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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
MLP | # 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_4/inductor_cache/cc/ccc3rfqzulhvn4nmibqhwjsomjzanxwwc34nr7d7enltyb7czw67.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=[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_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_relu_threshold_backward_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 32000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 500
x2 = (xindex // 2000)
x3 = xindex % 2000
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 + (2016*x2)), tmp4, xmask)
tl.store(out_ptr1 + (x3 + (2048*x2)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/od/cod4chcodxhfkpwgvalpo3wjcotr3qr3momwpd2myrvazuq2fy67.py
# Topologically Sorted Source Nodes: [relu, linear_1], Original ATen: [aten.relu, aten.view]
# Source node to ATen node mapping:
# linear_1 => view_2
# relu => relu
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %view_2 : [num_users=2] = call_function[target=torch.ops.aten.reshape.default](args = (%relu, [64, 500]), kwargs = {})
triton_poi_fused_relu_view_1 = async_compile.triton('triton_poi_fused_relu_view_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: '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_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_relu_view_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 32000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 500
x1 = (xindex // 500)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (500*(x1 % 4)) + (2016*(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 = args
args.clear()
assert_size_stride(primals_1, (500, 4), (4, 1))
assert_size_stride(primals_2, (500, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 500), (500, 1))
assert_size_stride(primals_5, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 500), (500, 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, 500), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 500), (8064, 2016, 500, 1), torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 500), (8192, 2048, 500, 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(buf0, primals_2, buf1, buf4, 32000, grid=grid(32000), stream=stream0)
del primals_2
buf2 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [relu, linear_1], Original ATen: [aten.relu, aten.view]
triton_poi_fused_relu_view_1.run(buf1, buf2, 32000, grid=grid(32000), stream=stream0)
del buf1
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, buf2, reinterpret_tensor(primals_4, (500, 4), (1, 500), 0), alpha=1, beta=1, out=buf3)
del primals_5
return (reinterpret_tensor(buf3, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf2, primals_4, 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((500, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((500, ), (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, 500), (500, 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 MLP(nn.Module):
def __init__(self, input_size, output_size, hidden_size=500,
weight_decay=0.0):
super(MLP, self).__init__()
self.i2h = nn.Linear(in_features=input_size, out_features=hidden_size)
self.Dropout = nn.Dropout(p=0.5)
self.h2o = nn.Linear(in_features=hidden_size, out_features=output_size)
self.loss = nn.functional.cross_entropy
self.weight_decay = weight_decay
def forward(self, x):
return self.h2o(torch.relu(self.Dropout(self.i2h(x))))
def predict(self, x):
return self.forward(x).argmax(dim=1)
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
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_relu_threshold_backward_0(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 32000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 500
x2 = xindex // 2000
x3 = xindex % 2000
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 + 2016 * x2), tmp4, xmask)
tl.store(out_ptr1 + (x3 + 2048 * x2), tmp6, xmask)
@triton.jit
def triton_poi_fused_relu_view_1(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 32000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 500
x1 = xindex // 500
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 500 * (x1 % 4) + 2016 * (x1 // 4)), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (500, 4), (4, 1))
assert_size_stride(primals_2, (500,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (4, 500), (500, 1))
assert_size_stride(primals_5, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 500), (500, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 500), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 500), (8064, 2016, 500, 1),
torch.float32)
buf4 = empty_strided_cuda((4, 4, 4, 500), (8192, 2048, 500, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(32000)](buf0,
primals_2, buf1, buf4, 32000, XBLOCK=256, num_warps=4, num_stages=1
)
del primals_2
buf2 = buf0
del buf0
triton_poi_fused_relu_view_1[grid(32000)](buf1, buf2, 32000, XBLOCK
=256, num_warps=4, num_stages=1)
del buf1
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, buf2, reinterpret_tensor(primals_4,
(500, 4), (1, 500), 0), alpha=1, beta=1, out=buf3)
del primals_5
return reinterpret_tensor(buf3, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf2, primals_4, buf4
class MLPNew(nn.Module):
def __init__(self, input_size, output_size, hidden_size=500,
weight_decay=0.0):
super(MLPNew, self).__init__()
self.i2h = nn.Linear(in_features=input_size, out_features=hidden_size)
self.Dropout = nn.Dropout(p=0.5)
self.h2o = nn.Linear(in_features=hidden_size, out_features=output_size)
self.loss = nn.functional.cross_entropy
self.weight_decay = weight_decay
def predict(self, x):
return self.forward(x).argmax(dim=1)
def forward(self, input_0):
primals_1 = self.i2h.weight
primals_2 = self.i2h.bias
primals_4 = self.h2o.weight
primals_5 = self.h2o.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| gchrupala/platalea | MLP | false | 6,728 | [
"Apache-2.0"
] | 1 | 65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | https://github.com/gchrupala/platalea/tree/65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def __init__(self, input_size, output_size, hidden_size=500,
weight_decay=0.0):
super().__init__()
self.i2h = nn.Linear(in_features=input_size, out_features=hidden_size)
self.Dropout = nn.Dropout(p=0.5)
self.h2o = nn.Linear(in_features=hidden_size, out_features=output_size)
self.loss = nn.functional.cross_entropy
self.weight_decay = weight_decay
def forward(self, x):
return self.h2o(torch.relu(self.Dropout(self.i2h(x))))
def predict(self, x):
return self.forward(x).argmax(dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
MeanPool | # 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_4/inductor_cache/hh/chh6c5w5qa6uf7vojzls7kg4by5riqn4sgtlt67ukhrqv4nd6zcl.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# x => mean
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%arg0_1, [1]), 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=[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_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_mean_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 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / 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), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mean]
stream0 = get_raw_stream(0)
triton_poi_fused_mean_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
class MeanPool(nn.Module):
def __init__(self):
super(MeanPool, self).__init__()
def forward(self, input):
x = input.mean(dim=1)
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
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_mean_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 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / 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), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mean_0[grid(64)](arg0_1, buf0, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del arg0_1
return buf0,
class MeanPoolNew(nn.Module):
def __init__(self):
super(MeanPoolNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| gchrupala/platalea | MeanPool | false | 6,729 | [
"Apache-2.0"
] | 1 | 65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | https://github.com/gchrupala/platalea/tree/65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def __init__(self):
super().__init__()
def forward(self, input):
x = input.mean(dim=1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
NNet | # 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_4/inductor_cache/nq/cnqjufcqn3ur3s7xvlb2i747nyf24md4zaiatlwgkasynplfjstu.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=[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_4/inductor_cache/e7/ce7ewq7bv76ie5hdmfxjj46viiuxlajdhtbost7f4gwclfa3hk4i.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=[16384],
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 = 16384
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)
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, 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, (256, 64), (64, 1))
assert_size_stride(primals_5, (256, ), (1, ))
assert_size_stride(primals_6, (4, 256), (256, 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
buf6 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 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, buf6, 4096, grid=grid(4096), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 256), (256, 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, 256), (1, 64), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 256), (4096, 1024, 256, 1), 0); del buf2 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf3, primals_5, buf5, 16384, grid=grid(16384), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 256), (256, 1), 0), reinterpret_tensor(primals_6, (256, 4), (1, 256), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 64), (64, 1), 0), reinterpret_tensor(buf3, (64, 256), (256, 1), 0), primals_6, buf5, 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((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((256, 64), (64, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 256), (256, 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
import torch.nn.functional as F
class NNet(nn.Module):
def __init__(self, input_dim, output_dim):
super(NNet, self).__init__()
self.linear1 = nn.Linear(input_dim, 64)
self.linear2 = nn.Linear(64, 256)
self.linear3 = nn.Linear(256, output_dim)
def forward(self, x):
x = F.relu(self.linear1(x))
x = F.relu(self.linear2(x))
x = self.linear3(x)
return 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 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_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_relu_threshold_backward_1(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 % 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)
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, 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, (256, 64), (64, 1))
assert_size_stride(primals_5, (256,), (1,))
assert_size_stride(primals_6, (4, 256), (256, 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
buf6 = 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, buf6, 4096, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 64), (64, 1), 0),
reinterpret_tensor(primals_4, (64, 256), (1, 64), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 256), (4096, 1024, 256, 1), 0
)
del buf2
buf5 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_1[grid(16384)](buf3,
primals_5, buf5, 16384, XBLOCK=256, 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, 256),
(256, 1), 0), reinterpret_tensor(primals_6, (256, 4), (1, 256),
0), alpha=1, beta=1, out=buf4)
del primals_7
return reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 64), (64, 1), 0), reinterpret_tensor(
buf3, (64, 256), (256, 1), 0), primals_6, buf5, primals_4, buf6
class NNetNew(nn.Module):
def __init__(self, input_dim, output_dim):
super(NNetNew, self).__init__()
self.linear1 = nn.Linear(input_dim, 64)
self.linear2 = nn.Linear(64, 256)
self.linear3 = nn.Linear(256, output_dim)
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.linear3.weight
primals_7 = self.linear3.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| gautam-sharma1/openRL | NNet | false | 6,730 | [
"MIT"
] | 1 | 14310a97a328fe5682a01ee85d83a6b5e1ae29ca | https://github.com/gautam-sharma1/openRL/tree/14310a97a328fe5682a01ee85d83a6b5e1ae29ca | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, input_dim, output_dim):
super().__init__()
self.linear1 = nn.Linear(input_dim, 64)
self.linear2 = nn.Linear(64, 256)
self.linear3 = nn.Linear(256, output_dim)
def forward(self, x):
x = F.relu(self.linear1(x))
x = F.relu(self.linear2(x))
x = self.linear3(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ELBO | # 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_4/inductor_cache/tn/ctnbp7njmu7nuaon7mzem6bmeplmdezc4szztfhl3m7mt2jg5dkh.py
# Topologically Sorted Source Nodes: [likelihood_cost, complexity_cost, total_loss], Original ATen: [aten.mse_loss, aten.mul, aten.add]
# Source node to ATen node mapping:
# complexity_cost => mul
# likelihood_cost => mean, pow_1, sub
# total_loss => add
# 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=2] = call_function[target=torch.ops.aten.mean.default](args = (%pow_1,), kwargs = {})
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg2_1, 1.0), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, %mul), kwargs = {})
triton_per_fused_add_mse_loss_mul_0 = async_compile.triton('triton_per_fused_add_mse_loss_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.persistent_reduction(
size_hints=[1, 256],
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': {6: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 7), equal_to_1=(6,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mse_loss_mul_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, '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_add_mse_loss_mul_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_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)
tmp9 = tl.load(in_ptr2 + (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
tmp10 = 1.0
tmp11 = tmp9 * tmp10
tmp12 = tmp8 + tmp11
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp8, None)
tl.store(out_ptr0 + (tl.broadcast_to(r0, [RBLOCK])), tmp11, None)
tl.store(out_ptr1 + (tl.broadcast_to(r0, [RBLOCK])), tmp12, 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((), (), torch.float32)
buf1 = buf0; 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.float32)
# Topologically Sorted Source Nodes: [likelihood_cost, complexity_cost, total_loss], Original ATen: [aten.mse_loss, aten.mul, aten.add]
stream0 = get_raw_stream(0)
triton_per_fused_add_mse_loss_mul_0.run(buf1, arg1_1, arg0_1, arg2_1, buf2, buf3, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
return (buf3, buf1, 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 torch
import torch.nn.functional
import torch.nn as nn
class ELBO(nn.Module):
def __init__(self, train_size, loss_function=nn.MSELoss()):
"""
Quantify the Evidence Lower Bound (ELBO) and provide the total loss.
"""
super(ELBO, self).__init__()
self.train_size = train_size
self.loss_function = loss_function
def forward(self, input, target, kl, kl_weight=1.0):
"""
Kullback-Leibler divergence. This comes from
the euqation (4) in Shridhar et. al. 2019, which consists of likelihood cost
(is dependent on data) and complexity cost (id dependent on distribution).
"""
assert not target.requires_grad
likelihood_cost = self.loss_function(input, target)
complexity_cost = kl_weight * kl
total_loss = likelihood_cost + complexity_cost
return total_loss, likelihood_cost, complexity_cost
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 [[], {'train_size': False}]
| 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.functional
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_per_fused_add_mse_loss_mul_0(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_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)
tmp9 = tl.load(in_ptr2 + 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
tmp10 = 1.0
tmp11 = tmp9 * tmp10
tmp12 = tmp8 + tmp11
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp8, None)
tl.store(out_ptr0 + tl.broadcast_to(r0, [RBLOCK]), tmp11, None)
tl.store(out_ptr1 + tl.broadcast_to(r0, [RBLOCK]), tmp12, 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((), (), torch.float32)
buf1 = buf0
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.float32)
get_raw_stream(0)
triton_per_fused_add_mse_loss_mul_0[grid(1)](buf1, arg1_1, arg0_1,
arg2_1, buf2, buf3, 1, 256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
del arg2_1
return buf3, buf1, buf2
class ELBONew(nn.Module):
def __init__(self, train_size, loss_function=nn.MSELoss()):
"""
Quantify the Evidence Lower Bound (ELBO) and provide the total loss.
"""
super(ELBONew, self).__init__()
self.train_size = train_size
self.loss_function = loss_function
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], output[2]
| geek-yang/NEmo | ELBO | false | 6,731 | [
"Apache-2.0"
] | 1 | 4f310535c4865f3816155b99b4a2bbb891672cc9 | https://github.com/geek-yang/NEmo/tree/4f310535c4865f3816155b99b4a2bbb891672cc9 | import torch
import torch.nn.functional
import torch.nn as nn
class Model(nn.Module):
def __init__(self, train_size, loss_function=nn.MSELoss()):
"""
Quantify the Evidence Lower Bound (ELBO) and provide the total loss.
"""
super().__init__()
self.train_size = train_size
self.loss_function = loss_function
def forward(self, input, target, kl, kl_weight=1.0):
"""
Kullback-Leibler divergence. This comes from
the euqation (4) in Shridhar et. al. 2019, which consists of likelihood cost
(is dependent on data) and complexity cost (id dependent on distribution).
