1, primals_3, primals_4, primals_6, primals_7, primals_9, primals_10, primals_12, primals_13, primals_15, primals_17, primals_19, primals_21, primals_23, primals_25, primals_27, primals_29, buf1, buf3, buf4, buf5, buf6, buf7, buf9, buf10, buf13, buf16, buf18, buf21, buf24, buf27, buf35, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((256, 2048, 1, 1), (2048, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 2048, 64, 64), (8388608, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((256, 1024, 1, 1), (1024, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 1024, 64, 64), (4194304, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((256, 512, 1, 1), (512, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 512, 64, 64), (2097152, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((256, 256, 1, 1), (256, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, 256, 64, 64), (1048576, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((128, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_19 = rand_strided((128, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_20 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_21 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_22 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_23 = rand_strided((128, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_24 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_25 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_26 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_27 = rand_strided((128, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_28 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_29 = rand_strided((4, 512, 3, 3), (4608, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_30 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25, primals_26, primals_27, primals_28, primals_29, primals_30])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
import torch.nn as nn
class fpn_module(nn.Module):
def __init__(self, numClass):
super(fpn_module, self).__init__()
self.toplayer = nn.Conv2d(2048, 256, kernel_size=1, stride=1, padding=0
)
self.smooth1_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth2_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth3_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth4_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth1_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth2_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth3_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth4_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.latlayer1 = nn.Conv2d(1024, 256, kernel_size=1, stride=1,
padding=0)
self.latlayer2 = nn.Conv2d(512, 256, kernel_size=1, stride=1, padding=0
)
self.latlayer3 = nn.Conv2d(256, 256, kernel_size=1, stride=1, padding=0
)
self.classify = nn.Conv2d(128 * 4, numClass, kernel_size=3, stride=
1, padding=1)
def _concatenate(self, p5, p4, p3, p2):
_, _, H, W = p2.size()
p5 = F.interpolate(p5, size=(H, W), mode='bilinear', align_corners=True
)
p4 = F.interpolate(p4, size=(H, W), mode='bilinear', align_corners=True
)
p3 = F.interpolate(p3, size=(H, W), mode='bilinear', align_corners=True
)
return torch.cat([p5, p4, p3, p2], dim=1)
def _upsample_add(self, x, y):
_, _, H, W = y.size()
return F.interpolate(x, size=(H, W), mode='bilinear', align_corners
=True) + y
def forward(self, c2, c3, c4, c5):
p5 = self.toplayer(c5)
p4 = self._upsample_add(p5, self.latlayer1(c4))
p3 = self._upsample_add(p4, self.latlayer2(c3))
p2 = self._upsample_add(p3, self.latlayer3(c2))
p5 = self.smooth1_2(self.smooth1_1(p5))
p4 = self.smooth2_2(self.smooth2_1(p4))
p3 = self.smooth3_2(self.smooth3_1(p3))
p2 = self.smooth4_2(self.smooth4_1(p2))
output = self.classify(self._concatenate(p5, p4, p3, p2))
return output
def get_inputs():
return [torch.rand([4, 256, 64, 64]), torch.rand([4, 512, 64, 64]),
torch.rand([4, 1024, 64, 64]), torch.rand([4, 2048, 64, 64])]
def get_init_inputs():
return [[], {'numClass': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
import torch.nn.functional as F
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_convolution_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = xindex // 4096 % 256
tmp0 = tl.load(in_out_ptr0 + x3, None)
tmp1 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, None)
@triton.jit
def triton_poi_fused__to_copy_1(out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp5.to(tl.int32)
tl.store(out_ptr0 + x0, tmp6, xmask)
@triton.jit
def triton_poi_fused_add_clamp_2(out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp5.to(tl.int32)
tmp7 = tl.full([1], 1, tl.int64)
tmp8 = tmp6 + tmp7
tmp9 = tl.full([1], 63, tl.int64)
tmp10 = triton_helpers.minimum(tmp8, tmp9)
tl.store(out_ptr0 + x0, tmp10, xmask)
@triton.jit
def triton_poi_fused__to_copy_arange_clamp_mul_sub_3(out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 0.0
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp5.to(tl.int32)
tmp7 = tmp6.to(tl.float32)
tmp8 = tmp5 - tmp7
tmp9 = triton_helpers.maximum(tmp8, tmp4)
tmp10 = triton_helpers.minimum(tmp9, tmp2)
tl.store(out_ptr0 + x0, tmp10, xmask)
@triton.jit
def triton_poi_fused__unsafe_index_add_convolution_mul_sub_4(in_out_ptr0,
in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7,
xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x1 = xindex // 64 % 64
x0 = xindex % 64
x2 = xindex // 4096
x6 = xindex
x3 = xindex // 4096 % 256
tmp0 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + x0, None, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr3 + x0, None, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr4 + x0, None, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr5 + x1, None, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr6 + x1, None, eviction_policy='evict_last')
tmp32 = tl.load(in_out_ptr0 + x6, None)
tmp33 = tl.load(in_ptr7 + x3, None, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 64, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tmp6 = tmp5 + tmp1
tmp7 = tmp5 < 0
tmp8 = tl.where(tmp7, tmp6, tmp5)
tmp9 = tl.load(in_ptr2 + (tmp8 + 64 * tmp4 + 4096 * x2), None,
eviction_policy='evict_last')
tmp11 = tmp10 + tmp1
tmp12 = tmp10 < 0
tmp13 = tl.where(tmp12, tmp11, tmp10)
tmp14 = tl.load(in_ptr2 + (tmp13 + 64 * tmp4 + 4096 * x2), None,
eviction_policy='evict_last')
tmp15 = tmp14 - tmp9
tmp17 = tmp15 * tmp16
tmp18 = tmp9 + tmp17
tmp20 = tmp19 + tmp1
tmp21 = tmp19 < 0
tmp22 = tl.where(tmp21, tmp20, tmp19)
tmp23 = tl.load(in_ptr2 + (tmp8 + 64 * tmp22 + 4096 * x2), None,
eviction_policy='evict_last')
tmp24 = tl.load(in_ptr2 + (tmp13 + 64 * tmp22 + 4096 * x2), None,
eviction_policy='evict_last')
tmp25 = tmp24 - tmp23
tmp26 = tmp25 * tmp16
tmp27 = tmp23 + tmp26
tmp28 = tmp27 - tmp18
tmp30 = tmp28 * tmp29
tmp31 = tmp18 + tmp30
tmp34 = tmp32 + tmp33
tmp35 = tmp31 + tmp34
tl.store(in_out_ptr0 + x6, tmp35, None)
@triton.jit
def triton_poi_fused__unsafe_index_add_convolution_mul_sub_5(in_ptr0,
in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, in_ptr8,
in_ptr9, in_ptr10, in_ptr11, out_ptr0, out_ptr1, out_ptr2, out_ptr3,
out_ptr4, out_ptr5, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x1 = xindex // 64 % 64
x0 = xindex % 64
x5 = xindex // 4096
x2 = xindex // 4096 % 128
x6 = xindex
tmp0 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + x0, None, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr3 + x2, None, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr4 + x0, None, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr5 + x0, None, eviction_policy='evict_last')
tmp22 = tl.load(in_ptr6 + x1, None, eviction_policy='evict_last')
tmp34 = tl.load(in_ptr7 + x1, None, eviction_policy='evict_last')
tmp37 = tl.load(in_ptr9 + x2, None, eviction_policy='evict_last')
tmp54 = tl.load(in_ptr11 + x2, None, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 64, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tmp6 = tmp5 + tmp1
tmp7 = tmp5 < 0
tmp8 = tl.where(tmp7, tmp6, tmp5)
tmp9 = tl.load(in_ptr2 + (tmp8 + 64 * tmp4 + 4096 * x5), None,
eviction_policy='evict_last')
tmp11 = tmp9 + tmp10
tmp13 = tmp12 + tmp1
tmp14 = tmp12 < 0
tmp15 = tl.where(tmp14, tmp13, tmp12)
tmp16 = tl.load(in_ptr2 + (tmp15 + 64 * tmp4 + 4096 * x5), None,
eviction_policy='evict_last')
tmp17 = tmp16 + tmp10
tmp18 = tmp17 - tmp11
tmp20 = tmp18 * tmp19
tmp21 = tmp11 + tmp20
tmp23 = tmp22 + tmp1
tmp24 = tmp22 < 0
tmp25 = tl.where(tmp24, tmp23, tmp22)
tmp26 = tl.load(in_ptr2 + (tmp8 + 64 * tmp25 + 4096 * x5), None,
eviction_policy='evict_last')
tmp27 = tmp26 + tmp10
tmp28 = tl.load(in_ptr2 + (tmp15 + 64 * tmp25 + 4096 * x5), None,
eviction_policy='evict_last')
tmp29 = tmp28 + tmp10
tmp30 = tmp29 - tmp27
tmp31 = tmp30 * tmp19
tmp32 = tmp27 + tmp31
tmp33 = tmp32 - tmp21
tmp35 = tmp33 * tmp34
tmp36 = tl.load(in_ptr8 + (tmp8 + 64 * tmp4 + 4096 * x5), None,
eviction_policy='evict_last')
tmp38 = tmp36 + tmp37
tmp39 = tl.load(in_ptr8 + (tmp15 + 64 * tmp4 + 4096 * x5), None,
eviction_policy='evict_last')
tmp40 = tmp39 + tmp37
tmp41 = tmp40 - tmp38
tmp42 = tmp41 * tmp19
tmp43 = tmp38 + tmp42
tmp44 = tl.load(in_ptr8 + (tmp8 + 64 * tmp25 + 4096 * x5), None,
eviction_policy='evict_last')
tmp45 = tmp44 + tmp37
tmp46 = tl.load(in_ptr8 + (tmp15 + 64 * tmp25 + 4096 * x5), None,
eviction_policy='evict_last')
tmp47 = tmp46 + tmp37
tmp48 = tmp47 - tmp45
tmp49 = tmp48 * tmp19
tmp50 = tmp45 + tmp49
tmp51 = tmp50 - tmp43
tmp52 = tmp51 * tmp34
tmp53 = tl.load(in_ptr10 + (tmp8 + 64 * tmp4 + 4096 * x5), None,
eviction_policy='evict_last')
tmp55 = tmp53 + tmp54
tmp56 = tl.load(in_ptr10 + (tmp15 + 64 * tmp4 + 4096 * x5), None,
eviction_policy='evict_last')
tmp57 = tmp56 + tmp54
tmp58 = tmp57 - tmp55
tmp59 = tmp58 * tmp19
tmp60 = tmp55 + tmp59
tmp61 = tl.load(in_ptr10 + (tmp8 + 64 * tmp25 + 4096 * x5), None,
eviction_policy='evict_last')
tmp62 = tmp61 + tmp54
tmp63 = tl.load(in_ptr10 + (tmp15 + 64 * tmp25 + 4096 * x5), None,
eviction_policy='evict_last')
tmp64 = tmp63 + tmp54
tmp65 = tmp64 - tmp62
tmp66 = tmp65 * tmp19
tmp67 = tmp62 + tmp66
tmp68 = tmp67 - tmp60
tmp69 = tmp68 * tmp34
tl.store(out_ptr0 + x6, tmp21, None)
tl.store(out_ptr1 + x6, tmp35, None)
tl.store(out_ptr2 + x6, tmp43, None)
tl.store(out_ptr3 + x6, tmp52, None)
tl.store(out_ptr4 + x6, tmp60, None)
tl.store(out_ptr5 + x6, tmp69, None)
@triton.jit
def triton_poi_fused_cat_6(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
in_ptr5, in_ptr6, in_ptr7, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x1 = xindex // 4096 % 512
x0 = xindex % 4096
x2 = xindex // 2097152
x3 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 128, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 4096 * x1 + 524288 * x2), tmp4, other=0.0)
tmp6 = tl.load(in_ptr1 + (x0 + 4096 * x1 + 524288 * x2), tmp4, other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full(tmp7.shape, 0.0, tmp7.dtype)
tmp9 = tl.where(tmp4, tmp7, tmp8)
tmp10 = tmp0 >= tmp3
tmp11 = tl.full([1], 256, tl.int64)
tmp12 = tmp0 < tmp11
tmp13 = tmp10 & tmp12
tmp14 = tl.load(in_ptr2 + (x0 + 4096 * (-128 + x1) + 524288 * x2),
tmp13, other=0.0)
tmp15 = tl.load(in_ptr3 + (x0 + 4096 * (-128 + x1) + 524288 * x2),
tmp13, other=0.0)
tmp16 = tmp14 + tmp15
tmp17 = tl.full(tmp16.shape, 0.0, tmp16.dtype)
tmp18 = tl.where(tmp13, tmp16, tmp17)
tmp19 = tmp0 >= tmp11
tmp20 = tl.full([1], 384, tl.int64)
tmp21 = tmp0 < tmp20
tmp22 = tmp19 & tmp21
tmp23 = tl.load(in_ptr4 + (x0 + 4096 * (-256 + x1) + 524288 * x2),
tmp22, other=0.0)
tmp24 = tl.load(in_ptr5 + (x0 + 4096 * (-256 + x1) + 524288 * x2),
tmp22, other=0.0)
tmp25 = tmp23 + tmp24
tmp26 = tl.full(tmp25.shape, 0.0, tmp25.dtype)
tmp27 = tl.where(tmp22, tmp25, tmp26)
tmp28 = tmp0 >= tmp20
tl.full([1], 512, tl.int64)
tmp31 = tl.load(in_ptr6 + (x0 + 4096 * (-384 + x1) + 524288 * x2),
tmp28, other=0.0)
tmp32 = tl.load(in_ptr7 + (-384 + x1), tmp28, eviction_policy=
'evict_last', other=0.0)
tmp33 = tmp31 + tmp32
tmp34 = tl.full(tmp33.shape, 0.0, tmp33.dtype)
tmp35 = tl.where(tmp28, tmp33, tmp34)
tmp36 = tl.where(tmp22, tmp27, tmp35)
tmp37 = tl.where(tmp13, tmp18, tmp36)
tmp38 = tl.where(tmp4, tmp9, tmp37)
tl.store(out_ptr0 + x3, tmp38, None)
@triton.jit
def triton_poi_fused_convolution_7(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = xindex // 4096 % 4
tmp0 = tl.load(in_out_ptr0 + x3, None)
tmp1 = tl.load(in_ptr0 + x1, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, None)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17,
primals_18, primals_19, primals_20, primals_21, primals_22,
primals_23, primals_24, primals_25, primals_26, primals_27,
primals_28, primals_29, primals_30) = args
args.clear()
assert_size_stride(primals_1, (256, 2048, 1, 1), (2048, 1, 1, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 2048, 64, 64), (8388608, 4096, 64, 1))
assert_size_stride(primals_4, (256, 1024, 1, 1), (1024, 1, 1, 1))
assert_size_stride(primals_5, (256,), (1,))
assert_size_stride(primals_6, (4, 1024, 64, 64), (4194304, 4096, 64, 1))
assert_size_stride(primals_7, (256, 512, 1, 1), (512, 1, 1, 1))
assert_size_stride(primals_8, (256,), (1,))
assert_size_stride(primals_9, (4, 512, 64, 64), (2097152, 4096, 64, 1))
assert_size_stride(primals_10, (256, 256, 1, 1), (256, 1, 1, 1))
assert_size_stride(primals_11, (256,), (1,))
assert_size_stride(primals_12, (4, 256, 64, 64), (1048576, 4096, 64, 1))
assert_size_stride(primals_13, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_14, (256,), (1,))
assert_size_stride(primals_15, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_16, (128,), (1,))
assert_size_stride(primals_17, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_18, (256,), (1,))
assert_size_stride(primals_19, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_20, (128,), (1,))
assert_size_stride(primals_21, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_22, (256,), (1,))
assert_size_stride(primals_23, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_24, (128,), (1,))
assert_size_stride(primals_25, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_26, (256,), (1,))
assert_size_stride(primals_27, (128, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_28, (128,), (1,))
assert_size_stride(primals_29, (4, 512, 3, 3), (4608, 9, 3, 1))
assert_size_stride(primals_30, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(4194304)](buf1, primals_2,
4194304, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_2
buf2 = extern_kernels.convolution(primals_6, primals_4, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf2, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf3 = empty_strided_cuda((64, 1), (1, 1), torch.int64)
triton_poi_fused__to_copy_1[grid(64)](buf3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf4 = empty_strided_cuda((64, 1), (1, 1), torch.int64)
triton_poi_fused_add_clamp_2[grid(64)](buf4, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((64,), (1,), torch.int64)
triton_poi_fused__to_copy_1[grid(64)](buf5, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf6 = empty_strided_cuda((64,), (1,), torch.int64)
triton_poi_fused_add_clamp_2[grid(64)](buf6, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf7 = empty_strided_cuda((64,), (1,), torch.float32)
triton_poi_fused__to_copy_arange_clamp_mul_sub_3[grid(64)](buf7, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf9 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
triton_poi_fused__to_copy_arange_clamp_mul_sub_3[grid(64)](buf9, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf10 = buf2
del buf2
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4[grid(4194304)
](buf10, buf3, buf5, buf1, buf6, buf7, buf4, buf9, primals_5,
4194304, XBLOCK=512, num_warps=8, num_stages=1)
del primals_5
buf11 = extern_kernels.convolution(primals_9, primals_7, stride=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf11, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf13 = buf11
del buf11
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4[grid(4194304)
](buf13, buf3, buf5, buf10, buf6, buf7, buf4, buf9, primals_8,
4194304, XBLOCK=512, num_warps=8, num_stages=1)
del primals_8
buf14 = extern_kernels.convolution(primals_12, primals_10, stride=(
1, 1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf14, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf16 = buf14
del buf14
triton_poi_fused__unsafe_index_add_convolution_mul_sub_4[grid(4194304)
](buf16, buf3, buf5, buf13, buf6, buf7, buf4, buf9, primals_11,
4194304, XBLOCK=512, num_warps=8, num_stages=1)
del primals_11
buf17 = extern_kernels.convolution(buf1, primals_13, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf17, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf18 = buf17
del buf17
triton_poi_fused_convolution_0[grid(4194304)](buf18, primals_14,
4194304, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_14
buf19 = extern_kernels.convolution(buf18, primals_15, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf19, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf20 = extern_kernels.convolution(buf10, primals_17, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf20, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf21 = buf20
del buf20
triton_poi_fused_convolution_0[grid(4194304)](buf21, primals_18,
4194304, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_18
buf22 = extern_kernels.convolution(buf21, primals_19, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf22, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf23 = extern_kernels.convolution(buf13, primals_21, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf23, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf24 = buf23
del buf23
triton_poi_fused_convolution_0[grid(4194304)](buf24, primals_22,
4194304, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_22
buf25 = extern_kernels.convolution(buf24, primals_23, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf25, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf26 = extern_kernels.convolution(buf16, primals_25, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf26, (4, 256, 64, 64), (1048576, 4096, 64, 1))
buf27 = buf26
del buf26
triton_poi_fused_convolution_0[grid(4194304)](buf27, primals_26,
4194304, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_26
buf28 = extern_kernels.convolution(buf27, primals_27, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf28, (4, 128, 64, 64), (524288, 4096, 64, 1))
buf29 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1),
torch.float32)
buf30 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1),
torch.float32)
buf31 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1),
torch.float32)
buf32 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1),
torch.float32)
buf33 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1),
torch.float32)
buf34 = empty_strided_cuda((4, 128, 64, 64), (524288, 4096, 64, 1),
torch.float32)
triton_poi_fused__unsafe_index_add_convolution_mul_sub_5[grid(2097152)
](buf3, buf5, buf19, primals_16, buf6, buf7, buf4, buf9, buf22,
primals_20, buf25, primals_24, buf29, buf30, buf31, buf32,
buf33, buf34, 2097152, XBLOCK=512, num_warps=8, num_stages=1)
del buf19
del buf22
del buf25
del primals_16
del primals_20
del primals_24
buf35 = empty_strided_cuda((4, 512, 64, 64), (2097152, 4096, 64, 1),
torch.float32)
triton_poi_fused_cat_6[grid(8388608)](buf29, buf30, buf31, buf32,
buf33, buf34, buf28, primals_28, buf35, 8388608, XBLOCK=1024,
num_warps=4, num_stages=1)
del buf28
del buf29
del buf30
del buf31
del buf32
del buf33
del buf34
del primals_28
buf36 = extern_kernels.convolution(buf35, primals_29, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf36, (4, 4, 64, 64), (16384, 4096, 64, 1))
buf37 = buf36
del buf36
triton_poi_fused_convolution_7[grid(65536)](buf37, primals_30,
65536, XBLOCK=256, num_warps=4, num_stages=1)
del primals_30
return (buf37, primals_1, primals_3, primals_4, primals_6, primals_7,
primals_9, primals_10, primals_12, primals_13, primals_15,
primals_17, primals_19, primals_21, primals_23, primals_25,
primals_27, primals_29, buf1, buf3, buf4, buf5, buf6, buf7, buf9,
buf10, buf13, buf16, buf18, buf21, buf24, buf27, buf35)
class fpn_moduleNew(nn.Module):
def __init__(self, numClass):
super(fpn_moduleNew, self).__init__()
self.toplayer = nn.Conv2d(2048, 256, kernel_size=1, stride=1, padding=0
)
self.smooth1_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth2_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth3_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth4_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth1_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth2_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth3_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth4_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.latlayer1 = nn.Conv2d(1024, 256, kernel_size=1, stride=1,
padding=0)
self.latlayer2 = nn.Conv2d(512, 256, kernel_size=1, stride=1, padding=0
)
self.latlayer3 = nn.Conv2d(256, 256, kernel_size=1, stride=1, padding=0
)
self.classify = nn.Conv2d(128 * 4, numClass, kernel_size=3, stride=
1, padding=1)
def _concatenate(self, p5, p4, p3, p2):
_, _, H, W = p2.size()
p5 = F.interpolate(p5, size=(H, W), mode='bilinear', align_corners=True
)
p4 = F.interpolate(p4, size=(H, W), mode='bilinear', align_corners=True
)
p3 = F.interpolate(p3, size=(H, W), mode='bilinear', align_corners=True
)
return torch.cat([p5, p4, p3, p2], dim=1)
def _upsample_add(self, x, y):
_, _, H, W = y.size()
return F.interpolate(x, size=(H, W), mode='bilinear', align_corners
=True) + y
def forward(self, input_0, input_1, input_2, input_3):
primals_1 = self.toplayer.weight
primals_2 = self.toplayer.bias
primals_13 = self.smooth1_1.weight
primals_5 = self.smooth1_1.bias
primals_17 = self.smooth2_1.weight
primals_8 = self.smooth2_1.bias
primals_21 = self.smooth3_1.weight
primals_11 = self.smooth3_1.bias
primals_25 = self.smooth4_1.weight
primals_14 = self.smooth4_1.bias
primals_15 = self.smooth1_2.weight
primals_16 = self.smooth1_2.bias
primals_19 = self.smooth2_2.weight
primals_20 = self.smooth2_2.bias
primals_23 = self.smooth3_2.weight
primals_24 = self.smooth3_2.bias
primals_27 = self.smooth4_2.weight
primals_28 = self.smooth4_2.bias
primals_4 = self.latlayer1.weight
primals_18 = self.latlayer1.bias
primals_7 = self.latlayer2.weight
primals_22 = self.latlayer2.bias
primals_10 = self.latlayer3.weight
primals_26 = self.latlayer3.bias
primals_29 = self.classify.weight
primals_30 = self.classify.bias
primals_12 = input_0
primals_9 = input_1
primals_6 = input_2
primals_3 = input_3
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18, primals_19,
primals_20, primals_21, primals_22, primals_23, primals_24,
primals_25, primals_26, primals_27, primals_28, primals_29,
primals_30])
return output[0]
| LOUEY233/CPS3320_python | fpn_module | false | 1,113 | [
"MIT"
] | 0 | 3cc1733d91c3a8f680eeb984348e2a52ae3285ec | https://github.com/LOUEY233/CPS3320_python/tree/3cc1733d91c3a8f680eeb984348e2a52ae3285ec | import torch
import torch.nn.functional as F
import torch.nn as nn
class Model(nn.Module):
def __init__(self, numClass):
super().__init__()
self.toplayer = nn.Conv2d(2048, 256, kernel_size=1, stride=1, padding=0
)
self.smooth1_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth2_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth3_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth4_1 = nn.Conv2d(256, 256, kernel_size=3, stride=1, padding=1
)
self.smooth1_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth2_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth3_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.smooth4_2 = nn.Conv2d(256, 128, kernel_size=3, stride=1, padding=1
)
self.latlayer1 = nn.Conv2d(1024, 256, kernel_size=1, stride=1,
padding=0)
self.latlayer2 = nn.Conv2d(512, 256, kernel_size=1, stride=1, padding=0
)
self.latlayer3 = nn.Conv2d(256, 256, kernel_size=1, stride=1, padding=0
)
self.classify = nn.Conv2d(128 * 4, numClass, kernel_size=3, stride=
1, padding=1)
def _concatenate(self, p5, p4, p3, p2):
_, _, H, W = p2.size()
p5 = F.interpolate(p5, size=(H, W), mode='bilinear', align_corners=True
)
p4 = F.interpolate(p4, size=(H, W), mode='bilinear', align_corners=True
)
p3 = F.interpolate(p3, size=(H, W), mode='bilinear', align_corners=True
)
return torch.cat([p5, p4, p3, p2], dim=1)
def _upsample_add(self, x, y):
_, _, H, W = y.size()
return F.interpolate(x, size=(H, W), mode='bilinear', align_corners
=True) + y
def forward(self, c2, c3, c4, c5):
p5 = self.toplayer(c5)
p4 = self._upsample_add(p5, self.latlayer1(c4))
p3 = self._upsample_add(p4, self.latlayer2(c3))
p2 = self._upsample_add(p3, self.latlayer3(c2))
p5 = self.smooth1_2(self.smooth1_1(p5))
p4 = self.smooth2_2(self.smooth2_1(p4))
p3 = self.smooth3_2(self.smooth3_1(p3))
p2 = self.smooth4_2(self.smooth4_1(p2))
output = self.classify(self._concatenate(p5, p4, p3, p2))
return output
def get_inputs():
return [torch.rand([4, 256, 64, 64]), torch.rand([4, 512, 64, 64]),
torch.rand([4, 1024, 64, 64]), torch.rand([4, 2048, 64, 64])]
def get_init_inputs():
return [4]
|
Bilinear | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/55/c556jlky4dn3zysohrmkzzgx6ib2mru2umit7utahhvakbnltdhb.py
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# matmul_1 => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 64
x2 = (xindex // 256)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/c5/cc5niyk56fa3m2oc45eqgnf3eeposiu56eefgo6q2ozqlq3nwxnc.py
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# matmul_1 => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_1,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = (xindex // 64)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(buf0, buf1, 1024, grid=grid(1024), stream=stream0)
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(primals_3, buf2, 1024, grid=grid(1024), stream=stream0)
del primals_3
buf3 = empty_strided_cuda((64, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf1, (64, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf2, (64, 4, 4), (16, 4, 1), 0), out=buf3)
del buf1
return (reinterpret_tensor(buf3, (1024, 1), (1, 1), 0), reinterpret_tensor(buf2, (64, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_2, (4, 64), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class Bilinear(nn.Module):
def __init__(self, size):
super(Bilinear, self).__init__()
self.size = size
self.mat = nn.Parameter(torch.FloatTensor(self.size, self.size))
self.reset_parameters()
def reset_parameters(self):
params = [p for p in self.parameters() if p.requires_grad]
for i, param in enumerate(params):
param.data.normal_()
def forward(self, vector1, vector2):
bma = torch.matmul(vector1, self.mat).unsqueeze(1)
ba = torch.matmul(bma, vector2.unsqueeze(2)).view(-1, 1)
return ba
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 64
x2 = xindex // 256
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x2 = xindex // 64
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tl.store(out_ptr0 + x3, tmp0, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(1024)](buf0, buf1, 1024, XBLOCK=256,
num_warps=4, num_stages=1)
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_poi_fused_clone_1[grid(1024)](primals_3, buf2, 1024, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_3
buf3 = empty_strided_cuda((64, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf1, (64, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf2, (64, 4, 4), (16, 4, 1), 0), out=buf3)
del buf1
return reinterpret_tensor(buf3, (1024, 1), (1, 1), 0), reinterpret_tensor(
buf2, (64, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_2, (4,
64), (1, 4), 0)
class BilinearNew(nn.Module):
def __init__(self, size):
super(BilinearNew, self).__init__()
self.size = size
self.mat = nn.Parameter(torch.FloatTensor(self.size, self.size))
self.reset_parameters()
def reset_parameters(self):
params = [p for p in self.parameters() if p.requires_grad]
for i, param in enumerate(params):
param.data.normal_()
def forward(self, input_0, input_1):
primals_1 = self.mat
primals_2 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3])
return output[0]
| TRUMANCFY/VL-DIORA | Bilinear | false | 1,114 | [
"Apache-2.0"
] | 0 | cef398e05842d4a30345260d8e27d1c362671834 | https://github.com/TRUMANCFY/VL-DIORA/tree/cef398e05842d4a30345260d8e27d1c362671834 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, size):
super().__init__()
self.size = size
self.mat = nn.Parameter(torch.FloatTensor(self.size, self.size))
self.reset_parameters()
def reset_parameters(self):
params = [p for p in self.parameters() if p.requires_grad]
for i, param in enumerate(params):
param.data.normal_()
def forward(self, vector1, vector2):
bma = torch.matmul(vector1, self.mat).unsqueeze(1)
ba = torch.matmul(bma, vector2.unsqueeze(2)).view(-1, 1)
return ba
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
N_TransE | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/w2/cw2psuktwvgqiciur5rr6hsrv47ce76ixqoodmx4iymb6o43ia3n.py
# Topologically Sorted Source Nodes: [add, sub, norm, pred], Original ATen: [aten.add, aten.sub, aten.linalg_vector_norm, aten.neg]
# Source node to ATen node mapping:
# add => add
# norm => pow_1, pow_2, sum_1
# pred => neg
# sub => sub
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %arg2_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 4), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1]), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.25), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%pow_2,), kwargs = {})
triton_poi_fused_add_linalg_vector_norm_neg_sub_0 = async_compile.triton('triton_poi_fused_add_linalg_vector_norm_neg_sub_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_linalg_vector_norm_neg_sub_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 12, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_linalg_vector_norm_neg_sub_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr2 + (x0 + (64*x1)), xmask)
tmp7 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp8 = tl.load(in_ptr1 + (16 + x0 + (64*x1)), xmask)
tmp10 = tl.load(in_ptr2 + (16 + x0 + (64*x1)), xmask)
tmp15 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp16 = tl.load(in_ptr1 + (32 + x0 + (64*x1)), xmask)
tmp18 = tl.load(in_ptr2 + (32 + x0 + (64*x1)), xmask)
tmp23 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp24 = tl.load(in_ptr1 + (48 + x0 + (64*x1)), xmask)
tmp26 = tl.load(in_ptr2 + (48 + x0 + (64*x1)), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp5 = tmp4 * tmp4
tmp6 = tmp5 * tmp5
tmp9 = tmp7 + tmp8
tmp11 = tmp9 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tmp12 * tmp12
tmp14 = tmp6 + tmp13
tmp17 = tmp15 + tmp16
tmp19 = tmp17 - tmp18
tmp20 = tmp19 * tmp19
tmp21 = tmp20 * tmp20
tmp22 = tmp14 + tmp21
tmp25 = tmp23 + tmp24
tmp27 = tmp25 - tmp26
tmp28 = tmp27 * tmp27
tmp29 = tmp28 * tmp28
tmp30 = tmp22 + tmp29
tmp31 = 0.25
tmp32 = libdevice.pow(tmp30, tmp31)
tmp33 = -tmp32
tl.store(in_out_ptr0 + (x2), tmp33, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [add, sub, norm, pred], Original ATen: [aten.add, aten.sub, aten.linalg_vector_norm, aten.neg]
stream0 = get_raw_stream(0)
triton_poi_fused_add_linalg_vector_norm_neg_sub_0.run(buf1, arg0_1, arg1_1, arg2_1, 64, grid=grid(64), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
class N_TransE(torch.nn.Module):
def __init__(self, p, params):
super(N_TransE, self).__init__()
self.p = p
self.params = params
def forward(self, e1, r, e2):
pred = -torch.norm(e1 + r - e2, p=self.p, dim=1)
return pred
def loss(self, pos_score, neg_score, target):
return F.relu(pos_score + self.params[0] - neg_score).sum(
) + self.params[1] * F.relu(pos_score - self.params[2]).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [[], {'p': 4, 'params': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_linalg_vector_norm_neg_sub_0(in_out_ptr0, in_ptr0,
in_ptr1, in_ptr2, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr1 + (x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr2 + (x0 + 64 * x1), xmask)
tmp7 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp8 = tl.load(in_ptr1 + (16 + x0 + 64 * x1), xmask)
tmp10 = tl.load(in_ptr2 + (16 + x0 + 64 * x1), xmask)
tmp15 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp16 = tl.load(in_ptr1 + (32 + x0 + 64 * x1), xmask)
tmp18 = tl.load(in_ptr2 + (32 + x0 + 64 * x1), xmask)
tmp23 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp24 = tl.load(in_ptr1 + (48 + x0 + 64 * x1), xmask)
tmp26 = tl.load(in_ptr2 + (48 + x0 + 64 * x1), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 - tmp3
tmp5 = tmp4 * tmp4
tmp6 = tmp5 * tmp5
tmp9 = tmp7 + tmp8
tmp11 = tmp9 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tmp12 * tmp12
tmp14 = tmp6 + tmp13
tmp17 = tmp15 + tmp16
tmp19 = tmp17 - tmp18
tmp20 = tmp19 * tmp19
tmp21 = tmp20 * tmp20
tmp22 = tmp14 + tmp21
tmp25 = tmp23 + tmp24
tmp27 = tmp25 - tmp26
tmp28 = tmp27 * tmp27
tmp29 = tmp28 * tmp28
tmp30 = tmp22 + tmp29
tmp31 = 0.25
tmp32 = libdevice.pow(tmp30, tmp31)
tmp33 = -tmp32
tl.store(in_out_ptr0 + x2, tmp33, xmask)
def call(args):
arg0_1, arg1_1, arg2_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_add_linalg_vector_norm_neg_sub_0[grid(64)](buf1,
arg0_1, arg1_1, arg2_1, 64, XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
del arg1_1
del arg2_1
return buf1,
class N_TransENew(torch.nn.Module):
def __init__(self, p, params):
super(N_TransENew, self).__init__()
self.p = p
self.params = params
def loss(self, pos_score, neg_score, target):
return F.relu(pos_score + self.params[0] - neg_score).sum(
) + self.params[1] * F.relu(pos_score - self.params[2]).sum()
def forward(self, input_0, input_1, input_2):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
output = call([arg0_1, arg1_1, arg2_1])
return output[0]
| TMUITLab/EAFR | N_TransE | false | 1,115 | [
"MIT"
] | 0 | dadb6485d48711ccb8aa2f03760aeb437645f1ff | https://github.com/TMUITLab/EAFR/tree/dadb6485d48711ccb8aa2f03760aeb437645f1ff | import torch
import torch.nn.functional as F
class Model(torch.nn.Module):
def __init__(self, p, params):
super().__init__()
self.p = p
self.params = params
def forward(self, e1, r, e2):
pred = -torch.norm(e1 + r - e2, p=self.p, dim=1)
return pred
def loss(self, pos_score, neg_score, target):
return F.relu(pos_score + self.params[0] - neg_score).sum(
) + self.params[1] * F.relu(pos_score - self.params[2]).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
FM | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/5v/c5vqsotjiyyydjenxs3ttprtusysjuigcmpcuavw4a5cbsh4movc.py
# Topologically Sorted Source Nodes: [sum_1, square_of_sum, mul, sum_of_square, cross_term, sum_3, cross_term_1], Original ATen: [aten.sum, aten.pow, aten.mul, aten.sub]
# Source node to ATen node mapping:
# cross_term => sub
# cross_term_1 => mul_1
# mul => mul
# square_of_sum => pow_1
# sum_1 => sum_1
# sum_3 => sum_3
# sum_of_square => sum_2
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%arg0_1, [1], True), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 2), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %arg0_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%pow_1, %sum_2), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%sub, [2]), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_3, 0.5), kwargs = {})
triton_poi_fused_mul_pow_sub_sum_0 = async_compile.triton('triton_poi_fused_mul_pow_sub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_pow_sub_sum_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 16, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_pow_sub_sum_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp16 = tl.load(in_ptr0 + (4 + x0 + (64*x1)), xmask)
tmp17 = tl.load(in_ptr0 + (20 + x0 + (64*x1)), xmask)
tmp19 = tl.load(in_ptr0 + (36 + x0 + (64*x1)), xmask)
tmp21 = tl.load(in_ptr0 + (52 + x0 + (64*x1)), xmask)
tmp33 = tl.load(in_ptr0 + (8 + x0 + (64*x1)), xmask)
tmp34 = tl.load(in_ptr0 + (24 + x0 + (64*x1)), xmask)
tmp36 = tl.load(in_ptr0 + (40 + x0 + (64*x1)), xmask)
tmp38 = tl.load(in_ptr0 + (56 + x0 + (64*x1)), xmask)
tmp50 = tl.load(in_ptr0 + (12 + x0 + (64*x1)), xmask)
tmp51 = tl.load(in_ptr0 + (28 + x0 + (64*x1)), xmask)
tmp53 = tl.load(in_ptr0 + (44 + x0 + (64*x1)), xmask)
tmp55 = tl.load(in_ptr0 + (60 + x0 + (64*x1)), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = tmp6 * tmp6
tmp8 = tmp0 * tmp0
tmp9 = tmp1 * tmp1
tmp10 = tmp8 + tmp9
tmp11 = tmp3 * tmp3
tmp12 = tmp10 + tmp11
tmp13 = tmp5 * tmp5
tmp14 = tmp12 + tmp13
tmp15 = tmp7 - tmp14
tmp18 = tmp16 + tmp17
tmp20 = tmp18 + tmp19
tmp22 = tmp20 + tmp21
tmp23 = tmp22 * tmp22
tmp24 = tmp16 * tmp16
tmp25 = tmp17 * tmp17
tmp26 = tmp24 + tmp25
tmp27 = tmp19 * tmp19
tmp28 = tmp26 + tmp27
tmp29 = tmp21 * tmp21
tmp30 = tmp28 + tmp29
tmp31 = tmp23 - tmp30
tmp32 = tmp15 + tmp31
tmp35 = tmp33 + tmp34
tmp37 = tmp35 + tmp36
tmp39 = tmp37 + tmp38
tmp40 = tmp39 * tmp39
tmp41 = tmp33 * tmp33
tmp42 = tmp34 * tmp34
tmp43 = tmp41 + tmp42
tmp44 = tmp36 * tmp36
tmp45 = tmp43 + tmp44
tmp46 = tmp38 * tmp38
tmp47 = tmp45 + tmp46
tmp48 = tmp40 - tmp47
tmp49 = tmp32 + tmp48
tmp52 = tmp50 + tmp51
tmp54 = tmp52 + tmp53
tmp56 = tmp54 + tmp55
tmp57 = tmp56 * tmp56
tmp58 = tmp50 * tmp50
tmp59 = tmp51 * tmp51
tmp60 = tmp58 + tmp59
tmp61 = tmp53 * tmp53
tmp62 = tmp60 + tmp61
tmp63 = tmp55 * tmp55
tmp64 = tmp62 + tmp63
tmp65 = tmp57 - tmp64
tmp66 = tmp49 + tmp65
tmp67 = 0.5
tmp68 = tmp66 * tmp67
tl.store(in_out_ptr0 + (x2), tmp68, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 4), (4, 16, 1), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 1, 4), (4, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [sum_1, square_of_sum, mul, sum_of_square, cross_term, sum_3, cross_term_1], Original ATen: [aten.sum, aten.pow, aten.mul, aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_pow_sub_sum_0.run(buf1, arg0_1, 16, grid=grid(16), stream=stream0)
del arg0_1
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
from sklearn.metrics import *
class FM(nn.Module):
"""Factorization Machine models pairwise (order-2) feature interactions
without linear term and bias.
Input shape
- 3D tensor with shape: ``(batch_size,field_size,embedding_size)``.
Output shape
- 2D tensor with shape: ``(batch_size, 1)``.
References
- [Factorization Machines](https://www.csie.ntu.edu.tw/~b97053/paper/Rendle2010FM.pdf)
"""
def __init__(self):
super(FM, self).__init__()
def forward(self, inputs):
fm_input = inputs
square_of_sum = torch.pow(torch.sum(fm_input, dim=1, keepdim=True), 2)
sum_of_square = torch.sum(fm_input * fm_input, dim=1, keepdim=True)
cross_term = square_of_sum - sum_of_square
cross_term = 0.5 * torch.sum(cross_term, dim=2, keepdim=False)
return cross_term
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
from sklearn.metrics import *
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_mul_pow_sub_sum_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp16 = tl.load(in_ptr0 + (4 + x0 + 64 * x1), xmask)
tmp17 = tl.load(in_ptr0 + (20 + x0 + 64 * x1), xmask)
tmp19 = tl.load(in_ptr0 + (36 + x0 + 64 * x1), xmask)
tmp21 = tl.load(in_ptr0 + (52 + x0 + 64 * x1), xmask)
tmp33 = tl.load(in_ptr0 + (8 + x0 + 64 * x1), xmask)
tmp34 = tl.load(in_ptr0 + (24 + x0 + 64 * x1), xmask)
tmp36 = tl.load(in_ptr0 + (40 + x0 + 64 * x1), xmask)
tmp38 = tl.load(in_ptr0 + (56 + x0 + 64 * x1), xmask)
tmp50 = tl.load(in_ptr0 + (12 + x0 + 64 * x1), xmask)
tmp51 = tl.load(in_ptr0 + (28 + x0 + 64 * x1), xmask)
tmp53 = tl.load(in_ptr0 + (44 + x0 + 64 * x1), xmask)
tmp55 = tl.load(in_ptr0 + (60 + x0 + 64 * x1), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = tmp6 * tmp6
tmp8 = tmp0 * tmp0
tmp9 = tmp1 * tmp1
tmp10 = tmp8 + tmp9
tmp11 = tmp3 * tmp3
tmp12 = tmp10 + tmp11
tmp13 = tmp5 * tmp5
tmp14 = tmp12 + tmp13
tmp15 = tmp7 - tmp14
tmp18 = tmp16 + tmp17
tmp20 = tmp18 + tmp19
tmp22 = tmp20 + tmp21
tmp23 = tmp22 * tmp22
tmp24 = tmp16 * tmp16
tmp25 = tmp17 * tmp17
tmp26 = tmp24 + tmp25
tmp27 = tmp19 * tmp19
tmp28 = tmp26 + tmp27
tmp29 = tmp21 * tmp21
tmp30 = tmp28 + tmp29
tmp31 = tmp23 - tmp30
tmp32 = tmp15 + tmp31
tmp35 = tmp33 + tmp34
tmp37 = tmp35 + tmp36
tmp39 = tmp37 + tmp38
tmp40 = tmp39 * tmp39
tmp41 = tmp33 * tmp33
tmp42 = tmp34 * tmp34
tmp43 = tmp41 + tmp42
tmp44 = tmp36 * tmp36
tmp45 = tmp43 + tmp44
tmp46 = tmp38 * tmp38
tmp47 = tmp45 + tmp46
tmp48 = tmp40 - tmp47
tmp49 = tmp32 + tmp48
tmp52 = tmp50 + tmp51
tmp54 = tmp52 + tmp53
tmp56 = tmp54 + tmp55
tmp57 = tmp56 * tmp56
tmp58 = tmp50 * tmp50
tmp59 = tmp51 * tmp51
tmp60 = tmp58 + tmp59
tmp61 = tmp53 * tmp53
tmp62 = tmp60 + tmp61
tmp63 = tmp55 * tmp55
tmp64 = tmp62 + tmp63
tmp65 = tmp57 - tmp64
tmp66 = tmp49 + tmp65
tmp67 = 0.5
tmp68 = tmp66 * tmp67
tl.store(in_out_ptr0 + x2, tmp68, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 4), (4, 16, 1), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 1, 4), (4, 4, 1), 0)
del buf0
get_raw_stream(0)
triton_poi_fused_mul_pow_sub_sum_0[grid(16)](buf1, arg0_1, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del arg0_1
return buf1,
class FMNew(nn.Module):
"""Factorization Machine models pairwise (order-2) feature interactions
without linear term and bias.
Input shape
- 3D tensor with shape: ``(batch_size,field_size,embedding_size)``.
Output shape
- 2D tensor with shape: ``(batch_size, 1)``.
References
- [Factorization Machines](https://www.csie.ntu.edu.tw/~b97053/paper/Rendle2010FM.pdf)
"""
def __init__(self):
super(FMNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Sunmyunghan/Final_Project | FM | false | 1,117 | [
"MIT"
] | 0 | 28cde293dc6d07521b2e1c5613b20444aea91d21 | https://github.com/Sunmyunghan/Final_Project/tree/28cde293dc6d07521b2e1c5613b20444aea91d21 | import torch
import torch.nn as nn
from sklearn.metrics import *
class Model(nn.Module):
"""Factorization Machine models pairwise (order-2) feature interactions
without linear term and bias.
Input shape
- 3D tensor with shape: ``(batch_size,field_size,embedding_size)``.
Output shape
- 2D tensor with shape: ``(batch_size, 1)``.
References
- [Factorization Machines](https://www.csie.ntu.edu.tw/~b97053/paper/Rendle2010FM.pdf)
"""
def __init__(self):
super().__init__()
def forward(self, inputs):
fm_input = inputs
square_of_sum = torch.pow(torch.sum(fm_input, dim=1, keepdim=True), 2)
sum_of_square = torch.sum(fm_input * fm_input, dim=1, keepdim=True)
cross_term = square_of_sum - sum_of_square
cross_term = 0.5 * torch.sum(cross_term, dim=2, keepdim=False)
return cross_term
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
VertexDirectEmbedder | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/xq/cxqinuparlha25j4geyv6tolvpah7qdqdkpecjesyn3kblysszql.py
# Topologically Sorted Source Nodes: [norm, clamp, truediv], Original ATen: [aten.linalg_vector_norm, aten.clamp, aten.div]
# Source node to ATen node mapping:
# clamp => clamp_min
# norm => pow_1, pow_2, sum_1
# truediv => div
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_1, 2.0), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1], True), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%pow_2, 1e-06), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_1, %clamp_min), kwargs = {})
triton_poi_fused_clamp_div_linalg_vector_norm_0 = async_compile.triton('triton_poi_fused_clamp_div_linalg_vector_norm_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clamp_div_linalg_vector_norm_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clamp_div_linalg_vector_norm_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-06
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x2), tmp15, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [norm, clamp, truediv], Original ATen: [aten.linalg_vector_norm, aten.clamp, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_clamp_div_linalg_vector_norm_0.run(primals_1, buf0, 16, grid=grid(16), stream=stream0)
return (buf0, primals_1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.utils.data
from torch import nn
def normalize_embeddings(embeddings: 'torch.Tensor', epsilon: 'float'=1e-06
) ->torch.Tensor:
"""
Normalize N D-dimensional embedding vectors arranged in a tensor [N, D]
Args:
embeddings (tensor [N, D]): N D-dimensional embedding vectors
epsilon (float): minimum value for a vector norm
Return:
Normalized embeddings (tensor [N, D]), such that L2 vector norms are all equal to 1.
"""
return embeddings / torch.clamp(embeddings.norm(p=None, dim=1, keepdim=
True), min=epsilon)
class VertexDirectEmbedder(nn.Module):
"""
Class responsible for embedding vertices. Vertex embeddings take
the form of a tensor of size [N, D], where
N = number of vertices
D = number of dimensions in the embedding space
"""
def __init__(self, num_vertices: 'int', embed_dim: 'int'):
"""
Initialize embedder, set random embeddings
Args:
num_vertices (int): number of vertices to embed
embed_dim (int): number of dimensions in the embedding space
"""
super(VertexDirectEmbedder, self).__init__()
self.embeddings = nn.Parameter(torch.Tensor(num_vertices, embed_dim))
self.reset_parameters()
@torch.no_grad()
def reset_parameters(self):
"""
Reset embeddings to random values
"""
torch.nn.init.uniform_(self.embeddings, a=-0.5, b=0.5)
def forward(self) ->torch.Tensor:
"""
Produce vertex embeddings, a tensor of shape [N, D] where:
N = number of vertices
D = number of dimensions in the embedding space
Return:
Full vertex embeddings, a tensor of shape [N, D]
"""
return normalize_embeddings(self.embeddings)
@torch.no_grad()
def load(self, fpath: 'str'):
"""
Load data from a file
Args:
fpath (str): file path to load data from
"""
with PathManager.open(fpath, 'rb') as hFile:
data = pickle.load(hFile)
for name in ['embeddings']:
if name in data:
getattr(self, name).copy_(torch.tensor(data[name]).float())
def get_inputs():
return []
def get_init_inputs():
return [[], {'num_vertices': 4, 'embed_dim': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.utils.data
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_clamp_div_linalg_vector_norm_0(in_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-06
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + x2, tmp15, xmask)
def call(args):
primals_1, = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clamp_div_linalg_vector_norm_0[grid(16)](primals_1,
buf0, 16, XBLOCK=16, num_warps=1, num_stages=1)
return buf0, primals_1
def normalize_embeddings(embeddings: 'torch.Tensor', epsilon: 'float'=1e-06
) ->torch.Tensor:
"""
Normalize N D-dimensional embedding vectors arranged in a tensor [N, D]
Args:
embeddings (tensor [N, D]): N D-dimensional embedding vectors
epsilon (float): minimum value for a vector norm
Return:
Normalized embeddings (tensor [N, D]), such that L2 vector norms are all equal to 1.
"""
return embeddings / torch.clamp(embeddings.norm(p=None, dim=1, keepdim=
True), min=epsilon)
class VertexDirectEmbedderNew(nn.Module):
"""
Class responsible for embedding vertices. Vertex embeddings take
the form of a tensor of size [N, D], where
N = number of vertices
D = number of dimensions in the embedding space
"""
def __init__(self, num_vertices: 'int', embed_dim: 'int'):
"""
Initialize embedder, set random embeddings
Args:
num_vertices (int): number of vertices to embed
embed_dim (int): number of dimensions in the embedding space
"""
super(VertexDirectEmbedderNew, self).__init__()
self.embeddings = nn.Parameter(torch.Tensor(num_vertices, embed_dim))
self.reset_parameters()
@torch.no_grad()
def reset_parameters(self):
"""
Reset embeddings to random values
"""
torch.nn.init.uniform_(self.embeddings, a=-0.5, b=0.5)
@torch.no_grad()
def load(self, fpath: 'str'):
"""
Load data from a file
Args:
fpath (str): file path to load data from
"""
with PathManager.open(fpath, 'rb') as hFile:
data = pickle.load(hFile)
for name in ['embeddings']:
if name in data:
getattr(self, name).copy_(torch.tensor(data[name]).float())
def forward(self):
primals_1 = self.embeddings
output = call([primals_1])
return output[0]
| TWJianNuo/detectron2 | VertexDirectEmbedder | false | 1,118 | [
"Apache-2.0"
] | 0 | 091bc43e85b8f7cefdccebf8d85afb7cfff2a3f0 | https://github.com/TWJianNuo/detectron2/tree/091bc43e85b8f7cefdccebf8d85afb7cfff2a3f0 | import torch
import torch.utils.data
from torch import nn
def normalize_embeddings(embeddings: 'torch.Tensor', epsilon: 'float'=1e-06
) ->torch.Tensor:
"""
Normalize N D-dimensional embedding vectors arranged in a tensor [N, D]
Args:
embeddings (tensor [N, D]): N D-dimensional embedding vectors
epsilon (float): minimum value for a vector norm
Return:
Normalized embeddings (tensor [N, D]), such that L2 vector norms are all equal to 1.