"""
assert not target.requires_grad
likelihood_cost = self.loss_function(input, target)
complexity_cost = kl_weight * kl
total_loss = likelihood_cost + complexity_cost
return total_loss, likelihood_cost, complexity_cost
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 [False]
|
SCNLayer | # 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_4/inductor_cache/l4/cl4boort6vfsvh6h6bfd4lck36kbmtipkqcrnhckuuxer6sfib77.py
# Topologically Sorted Source Nodes: [X], Original ATen: [aten.zeros]
# Source node to ATen node mapping:
# X => full
# Graph fragment:
# %full : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
triton_poi_fused_zeros_0 = async_compile.triton('triton_poi_fused_zeros_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_zeros_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_zeros_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
tmp0 = 0.0
tl.store(out_ptr0 + (x0), tmp0, 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, ), (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: [X], Original ATen: [aten.zeros]
stream0 = get_raw_stream(0)
triton_poi_fused_zeros_0.run(buf0, 16, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [X], Original ATen: [aten._sparse_addmm]
buf1 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf0, (4, 4), (1, 4), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), beta=0)
del primals_1
del primals_2
buf2 = buf1
del buf1
buf3 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_4, reinterpret_tensor(buf2, (4, 4), (1, 4), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf3)
del primals_3
del primals_4
return (buf3, reinterpret_tensor(buf2, (4, 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), (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, ), (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 chebyshev(L, X, k=3):
if k == 1:
return torch.sparse.mm(L, X)
dp = [X, torch.sparse.mm(L, X)]
for i in range(2, k):
nxt = 2 * torch.sparse.mm(L, dp[i - 1])
dp.append(torch.sparse.FloatTensor.add(nxt, -dp[i - 2]))
return torch.cat(dp, dim=1)
class SCNLayer(nn.Module):
def __init__(self, feature_size, output_size, enable_bias=True, k=1):
super().__init__()
self.k = k
self.conv = nn.Linear(k * feature_size, output_size, bias=enable_bias)
def forward(self, L, x):
X = chebyshev(L, x, self.k)
return self.conv(X)
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'feature_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
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_zeros_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
tmp0 = 0.0
tl.store(out_ptr0 + x0, tmp0, 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,), (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_zeros_0[grid(16)](buf0, 16, XBLOCK=16, num_warps=1,
num_stages=1)
buf1 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf0,
(4, 4), (1, 4), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4
), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), beta=0)
del primals_1
del primals_2
buf2 = buf1
del buf1
buf3 = buf0
del buf0
extern_kernels.addmm(primals_4, reinterpret_tensor(buf2, (4, 4), (1,
4), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), alpha
=1, beta=1, out=buf3)
del primals_3
del primals_4
return buf3, reinterpret_tensor(buf2, (4, 4), (1, 4), 0)
def chebyshev(L, X, k=3):
if k == 1:
return torch.sparse.mm(L, X)
dp = [X, torch.sparse.mm(L, X)]
for i in range(2, k):
nxt = 2 * torch.sparse.mm(L, dp[i - 1])
dp.append(torch.sparse.FloatTensor.add(nxt, -dp[i - 2]))
return torch.cat(dp, dim=1)
class SCNLayerNew(nn.Module):
def __init__(self, feature_size, output_size, enable_bias=True, k=1):
super().__init__()
self.k = k
self.conv = nn.Linear(k * feature_size, output_size, bias=enable_bias)
def forward(self, input_0, input_1):
primals_1 = self.conv.weight
primals_4 = self.conv.bias
primals_2 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| ggoh29/Simplicial-neural-network-benchmark | SCNLayer | false | 6,733 | [
"MIT"
] | 1 | 9a12bcd054251790d85e3971f5473dcffaa5664b | https://github.com/ggoh29/Simplicial-neural-network-benchmark/tree/9a12bcd054251790d85e3971f5473dcffaa5664b | import torch
import torch.nn as nn
def chebyshev(L, X, k=3):
if k == 1:
return torch.sparse.mm(L, X)
dp = [X, torch.sparse.mm(L, X)]
for i in range(2, k):
nxt = 2 * torch.sparse.mm(L, dp[i - 1])
dp.append(torch.sparse.FloatTensor.add(nxt, -dp[i - 2]))
return torch.cat(dp, dim=1)
class Model(nn.Module):
def __init__(self, feature_size, output_size, enable_bias=True, k=1):
super().__init__()
self.k = k
self.conv = nn.Linear(k * feature_size, output_size, bias=enable_bias)
def forward(self, L, x):
X = chebyshev(L, x, self.k)
return self.conv(X)
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [4, 4]
|
Temp | # 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_4/inductor_cache/eq/ceqlzwqlgqnd5jih74vktx4hlpfsbm6lcksehyx7mnuogxj6jtqg.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.leaky_relu]
# Source node to ATen node mapping:
# x => gt, mul, where
# Graph fragment:
# %gt : [num_users=2] = 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, 0.01), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %view_1, %mul), kwargs = {})
triton_poi_fused_leaky_relu_0 = async_compile.triton('triton_poi_fused_leaky_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=[16384],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*i1', 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_leaky_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_leaky_relu_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
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_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr1 + (x0), None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.01
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tl.store(out_ptr0 + (x2), tmp4, None)
tl.store(out_ptr1 + (x2), tmp7, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/2p/c2pyneono2s7z7jcciz35lx44jj7lfjdiqgca5oj636u4ytl7k6i.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# x_1 => sigmoid
# Graph fragment:
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%view_3,), kwargs = {})
triton_poi_fused_sigmoid_1 = async_compile.triton('triton_poi_fused_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=[16384],
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_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_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
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.sigmoid(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, (256, 4), (4, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (256, 256), (256, 1))
assert_size_stride(primals_5, (256, ), (1, ))
assert_size_stride(primals_6, (256, 256), (256, 1))
assert_size_stride(primals_7, (256, ), (1, ))
assert_size_stride(primals_8, (4, 256), (256, 1))
assert_size_stride(primals_9, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 256), (256, 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, 256), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.bool)
buf2 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.leaky_relu]
stream0 = get_raw_stream(0)
triton_poi_fused_leaky_relu_0.run(buf0, primals_2, buf1, buf2, 16384, grid=grid(16384), stream=stream0)
del primals_2
buf3 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf2, (64, 256), (256, 1), 0), reinterpret_tensor(primals_4, (256, 256), (1, 256), 0), out=buf3)
buf4 = reinterpret_tensor(buf3, (4, 4, 4, 256), (4096, 1024, 256, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_1.run(buf4, primals_5, 16384, grid=grid(16384), stream=stream0)
del primals_5
buf5 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf4, (64, 256), (256, 1), 0), reinterpret_tensor(primals_6, (256, 256), (1, 256), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.bool)
buf7 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.leaky_relu]
triton_poi_fused_leaky_relu_0.run(buf5, primals_7, buf6, buf7, 16384, grid=grid(16384), stream=stream0)
del buf5
del primals_7
buf8 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf7, (64, 256), (256, 1), 0), reinterpret_tensor(primals_8, (256, 4), (1, 256), 0), alpha=1, beta=1, out=buf8)
del primals_9
return (reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf1, reinterpret_tensor(buf2, (64, 256), (256, 1), 0), buf4, buf6, reinterpret_tensor(buf7, (64, 256), (256, 1), 0), 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((256, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (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((256, 256), (256, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((256, 256), (256, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 256), (256, 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
import torch.nn.functional as F
class Temp(nn.Module):
def __init__(self, input_dim, output_dim):
super(Temp, self).__init__()
self.linear1 = nn.Linear(input_dim, 256)
self.linear2 = nn.Linear(256, 256)
self.linear3 = nn.Linear(256, 256)
self.linear4 = nn.Linear(256, output_dim)
def forward(self, x):
x = F.leaky_relu(self.linear1(x))
x = F.sigmoid(self.linear2(x))
x = F.leaky_relu(self.linear3(x))
x = self.linear4(x)
return 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
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_0(in_ptr0, in_ptr1, 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)
x2 = xindex
x0 = xindex % 256
tmp0 = tl.load(in_ptr0 + x2, None)
tmp1 = tl.load(in_ptr1 + x0, None, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.01
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tl.store(out_ptr0 + x2, tmp4, None)
tl.store(out_ptr1 + x2, tmp7, None)
@triton.jit
def triton_poi_fused_sigmoid_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)
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.sigmoid(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, (256, 4), (4, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (256, 256), (256, 1))
assert_size_stride(primals_5, (256,), (1,))
assert_size_stride(primals_6, (256, 256), (256, 1))
assert_size_stride(primals_7, (256,), (1,))
assert_size_stride(primals_8, (4, 256), (256, 1))
assert_size_stride(primals_9, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 256), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.bool)
buf2 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.float32)
get_raw_stream(0)
triton_poi_fused_leaky_relu_0[grid(16384)](buf0, primals_2, buf1,
buf2, 16384, XBLOCK=256, num_warps=4, num_stages=1)
del primals_2
buf3 = buf0
del buf0
extern_kernels.mm(reinterpret_tensor(buf2, (64, 256), (256, 1), 0),
reinterpret_tensor(primals_4, (256, 256), (1, 256), 0), out=buf3)
buf4 = reinterpret_tensor(buf3, (4, 4, 4, 256), (4096, 1024, 256, 1), 0
)
del buf3
triton_poi_fused_sigmoid_1[grid(16384)](buf4, primals_5, 16384,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_5
buf5 = empty_strided_cuda((64, 256), (256, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf4, (64, 256), (256, 1), 0),
reinterpret_tensor(primals_6, (256, 256), (1, 256), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.bool)
buf7 = empty_strided_cuda((4, 4, 4, 256), (4096, 1024, 256, 1),
torch.float32)
triton_poi_fused_leaky_relu_0[grid(16384)](buf5, primals_7, buf6,
buf7, 16384, XBLOCK=256, num_warps=4, num_stages=1)
del buf5
del primals_7
buf8 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_9, reinterpret_tensor(buf7, (64, 256),
(256, 1), 0), reinterpret_tensor(primals_8, (256, 4), (1, 256),
0), alpha=1, beta=1, out=buf8)
del primals_9
return reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf1, reinterpret_tensor(buf2, (64, 256), (256, 1), 0
), buf4, buf6, reinterpret_tensor(buf7, (64, 256), (256, 1), 0
), primals_8, primals_6, primals_4
class TempNew(nn.Module):
def __init__(self, input_dim, output_dim):
super(TempNew, self).__init__()
self.linear1 = nn.Linear(input_dim, 256)
self.linear2 = nn.Linear(256, 256)
self.linear3 = nn.Linear(256, 256)
self.linear4 = nn.Linear(256, output_dim)
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.linear3.weight
primals_7 = self.linear3.bias
primals_8 = self.linear4.weight
primals_9 = self.linear4.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]
| gautam-sharma1/openRL | Temp | false | 6,734 | [
"MIT"
] | 1 | 14310a97a328fe5682a01ee85d83a6b5e1ae29ca | https://github.com/gautam-sharma1/openRL/tree/14310a97a328fe5682a01ee85d83a6b5e1ae29ca | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, input_dim, output_dim):
super().__init__()
self.linear1 = nn.Linear(input_dim, 256)
self.linear2 = nn.Linear(256, 256)
self.linear3 = nn.Linear(256, 256)
self.linear4 = nn.Linear(256, output_dim)
def forward(self, x):
x = F.leaky_relu(self.linear1(x))
x = F.sigmoid(self.linear2(x))
x = F.leaky_relu(self.linear3(x))
x = self.linear4(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
GroupNorm | # 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_4/inductor_cache/gz/cgz3n2vpsv7nvzf5brt64krpzobahh2rku4tjvkm4coqb2auptpq.py
# Topologically Sorted Source Nodes: [mul, add_1], Original ATen: [aten.mul, aten.add]
# Source node to ATen node mapping:
# add_1 => add_1
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %primals_2), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %primals_3), kwargs = {})
triton_poi_fused_add_mul_0 = async_compile.triton('triton_poi_fused_add_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: '*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_0', '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_mul_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
x4 = xindex
x5 = (xindex // 4)
x2 = (xindex // 16) % 4
tmp0 = tl.load(in_ptr0 + (x4), xmask)
tmp1 = tl.load(in_ptr0 + (4*x5), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (4*x5)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (4*x5)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + (4*x5)), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + (x2), xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 4.0
tmp9 = tmp7 / tmp8
tmp10 = tmp0 - tmp9
tmp11 = tmp1 - tmp9
tmp12 = tmp11 * tmp11
tmp13 = tmp2 - tmp9
tmp14 = tmp13 * tmp13
tmp15 = tmp12 + tmp14
tmp16 = tmp4 - tmp9
tmp17 = tmp16 * tmp16
tmp18 = tmp15 + tmp17
tmp19 = tmp6 - tmp9
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = 3.0
tmp23 = tmp21 / tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = 1e-10
tmp26 = tmp24 + tmp25
tmp27 = tmp10 / tmp26
tmp29 = tmp27 * tmp28
tmp31 = tmp29 + tmp30
tl.store(out_ptr0 + (x4), tmp31, 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, 1), (1, 1, 1))
assert_size_stride(primals_3, (4, 1, 1), (1, 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: [mul, add_1], Original ATen: [aten.mul, aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_mul_0.run(primals_1, primals_2, primals_3, buf0, 256, grid=grid(256), stream=stream0)
del primals_2
del primals_3
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, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 1, 1), (1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((4, 1, 1), (1, 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 GroupNorm(nn.Module):
def __init__(self, c_num, group_num=16, eps=1e-10):
"""
The groupnorm layer from https://arxiv.org/abs/1803.08494
Args:
c_num (int): Number of input channels
group_num (int): Number of group by which to divide the channels
eps (float): Epsilon
"""
super(GroupNorm, self).__init__()
self.group_num = group_num
self.gamma = nn.Parameter(torch.ones(c_num, 1, 1))
self.beta = nn.Parameter(torch.zeros(c_num, 1, 1))
self.eps = eps
def forward(self, x):
batch_size, channels, height, width = x.size()
x = x.view(batch_size, self.group_num, -1)
mean = x.mean(dim=2, keepdim=True)
std = x.std(dim=2, keepdim=True)
x = (x - mean) / (std + self.eps)
x = x.view(batch_size, channels, height, width)
return x * self.gamma + self.beta
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'c_num': 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
@triton.jit
def triton_poi_fused_add_mul_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
x4 = xindex
x5 = xindex // 4
x2 = xindex // 16 % 4
tmp0 = tl.load(in_ptr0 + x4, xmask)
tmp1 = tl.load(in_ptr0 + 4 * x5, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 4 * x5), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 4 * x5), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (3 + 4 * x5), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + x2, xmask, eviction_policy='evict_last')
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = 4.0
tmp9 = tmp7 / tmp8
tmp10 = tmp0 - tmp9
tmp11 = tmp1 - tmp9
tmp12 = tmp11 * tmp11
tmp13 = tmp2 - tmp9
tmp14 = tmp13 * tmp13
tmp15 = tmp12 + tmp14
tmp16 = tmp4 - tmp9
tmp17 = tmp16 * tmp16
tmp18 = tmp15 + tmp17
tmp19 = tmp6 - tmp9
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = 3.0
tmp23 = tmp21 / tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = 1e-10
tmp26 = tmp24 + tmp25
tmp27 = tmp10 / tmp26
tmp29 = tmp27 * tmp28
tmp31 = tmp29 + tmp30
tl.store(out_ptr0 + x4, tmp31, 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, 1), (1, 1, 1))
assert_size_stride(primals_3, (4, 1, 1), (1, 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_add_mul_0[grid(256)](primals_1, primals_2,
primals_3, buf0, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
del primals_3
return buf0, primals_1
class GroupNormNew(nn.Module):
def __init__(self, c_num, group_num=16, eps=1e-10):
"""
The groupnorm layer from https://arxiv.org/abs/1803.08494
Args:
c_num (int): Number of input channels
group_num (int): Number of group by which to divide the channels
eps (float): Epsilon
"""
super(GroupNormNew, self).__init__()
self.group_num = group_num
self.gamma = nn.Parameter(torch.ones(c_num, 1, 1))
self.beta = nn.Parameter(torch.zeros(c_num, 1, 1))
self.eps = eps
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]
| generall/Torchlite | GroupNorm | false | 6,735 | [
"MIT"
] | 1 | 2eb3e2a20b7619bd58b0b0fca120e2aefca0e79a | https://github.com/generall/Torchlite/tree/2eb3e2a20b7619bd58b0b0fca120e2aefca0e79a | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, c_num, group_num=16, eps=1e-10):
"""
The groupnorm layer from https://arxiv.org/abs/1803.08494
Args:
c_num (int): Number of input channels
group_num (int): Number of group by which to divide the channels
eps (float): Epsilon
"""
super().__init__()
self.group_num = group_num
self.gamma = nn.Parameter(torch.ones(c_num, 1, 1))
self.beta = nn.Parameter(torch.zeros(c_num, 1, 1))
self.eps = eps
def forward(self, x):
batch_size, channels, height, width = x.size()
x = x.view(batch_size, self.group_num, -1)
mean = x.mean(dim=2, keepdim=True)
std = x.std(dim=2, keepdim=True)
x = (x - mean) / (std + self.eps)
x = x.view(batch_size, channels, height, width)
return x * self.gamma + self.beta
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
LinearAttention | # 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_4/inductor_cache/ud/cudbnoor4dpxxdmeqssolgpxcs76nh6jnfj2xf7auyk3sidsohnf.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_1, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_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=[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_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 = 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_4/inductor_cache/sx/csxi2lkjrgsyc43k5vauxutocx5sik4gnrglbjfjmvytt3alfq7w.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=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=[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 = 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_4/inductor_cache/kp/ckpt5cgd2ge6wf33wolxa3xllh3mqghcffsloi7ivxt46oqj5pri.py
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.sum]
# Source node to ATen node mapping:
# mul => mul
# x => sum_2
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expand, %primals_3), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
triton_poi_fused_mul_sum_2 = async_compile.triton('triton_poi_fused_mul_sum_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: '*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_2', '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_2(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
x1 = (xindex // 4) % 4
x2 = (xindex // 16)
x3 = xindex % 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x3 + (64*x2)), xmask)
tmp3 = tl.load(in_ptr0 + (4 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (16 + x3 + (64*x2)), xmask)
tmp7 = tl.load(in_ptr0 + (8 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (32 + x3 + (64*x2)), xmask)
tmp11 = tl.load(in_ptr0 + (12 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (48 + x3 + (64*x2)), xmask)
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 = args
args.clear()
assert_size_stride(primals_1, (1, 4), (4, 1))
assert_size_stride(primals_2, (1, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_1
del primals_2
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(buf1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf2, buf3, 64, grid=grid(64), stream=stream0)
buf4 = reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.sum]
triton_poi_fused_mul_sum_2.run(buf3, primals_3, buf4, 64, grid=grid(64), stream=stream0)
return (buf4, buf3, primals_3, 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((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((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)
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 LinearAttention(nn.Module):
def __init__(self, in_size):
super(LinearAttention, self).__init__()
self.out = nn.Linear(in_size, 1)
nn.init.orthogonal_(self.out.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input):
self.alpha = self.softmax(self.out(input))
x = (self.alpha.expand_as(input) * input).sum(dim=1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_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
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__softmax_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
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_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 = 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_2(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
x1 = xindex // 4 % 4
x2 = xindex // 16
x3 = xindex % 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + (x3 + 64 * x2), xmask)
tmp3 = tl.load(in_ptr0 + (4 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr1 + (16 + x3 + 64 * x2), xmask)
tmp7 = tl.load(in_ptr0 + (8 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp8 = tl.load(in_ptr1 + (32 + x3 + 64 * x2), xmask)
tmp11 = tl.load(in_ptr0 + (12 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp12 = tl.load(in_ptr1 + (48 + x3 + 64 * x2), xmask)
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 = args
args.clear()
assert_size_stride(primals_1, (1, 4), (4, 1))
assert_size_stride(primals_2, (1,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 1), (1, 4), 0
), alpha=1, beta=1, out=buf1)
del primals_1
del primals_2
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(64)](buf1, buf2, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf3 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused__softmax_1[grid(64)](buf2, buf3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf4 = reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0)
del buf2
triton_poi_fused_mul_sum_2[grid(64)](buf3, primals_3, buf4, 64,
XBLOCK=64, num_warps=1, num_stages=1)
return buf4, buf3, primals_3, buf3
class LinearAttentionNew(nn.Module):
def __init__(self, in_size):
super(LinearAttentionNew, self).__init__()
self.out = nn.Linear(in_size, 1)
nn.init.orthogonal_(self.out.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input_0):
primals_1 = self.out.weight
primals_2 = self.out.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| gchrupala/platalea | LinearAttention | false | 6,736 | [
"Apache-2.0"
] | 1 | 65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | https://github.com/gchrupala/platalea/tree/65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def __init__(self, in_size):
super().__init__()
self.out = nn.Linear(in_size, 1)
nn.init.orthogonal_(self.out.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input):
self.alpha = self.softmax(self.out(input))
x = (self.alpha.expand_as(input) * input).sum(dim=1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
GreedyCTCDecoder | # 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_4/inductor_cache/fl/cflouszn4d5pmmt3birh6hthvba6xg5vypczyfvcr7hxqs7k2c7z.py
# Topologically Sorted Source Nodes: [argmax, argmx], Original ATen: [aten.argmax, aten._to_copy]
# Source node to ATen node mapping:
# argmax => argmax
# argmx => convert_element_type
# Graph fragment:
# %argmax : [num_users=1] = call_function[target=torch.ops.aten.argmax.default](args = (%arg0_1, -1), kwargs = {})
# %convert_element_type : [num_users=1] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%argmax, torch.int32), kwargs = {})
triton_poi_fused__to_copy_argmax_0 = async_compile.triton('triton_poi_fused__to_copy_argmax_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: '*i32', 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_argmax_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__to_copy_argmax_0(in_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')
tmp17 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
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 = tmp46.to(tl.int32)
tl.store(out_ptr1 + (x0), tmp47, 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)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.int32)
# Topologically Sorted Source Nodes: [argmax, argmx], Original ATen: [aten.argmax, aten._to_copy]
stream0 = get_raw_stream(0)
triton_poi_fused__to_copy_argmax_0.run(arg0_1, buf1, 64, grid=grid(64), 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
class GreedyCTCDecoder(nn.Module):
""" Greedy CTC Decoder
"""
def __init__(self, **kwargs):
nn.Module.__init__(self)
def forward(self, log_probs):
with torch.no_grad():
argmx = log_probs.argmax(dim=-1, keepdim=False).int()
return argmx
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__to_copy_argmax_0(in_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')
tmp17 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp32 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
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 = tmp46.to(tl.int32)
tl.store(out_ptr1 + x0, tmp47, 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)
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.int32)
get_raw_stream(0)
triton_poi_fused__to_copy_argmax_0[grid(64)](arg0_1, buf1, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
return buf1,
class GreedyCTCDecoderNew(nn.Module):
""" Greedy CTC Decoder
"""
def __init__(self, **kwargs):
nn.Module.__init__(self)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| ghdrl95/Naver-Speech-Hackathon | GreedyCTCDecoder | false | 6,737 | [
"Apache-2.0"
] | 1 | 10b4526d98ce535415cb91d24338790d9c175b63 | https://github.com/ghdrl95/Naver-Speech-Hackathon/tree/10b4526d98ce535415cb91d24338790d9c175b63 | import torch
import torch.nn as nn
class Model(nn.Module):
""" Greedy CTC Decoder
"""
def __init__(self, **kwargs):
nn.Module.__init__(self)
def forward(self, log_probs):
with torch.no_grad():
argmx = log_probs.argmax(dim=-1, keepdim=False).int()
return argmx
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
RegionPenaltyLoss | # 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_4/inductor_cache/5x/c5xxthxs5ccf2qzrrz6iexmfuocmiqozcnvhzn2towb5tno7vjnn.py
# Topologically Sorted Source Nodes: [max_1, sub, penalty, sum_1], Original ATen: [aten.max, aten.sub, aten.abs, aten.sum]
# Source node to ATen node mapping:
# max_1 => max_1
# penalty => abs_1
# sub => sub
# sum_1 => sum_1
# Graph fragment:
# %max_1 : [num_users=1] = call_function[target=torch.ops.aten.max.default](args = (%arg1_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg1_1, %max_1), kwargs = {})
# %abs_1 : [num_users=2] = call_function[target=torch.ops.aten.abs.default](args = (%sub,), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%abs_1,), kwargs = {})
triton_per_fused_abs_max_sub_sum_0 = async_compile.triton('triton_per_fused_abs_max_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_abs_max_sub_sum_0', 'mutated_arg_names': [], 'no_x_dim': True, 'num_load': 1, '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_abs_max_sub_sum_0(in_ptr0, out_ptr0, out_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.broadcast_to(tmp0, [RBLOCK])
tmp3 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp1, 0))
tmp4 = tmp0 - tmp3
tmp5 = tl_math.abs(tmp4)
tmp6 = tl.broadcast_to(tmp5, [RBLOCK])
tmp8 = triton_helpers.promote_to_tensor(tl.sum(tmp6, 0))
tl.store(out_ptr0 + (tl.full([1], 0, tl.int32)), tmp3, None)
tl.store(out_ptr1 + (tl.full([1], 0, tl.int32)), tmp8, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/h7/ch74yqfi4ediwkonn4zsaw2lllhh4zlb574jitl7kfkfnnxz7i6b.py
# Topologically Sorted Source Nodes: [sum_2, mul_1], Original ATen: [aten.sum, aten.mul]
# Source node to ATen node mapping:
# mul_1 => mul_1
# sum_2 => sum_2
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view, [1]), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_2, 1.0), kwargs = {})
triton_per_fused_mul_sum_1 = async_compile.triton('triton_per_fused_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=[4, 64],
reduction_hint=ReductionHint.INNER,
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_mul_sum_1', 'mutated_arg_names': ['in_out_ptr0'], '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_mul_sum_1(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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 + (r1 + (64*x0)), xmask, other=0.0)
tmp2 = tl.load(in_ptr2 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp6 = tl.load(in_ptr3 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp4 = tmp1 - tmp3
tmp5 = tl_math.abs(tmp4)
tmp8 = tmp5 / tmp7
tmp9 = tmp0 * tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tmp14 = 1.0
tmp15 = tmp13 * tmp14
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp15, 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((), (), torch.float32)
buf1 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [max_1, sub, penalty, sum_1], Original ATen: [aten.max, aten.sub, aten.abs, aten.sum]
stream0 = get_raw_stream(0)
triton_per_fused_abs_max_sub_sum_0.run(arg1_1, buf0, buf1, 1, 256, grid=grid(1), stream=stream0)
buf2 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [sum_2, mul_1], Original ATen: [aten.sum, aten.mul]
triton_per_fused_mul_sum_1.run(buf3, arg0_1, arg1_1, buf0, buf1, 4, 64, grid=grid(4), stream=stream0)
del arg0_1
del arg1_1
del buf0
del buf1
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
from torch import nn
class RegionPenaltyLoss(nn.Module):
def __init__(self, scale=1.0):
"""
Multiplicative penalty.