"""
return embeddings / torch.clamp(embeddings.norm(p=None, dim=1, keepdim=
True), min=epsilon)
class Model(nn.Module):
"""
Class responsible for embedding vertices. Vertex embeddings take
the form of a tensor of size [N, D], where
N = number of vertices
D = number of dimensions in the embedding space
"""
def __init__(self, num_vertices: 'int', embed_dim: 'int'):
"""
Initialize embedder, set random embeddings
Args:
num_vertices (int): number of vertices to embed
embed_dim (int): number of dimensions in the embedding space
"""
super().__init__()
self.embeddings = nn.Parameter(torch.Tensor(num_vertices, embed_dim))
self.reset_parameters()
@torch.no_grad()
def reset_parameters(self):
"""
Reset embeddings to random values
"""
torch.nn.init.uniform_(self.embeddings, a=-0.5, b=0.5)
def forward(self) ->torch.Tensor:
"""
Produce vertex embeddings, a tensor of shape [N, D] where:
N = number of vertices
D = number of dimensions in the embedding space
Return:
Full vertex embeddings, a tensor of shape [N, D]
"""
return normalize_embeddings(self.embeddings)
@torch.no_grad()
def load(self, fpath: 'str'):
"""
Load data from a file
Args:
fpath (str): file path to load data from
"""
with PathManager.open(fpath, 'rb') as hFile:
data = pickle.load(hFile)
for name in ['embeddings']:
if name in data:
getattr(self, name).copy_(torch.tensor(data[name]).float())
def get_inputs():
return []
def get_init_inputs():
return [4, 4]
|
HighWay | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/mq/cmqzn4tuak4dmmt5haqtljb4thsuzy6lq5otm3atysewj5viuiif.py
# Topologically Sorted Source Nodes: [t_1, gate, mul, sub, mul_1, add_1], Original ATen: [aten.add, aten.sigmoid, aten.mul, aten.rsub]
# Source node to ATen node mapping:
# add_1 => add_1
# gate => sigmoid
# mul => mul
# mul_1 => mul_1
# sub => sub
# t_1 => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm, %primals_3), kwargs = {})
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, %primals_4), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %sigmoid), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %primals_2), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
triton_poi_fused_add_mul_rsub_sigmoid_0 = async_compile.triton('triton_poi_fused_add_mul_rsub_sigmoid_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_rsub_sigmoid_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + (x2), xmask)
tmp8 = tl.load(in_ptr3 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tmp5 = tmp3 * tmp4
tmp6 = 1.0
tmp7 = tmp6 - tmp3
tmp9 = tmp7 * tmp8
tmp10 = tmp5 + tmp9
tl.store(out_ptr0 + (x2), tmp10, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [t], Original ATen: [aten.mm]
extern_kernels.mm(primals_2, primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [t_1, gate, mul, sub, mul_1, add_1], Original ATen: [aten.add, aten.sigmoid, aten.mul, aten.rsub]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_rsub_sigmoid_0.run(buf0, primals_3, primals_4, primals_2, buf1, 16, grid=grid(16), stream=stream0)
return (buf1, primals_2, primals_3, primals_4, buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
from torch.nn import Parameter
class HighWay(torch.nn.Module):
def __init__(self, f_in, f_out, bias=True):
super(HighWay, self).__init__()
self.w = Parameter(torch.Tensor(f_in, f_out))
nn.init.xavier_uniform_(self.w)
if bias:
self.bias = Parameter(torch.Tensor(f_out))
nn.init.constant_(self.bias, 0)
else:
self.register_parameter('bias', None)
def forward(self, in_1, in_2):
t = torch.mm(in_1, self.w)
if self.bias is not None:
t = t + self.bias
gate = torch.sigmoid(t)
return gate * in_2 + (1.0 - gate) * in_1
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'f_in': 4, 'f_out': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
from torch.nn import Parameter
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_mul_rsub_sigmoid_0(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + x2, xmask)
tmp8 = tl.load(in_ptr3 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp3 = tl.sigmoid(tmp2)
tmp5 = tmp3 * tmp4
tmp6 = 1.0
tmp7 = tmp6 - tmp3
tmp9 = tmp7 * tmp8
tmp10 = tmp5 + tmp9
tl.store(out_ptr0 + x2, tmp10, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(primals_2, primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_mul_rsub_sigmoid_0[grid(16)](buf0, primals_3,
primals_4, primals_2, buf1, 16, XBLOCK=16, num_warps=1,
num_stages=1)
return buf1, primals_2, primals_3, primals_4, buf0
class HighWayNew(torch.nn.Module):
def __init__(self, f_in, f_out, bias=True):
super(HighWayNew, self).__init__()
self.w = Parameter(torch.Tensor(f_in, f_out))
nn.init.xavier_uniform_(self.w)
if bias:
self.bias = Parameter(torch.Tensor(f_out))
nn.init.constant_(self.bias, 0)
else:
self.register_parameter('bias', None)
def forward(self, input_0, input_1):
primals_1 = self.w
primals_3 = self.bias
primals_2 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| TMUITLab/EAFR | HighWay | false | 1,119 | [
"MIT"
] | 0 | dadb6485d48711ccb8aa2f03760aeb437645f1ff | https://github.com/TMUITLab/EAFR/tree/dadb6485d48711ccb8aa2f03760aeb437645f1ff | import torch
import torch.nn as nn
from torch.nn import Parameter
class Model(torch.nn.Module):
def __init__(self, f_in, f_out, bias=True):
super().__init__()
self.w = Parameter(torch.Tensor(f_in, f_out))
nn.init.xavier_uniform_(self.w)
if bias:
self.bias = Parameter(torch.Tensor(f_out))
nn.init.constant_(self.bias, 0)
else:
self.register_parameter('bias', None)
def forward(self, in_1, in_2):
t = torch.mm(in_1, self.w)
if self.bias is not None:
t = t + self.bias
gate = torch.sigmoid(t)
return gate * in_2 + (1.0 - gate) * in_1
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [4, 4]
|
Network | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/nq/cnqjufcqn3ur3s7xvlb2i747nyf24md4zaiatlwgkasynplfjstu.py
# Topologically Sorted Source Nodes: [t], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# t => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4096
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr0 + (x2), tmp6, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/j4/cj4miacghwuwo6tmp3hylr7yjqyun32g4pisr65oc2dtlcxfwv2f.py
# Topologically Sorted Source Nodes: [t_2], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# t_2 => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_5, [0], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_5, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_1 = async_compile.triton('triton_poi_fused__softmax_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (64 + x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (128 + x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (192 + x0), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/uy/cuylqrd7ye33ogvvpsnxb3skali4boxth4tryw5hn4czjzyh4a34.py
# Topologically Sorted Source Nodes: [t_2], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# t_2 => div, sum_1
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [0], True), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (64 + x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (128 + x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (192 + x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (64, 4), (4, 1))
assert_size_stride(primals_2, (64, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (64, 64), (64, 1))
assert_size_stride(primals_5, (64, ), (1, ))
assert_size_stride(primals_6, (4, 64), (64, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 64), (64, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 64), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 64), (1024, 256, 64, 1), 0); del buf0 # reuse
buf8 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool)
# Topologically Sorted Source Nodes: [t], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf8, 4096, grid=grid(4096), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 64), (64, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 64), (64, 1), 0), reinterpret_tensor(primals_4, (64, 64), (1, 64), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 64), (1024, 256, 64, 1), 0); del buf2 # reuse
buf7 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool)
# Topologically Sorted Source Nodes: [t_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_0.run(buf3, primals_5, buf7, 4096, grid=grid(4096), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 64), (64, 1), 0), reinterpret_tensor(primals_6, (64, 4), (1, 64), 0), alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [t_2], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf4, buf5, 256, grid=grid(256), stream=stream0)
buf6 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [t_2], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
del buf5
return (buf6, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 64), (64, 1), 0), reinterpret_tensor(buf3, (64, 64), (64, 1), 0), buf6, primals_6, buf7, primals_4, buf8, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((64, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((64, 64), (64, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 64), (64, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
from torch.nn.functional import relu
from torch.nn.functional import softmax
class Network(nn.Module):
def __init__(self, input_size, output_size):
super().__init__()
self.input_size = input_size
self.output_size = output_size
self.fc1 = nn.Linear(in_features=input_size, out_features=64)
self.fc2 = nn.Linear(in_features=64, out_features=64)
self.out = nn.Linear(in_features=64, out_features=output_size)
def forward(self, t):
t = relu(self.fc1(t))
t = relu(self.fc2(t))
t = softmax(self.out(t), dim=0)
return t
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'output_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, None)
tl.store(out_ptr0 + x2, tmp6, None)
@triton.jit
def triton_poi_fused__softmax_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (64 + x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (128 + x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (192 + x0), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tmp9 = tl_math.exp(tmp8)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (64 + x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (128 + x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (192 + x0), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (64, 4), (4, 1))
assert_size_stride(primals_2, (64,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (64, 64), (64, 1))
assert_size_stride(primals_5, (64,), (1,))
assert_size_stride(primals_6, (4, 64), (64, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 64), (64, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 64), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 64), (1024, 256, 64, 1), 0)
del buf0
buf8 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool
)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(4096)](buf1,
primals_2, buf8, 4096, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 64), (64, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 64), (64, 1), 0),
reinterpret_tensor(primals_4, (64, 64), (1, 64), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 64), (1024, 256, 64, 1), 0)
del buf2
buf7 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool
)
triton_poi_fused_relu_threshold_backward_0[grid(4096)](buf3,
primals_5, buf7, 4096, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 64),
(64, 1), 0), reinterpret_tensor(primals_6, (64, 4), (1, 64), 0),
alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf4, buf5, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf6 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf4
triton_poi_fused__softmax_2[grid(256)](buf5, buf6, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf5
return buf6, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 64), (64, 1), 0), reinterpret_tensor(
buf3, (64, 64), (64, 1), 0), buf6, primals_6, buf7, primals_4, buf8
class NetworkNew(nn.Module):
def __init__(self, input_size, output_size):
super().__init__()
self.input_size = input_size
self.output_size = output_size
self.fc1 = nn.Linear(in_features=input_size, out_features=64)
self.fc2 = nn.Linear(in_features=64, out_features=64)
self.out = nn.Linear(in_features=64, out_features=output_size)
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_6 = self.out.weight
primals_7 = self.out.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| THE-RAF/Reinforcement-Learning | Network | false | 1,120 | [
"MIT"
] | 0 | 36b4c5330740b533fb8170263f995afb91a1d021 | https://github.com/THE-RAF/Reinforcement-Learning/tree/36b4c5330740b533fb8170263f995afb91a1d021 | import torch
import torch.nn as nn
from torch.nn.functional import relu
from torch.nn.functional import softmax
class Model(nn.Module):
def __init__(self, input_size, output_size):
super().__init__()
self.input_size = input_size
self.output_size = output_size
self.fc1 = nn.Linear(in_features=input_size, out_features=64)
self.fc2 = nn.Linear(in_features=64, out_features=64)
self.out = nn.Linear(in_features=64, out_features=output_size)
def forward(self, t):
t = relu(self.fc1(t))
t = relu(self.fc2(t))
t = softmax(self.out(t), dim=0)
return t
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
SpatialCrossMapLRN | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/t4/ct42qpaygn7av2p6rystjl4hk3ybzwp5jyvmk3jaiukfiri3pq65.py
# Topologically Sorted Source Nodes: [mul, add, div_2, x], Original ATen: [aten.mul, aten.add, aten.pow, aten.div]
# Source node to ATen node mapping:
# add => add
# div_2 => pow_2
# mul => mul
# x => div
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%squeeze, 1.0), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 1), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%add, 0.75), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg0_1, %pow_2), kwargs = {})
triton_poi_fused_add_div_mul_pow_0 = async_compile.triton('triton_poi_fused_add_div_mul_pow_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_mul_pow_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_mul_pow_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tmp0 * tmp0
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = tmp3 * tmp2
tmp5 = tmp4 + tmp2
tmp6 = 0.75
tmp7 = libdevice.pow(tmp5, tmp6)
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, add, div_2, x], Original ATen: [aten.mul, aten.add, aten.pow, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_mul_pow_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
class SpatialCrossMapLRN(nn.Module):
def __init__(self, local_size=1, alpha=1.0, beta=0.75, k=1,
ACROSS_CHANNELS=True):
super(SpatialCrossMapLRN, self).__init__()
self.ACROSS_CHANNELS = ACROSS_CHANNELS
if ACROSS_CHANNELS:
self.average = nn.AvgPool3d(kernel_size=(local_size, 1, 1),
stride=1, padding=(int((local_size - 1.0) / 2), 0, 0))
else:
self.average = nn.AvgPool2d(kernel_size=local_size, stride=1,
padding=int((local_size - 1.0) / 2))
self.alpha = alpha
self.beta = beta
self.k = k
def forward(self, x):
if self.ACROSS_CHANNELS:
div = x.pow(2).unsqueeze(1)
div = self.average(div).squeeze(1)
div = div.mul(self.alpha).add(self.k).pow(self.beta)
else:
div = x.pow(2)
div = self.average(div)
div = div.mul(self.alpha).add(self.k).pow(self.beta)
x = x.div(div)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_div_mul_pow_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tmp0 * tmp0
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = tmp3 * tmp2
tmp5 = tmp4 + tmp2
tmp6 = 0.75
tmp7 = libdevice.pow(tmp5, tmp6)
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x0, tmp8, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_mul_pow_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class SpatialCrossMapLRNNew(nn.Module):
def __init__(self, local_size=1, alpha=1.0, beta=0.75, k=1,
ACROSS_CHANNELS=True):
super(SpatialCrossMapLRNNew, self).__init__()
self.ACROSS_CHANNELS = ACROSS_CHANNELS
if ACROSS_CHANNELS:
self.average = nn.AvgPool3d(kernel_size=(local_size, 1, 1),
stride=1, padding=(int((local_size - 1.0) / 2), 0, 0))
else:
self.average = nn.AvgPool2d(kernel_size=local_size, stride=1,
padding=int((local_size - 1.0) / 2))
self.alpha = alpha
self.beta = beta
self.k = k
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Tagussan/pretrained-models.pytorch | SpatialCrossMapLRN | false | 1,121 | [
"BSD-3-Clause"
] | 0 | 854e6c153c2534dd7cf76a5ec102307ea5171167 | https://github.com/Tagussan/pretrained-models.pytorch/tree/854e6c153c2534dd7cf76a5ec102307ea5171167 | import torch
import torch.nn as nn
import torch.nn.parallel
import torch.optim
import torch.utils.data
class Model(nn.Module):
def __init__(self, local_size=1, alpha=1.0, beta=0.75, k=1,
ACROSS_CHANNELS=True):
super().__init__()
self.ACROSS_CHANNELS = ACROSS_CHANNELS
if ACROSS_CHANNELS:
self.average = nn.AvgPool3d(kernel_size=(local_size, 1, 1),
stride=1, padding=(int((local_size - 1.0) / 2), 0, 0))
else:
self.average = nn.AvgPool2d(kernel_size=local_size, stride=1,
padding=int((local_size - 1.0) / 2))
self.alpha = alpha
self.beta = beta
self.k = k
def forward(self, x):
if self.ACROSS_CHANNELS:
div = x.pow(2).unsqueeze(1)
div = self.average(div).squeeze(1)
div = div.mul(self.alpha).add(self.k).pow(self.beta)
else:
div = x.pow(2)
div = self.average(div)
div = div.mul(self.alpha).add(self.k).pow(self.beta)
x = x.div(div)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
MLPBase | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/ix/cixxyusyg44s2hkoufcgbrv3ix5ookwqjl4ia3xkv7bdqi4yrzus.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 400
x2 = xindex % 1600
x3 = (xindex // 1600)
tmp0 = tl.load(in_out_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x4), tmp4, xmask)
tl.store(out_ptr0 + (x2 + (1664*x3)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/op/coptu6xep3awc4lajb4xivopppqmjtx3zy7ebtazm45rqvyeknds.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x_1 => relu_1
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_3,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 300
x2 = (xindex // 1200)
x3 = xindex % 1200
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x3 + (1216*x2)), tmp4, xmask)
tl.store(out_ptr1 + (x3 + (1280*x2)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/as/casrc7bf7ghsendgi7tkqxk3hj4ic6aqb4rmkxzuk5dhbidznia7.py
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.relu, aten.view]
# Source node to ATen node mapping:
# x_1 => relu_1
# x_2 => view_4
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_3,), kwargs = {})
# %view_4 : [num_users=2] = call_function[target=torch.ops.aten.reshape.default](args = (%relu_1, [64, 300]), kwargs = {})
triton_poi_fused_relu_view_2 = async_compile.triton('triton_poi_fused_relu_view_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_view_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_view_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 300
x1 = (xindex // 300)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (300*(x1 % 4)) + (1216*(x1 // 4))), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (400, 4), (4, 1))
assert_size_stride(primals_2, (400, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (300, 400), (400, 1))
assert_size_stride(primals_5, (300, ), (1, ))
assert_size_stride(primals_6, (4, 300), (300, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 400), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 400), (6400, 1600, 400, 1), 0); del buf0 # reuse
buf7 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf7, 25600, grid=grid(25600), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 400), (400, 1), 0), reinterpret_tensor(primals_4, (400, 300), (1, 400), 0), out=buf2)
buf3 = empty_strided_cuda((4, 4, 4, 300), (4864, 1216, 300, 1), torch.float32)
buf6 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf2, primals_5, buf3, buf6, 19200, grid=grid(19200), stream=stream0)
del primals_5
buf4 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.relu, aten.view]
triton_poi_fused_relu_view_2.run(buf3, buf4, 19200, grid=grid(19200), stream=stream0)
del buf3
buf5 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf4, reinterpret_tensor(primals_6, (300, 4), (1, 300), 0), alpha=1, beta=1, out=buf5)
del primals_7
return (reinterpret_tensor(buf5, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 400), (400, 1), 0), buf4, primals_6, buf6, primals_4, buf7, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((400, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((300, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((300, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 300), (300, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
import torch.nn.functional as F
class MLPBase(nn.Module):
def __init__(self, num_inputs, num_outputs):
super(MLPBase, self).__init__()
self.l1 = nn.Linear(num_inputs, 400)
self.l2 = nn.Linear(400, 300)
self.l3 = nn.Linear(300, num_outputs)
def forward(self, inputs):
x = F.relu(self.l1(inputs))
x = F.relu(self.l2(x))
x = self.l3(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_inputs': 4, 'num_outputs': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 400
x2 = xindex % 1600
x3 = xindex // 1600
tmp0 = tl.load(in_out_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x4, tmp4, xmask)
tl.store(out_ptr0 + (x2 + 1664 * x3), tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 300
x2 = xindex // 1200
x3 = xindex % 1200
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x3 + 1216 * x2), tmp4, xmask)
tl.store(out_ptr1 + (x3 + 1280 * x2), tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_view_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 300
x1 = xindex // 300
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 300 * (x1 % 4) + 1216 * (x1 // 4)), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7) = args
args.clear()
assert_size_stride(primals_1, (400, 4), (4, 1))
assert_size_stride(primals_2, (400,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (300, 400), (400, 1))
assert_size_stride(primals_5, (300,), (1,))
assert_size_stride(primals_6, (4, 300), (300, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 400), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 400), (6400, 1600, 400, 1), 0
)
del buf0
buf7 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(25600)](buf1,
primals_2, buf7, 25600, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 400), (400, 1), 0),
reinterpret_tensor(primals_4, (400, 300), (1, 400), 0), out=buf2)
buf3 = empty_strided_cuda((4, 4, 4, 300), (4864, 1216, 300, 1),
torch.float32)
buf6 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_1[grid(19200)](buf2,
primals_5, buf3, buf6, 19200, XBLOCK=256, num_warps=4, num_stages=1
)
del primals_5
buf4 = buf2
del buf2
triton_poi_fused_relu_view_2[grid(19200)](buf3, buf4, 19200, XBLOCK
=256, num_warps=4, num_stages=1)
del buf3
buf5 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, buf4, reinterpret_tensor(primals_6,
(300, 4), (1, 300), 0), alpha=1, beta=1, out=buf5)
del primals_7
return reinterpret_tensor(buf5, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 400), (400, 1), 0
), buf4, primals_6, buf6, primals_4, buf7
class MLPBaseNew(nn.Module):
def __init__(self, num_inputs, num_outputs):
super(MLPBaseNew, self).__init__()
self.l1 = nn.Linear(num_inputs, 400)
self.l2 = nn.Linear(400, 300)
self.l3 = nn.Linear(300, num_outputs)
def forward(self, input_0):
primals_1 = self.l1.weight
primals_2 = self.l1.bias
primals_4 = self.l2.weight
primals_5 = self.l2.bias
primals_6 = self.l3.weight
primals_7 = self.l3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| TachikakaMin/dreamer-torch | MLPBase | false | 1,122 | [
"MIT"
] | 0 | 3c99526f4507e28cf8b34ada0321001adcf8ae1f | https://github.com/TachikakaMin/dreamer-torch/tree/3c99526f4507e28cf8b34ada0321001adcf8ae1f | import torch
from torch import nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, num_inputs, num_outputs):
super().__init__()
self.l1 = nn.Linear(num_inputs, 400)
self.l2 = nn.Linear(400, 300)
self.l3 = nn.Linear(300, num_outputs)
def forward(self, inputs):
x = F.relu(self.l1(inputs))
x = F.relu(self.l2(x))
x = self.l3(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
N_R_Align | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/dl/cdl26m2hy376ftir63aeqndptbftlfcqva4itwmmgwwbyrs6eth2.py
# Topologically Sorted Source Nodes: [cosine_similarity], Original ATen: [aten.linalg_vector_norm, aten.clamp_min, aten.div, aten.mul]
# Source node to ATen node mapping:
# cosine_similarity => clamp_min, clamp_min_1, div, div_1, mul, pow_1, pow_2, pow_3, pow_4, sum_1, sum_2
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg1_1, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1], True), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%pow_2, 1e-06), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg1_1, %clamp_min), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg0_1, 2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_3, [1], True), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_2, 0.5), kwargs = {})
# %clamp_min_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%pow_4, 1e-06), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg0_1, %clamp_min_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div_1, %div), kwargs = {})
triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0 = async_compile.triton('triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 10, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr1 + (x3), xmask)
tmp17 = tl.load(in_ptr1 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr1 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp22 = tl.load(in_ptr1 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr1 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-06
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tmp18 = tmp17 * tmp17
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = libdevice.sqrt(tmp27)
tmp29 = triton_helpers.maximum(tmp28, tmp13)
tmp30 = tmp16 / tmp29
tmp31 = tmp15 * tmp30
tl.store(out_ptr0 + (x3), tmp31, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/eu/ceuwdky2xdaefd6i42mj44mw54md2rx7qikfobocu3ikagjrvotq.py
# Topologically Sorted Source Nodes: [cosine_similarity, sigmoid, mul, cosine_similarity_1, sigmoid_1, mul_1, add], Original ATen: [aten.sum, aten.sigmoid, aten.mul, aten.add]
# Source node to ATen node mapping:
# add => add
# cosine_similarity => sum_3
# cosine_similarity_1 => sum_6
# mul => mul_1
# mul_1 => mul_3
# sigmoid => sigmoid
# sigmoid_1 => sigmoid_1
# Graph fragment:
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%sum_3,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid, 4), kwargs = {})
# %sum_6 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_2, [1]), kwargs = {})
# %sigmoid_1 : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%sum_6,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sigmoid_1, -3), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %mul_3), kwargs = {})
triton_poi_fused_add_mul_sigmoid_sum_1 = async_compile.triton('triton_poi_fused_add_mul_sigmoid_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mul_sigmoid_sum_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sum_1(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp10 = tl.load(in_ptr1 + (x0 + (64*x1)), xmask)
tmp11 = tl.load(in_ptr1 + (16 + x0 + (64*x1)), xmask)
tmp13 = tl.load(in_ptr1 + (32 + x0 + (64*x1)), xmask)
tmp15 = tl.load(in_ptr1 + (48 + x0 + (64*x1)), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp8 = 4.0
tmp9 = tmp7 * tmp8
tmp12 = tmp10 + tmp11
tmp14 = tmp12 + tmp13
tmp16 = tmp14 + tmp15
tmp17 = tl.sigmoid(tmp16)
tmp18 = -3.0
tmp19 = tmp17 * tmp18
tmp20 = tmp9 + tmp19
tl.store(out_ptr0 + (x2), tmp20, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1, arg2_1, arg3_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg3_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [cosine_similarity], Original ATen: [aten.linalg_vector_norm, aten.clamp_min, aten.div, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0.run(arg1_1, arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [cosine_similarity_1], Original ATen: [aten.linalg_vector_norm, aten.clamp_min, aten.div, aten.mul]
triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0.run(arg3_1, arg2_1, buf1, 256, grid=grid(256), stream=stream0)
del arg2_1
del arg3_1
buf2 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [cosine_similarity, sigmoid, mul, cosine_similarity_1, sigmoid_1, mul_1, add], Original ATen: [aten.sum, aten.sigmoid, aten.mul, aten.add]
triton_poi_fused_add_mul_sigmoid_sum_1.run(buf0, buf1, buf2, 64, grid=grid(64), stream=stream0)
del buf0
del buf1
return (buf2, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg3_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1, arg2_1, arg3_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class N_R_Align(torch.nn.Module):
def __init__(self, params):
super(N_R_Align, self).__init__()
self.params = params
self.cos_sim = nn.CosineSimilarity(dim=1, eps=1e-06)
def forward(self, e1, e2, n1, n2):
return self.params * torch.sigmoid(self.cos_sim(n1, n2)) + (1 -
self.params) * torch.sigmoid(self.cos_sim(e1, e2))
def loss(self, pos_score, neg_score, target):
return -torch.log(pos_score).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'params': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0(in_ptr0,
in_ptr1, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp16 = tl.load(in_ptr1 + x3, xmask)
tmp17 = tl.load(in_ptr1 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp19 = tl.load(in_ptr1 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp22 = tl.load(in_ptr1 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp25 = tl.load(in_ptr1 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-06
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tmp18 = tmp17 * tmp17
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp23 = tmp22 * tmp22
tmp24 = tmp21 + tmp23
tmp26 = tmp25 * tmp25
tmp27 = tmp24 + tmp26
tmp28 = libdevice.sqrt(tmp27)
tmp29 = triton_helpers.maximum(tmp28, tmp13)
tmp30 = tmp16 / tmp29
tmp31 = tmp15 * tmp30
tl.store(out_ptr0 + x3, tmp31, xmask)
@triton.jit
def triton_poi_fused_add_mul_sigmoid_sum_1(in_ptr0, in_ptr1, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp10 = tl.load(in_ptr1 + (x0 + 64 * x1), xmask)
tmp11 = tl.load(in_ptr1 + (16 + x0 + 64 * x1), xmask)
tmp13 = tl.load(in_ptr1 + (32 + x0 + 64 * x1), xmask)
tmp15 = tl.load(in_ptr1 + (48 + x0 + 64 * x1), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = tl.sigmoid(tmp6)
tmp8 = 4.0
tmp9 = tmp7 * tmp8
tmp12 = tmp10 + tmp11
tmp14 = tmp12 + tmp13
tmp16 = tmp14 + tmp15
tmp17 = tl.sigmoid(tmp16)
tmp18 = -3.0
tmp19 = tmp17 * tmp18
tmp20 = tmp9 + tmp19
tl.store(out_ptr0 + x2, tmp20, xmask)
def call(args):
arg0_1, arg1_1, arg2_1, arg3_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg3_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0[grid(256)](
arg1_1, arg0_1, buf0, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_clamp_min_div_linalg_vector_norm_mul_0[grid(256)](