Penalizes "forbidden" regions instead of exact distribution matches.
Optionally used in tandem with MTCrossEntropyRegionAwareLoss.
`scale` param allows caller to scale the loss in order to match
magnitude of other loss terms
"""
super().__init__()
self.scale = scale
def forward(self, preds, targets):
"""
"""
batch_size = preds.shape[0]
penalty = torch.abs(targets - targets.max())
penalty /= torch.sum(penalty)
loss = preds * penalty
loss = loss.view(batch_size, -1)
return torch.sum(loss, dim=1) * self.scale
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
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_per_fused_abs_max_sub_sum_0(in_ptr0, out_ptr0, out_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.broadcast_to(tmp0, [RBLOCK])
tmp3 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp1, 0))
tmp4 = tmp0 - tmp3
tmp5 = tl_math.abs(tmp4)
tmp6 = tl.broadcast_to(tmp5, [RBLOCK])
tmp8 = triton_helpers.promote_to_tensor(tl.sum(tmp6, 0))
tl.store(out_ptr0 + tl.full([1], 0, tl.int32), tmp3, None)
tl.store(out_ptr1 + tl.full([1], 0, tl.int32), tmp8, None)
@triton.jit
def triton_per_fused_mul_sum_1(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2,
in_ptr3, 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 + (r1 + 64 * x0), xmask, other=0.0)
tmp2 = tl.load(in_ptr2 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp6 = tl.load(in_ptr3 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp4 = tmp1 - tmp3
tmp5 = tl_math.abs(tmp4)
tmp8 = tmp5 / tmp7
tmp9 = tmp0 * tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tmp14 = 1.0
tmp15 = tmp13 * tmp14
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp15, 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((), (), torch.float32)
buf1 = empty_strided_cuda((), (), torch.float32)
get_raw_stream(0)
triton_per_fused_abs_max_sub_sum_0[grid(1)](arg1_1, buf0, buf1, 1,
256, num_warps=2, num_stages=1)
buf2 = empty_strided_cuda((4,), (1,), torch.float32)
buf3 = buf2
del buf2
triton_per_fused_mul_sum_1[grid(4)](buf3, arg0_1, arg1_1, buf0,
buf1, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
del buf0
del buf1
return buf3,
class RegionPenaltyLossNew(nn.Module):
def __init__(self, scale=1.0):
"""
Multiplicative penalty.
Penalizes "forbidden" regions instead of exact distribution matches.
Optionally used in tandem with MTCrossEntropyRegionAwareLoss.
`scale` param allows caller to scale the loss in order to match
magnitude of other loss terms
"""
super().__init__()
self.scale = scale
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| geoffreyangus/pet-ct | RegionPenaltyLoss | false | 6,738 | [
"Apache-2.0"
] | 1 | fa96a07734afade475f6a1e1587ec14965fe2de3 | https://github.com/geoffreyangus/pet-ct/tree/fa96a07734afade475f6a1e1587ec14965fe2de3 | import torch
from torch import nn
class Model(nn.Module):
def __init__(self, scale=1.0):
"""
Multiplicative penalty.
Penalizes "forbidden" regions instead of exact distribution matches.
Optionally used in tandem with MTCrossEntropyRegionAwareLoss.
`scale` param allows caller to scale the loss in order to match
magnitude of other loss terms
"""
super().__init__()
self.scale = scale
def forward(self, preds, targets):
"""
"""
batch_size = preds.shape[0]
penalty = torch.abs(targets - targets.max())
penalty /= torch.sum(penalty)
loss = preds * penalty
loss = loss.view(batch_size, -1)
return torch.sum(loss, dim=1) * self.scale
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
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_4/inductor_cache/5y/c5yq7wkgmmcygrawripwacy566sggsmh2mzk5izw35wk7ferohhu.py
# Topologically Sorted Source Nodes: [layer_1_output], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# layer_1_output => 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=[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 = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
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_4/inductor_cache/ld/cld3palaxnfc7osbafgwkvp6zk52xgffwvw26mvcpxisjgbhf4qn.py
# Topologically Sorted Source Nodes: [layer_2_output], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# layer_2_output => 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=[16384],
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 = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 200
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 = args
args.clear()
assert_size_stride(primals_1, (100, 4), (4, 1))
assert_size_stride(primals_2, (100, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (200, 100), (100, 1))
assert_size_stride(primals_5, (200, ), (1, ))
assert_size_stride(primals_6, (4, 200), (200, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 100), (100, 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, 100), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 100), (1600, 400, 100, 1), 0); del buf0 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1), torch.bool)
# Topologically Sorted Source Nodes: [layer_1_output], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, primals_2, buf6, 6400, grid=grid(6400), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 200), (200, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf1, (64, 100), (100, 1), 0), reinterpret_tensor(primals_4, (100, 200), (1, 100), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 200), (3200, 800, 200, 1), 0); del buf2 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1), torch.bool)
# Topologically Sorted Source Nodes: [layer_2_output], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf3, primals_5, buf5, 12800, grid=grid(12800), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 200), (200, 1), 0), reinterpret_tensor(primals_6, (200, 4), (1, 200), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 100), (100, 1), 0), reinterpret_tensor(buf3, (64, 200), (200, 1), 0), primals_6, buf5, 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((100, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((100, ), (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((200, 100), (100, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((200, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 200), (200, 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
class Network(torch.nn.Module):
def __init__(self, input_dimension, output_dimension):
super(Network, self).__init__()
self.layer_1 = torch.nn.Linear(in_features=input_dimension,
out_features=100)
self.layer_2 = torch.nn.Linear(in_features=100, out_features=200)
self.output_layer = torch.nn.Linear(in_features=200, out_features=
output_dimension)
self.relu = torch.nn.ReLU()
def forward(self, input):
layer_1_output = self.relu(self.layer_1(input))
layer_2_output = self.relu(self.layer_2(layer_1_output))
output = self.output_layer(layer_2_output)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dimension': 4, 'output_dimension': 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
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 = 6400
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 100
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_out_ptr0, in_ptr0,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 12800
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 200
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) = args
args.clear()
assert_size_stride(primals_1, (100, 4), (4, 1))
assert_size_stride(primals_2, (100,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (200, 100), (100, 1))
assert_size_stride(primals_5, (200,), (1,))
assert_size_stride(primals_6, (4, 200), (200, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 100), (100, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 100), (1, 4), 0), out=buf0)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4, 100), (1600, 400, 100, 1), 0)
del buf0
buf6 = empty_strided_cuda((4, 4, 4, 100), (1664, 400, 100, 1),
torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(6400)](buf1,
primals_2, buf6, 6400, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 200), (200, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 100), (100, 1), 0),
reinterpret_tensor(primals_4, (100, 200), (1, 100), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 200), (3200, 800, 200, 1), 0)
del buf2
buf5 = empty_strided_cuda((4, 4, 4, 200), (3200, 800, 200, 1),
torch.bool)
triton_poi_fused_relu_threshold_backward_1[grid(12800)](buf3,
primals_5, buf5, 12800, 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, 200),
(200, 1), 0), reinterpret_tensor(primals_6, (200, 4), (1, 200),
0), alpha=1, beta=1, out=buf4)
del primals_7
return reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 100), (100, 1), 0
), reinterpret_tensor(buf3, (64, 200), (200, 1), 0
), primals_6, buf5, primals_4, buf6
class NetworkNew(torch.nn.Module):
def __init__(self, input_dimension, output_dimension):
super(NetworkNew, self).__init__()
self.layer_1 = torch.nn.Linear(in_features=input_dimension,
out_features=100)
self.layer_2 = torch.nn.Linear(in_features=100, out_features=200)
self.output_layer = torch.nn.Linear(in_features=200, out_features=
output_dimension)
self.relu = torch.nn.ReLU()
def forward(self, input_0):
primals_1 = self.layer_1.weight
primals_2 = self.layer_1.bias
primals_4 = self.layer_2.weight
primals_5 = self.layer_2.bias
primals_6 = self.output_layer.weight
primals_7 = self.output_layer.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| georgeyiasemis/Deep-Reinforcement-Learning-on-a-Grid-World- | Network | false | 6,739 | [
"MIT"
] | 1 | f32ceac5f4e29cba212d6fd1b8a25c08ac733666 | https://github.com/georgeyiasemis/Deep-Reinforcement-Learning-on-a-Grid-World-/tree/f32ceac5f4e29cba212d6fd1b8a25c08ac733666 | import torch
class Model(torch.nn.Module):
def __init__(self, input_dimension, output_dimension):
super().__init__()
self.layer_1 = torch.nn.Linear(in_features=input_dimension,
out_features=100)
self.layer_2 = torch.nn.Linear(in_features=100, out_features=200)
self.output_layer = torch.nn.Linear(in_features=200, out_features=
output_dimension)
self.relu = torch.nn.ReLU()
def forward(self, input):
layer_1_output = self.relu(self.layer_1(input))
layer_2_output = self.relu(self.layer_2(layer_1_output))
output = self.output_layer(layer_2_output)
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ScalarAttention | # 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_4/inductor_cache/nc/cncwsucylpsg2zmlivjfxu6vbd64ztxjndlsix2ysjtby3xohgk4.py
# Topologically Sorted Source Nodes: [tanh], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# tanh => 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=[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': ['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 = 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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/dm/cdmkcxuzpnailvibeivaikqdr4zvashgzwju7qijhq5aizlo3aor.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_3, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %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_4/inductor_cache/kt/cktghousutx6xui2sl2rvevzmb7gkacvfhntjq5n2xzeu7v57oz6.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=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_4/inductor_cache/fj/cfjxsjft32uejuiclvkwroildivviknko2qfji2c5dxkqu3gx26a.py
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.sum]
# Source node to ATen node mapping:
# mul => mul
# x => sum_2
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%expand, %primals_3), 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
x1 = (xindex // 4) % 4
x2 = (xindex // 16)
x3 = xindex % 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x3 + (64*x2)), xmask)
tmp3 = tl.load(in_ptr0 + (4 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (16 + x3 + (64*x2)), xmask)
tmp7 = tl.load(in_ptr0 + (8 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (32 + x3 + (64*x2)), xmask)
tmp11 = tl.load(in_ptr0 + (12 + x1 + (16*x2)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (48 + x3 + (64*x2)), xmask)
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 = 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, (1, 4), (4, 1))
assert_size_stride(primals_5, (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: [], 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
# Topologically Sorted Source Nodes: [tanh], Original ATen: [aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_tanh_0.run(buf1, primals_2, 256, grid=grid(256), stream=stream0)
del primals_2
buf3 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf3)
del primals_5
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [softmax], 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: [softmax], 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: [mul, x], Original ATen: [aten.mul, aten.sum]
triton_poi_fused_mul_sum_3.run(buf5, primals_3, buf6, 64, grid=grid(64), stream=stream0)
return (buf6, buf5, primals_3, buf1, buf5, 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, 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((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, ), (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 ScalarAttention(nn.Module):
def __init__(self, in_size, hidden_size):
super(ScalarAttention, self).__init__()
self.hidden = nn.Linear(in_size, hidden_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.out = nn.Linear(hidden_size, 1)
nn.init.orthogonal_(self.hidden.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input):
self.alpha = self.softmax(self.out(torch.tanh(self.hidden(input))))
x = (self.alpha.expand_as(input) * input).sum(dim=1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_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 import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, 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
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
):
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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, 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
x1 = xindex // 4 % 4
x2 = xindex // 16
x3 = xindex % 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr1 + (x3 + 64 * x2), xmask)
tmp3 = tl.load(in_ptr0 + (4 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr1 + (16 + x3 + 64 * x2), xmask)
tmp7 = tl.load(in_ptr0 + (8 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp8 = tl.load(in_ptr1 + (32 + x3 + 64 * x2), xmask)
tmp11 = tl.load(in_ptr0 + (12 + x1 + 16 * x2), xmask, eviction_policy=
'evict_last')
tmp12 = tl.load(in_ptr1 + (48 + x3 + 64 * x2), xmask)
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 = 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, (1, 4), (4, 1))
assert_size_stride(primals_5, (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_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
get_raw_stream(0)
triton_poi_fused_tanh_0[grid(256)](buf1, primals_2, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_2
buf3 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf1, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_4, (4, 1), (1, 4), 0),
alpha=1, beta=1, out=buf3)
del primals_5
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)](buf5, primals_3, buf6, 64,
XBLOCK=64, num_warps=1, num_stages=1)
return buf6, buf5, primals_3, buf1, buf5, primals_4
class ScalarAttentionNew(nn.Module):
def __init__(self, in_size, hidden_size):
super(ScalarAttentionNew, self).__init__()
self.hidden = nn.Linear(in_size, hidden_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.out = nn.Linear(hidden_size, 1)
nn.init.orthogonal_(self.hidden.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input_0):
primals_1 = self.hidden.weight
primals_2 = self.hidden.bias
primals_4 = self.out.weight
primals_5 = self.out.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| gchrupala/platalea | ScalarAttention | false | 6,740 | [
"Apache-2.0"
] | 1 | 65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | https://github.com/gchrupala/platalea/tree/65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def __init__(self, in_size, hidden_size):
super().__init__()
self.hidden = nn.Linear(in_size, hidden_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.out = nn.Linear(hidden_size, 1)
nn.init.orthogonal_(self.hidden.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input):
self.alpha = self.softmax(self.out(torch.tanh(self.hidden(input))))
x = (self.alpha.expand_as(input) * input).sum(dim=1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ODEfunc | # 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_4/inductor_cache/ew/cewlcpr2jhkktbpmzbbjxdsiykdntmypm237lc34qynaxm2ln5ee.py
# Topologically Sorted Source Nodes: [out, out_1], Original ATen: [aten.native_group_norm, aten.relu]
# Source node to ATen node mapping:
# out => add, add_1, mul_1, rsqrt, var_mean
# out_1 => relu
# 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=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %unsqueeze_2), kwargs = {})
# %relu : [num_users=1] = call_function[target=torch.ops.aten.relu.default](args = (%add_1,), kwargs = {})
triton_per_fused_native_group_norm_relu_0 = async_compile.triton('triton_per_fused_native_group_norm_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: '*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, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_native_group_norm_relu_0', 'mutated_arg_names': ['in_out_ptr0'], '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_native_group_norm_relu_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, 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
x2 = xindex % 4
x3 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (r1 + (16*x0)), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + (x2), xmask, 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], 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 = 16.0
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 = tl.full([1, 1], 0, tl.int32)
tmp29 = triton_helpers.maximum(tmp28, tmp27)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp21, xmask)
tl.store(out_ptr1 + (r1 + (16*x2) + (80*x3)), tmp29, xmask)
tl.store(out_ptr0 + (x0), tmp10, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/yl/cyltj4xe7bwa5jmotmsxfdzwedvvrytkhaf3f2qw62sd4zn5rnro.py
# Topologically Sorted Source Nodes: [ttx, ttx_1], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# ttx => cat
# ttx_1 => cat_1
# Graph fragment:
# %cat : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_4, %relu], 1), kwargs = {})
# %cat_1 : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_4, %relu_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=[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_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, 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
x0 = xindex % 16
x1 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tl.store(out_ptr0 + (x0 + (80*x1)), tmp0, xmask)
tl.store(out_ptr1 + (x0 + (80*x1)), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/mr/cmr56lkwxw77qikvfa54yx4b56plsu5zod4pwpjjr4x2wgpvy3h6.py
# Topologically Sorted Source Nodes: [out_2, out_3, out_4], Original ATen: [aten.convolution, aten.native_group_norm, aten.relu]
# Source node to ATen node mapping:
# out_2 => convolution
# out_3 => add_2, add_3, mul_4, rsqrt_1, var_mean_1
# out_4 => relu_1
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%cat, %primals_5, %primals_6, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_2, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add_2 : [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_2,), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, %unsqueeze_11), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_4, %unsqueeze_8), kwargs = {})
# %relu_1 : [num_users=1] = call_function[target=torch.ops.aten.relu.default](args = (%add_3,), kwargs = {})
triton_per_fused_convolution_native_group_norm_relu_2 = async_compile.triton('triton_per_fused_convolution_native_group_norm_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.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32', 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_per_fused_convolution_native_group_norm_relu_2', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 4, '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_convolution_native_group_norm_relu_2(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
x1 = (xindex // 4)
tmp0 = tl.load(in_out_ptr0 + (r2 + (16*x3)), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 16.0
tmp20 = tmp18 / tmp19
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tmp24 = tmp2 - tmp12
tmp25 = tmp24 * tmp23
tmp27 = tmp25 * tmp26
tmp29 = tmp27 + tmp28
tmp30 = tl.full([1, 1], 0, tl.int32)
tmp31 = triton_helpers.maximum(tmp30, tmp29)
tl.store(in_out_ptr0 + (r2 + (16*x3)), tmp2, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x3), tmp23, xmask)
tl.store(out_ptr1 + (r2 + (16*x0) + (80*x1)), tmp31, xmask)
tl.store(out_ptr0 + (x3), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/lj/cljbnzt4e5mf4f235sbsd7nao5p35wmgsn35efjytvld4hyxvgz4.py
# Topologically Sorted Source Nodes: [out_5, out_6], Original ATen: [aten.convolution, aten.native_group_norm]
# Source node to ATen node mapping:
# out_5 => convolution_1
# out_6 => add_4, add_5, mul_7, rsqrt_2, var_mean_2
# Graph fragment:
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%cat_1, %primals_9, %primals_10, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %var_mean_2 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_4, [2, 3]), kwargs = {correction: 0, keepdim: True})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_4, 1e-05), kwargs = {})
# %rsqrt_2 : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_4,), kwargs = {})
# %mul_7 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_5, %unsqueeze_17), kwargs = {})
# %add_5 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_7, %unsqueeze_14), kwargs = {})
triton_per_fused_convolution_native_group_norm_3 = async_compile.triton('triton_per_fused_convolution_native_group_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.persistent_reduction(
size_hints=[16, 16],
reduction_hint=ReductionHint.INNER,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: 'i32', 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_per_fused_convolution_native_group_norm_3', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 4, '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_convolution_native_group_norm_3(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (r2 + (16*x3)), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = tmp2 - tmp12
tmp20 = 16.0
tmp21 = tmp18 / tmp20
tmp22 = 1e-05
tmp23 = tmp21 + tmp22
tmp24 = libdevice.rsqrt(tmp23)
tmp25 = tmp19 * tmp24
tmp27 = tmp25 * tmp26
tmp29 = tmp27 + tmp28
tl.store(in_out_ptr0 + (r2 + (16*x3)), tmp2, xmask)
tl.store(out_ptr2 + (r2 + (16*x3)), tmp29, xmask)
tl.store(out_ptr3 + (x3), tmp24, xmask)
tl.store(out_ptr0 + (x3), tmp12, 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 = args
args.clear()
assert_size_stride(primals_1, (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, 4, 4), (16, 16, 4, 1))
assert_size_stride(primals_5, (4, 5, 3, 3), (45, 9, 3, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, ), (1, ))
assert_size_stride(primals_9, (4, 5, 3, 3), (45, 9, 3, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, ), (1, ))
assert_size_stride(primals_12, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
buf1 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf3 = reinterpret_tensor(buf1, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf1 # reuse
buf6 = empty_strided_cuda((4, 5, 4, 4), (80, 16, 4, 1), torch.float32)
buf5 = reinterpret_tensor(buf6, (4, 4, 4, 4), (80, 16, 4, 1), 16) # alias
# Topologically Sorted Source Nodes: [out, out_1], Original ATen: [aten.native_group_norm, aten.relu]
stream0 = get_raw_stream(0)
triton_per_fused_native_group_norm_relu_0.run(buf3, primals_3, primals_1, primals_2, buf0, buf5, 16, 16, grid=grid(16), stream=stream0)
buf4 = reinterpret_tensor(buf6, (4, 1, 4, 4), (80, 16, 4, 1), 0) # alias
buf15 = empty_strided_cuda((4, 5, 4, 4), (80, 16, 4, 1), torch.float32)
buf13 = reinterpret_tensor(buf15, (4, 1, 4, 4), (80, 16, 4, 1), 0) # alias
# Topologically Sorted Source Nodes: [ttx, ttx_1], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(primals_4, buf4, buf13, 64, grid=grid(64), stream=stream0)
del primals_4
# Topologically Sorted Source Nodes: [out_2], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(buf6, primals_5, stride=(1, 1), padding=(1, 1), 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
buf9 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
buf10 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf12 = reinterpret_tensor(buf10, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf10 # reuse
buf14 = reinterpret_tensor(buf15, (4, 4, 4, 4), (80, 16, 4, 1), 16) # alias
# Topologically Sorted Source Nodes: [out_2, out_3, out_4], Original ATen: [aten.convolution, aten.native_group_norm, aten.relu]
triton_per_fused_convolution_native_group_norm_relu_2.run(buf8, buf12, primals_6, primals_7, primals_8, buf9, buf14, 16, 16, grid=grid(16), stream=stream0)
del primals_6
# Topologically Sorted Source Nodes: [out_5], Original ATen: [aten.convolution]
buf16 = extern_kernels.convolution(buf15, primals_9, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf16, (4, 4, 4, 4), (64, 16, 4, 1))
buf17 = buf16; del buf16 # reuse
buf18 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf21 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf22 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [out_5, out_6], Original ATen: [aten.convolution, aten.native_group_norm]
triton_per_fused_convolution_native_group_norm_3.run(buf17, primals_10, primals_11, primals_12, buf18, buf21, buf22, 16, 16, grid=grid(16), stream=stream0)
del primals_10
del primals_12
return (buf21, primals_1, primals_2, primals_3, primals_5, primals_7, primals_8, primals_9, primals_11, buf0, buf3, buf6, buf8, buf9, buf12, buf15, buf17, reinterpret_tensor(buf18, (4, 4), (4, 1), 0), reinterpret_tensor(buf22, (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, ), (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, 4, 4), (16, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 5, 3, 3), (45, 9, 3, 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)
primals_8 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((4, 5, 3, 3), (45, 9, 3, 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)
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])
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 norm(dim):
return nn.GroupNorm(min(32, dim), dim)
class ConcatConv2d(nn.Module):
def __init__(self, dim_in, dim_out, ksize=3, stride=1, padding=0,
dilation=1, groups=1, bias=True, transpose=False):
super(ConcatConv2d, self).__init__()
module = nn.ConvTranspose2d if transpose else nn.Conv2d
self._layer = module(dim_in + 1, dim_out, kernel_size=ksize, stride
=stride, padding=padding, dilation=dilation, groups=groups,
bias=bias)
def forward(self, t, x):
tt = torch.ones_like(x[:, :1, :, :]) * t
ttx = torch.cat([tt, x], 1)
return self._layer(ttx)
class ODEfunc(nn.Module):
def __init__(self, dim):
super(ODEfunc, self).__init__()
self.norm1 = norm(dim)
self.relu = nn.ReLU(inplace=True)
self.conv1 = ConcatConv2d(dim, dim, 3, 1, 1)
self.norm2 = norm(dim)
self.conv2 = ConcatConv2d(dim, dim, 3, 1, 1)
self.norm3 = norm(dim)
self.nfe = 0
def forward(self, t, x):
self.nfe += 1
out = self.norm1(x)
out = self.relu(out)
out = self.conv1(t, out)
out = self.norm2(out)
out = self.relu(out)
out = self.conv2(t, out)
out = self.norm3(out)
return out
def get_inputs():
return [torch.rand([4, 1, 4, 4]), 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
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
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_per_fused_native_group_norm_relu_0(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr1, 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
x2 = xindex % 4
x3 = xindex // 4
tmp0 = tl.load(in_ptr0 + (r1 + 16 * x0), xmask, other=0.0)
tmp24 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr2 + x2, xmask, 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], 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 = 16.0
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 = tl.full([1, 1], 0, tl.int32)
tmp29 = triton_helpers.maximum(tmp28, tmp27)
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp21, xmask)
tl.store(out_ptr1 + (r1 + 16 * x2 + 80 * x3), tmp29, xmask)
tl.store(out_ptr0 + x0, tmp10, xmask)
@triton.jit
def triton_poi_fused_cat_1(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
x2 = xindex
x0 = xindex % 16
x1 = xindex // 16
tmp0 = tl.load(in_ptr0 + x2, xmask)
tl.store(out_ptr0 + (x0 + 80 * x1), tmp0, xmask)
tl.store(out_ptr1 + (x0 + 80 * x1), tmp0, xmask)
@triton.jit
def triton_per_fused_convolution_native_group_norm_relu_2(in_out_ptr0,
in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
x1 = xindex // 4
tmp0 = tl.load(in_out_ptr0 + (r2 + 16 * x3), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 16.0
tmp20 = tmp18 / tmp19
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tmp24 = tmp2 - tmp12
tmp25 = tmp24 * tmp23
tmp27 = tmp25 * tmp26
tmp29 = tmp27 + tmp28
tmp30 = tl.full([1, 1], 0, tl.int32)
tmp31 = triton_helpers.maximum(tmp30, tmp29)
tl.store(in_out_ptr0 + (r2 + 16 * x3), tmp2, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + x3, tmp23, xmask)
tl.store(out_ptr1 + (r2 + 16 * x0 + 80 * x1), tmp31, xmask)
tl.store(out_ptr0 + x3, tmp12, xmask)
@triton.jit
def triton_per_fused_convolution_native_group_norm_3(in_out_ptr0, in_ptr0,
in_ptr1, in_ptr2, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (r2 + 16 * x3), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = tmp2 - tmp12
tmp20 = 16.0
tmp21 = tmp18 / tmp20
tmp22 = 1e-05
tmp23 = tmp21 + tmp22
tmp24 = libdevice.rsqrt(tmp23)
tmp25 = tmp19 * tmp24
tmp27 = tmp25 * tmp26
tmp29 = tmp27 + tmp28
tl.store(in_out_ptr0 + (r2 + 16 * x3), tmp2, xmask)
tl.store(out_ptr2 + (r2 + 16 * x3), tmp29, xmask)
tl.store(out_ptr3 + x3, tmp24, xmask)
tl.store(out_ptr0 + x3, tmp12, 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
) = args
args.clear()
assert_size_stride(primals_1, (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, 4, 4), (16, 16, 4, 1))
assert_size_stride(primals_5, (4, 5, 3, 3), (45, 9, 3, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4,), (1,))
assert_size_stride(primals_9, (4, 5, 3, 3), (45, 9, 3, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4,), (1,))
assert_size_stride(primals_12, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
buf1 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf3 = reinterpret_tensor(buf1, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf1
buf6 = empty_strided_cuda((4, 5, 4, 4), (80, 16, 4, 1), torch.float32)
buf5 = reinterpret_tensor(buf6, (4, 4, 4, 4), (80, 16, 4, 1), 16)
get_raw_stream(0)
triton_per_fused_native_group_norm_relu_0[grid(16)](buf3, primals_3,
primals_1, primals_2, buf0, buf5, 16, 16, XBLOCK=8, num_warps=2,
num_stages=1)
buf4 = reinterpret_tensor(buf6, (4, 1, 4, 4), (80, 16, 4, 1), 0)
buf15 = empty_strided_cuda((4, 5, 4, 4), (80, 16, 4, 1), torch.float32)
buf13 = reinterpret_tensor(buf15, (4, 1, 4, 4), (80, 16, 4, 1), 0)
triton_poi_fused_cat_1[grid(64)](primals_4, buf4, buf13, 64, XBLOCK
=64, num_warps=1, num_stages=1)
del primals_4
buf7 = extern_kernels.convolution(buf6, primals_5, stride=(1, 1),
padding=(1, 1), 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
buf9 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
buf10 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf12 = reinterpret_tensor(buf10, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf10
buf14 = reinterpret_tensor(buf15, (4, 4, 4, 4), (80, 16, 4, 1), 16)
triton_per_fused_convolution_native_group_norm_relu_2[grid(16)](buf8,
buf12, primals_6, primals_7, primals_8, buf9, buf14, 16, 16,
XBLOCK=1, num_warps=2, num_stages=1)
del primals_6
buf16 = extern_kernels.convolution(buf15, primals_9, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf16, (4, 4, 4, 4), (64, 16, 4, 1))
buf17 = buf16
del buf16
buf18 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
buf21 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf22 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
triton_per_fused_convolution_native_group_norm_3[grid(16)](buf17,
primals_10, primals_11, primals_12, buf18, buf21, buf22, 16, 16,
XBLOCK=1, num_warps=2, num_stages=1)
del primals_10
del primals_12
return (buf21, primals_1, primals_2, primals_3, primals_5, primals_7,
primals_8, primals_9, primals_11, buf0, buf3, buf6, buf8, buf9,
buf12, buf15, buf17, reinterpret_tensor(buf18, (4, 4), (4, 1), 0),
reinterpret_tensor(buf22, (4, 4), (4, 1), 0))
def norm(dim):
return nn.GroupNorm(min(32, dim), dim)
class ConcatConv2d(nn.Module):
def __init__(self, dim_in, dim_out, ksize=3, stride=1, padding=0,
dilation=1, groups=1, bias=True, transpose=False):
super(ConcatConv2d, self).__init__()
module = nn.ConvTranspose2d if transpose else nn.Conv2d
self._layer = module(dim_in + 1, dim_out, kernel_size=ksize, stride
=stride, padding=padding, dilation=dilation, groups=groups,
bias=bias)
def forward(self, t, x):
tt = torch.ones_like(x[:, :1, :, :]) * t
ttx = torch.cat([tt, x], 1)
return self._layer(ttx)
class ODEfuncNew(nn.Module):
def __init__(self, dim):
super(ODEfuncNew, self).__init__()
self.norm1 = norm(dim)
self.relu = nn.ReLU(inplace=True)
self.conv1 = ConcatConv2d(dim, dim, 3, 1, 1)
self.norm2 = norm(dim)
self.conv2 = ConcatConv2d(dim, dim, 3, 1, 1)
self.norm3 = norm(dim)
self.nfe = 0
def forward(self, input_0, input_1):
primals_1 = self.norm1.weight
primals_2 = self.norm1.bias
primals_5 = self.conv1._layer.weight
primals_6 = self.conv1._layer.bias
primals_7 = self.norm2.weight
primals_8 = self.norm2.bias
primals_9 = self.conv2._layer.weight
primals_10 = self.conv2._layer.bias
primals_11 = self.norm3.weight
primals_12 = self.norm3.bias
primals_4 = input_0
primals_3 = 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])
return output[0]
| gaozhihan/torchdiffeq | ODEfunc | false | 6,741 | [
"MIT"
] | 1 | 414781617d595ba01cc3f23382e25ab890f4ca66 | https://github.com/gaozhihan/torchdiffeq/tree/414781617d595ba01cc3f23382e25ab890f4ca66 | import torch
import torch.nn as nn
def norm(dim):
return nn.GroupNorm(min(32, dim), dim)
class ConcatConv2d(nn.Module):
def __init__(self, dim_in, dim_out, ksize=3, stride=1, padding=0,
dilation=1, groups=1, bias=True, transpose=False):
super().__init__()
module = nn.ConvTranspose2d if transpose else nn.Conv2d
self._layer = module(dim_in + 1, dim_out, kernel_size=ksize, stride
=stride, padding=padding, dilation=dilation, groups=groups,
bias=bias)
def forward(self, t, x):
tt = torch.ones_like(x[:, :1, :, :]) * t
ttx = torch.cat([tt, x], 1)
return self._layer(ttx)
class Model(nn.Module):
def __init__(self, dim):
super().__init__()
self.norm1 = norm(dim)
self.relu = nn.ReLU(inplace=True)
self.conv1 = ConcatConv2d(dim, dim, 3, 1, 1)
self.norm2 = norm(dim)
self.conv2 = ConcatConv2d(dim, dim, 3, 1, 1)
self.norm3 = norm(dim)
self.nfe = 0
def forward(self, t, x):
self.nfe += 1
out = self.norm1(x)
out = self.relu(out)
out = self.conv1(t, out)
out = self.norm2(out)
out = self.relu(out)
out = self.conv2(t, out)
out = self.norm3(out)
return out
def get_inputs():
return [torch.rand([4, 1, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
MSELoss2d | # 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_4/inductor_cache/qz/cqza6p5fjiie2hfiu5dfjqqugrnzziwuwxzlhzy2aa7khopxjbym.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 = (%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=[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
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_4/inductor_cache/m4/cm4xtb3bnuq3knchbeowql4btzwa7fgl5fl2ojphcrhtylwyafoa.py
# Topologically Sorted Source Nodes: [softmax, loss], Original ATen: [aten._softmax, aten.mse_loss]
# Source node to ATen node mapping:
# loss => mean, pow_1, sub_1
# 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 = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%div, %arg1_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub_1, 2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%pow_1,), kwargs = {})
triton_per_fused__softmax_mse_loss_1 = async_compile.triton('triton_per_fused__softmax_mse_loss_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__softmax_mse_loss_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 6, '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__softmax_mse_loss_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)
r3 = rindex
r0 = rindex % 16
r2 = (rindex // 64)
tmp0 = tl.load(in_ptr0 + (r3), None)
tmp1 = tl.load(in_ptr0 + (r0 + (64*r2)), None, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (16 + r0 + (64*r2)), None, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (32 + r0 + (64*r2)), None, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (48 + r0 + (64*r2)), None, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + (r3), None)
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tmp10 = tmp8 - tmp9
tmp11 = tmp10 * tmp10
tmp12 = tl.broadcast_to(tmp11, [RBLOCK])
tmp14 = triton_helpers.promote_to_tensor(tl.sum(tmp12, 0))
tmp15 = 256.0
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, 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: [softmax], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [softmax, loss], Original ATen: [aten._softmax, aten.mse_loss]
triton_per_fused__softmax_mse_loss_1.run(buf2, buf0, arg1_1, 1, 256, grid=grid(1), stream=stream0)
del arg1_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
from torch import nn
class MSELoss2d(nn.Module):
def __init__(self, size_average=None, reduce=None, reduction='mean',
ignore_index=255):
super(MSELoss2d, self).__init__()
self.MSE = nn.MSELoss(size_average=size_average, reduce=reduce,
reduction=reduction)
def forward(self, output, target):
loss = self.MSE(torch.softmax(output, dim=1), target)
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
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import 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
@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