arg3_1, arg2_1, buf1, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg2_1
del arg3_1
buf2 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_mul_sigmoid_sum_1[grid(64)](buf0, buf1, buf2,
64, XBLOCK=64, num_warps=1, num_stages=1)
del buf0
del buf1
return buf2,
class N_R_AlignNew(torch.nn.Module):
def __init__(self, params):
super(N_R_AlignNew, self).__init__()
self.params = params
self.cos_sim = nn.CosineSimilarity(dim=1, eps=1e-06)
def loss(self, pos_score, neg_score, target):
return -torch.log(pos_score).sum()
def forward(self, input_0, input_1, input_2, input_3):
arg0_1 = input_0
arg1_1 = input_1
arg2_1 = input_2
arg3_1 = input_3
output = call([arg0_1, arg1_1, arg2_1, arg3_1])
return output[0]
| TMUITLab/EAFR | N_R_Align | false | 1,123 | [
"MIT"
] | 0 | dadb6485d48711ccb8aa2f03760aeb437645f1ff | https://github.com/TMUITLab/EAFR/tree/dadb6485d48711ccb8aa2f03760aeb437645f1ff | import torch
import torch.nn as nn
class Model(torch.nn.Module):
def __init__(self, params):
super().__init__()
self.params = params
self.cos_sim = nn.CosineSimilarity(dim=1, eps=1e-06)
def forward(self, e1, e2, n1, n2):
return self.params * torch.sigmoid(self.cos_sim(n1, n2)) + (1 -
self.params) * torch.sigmoid(self.cos_sim(e1, e2))
def loss(self, pos_score, neg_score, target):
return -torch.log(pos_score).sum()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4]), torch.rand(
[4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
FC | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/zk/czkecze4wfcslvwkeplzfjv4ih2ndbfnpgq55lpnflnsjgnc72g4.py
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.leaky_relu, aten.view, aten.leaky_relu_backward]
# Source node to ATen node mapping:
# out_1 => gt, mul, view_3, where
# Graph fragment:
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%view_1, 0), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, 0.2), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %view_1, %mul), kwargs = {})
# %view_3 : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%view_2, [4, 4, 4, 4]), kwargs = {})
# %gt_1 : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%view_8, 0), kwargs = {})
triton_poi_fused_leaky_relu_leaky_relu_backward_view_0 = async_compile.triton('triton_poi_fused_leaky_relu_leaky_relu_backward_view_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_leaky_relu_leaky_relu_backward_view_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_leaky_relu_leaky_relu_backward_view_0(in_out_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (x0), xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 0.2
tmp4 = tmp0 * tmp3
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp6 = tmp5 > tmp1
tl.store(out_ptr0 + (x0), tmp5, xmask)
tl.store(out_ptr1 + (x0), tmp6, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf0 # reuse
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.leaky_relu, aten.view, aten.leaky_relu_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_leaky_relu_leaky_relu_backward_view_0.run(buf1, buf2, buf3, 256, grid=grid(256), stream=stream0)
del buf1
return (buf2, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.nn.functional as F
class FC(nn.Module):
def __init__(self, in_channels, out_channels, use_bias=False,
activation='LR', gain=2 ** 0.5):
super(FC, self).__init__()
self.he_std = in_channels * -0.5 * gain
self.weight = torch.nn.Parameter(torch.randn(out_channels,
in_channels) * self.he_std)
if use_bias:
self.bias = torch.nn.Parameter(torch.zeros(out_channels))
else:
self.bias = None
if activation == 'LR':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'Sigmoid':
self.activation = nn.Sigmoid()
elif activation is None:
self.activation = None
else:
assert 0, " STGAN's FC reruires LR or Sigmoid, not{}".format(
activation)
def forward(self, x):
if self.bias is not None:
out = F.linear(x, self.weight, self.bias)
else:
out = F.linear(x, self.weight)
if self.activation:
out = self.activation(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_leaky_relu_leaky_relu_backward_view_0(in_out_ptr0,
out_ptr0, out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + x0, xmask)
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp3 = 0.2
tmp4 = tmp0 * tmp3
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp6 = tmp5 > tmp1
tl.store(out_ptr0 + x0, tmp5, xmask)
tl.store(out_ptr1 + x0, tmp6, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf0
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_leaky_relu_leaky_relu_backward_view_0[grid(256)](buf1,
buf2, buf3, 256, XBLOCK=128, num_warps=4, num_stages=1)
del buf1
return buf2, reinterpret_tensor(primals_2, (64, 4), (4, 1), 0), buf3
class FCNew(nn.Module):
def __init__(self, in_channels, out_channels, use_bias=False,
activation='LR', gain=2 ** 0.5):
super(FCNew, self).__init__()
self.he_std = in_channels * -0.5 * gain
self.weight = torch.nn.Parameter(torch.randn(out_channels,
in_channels) * self.he_std)
if use_bias:
self.bias = torch.nn.Parameter(torch.zeros(out_channels))
else:
self.bias = None
if activation == 'LR':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'Sigmoid':
self.activation = nn.Sigmoid()
elif activation is None:
self.activation = None
else:
assert 0, " STGAN's FC reruires LR or Sigmoid, not{}".format(
activation)
def forward(self, input_0):
primals_1 = self.weight
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| TOMeoww/STGAN | FC | false | 1,124 | [
"MIT"
] | 0 | 090a4024999e68f017140312ecfdd0d4dc3dc425 | https://github.com/TOMeoww/STGAN/tree/090a4024999e68f017140312ecfdd0d4dc3dc425 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, in_channels, out_channels, use_bias=False,
activation='LR', gain=2 ** 0.5):
super().__init__()
self.he_std = in_channels * -0.5 * gain
self.weight = torch.nn.Parameter(torch.randn(out_channels,
in_channels) * self.he_std)
if use_bias:
self.bias = torch.nn.Parameter(torch.zeros(out_channels))
else:
self.bias = None
if activation == 'LR':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'Sigmoid':
self.activation = nn.Sigmoid()
elif activation is None:
self.activation = None
else:
assert 0, " STGAN's FC reruires LR or Sigmoid, not{}".format(
activation)
def forward(self, x):
if self.bias is not None:
out = F.linear(x, self.weight, self.bias)
else:
out = F.linear(x, self.weight)
if self.activation:
out = self.activation(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Mean | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/7g/c7gvt7fcrwihu3eypgqeru4bybzepmpynh45oes44jy2ps6isyr7.py
# Topologically Sorted Source Nodes: [std_mean], Original ATen: [aten.std_mean]
# Source node to ATen node mapping:
# std_mean => var_mean
# Graph fragment:
# %var_mean : [num_users=1] = call_function[target=torch.ops.aten.var_mean.correction](args = (%arg0_1, [4]), kwargs = {correction: 1.0})
triton_poi_fused_std_mean_0 = async_compile.triton('triton_poi_fused_std_mean_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_std_mean_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_std_mean_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tl.store(out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4, 4), (256, 64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 256), torch.float32)
# Topologically Sorted Source Nodes: [std_mean], Original ATen: [aten.std_mean]
stream0 = get_raw_stream(0)
triton_poi_fused_std_mean_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
class Mean(torch.nn.Module):
def __init__(self, dim):
super().__init__()
self.dim = dim
def forward(self, x):
_std, mean = torch.std_mean(x, self.dim)
return mean
def get_inputs():
return [torch.rand([4, 4, 4, 4, 4])]
def get_init_inputs():
return [[], {'dim': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_std_mean_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tl.store(out_ptr0 + x0, tmp8, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4, 4), (256, 64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 256),
torch.float32)
get_raw_stream(0)
triton_poi_fused_std_mean_0[grid(256)](arg0_1, buf0, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg0_1
return reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0),
class MeanNew(torch.nn.Module):
def __init__(self, dim):
super().__init__()
self.dim = dim
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Tahlor/glom-pytorch | Mean | false | 1,125 | [
"MIT"
] | 0 | 45b2fc52af5288cd53611e497a70d53ffa303410 | https://github.com/Tahlor/glom-pytorch/tree/45b2fc52af5288cd53611e497a70d53ffa303410 | import torch
class Model(torch.nn.Module):
def __init__(self, dim):
super().__init__()
self.dim = dim
def forward(self, x):
_std, mean = torch.std_mean(x, self.dim)
return mean
def get_inputs():
return [torch.rand([4, 4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
LinearModel | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/yy/cyya3js6wt64vdji3sfisvrqyfvqxwkwqq5mzg5bqjl2crzjs4t3.py
# Topologically Sorted Source Nodes: [hidden], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# hidden => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%select,), kwargs = {})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = (xindex // 16)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (64*x1)), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [hidden], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(primals_1, buf0, 64, grid=grid(64), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_2
del primals_3
return (reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf0, (16, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
class LinearModel(torch.nn.Module):
def __init__(self, input_size: 'int', output_size: 'int', dropout: 'float'
):
super().__init__()
self.linear = torch.nn.Linear(input_size, output_size)
self.dropout = torch.nn.Dropout(dropout)
def forward(self, data):
data = data[:, 0, :]
hidden = self.dropout(data)
output = self.linear(hidden)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'output_size': 4, 'dropout': 0.5}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 16
x1 = xindex // 16
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 64 * x1), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(64)](primals_1, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(buf0, (16, 4), (
4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf1)
del primals_2
del primals_3
return reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0
), reinterpret_tensor(buf0, (16, 4), (4, 1), 0)
class LinearModelNew(torch.nn.Module):
def __init__(self, input_size: 'int', output_size: 'int', dropout: 'float'
):
super().__init__()
self.linear = torch.nn.Linear(input_size, output_size)
self.dropout = torch.nn.Dropout(dropout)
def forward(self, input_0):
primals_2 = self.linear.weight
primals_3 = self.linear.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| TDteach/SEAM | LinearModel | false | 1,126 | [
"MIT"
] | 0 | 231447dad15403e7620adcf6629b6e7fccc4b809 | https://github.com/TDteach/SEAM/tree/231447dad15403e7620adcf6629b6e7fccc4b809 | import torch
class Model(torch.nn.Module):
def __init__(self, input_size: 'int', output_size: 'int', dropout: 'float'
):
super().__init__()
self.linear = torch.nn.Linear(input_size, output_size)
self.dropout = torch.nn.Dropout(dropout)
def forward(self, data):
data = data[:, 0, :]
hidden = self.dropout(data)
output = self.linear(hidden)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 0.5]
|
GeometricMean | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/53/c536dq7bncdpsnuyp7jtn2kgbqzlbq4i6q7p4bjliwwyfzwkz7sq.py
# Topologically Sorted Source Nodes: [relu, log_x, mean, exp], Original ATen: [aten.relu, aten.log, aten.mean, aten.exp]
# Source node to ATen node mapping:
# exp => exp
# log_x => log
# mean => mean
# relu => relu
# Graph fragment:
# %relu : [num_users=1] = call_function[target=torch.ops.aten.relu.default](args = (%arg0_1,), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%relu,), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%log, [4]), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%mean,), kwargs = {})
triton_poi_fused_exp_log_mean_relu_0 = async_compile.triton('triton_poi_fused_exp_log_mean_relu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_exp_log_mean_relu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_exp_log_mean_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (4*x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = tl_math.log(tmp2)
tmp5 = triton_helpers.maximum(tmp1, tmp4)
tmp6 = tl_math.log(tmp5)
tmp7 = tmp3 + tmp6
tmp9 = triton_helpers.maximum(tmp1, tmp8)
tmp10 = tl_math.log(tmp9)
tmp11 = tmp7 + tmp10
tmp13 = triton_helpers.maximum(tmp1, tmp12)
tmp14 = tl_math.log(tmp13)
tmp15 = tmp11 + tmp14
tmp16 = 4.0
tmp17 = tmp15 / tmp16
tmp18 = tl_math.exp(tmp17)
tl.store(out_ptr0 + (x0), tmp18, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4, 4), (256, 64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [relu, log_x, mean, exp], Original ATen: [aten.relu, aten.log, aten.mean, aten.exp]
stream0 = get_raw_stream(0)
triton_poi_fused_exp_log_mean_relu_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
class GeometricMean(torch.nn.Module):
def __init__(self, dim):
super().__init__()
self.dim = dim
def forward(self, x):
log_x = torch.log(F.relu(x))
return torch.exp(torch.mean(log_x, dim=self.dim))
def get_inputs():
return [torch.rand([4, 4, 4, 4, 4])]
def get_init_inputs():
return [[], {'dim': 4}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_exp_log_mean_relu_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x0, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = tl_math.log(tmp2)
tmp5 = triton_helpers.maximum(tmp1, tmp4)
tmp6 = tl_math.log(tmp5)
tmp7 = tmp3 + tmp6
tmp9 = triton_helpers.maximum(tmp1, tmp8)
tmp10 = tl_math.log(tmp9)
tmp11 = tmp7 + tmp10
tmp13 = triton_helpers.maximum(tmp1, tmp12)
tmp14 = tl_math.log(tmp13)
tmp15 = tmp11 + tmp14
tmp16 = 4.0
tmp17 = tmp15 / tmp16
tmp18 = tl_math.exp(tmp17)
tl.store(out_ptr0 + x0, tmp18, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4, 4), (256, 64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_exp_log_mean_relu_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class GeometricMeanNew(torch.nn.Module):
def __init__(self, dim):
super().__init__()
self.dim = dim
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Tahlor/glom-pytorch | GeometricMean | false | 1,127 | [
"MIT"
] | 0 | 45b2fc52af5288cd53611e497a70d53ffa303410 | https://github.com/Tahlor/glom-pytorch/tree/45b2fc52af5288cd53611e497a70d53ffa303410 | import torch
import torch.nn.functional as F
class Model(torch.nn.Module):
def __init__(self, dim):
super().__init__()
self.dim = dim
def forward(self, x):
log_x = torch.log(F.relu(x))
return torch.exp(torch.mean(log_x, dim=self.dim))
def get_inputs():
return [torch.rand([4, 4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
MinibatchStd | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/fe/cfe7ai6nrnajldrnau4a6rtqqrk2wdm7opgj7c5duuavburs6m7o.py
# Topologically Sorted Source Nodes: [std, mean, repeat], Original ATen: [aten.std, aten.mean, aten.repeat]
# Source node to ATen node mapping:
# mean => mean
# repeat => repeat
# std => sqrt, var
# Graph fragment:
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%arg0_1, [0]), kwargs = {correction: 1.0})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%var,), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sqrt,), kwargs = {})
# %repeat : [num_users=1] = call_function[target=torch.ops.aten.repeat.default](args = (%mean, [4, 1, 4, 4]), kwargs = {})
triton_per_fused_mean_repeat_std_0 = async_compile.triton('triton_per_fused_mean_repeat_std_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 64],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {2: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 3), equal_to_1=(2,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_mean_repeat_std_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_mean_repeat_std_0(in_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
r1 = rindex % 16
r2 = (rindex // 16)
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.load(in_ptr0 + (64 + r0), None)
tmp3 = tl.load(in_ptr0 + (128 + r0), None)
tmp5 = tl.load(in_ptr0 + (192 + r0), None)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = 3.0
tmp21 = tmp19 / tmp20
tmp22 = libdevice.sqrt(tmp21)
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.sum(tmp23, 1)[:, None]
tmp26 = 64.0
tmp27 = tmp25 / tmp26
tl.store(out_ptr1 + (tl.broadcast_to(r1 + (80*r2), [XBLOCK, RBLOCK])), tmp27, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/yi/cyidf2yj3fms5jdxlfe7fdijzfj6p5a5q2qxo4llkuxnpqh6fj5o.py
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# cat => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%arg0_1, %repeat], 1), kwargs = {})
triton_poi_fused_cat_1 = async_compile.triton('triton_poi_fused_cat_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
x1 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tl.store(out_ptr0 + (x0 + (80*x1)), tmp0, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf3 = empty_strided_cuda((4, 5, 4, 4), (80, 16, 4, 1), torch.float32)
buf2 = reinterpret_tensor(buf3, (4, 1, 4, 4), (80, 16, 4, 1), 64) # alias
# Topologically Sorted Source Nodes: [std, mean, repeat], Original ATen: [aten.std, aten.mean, aten.repeat]
stream0 = get_raw_stream(0)
triton_per_fused_mean_repeat_std_0.run(arg0_1, buf2, 1, 64, grid=grid(1), stream=stream0)
buf1 = reinterpret_tensor(buf3, (4, 4, 4, 4), (80, 16, 4, 1), 0) # alias
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(arg0_1, buf1, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class MinibatchStd(nn.Module):
"""
calculate minibatch std to avoid mode collapse
"""
def __init__(self):
super(MinibatchStd, self).__init__()
def forward(self, x):
size = list(x.size())
size[1] = 1
std = torch.std(x, dim=0)
mean = torch.mean(std)
return torch.cat((x, mean.repeat(size)), dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_mean_repeat_std_0(in_ptr0, out_ptr1, xnumel, rnumel,
XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
r1 = rindex % 16
r2 = rindex // 16
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.load(in_ptr0 + (64 + r0), None)
tmp3 = tl.load(in_ptr0 + (128 + r0), None)
tmp5 = tl.load(in_ptr0 + (192 + r0), None)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tmp9 = tmp0 - tmp8
tmp10 = tmp9 * tmp9
tmp11 = tmp1 - tmp8
tmp12 = tmp11 * tmp11
tmp13 = tmp10 + tmp12
tmp14 = tmp3 - tmp8
tmp15 = tmp14 * tmp14
tmp16 = tmp13 + tmp15
tmp17 = tmp5 - tmp8
tmp18 = tmp17 * tmp17
tmp19 = tmp16 + tmp18
tmp20 = 3.0
tmp21 = tmp19 / tmp20
tmp22 = libdevice.sqrt(tmp21)
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.sum(tmp23, 1)[:, None]
tmp26 = 64.0
tmp27 = tmp25 / tmp26
tl.store(out_ptr1 + tl.broadcast_to(r1 + 80 * r2, [XBLOCK, RBLOCK]),
tmp27, None)
@triton.jit
def triton_poi_fused_cat_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 64
x1 = xindex // 64
tmp0 = tl.load(in_ptr0 + x2, xmask)
tl.store(out_ptr0 + (x0 + 80 * x1), tmp0, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf3 = empty_strided_cuda((4, 5, 4, 4), (80, 16, 4, 1), torch.float32)
buf2 = reinterpret_tensor(buf3, (4, 1, 4, 4), (80, 16, 4, 1), 64)
get_raw_stream(0)
triton_per_fused_mean_repeat_std_0[grid(1)](arg0_1, buf2, 1, 64,
XBLOCK=1, num_warps=2, num_stages=1)
buf1 = reinterpret_tensor(buf3, (4, 4, 4, 4), (80, 16, 4, 1), 0)
triton_poi_fused_cat_1[grid(256)](arg0_1, buf1, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del arg0_1
return buf3,
class MinibatchStdNew(nn.Module):
"""
calculate minibatch std to avoid mode collapse
"""
def __init__(self):
super(MinibatchStdNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Tak-jae-ho/RGBD-GAN-pytorch | MinibatchStd | false | 1,128 | [
"MIT"
] | 0 | 4fb1bc1de7b7807fd4f2d346d9b688a2d257eedb | https://github.com/Tak-jae-ho/RGBD-GAN-pytorch/tree/4fb1bc1de7b7807fd4f2d346d9b688a2d257eedb | import torch
import torch.nn as nn
class Model(nn.Module):
"""
calculate minibatch std to avoid mode collapse
"""
def __init__(self):
super().__init__()
def forward(self, x):
size = list(x.size())
size[1] = 1
std = torch.std(x, dim=0)
mean = torch.mean(std)
return torch.cat((x, mean.repeat(size)), dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
PixelwiseNorm | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/ao/caol5mhvrwgvqsgq2tai5vbpmasyjuuavozutxehikyh6g7bgyvg.py
# Topologically Sorted Source Nodes: [pow_1, sum_1, add, sqrt, truediv], Original ATen: [aten.pow, aten.sum, aten.add, aten.sqrt, aten.div]
# Source node to ATen node mapping:
# add => add
# pow_1 => pow_1
# sqrt => sqrt
# sum_1 => sum_1
# truediv => div
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%arg0_1, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1], True), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, 1e-07), kwargs = {})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%add,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg0_1, %sqrt), kwargs = {})
triton_poi_fused_add_div_pow_sqrt_sum_0 = async_compile.triton('triton_poi_fused_add_div_pow_sqrt_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_div_pow_sqrt_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_div_pow_sqrt_sum_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = 1e-07
tmp13 = tmp11 + tmp12
tmp14 = libdevice.sqrt(tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x3), tmp15, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pow_1, sum_1, add, sqrt, truediv], Original ATen: [aten.pow, aten.sum, aten.add, aten.sqrt, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_pow_sqrt_sum_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
return (buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class PixelwiseNorm(nn.Module):
"""
layer pixelwise normalization
"""
def __init__(self, eps=1e-07):
super(PixelwiseNorm, self).__init__()
self.eps = eps
def forward(self, x):
return x / torch.sqrt(torch.sum(x ** 2, dim=1, keepdim=True) + self.eps
)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused_add_div_pow_sqrt_sum_0(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = 1e-07
tmp13 = tmp11 + tmp12
tmp14 = libdevice.sqrt(tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + x3, tmp15, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_pow_sqrt_sum_0[grid(256)](arg0_1, buf0,
256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class PixelwiseNormNew(nn.Module):
"""
layer pixelwise normalization
"""
def __init__(self, eps=1e-07):
super(PixelwiseNormNew, self).__init__()
self.eps = eps
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Tak-jae-ho/RGBD-GAN-pytorch | PixelwiseNorm | false | 1,129 | [
"MIT"
] | 0 | 4fb1bc1de7b7807fd4f2d346d9b688a2d257eedb | https://github.com/Tak-jae-ho/RGBD-GAN-pytorch/tree/4fb1bc1de7b7807fd4f2d346d9b688a2d257eedb | import torch
import torch.nn as nn
class Model(nn.Module):
"""
layer pixelwise normalization
"""
def __init__(self, eps=1e-07):
super().__init__()
self.eps = eps
def forward(self, x):
return x / torch.sqrt(torch.sum(x ** 2, dim=1, keepdim=True) + self.eps
)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
ConsensusAttention | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/qd/cqdnz7hcxxkadyci2slws6wn222hkkzcp6hb3x3b5ttuny2wks7b.py
# Topologically Sorted Source Nodes: [einsum, out], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# einsum => clone
# out => clone_2
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_2,), kwargs = {memory_format: torch.contiguous_format})
# %clone_2 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_8,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_0 = async_compile.triton('triton_poi_fused_clone_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = (xindex // 4) % 4
x2 = (xindex // 16) % 4
x3 = (xindex // 64)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2) + (16*x1) + (64*x3)), xmask)
tl.store(out_ptr0 + (x4), tmp0, xmask)
tl.store(out_ptr1 + (x4), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/w6/cw6bzt4yaqzfq7jhtvpqzm42cbssdg77yhymaouoko5jz36q3rhl.py
# Topologically Sorted Source Nodes: [einsum], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# einsum => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_3,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_1 = async_compile.triton('triton_poi_fused_clone_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64, 4], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_clone_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x3 = xindex
y2 = (yindex // 16)
y4 = yindex % 16
y1 = (yindex // 4) % 4
y5 = yindex
tmp0 = tl.load(in_ptr0 + (y4 + (16*x3) + (64*y2)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + ((4*y1) + (16*x3) + (64*y2)), xmask & ymask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*y1) + (16*x3) + (64*y2)), xmask & ymask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + (4*y1) + (16*x3) + (64*y2)), xmask & ymask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + (4*y1) + (16*x3) + (64*y2)), xmask & ymask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-12
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x3 + (4*y5)), tmp15, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/7n/c7nhob3voygbzg73wacfq2bbcquxe5ozjclkx7uda5hunwppje36.py
# Topologically Sorted Source Nodes: [attn], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attn => exp
# Graph fragment:
# %mul_tensor : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, 1), kwargs = {})
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%mul_tensor, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_tensor, %amax_default), kwargs = {})
# %mul_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_tensor, 0.5), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%mul_tensor_1,), kwargs = {})
triton_poi_fused__softmax_2 = async_compile.triton('triton_poi_fused__softmax_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = 0.5
tmp16 = tmp14 * tmp15
tmp17 = tl_math.exp(tmp16)
tl.store(out_ptr0 + (x2), tmp17, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ry/cryn7ntc2gpkbfzbre3xh7lffx7zkbskw6oihbzsekkgajmdbki6.py
# Topologically Sorted Source Nodes: [attn], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attn => div_1, sum_2
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_2), kwargs = {})
triton_poi_fused__softmax_3 = async_compile.triton('triton_poi_fused__softmax_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__softmax_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [einsum, out], Original ATen: [aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_0.run(arg0_1, buf0, buf5, 256, grid=grid(256), stream=stream0)
buf1 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [einsum], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(arg0_1, buf1, 64, 4, grid=grid(64, 4), stream=stream0)
del arg0_1
buf2 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [einsum], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0), out=buf2)
del buf0
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [attn], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [attn], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
buf6 = reinterpret_tensor(buf3, (16, 4, 4), (16, 4, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4, 4), (16, 4, 1), 0), out=buf6)
del buf4
del buf5
return (reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 4, 16, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn.functional as F
from torch import nn
from torch import einsum
class ConsensusAttention(nn.Module):
def __init__(self, num_patches_side, attend_self=True,
local_consensus_radius=0):
super().__init__()
self.attend_self = attend_self
self.local_consensus_radius = local_consensus_radius
if self.local_consensus_radius > 0:
coors = torch.stack(torch.meshgrid(torch.arange(
num_patches_side), torch.arange(num_patches_side))).float()
coors = rearrange(coors, 'c h w -> (h w) c')
dist = torch.cdist(coors, coors)
mask_non_local = dist > self.local_consensus_radius
mask_non_local = rearrange(mask_non_local, 'i j -> () i j')
self.register_buffer('non_local_mask', mask_non_local)
def forward(self, levels):
_, n, _, d, device = *levels.shape, levels.device
q, k, _v = levels, F.normalize(levels, dim=-1), levels
sim = einsum('b i l d, b j l d -> b l i j', q, k
) * d ** -0.5
if not self.attend_self:
self_mask = torch.eye(n, device=device, dtype=torch.bool)
self_mask = rearrange(self_mask, 'i j -> () () i j')
sim.masked_fill_(self_mask, TOKEN_ATTEND_SELF_VALUE)
if self.local_consensus_radius > 0:
max_neg_value = -torch.finfo(sim.dtype).max
sim.masked_fill_(self.non_local_mask, max_neg_value)
attn = sim.softmax(dim=-1)
out = einsum('b l i j, b j l d -> b i l d', attn, levels)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_patches_side': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_clone_0(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x1 = xindex // 4 % 4
x2 = xindex // 16 % 4
x3 = xindex // 64
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2 + 16 * x1 + 64 * x3), xmask)
tl.store(out_ptr0 + x4, tmp0, xmask)
tl.store(out_ptr1 + x4, tmp0, xmask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 4
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x3 = xindex
y2 = yindex // 16
y4 = yindex % 16
y1 = yindex // 4 % 4
y5 = yindex
tmp0 = tl.load(in_ptr0 + (y4 + 16 * x3 + 64 * y2), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (4 * y1 + 16 * x3 + 64 * y2), xmask & ymask,
eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * y1 + 16 * x3 + 64 * y2), xmask &
ymask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + 4 * y1 + 16 * x3 + 64 * y2), xmask &
ymask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + 4 * y1 + 16 * x3 + 64 * y2), xmask &
ymask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-12
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x3 + 4 * y5), tmp15, xmask & ymask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tmp6 = tmp5 * tmp1
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = tmp8 * tmp1
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = tmp11 * tmp1
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = 0.5
tmp16 = tmp14 * tmp15
tmp17 = tl_math.exp(tmp16)
tl.store(out_ptr0 + x2, tmp17, xmask)
@triton.jit
def triton_poi_fused__softmax_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
def call(args):
arg0_1, = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
get_raw_stream(0)
triton_poi_fused_clone_0[grid(256)](arg0_1, buf0, buf5, 256, XBLOCK
=256, num_warps=4, num_stages=1)
buf1 = empty_strided_cuda((4, 4, 4, 4, 1), (64, 16, 4, 1, 1), torch
.float32)
triton_poi_fused_clone_1[grid(64, 4)](arg0_1, buf1, 64, 4, XBLOCK=4,
YBLOCK=32, num_warps=4, num_stages=1)
del arg0_1
buf2 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf1, (16, 4, 4), (16, 4, 1), 0), out=buf2)
del buf0
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf1
triton_poi_fused__softmax_2[grid(256)](buf2, buf3, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf4 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf2