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_per_fused__softmax_mse_loss_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)
r3 = rindex
r0 = rindex % 16
r2 = rindex // 64
tmp0 = tl.load(in_ptr0 + r3, None)
tmp1 = tl.load(in_ptr0 + (r0 + 64 * r2), None, eviction_policy='evict_last'
)
tmp2 = tl.load(in_ptr0 + (16 + r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr0 + (32 + r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp6 = tl.load(in_ptr0 + (48 + r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp9 = tl.load(in_ptr1 + r3, None)
tmp3 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tmp10 = tmp8 - tmp9
tmp11 = tmp10 * tmp10
tmp12 = tl.broadcast_to(tmp11, [RBLOCK])
tmp14 = triton_helpers.promote_to_tensor(tl.sum(tmp12, 0))
tmp15 = 256.0
tmp16 = tmp14 / tmp15
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp16, 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__softmax_0[grid(256)](arg0_1, buf0, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg0_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
triton_per_fused__softmax_mse_loss_1[grid(1)](buf2, buf0, arg1_1, 1,
256, num_warps=2, num_stages=1)
del arg1_1
del buf0
return buf2,
class MSELoss2dNew(nn.Module):
def __init__(self, size_average=None, reduce=None, reduction='mean',
ignore_index=255):
super(MSELoss2dNew, self).__init__()
self.MSE = nn.MSELoss(size_average=size_average, reduce=reduce,
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]
| giannifranchi/deeplabv3-superpixelmix | MSELoss2d | false | 6,742 | [
"MIT"
] | 1 | db52bf83b3b242af05bde5e39ee3de896e44c264 | https://github.com/giannifranchi/deeplabv3-superpixelmix/tree/db52bf83b3b242af05bde5e39ee3de896e44c264 | import torch
from torch import nn
class Model(nn.Module):
def __init__(self, size_average=None, reduce=None, reduction='mean',
ignore_index=255):
super().__init__()
self.MSE = nn.MSELoss(size_average=size_average, reduce=reduce,
reduction=reduction)
def forward(self, output, target):
loss = self.MSE(torch.softmax(output, dim=1), target)
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
ResidualBlock | # 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_4/inductor_cache/d2/cd26tq7ga52mdxogiuulvwbmbjkpkq2yqyx64cyua3djeh4ftabe.py
# Topologically Sorted Source Nodes: [residual, residual_1, residual_2], Original ATen: [aten.convolution, aten._native_batch_norm_legit, aten._prelu_kernel]
# Source node to ATen node mapping:
# residual => convolution
# residual_1 => add, rsqrt, var_mean
# residual_2 => gt, mul_1, 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 = {})
# %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 = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%view_1, 0), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_2, %view_1), kwargs = {})
# %where : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%gt, %view_1, %mul_1), kwargs = {})
triton_per_fused__native_batch_norm_legit__prelu_kernel_convolution_0 = async_compile.triton('triton_per_fused__native_batch_norm_legit__prelu_kernel_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.persistent_reduction(
size_hints=[16, 16],
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, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__native_batch_norm_legit__prelu_kernel_convolution_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__native_batch_norm_legit__prelu_kernel_convolution_0(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, out_ptr0, out_ptr1, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (r2 + (16*x3)), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr1 + (0))
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 16.0
tmp20 = tmp18 / tmp19
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tmp24 = tmp2 - tmp12
tmp25 = tmp24 * tmp23
tmp26 = 0.0
tmp27 = tmp25 > tmp26
tmp30 = tmp29 * tmp25
tmp31 = tl.where(tmp27, tmp25, tmp30)
tl.store(in_out_ptr0 + (r2 + (16*x3)), tmp2, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x3), tmp23, xmask)
tl.store(out_ptr1 + (r2 + (16*x3)), tmp31, xmask)
tl.store(out_ptr0 + (x3), tmp12, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/lh/clhmsqir75rcwu2chfhikphjpocqjy67qzkdc6fbtmlum5jehkje.py
# Topologically Sorted Source Nodes: [residual_3, residual_4, add], Original ATen: [aten.convolution, aten._native_batch_norm_legit, aten.add]
# Source node to ATen node mapping:
# add => add_2
# residual_3 => convolution_1
# residual_4 => add_1, rsqrt_1, var_mean_1
# Graph fragment:
# %convolution_1 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%where, %primals_5, %primals_6, [1, 1], [1, 1], [1, 1], False, [0, 0], 1), kwargs = {})
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_3, [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 = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_3, %view_4), kwargs = {})
triton_per_fused__native_batch_norm_legit_add_convolution_1 = async_compile.triton('triton_per_fused__native_batch_norm_legit_add_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.persistent_reduction(
size_hints=[16, 16],
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, 6, 7), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__native_batch_norm_legit_add_convolution_1', 'mutated_arg_names': ['in_out_ptr0'], '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__native_batch_norm_legit_add_convolution_1(in_out_ptr0, in_ptr0, in_ptr1, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (r2 + (16*x3)), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr1 + (r2 + (16*x3)), xmask, other=0.0)
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp20 = tmp2 - tmp12
tmp21 = 16.0
tmp22 = tmp18 / tmp21
tmp23 = 1e-05
tmp24 = tmp22 + tmp23
tmp25 = libdevice.rsqrt(tmp24)
tmp26 = tmp20 * tmp25
tmp27 = tmp19 + tmp26
tl.store(in_out_ptr0 + (r2 + (16*x3)), tmp2, xmask)
tl.store(out_ptr2 + (r2 + (16*x3)), tmp27, xmask)
tl.store(out_ptr3 + (x3), tmp25, xmask)
tl.store(out_ptr0 + (x3), tmp12, 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, 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, (1, ), (1, ))
assert_size_stride(primals_5, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_6, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [residual], 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
buf2 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 1, 1), torch.float32)
buf3 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
buf5 = reinterpret_tensor(buf3, (1, 16, 1, 1), (16, 1, 1, 1), 0); del buf3 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [residual, residual_1, residual_2], Original ATen: [aten.convolution, aten._native_batch_norm_legit, aten._prelu_kernel]
stream0 = get_raw_stream(0)
triton_per_fused__native_batch_norm_legit__prelu_kernel_convolution_0.run(buf1, buf5, primals_2, primals_4, buf2, buf6, 16, 16, grid=grid(16), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [residual_3], Original ATen: [aten.convolution]
buf7 = extern_kernels.convolution(buf6, primals_5, stride=(1, 1), padding=(1, 1), 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
buf9 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
buf13 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf12 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [residual_3, residual_4, add], Original ATen: [aten.convolution, aten._native_batch_norm_legit, aten.add]
triton_per_fused__native_batch_norm_legit_add_convolution_1.run(buf8, primals_6, primals_3, buf9, buf13, buf12, 16, 16, grid=grid(16), stream=stream0)
del primals_6
return (buf13, primals_1, primals_3, primals_4, primals_5, buf1, buf2, buf5, buf6, buf8, reinterpret_tensor(buf12, (16, ), (1, ), 0), reinterpret_tensor(buf9, (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, 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((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4, 3, 3), (36, 9, 3, 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
import torch.nn as nn
class ResidualBlock(nn.Module):
def __init__(self, channels):
super(ResidualBlock, self).__init__()
self.conv1 = nn.Conv2d(channels, channels, kernel_size=3, padding=1)
self.in1 = nn.InstanceNorm2d(channels)
self.prelu = nn.PReLU()
self.conv2 = nn.Conv2d(channels, channels, kernel_size=3, padding=1)
self.in2 = nn.InstanceNorm2d(channels)
def forward(self, x):
residual = self.conv1(x)
residual = self.in1(residual)
residual = self.prelu(residual)
residual = self.conv2(residual)
residual = self.in2(residual)
return x + residual
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'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
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__prelu_kernel_convolution_0(
in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, out_ptr0, out_ptr1, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (r2 + 16 * x3), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr1 + 0)
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 16.0
tmp20 = tmp18 / tmp19
tmp21 = 1e-05
tmp22 = tmp20 + tmp21
tmp23 = libdevice.rsqrt(tmp22)
tmp24 = tmp2 - tmp12
tmp25 = tmp24 * tmp23
tmp26 = 0.0
tmp27 = tmp25 > tmp26
tmp30 = tmp29 * tmp25
tmp31 = tl.where(tmp27, tmp25, tmp30)
tl.store(in_out_ptr0 + (r2 + 16 * x3), tmp2, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + x3, tmp23, xmask)
tl.store(out_ptr1 + (r2 + 16 * x3), tmp31, xmask)
tl.store(out_ptr0 + x3, tmp12, xmask)
@triton.jit
def triton_per_fused__native_batch_norm_legit_add_convolution_1(in_out_ptr0,
in_ptr0, in_ptr1, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_out_ptr0 + (r2 + 16 * x3), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr1 + (r2 + 16 * x3), xmask, other=0.0)
tmp2 = tmp0 + tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tl.where(xmask, tmp3, 0)
tmp6 = tl.broadcast_to(tmp3, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp3 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp20 = tmp2 - tmp12
tmp21 = 16.0
tmp22 = tmp18 / tmp21
tmp23 = 1e-05
tmp24 = tmp22 + tmp23
tmp25 = libdevice.rsqrt(tmp24)
tmp26 = tmp20 * tmp25
tmp27 = tmp19 + tmp26
tl.store(in_out_ptr0 + (r2 + 16 * x3), tmp2, xmask)
tl.store(out_ptr2 + (r2 + 16 * x3), tmp27, xmask)
tl.store(out_ptr3 + x3, tmp25, xmask)
tl.store(out_ptr0 + x3, tmp12, 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, 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, (1,), (1,))
assert_size_stride(primals_5, (4, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_6, (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
buf2 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 1, 1), torch.float32)
buf3 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32
)
buf5 = reinterpret_tensor(buf3, (1, 16, 1, 1), (16, 1, 1, 1), 0)
del buf3
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused__native_batch_norm_legit__prelu_kernel_convolution_0[
grid(16)](buf1, buf5, primals_2, primals_4, buf2, buf6, 16, 16,
XBLOCK=8, num_warps=2, num_stages=1)
del primals_2
buf7 = extern_kernels.convolution(buf6, primals_5, stride=(1, 1),
padding=(1, 1), 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
buf9 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.float32
)
buf13 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf12 = empty_strided_cuda((1, 16, 1, 1), (16, 1, 16, 16), torch.
float32)
triton_per_fused__native_batch_norm_legit_add_convolution_1[grid(16)](
buf8, primals_6, primals_3, buf9, buf13, buf12, 16, 16, XBLOCK=
1, num_warps=2, num_stages=1)
del primals_6
return (buf13, primals_1, primals_3, primals_4, primals_5, buf1, buf2,
buf5, buf6, buf8, reinterpret_tensor(buf12, (16,), (1,), 0),
reinterpret_tensor(buf9, (1, 16, 1, 1), (16, 1, 1, 1), 0))
class ResidualBlockNew(nn.Module):
def __init__(self, channels):
super(ResidualBlockNew, self).__init__()
self.conv1 = nn.Conv2d(channels, channels, kernel_size=3, padding=1)
self.in1 = nn.InstanceNorm2d(channels)
self.prelu = nn.PReLU()
self.conv2 = nn.Conv2d(channels, channels, kernel_size=3, padding=1)
self.in2 = nn.InstanceNorm2d(channels)
def forward(self, input_0):
primals_1 = self.conv1.weight
primals_2 = self.conv1.bias
primals_4 = self.prelu.weight
primals_5 = self.conv2.weight
primals_6 = self.conv2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6])
return output[0]
| generall/Torchlite | ResidualBlock | false | 6,743 | [
"MIT"
] | 1 | 2eb3e2a20b7619bd58b0b0fca120e2aefca0e79a | https://github.com/generall/Torchlite/tree/2eb3e2a20b7619bd58b0b0fca120e2aefca0e79a | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, channels):
super().__init__()
self.conv1 = nn.Conv2d(channels, channels, kernel_size=3, padding=1)
self.in1 = nn.InstanceNorm2d(channels)
self.prelu = nn.PReLU()
self.conv2 = nn.Conv2d(channels, channels, kernel_size=3, padding=1)
self.in2 = nn.InstanceNorm2d(channels)
def forward(self, x):
residual = self.conv1(x)
residual = self.in1(residual)
residual = self.prelu(residual)
residual = self.conv2(residual)
residual = self.in2(residual)
return x + residual
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [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_4/inductor_cache/2z/c2zxna47l7bxhawt2r7o2srls6sac3it5g6jexlt3seylwwxl2n2.py
# Topologically Sorted Source Nodes: [x, leaky_relu], Original ATen: [aten.convolution, aten.leaky_relu]
# Source node to ATen node mapping:
# leaky_relu => gt, mul, where
# x => convolution
# Graph fragment:
# %convolution : [num_users=3] = 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 = {})
# %gt : [num_users=2] = call_function[target=torch.ops.aten.gt.Scalar](args = (%convolution, 0), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%convolution, 0.01), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %convolution, %mul), kwargs = {})
triton_poi_fused_convolution_leaky_relu_0 = async_compile.triton('triton_poi_fused_convolution_leaky_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: '*i1', 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_convolution_leaky_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_convolution_leaky_relu_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, 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_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.01
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tl.store(out_ptr0 + (x3), tmp4, xmask)
tl.store(out_ptr1 + (x3), tmp7, 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)
# 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))
buf1 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.bool)
buf2 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.float32)
# Topologically Sorted Source Nodes: [x, leaky_relu], Original ATen: [aten.convolution, aten.leaky_relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_leaky_relu_0.run(buf0, primals_2, buf1, buf2, 144, grid=grid(144), stream=stream0)
del buf0
del primals_2
return (buf2, primals_1, primals_3, 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, 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
class ConvBlock(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride):
super(ConvBlock, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=1)
self.lr = nn.LeakyReLU()
def forward(self, x):
x = self.conv(x)
return self.lr(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}]
| 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_convolution_leaky_relu_0(in_ptr0, in_ptr1, out_ptr0,
out_ptr1, 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_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.01
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tl.store(out_ptr0 + x3, tmp4, xmask)
tl.store(out_ptr1 + x3, tmp7, 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 = 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))
buf1 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.bool)
buf2 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_convolution_leaky_relu_0[grid(144)](buf0,
primals_2, buf1, buf2, 144, XBLOCK=128, num_warps=4, num_stages=1)
del buf0
del primals_2
return buf2, primals_1, primals_3, buf1
class ConvBlockNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride):
super(ConvBlockNew, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=1)
self.lr = nn.LeakyReLU()
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]
| gle-bellier/DuelingNetwork | ConvBlock | false | 6,744 | [
"MIT"
] | 1 | 8909fe1ba6aee08b6249cb6ca3287752039c6410 | https://github.com/gle-bellier/DuelingNetwork/tree/8909fe1ba6aee08b6249cb6ca3287752039c6410 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size, stride):
super().__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size,
stride, padding=1)
self.lr = nn.LeakyReLU()
def forward(self, x):
x = self.conv(x)
return self.lr(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}]
|
WordPredictor | # 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_4/inductor_cache/vj/cvjsxyxn6gyyvtipzzz4vzmiphntjftekqenbhs4uunn7tb5u3jd.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mean]
# Source node to ATen node mapping:
# x => mean
# Graph fragment:
# %mean : [num_users=2] = call_function[target=torch.ops.aten.mean.dim](args = (%select, [0]), 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': 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_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
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0), xmask)
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')
# kernel path: runs/run_shard_4/inductor_cache/5i/c5iaqsrsfvsdc63zdr5cg7f37p7opjrglg57oxsv7o3xmexazhkl.py
# Topologically Sorted Source Nodes: [attn_input], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# attn_input => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%expand, %select], 2), 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=[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_cat_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_cat_1(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 % 8
x1 = (xindex // 8) % 4
x3 = (xindex // 8)
x4 = 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 = tl.load(in_ptr1 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full([1], 0, tl.int32)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp10 = tl.full(tmp9.shape, 0.0, tmp9.dtype)
tmp11 = tl.where(tmp4, tmp9, tmp10)
tmp12 = tmp0 >= tmp3
tmp13 = tl.full([1], 8, tl.int64)
tmp14 = tmp0 < tmp13
tmp15 = tl.load(in_ptr2 + ((4*x3) + ((-4) + x0)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp16 = tl.where(tmp4, tmp11, tmp15)
tl.store(out_ptr0 + (x4), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/dj/cdjzmtk5emc6dzqhkvu5c2hunn5lmuxuc47c6cxqz34wv4kuxdx7.py
# Topologically Sorted Source Nodes: [attn_scores, attn_scores_1], Original ATen: [aten.relu, aten._softmax]
# Source node to ATen node mapping:
# attn_scores => relu_1
# attn_scores_1 => amax, exp, sub
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%relu_1, [0], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%relu_1, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), kwargs = {})
triton_poi_fused__softmax_relu_2 = async_compile.triton('triton_poi_fused__softmax_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=[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_relu_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_relu_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
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp3 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (4 + x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (8 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (12 + x0), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp1, tmp3)
tmp6 = triton_helpers.maximum(tmp1, tmp5)
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = triton_helpers.maximum(tmp1, tmp8)
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = triton_helpers.maximum(tmp1, tmp11)
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = tl_math.exp(tmp14)
tl.store(out_ptr0 + (x2), tmp15, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ny/cny2tn74oditrr2kjk2zdrdpqulok6k67hyn3d3scy77clhd6evl.py
# Topologically Sorted Source Nodes: [attn_scores_1], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attn_scores_1 => 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_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=[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_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 = 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_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (4 + x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (8 + x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (12 + 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')
# kernel path: runs/run_shard_4/inductor_cache/7i/c7i7xxsu2jfndntuz754dt7qepn6ntejzl6j4mrnolhaicepomou.py
# Topologically Sorted Source Nodes: [pred_input], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# pred_input => cat_1
# Graph fragment:
# %cat_1 : [num_users=2] = call_function[target=torch.ops.aten.cat.default](args = ([%relu, %sum_2], 1), kwargs = {})
triton_poi_fused_cat_4 = async_compile.triton('triton_poi_fused_cat_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=[32],
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_cat_4', '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_cat_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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 = tl.load(in_ptr1 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full([1], 0, tl.int32)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp10 = tl.full(tmp9.shape, 0.0, tmp9.dtype)
tmp11 = tl.where(tmp4, tmp9, tmp10)
tmp12 = tmp0 >= tmp3
tmp13 = tl.full([1], 8, tl.int64)
tmp14 = tmp0 < tmp13
tmp15 = tl.load(in_ptr2 + ((4*x1) + ((-4) + x0)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp16 = tl.load(in_ptr3 + (x1), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp17 = tmp15 * tmp16
tmp18 = tl.load(in_ptr2 + (16 + (4*x1) + ((-4) + x0)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp19 = tl.load(in_ptr3 + (4 + x1), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp20 = tmp18 * tmp19
tmp21 = tmp17 + tmp20
tmp22 = tl.load(in_ptr2 + (32 + (4*x1) + ((-4) + x0)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp23 = tl.load(in_ptr3 + (8 + x1), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp24 = tmp22 * tmp23
tmp25 = tmp21 + tmp24
tmp26 = tl.load(in_ptr2 + (48 + (4*x1) + ((-4) + x0)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp27 = tl.load(in_ptr3 + (12 + x1), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp28 = tmp26 * tmp27
tmp29 = tmp25 + tmp28
tmp30 = tl.full(tmp29.shape, 0.0, tmp29.dtype)
tmp31 = tl.where(tmp12, tmp29, tmp30)
tmp32 = tl.where(tmp4, tmp11, tmp31)
tl.store(out_ptr0 + (x2), tmp32, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/v4/cv4ko24f26un3axamp426zbnugqu4jhirnvlxjjqhipcugzzgcbk.py
# Topologically Sorted Source Nodes: [pred_hidden], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# pred_hidden => relu_2
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_7), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_relu_5 = async_compile.triton('triton_poi_fused_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=[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_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_5(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
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_4/inductor_cache/oo/coooss73fnhko7ikpecnkcvpqw5bmlwugiidvseo44i3dgysybgg.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:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_3), kwargs = {})
# %relu : [num_users=3] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor_1,), kwargs = {})
# %le_2 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_6 = async_compile.triton('triton_poi_fused_relu_threshold_backward_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: '*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_6', '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_6(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 = 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 + (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 = 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, (1, 8), (8, 1))
assert_size_stride(primals_5, (1, ), (1, ))
assert_size_stride(primals_6, (4, 8), (8, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (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: [x], 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, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf0, reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf1)
del primals_2
buf2 = empty_strided_cuda((4, 4, 8), (32, 8, 1), torch.float32)
# Topologically Sorted Source Nodes: [attn_input], Original ATen: [aten.cat]
triton_poi_fused_cat_1.run(buf1, primals_3, primals_1, buf2, 128, grid=grid(128), stream=stream0)
buf4 = empty_strided_cuda((16, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf2, (16, 8), (8, 1), 0), reinterpret_tensor(primals_4, (8, 1), (1, 8), 0), alpha=1, beta=1, out=buf4)
del primals_5
buf5 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
# Topologically Sorted Source Nodes: [attn_scores, attn_scores_1], Original ATen: [aten.relu, aten._softmax]
triton_poi_fused__softmax_relu_2.run(buf4, buf5, 16, grid=grid(16), stream=stream0)
buf6 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
# Topologically Sorted Source Nodes: [attn_scores_1], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf5, buf6, 16, grid=grid(16), stream=stream0)
buf7 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [pred_input], Original ATen: [aten.cat]
triton_poi_fused_cat_4.run(buf1, primals_3, primals_1, buf6, buf7, 32, grid=grid(32), stream=stream0)
buf8 = reinterpret_tensor(buf6, (4, 4), (4, 1), 0); del buf6 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf7, reinterpret_tensor(primals_6, (8, 4), (1, 8), 0), out=buf8)
buf9 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [pred_hidden], Original ATen: [aten.relu]
triton_poi_fused_relu_5.run(buf9, primals_7, 16, grid=grid(16), stream=stream0)
del primals_7
buf10 = reinterpret_tensor(buf5, (4, 4), (4, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [pred_logit], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, buf9, reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf10)
del primals_9
buf11 = empty_strided_cuda((4, 4), (4, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_6.run(buf1, primals_3, buf11, 16, grid=grid(16), stream=stream0)
del buf1
del primals_3
return (buf10, reinterpret_tensor(primals_1, (4, 4, 4), (16, 4, 1), 0), buf0, reinterpret_tensor(buf2, (16, 8), (8, 1), 0), buf4, buf7, buf9, primals_8, primals_6, primals_4, buf11, )
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((1, 8), (8, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 8), (8, 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, 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.functional as F
import torch.nn as nn
import torch.onnx.operators
class WordPredictor(nn.Module):
def __init__(self, encoder_output_dim, hidden_dim, output_dim):
super().__init__()
self.encoder_output_dim = encoder_output_dim
self.hidden_dim = hidden_dim
self.output_dim = output_dim
self.init_layer = nn.Linear(encoder_output_dim, encoder_output_dim)
self.attn_layer = nn.Linear(2 * encoder_output_dim, 1)
self.hidden_layer = nn.Linear(2 * encoder_output_dim, hidden_dim)
self.output_layer = nn.Linear(hidden_dim, output_dim)
def forward(self, encoder_output):
encoder_hiddens, *_ = encoder_output
assert encoder_hiddens.dim() == 3
init_state = self._get_init_state(encoder_hiddens)
attn_scores = self._attention(encoder_hiddens, init_state)
attned_state = (encoder_hiddens * attn_scores).sum(0)
pred_input = torch.cat([init_state, attned_state], 1)
pred_hidden = F.relu(self.hidden_layer(pred_input))
pred_logit = self.output_layer(pred_hidden)
return pred_logit
def _get_init_state(self, encoder_hiddens):
x = torch.mean(encoder_hiddens, 0)
x = F.relu(self.init_layer(x))
return x
def _attention(self, encoder_hiddens, init_state):
init_state = init_state.unsqueeze(0).expand_as(encoder_hiddens)
attn_input = torch.cat([init_state, encoder_hiddens], 2)
attn_scores = F.relu(self.attn_layer(attn_input))
attn_scores = F.softmax(attn_scores, 0)
return attn_scores
def get_normalized_probs(self, net_output, log_probs):
"""Get normalized probabilities (or log probs) from a net's output."""
logits = net_output
if log_probs:
return F.log_softmax(logits, dim=1)
else:
return F.softmax(logits, dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'encoder_output_dim': 4, 'hidden_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 import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn.functional as F
import torch.nn as nn
import torch.onnx.operators
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_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
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr0 + (16 + x0), xmask)
tmp3 = tl.load(in_ptr0 + (32 + x0), xmask)
tmp5 = tl.load(in_ptr0 + (48 + x0), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp7 = 4.0
tmp8 = tmp6 / tmp7
tl.store(out_ptr0 + x0, tmp8, xmask)
@triton.jit
def triton_poi_fused_cat_1(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 % 8
x1 = xindex // 8 % 4
x3 = xindex // 8
x4 = 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 = tl.load(in_ptr1 + x0, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full([1], 0, tl.int32)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp10 = tl.full(tmp9.shape, 0.0, tmp9.dtype)
tmp11 = tl.where(tmp4, tmp9, tmp10)
tmp12 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp15 = tl.load(in_ptr2 + (4 * x3 + (-4 + x0)), tmp12 & xmask,
eviction_policy='evict_last', other=0.0)
tmp16 = tl.where(tmp4, tmp11, tmp15)
tl.store(out_ptr0 + x4, tmp16, xmask)
@triton.jit
def triton_poi_fused__softmax_relu_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
x2 = xindex
x0 = xindex % 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp3 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (4 + x0), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (8 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (12 + x0), xmask, eviction_policy='evict_last')
tmp1 = tl.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp4 = triton_helpers.maximum(tmp1, tmp3)
tmp6 = triton_helpers.maximum(tmp1, tmp5)
tmp7 = triton_helpers.maximum(tmp4, tmp6)
tmp9 = triton_helpers.maximum(tmp1, tmp8)
tmp10 = triton_helpers.maximum(tmp7, tmp9)
tmp12 = triton_helpers.maximum(tmp1, tmp11)
tmp13 = triton_helpers.maximum(tmp10, tmp12)
tmp14 = tmp2 - tmp13
tmp15 = tl_math.exp(tmp14)
tl.store(out_ptr0 + x2, tmp15, xmask)
@triton.jit
def triton_poi_fused__softmax_3(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
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (4 + x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (8 + x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (12 + 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)
@triton.jit
def triton_poi_fused_cat_4(in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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 = tl.load(in_ptr1 + x0, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp7 = tmp5 + tmp6
tmp8 = tl.full([1], 0, tl.int32)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp10 = tl.full(tmp9.shape, 0.0, tmp9.dtype)
tmp11 = tl.where(tmp4, tmp9, tmp10)
tmp12 = tmp0 >= tmp3
tl.full([1], 8, tl.int64)
tmp15 = tl.load(in_ptr2 + (4 * x1 + (-4 + x0)), tmp12 & xmask,
eviction_policy='evict_last', other=0.0)
tmp16 = tl.load(in_ptr3 + x1, tmp12 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp17 = tmp15 * tmp16
tmp18 = tl.load(in_ptr2 + (16 + 4 * x1 + (-4 + x0)), tmp12 & xmask,
eviction_policy='evict_last', other=0.0)
tmp19 = tl.load(in_ptr3 + (4 + x1), tmp12 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp20 = tmp18 * tmp19
tmp21 = tmp17 + tmp20
tmp22 = tl.load(in_ptr2 + (32 + 4 * x1 + (-4 + x0)), tmp12 & xmask,
eviction_policy='evict_last', other=0.0)
tmp23 = tl.load(in_ptr3 + (8 + x1), tmp12 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp24 = tmp22 * tmp23
tmp25 = tmp21 + tmp24
tmp26 = tl.load(in_ptr2 + (48 + 4 * x1 + (-4 + x0)), tmp12 & xmask,
eviction_policy='evict_last', other=0.0)
tmp27 = tl.load(in_ptr3 + (12 + x1), tmp12 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp28 = tmp26 * tmp27
tmp29 = tmp25 + tmp28
tmp30 = tl.full(tmp29.shape, 0.0, tmp29.dtype)
tmp31 = tl.where(tmp12, tmp29, tmp30)
tmp32 = tl.where(tmp4, tmp11, tmp31)
tl.store(out_ptr0 + x2, tmp32, xmask)
@triton.jit
def triton_poi_fused_relu_5(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
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_relu_threshold_backward_6(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 = 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 + x2, tmp6, 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, (4, 4), (4, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (1, 8), (8, 1))
assert_size_stride(primals_5, (1,), (1,))
assert_size_stride(primals_6, (4, 8), (8, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (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_mean_0[grid(16)](primals_1, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf1 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(buf0, reinterpret_tensor(primals_2, (4, 4), (1, 4
), 0), out=buf1)
del primals_2
buf2 = empty_strided_cuda((4, 4, 8), (32, 8, 1), torch.float32)
triton_poi_fused_cat_1[grid(128)](buf1, primals_3, primals_1, buf2,
128, XBLOCK=128, num_warps=4, num_stages=1)
buf4 = empty_strided_cuda((16, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf2, (16, 8), (
8, 1), 0), reinterpret_tensor(primals_4, (8, 1), (1, 8), 0),
alpha=1, beta=1, out=buf4)
del primals_5
buf5 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
triton_poi_fused__softmax_relu_2[grid(16)](buf4, buf5, 16, XBLOCK=
16, num_warps=1, num_stages=1)
buf6 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
triton_poi_fused__softmax_3[grid(16)](buf5, buf6, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf7 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
triton_poi_fused_cat_4[grid(32)](buf1, primals_3, primals_1, buf6,
buf7, 32, XBLOCK=32, num_warps=1, num_stages=1)
buf8 = reinterpret_tensor(buf6, (4, 4), (4, 1), 0)
del buf6
extern_kernels.mm(buf7, reinterpret_tensor(primals_6, (8, 4), (1, 8
), 0), out=buf8)
buf9 = buf8
del buf8
triton_poi_fused_relu_5[grid(16)](buf9, primals_7, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_7
buf10 = reinterpret_tensor(buf5, (4, 4), (4, 1), 0)
del buf5
extern_kernels.addmm(primals_9, buf9, reinterpret_tensor(primals_8,
(4, 4), (1, 4), 0), alpha=1, beta=1, out=buf10)
del primals_9
buf11 = empty_strided_cuda((4, 4), (4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_6[grid(16)](buf1,
primals_3, buf11, 16, XBLOCK=16, num_warps=1, num_stages=1)
del buf1
del primals_3
return buf10, reinterpret_tensor(primals_1, (4, 4, 4), (16, 4, 1), 0
), buf0, reinterpret_tensor(buf2, (16, 8), (8, 1), 0
), buf4, buf7, buf9, primals_8, primals_6, primals_4, buf11
class WordPredictorNew(nn.Module):
def __init__(self, encoder_output_dim, hidden_dim, output_dim):
super().__init__()
self.encoder_output_dim = encoder_output_dim
self.hidden_dim = hidden_dim
self.output_dim = output_dim
self.init_layer = nn.Linear(encoder_output_dim, encoder_output_dim)
self.attn_layer = nn.Linear(2 * encoder_output_dim, 1)
self.hidden_layer = nn.Linear(2 * encoder_output_dim, hidden_dim)
self.output_layer = nn.Linear(hidden_dim, output_dim)
def _get_init_state(self, encoder_hiddens):
x = torch.mean(encoder_hiddens, 0)
x = F.relu(self.init_layer(x))
return x
def _attention(self, encoder_hiddens, init_state):
init_state = init_state.unsqueeze(0).expand_as(encoder_hiddens)
attn_input = torch.cat([init_state, encoder_hiddens], 2)
attn_scores = F.relu(self.attn_layer(attn_input))
attn_scores = F.softmax(attn_scores, 0)
return attn_scores
def get_normalized_probs(self, net_output, log_probs):
"""Get normalized probabilities (or log probs) from a net's output."""
logits = net_output
if log_probs:
return F.log_softmax(logits, dim=1)
else:
return F.softmax(logits, dim=1)
def forward(self, input_0):
primals_2 = self.init_layer.weight
primals_3 = self.init_layer.bias
primals_4 = self.attn_layer.weight
primals_5 = self.attn_layer.bias
primals_6 = self.hidden_layer.weight
primals_7 = self.hidden_layer.bias
primals_8 = self.output_layer.weight
primals_9 = self.output_layer.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]
| gardenia22/translate | WordPredictor | false | 6,745 | [
"BSD-3-Clause"
] | 1 | 0be57c8f55b52fc9d39197efa02e05d1c1cda024 | https://github.com/gardenia22/translate/tree/0be57c8f55b52fc9d39197efa02e05d1c1cda024 | import torch
import torch.nn.functional as F
import torch.nn as nn
import torch.onnx.operators
class Model(nn.Module):
def __init__(self, encoder_output_dim, hidden_dim, output_dim):
super().__init__()
self.encoder_output_dim = encoder_output_dim
self.hidden_dim = hidden_dim
self.output_dim = output_dim
self.init_layer = nn.Linear(encoder_output_dim, encoder_output_dim)
self.attn_layer = nn.Linear(2 * encoder_output_dim, 1)
self.hidden_layer = nn.Linear(2 * encoder_output_dim, hidden_dim)
self.output_layer = nn.Linear(hidden_dim, output_dim)
def forward(self, encoder_output):
encoder_hiddens, *_ = encoder_output
assert encoder_hiddens.dim() == 3
init_state = self._get_init_state(encoder_hiddens)
attn_scores = self._attention(encoder_hiddens, init_state)
attned_state = (encoder_hiddens * attn_scores).sum(0)
pred_input = torch.cat([init_state, attned_state], 1)
pred_hidden = F.relu(self.hidden_layer(pred_input))
pred_logit = self.output_layer(pred_hidden)
return pred_logit
def _get_init_state(self, encoder_hiddens):
x = torch.mean(encoder_hiddens, 0)
x = F.relu(self.init_layer(x))
return x
def _attention(self, encoder_hiddens, init_state):
init_state = init_state.unsqueeze(0).expand_as(encoder_hiddens)
attn_input = torch.cat([init_state, encoder_hiddens], 2)
attn_scores = F.relu(self.attn_layer(attn_input))
attn_scores = F.softmax(attn_scores, 0)
return attn_scores
def get_normalized_probs(self, net_output, log_probs):
"""Get normalized probabilities (or log probs) from a net's output."""