triton_poi_fused__softmax_3[grid(256)](buf3, buf4, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf6 = reinterpret_tensor(buf3, (16, 4, 4), (16, 4, 1), 0)
del buf3
extern_kernels.bmm(reinterpret_tensor(buf4, (16, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf5, (16, 4, 4), (16, 4, 1), 0), out=buf6)
del buf4
del buf5
return reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 4, 16, 1), 0),
class ConsensusAttentionNew(nn.Module):
def __init__(self, num_patches_side, attend_self=True,
local_consensus_radius=0):
super().__init__()
self.attend_self = attend_self
self.local_consensus_radius = local_consensus_radius
if self.local_consensus_radius > 0:
coors = torch.stack(torch.meshgrid(torch.arange(
num_patches_side), torch.arange(num_patches_side))).float()
coors = rearrange(coors, 'c h w -> (h w) c')
dist = torch.cdist(coors, coors)
mask_non_local = dist > self.local_consensus_radius
mask_non_local = rearrange(mask_non_local, 'i j -> () i j')
self.register_buffer('non_local_mask', mask_non_local)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| Tahlor/glom-pytorch | ConsensusAttention | false | 1,130 | [
"MIT"
] | 0 | 45b2fc52af5288cd53611e497a70d53ffa303410 | https://github.com/Tahlor/glom-pytorch/tree/45b2fc52af5288cd53611e497a70d53ffa303410 | import torch
import torch.nn.functional as F
from torch import nn
from torch import einsum
class Model(nn.Module):
def __init__(self, num_patches_side, attend_self=True,
local_consensus_radius=0):
super().__init__()
self.attend_self = attend_self
self.local_consensus_radius = local_consensus_radius
if self.local_consensus_radius > 0:
coors = torch.stack(torch.meshgrid(torch.arange(
num_patches_side), torch.arange(num_patches_side))).float()
coors = rearrange(coors, 'c h w -> (h w) c')
dist = torch.cdist(coors, coors)
mask_non_local = dist > self.local_consensus_radius
mask_non_local = rearrange(mask_non_local, 'i j -> () i j')
self.register_buffer('non_local_mask', mask_non_local)
def forward(self, levels):
_, n, _, d, device = *levels.shape, levels.device
q, k, _v = levels, F.normalize(levels, dim=-1), levels
sim = einsum('b i l d, b j l d -> b l i j', q, k
) * d ** -0.5
if not self.attend_self:
self_mask = torch.eye(n, device=device, dtype=torch.bool)
self_mask = rearrange(self_mask, 'i j -> () () i j')
sim.masked_fill_(self_mask, TOKEN_ATTEND_SELF_VALUE)
if self.local_consensus_radius > 0:
max_neg_value = -torch.finfo(sim.dtype).max
sim.masked_fill_(self.non_local_mask, max_neg_value)
attn = sim.softmax(dim=-1)
out = einsum('b l i j, b j l d -> b i l d', attn, levels)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
DenseCrossEntropy | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/nr/cnrkptzsuv7qm3ss6i6xgoxkou23z76h2vmwqkwz2zkgpdbxhedc.py
# Topologically Sorted Source Nodes: [logprobs], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# logprobs => amax, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%arg0_1, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %amax), kwargs = {})
triton_poi_fused__log_softmax_0 = async_compile.triton('triton_poi_fused__log_softmax_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__log_softmax_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/nb/cnbdxy34iv6vkig4bfuqrxbegug3ek6lhyugevz3qctt7efdvtge.py
# Topologically Sorted Source Nodes: [logprobs, neg, loss, loss_1, mean], Original ATen: [aten._log_softmax, aten.neg, aten.mul, aten.sum, aten.mean]
# Source node to ATen node mapping:
# logprobs => exp, log, sub_1, sum_1
# loss => mul
# loss_1 => sum_2
# mean => mean
# neg => neg
# Graph fragment:
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp, [-1], True), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %log), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%sub_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%neg, %arg1_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [-1]), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sum_2,), kwargs = {})
triton_per_fused__log_softmax_mean_mul_neg_sum_1 = async_compile.triton('triton_per_fused__log_softmax_mean_mul_neg_sum_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 64],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__log_softmax_mean_mul_neg_sum_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused__log_softmax_mean_mul_neg_sum_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 1
rnumel = 64
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (4*r0), None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr1 + (4*r0), None, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr1 + (1 + (4*r0)), None, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr1 + (2 + (4*r0)), None, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr1 + (3 + (4*r0)), None, eviction_policy='evict_last')
tmp1 = tl_math.exp(tmp0)
tmp3 = tl_math.exp(tmp2)
tmp4 = tmp1 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp13 = -tmp12
tmp15 = tmp13 * tmp14
tmp16 = tmp2 - tmp11
tmp17 = -tmp16
tmp19 = tmp17 * tmp18
tmp20 = tmp15 + tmp19
tmp21 = tmp5 - tmp11
tmp22 = -tmp21
tmp24 = tmp22 * tmp23
tmp25 = tmp20 + tmp24
tmp26 = tmp8 - tmp11
tmp27 = -tmp26
tmp29 = tmp27 * tmp28
tmp30 = tmp25 + tmp29
tmp31 = tl.broadcast_to(tmp30, [XBLOCK, RBLOCK])
tmp33 = tl.sum(tmp31, 1)[:, None]
tmp34 = 64.0
tmp35 = tmp33 / tmp34
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp35, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [logprobs], Original ATen: [aten._log_softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
buf2 = empty_strided_cuda((), (), torch.float32)
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [logprobs, neg, loss, loss_1, mean], Original ATen: [aten._log_softmax, aten.neg, aten.mul, aten.sum, aten.mean]
triton_per_fused__log_softmax_mean_mul_neg_sum_1.run(buf3, buf0, arg1_1, 1, 64, grid=grid(1), stream=stream0)
del arg1_1
del buf0
return (buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data
import torch.nn.parallel
class DenseCrossEntropy(nn.Module):
def forward(self, x, target):
x = x.float()
target = target.float()
logprobs = torch.nn.functional.log_softmax(x, dim=-1)
loss = -logprobs * target
loss = loss.sum(-1)
return loss.mean()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.utils.data
import torch.nn.parallel
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp0 - tmp7
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_per_fused__log_softmax_mean_mul_neg_sum_1(in_out_ptr0, in_ptr0,
in_ptr1, xnumel, rnumel, XBLOCK: tl.constexpr):
RBLOCK: tl.constexpr = 64
xoffset = tl.program_id(0) * XBLOCK
xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + 4 * r0, None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (2 + 4 * r0), None, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (3 + 4 * r0), None, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr1 + 4 * r0, None, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr1 + (1 + 4 * r0), None, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr1 + (2 + 4 * r0), None, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr1 + (3 + 4 * r0), None, eviction_policy='evict_last')
tmp1 = tl_math.exp(tmp0)
tmp3 = tl_math.exp(tmp2)
tmp4 = tmp1 + tmp3
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp4 + tmp6
tmp9 = tl_math.exp(tmp8)
tmp10 = tmp7 + tmp9
tmp11 = tl_math.log(tmp10)
tmp12 = tmp0 - tmp11
tmp13 = -tmp12
tmp15 = tmp13 * tmp14
tmp16 = tmp2 - tmp11
tmp17 = -tmp16
tmp19 = tmp17 * tmp18
tmp20 = tmp15 + tmp19
tmp21 = tmp5 - tmp11
tmp22 = -tmp21
tmp24 = tmp22 * tmp23
tmp25 = tmp20 + tmp24
tmp26 = tmp8 - tmp11
tmp27 = -tmp26
tmp29 = tmp27 * tmp28
tmp30 = tmp25 + tmp29
tmp31 = tl.broadcast_to(tmp30, [XBLOCK, RBLOCK])
tmp33 = tl.sum(tmp31, 1)[:, None]
tmp34 = 64.0
tmp35 = tmp33 / tmp34
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp35, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
buf2 = empty_strided_cuda((), (), torch.float32)
buf3 = buf2
del buf2
triton_per_fused__log_softmax_mean_mul_neg_sum_1[grid(1)](buf3,
buf0, arg1_1, 1, 64, XBLOCK=1, num_warps=2, num_stages=1)
del arg1_1
del buf0
return buf3,
class DenseCrossEntropyNew(nn.Module):
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| Tanmengxuan/Google-Landmark-Recognition-2020-3rd-Place-Solution | DenseCrossEntropy | false | 1,131 | [
"Apache-2.0"
] | 0 | 8e2d9056d5c88c6415827086809e73522b336fbb | https://github.com/Tanmengxuan/Google-Landmark-Recognition-2020-3rd-Place-Solution/tree/8e2d9056d5c88c6415827086809e73522b336fbb | import torch
import torch.nn as nn
import torch.utils.data
import torch.nn.parallel
class Model(nn.Module):
def forward(self, x, target):
x = x.float()
target = target.float()
logprobs = torch.nn.functional.log_softmax(x, dim=-1)
loss = -logprobs * target
loss = loss.sum(-1)
return loss.mean()
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
HalfMSELoss | # AOT ID: ['0_inference']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/wj/cwj2vuvocpmuwf7kyymuynpm2f23fwqwm7fmbyjxaws24h4obhbj.py
# Topologically Sorted Source Nodes: [mse_loss, truediv], Original ATen: [aten.mse_loss, aten.div]
# Source node to ATen node mapping:
# mse_loss => mean, pow_1, sub
# truediv => div
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %arg0_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%pow_1,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mean, 2), kwargs = {})
triton_per_fused_div_mse_loss_0 = async_compile.triton('triton_per_fused_div_mse_loss_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[1, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {3: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=(3,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_mse_loss_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_div_mse_loss_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel, rnumel):
xnumel = 1
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + (r0), None)
tmp1 = tl.load(in_ptr1 + (r0), None)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp4 = tl.broadcast_to(tmp3, [RBLOCK])
tmp6 = triton_helpers.promote_to_tensor(tl.sum(tmp4, 0))
tmp7 = 256.0
tmp8 = tmp6 / tmp7
tmp9 = 0.5
tmp10 = tmp8 * tmp9
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp10, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [mse_loss, truediv], Original ATen: [aten.mse_loss, aten.div]
stream0 = get_raw_stream(0)
triton_per_fused_div_mse_loss_0.run(buf1, arg1_1, arg0_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
return (buf1, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
arg0_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([arg0_1, arg1_1])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch.nn.modules.loss import MSELoss
class HalfMSELoss(MSELoss):
def __init__(self, reduction='mean'):
super().__init__(reduction=reduction)
def forward(self, input, target):
return super().forward(input, target) / 2
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {}]
| import torch
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch.nn.modules.loss import MSELoss
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_per_fused_div_mse_loss_0(in_out_ptr0, in_ptr0, in_ptr1, xnumel,
rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r0 = rindex
tmp0 = tl.load(in_ptr0 + r0, None)
tmp1 = tl.load(in_ptr1 + r0, None)
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp4 = tl.broadcast_to(tmp3, [RBLOCK])
tmp6 = triton_helpers.promote_to_tensor(tl.sum(tmp4, 0))
tmp7 = 256.0
tmp8 = tmp6 / tmp7
tmp9 = 0.5
tmp10 = tmp8 * tmp9
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp10, None)
def call(args):
arg0_1, arg1_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((), (), torch.float32)
buf1 = buf0
del buf0
get_raw_stream(0)
triton_per_fused_div_mse_loss_0[grid(1)](buf1, arg1_1, arg0_1, 1,
256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf1,
class HalfMSELossNew(MSELoss):
def __init__(self, reduction='mean'):
super().__init__(reduction=reduction)
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| ThayaFluss/candle | HalfMSELoss | false | 1,132 | [
"MIT"
] | 0 | 4a12fde60ffbbf0cb688617fee81aded94c0b613 | https://github.com/ThayaFluss/candle/tree/4a12fde60ffbbf0cb688617fee81aded94c0b613 | import torch
from torch.nn.modules.loss import MSELoss
class Model(MSELoss):
def __init__(self, reduction='mean'):
super().__init__(reduction=reduction)
def forward(self, input, target):
return super().forward(input, target) / 2
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
EqualLinear | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/fl/cflw6zjzdk2wqtau7m6nsei5vavjfijzxhb37zaa3xp4yxpw5yb2.py
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, 1), kwargs = {})
triton_poi_fused_mul_0 = async_compile.triton('triton_poi_fused_mul_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/2o/c2oqkq7zaubqmw7vuixxlseb2ff5jzqqbyczicxlmsahuxwdpdyp.py
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul_1 => mul_1
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, 1), kwargs = {})
triton_poi_fused_mul_1 = async_compile.triton('triton_poi_fused_mul_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_1, buf0, 16, grid=grid(16), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul_1], Original ATen: [aten.mul]
triton_poi_fused_mul_1.run(primals_2, buf1, 4, grid=grid(4), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1, linear], Original ATen: [aten.mul, aten.addmm]
extern_kernels.addmm(buf1, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf0
del buf1
return (reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
import torch.nn.functional as F
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, lr_mul=1, bias=True):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim))
self.lr_mul = lr_mul
def forward(self, input):
return F.linear(input, self.weight * self.lr_mul, bias=self.bias *
self.lr_mul)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_dim': 4, 'out_dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_mul_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
@triton.jit
def triton_poi_fused_mul_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 4
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 1.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(16)](primals_1, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
triton_poi_fused_mul_1[grid(4)](primals_2, buf1, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(buf1, reinterpret_tensor(primals_3, (64, 4), (
4, 1), 0), reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1,
beta=1, out=buf2)
del buf0
del buf1
return reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0)
class EqualLinearNew(nn.Module):
def __init__(self, in_dim, out_dim, lr_mul=1, bias=True):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim))
self.lr_mul = lr_mul
def forward(self, input_0):
primals_1 = self.weight
primals_2 = self.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| TheSignPainter/AGGAN | EqualLinear | false | 1,133 | [
"Apache-2.0"
] | 0 | d75144f81df3f5a0a761d48c6285c38e74002be3 | https://github.com/TheSignPainter/AGGAN/tree/d75144f81df3f5a0a761d48c6285c38e74002be3 | import torch
from torch import nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, in_dim, out_dim, lr_mul=1, bias=True):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim))
self.lr_mul = lr_mul
def forward(self, input):
return F.linear(input, self.weight * self.lr_mul, bias=self.bias *
self.lr_mul)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
SuperPointNet | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/pn/cpng7gl7lqxvqafyqlu5mbr4lc7m2sgi4l5ulbiv46djlkgyencv.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_0 = async_compile.triton('triton_poi_fused_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[4096, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 4096
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = (yindex // 64)
tmp0 = tl.load(in_ptr0 + (x2 + (9*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (64*x2) + (576*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ne/cnepmjd66uu3laeexeusfxab3aayptiri2wp2knrgtgmx52tvzxj.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_1 = async_compile.triton('triton_poi_fused_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 8192
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = (yindex // 64)
tmp0 = tl.load(in_ptr0 + (x2 + (9*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (64*x2) + (576*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ba/cbayuw2by4w6xwduhs5qdriinmydiep6bpw7fyi37s377up7lrcm.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_2 = async_compile.triton('triton_poi_fused_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[16384, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_2(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16384
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = (yindex // 128)
tmp0 = tl.load(in_ptr0 + (x2 + (9*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (128*x2) + (1152*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/yd/cydvmxsmzwizyj5fbgjnjeeo27as6zdlft5s5uj57ovvcxtlbfhh.py
# Unsorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
triton_poi_fused_3 = async_compile.triton('triton_poi_fused_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768, 16], tile_hint=TileHint.SQUARE,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 32768
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = (yindex // 128)
tmp0 = tl.load(in_ptr0 + (x2 + (9*y3)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (y0 + (128*x2) + (1152*y1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/vn/cvnnodtrripz7gtommdv4wbjjfexefcdjq3t2xglmrxcj2g7mll4.py
# Topologically Sorted Source Nodes: [conv2d, x], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d => convolution
# x => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
triton_poi_fused_convolution_relu_4 = async_compile.triton('triton_poi_fused_convolution_relu_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256, 4096], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_4(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 256
xnumel = 4096
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = (yindex // 64)
tmp0 = tl.load(in_ptr0 + (x2 + (4096*y3)), ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1, 1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(out_ptr0 + (y0 + (64*x2) + (262144*y1)), tmp4, ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/7c/c7caq6stn5xhqphn2xnmwbpvxspyfj5wahntqw4tlpltw5xu6ktg.py
# Topologically Sorted Source Nodes: [conv2d_1, x_1], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d_1 => convolution_1
# x_1 => relu_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu, %primals_4, %primals_5, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_1,), kwargs = {})
triton_poi_fused_convolution_relu_5 = async_compile.triton('triton_poi_fused_convolution_relu_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1048576],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_5(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1048576
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/p7/cp7hkfs7dzspvks5o4gggcw3s4o5jb3vqo372n6r4xcl5tx3xupa.py
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# x_2 => getitem, getitem_1
# Graph fragment:
# %getitem : [num_users=2] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 0), kwargs = {})
# %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets, 1), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_6 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_6', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[262144],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i8', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_6', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_6(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 262144
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 64
x1 = (xindex // 64) % 32
x2 = (xindex // 2048)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (128*x1) + (8192*x2)), None)
tmp1 = tl.load(in_ptr0 + (64 + x0 + (128*x1) + (8192*x2)), None)
tmp3 = tl.load(in_ptr0 + (4096 + x0 + (128*x1) + (8192*x2)), None)
tmp5 = tl.load(in_ptr0 + (4160 + x0 + (128*x1) + (8192*x2)), None)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + (x3), tmp6, None)
tl.store(out_ptr1 + (x3), tmp16, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/ya/cya72dxxug7bvahrgkiz2tev6wxbq4aissg3wd3pl37yen37nb4b.py
# Topologically Sorted Source Nodes: [conv2d_2, x_3], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d_2 => convolution_2
# x_3 => relu_2
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%getitem, %primals_6, %primals_7, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_2,), kwargs = {})
triton_poi_fused_convolution_relu_7 = async_compile.triton('triton_poi_fused_convolution_relu_7', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[262144],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_7', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_7(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 262144
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/xj/cxjptd7j2qxrb3kjd7zlgxmewdvhhkbw3tukgvay2kmhnxcajkzw.py
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# x_5 => getitem_2, getitem_3
# Graph fragment:
# %getitem_2 : [num_users=2] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_1, 0), kwargs = {})
# %getitem_3 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_1, 1), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_8 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_8', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i8', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_8', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_8(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 65536
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 64
x1 = (xindex // 64) % 16
x2 = (xindex // 1024)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (128*x1) + (4096*x2)), None)
tmp1 = tl.load(in_ptr0 + (64 + x0 + (128*x1) + (4096*x2)), None)
tmp3 = tl.load(in_ptr0 + (2048 + x0 + (128*x1) + (4096*x2)), None)
tmp5 = tl.load(in_ptr0 + (2112 + x0 + (128*x1) + (4096*x2)), None)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + (x3), tmp6, None)
tl.store(out_ptr1 + (x3), tmp16, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/v5/cv5bres457ho44iaqr63mi3bbgzezc3pxml5sotwsljao2g5whrl.py
# Topologically Sorted Source Nodes: [conv2d_4, x_6], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d_4 => convolution_4
# x_6 => relu_4
# Graph fragment:
# %convolution_4 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%getitem_2, %primals_10, %primals_11, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_4 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_4,), kwargs = {})
triton_poi_fused_convolution_relu_9 = async_compile.triton('triton_poi_fused_convolution_relu_9', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[131072],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_9', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_9(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 131072
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/sr/csr7z2afgh7gbn3y7lq2wp2sva4b7imt3iniu36uxe33zsilt4x7.py
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.max_pool2d_with_indices]
# Source node to ATen node mapping:
# x_8 => getitem_4, getitem_5
# Graph fragment:
# %getitem_4 : [num_users=2] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_2, 0), kwargs = {})
# %getitem_5 : [num_users=1] = call_function[target=operator.getitem](args = (%_low_memory_max_pool2d_with_offsets_2, 1), kwargs = {})
triton_poi_fused_max_pool2d_with_indices_10 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_10', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i8', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_pool2d_with_indices_10', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 4, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_10(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 32768
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 128
x1 = (xindex // 128) % 8
x2 = (xindex // 1024)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (256*x1) + (4096*x2)), None)
tmp1 = tl.load(in_ptr0 + (128 + x0 + (256*x1) + (4096*x2)), None)
tmp3 = tl.load(in_ptr0 + (2048 + x0 + (256*x1) + (4096*x2)), None)
tmp5 = tl.load(in_ptr0 + (2176 + x0 + (256*x1) + (4096*x2)), None)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + (x3), tmp6, None)
tl.store(out_ptr1 + (x3), tmp16, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/g6/cg6bxfpbjt7yt2cidmobe46sxen6spgw4gul66mxxotjhzxvzddf.py
# Topologically Sorted Source Nodes: [conv2d_6, x_9], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# conv2d_6 => convolution_6
# x_9 => relu_6
# Graph fragment:
# %convolution_6 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%getitem_4, %primals_14, %primals_15, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_6 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_6,), kwargs = {})
triton_poi_fused_convolution_relu_11 = async_compile.triton('triton_poi_fused_convolution_relu_11', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_11', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_11(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32768
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/l5/cl5g2w4v5osla5sy5d2al2y6dspf2ipfcnmfehdgyecqjlqqwxp5.py
# Topologically Sorted Source Nodes: [conv2d_8, cPa], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# cPa => relu_8
# conv2d_8 => convolution_8
# Graph fragment:
# %convolution_8 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_7, %primals_18, %primals_19, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu_8 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution_8,), kwargs = {})
triton_poi_fused_convolution_relu_12 = async_compile.triton('triton_poi_fused_convolution_relu_12', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[65536],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_relu_12', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_relu_12(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 65536
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/q3/cq35pbgnsnzjuhpsen4p6ua7wlqsoqqkc5hvjqqupede7xjns4pl.py
# Topologically Sorted Source Nodes: [semi], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# semi => convolution_9
# Graph fragment:
# %convolution_9 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_8, %primals_20, %primals_21, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_13 = async_compile.triton('triton_poi_fused_convolution_13', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512, 64], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_13', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_13(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 260
xnumel = 64
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 65
y1 = (yindex // 65)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (65*x2) + (4160*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2 + (64*y3)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/le/clesx3fq4nvfilcvyfxgg66sqkyn3nl3mexlek76x5apn2ediyvi.py
# Topologically Sorted Source Nodes: [desc, dn], Original ATen: [aten.convolution, aten.linalg_vector_norm]
# Source node to ATen node mapping:
# desc => convolution_11
# dn => pow_1, pow_2, sum_1
# Graph fragment:
# %convolution_11 : [num_users=3] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_9, %primals_24, %primals_25, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%convolution_11, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [1]), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
triton_per_fused_convolution_linalg_vector_norm_14 = async_compile.triton('triton_per_fused_convolution_linalg_vector_norm_14', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[256, 256],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_convolution_linalg_vector_norm_14', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': True, 'num_load': 2, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_convolution_linalg_vector_norm_14(in_out_ptr0, in_out_ptr1, in_ptr0, xnumel, rnumel):
xnumel = 256
XBLOCK: tl.constexpr = 1
rnumel = 256
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
xmask = tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
roffset = 0
rmask = tl.full([RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (r1 + (256*x0)), None)
tmp1 = tl.load(in_ptr0 + (r1), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tmp2 * tmp2
tmp4 = tl.broadcast_to(tmp3, [RBLOCK])
tmp6 = triton_helpers.promote_to_tensor(tl.sum(tmp4, 0))
tmp7 = libdevice.sqrt(tmp6)
tl.store(in_out_ptr0 + (r1 + (256*x0)), tmp2, None)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x0), tmp7, None)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/f4/cf4ceh57l4kazdvaqbryn4i5xohp6gmuuxrk5bvsv4ct3wlef3om.py
# Topologically Sorted Source Nodes: [desc_1], Original ATen: [aten.div]
# Source node to ATen node mapping:
# desc_1 => div
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%convolution_11, %unsqueeze), kwargs = {})
triton_poi_fused_div_15 = async_compile.triton('triton_poi_fused_div_15', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024, 64], tile_hint=TileHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_15', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_15(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 1024
xnumel = 64
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 256
y1 = (yindex // 256)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (256*x2) + (16384*y1)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2 + (64*y1)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 / tmp1
tl.store(out_ptr0 + (x2 + (64*y3)), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25 = args
args.clear()
assert_size_stride(primals_1, (64, 1, 3, 3), (9, 9, 3, 1))
assert_size_stride(primals_2, (64, ), (1, ))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_4, (64, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_5, (64, ), (1, ))
assert_size_stride(primals_6, (64, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_7, (64, ), (1, ))
assert_size_stride(primals_8, (64, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_9, (64, ), (1, ))
assert_size_stride(primals_10, (128, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_11, (128, ), (1, ))
assert_size_stride(primals_12, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_13, (128, ), (1, ))
assert_size_stride(primals_14, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_15, (128, ), (1, ))
assert_size_stride(primals_16, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_17, (128, ), (1, ))
assert_size_stride(primals_18, (256, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_19, (256, ), (1, ))
assert_size_stride(primals_20, (65, 256, 1, 1), (256, 1, 1, 1))
assert_size_stride(primals_21, (65, ), (1, ))