logits = net_output
if log_probs:
return F.log_softmax(logits, dim=1)
else:
return F.softmax(logits, dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
VideoNormalizer | # 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_4/inductor_cache/pg/cpg32ajhfk5rtoj45a2u342syo2aujsz5vssd25nxwxjrtwnybqk.py
# Topologically Sorted Source Nodes: [truediv, sub, video_1], Original ATen: [aten.div, aten.sub]
# Source node to ATen node mapping:
# sub => sub
# truediv => div
# video_1 => div_1
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%div, %arg2_1), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %arg3_1), kwargs = {})
triton_poi_fused_div_sub_0 = async_compile.triton('triton_poi_fused_div_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: '*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_div_sub_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_div_sub_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 3
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp4 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tmp0 / tmp2
tmp5 = tmp3 - tmp4
tmp7 = tmp5 / tmp6
tl.store(out_ptr0 + (x2), tmp7, 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, 3), (48, 12, 3, 1))
assert_size_stride(arg1_1, (1, ), (1, ))
assert_size_stride(arg2_1, (3, ), (1, ))
assert_size_stride(arg3_1, (3, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 3), (48, 12, 3, 1), torch.float32)
# Topologically Sorted Source Nodes: [truediv, sub, video_1], Original ATen: [aten.div, aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_div_sub_0.run(arg0_1, arg1_1, arg2_1, arg3_1, buf0, 192, grid=grid(192), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
del arg3_1
return (reinterpret_tensor(buf0, (4, 3, 4, 4), (48, 1, 12, 3), 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, 3), (48, 12, 3, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((3, ), (1, ), device='cuda:0', dtype=torch.float32)
arg3_1 = rand_strided((3, ), (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 VideoNormalizer(nn.Module):
def __init__(self):
super(VideoNormalizer, self).__init__()
self.scale = nn.Parameter(torch.Tensor([255.0]), requires_grad=False)
self.mean = nn.Parameter(torch.Tensor([0.485, 0.456, 0.406]),
requires_grad=False)
self.std = nn.Parameter(torch.Tensor([0.229, 0.224, 0.225]),
requires_grad=False)
def forward(self, video):
video = video.float()
video = (video / self.scale - self.mean) / self.std
return video.permute(0, 3, 1, 2)
def get_inputs():
return [torch.rand([4, 4, 4, 3])]
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
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused_div_sub_0(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 3
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp4 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + x0, xmask, eviction_policy='evict_last')
tmp3 = tmp0 / tmp2
tmp5 = tmp3 - tmp4
tmp7 = tmp5 / tmp6
tl.store(out_ptr0 + x2, tmp7, xmask)
def call(args):
arg0_1, arg1_1, arg2_1, arg3_1 = args
args.clear()
assert_size_stride(arg0_1, (4, 4, 4, 3), (48, 12, 3, 1))
assert_size_stride(arg1_1, (1,), (1,))
assert_size_stride(arg2_1, (3,), (1,))
assert_size_stride(arg3_1, (3,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 3), (48, 12, 3, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_div_sub_0[grid(192)](arg0_1, arg1_1, arg2_1,
arg3_1, buf0, 192, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
del arg1_1
del arg2_1
del arg3_1
return reinterpret_tensor(buf0, (4, 3, 4, 4), (48, 1, 12, 3), 0),
class VideoNormalizerNew(nn.Module):
def __init__(self):
super(VideoNormalizerNew, self).__init__()
self.scale = nn.Parameter(torch.Tensor([255.0]), requires_grad=False)
self.mean = nn.Parameter(torch.Tensor([0.485, 0.456, 0.406]),
requires_grad=False)
self.std = nn.Parameter(torch.Tensor([0.229, 0.224, 0.225]),
requires_grad=False)
def forward(self, input_0):
arg1_1 = self.scale
arg2_1 = self.mean
arg3_1 = self.std
arg0_1 = input_0
output = call([arg0_1, arg1_1, arg2_1, arg3_1])
return output[0]
| glee1228/segment_temporal_context_aggregation | VideoNormalizer | false | 6,746 | [
"Apache-2.0"
] | 1 | e5778f848f1cfd89bd1f77beb5e1b38a66a2f13d | https://github.com/glee1228/segment_temporal_context_aggregation/tree/e5778f848f1cfd89bd1f77beb5e1b38a66a2f13d | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self):
super().__init__()
self.scale = nn.Parameter(torch.Tensor([255.0]), requires_grad=False)
self.mean = nn.Parameter(torch.Tensor([0.485, 0.456, 0.406]),
requires_grad=False)
self.std = nn.Parameter(torch.Tensor([0.229, 0.224, 0.225]),
requires_grad=False)
def forward(self, video):
video = video.float()
video = (video / self.scale - self.mean) / self.std
return video.permute(0, 3, 1, 2)
def get_inputs():
return [torch.rand([4, 4, 4, 3])]
def get_init_inputs():
return []
|
LogSparsemax | # 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_4/inductor_cache/xr/cxriojqnst2q32kjzwrpwltjdtloxmuqf3cbxuvng6r4ts5aacce.py
# Topologically Sorted Source Nodes: [max_1, input_1], Original ATen: [aten.max, aten.sub]
# Source node to ATen node mapping:
# input_1 => sub
# max_1 => max_1
# Graph fragment:
# %max_1 : [num_users=1] = call_function[target=torch.ops.aten.max.dim](args = (%arg0_1, 0, True), kwargs = {})
# %sub : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %getitem), kwargs = {})
triton_poi_fused_max_sub_0 = async_compile.triton('triton_poi_fused_max_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_max_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_max_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
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
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/43/c43uhskbeaaf4eboq6purohj7g5a6yrwr4b6mkhwl4lu2iqcj4cs.py
# Topologically Sorted Source Nodes: [sort, cumsum], Original ATen: [aten.sort, aten.cumsum]
# Source node to ATen node mapping:
# cumsum => cumsum
# sort => sort
# Graph fragment:
# %sort : [num_users=1] = call_function[target=torch.ops.aten.sort.default](args = (%sub, 0, True), kwargs = {})
# %cumsum : [num_users=1] = call_function[target=torch.ops.aten.cumsum.default](args = (%getitem_2, 0), kwargs = {})
triton_per_fused_cumsum_sort_1 = async_compile.triton('triton_per_fused_cumsum_sort_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.jit
def _triton_helper_fn_add0(arg0_0, arg1_0):
tmp0 = arg0_0 + arg1_0
return tmp0
@triton_heuristics.persistent_reduction(
size_hints=[64, 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': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cumsum_sort_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}
)
@triton.jit
def triton_per_fused_cumsum_sort_1(in_ptr0, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 4
RBLOCK: tl.constexpr = 4
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 + (x0 + (64*r1)), xmask, other=0.0)
tmp1 = r1
tmp2 = tmp1.to(tl.int16)
tmp3 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp4 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5, tmp6, = triton_helpers.sort_with_index(tmp3, tmp4, None, 1, stable=False, descending=True)
tmp7 = tmp5.to(tl.float32)
tmp8 = tl.broadcast_to(tmp7, [XBLOCK, RBLOCK])
tmp9, = tl.associative_scan((tmp8,), 1, _triton_helper_fn_add0)
tl.store(out_ptr0 + (x0 + (64*r1)), tmp5, xmask)
tl.store(out_ptr1 + (x0 + (64*r1)), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/rc/crclwvykuaxrcwig3dfr3ygs4skmnzpu4fkszzcyzp6jebodtdn6.py
# Topologically Sorted Source Nodes: [mul, input_cumsum, support, sum_1], Original ATen: [aten.mul, aten.sub, aten.gt, aten.sum]
# Source node to ATen node mapping:
# input_cumsum => sub_1
# mul => mul_1
# sum_1 => sum_1
# support => gt
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %getitem_2), kwargs = {})
# %sub_1 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%cumsum, 1), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Tensor](args = (%mul_1, %sub_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%gt, [0]), kwargs = {})
triton_poi_fused_gt_mul_sub_sum_2 = async_compile.triton('triton_poi_fused_gt_mul_sub_sum_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: '*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_gt_mul_sub_sum_2', '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_gt_mul_sub_sum_2(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
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask)
tmp7 = tl.load(in_ptr0 + (64 + x0), xmask)
tmp10 = tl.load(in_ptr1 + (64 + x0), xmask)
tmp15 = tl.load(in_ptr0 + (128 + x0), xmask)
tmp18 = tl.load(in_ptr1 + (128 + x0), xmask)
tmp23 = tl.load(in_ptr0 + (192 + x0), xmask)
tmp26 = tl.load(in_ptr1 + (192 + x0), xmask)
tmp1 = 1.0
tmp2 = tmp1 * tmp0
tmp4 = tmp3 - tmp1
tmp5 = tmp2 > tmp4
tmp6 = tmp5.to(tl.int64)
tmp8 = 2.0
tmp9 = tmp8 * tmp7
tmp11 = tmp10 - tmp1
tmp12 = tmp9 > tmp11
tmp13 = tmp12.to(tl.int64)
tmp14 = tmp6 + tmp13
tmp16 = 3.0
tmp17 = tmp16 * tmp15
tmp19 = tmp18 - tmp1
tmp20 = tmp17 > tmp19
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp14 + tmp21
tmp24 = 4.0
tmp25 = tmp24 * tmp23
tmp27 = tmp26 - tmp1
tmp28 = tmp25 > tmp27
tmp29 = tmp28.to(tl.int64)
tmp30 = tmp22 + tmp29
tl.store(out_ptr0 + (x0), tmp30, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/en/cencjosxr2pioverno5ax44si2tzeunwbyipd7u4r6velk3qps3r.py
# Topologically Sorted Source Nodes: [input_cumsum, sub_1, tau, tau_1, sub_2, output, log], Original ATen: [aten.sub, aten.gather, aten.div, aten.clamp, aten.log]
# Source node to ATen node mapping:
# input_cumsum => sub_1
# log => log
# output => clamp_min
# sub_1 => sub_2
# sub_2 => sub_3
# tau => gather
# tau_1 => div
# Graph fragment:
# %sub_1 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%cumsum, 1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%unsqueeze, 1), kwargs = {})
# %gather : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%sub_1, 0, %sub_2), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%gather, %unsqueeze), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %div), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sub_3, 0), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%clamp_min,), kwargs = {})
# %copy_ : [num_users=0] = call_function[target=torch.ops.aten.copy_.default](args = (%arg0_1, %sub), kwargs = {})
triton_poi_fused_clamp_div_gather_log_sub_3 = async_compile.triton('triton_poi_fused_clamp_div_gather_log_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=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i64', 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_clamp_div_gather_log_sub_3', 'mutated_arg_names': ['out_ptr1'], '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_clamp_div_gather_log_sub_3(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
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.full([1], 1, tl.int64)
tmp3 = tmp1 - tmp2
tmp4 = tl.full([XBLOCK], 4, tl.int32)
tmp5 = tmp3 + tmp4
tmp6 = tmp3 < 0
tmp7 = tl.where(tmp6, tmp5, tmp3)
tl.device_assert(((0 <= tmp7) & (tmp7 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp7 < 4")
tmp9 = tl.load(in_ptr2 + (x0 + (64*tmp7)), xmask)
tmp10 = 1.0
tmp11 = tmp9 - tmp10
tmp12 = tmp1.to(tl.float32)
tmp13 = tmp11 / tmp12
tmp14 = tmp0 - tmp13
tmp15 = 0.0
tmp16 = triton_helpers.maximum(tmp14, tmp15)
tmp17 = tl_math.log(tmp16)
tl.store(out_ptr0 + (x2), tmp17, xmask)
tl.store(out_ptr1 + (x2), 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: [max_1, input_1], Original ATen: [aten.max, aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_max_sub_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
buf1 = 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.float32)
# Topologically Sorted Source Nodes: [sort, cumsum], Original ATen: [aten.sort, aten.cumsum]
triton_per_fused_cumsum_sort_1.run(buf0, buf1, buf3, 64, 4, grid=grid(64), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.int64)
# Topologically Sorted Source Nodes: [mul, input_cumsum, support, sum_1], Original ATen: [aten.mul, aten.sub, aten.gt, aten.sum]
triton_poi_fused_gt_mul_sub_sum_2.run(buf1, buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [input_cumsum, sub_1, tau, tau_1, sub_2, output, log], Original ATen: [aten.sub, aten.gather, aten.div, aten.clamp, aten.log]
triton_poi_fused_clamp_div_gather_log_sub_3.run(buf0, buf4, buf3, buf5, arg0_1, 256, grid=grid(256), stream=stream0)
del arg0_1
del buf0
del buf3
del buf4
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)
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.autograd import Function
import torch
import torch.nn.init
import torch.nn as nn
def _make_ix_like(input, dim=0):
d = input.size(dim)
rho = torch.arange(1, d + 1, device=input.device, dtype=input.dtype)
view = [1] * input.dim()
view[0] = -1
return rho.view(view).transpose(0, dim)
def _threshold_and_support(input, dim=0):
"""Sparsemax building block: compute the threshold
Args:
input: any dimension
dim: dimension along which to apply the sparsemax
Returns:
the threshold value
"""
input_srt, _ = torch.sort(input, descending=True, dim=dim)
input_cumsum = input_srt.cumsum(dim) - 1
rhos = _make_ix_like(input, dim)
support = rhos * input_srt > input_cumsum
support_size = support.sum(dim=dim).unsqueeze(dim)
tau = input_cumsum.gather(dim, support_size - 1)
tau /= support_size
return tau, support_size
def sparsemax(input, dim=0):
return SparsemaxFunction.apply(input, dim)
class SparsemaxFunction(Function):
@staticmethod
def forward(ctx, input, dim=0):
"""sparsemax: normalizing sparse transform (a la softmax)
Parameters:
input (Tensor): any shape
dim: dimension along which to apply sparsemax
Returns:
output (Tensor): same shape as input
"""
ctx.dim = dim
max_val, _ = input.max(dim=dim, keepdim=True)
input -= max_val
tau, supp_size = _threshold_and_support(input, dim=dim)
output = torch.clamp(input - tau, min=0)
ctx.save_for_backward(supp_size, output)
return output
@staticmethod
def backward(ctx, grad_output):
supp_size, output = ctx.saved_tensors
dim = ctx.dim
grad_input = grad_output.clone()
grad_input[output == 0] = 0
v_hat = grad_input.sum(dim=dim) / supp_size.squeeze()
v_hat = v_hat.unsqueeze(dim)
grad_input = torch.where(output != 0, grad_input - v_hat, grad_input)
return grad_input, None
class LogSparsemax(nn.Module):
def __init__(self, dim=0):
self.dim = dim
super(LogSparsemax, self).__init__()
def forward(self, input):
return torch.log(sparsemax(input, 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
from torch._inductor.runtime.triton_helpers import math as tl_math
from torch.autograd import Function
import torch.nn.init
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_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
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
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def _triton_helper_fn_add0(arg0_0, arg1_0):
tmp0 = arg0_0 + arg1_0
return tmp0
@triton.jit
def triton_per_fused_cumsum_sort_1(in_ptr0, out_ptr0, out_ptr1, xnumel,
rnumel, XBLOCK: tl.constexpr):
xnumel = 64
RBLOCK: tl.constexpr = 4
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 + (x0 + 64 * r1), xmask, other=0.0)
tmp1 = r1
tmp2 = tmp1.to(tl.int16)
tmp3 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp4 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5, _tmp6 = triton_helpers.sort_with_index(tmp3, tmp4, None, 1,
stable=False, descending=True)
tmp7 = tmp5.to(tl.float32)
tmp8 = tl.broadcast_to(tmp7, [XBLOCK, RBLOCK])
tmp9, = tl.associative_scan((tmp8,), 1, _triton_helper_fn_add0)
tl.store(out_ptr0 + (x0 + 64 * r1), tmp5, xmask)
tl.store(out_ptr1 + (x0 + 64 * r1), tmp9, xmask)
@triton.jit
def triton_poi_fused_gt_mul_sub_sum_2(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
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask)
tmp7 = tl.load(in_ptr0 + (64 + x0), xmask)
tmp10 = tl.load(in_ptr1 + (64 + x0), xmask)
tmp15 = tl.load(in_ptr0 + (128 + x0), xmask)
tmp18 = tl.load(in_ptr1 + (128 + x0), xmask)
tmp23 = tl.load(in_ptr0 + (192 + x0), xmask)
tmp26 = tl.load(in_ptr1 + (192 + x0), xmask)
tmp1 = 1.0
tmp2 = tmp1 * tmp0
tmp4 = tmp3 - tmp1
tmp5 = tmp2 > tmp4
tmp6 = tmp5.to(tl.int64)
tmp8 = 2.0
tmp9 = tmp8 * tmp7
tmp11 = tmp10 - tmp1
tmp12 = tmp9 > tmp11
tmp13 = tmp12.to(tl.int64)
tmp14 = tmp6 + tmp13
tmp16 = 3.0
tmp17 = tmp16 * tmp15
tmp19 = tmp18 - tmp1
tmp20 = tmp17 > tmp19
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp14 + tmp21
tmp24 = 4.0
tmp25 = tmp24 * tmp23
tmp27 = tmp26 - tmp1
tmp28 = tmp25 > tmp27
tmp29 = tmp28.to(tl.int64)
tmp30 = tmp22 + tmp29
tl.store(out_ptr0 + x0, tmp30, xmask)
@triton.jit
def triton_poi_fused_clamp_div_gather_log_sub_3(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
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.full([1], 1, tl.int64)
tmp3 = tmp1 - tmp2
tmp4 = tl.full([XBLOCK], 4, tl.int32)
tmp5 = tmp3 + tmp4
tmp6 = tmp3 < 0
tmp7 = tl.where(tmp6, tmp5, tmp3)
tl.device_assert((0 <= tmp7) & (tmp7 < 4) | ~xmask,
'index out of bounds: 0 <= tmp7 < 4')
tmp9 = tl.load(in_ptr2 + (x0 + 64 * tmp7), xmask)
tmp10 = 1.0
tmp11 = tmp9 - tmp10
tmp12 = tmp1.to(tl.float32)
tmp13 = tmp11 / tmp12
tmp14 = tmp0 - tmp13
tmp15 = 0.0
tmp16 = triton_helpers.maximum(tmp14, tmp15)
tmp17 = tl_math.log(tmp16)
tl.store(out_ptr0 + x2, tmp17, xmask)
tl.store(out_ptr1 + x2, 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_max_sub_0[grid(256)](arg0_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)
buf3 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_per_fused_cumsum_sort_1[grid(64)](buf0, buf1, buf3, 64, 4,
XBLOCK=32, num_warps=2, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.int64)
triton_poi_fused_gt_mul_sub_sum_2[grid(64)](buf1, buf3, buf4, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf5 = buf1
del buf1
triton_poi_fused_clamp_div_gather_log_sub_3[grid(256)](buf0, buf4,
buf3, buf5, arg0_1, 256, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
del buf0
del buf3
del buf4
return buf5,
def _make_ix_like(input, dim=0):
d = input.size(dim)
rho = torch.arange(1, d + 1, device=input.device, dtype=input.dtype)
view = [1] * input.dim()
view[0] = -1
return rho.view(view).transpose(0, dim)
def _threshold_and_support(input, dim=0):
"""Sparsemax building block: compute the threshold
Args:
input: any dimension
dim: dimension along which to apply the sparsemax
Returns:
the threshold value
"""
input_srt, _ = torch.sort(input, descending=True, dim=dim)
input_cumsum = input_srt.cumsum(dim) - 1
rhos = _make_ix_like(input, dim)
support = rhos * input_srt > input_cumsum
support_size = support.sum(dim=dim).unsqueeze(dim)
tau = input_cumsum.gather(dim, support_size - 1)
tau /= support_size
return tau, support_size
def sparsemax(input, dim=0):
return SparsemaxFunction.apply(input, dim)
class SparsemaxFunction(Function):
@staticmethod
def forward(ctx, input, dim=0):
"""sparsemax: normalizing sparse transform (a la softmax)
Parameters:
input (Tensor): any shape
dim: dimension along which to apply sparsemax
Returns:
output (Tensor): same shape as input
"""
ctx.dim = dim
max_val, _ = input.max(dim=dim, keepdim=True)
input -= max_val
tau, supp_size = _threshold_and_support(input, dim=dim)
output = torch.clamp(input - tau, min=0)
ctx.save_for_backward(supp_size, output)
return output
@staticmethod
def backward(ctx, grad_output):
supp_size, output = ctx.saved_tensors
dim = ctx.dim
grad_input = grad_output.clone()
grad_input[output == 0] = 0
v_hat = grad_input.sum(dim=dim) / supp_size.squeeze()
v_hat = v_hat.unsqueeze(dim)
grad_input = torch.where(output != 0, grad_input - v_hat, grad_input)
return grad_input, None
class LogSparsemaxNew(nn.Module):
def __init__(self, dim=0):
self.dim = dim
super(LogSparsemaxNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| gililior/qasrl-modeling | LogSparsemax | false | 6,747 | [
"MIT"
] | 1 | 2f9684536f6d5f0283b0e4b90a911ea12fa72f72 | https://github.com/gililior/qasrl-modeling/tree/2f9684536f6d5f0283b0e4b90a911ea12fa72f72 | from torch.autograd import Function
import torch
import torch.nn.init
import torch.nn as nn
def _make_ix_like(input, dim=0):
d = input.size(dim)
rho = torch.arange(1, d + 1, device=input.device, dtype=input.dtype)
view = [1] * input.dim()
view[0] = -1
return rho.view(view).transpose(0, dim)
def _threshold_and_support(input, dim=0):
"""Sparsemax building block: compute the threshold
Args:
input: any dimension
dim: dimension along which to apply the sparsemax
Returns:
the threshold value
"""
input_srt, _ = torch.sort(input, descending=True, dim=dim)
input_cumsum = input_srt.cumsum(dim) - 1
rhos = _make_ix_like(input, dim)
support = rhos * input_srt > input_cumsum
support_size = support.sum(dim=dim).unsqueeze(dim)
tau = input_cumsum.gather(dim, support_size - 1)
tau /= support_size