assert_size_stride(primals_22, (256, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_23, (256, ), (1, ))
assert_size_stride(primals_24, (256, 256, 1, 1), (256, 1, 1, 1))
assert_size_stride(primals_25, (256, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 64, 3, 3), (576, 1, 192, 64), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
stream0 = get_raw_stream(0)
triton_poi_fused_0.run(primals_4, buf0, 4096, 9, grid=grid(4096, 9), stream=stream0)
del primals_4
buf1 = empty_strided_cuda((64, 64, 3, 3), (576, 1, 192, 64), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_0.run(primals_6, buf1, 4096, 9, grid=grid(4096, 9), stream=stream0)
del primals_6
buf2 = empty_strided_cuda((64, 64, 3, 3), (576, 1, 192, 64), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_0.run(primals_8, buf2, 4096, 9, grid=grid(4096, 9), stream=stream0)
del primals_8
buf3 = empty_strided_cuda((128, 64, 3, 3), (576, 1, 192, 64), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_1.run(primals_10, buf3, 8192, 9, grid=grid(8192, 9), stream=stream0)
del primals_10
buf4 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_2.run(primals_12, buf4, 16384, 9, grid=grid(16384, 9), stream=stream0)
del primals_12
buf5 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_2.run(primals_14, buf5, 16384, 9, grid=grid(16384, 9), stream=stream0)
del primals_14
buf6 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_2.run(primals_16, buf6, 16384, 9, grid=grid(16384, 9), stream=stream0)
del primals_16
buf7 = empty_strided_cuda((256, 128, 3, 3), (1152, 1, 384, 128), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_3.run(primals_18, buf7, 32768, 9, grid=grid(32768, 9), stream=stream0)
del primals_18
buf8 = empty_strided_cuda((256, 128, 3, 3), (1152, 1, 384, 128), torch.float32)
# Unsorted Source Nodes: [], Original ATen: []
triton_poi_fused_3.run(primals_22, buf8, 32768, 9, grid=grid(32768, 9), stream=stream0)
del primals_22
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
buf9 = extern_kernels.convolution(primals_3, primals_1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 64, 64, 64), (262144, 4096, 64, 1))
buf10 = empty_strided_cuda((4, 64, 64, 64), (262144, 1, 4096, 64), torch.float32)
# Topologically Sorted Source Nodes: [conv2d, x], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_4.run(buf9, primals_2, buf10, 256, 4096, grid=grid(256, 4096), stream=stream0)
del buf9
del primals_2
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf11 = extern_kernels.convolution(buf10, buf0, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf11, (4, 64, 64, 64), (262144, 1, 4096, 64))
buf12 = buf11; del buf11 # reuse
# Topologically Sorted Source Nodes: [conv2d_1, x_1], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_5.run(buf12, primals_5, 1048576, grid=grid(1048576), stream=stream0)
del primals_5
buf13 = empty_strided_cuda((4, 64, 32, 32), (65536, 1, 2048, 64), torch.float32)
buf14 = empty_strided_cuda((4, 64, 32, 32), (65536, 1, 2048, 64), torch.int8)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.max_pool2d_with_indices]
triton_poi_fused_max_pool2d_with_indices_6.run(buf12, buf13, buf14, 262144, grid=grid(262144), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_2], Original ATen: [aten.convolution]
buf15 = extern_kernels.convolution(buf13, buf1, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf15, (4, 64, 32, 32), (65536, 1, 2048, 64))
buf16 = buf15; del buf15 # reuse
# Topologically Sorted Source Nodes: [conv2d_2, x_3], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_7.run(buf16, primals_7, 262144, grid=grid(262144), stream=stream0)
del primals_7
# Topologically Sorted Source Nodes: [conv2d_3], Original ATen: [aten.convolution]
buf17 = extern_kernels.convolution(buf16, buf2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf17, (4, 64, 32, 32), (65536, 1, 2048, 64))
buf18 = buf17; del buf17 # reuse
# Topologically Sorted Source Nodes: [conv2d_3, x_4], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_7.run(buf18, primals_9, 262144, grid=grid(262144), stream=stream0)
del primals_9
buf19 = empty_strided_cuda((4, 64, 16, 16), (16384, 1, 1024, 64), torch.float32)
buf20 = empty_strided_cuda((4, 64, 16, 16), (16384, 1, 1024, 64), torch.int8)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.max_pool2d_with_indices]
triton_poi_fused_max_pool2d_with_indices_8.run(buf18, buf19, buf20, 65536, grid=grid(65536), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_4], Original ATen: [aten.convolution]
buf21 = extern_kernels.convolution(buf19, buf3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf21, (4, 128, 16, 16), (32768, 1, 2048, 128))
buf22 = buf21; del buf21 # reuse
# Topologically Sorted Source Nodes: [conv2d_4, x_6], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_9.run(buf22, primals_11, 131072, grid=grid(131072), stream=stream0)
del primals_11
# Topologically Sorted Source Nodes: [conv2d_5], Original ATen: [aten.convolution]
buf23 = extern_kernels.convolution(buf22, buf4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf23, (4, 128, 16, 16), (32768, 1, 2048, 128))
buf24 = buf23; del buf23 # reuse
# Topologically Sorted Source Nodes: [conv2d_5, x_7], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_9.run(buf24, primals_13, 131072, grid=grid(131072), stream=stream0)
del primals_13
buf25 = empty_strided_cuda((4, 128, 8, 8), (8192, 1, 1024, 128), torch.float32)
buf26 = empty_strided_cuda((4, 128, 8, 8), (8192, 1, 1024, 128), torch.int8)
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.max_pool2d_with_indices]
triton_poi_fused_max_pool2d_with_indices_10.run(buf24, buf25, buf26, 32768, grid=grid(32768), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_6], Original ATen: [aten.convolution]
buf27 = extern_kernels.convolution(buf25, buf5, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf27, (4, 128, 8, 8), (8192, 1, 1024, 128))
buf28 = buf27; del buf27 # reuse
# Topologically Sorted Source Nodes: [conv2d_6, x_9], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_11.run(buf28, primals_15, 32768, grid=grid(32768), stream=stream0)
del primals_15
# Topologically Sorted Source Nodes: [conv2d_7], Original ATen: [aten.convolution]
buf29 = extern_kernels.convolution(buf28, buf6, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf29, (4, 128, 8, 8), (8192, 1, 1024, 128))
buf30 = buf29; del buf29 # reuse
# Topologically Sorted Source Nodes: [conv2d_7, x_10], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_11.run(buf30, primals_17, 32768, grid=grid(32768), stream=stream0)
del primals_17
# Topologically Sorted Source Nodes: [conv2d_8], Original ATen: [aten.convolution]
buf31 = extern_kernels.convolution(buf30, buf7, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf31, (4, 256, 8, 8), (16384, 1, 2048, 256))
buf32 = buf31; del buf31 # reuse
# Topologically Sorted Source Nodes: [conv2d_8, cPa], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_12.run(buf32, primals_19, 65536, grid=grid(65536), stream=stream0)
del primals_19
# Topologically Sorted Source Nodes: [semi], Original ATen: [aten.convolution]
buf33 = extern_kernels.convolution(buf32, primals_20, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf33, (4, 65, 8, 8), (4160, 1, 520, 65))
buf34 = empty_strided_cuda((4, 65, 8, 8), (4160, 64, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [semi], Original ATen: [aten.convolution]
triton_poi_fused_convolution_13.run(buf33, primals_21, buf34, 260, 64, grid=grid(260, 64), stream=stream0)
del buf33
del primals_21
# Topologically Sorted Source Nodes: [conv2d_10], Original ATen: [aten.convolution]
buf35 = extern_kernels.convolution(buf30, buf8, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf35, (4, 256, 8, 8), (16384, 1, 2048, 256))
buf36 = buf35; del buf35 # reuse
# Topologically Sorted Source Nodes: [conv2d_10, cDa], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_12.run(buf36, primals_23, 65536, grid=grid(65536), stream=stream0)
del primals_23
# Topologically Sorted Source Nodes: [desc], Original ATen: [aten.convolution]
buf37 = extern_kernels.convolution(buf36, primals_24, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf37, (4, 256, 8, 8), (16384, 1, 2048, 256))
buf38 = buf37; del buf37 # reuse
buf39 = empty_strided_cuda((4, 8, 8), (64, 8, 1), torch.float32)
buf40 = buf39; del buf39 # reuse
# Topologically Sorted Source Nodes: [desc, dn], Original ATen: [aten.convolution, aten.linalg_vector_norm]
triton_per_fused_convolution_linalg_vector_norm_14.run(buf38, buf40, primals_25, 256, 256, grid=grid(256), stream=stream0)
del primals_25
buf41 = empty_strided_cuda((4, 256, 8, 8), (16384, 64, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [desc_1], Original ATen: [aten.div]
triton_poi_fused_div_15.run(buf38, buf40, buf41, 1024, 64, grid=grid(1024, 64), stream=stream0)
return (buf34, buf41, primals_1, primals_3, buf0, buf1, buf2, buf3, buf4, buf5, buf6, buf7, primals_20, buf8, primals_24, buf10, buf12, buf13, buf14, buf16, buf18, buf19, buf20, buf22, buf24, buf25, buf26, buf28, buf30, buf32, buf36, buf38, reinterpret_tensor(buf40, (4, 1, 8, 8), (64, 64, 8, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((64, 1, 3, 3), (9, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 1, 64, 64), (4096, 4096, 64, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((64, 64, 3, 3), (576, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((64, 64, 3, 3), (576, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((64, 64, 3, 3), (576, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((128, 64, 3, 3), (576, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((128, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((128, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_15 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_16 = rand_strided((128, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_17 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_18 = rand_strided((256, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_19 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_20 = rand_strided((65, 256, 1, 1), (256, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_21 = rand_strided((65, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_22 = rand_strided((256, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_23 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_24 = rand_strided((256, 256, 1, 1), (256, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_25 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14, primals_15, primals_16, primals_17, primals_18, primals_19, primals_20, primals_21, primals_22, primals_23, primals_24, primals_25])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.optim
import torch.utils.data
class SuperPointNet(torch.nn.Module):
""" Pytorch definition of SuperPoint Network. """
def __init__(self):
super(SuperPointNet, self).__init__()
self.relu = torch.nn.ReLU(inplace=True)
self.pool = torch.nn.MaxPool2d(kernel_size=2, stride=2)
c1, c2, c3, c4, c5, d1 = 64, 64, 128, 128, 256, 256
self.conv1a = torch.nn.Conv2d(1, c1, kernel_size=3, stride=1, padding=1
)
self.conv1b = torch.nn.Conv2d(c1, c1, kernel_size=3, stride=1,
padding=1)
self.conv2a = torch.nn.Conv2d(c1, c2, kernel_size=3, stride=1,
padding=1)
self.conv2b = torch.nn.Conv2d(c2, c2, kernel_size=3, stride=1,
padding=1)
self.conv3a = torch.nn.Conv2d(c2, c3, kernel_size=3, stride=1,
padding=1)
self.conv3b = torch.nn.Conv2d(c3, c3, kernel_size=3, stride=1,
padding=1)
self.conv4a = torch.nn.Conv2d(c3, c4, kernel_size=3, stride=1,
padding=1)
self.conv4b = torch.nn.Conv2d(c4, c4, kernel_size=3, stride=1,
padding=1)
self.convPa = torch.nn.Conv2d(c4, c5, kernel_size=3, stride=1,
padding=1)
self.convPb = torch.nn.Conv2d(c5, 65, kernel_size=1, stride=1,
padding=0)
self.convDa = torch.nn.Conv2d(c4, c5, kernel_size=3, stride=1,
padding=1)
self.convDb = torch.nn.Conv2d(c5, d1, kernel_size=1, stride=1,
padding=0)
def forward(self, x):
""" Forward pass that jointly computes unprocessed point and descriptor
tensors.
Input
x: Image pytorch tensor shaped N x 1 x H x W.
Output
semi: Output point pytorch tensor shaped N x 65 x H/8 x W/8.
desc: Output descriptor pytorch tensor shaped N x 256 x H/8 x W/8.
"""
x = self.relu(self.conv1a(x))
x = self.relu(self.conv1b(x))
x = self.pool(x)
x = self.relu(self.conv2a(x))
x = self.relu(self.conv2b(x))
x = self.pool(x)
x = self.relu(self.conv3a(x))
x = self.relu(self.conv3b(x))
x = self.pool(x)
x = self.relu(self.conv4a(x))
x = self.relu(self.conv4b(x))
cPa = self.relu(self.convPa(x))
semi = self.convPb(cPa)
cDa = self.relu(self.convDa(x))
desc = self.convDb(cDa)
dn = torch.norm(desc, p=2, dim=1)
desc = desc.div(torch.unsqueeze(dn, 1))
return semi, desc
def get_inputs():
return [torch.rand([4, 1, 64, 64])]
def get_init_inputs():
return [[], {}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.optim
import torch.utils.data
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = yindex // 64
tmp0 = tl.load(in_ptr0 + (x2 + 9 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 64 * x2 + 576 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_1(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = yindex // 64
tmp0 = tl.load(in_ptr0 + (x2 + 9 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 64 * x2 + 576 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_2(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = yindex // 128
tmp0 = tl.load(in_ptr0 + (x2 + 9 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 128 * x2 + 1152 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_3(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
xnumel = 9
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y3 = yindex
y0 = yindex % 128
y1 = yindex // 128
tmp0 = tl.load(in_ptr0 + (x2 + 9 * y3), xmask, eviction_policy='evict_last'
)
tl.store(out_ptr0 + (y0 + 128 * x2 + 1152 * y1), tmp0, xmask)
@triton.jit
def triton_poi_fused_convolution_relu_4(in_ptr0, in_ptr1, out_ptr0, ynumel,
xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 256
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
x2 = xindex
y3 = yindex
y0 = yindex % 64
y1 = yindex // 64
tmp0 = tl.load(in_ptr0 + (x2 + 4096 * y3), ymask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1, 1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(out_ptr0 + (y0 + 64 * x2 + 262144 * y1), tmp4, ymask)
@triton.jit
def triton_poi_fused_convolution_relu_5(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_6(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 64
x1 = xindex // 64 % 32
x2 = xindex // 2048
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 128 * x1 + 8192 * x2), None)
tmp1 = tl.load(in_ptr0 + (64 + x0 + 128 * x1 + 8192 * x2), None)
tmp3 = tl.load(in_ptr0 + (4096 + x0 + 128 * x1 + 8192 * x2), None)
tmp5 = tl.load(in_ptr0 + (4160 + x0 + 128 * x1 + 8192 * x2), None)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + x3, tmp6, None)
tl.store(out_ptr1 + x3, tmp16, None)
@triton.jit
def triton_poi_fused_convolution_relu_7(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_8(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 64
x1 = xindex // 64 % 16
x2 = xindex // 1024
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 128 * x1 + 4096 * x2), None)
tmp1 = tl.load(in_ptr0 + (64 + x0 + 128 * x1 + 4096 * x2), None)
tmp3 = tl.load(in_ptr0 + (2048 + x0 + 128 * x1 + 4096 * x2), None)
tmp5 = tl.load(in_ptr0 + (2112 + x0 + 128 * x1 + 4096 * x2), None)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + x3, tmp6, None)
tl.store(out_ptr1 + x3, tmp16, None)
@triton.jit
def triton_poi_fused_convolution_relu_9(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
@triton.jit
def triton_poi_fused_max_pool2d_with_indices_10(in_ptr0, out_ptr0, out_ptr1,
xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 128
x1 = xindex // 128 % 8
x2 = xindex // 1024
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 256 * x1 + 4096 * x2), None)
tmp1 = tl.load(in_ptr0 + (128 + x0 + 256 * x1 + 4096 * x2), None)
tmp3 = tl.load(in_ptr0 + (2048 + x0 + 256 * x1 + 4096 * x2), None)
tmp5 = tl.load(in_ptr0 + (2176 + x0 + 256 * x1 + 4096 * x2), None)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = tmp1 > tmp0
tmp8 = tl.full([1], 1, tl.int8)
tmp9 = tl.full([1], 0, tl.int8)
tmp10 = tl.where(tmp7, tmp8, tmp9)
tmp11 = tmp3 > tmp2
tmp12 = tl.full([1], 2, tl.int8)
tmp13 = tl.where(tmp11, tmp12, tmp10)
tmp14 = tmp5 > tmp4
tmp15 = tl.full([1], 3, tl.int8)
tmp16 = tl.where(tmp14, tmp15, tmp13)
tl.store(out_ptr0 + x3, tmp6, None)
tl.store(out_ptr1 + x3, tmp16, None)
@triton.jit
def triton_poi_fused_convolution_relu_11(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
@triton.jit
def triton_poi_fused_convolution_relu_12(in_out_ptr0, in_ptr0, xnumel,
XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(in_out_ptr0 + x2, tmp4, None)
@triton.jit
def triton_poi_fused_convolution_13(in_ptr0, in_ptr1, out_ptr0, ynumel,
xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 260
xnumel = 64
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
ymask = yindex < ynumel
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 65
y1 = yindex // 65
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 65 * x2 + 4160 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(out_ptr0 + (x2 + 64 * y3), tmp2, xmask & ymask)
@triton.jit
def triton_per_fused_convolution_linalg_vector_norm_14(in_out_ptr0,
in_out_ptr1, in_ptr0, xnumel, rnumel):
XBLOCK: tl.constexpr = 1
RBLOCK: tl.constexpr = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = tl.full([1], xoffset, tl.int32)
tl.full([RBLOCK], True, tl.int1)
rindex = tl.arange(0, RBLOCK)[:]
tl.full([RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_out_ptr0 + (r1 + 256 * x0), None)
tmp1 = tl.load(in_ptr0 + r1, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tmp2 * tmp2
tmp4 = tl.broadcast_to(tmp3, [RBLOCK])
tmp6 = triton_helpers.promote_to_tensor(tl.sum(tmp4, 0))
tmp7 = libdevice.sqrt(tmp6)
tl.store(in_out_ptr0 + (r1 + 256 * x0), tmp2, None)
tl.debug_barrier()
tl.store(in_out_ptr1 + x0, tmp7, None)
@triton.jit
def triton_poi_fused_div_15(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
xnumel = 64
yoffset = tl.program_id(1) * YBLOCK
yindex = yoffset + tl.arange(0, YBLOCK)[None, :]
tl.full([XBLOCK, YBLOCK], True, tl.int1)
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
x2 = xindex
y0 = yindex % 256
y1 = yindex // 256
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 256 * x2 + 16384 * y1), xmask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2 + 64 * y1), xmask, eviction_policy=
'evict_last')
tmp2 = tmp0 / tmp1
tl.store(out_ptr0 + (x2 + 64 * y3), tmp2, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14, primals_15, primals_16, primals_17,
primals_18, primals_19, primals_20, primals_21, primals_22,
primals_23, primals_24, primals_25) = args
args.clear()
assert_size_stride(primals_1, (64, 1, 3, 3), (9, 9, 3, 1))
assert_size_stride(primals_2, (64,), (1,))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_4, (64, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_5, (64,), (1,))
assert_size_stride(primals_6, (64, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_7, (64,), (1,))
assert_size_stride(primals_8, (64, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_9, (64,), (1,))
assert_size_stride(primals_10, (128, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_11, (128,), (1,))
assert_size_stride(primals_12, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_13, (128,), (1,))
assert_size_stride(primals_14, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_15, (128,), (1,))
assert_size_stride(primals_16, (128, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_17, (128,), (1,))
assert_size_stride(primals_18, (256, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_19, (256,), (1,))
assert_size_stride(primals_20, (65, 256, 1, 1), (256, 1, 1, 1))
assert_size_stride(primals_21, (65,), (1,))
assert_size_stride(primals_22, (256, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_23, (256,), (1,))
assert_size_stride(primals_24, (256, 256, 1, 1), (256, 1, 1, 1))
assert_size_stride(primals_25, (256,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 64, 3, 3), (576, 1, 192, 64), torch.
float32)
get_raw_stream(0)
triton_poi_fused_0[grid(4096, 9)](primals_4, buf0, 4096, 9, XBLOCK=
16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_4
buf1 = empty_strided_cuda((64, 64, 3, 3), (576, 1, 192, 64), torch.
float32)
triton_poi_fused_0[grid(4096, 9)](primals_6, buf1, 4096, 9, XBLOCK=
16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_6
buf2 = empty_strided_cuda((64, 64, 3, 3), (576, 1, 192, 64), torch.
float32)
triton_poi_fused_0[grid(4096, 9)](primals_8, buf2, 4096, 9, XBLOCK=
16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_8
buf3 = empty_strided_cuda((128, 64, 3, 3), (576, 1, 192, 64), torch
.float32)
triton_poi_fused_1[grid(8192, 9)](primals_10, buf3, 8192, 9, XBLOCK
=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_10
buf4 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128),
torch.float32)
triton_poi_fused_2[grid(16384, 9)](primals_12, buf4, 16384, 9,
XBLOCK=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_12
buf5 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128),
torch.float32)
triton_poi_fused_2[grid(16384, 9)](primals_14, buf5, 16384, 9,
XBLOCK=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_14
buf6 = empty_strided_cuda((128, 128, 3, 3), (1152, 1, 384, 128),
torch.float32)
triton_poi_fused_2[grid(16384, 9)](primals_16, buf6, 16384, 9,
XBLOCK=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_16
buf7 = empty_strided_cuda((256, 128, 3, 3), (1152, 1, 384, 128),
torch.float32)
triton_poi_fused_3[grid(32768, 9)](primals_18, buf7, 32768, 9,
XBLOCK=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_18
buf8 = empty_strided_cuda((256, 128, 3, 3), (1152, 1, 384, 128),
torch.float32)
triton_poi_fused_3[grid(32768, 9)](primals_22, buf8, 32768, 9,
XBLOCK=16, YBLOCK=64, num_warps=4, num_stages=1)
del primals_22
buf9 = extern_kernels.convolution(primals_3, primals_1, stride=(1,
1), padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf9, (4, 64, 64, 64), (262144, 4096, 64, 1))
buf10 = empty_strided_cuda((4, 64, 64, 64), (262144, 1, 4096, 64),
torch.float32)
triton_poi_fused_convolution_relu_4[grid(256, 4096)](buf9,
primals_2, buf10, 256, 4096, XBLOCK=32, YBLOCK=32, num_warps=4,
num_stages=1)
del buf9
del primals_2
buf11 = extern_kernels.convolution(buf10, buf0, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf11, (4, 64, 64, 64), (262144, 1, 4096, 64))
buf12 = buf11
del buf11
triton_poi_fused_convolution_relu_5[grid(1048576)](buf12, primals_5,
1048576, XBLOCK=1024, num_warps=4, num_stages=1)
del primals_5
buf13 = empty_strided_cuda((4, 64, 32, 32), (65536, 1, 2048, 64),
torch.float32)
buf14 = empty_strided_cuda((4, 64, 32, 32), (65536, 1, 2048, 64),
torch.int8)
triton_poi_fused_max_pool2d_with_indices_6[grid(262144)](buf12,
buf13, buf14, 262144, XBLOCK=1024, num_warps=4, num_stages=1)
buf15 = extern_kernels.convolution(buf13, buf1, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf15, (4, 64, 32, 32), (65536, 1, 2048, 64))
buf16 = buf15
del buf15
triton_poi_fused_convolution_relu_7[grid(262144)](buf16, primals_7,
262144, XBLOCK=512, num_warps=8, num_stages=1)
del primals_7
buf17 = extern_kernels.convolution(buf16, buf2, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf17, (4, 64, 32, 32), (65536, 1, 2048, 64))
buf18 = buf17
del buf17
triton_poi_fused_convolution_relu_7[grid(262144)](buf18, primals_9,
262144, XBLOCK=512, num_warps=8, num_stages=1)
del primals_9
buf19 = empty_strided_cuda((4, 64, 16, 16), (16384, 1, 1024, 64),
torch.float32)
buf20 = empty_strided_cuda((4, 64, 16, 16), (16384, 1, 1024, 64),
torch.int8)
triton_poi_fused_max_pool2d_with_indices_8[grid(65536)](buf18,
buf19, buf20, 65536, XBLOCK=256, num_warps=4, num_stages=1)
buf21 = extern_kernels.convolution(buf19, buf3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf21, (4, 128, 16, 16), (32768, 1, 2048, 128))
buf22 = buf21
del buf21
triton_poi_fused_convolution_relu_9[grid(131072)](buf22, primals_11,
131072, XBLOCK=512, num_warps=8, num_stages=1)
del primals_11
buf23 = extern_kernels.convolution(buf22, buf4, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf23, (4, 128, 16, 16), (32768, 1, 2048, 128))
buf24 = buf23
del buf23
triton_poi_fused_convolution_relu_9[grid(131072)](buf24, primals_13,
131072, XBLOCK=512, num_warps=8, num_stages=1)
del primals_13
buf25 = empty_strided_cuda((4, 128, 8, 8), (8192, 1, 1024, 128),
torch.float32)
buf26 = empty_strided_cuda((4, 128, 8, 8), (8192, 1, 1024, 128),
torch.int8)
triton_poi_fused_max_pool2d_with_indices_10[grid(32768)](buf24,
buf25, buf26, 32768, XBLOCK=128, num_warps=4, num_stages=1)
buf27 = extern_kernels.convolution(buf25, buf5, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf27, (4, 128, 8, 8), (8192, 1, 1024, 128))
buf28 = buf27
del buf27
triton_poi_fused_convolution_relu_11[grid(32768)](buf28, primals_15,
32768, XBLOCK=256, num_warps=4, num_stages=1)
del primals_15
buf29 = extern_kernels.convolution(buf28, buf6, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf29, (4, 128, 8, 8), (8192, 1, 1024, 128))
buf30 = buf29
del buf29
triton_poi_fused_convolution_relu_11[grid(32768)](buf30, primals_17,
32768, XBLOCK=256, num_warps=4, num_stages=1)
del primals_17
buf31 = extern_kernels.convolution(buf30, buf7, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf31, (4, 256, 8, 8), (16384, 1, 2048, 256))
buf32 = buf31
del buf31
triton_poi_fused_convolution_relu_12[grid(65536)](buf32, primals_19,
65536, XBLOCK=512, num_warps=4, num_stages=1)
del primals_19
buf33 = extern_kernels.convolution(buf32, primals_20, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf33, (4, 65, 8, 8), (4160, 1, 520, 65))
buf34 = empty_strided_cuda((4, 65, 8, 8), (4160, 64, 8, 1), torch.
float32)
triton_poi_fused_convolution_13[grid(260, 64)](buf33, primals_21,
buf34, 260, 64, XBLOCK=64, YBLOCK=4, num_warps=4, num_stages=1)
del buf33
del primals_21
buf35 = extern_kernels.convolution(buf30, buf8, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf35, (4, 256, 8, 8), (16384, 1, 2048, 256))
buf36 = buf35
del buf35
triton_poi_fused_convolution_relu_12[grid(65536)](buf36, primals_23,
65536, XBLOCK=512, num_warps=4, num_stages=1)
del primals_23
buf37 = extern_kernels.convolution(buf36, primals_24, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf37, (4, 256, 8, 8), (16384, 1, 2048, 256))
buf38 = buf37
del buf37
buf39 = empty_strided_cuda((4, 8, 8), (64, 8, 1), torch.float32)
buf40 = buf39
del buf39
triton_per_fused_convolution_linalg_vector_norm_14[grid(256)](buf38,
buf40, primals_25, 256, 256, num_warps=2, num_stages=1)
del primals_25
buf41 = empty_strided_cuda((4, 256, 8, 8), (16384, 64, 8, 1), torch
.float32)
triton_poi_fused_div_15[grid(1024, 64)](buf38, buf40, buf41, 1024,
64, XBLOCK=32, YBLOCK=32, num_warps=4, num_stages=1)
return (buf34, buf41, primals_1, primals_3, buf0, buf1, buf2, buf3,
buf4, buf5, buf6, buf7, primals_20, buf8, primals_24, buf10, buf12,
buf13, buf14, buf16, buf18, buf19, buf20, buf22, buf24, buf25,
buf26, buf28, buf30, buf32, buf36, buf38, reinterpret_tensor(buf40,
(4, 1, 8, 8), (64, 64, 8, 1), 0))
class SuperPointNetNew(torch.nn.Module):
""" Pytorch definition of SuperPoint Network. """
def __init__(self):
super(SuperPointNetNew, self).__init__()
self.relu = torch.nn.ReLU(inplace=True)
self.pool = torch.nn.MaxPool2d(kernel_size=2, stride=2)
c1, c2, c3, c4, c5, d1 = 64, 64, 128, 128, 256, 256
self.conv1a = torch.nn.Conv2d(1, c1, kernel_size=3, stride=1, padding=1
)
self.conv1b = torch.nn.Conv2d(c1, c1, kernel_size=3, stride=1,
padding=1)
self.conv2a = torch.nn.Conv2d(c1, c2, kernel_size=3, stride=1,
padding=1)
self.conv2b = torch.nn.Conv2d(c2, c2, kernel_size=3, stride=1,
padding=1)
self.conv3a = torch.nn.Conv2d(c2, c3, kernel_size=3, stride=1,
padding=1)
self.conv3b = torch.nn.Conv2d(c3, c3, kernel_size=3, stride=1,
padding=1)
self.conv4a = torch.nn.Conv2d(c3, c4, kernel_size=3, stride=1,
padding=1)
self.conv4b = torch.nn.Conv2d(c4, c4, kernel_size=3, stride=1,
padding=1)
self.convPa = torch.nn.Conv2d(c4, c5, kernel_size=3, stride=1,
padding=1)
self.convPb = torch.nn.Conv2d(c5, 65, kernel_size=1, stride=1,
padding=0)
self.convDa = torch.nn.Conv2d(c4, c5, kernel_size=3, stride=1,
padding=1)
self.convDb = torch.nn.Conv2d(c5, d1, kernel_size=1, stride=1,
padding=0)
def forward(self, input_0):
primals_1 = self.conv1a.weight
primals_2 = self.conv1a.bias
primals_4 = self.conv1b.weight
primals_5 = self.conv1b.bias
primals_6 = self.conv2a.weight
primals_7 = self.conv2a.bias
primals_8 = self.conv2b.weight
primals_9 = self.conv2b.bias
primals_10 = self.conv3a.weight
primals_11 = self.conv3a.bias
primals_12 = self.conv3b.weight
primals_13 = self.conv3b.bias
primals_14 = self.conv4a.weight
primals_15 = self.conv4a.bias
primals_16 = self.conv4b.weight
primals_17 = self.conv4b.bias
primals_18 = self.convPa.weight
primals_19 = self.convPa.bias
primals_20 = self.convPb.weight
primals_21 = self.convPb.bias
primals_22 = self.convDa.weight
primals_23 = self.convDa.bias
primals_24 = self.convDb.weight
primals_25 = self.convDb.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14,
primals_15, primals_16, primals_17, primals_18, primals_19,
primals_20, primals_21, primals_22, primals_23, primals_24,
primals_25])
return output[0], output[1]
| Sunny-Qin-0314/pytorch-superpoint | SuperPointNet | false | 1,134 | [
"MIT"
] | 0 | 5c5325a1e5917afcc7469e137206990a8cd33725 | https://github.com/Sunny-Qin-0314/pytorch-superpoint/tree/5c5325a1e5917afcc7469e137206990a8cd33725 | import torch
import torch.optim
import torch.utils.data
class Model(torch.nn.Module):
""" Pytorch definition of SuperPoint Network. """
def __init__(self):
super().__init__()
self.relu = torch.nn.ReLU(inplace=True)
self.pool = torch.nn.MaxPool2d(kernel_size=2, stride=2)
c1, c2, c3, c4, c5, d1 = 64, 64, 128, 128, 256, 256
self.conv1a = torch.nn.Conv2d(1, c1, kernel_size=3, stride=1, padding=1
)
self.conv1b = torch.nn.Conv2d(c1, c1, kernel_size=3, stride=1,
padding=1)
self.conv2a = torch.nn.Conv2d(c1, c2, kernel_size=3, stride=1,
padding=1)
self.conv2b = torch.nn.Conv2d(c2, c2, kernel_size=3, stride=1,
padding=1)
self.conv3a = torch.nn.Conv2d(c2, c3, kernel_size=3, stride=1,
padding=1)
self.conv3b = torch.nn.Conv2d(c3, c3, kernel_size=3, stride=1,
padding=1)
self.conv4a = torch.nn.Conv2d(c3, c4, kernel_size=3, stride=1,
padding=1)
self.conv4b = torch.nn.Conv2d(c4, c4, kernel_size=3, stride=1,
padding=1)
self.convPa = torch.nn.Conv2d(c4, c5, kernel_size=3, stride=1,
padding=1)
self.convPb = torch.nn.Conv2d(c5, 65, kernel_size=1, stride=1,
padding=0)
self.convDa = torch.nn.Conv2d(c4, c5, kernel_size=3, stride=1,
padding=1)
self.convDb = torch.nn.Conv2d(c5, d1, kernel_size=1, stride=1,
padding=0)
def forward(self, x):
""" Forward pass that jointly computes unprocessed point and descriptor
tensors.
Input
x: Image pytorch tensor shaped N x 1 x H x W.
Output
semi: Output point pytorch tensor shaped N x 65 x H/8 x W/8.
desc: Output descriptor pytorch tensor shaped N x 256 x H/8 x W/8.
"""
x = self.relu(self.conv1a(x))
x = self.relu(self.conv1b(x))
x = self.pool(x)
x = self.relu(self.conv2a(x))
x = self.relu(self.conv2b(x))
x = self.pool(x)
x = self.relu(self.conv3a(x))
x = self.relu(self.conv3b(x))
x = self.pool(x)
x = self.relu(self.conv4a(x))
x = self.relu(self.conv4b(x))
cPa = self.relu(self.convPa(x))
semi = self.convPb(cPa)
cDa = self.relu(self.convDa(x))
desc = self.convDb(cDa)
dn = torch.norm(desc, p=2, dim=1)
desc = desc.div(torch.unsqueeze(dn, 1))
return semi, desc
def get_inputs():
return [torch.rand([4, 1, 64, 64])]
def get_init_inputs():
return []
|
ArcMarginProduct_subcenter | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/fh/cfhnguw4v6uy4ysjg54ojclakwi3bj2lte6oqizl4rpf4lcxpiyp.py
# Topologically Sorted Source Nodes: [normalize], Original ATen: [aten.div]
# Source node to ATen node mapping:
# normalize => div
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_1, %expand), kwargs = {})
triton_poi_fused_div_0 = async_compile.triton('triton_poi_fused_div_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = (xindex // 64)
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + (64*x2)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-12
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x3), tmp15, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/2s/c2s7vvaz25xjuqqa7ymam5rco6ey7ycspmykfxm67txox57kysz2.py
# Topologically Sorted Source Nodes: [normalize_1], Original ATen: [aten.div]
# Source node to ATen node mapping:
# normalize_1 => div_1
# Graph fragment:
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_2, %expand_1), kwargs = {})
triton_poi_fused_div_1 = async_compile.triton('triton_poi_fused_div_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_div_1', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 5, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_div_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 48
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-12
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x2), tmp15, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/o7/co7bzw6oskv7grmzjusyhgf3movzeocopdg3tnslwlwhxsiewsyw.py
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
# Source node to ATen node mapping:
# max_1 => getitem, max_1
# Graph fragment:
# %max_1 : [num_users=2] = call_function[target=torch.ops.aten.max.dim](args = (%view_2, 2), kwargs = {})
# %getitem : [num_users=1] = call_function[target=operator.getitem](args = (%max_1, 0), kwargs = {})
triton_poi_fused_max_2 = async_compile.triton('triton_poi_fused_max_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i64', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_max_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_max_2(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (3*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (3*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (3*x0)), xmask, eviction_policy='evict_last')
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = tmp0 > tmp1
tmp6 = tmp0 == tmp1
tmp7 = tmp0 != tmp0
tmp8 = tmp1 != tmp1
tmp9 = tmp7 > tmp8
tmp10 = tmp5 | tmp9
tmp11 = tmp7 & tmp8
tmp12 = tmp6 | tmp11
tmp13 = tl.full([1], 0, tl.int64)
tmp14 = tl.full([1], 1, tl.int64)
tmp15 = tmp13 < tmp14
tmp16 = tmp12 & tmp15
tmp17 = tmp10 | tmp16
tmp18 = tl.where(tmp17, tmp0, tmp1)
tmp19 = tl.where(tmp17, tmp13, tmp14)
tmp20 = tmp18 > tmp3
tmp21 = tmp18 == tmp3
tmp22 = tmp18 != tmp18
tmp23 = tmp3 != tmp3
tmp24 = tmp22 > tmp23
tmp25 = tmp20 | tmp24
tmp26 = tmp22 & tmp23
tmp27 = tmp21 | tmp26
tmp28 = tl.full([1], 2, tl.int64)
tmp29 = tmp19 < tmp28
tmp30 = tmp27 & tmp29
tmp31 = tmp25 | tmp30
tmp32 = tl.where(tmp31, tmp18, tmp3)
tmp33 = tl.where(tmp31, tmp19, tmp28)
tl.store(out_ptr0 + (x0), tmp4, xmask)
tl.store(out_ptr1 + (x0), tmp33, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (12, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [normalize], Original ATen: [aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_div_0.run(primals_1, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
buf1 = empty_strided_cuda((12, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [normalize_1], Original ATen: [aten.div]
triton_poi_fused_div_1.run(primals_2, buf1, 48, grid=grid(48), stream=stream0)
buf2 = empty_strided_cuda((64, 12), (12, 1), torch.float32)
# Topologically Sorted Source Nodes: [cosine_all], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf0, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (4, 12), (1, 4), 0), out=buf2)
del buf1
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.int64)
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
triton_poi_fused_max_2.run(buf2, buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf2
return (buf3, primals_2, reinterpret_tensor(buf0, (64, 4), (4, 1), 0), reinterpret_tensor(buf4, (64, 4, 1), (4, 1, 1), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((12, 4), (4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import math
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.data
import torch.nn.parallel
class ArcMarginProduct_subcenter(nn.Module):
def __init__(self, in_features, out_features, k=3):
super().__init__()
self.weight = nn.Parameter(torch.FloatTensor(out_features * k,
in_features))
self.reset_parameters()
self.k = k
self.out_features = out_features
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.weight.size(1))
self.weight.data.uniform_(-stdv, stdv)
def forward(self, features):
cosine_all = F.linear(F.normalize(features), F.normalize(self.weight))
cosine_all = cosine_all.view(-1, self.out_features, self.k)
cosine, _ = torch.max(cosine_all, dim=2)
return cosine
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_features': 4, 'out_features': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import math
import torch.nn as nn
import torch.utils.data
import torch.nn.parallel
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_div_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x0 = xindex % 16
x2 = xindex // 64
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + (x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (32 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp9 = tl.load(in_ptr0 + (48 + x0 + 64 * x2), xmask, eviction_policy=
'evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-12
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + x3, tmp15, xmask)
@triton.jit
def triton_poi_fused_div_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 48
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp2 = tmp1 * tmp1
tmp4 = tmp3 * tmp3
tmp5 = tmp2 + tmp4
tmp7 = tmp6 * tmp6
tmp8 = tmp5 + tmp7
tmp10 = tmp9 * tmp9
tmp11 = tmp8 + tmp10
tmp12 = libdevice.sqrt(tmp11)
tmp13 = 1e-12
tmp14 = triton_helpers.maximum(tmp12, tmp13)
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + x2, tmp15, xmask)
@triton.jit
def triton_poi_fused_max_2(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK: tl.
constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + 3 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 3 * x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + 3 * x0), xmask, eviction_policy='evict_last')
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = tmp0 > tmp1
tmp6 = tmp0 == tmp1
tmp7 = tmp0 != tmp0
tmp8 = tmp1 != tmp1
tmp9 = tmp7 > tmp8
tmp10 = tmp5 | tmp9
tmp11 = tmp7 & tmp8
tmp12 = tmp6 | tmp11
tmp13 = tl.full([1], 0, tl.int64)
tmp14 = tl.full([1], 1, tl.int64)
tmp15 = tmp13 < tmp14
tmp16 = tmp12 & tmp15
tmp17 = tmp10 | tmp16
tmp18 = tl.where(tmp17, tmp0, tmp1)
tmp19 = tl.where(tmp17, tmp13, tmp14)
tmp20 = tmp18 > tmp3
tmp21 = tmp18 == tmp3
tmp22 = tmp18 != tmp18
tmp23 = tmp3 != tmp3
tmp24 = tmp22 > tmp23
tmp25 = tmp20 | tmp24
tmp26 = tmp22 & tmp23
tmp27 = tmp21 | tmp26
tmp28 = tl.full([1], 2, tl.int64)
tmp29 = tmp19 < tmp28
tmp30 = tmp27 & tmp29
tmp31 = tmp25 | tmp30
tl.where(tmp31, tmp18, tmp3)
tmp33 = tl.where(tmp31, tmp19, tmp28)
tl.store(out_ptr0 + x0, tmp4, xmask)
tl.store(out_ptr1 + x0, tmp33, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (12, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_div_0[grid(256)](primals_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_1
buf1 = empty_strided_cuda((12, 4), (4, 1), torch.float32)
triton_poi_fused_div_1[grid(48)](primals_2, buf1, 48, XBLOCK=64,
num_warps=1, num_stages=1)
buf2 = empty_strided_cuda((64, 12), (12, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (64, 4), (4, 1), 0),
reinterpret_tensor(buf1, (4, 12), (1, 4), 0), out=buf2)
del buf1
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.int64)
triton_poi_fused_max_2[grid(256)](buf2, buf3, buf4, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf2
return buf3, primals_2, reinterpret_tensor(buf0, (64, 4), (4, 1), 0
), reinterpret_tensor(buf4, (64, 4, 1), (4, 1, 1), 0)
class ArcMarginProduct_subcenterNew(nn.Module):
def __init__(self, in_features, out_features, k=3):
super().__init__()
self.weight = nn.Parameter(torch.FloatTensor(out_features * k,
in_features))
self.reset_parameters()
self.k = k
self.out_features = out_features
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.weight.size(1))
self.weight.data.uniform_(-stdv, stdv)
def forward(self, input_0):
primals_2 = self.weight
primals_1 = input_0
output = call([primals_1, primals_2])
return output[0]
| Tanmengxuan/Google-Landmark-Recognition-2020-3rd-Place-Solution | ArcMarginProduct_subcenter | false | 1,135 | [
"Apache-2.0"
] | 0 | 8e2d9056d5c88c6415827086809e73522b336fbb | https://github.com/Tanmengxuan/Google-Landmark-Recognition-2020-3rd-Place-Solution/tree/8e2d9056d5c88c6415827086809e73522b336fbb | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.data
import torch.nn.parallel
class Model(nn.Module):
def __init__(self, in_features, out_features, k=3):
super().__init__()
self.weight = nn.Parameter(torch.FloatTensor(out_features * k,
in_features))
self.reset_parameters()
self.k = k
self.out_features = out_features
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.weight.size(1))
self.weight.data.uniform_(-stdv, stdv)
def forward(self, features):
cosine_all = F.linear(F.normalize(features), F.normalize(self.weight))
cosine_all = cosine_all.view(-1, self.out_features, self.k)
cosine, _ = torch.max(cosine_all, dim=2)
return cosine
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ChannelMixer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/nh/cnhx37tsffx4r7taj3xi72s7yfpnnccem24fupfbht6b7bzliavu.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.gelu]
# Source node to ATen node mapping:
# x_1 => add, erf, mul, mul_1, mul_2
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, 0.5), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, 0.7071067811865476), kwargs = {})
# %erf : [num_users=1] = call_function[target=torch.ops.aten.erf.default](args = (%mul_1,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%erf, 1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %add), kwargs = {})
triton_poi_fused_gelu_0 = async_compile.triton('triton_poi_fused_gelu_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_gelu_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_gelu_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp3 = 0.7071067811865476
tmp4 = tmp0 * tmp3
tmp5 = libdevice.erf(tmp4)
tmp6 = 1.0
tmp7 = tmp5 + tmp6
tmp8 = tmp2 * tmp7
tl.store(out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/mz/cmz3wjq2uutgv7zzhrquuijmcstklp4wvd4q2ptdi3fpwbjqcpo6.py
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.add]
# Source node to ATen node mapping:
# x_3 => add_1
# Graph fragment:
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, %primals_1), kwargs = {})
triton_poi_fused_add_1 = async_compile.triton('triton_poi_fused_add_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 3, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_add_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.gelu]
stream0 = get_raw_stream(0)
triton_poi_fused_gelu_0.run(buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.add]
triton_poi_fused_add_1.run(buf3, primals_5, primals_1, 256, grid=grid(256), stream=stream0)
del primals_5
return (buf3, reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), buf0, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), primals_4, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class ChannelMixer(nn.Module):
def __init__(self, input_size, hidden_size, dropout=None):
super(ChannelMixer, self).__init__()
self.fc1 = nn.Linear(input_size, hidden_size)
self.fc2 = nn.Linear(hidden_size, input_size)
self.dropout = None
if dropout is not None:
self.dropout = nn.Dropout(dropout)
self.activation = nn.GELU()
def forward(self, x):
input = x
x = self.fc1(x)
x = self.activation(x)
if self.dropout is not None:
x = self.dropout(x)
x = self.fc2(x)
if self.dropout is not None:
x = self.dropout(x)
x = x + input
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'hidden_size': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_gelu_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp3 = 0.7071067811865476
tmp4 = tmp0 * tmp3
tmp5 = libdevice.erf(tmp4)
tmp6 = 1.0
tmp7 = tmp5 + tmp6
tmp8 = tmp2 * tmp7
tl.store(out_ptr0 + x0, tmp8, xmask)
@triton.jit
def triton_poi_fused_add_1(in_out_ptr0, in_ptr0, in_ptr1, xnumel, XBLOCK:
tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tl.store(in_out_ptr0 + x2, tmp4, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4), (4, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(primals_1, (64,
4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0
), alpha=1, beta=1, out=buf0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_gelu_0[grid(256)](buf0, buf1, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf2
triton_poi_fused_add_1[grid(256)](buf3, primals_5, primals_1, 256,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
return buf3, reinterpret_tensor(primals_1, (64, 4), (4, 1), 0
), buf0, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), primals_4
class ChannelMixerNew(nn.Module):
def __init__(self, input_size, hidden_size, dropout=None):
super(ChannelMixerNew, self).__init__()
self.fc1 = nn.Linear(input_size, hidden_size)
self.fc2 = nn.Linear(hidden_size, input_size)
self.dropout = None
if dropout is not None:
self.dropout = nn.Dropout(dropout)
self.activation = nn.GELU()
def forward(self, input_0):
primals_2 = self.fc1.weight
primals_3 = self.fc1.bias
primals_4 = self.fc2.weight
primals_5 = self.fc2.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| TheRealMarVin/mlp-mixer | ChannelMixer | false | 1,136 | [
"MIT"
] | 0 | 2124cb5c5adfc7af473cab535095471d4943adab | https://github.com/TheRealMarVin/mlp-mixer/tree/2124cb5c5adfc7af473cab535095471d4943adab | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, input_size, hidden_size, dropout=None):
super().__init__()
self.fc1 = nn.Linear(input_size, hidden_size)
self.fc2 = nn.Linear(hidden_size, input_size)
self.dropout = None
if dropout is not None:
self.dropout = nn.Dropout(dropout)
self.activation = nn.GELU()
def forward(self, x):
input = x
x = self.fc1(x)
x = self.activation(x)
if self.dropout is not None:
x = self.dropout(x)
x = self.fc2(x)
if self.dropout is not None:
x = self.dropout(x)
x = x + input
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Net | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/md/cmd3ewacyhu5w5hausgbjbmtnt5rr66cgczh4ibdypq7dz6p4v7g.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x_1 => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_relu_threshold_backward_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[8192],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr0 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr0 + (x2), tmp6, None)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (128, ), (1, ))
assert_size_stride(primals_3, (128, 4), (4, 1))
assert_size_stride(primals_4, (4, 128), (128, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 128), (1, 4), 0), out=buf0)
del primals_3
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0); del buf0 # reuse
buf3 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf3, 8192, grid=grid(8192), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 128), (128, 1), 0), reinterpret_tensor(primals_4, (128, 4), (1, 128), 0), alpha=1, beta=1, out=buf2)
del primals_5
return (reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 128), (128, 1), 0), primals_4, buf3, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((128, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 128), (128, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
from torch.nn import functional as F
class Net(nn.Module):
def __init__(self, obs_dim, act_dim):
super(Net, self).__init__()
self.fc0 = nn.Linear(obs_dim, 128)
self.fc1 = nn.Linear(128, act_dim)
def forward(self, x):
x = x.type_as(self.fc0.bias)
x = F.relu(self.fc0(x))
x = self.fc1(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'obs_dim': 4, 'act_dim': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch import nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_relu_threshold_backward_0(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
tl.full([XBLOCK], True, tl.int1)
x2 = xindex
x0 = xindex % 128
tmp0 = tl.load(in_out_ptr0 + x2, None)
tmp1 = tl.load(in_ptr0 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x2, tmp4, None)
tl.store(out_ptr0 + x2, tmp6, None)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (128,), (1,))
assert_size_stride(primals_3, (128, 4), (4, 1))
assert_size_stride(primals_4, (4, 128), (128, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 128), (128, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 128), (1, 4), 0), out=buf0)
del primals_3
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 128), (2048, 512, 128, 1), 0)
del buf0
buf3 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(8192)](buf1,
primals_2, buf3, 8192, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 128),
(128, 1), 0), reinterpret_tensor(primals_4, (128, 4), (1, 128),
0), alpha=1, beta=1, out=buf2)
del primals_5
return reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_1, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 128), (128, 1), 0), primals_4, buf3
class NetNew(nn.Module):
def __init__(self, obs_dim, act_dim):
super(NetNew, self).__init__()
self.fc0 = nn.Linear(obs_dim, 128)
self.fc1 = nn.Linear(128, act_dim)
def forward(self, input_0):
primals_3 = self.fc0.weight
primals_2 = self.fc0.bias
primals_4 = self.fc1.weight
primals_5 = self.fc1.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| TommeyChang/CS294-Homework | Net | false | 1,137 | [
"MIT"
] | 0 | 17b525bf4366034b45c4febd89f1053d44550237 | https://github.com/TommeyChang/CS294-Homework/tree/17b525bf4366034b45c4febd89f1053d44550237 | import torch
from torch import nn
from torch.nn import functional as F
class Model(nn.Module):
def __init__(self, obs_dim, act_dim):
super().__init__()
self.fc0 = nn.Linear(obs_dim, 128)
self.fc1 = nn.Linear(128, act_dim)
def forward(self, x):
x = x.type_as(self.fc0.bias)
x = F.relu(self.fc0(x))
x = self.fc1(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ActorDownAction | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/cz/cczdmlbbmfz5zwvgavqgsp7p2chtvjm2zzbxhjk7w5jaagtfot3j.py
# Topologically Sorted Source Nodes: [cat_1], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# cat_1 => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%tanh, %tanh_1], -1), kwargs = {})
triton_poi_fused_cat_0 = async_compile.triton('triton_poi_fused_cat_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_cat_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = (xindex // 8)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + ((4*x1) + x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = libdevice.tanh(tmp5)
tmp7 = tl.full(tmp6.shape, 0.0, tmp6.dtype)
tmp8 = tl.where(tmp4, tmp6, tmp7)
tmp9 = tmp0 >= tmp3
tmp10 = tl.full([1], 8, tl.int64)
tmp11 = tmp0 < tmp10
tmp12 = tl.load(in_ptr1 + ((4*x1) + ((-4) + x0)), tmp9 & xmask, eviction_policy='evict_last', other=0.0)
tmp13 = libdevice.tanh(tmp12)
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp9, tmp13, tmp14)
tmp16 = tl.where(tmp4, tmp8, tmp15)
tl.store(out_ptr0 + (x2), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/m6/cm6ozsdmt5vl54fxwk7cgktzswysgn2c37vsaybpucplzehkrnnz.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %le_3 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_relu_threshold_backward_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 3: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(in_out_ptr0, in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 400
x2 = xindex % 1600
x3 = (xindex // 1600)
tmp0 = tl.load(in_out_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + (x4), tmp4, xmask)
tl.store(out_ptr0 + (x2 + (1664*x3)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/64/c64g5uxk2a5hbzuhd6oikla2gb5eyfjjb6kbh7btzswha52gl5ex.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x_1 => relu_1
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_3,), kwargs = {})
# %le_2 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_2 = async_compile.triton('triton_poi_fused_relu_threshold_backward_2', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*i1', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_threshold_backward_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_threshold_backward_2(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 300
x2 = (xindex // 1200)
x3 = xindex % 1200
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x3 + (1216*x2)), tmp4, xmask)
tl.store(out_ptr1 + (x3 + (1280*x2)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/4h/c4h6r6vefoeuinm5eqv2d6wqmfj2mnjacalp633y3m6bnseb2bnk.py
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.relu, aten.view]
# Source node to ATen node mapping:
# x_1 => relu_1
# x_2 => view_4
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_3,), kwargs = {})
# %view_4 : [num_users=2] = call_function[target=torch.ops.aten.reshape.default](args = (%relu_1, [64, 300]), kwargs = {})
triton_poi_fused_relu_view_3 = async_compile.triton('triton_poi_fused_relu_view_3', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[32768],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_view_3', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_relu_view_3(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 300
x1 = (xindex // 300)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (300*(x1 % 4)) + (1216*(x1 // 4))), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/yl/cyl3pepuyms7had2pbaeyfxi5l2wn7v5miawhd67mloipgsmgxwf.py
# Topologically Sorted Source Nodes: [tanh_1, action], Original ATen: [aten.tanh, aten.mul]
# Source node to ATen node mapping:
# action => mul
# tanh_1 => tanh_2
# Graph fragment:
# %tanh_2 : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%view_5,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%tanh_2, 4), kwargs = {})
triton_poi_fused_mul_tanh_4 = async_compile.triton('triton_poi_fused_mul_tanh_4', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_mul_tanh_4', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_mul_tanh_4(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = libdevice.tanh(tmp0)
tmp2 = 4.0
tmp3 = tmp1 * tmp2
tl.store(out_ptr0 + (x0), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/tc/ctcyhc3j7gcndovmump6plsamm3m4336gv7k75wru64k3klwdqb5.py
# Topologically Sorted Source Nodes: [msg_down], Original ATen: [aten.linalg_vector_norm, aten.div]
# Source node to ATen node mapping:
# msg_down => div, pow_1, pow_2, sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_11, 2.0), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [-1], True), kwargs = {})
# %pow_2 : [num_users=2] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_11, %expand), kwargs = {})
triton_per_fused_div_linalg_vector_norm_5 = async_compile.triton('triton_per_fused_div_linalg_vector_norm_5', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.persistent_reduction(
size_hints=[64, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: 'i32', 4: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_linalg_vector_norm_5', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 1, 'num_reduction': 1, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False}
)
@triton.jit
def triton_per_fused_div_linalg_vector_norm_5(in_out_ptr0, in_ptr0, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 16
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
roffset = 0
rmask = tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp7 = 1e-12
tmp8 = triton_helpers.maximum(tmp6, tmp7)
tmp9 = tmp0 / tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr0 + (r1 + (16*x0)), tmp9, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (400, 8), (8, 1))
assert_size_stride(primals_4, (400, ), (1, ))
assert_size_stride(primals_5, (300, 400), (400, 1))
assert_size_stride(primals_6, (300, ), (1, ))
assert_size_stride(primals_7, (4, 300), (300, 1))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (400, 8), (8, 1))
assert_size_stride(primals_10, (400, ), (1, ))
assert_size_stride(primals_11, (300, 400), (400, 1))
assert_size_stride(primals_12, (300, ), (1, ))
assert_size_stride(primals_13, (16, 300), (300, 1))
assert_size_stride(primals_14, (16, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 8), (128, 32, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [cat_1], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(primals_1, primals_2, buf0, 512, grid=grid(512), stream=stream0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (64, 8), (8, 1), 0), reinterpret_tensor(primals_3, (8, 400), (1, 8), 0), out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf1, (4, 4, 4, 400), (6400, 1600, 400, 1), 0); del buf1 # reuse
buf20 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1), torch.bool)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf2, primals_4, buf20, 25600, grid=grid(25600), stream=stream0)
del primals_4
buf3 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf2, (64, 400), (400, 1), 0), reinterpret_tensor(primals_5, (400, 300), (1, 400), 0), out=buf3)
buf4 = empty_strided_cuda((4, 4, 4, 300), (4864, 1216, 300, 1), torch.float32)
buf19 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_2.run(buf3, primals_6, buf4, buf19, 19200, grid=grid(19200), stream=stream0)
del primals_6
buf5 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [x_1, x_2], Original ATen: [aten.relu, aten.view]
triton_poi_fused_relu_view_3.run(buf4, buf5, 19200, grid=grid(19200), stream=stream0)
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_8, buf5, reinterpret_tensor(primals_7, (300, 4), (1, 300), 0), alpha=1, beta=1, out=buf6)
del primals_8
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [tanh_1, action], Original ATen: [aten.tanh, aten.mul]
triton_poi_fused_mul_tanh_4.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
buf8 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (64, 8), (8, 1), 0), reinterpret_tensor(primals_9, (8, 400), (1, 8), 0), out=buf8)
del primals_9
buf9 = reinterpret_tensor(buf8, (4, 4, 4, 400), (6400, 1600, 400, 1), 0); del buf8 # reuse
buf18 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf9, primals_10, buf18, 25600, grid=grid(25600), stream=stream0)
del primals_10
buf10 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf9, (64, 400), (400, 1), 0), reinterpret_tensor(primals_11, (400, 300), (1, 400), 0), out=buf10)
buf11 = buf4; del buf4 # reuse
buf17 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_2.run(buf10, primals_12, buf11, buf17, 19200, grid=grid(19200), stream=stream0)
del primals_12
buf12 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [x_4, x_5], Original ATen: [aten.relu, aten.view]
triton_poi_fused_relu_view_3.run(buf11, buf12, 19200, grid=grid(19200), stream=stream0)
del buf11
buf13 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_14, buf12, reinterpret_tensor(primals_13, (300, 16), (1, 300), 0), alpha=1, beta=1, out=buf13)
del primals_14
buf14 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf15 = reinterpret_tensor(buf14, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf14 # reuse
buf16 = empty_strided_cuda((4, 4, 4, 16), (256, 64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [msg_down], Original ATen: [aten.linalg_vector_norm, aten.div]
triton_per_fused_div_linalg_vector_norm_5.run(buf15, buf13, buf16, 64, 16, grid=grid(64), stream=stream0)
return (buf7, buf16, reinterpret_tensor(buf0, (64, 8), (8, 1), 0), reinterpret_tensor(buf2, (64, 400), (400, 1), 0), buf5, buf6, reinterpret_tensor(buf9, (64, 400), (400, 1), 0), buf12, buf13, buf15, primals_13, buf17, primals_11, buf18, primals_7, buf19, primals_5, buf20, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((400, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((300, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((300, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((4, 300), (300, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((400, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((400, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((300, 400), (400, 1), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((300, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((16, 300), (300, 1), device='cuda:0', dtype=torch.float32)
primals_14 = rand_strided((16, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3, primals_4, primals_5, primals_6, primals_7, primals_8, primals_9, primals_10, primals_11, primals_12, primals_13, primals_14])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
from torch import nn
import torch.nn.functional as F
class MLPBase(nn.Module):
def __init__(self, num_inputs, num_outputs):
super(MLPBase, self).__init__()
self.l1 = nn.Linear(num_inputs, 400)
self.l2 = nn.Linear(400, 300)
self.l3 = nn.Linear(300, num_outputs)
def forward(self, inputs):
x = F.relu(self.l1(inputs))
x = F.relu(self.l2(x))
x = self.l3(x)
return x
class ActorDownAction(nn.Module):
"""a top-down module used in bothway message passing that passes messages to children and outputs action"""
def __init__(self, self_input_dim, action_dim, msg_dim, max_action,
max_children):
super(ActorDownAction, self).__init__()
self.max_action = max_action
self.action_base = MLPBase(self_input_dim + msg_dim, action_dim)
self.msg_base = MLPBase(self_input_dim + msg_dim, msg_dim *
max_children)
def forward(self, x, m):
xm = torch.cat((x, m), dim=-1)
xm = torch.tanh(xm)
action = self.max_action * torch.tanh(self.action_base(xm))
msg_down = self.msg_base(xm)
msg_down = F.normalize(msg_down, dim=-1)
return action, msg_down
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'self_input_dim': 4, 'action_dim': 4, 'msg_dim': 4,
'max_action': 4, 'max_children': 4}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
from torch import nn
import torch.nn.functional as F
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_cat_0(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 512
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 8
x1 = xindex // 8
x2 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (4 * x1 + x0), tmp4 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp6 = libdevice.tanh(tmp5)
tmp7 = tl.full(tmp6.shape, 0.0, tmp6.dtype)
tmp8 = tl.where(tmp4, tmp6, tmp7)
tmp9 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp12 = tl.load(in_ptr1 + (4 * x1 + (-4 + x0)), tmp9 & xmask,
eviction_policy='evict_last', other=0.0)
tmp13 = libdevice.tanh(tmp12)
tmp14 = tl.full(tmp13.shape, 0.0, tmp13.dtype)
tmp15 = tl.where(tmp9, tmp13, tmp14)
tmp16 = tl.where(tmp4, tmp8, tmp15)
tl.store(out_ptr0 + x2, tmp16, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 400
x2 = xindex % 1600
x3 = xindex // 1600
tmp0 = tl.load(in_out_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(in_out_ptr0 + x4, tmp4, xmask)
tl.store(out_ptr0 + (x2 + 1664 * x3), tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_2(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 300
x2 = xindex // 1200
x3 = xindex % 1200
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tl.store(out_ptr0 + (x3 + 1216 * x2), tmp4, xmask)
tl.store(out_ptr1 + (x3 + 1280 * x2), tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_view_3(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 19200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 300
x1 = xindex // 300
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 300 * (x1 % 4) + 1216 * (x1 // 4)), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
@triton.jit
def triton_poi_fused_mul_tanh_4(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = libdevice.tanh(tmp0)
tmp2 = 4.0
tmp3 = tmp1 * tmp2
tl.store(out_ptr0 + x0, tmp3, xmask)
@triton.jit
def triton_per_fused_div_linalg_vector_norm_5(in_out_ptr0, in_ptr0,
out_ptr0, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 64
RBLOCK: tl.constexpr = 16
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:, None]
xmask = xindex < xnumel
rindex = tl.arange(0, RBLOCK)[None, :]
tl.full([XBLOCK, RBLOCK], True, tl.int1)
r1 = rindex
x0 = xindex
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp1 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = libdevice.sqrt(tmp5)
tmp7 = 1e-12
tmp8 = triton_helpers.maximum(tmp6, tmp7)
tmp9 = tmp0 / tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr0 + (r1 + 16 * x0), tmp9, xmask)
def call(args):
(primals_1, primals_2, primals_3, primals_4, primals_5, primals_6,
primals_7, primals_8, primals_9, primals_10, primals_11, primals_12,
primals_13, primals_14) = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (400, 8), (8, 1))
assert_size_stride(primals_4, (400,), (1,))
assert_size_stride(primals_5, (300, 400), (400, 1))
assert_size_stride(primals_6, (300,), (1,))
assert_size_stride(primals_7, (4, 300), (300, 1))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (400, 8), (8, 1))
assert_size_stride(primals_10, (400,), (1,))
assert_size_stride(primals_11, (300, 400), (400, 1))
assert_size_stride(primals_12, (300,), (1,))
assert_size_stride(primals_13, (16, 300), (300, 1))
assert_size_stride(primals_14, (16,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 8), (128, 32, 8, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(512)](primals_1, primals_2, buf0, 512,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
del primals_2
buf1 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (64, 8), (8, 1), 0),
reinterpret_tensor(primals_3, (8, 400), (1, 8), 0), out=buf1)
del primals_3
buf2 = reinterpret_tensor(buf1, (4, 4, 4, 400), (6400, 1600, 400, 1), 0
)
del buf1
buf20 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_1[grid(25600)](buf2,
primals_4, buf20, 25600, XBLOCK=256, num_warps=4, num_stages=1)
del primals_4
buf3 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf2, (64, 400), (400, 1), 0),
reinterpret_tensor(primals_5, (400, 300), (1, 400), 0), out=buf3)
buf4 = empty_strided_cuda((4, 4, 4, 300), (4864, 1216, 300, 1),
torch.float32)
buf19 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_2[grid(19200)](buf3,
primals_6, buf4, buf19, 19200, XBLOCK=256, num_warps=4,
num_stages=1)
del primals_6
buf5 = buf3
del buf3
triton_poi_fused_relu_view_3[grid(19200)](buf4, buf5, 19200, XBLOCK
=128, num_warps=4, num_stages=1)
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_8, buf5, reinterpret_tensor(primals_7,
(300, 4), (1, 300), 0), alpha=1, beta=1, out=buf6)
del primals_8
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_mul_tanh_4[grid(256)](buf6, buf7, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf8 = empty_strided_cuda((64, 400), (400, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (64, 8), (8, 1), 0),
reinterpret_tensor(primals_9, (8, 400), (1, 8), 0), out=buf8)
del primals_9
buf9 = reinterpret_tensor(buf8, (4, 4, 4, 400), (6400, 1600, 400, 1), 0
)
del buf8
buf18 = empty_strided_cuda((4, 4, 4, 400), (6656, 1664, 400, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_1[grid(25600)](buf9,
primals_10, buf18, 25600, XBLOCK=256, num_warps=4, num_stages=1)
del primals_10
buf10 = empty_strided_cuda((64, 300), (300, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf9, (64, 400), (400, 1), 0),
reinterpret_tensor(primals_11, (400, 300), (1, 400), 0), out=buf10)
buf11 = buf4
del buf4
buf17 = empty_strided_cuda((4, 4, 4, 300), (5120, 1280, 300, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_2[grid(19200)](buf10,
primals_12, buf11, buf17, 19200, XBLOCK=256, num_warps=4,
num_stages=1)
del primals_12
buf12 = buf10
del buf10
triton_poi_fused_relu_view_3[grid(19200)](buf11, buf12, 19200,
XBLOCK=128, num_warps=4, num_stages=1)
del buf11
buf13 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.addmm(primals_14, buf12, reinterpret_tensor(
primals_13, (300, 16), (1, 300), 0), alpha=1, beta=1, out=buf13)
del primals_14
buf14 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
buf15 = reinterpret_tensor(buf14, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf14
buf16 = empty_strided_cuda((4, 4, 4, 16), (256, 64, 16, 1), torch.
float32)
triton_per_fused_div_linalg_vector_norm_5[grid(64)](buf15, buf13,
buf16, 64, 16, XBLOCK=32, num_warps=4, num_stages=1)
return (buf7, buf16, reinterpret_tensor(buf0, (64, 8), (8, 1), 0),
reinterpret_tensor(buf2, (64, 400), (400, 1), 0), buf5, buf6,
reinterpret_tensor(buf9, (64, 400), (400, 1), 0), buf12, buf13,
buf15, primals_13, buf17, primals_11, buf18, primals_7, buf19,
primals_5, buf20)
class MLPBase(nn.Module):
def __init__(self, num_inputs, num_outputs):
super(MLPBase, self).__init__()
self.l1 = nn.Linear(num_inputs, 400)
self.l2 = nn.Linear(400, 300)
self.l3 = nn.Linear(300, num_outputs)
def forward(self, inputs):
x = F.relu(self.l1(inputs))
x = F.relu(self.l2(x))
x = self.l3(x)
return x
class ActorDownActionNew(nn.Module):
"""a top-down module used in bothway message passing that passes messages to children and outputs action"""
def __init__(self, self_input_dim, action_dim, msg_dim, max_action,
max_children):
super(ActorDownActionNew, self).__init__()
self.max_action = max_action
self.action_base = MLPBase(self_input_dim + msg_dim, action_dim)
self.msg_base = MLPBase(self_input_dim + msg_dim, msg_dim *
max_children)
def forward(self, input_0, input_1):
primals_3 = self.action_base.l1.weight
primals_4 = self.action_base.l1.bias
primals_5 = self.action_base.l2.weight
primals_6 = self.action_base.l2.bias
primals_7 = self.action_base.l3.weight
primals_8 = self.action_base.l3.bias
primals_9 = self.msg_base.l1.weight
primals_10 = self.msg_base.l1.bias
primals_11 = self.msg_base.l2.weight
primals_12 = self.msg_base.l2.bias
primals_13 = self.msg_base.l3.weight
primals_14 = self.msg_base.l3.bias
primals_1 = input_0
primals_2 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7, primals_8, primals_9,
primals_10, primals_11, primals_12, primals_13, primals_14])
return output[0], output[1]
| TachikakaMin/dreamer-torch | ActorDownAction | false | 1,138 | [
"MIT"
] | 0 | 3c99526f4507e28cf8b34ada0321001adcf8ae1f | https://github.com/TachikakaMin/dreamer-torch/tree/3c99526f4507e28cf8b34ada0321001adcf8ae1f | import torch
from torch import nn
import torch.nn.functional as F
class MLPBase(nn.Module):
def __init__(self, num_inputs, num_outputs):
super().__init__()
self.l1 = nn.Linear(num_inputs, 400)
self.l2 = nn.Linear(400, 300)
self.l3 = nn.Linear(300, num_outputs)
def forward(self, inputs):
x = F.relu(self.l1(inputs))
x = F.relu(self.l2(x))
x = self.l3(x)
return x
class Model(nn.Module):
"""a top-down module used in bothway message passing that passes messages to children and outputs action"""
def __init__(self, self_input_dim, action_dim, msg_dim, max_action,
max_children):
super().__init__()
self.max_action = max_action
self.action_base = MLPBase(self_input_dim + msg_dim, action_dim)
self.msg_base = MLPBase(self_input_dim + msg_dim, msg_dim *
max_children)
def forward(self, x, m):
xm = torch.cat((x, m), dim=-1)
xm = torch.tanh(xm)
action = self.max_action * torch.tanh(self.action_base(xm))
msg_down = self.msg_base(xm)
msg_down = F.normalize(msg_down, dim=-1)
return action, msg_down
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'self_input_dim': 4, 'action_dim': 4, 'msg_dim': 4,
'max_action': 4, 'max_children': 4}]
|
UpsampleConvLayer | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/3g/c3g7vlanljv77gridfmcaf7xvc2k6yt5nwslumydb6jkvzhxp24z.py
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten._unsafe_index, aten.reflection_pad2d]
# Source node to ATen node mapping:
# x => _unsafe_index
# x_1 => _unsafe_index_1, _unsafe_index_2
# Graph fragment:
# %_unsafe_index : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%primals_1, [None, None, %unsqueeze, %convert_element_type_1]), kwargs = {})
# %_unsafe_index_1 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index, [None, None, %sub_1, None]), kwargs = {})
# %_unsafe_index_2 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index_1, [None, None, None, %sub_1]), kwargs = {})
triton_poi_fused__unsafe_index_reflection_pad2d_0 = async_compile.triton('triton_poi_fused__unsafe_index_reflection_pad2d_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[1024],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_index_reflection_pad2d_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 0, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused__unsafe_index_reflection_pad2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 8) % 8
x0 = xindex % 8
x2 = (xindex // 64)
x5 = xindex
tmp0 = 3 + ((-1)*(tl_math.abs((-3) + (tl_math.abs((-2) + x1)))))
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tmp5 = 3 + ((-1)*(tl_math.abs((-3) + (tl_math.abs((-2) + x0)))))
tmp6 = tmp5.to(tl.float32)
tmp7 = tmp6 * tmp2
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.load(in_ptr0 + (tmp8 + (4*tmp4) + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x5), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_6/inductor_cache/42/c42i6kggymcvforsoo45syfc6w3ujwnd3pxalcsjxkelshjyz7gv.py
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_2 => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_2, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
triton_poi_fused_convolution_1 = async_compile.triton('triton_poi_fused_convolution_1', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[512],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 2, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 25) % 4
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 8), (256, 64, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten._unsafe_index, aten.reflection_pad2d]
stream0 = get_raw_stream(0)
triton_poi_fused__unsafe_index_reflection_pad2d_0.run(primals_1, buf0, 1024, grid=grid(1024), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 5, 5), (100, 25, 5, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf2, primals_3, 400, grid=grid(400), stream=stream0)
del primals_3
return (buf2, primals_2, buf0, )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2, primals_3])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
class UpsampleConvLayer(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride,
scale_factor):
super(UpsampleConvLayer, self).__init__()
self._scale_factor = scale_factor
self._reflection_pad = nn.ReflectionPad2d(kernel_size // 2)
self._conv = nn.Conv2d(in_channels, out_channels, kernel_size, stride)
def forward(self, x):
x = nn.functional.interpolate(x, mode='nearest', scale_factor=self.
_scale_factor)
x = self._reflection_pad(x)
x = self._conv(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4,
'stride': 1, 'scale_factor': 1.0}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
@triton.jit
def triton_poi_fused__unsafe_index_reflection_pad2d_0(in_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 1024
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 8 % 8
x0 = xindex % 8
x2 = xindex // 64
x5 = xindex
tmp0 = 3 + -1 * tl_math.abs(-3 + tl_math.abs(-2 + x1))
tmp1 = tmp0.to(tl.float32)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tmp5 = 3 + -1 * tl_math.abs(-3 + tl_math.abs(-2 + x0))
tmp6 = tmp5.to(tl.float32)
tmp7 = tmp6 * tmp2
tmp8 = tmp7.to(tl.int32)
tmp9 = tl.load(in_ptr0 + (tmp8 + 4 * tmp4 + 16 * x2), xmask,
eviction_policy='evict_last')
tl.store(out_ptr0 + x5, tmp9, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 25 % 4
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_3, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 8, 8), (256, 64, 8, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__unsafe_index_reflection_pad2d_0[grid(1024)](primals_1
, buf0, 1024, XBLOCK=256, num_warps=4, num_stages=1)
del primals_1
buf1 = extern_kernels.convolution(buf0, primals_2, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 4, 5, 5), (100, 25, 5, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_1[grid(400)](buf2, primals_3, 400,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
return buf2, primals_2, buf0
class UpsampleConvLayerNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride,
scale_factor):
super(UpsampleConvLayerNew, self).__init__()
self._scale_factor = scale_factor
self._reflection_pad = nn.ReflectionPad2d(kernel_size // 2)
self._conv = nn.Conv2d(in_channels, out_channels, kernel_size, stride)
def forward(self, input_0):
primals_1 = self._conv.weight
primals_3 = self._conv.bias
primals_2 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| ThomasRanvier/cnn_style_transfer | UpsampleConvLayer | false | 1,139 | [
"MIT"
] | 0 | 90b6c76c20263c22f4e45184d572284726ecbd7b | https://github.com/ThomasRanvier/cnn_style_transfer/tree/90b6c76c20263c22f4e45184d572284726ecbd7b | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride,
scale_factor):
super().__init__()
self._scale_factor = scale_factor
self._reflection_pad = nn.ReflectionPad2d(kernel_size // 2)
self._conv = nn.Conv2d(in_channels, out_channels, kernel_size, stride)
def forward(self, x):
x = nn.functional.interpolate(x, mode='nearest', scale_factor=self.
_scale_factor)
x = self._reflection_pad(x)
x = self._conv(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_channels': 4, 'out_channels': 4, 'kernel_size': 4,
'stride': 1, 'scale_factor': 1.0}]
|
StructuralProbe | # AOT ID: ['0_forward']
from ctypes import c_void_p, c_long, c_int
import torch
import math
import random
import os
import tempfile
from math import inf, nan
from torch._inductor.hooks import run_intermediate_hooks
from torch._inductor.utils import maybe_profile
from torch._inductor.codegen.memory_planning import _align as align
from torch import device, empty_strided
from torch._inductor.async_compile import AsyncCompile
from torch._inductor.select_algorithm import extern_kernels
from torch._inductor.codegen.multi_kernel import MultiKernelCall
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid, split_scan_grid, grid_combo_kernels, start_graph, end_graph
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
aten = torch.ops.aten
inductor_ops = torch.ops.inductor
_quantized = torch.ops._quantized
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cpu = torch._C._dynamo.guards._empty_strided_cpu
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
empty_strided_xpu = torch._C._dynamo.guards._empty_strided_xpu
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
alloc_from_pool = torch.ops.inductor._alloc_from_pool
async_compile = AsyncCompile()
# kernel path: runs/run_shard_6/inductor_cache/lc/clc2yzg4fyrzlue3igzxk6yznmjxztgiekr3dkkq6awe43hpqv34.py
# Topologically Sorted Source Nodes: [diffs, squared_diffs, squared_distances], Original ATen: [aten.sub, aten.pow, aten.sum]
# Source node to ATen node mapping:
# diffs => sub
# squared_diffs => pow_1
# squared_distances => sum_1
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%expand, %permute), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [-1]), kwargs = {})
triton_poi_fused_pow_sub_sum_0 = async_compile.triton('triton_poi_fused_pow_sub_sum_0', '''
import triton
import triton.language as tl
from triton.compiler.compiler import AttrsDescriptor
from torch._inductor.runtime import triton_helpers, triton_heuristics
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch._inductor.runtime.hints import AutotuneHint, ReductionHint, TileHint, instance_descriptor, DeviceProperties
@triton_heuristics.pointwise(
size_hints=[64],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: 'i32'}, 'device': DeviceProperties(type='cuda', index=0, cc=80, major=8, regs_per_multiprocessor=65536, max_threads_per_multi_processor=2048, multi_processor_count=108), 'constants': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_pow_sub_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 8, 'num_reduction': 0, 'backend_hash': 'A9C866B4A14FD3277824029365D703C2427B2E685E54EC9B3EF4ADC8D1EEAC1D', 'are_deterministic_algorithms_enabled': False, 'assert_indirect_indexing': True, 'autotune_local_cache': True, 'autotune_pointwise': True, 'autotune_remote_cache': None, 'force_disable_caches': False, 'dynamic_scale_rblock': True, 'max_autotune': False, 'max_autotune_pointwise': False, 'min_split_scan_rblock': 256, 'spill_threshold': 16, 'store_cubin': False},
min_elem_per_thread=0
)
@triton.jit
def triton_poi_fused_pow_sub_sum_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = (xindex // 4)
x0 = xindex % 4
x2 = (xindex // 16)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (4*x3), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + ((4*x0) + (16*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (1 + (4*x3)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + (4*x0) + (16*x2)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (2 + (4*x3)), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr0 + (2 + (4*x0) + (16*x2)), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (3 + (4*x3)), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr0 + (3 + (4*x0) + (16*x2)), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp6 = tmp4 - tmp5
tmp7 = tmp6 * tmp6
tmp8 = tmp3 + tmp7
tmp11 = tmp9 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tmp8 + tmp12
tmp16 = tmp14 - tmp15
tmp17 = tmp16 * tmp16
tmp18 = tmp13 + tmp17
tl.store(out_ptr0 + (x4), tmp18, xmask)
''', device_str='cuda')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [transformed], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [diffs, squared_diffs, squared_distances], Original ATen: [aten.sub, aten.pow, aten.sum]
stream0 = get_raw_stream(0)
triton_poi_fused_pow_sub_sum_0.run(buf0, buf1, 64, grid=grid(64), stream=stream0)
return (buf1, buf0, reinterpret_tensor(primals_2, (4, 16), (1, 4), 0), )
def benchmark_compiled_module(times=10, repeat=10):
from torch._dynamo.testing import rand_strided
from torch._inductor.utils import print_performance
primals_1 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
fn = lambda: call([primals_1, primals_2])
return print_performance(fn, times=times, repeat=repeat)
if __name__ == "__main__":
from torch._inductor.wrapper_benchmark import compiled_module_main
compiled_module_main('None', benchmark_compiled_module)
| import torch
import torch.nn as nn
import torch.utils.data.dataloader
class StructuralProbe(nn.Module):
""" Computes squared L2 distance after projection by a matrix.
For a batch of sentences, computes all n^2 pairs of distances
for each sentence in the batch.
"""
def __init__(self, model_dim, rank, device):
super().__init__()
self.probe_rank = rank
self.model_dim = model_dim
self.proj = nn.Parameter(data=torch.zeros(self.model_dim, self.
probe_rank))
nn.init.uniform_(self.proj, -0.05, 0.05)
self
def forward(self, batch):
""" Computes all n^2 pairs of distances after projection
for each sentence in a batch.
Note that due to padding, some distances will be non-zero for pads.
Computes (B(h_i-h_j))^T(B(h_i-h_j)) for all i,j
Args:
batch: a batch of word representations of the shape
(batch_size, max_seq_len, representation_dim)
Returns:
A tensor of distances of shape (batch_size, max_seq_len, max_seq_len)
"""
transformed = torch.matmul(batch, self.proj)
_batchlen, seqlen, _rank = transformed.size()
transformed = transformed.unsqueeze(2)
transformed = transformed.expand(-1, -1, seqlen, -1)
transposed = transformed.transpose(1, 2)
diffs = transformed - transposed
squared_diffs = diffs.pow(2)
squared_distances = torch.sum(squared_diffs, -1)
return squared_distances
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'model_dim': 4, 'rank': 4, 'device': 0}]
| import torch
from torch._inductor.select_algorithm import extern_kernels
import triton
import triton.language as tl
from torch._inductor.runtime.triton_heuristics import grid
from torch._C import _cuda_getCurrentRawStream as get_raw_stream
import torch.nn as nn
import torch.utils.data.dataloader
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
empty_strided_cuda = torch._C._dynamo.guards._empty_strided_cuda
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_pow_sub_sum_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 64
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex // 4
x0 = xindex % 4
x2 = xindex // 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x3, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (4 * x0 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (1 + 4 * x3), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (1 + 4 * x0 + 16 * x2), xmask, eviction_policy
='evict_last')
tmp9 = tl.load(in_ptr0 + (2 + 4 * x3), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr0 + (2 + 4 * x0 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (3 + 4 * x3), xmask, eviction_policy='evict_last'
)
tmp15 = tl.load(in_ptr0 + (3 + 4 * x0 + 16 * x2), xmask,
eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp3 = tmp2 * tmp2
tmp6 = tmp4 - tmp5
tmp7 = tmp6 * tmp6
tmp8 = tmp3 + tmp7
tmp11 = tmp9 - tmp10
tmp12 = tmp11 * tmp11
tmp13 = tmp8 + tmp12
tmp16 = tmp14 - tmp15
tmp17 = tmp16 * tmp16
tmp18 = tmp13 + tmp17
tl.store(out_ptr0 + x4, tmp18, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 4), (4, 1))
assert_size_stride(primals_2, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0),
primals_1, out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_pow_sub_sum_0[grid(64)](buf0, buf1, 64, XBLOCK=64,
num_warps=1, num_stages=1)
return buf1, buf0, reinterpret_tensor(primals_2, (4, 16), (1, 4), 0)
class StructuralProbeNew(nn.Module):
""" Computes squared L2 distance after projection by a matrix.
For a batch of sentences, computes all n^2 pairs of distances
for each sentence in the batch.
"""
def __init__(self, model_dim, rank, device):
super().__init__()
self.probe_rank = rank
self.model_dim = model_dim
self.proj = nn.Parameter(data=torch.zeros(self.model_dim, self.
probe_rank))
nn.init.uniform_(self.proj, -0.05, 0.05)
self
def forward(self, input_0):
primals_1 = self.proj
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| TimO96/NLP2 | StructuralProbe | false | 1,140 | [
"MIT"
] | 0 | 83f65a385457f68397c641f38b53df0110282578 | https://github.com/TimO96/NLP2/tree/83f65a385457f68397c641f38b53df0110282578 | import torch
import torch.nn as nn
import torch.utils.data.dataloader
class Model(nn.Module):
""" Computes squared L2 distance after projection by a matrix.
For a batch of sentences, computes all n^2 pairs of distances
for each sentence in the batch.
"""
def __init__(self, model_dim, rank, device):
super().__init__()
self.probe_rank = rank
self.model_dim = model_dim
self.proj = nn.Parameter(data=torch.zeros(self.model_dim, self.
probe_rank))
nn.init.uniform_(self.proj, -0.05, 0.05)
self
def forward(self, batch):
""" Computes all n^2 pairs of distances after projection
for each sentence in a batch.
Note that due to padding, some distances will be non-zero for pads.
Computes (B(h_i-h_j))^T(B(h_i-h_j)) for all i,j
Args:
batch: a batch of word representations of the shape
(batch_size, max_seq_len, representation_dim)
Returns:
A tensor of distances of shape (batch_size, max_seq_len, max_seq_len)
"""
transformed = torch.matmul(batch, self.proj)
_batchlen, seqlen, _rank = transformed.size()
transformed = transformed.unsqueeze(2)
transformed = transformed.expand(-1, -1, seqlen, -1)
transposed = transformed.transpose(1, 2)
diffs = transformed - transposed
squared_diffs = diffs.pow(2)
squared_distances = torch.sum(squared_diffs, -1)
return squared_distances
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [4, 4, 0]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.