return tau, support_size
def sparsemax(input, dim=0):
return SparsemaxFunction.apply(input, dim)
class SparsemaxFunction(Function):
@staticmethod
def forward(ctx, input, dim=0):
"""sparsemax: normalizing sparse transform (a la softmax)
Parameters:
input (Tensor): any shape
dim: dimension along which to apply sparsemax
Returns:
output (Tensor): same shape as input
"""
ctx.dim = dim
max_val, _ = input.max(dim=dim, keepdim=True)
input -= max_val
tau, supp_size = _threshold_and_support(input, dim=dim)
output = torch.clamp(input - tau, min=0)
ctx.save_for_backward(supp_size, output)
return output
@staticmethod
def backward(ctx, grad_output):
supp_size, output = ctx.saved_tensors
dim = ctx.dim
grad_input = grad_output.clone()
grad_input[output == 0] = 0
v_hat = grad_input.sum(dim=dim) / supp_size.squeeze()
v_hat = v_hat.unsqueeze(dim)
grad_input = torch.where(output != 0, grad_input - v_hat, grad_input)
return grad_input, None
class Model(nn.Module):
def __init__(self, dim=0):
self.dim = dim
super().__init__()
def forward(self, input):
return torch.log(sparsemax(input, self.dim))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
RegressionSubNet | # 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_4/inductor_cache/au/caug6nesiygukdpkrndsclkfho3dygoeotjtbnihl4wlyyiuddug.py
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# x => convolution
# x_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], [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=[16384],
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 = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 16) % 256
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_4/inductor_cache/pa/cpahjig7y6mkcerz4vyyoal5i3jiipcxcxxekkaeapkt2lc7l2nj.py
# Topologically Sorted Source Nodes: [reshape], Original ATen: [aten.clone, aten._unsafe_view]
# Source node to ATen node mapping:
# reshape => clone, view
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute,), kwargs = {memory_format: torch.contiguous_format})
# %view : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%clone, [4, 144, 4]), kwargs = {})
triton_poi_fused__unsafe_view_clone_1 = async_compile.triton('triton_poi_fused__unsafe_view_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, 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), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__unsafe_view_clone_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__unsafe_view_clone_1(in_out_ptr0, in_ptr0, in_ptr1, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 36
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 % 16
y1 = (yindex // 16)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x2) + (576*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.debug_barrier()
tl.store(in_out_ptr0 + (x2 + (36*y3)), 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 = args
args.clear()
assert_size_stride(primals_1, (256, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_5, (256, ), (1, ))
assert_size_stride(primals_6, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_7, (256, ), (1, ))
assert_size_stride(primals_8, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_9, (256, ), (1, ))
assert_size_stride(primals_10, (36, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_11, (36, ), (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, 256, 4, 4), (4096, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.convolution, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_0.run(buf1, primals_2, 16384, grid=grid(16384), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [x_2], 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, 256, 4, 4), (4096, 16, 4, 1))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [x_2, x_3], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_0.run(buf3, primals_5, 16384, grid=grid(16384), stream=stream0)
del primals_5
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 256, 4, 4), (4096, 16, 4, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [x_4, x_5], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_0.run(buf5, primals_7, 16384, grid=grid(16384), stream=stream0)
del primals_7
# Topologically Sorted Source Nodes: [x_6], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 256, 4, 4), (4096, 16, 4, 1))
buf7 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [x_6, x_7], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_0.run(buf7, primals_9, 16384, grid=grid(16384), stream=stream0)
del primals_9
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.convolution]
buf8 = extern_kernels.convolution(buf7, primals_10, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 36, 4, 4), (576, 16, 4, 1))
buf9 = empty_strided_cuda((4, 4, 4, 36), (576, 144, 36, 1), torch.float32)
buf10 = reinterpret_tensor(buf9, (4, 144, 4), (576, 4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [reshape], Original ATen: [aten.clone, aten._unsafe_view]
triton_poi_fused__unsafe_view_clone_1.run(buf10, buf8, primals_11, 64, 36, grid=grid(64, 36), stream=stream0)
del buf8
del primals_11
return (buf10, primals_1, primals_3, primals_4, primals_6, primals_8, primals_10, buf1, buf3, buf5, 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((256, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (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((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((36, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((36, ), (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
class RegressionSubNet(nn.Module):
def __init__(self, in_channels, num_anchors=9):
super().__init__()
self.conv2d_1 = nn.Conv2d(in_channels, 256, 3, padding=1)
nn.init.normal_(self.conv2d_1.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_1.bias.data)
self.conv2d_2 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_2.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_2.bias.data)
self.conv2d_3 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_3.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_3.bias.data)
self.conv2d_4 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_4.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_4.bias.data)
self.conv2d_5 = nn.Conv2d(256, 4 * num_anchors, 3, padding=1)
nn.init.normal_(self.conv2d_5.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_5.bias.data)
def forward(self, x):
x = self.conv2d_1(x)
x = nn.functional.relu(x)
x = self.conv2d_2(x)
x = nn.functional.relu(x)
x = self.conv2d_3(x)
x = nn.functional.relu(x)
x = self.conv2d_4(x)
x = nn.functional.relu(x)
x = self.conv2d_5(x)
x = x.permute(0, 2, 3, 1)
return x.reshape(x.size(0), -1, 4)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_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
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_relu_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 // 16 % 256
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__unsafe_view_clone_1(in_out_ptr0, in_ptr0, in_ptr1,
ynumel, xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 36
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 % 16
y1 = yindex // 16
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x2 + 576 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tl.debug_barrier()
tl.store(in_out_ptr0 + (x2 + 36 * y3), 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) = args
args.clear()
assert_size_stride(primals_1, (256, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_5, (256,), (1,))
assert_size_stride(primals_6, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_7, (256,), (1,))
assert_size_stride(primals_8, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_9, (256,), (1,))
assert_size_stride(primals_10, (36, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_11, (36,), (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, 256, 4, 4), (4096, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_relu_0[grid(16384)](buf1, primals_2,
16384, 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, 256, 4, 4), (4096, 16, 4, 1))
buf3 = buf2
del buf2
triton_poi_fused_convolution_relu_0[grid(16384)](buf3, primals_5,
16384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 256, 4, 4), (4096, 16, 4, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_relu_0[grid(16384)](buf5, primals_7,
16384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_7
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 256, 4, 4), (4096, 16, 4, 1))
buf7 = buf6
del buf6
triton_poi_fused_convolution_relu_0[grid(16384)](buf7, primals_9,
16384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_9
buf8 = extern_kernels.convolution(buf7, primals_10, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 36, 4, 4), (576, 16, 4, 1))
buf9 = empty_strided_cuda((4, 4, 4, 36), (576, 144, 36, 1), torch.
float32)
buf10 = reinterpret_tensor(buf9, (4, 144, 4), (576, 4, 1), 0)
del buf9
triton_poi_fused__unsafe_view_clone_1[grid(64, 36)](buf10, buf8,
primals_11, 64, 36, XBLOCK=64, YBLOCK=4, num_warps=4, num_stages=1)
del buf8
del primals_11
return (buf10, primals_1, primals_3, primals_4, primals_6, primals_8,
primals_10, buf1, buf3, buf5, buf7)
class RegressionSubNetNew(nn.Module):
def __init__(self, in_channels, num_anchors=9):
super().__init__()
self.conv2d_1 = nn.Conv2d(in_channels, 256, 3, padding=1)
nn.init.normal_(self.conv2d_1.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_1.bias.data)
self.conv2d_2 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_2.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_2.bias.data)
self.conv2d_3 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_3.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_3.bias.data)
self.conv2d_4 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_4.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_4.bias.data)
self.conv2d_5 = nn.Conv2d(256, 4 * num_anchors, 3, padding=1)
nn.init.normal_(self.conv2d_5.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_5.bias.data)
def forward(self, input_0):
primals_1 = self.conv2d_1.weight
primals_2 = self.conv2d_1.bias
primals_4 = self.conv2d_2.weight
primals_5 = self.conv2d_2.bias
primals_6 = self.conv2d_3.weight
primals_7 = self.conv2d_3.bias
primals_8 = self.conv2d_4.weight
primals_9 = self.conv2d_4.bias
primals_10 = self.conv2d_5.weight
primals_11 = self.conv2d_5.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])
return output[0]
| geez0219/ARC | RegressionSubNet | false | 6,748 | [
"Apache-2.0"
] | 1 | f2176f0d442d4a2d6028f0770b1efc1a9ae982b8 | https://github.com/geez0219/ARC/tree/f2176f0d442d4a2d6028f0770b1efc1a9ae982b8 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_channels, num_anchors=9):
super().__init__()
self.conv2d_1 = nn.Conv2d(in_channels, 256, 3, padding=1)
nn.init.normal_(self.conv2d_1.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_1.bias.data)
self.conv2d_2 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_2.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_2.bias.data)
self.conv2d_3 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_3.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_3.bias.data)
self.conv2d_4 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_4.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_4.bias.data)
self.conv2d_5 = nn.Conv2d(256, 4 * num_anchors, 3, padding=1)
nn.init.normal_(self.conv2d_5.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_5.bias.data)
def forward(self, x):
x = self.conv2d_1(x)
x = nn.functional.relu(x)
x = self.conv2d_2(x)
x = nn.functional.relu(x)
x = self.conv2d_3(x)
x = nn.functional.relu(x)
x = self.conv2d_4(x)
x = nn.functional.relu(x)
x = self.conv2d_5(x)
x = x.permute(0, 2, 3, 1)
return x.reshape(x.size(0), -1, 4)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
Attention | # 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_4/inductor_cache/nc/cncwsucylpsg2zmlivjfxu6vbd64ztxjndlsix2ysjtby3xohgk4.py
# Topologically Sorted Source Nodes: [x_tr], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# x_tr => 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=[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': ['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 = 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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/2s/c2s2pec3vduo4xn2kfu53hypzbhir2ql56lmvazfmymhwv2ehhv5.py
# Topologically Sorted Source Nodes: [weights_1], Original ATen: [aten.sigmoid]
# Source node to ATen node mapping:
# weights_1 => sigmoid
# Graph fragment:
# %sigmoid : [num_users=2] = call_function[target=torch.ops.aten.sigmoid.default](args = (%view_3,), kwargs = {})
triton_poi_fused_sigmoid_1 = async_compile.triton('triton_poi_fused_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: '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_sigmoid_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_sigmoid_1(in_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_out_ptr0 + (x0), xmask)
tmp1 = tl.sigmoid(tmp0)
tl.store(in_out_ptr0 + (x0), tmp1, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/kn/cknh33objkrsgk5p4pkgfisw7sw6hhzeaun22c5v5blgmcgpx2ka.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# x => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_3, %sigmoid), 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=[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_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 = 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_ptr1 + (x1), 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):
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, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (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: [], 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
# Topologically Sorted Source Nodes: [x_tr], Original ATen: [aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_tanh_0.run(buf1, primals_2, 256, grid=grid(256), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [weights], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 1), (1, 4), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [weights_1], Original ATen: [aten.sigmoid]
triton_poi_fused_sigmoid_1.run(buf3, 64, grid=grid(64), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mul]
triton_poi_fused_mul_2.run(primals_3, buf3, buf4, 256, grid=grid(256), stream=stream0)
return (buf4, buf3, primals_3, buf1, buf3, 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, 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((1, 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
class Attention(nn.Module):
def __init__(self, dims, norm=False):
super(Attention, self).__init__()
self.norm = norm
if self.norm:
self.constrain = L2Constrain()
else:
self.transform = nn.Linear(dims, dims)
self.context_vector = nn.Linear(dims, 1, bias=False)
self.reset_parameters()
def forward(self, x):
if self.norm:
weights = self.context_vector(x)
weights = torch.add(torch.div(weights, 2.0), 0.5)
else:
x_tr = torch.tanh(self.transform(x))
weights = self.context_vector(x_tr)
weights = torch.sigmoid(weights)
x = x * weights
return x, weights
def reset_parameters(self):
if self.norm:
nn.init.normal_(self.context_vector.weight)
self.constrain(self.context_vector)
else:
nn.init.xavier_uniform_(self.context_vector.weight)
nn.init.xavier_uniform_(self.transform.weight)
nn.init.zeros_(self.transform.bias)
def apply_contraint(self):
if self.norm:
self.constrain(self.context_vector)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'dims': 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_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
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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, xmask)
@triton.jit
def triton_poi_fused_sigmoid_1(in_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_out_ptr0 + x0, xmask)
tmp1 = tl.sigmoid(tmp0)
tl.store(in_out_ptr0 + x0, tmp1, xmask)
@triton.jit
def triton_poi_fused_mul_2(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
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tl.store(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, 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, (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_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
get_raw_stream(0)
triton_poi_fused_tanh_0[grid(256)](buf1, primals_2, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 1), (1, 4), 0), out=buf2)
buf3 = reinterpret_tensor(buf2, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf2
triton_poi_fused_sigmoid_1[grid(64)](buf3, 64, XBLOCK=64, num_warps
=1, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_mul_2[grid(256)](primals_3, buf3, buf4, 256,
XBLOCK=128, num_warps=4, num_stages=1)
return buf4, buf3, primals_3, buf1, buf3, primals_4
class AttentionNew(nn.Module):
def __init__(self, dims, norm=False):
super(AttentionNew, self).__init__()
self.norm = norm
if self.norm:
self.constrain = L2Constrain()
else:
self.transform = nn.Linear(dims, dims)
self.context_vector = nn.Linear(dims, 1, bias=False)
self.reset_parameters()
def reset_parameters(self):
if self.norm:
nn.init.normal_(self.context_vector.weight)
self.constrain(self.context_vector)
else:
nn.init.xavier_uniform_(self.context_vector.weight)
nn.init.xavier_uniform_(self.transform.weight)
nn.init.zeros_(self.transform.bias)
def apply_contraint(self):
if self.norm:
self.constrain(self.context_vector)
def forward(self, input_0):
primals_1 = self.transform.weight
primals_2 = self.transform.bias
primals_4 = self.context_vector.weight
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0], output[1]
| glee1228/segment_temporal_context_aggregation | Attention | false | 6,749 | [
"Apache-2.0"
] | 1 | e5778f848f1cfd89bd1f77beb5e1b38a66a2f13d | https://github.com/glee1228/segment_temporal_context_aggregation/tree/e5778f848f1cfd89bd1f77beb5e1b38a66a2f13d | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, dims, norm=False):
super().__init__()
self.norm = norm
if self.norm:
self.constrain = L2Constrain()
else:
self.transform = nn.Linear(dims, dims)
self.context_vector = nn.Linear(dims, 1, bias=False)
self.reset_parameters()
def forward(self, x):
if self.norm:
weights = self.context_vector(x)
weights = torch.add(torch.div(weights, 2.0), 0.5)
else:
x_tr = torch.tanh(self.transform(x))
weights = self.context_vector(x_tr)
weights = torch.sigmoid(weights)
x = x * weights
return x, weights
def reset_parameters(self):
if self.norm:
nn.init.normal_(self.context_vector.weight)
self.constrain(self.context_vector)
else:
nn.init.xavier_uniform_(self.context_vector.weight)
nn.init.xavier_uniform_(self.transform.weight)
nn.init.zeros_(self.transform.bias)
def apply_contraint(self):
if self.norm:
self.constrain(self.context_vector)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
Sparsemax | # 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_4/inductor_cache/xr/cxriojqnst2q32kjzwrpwltjdtloxmuqf3cbxuvng6r4ts5aacce.py
# Topologically Sorted Source Nodes: [max_1, input_1], Original ATen: [aten.max, aten.sub]
# Source node to ATen node mapping:
# input_1 => sub
# max_1 => max_1
# Graph fragment:
# %max_1 : [num_users=1] = call_function[target=torch.ops.aten.max.dim](args = (%arg0_1, 0, True), kwargs = {})
# %sub : [num_users=3] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %getitem), kwargs = {})
triton_poi_fused_max_sub_0 = async_compile.triton('triton_poi_fused_max_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_max_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_max_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
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
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/43/c43uhskbeaaf4eboq6purohj7g5a6yrwr4b6mkhwl4lu2iqcj4cs.py
# Topologically Sorted Source Nodes: [sort, cumsum], Original ATen: [aten.sort, aten.cumsum]
# Source node to ATen node mapping:
# cumsum => cumsum
# sort => sort
# Graph fragment:
# %sort : [num_users=1] = call_function[target=torch.ops.aten.sort.default](args = (%sub, 0, True), kwargs = {})
# %cumsum : [num_users=1] = call_function[target=torch.ops.aten.cumsum.default](args = (%getitem_2, 0), kwargs = {})
triton_per_fused_cumsum_sort_1 = async_compile.triton('triton_per_fused_cumsum_sort_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.jit
def _triton_helper_fn_add0(arg0_0, arg1_0):
tmp0 = arg0_0 + arg1_0
return tmp0
@triton_heuristics.persistent_reduction(
size_hints=[64, 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': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_cumsum_sort_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}
)
@triton.jit
def triton_per_fused_cumsum_sort_1(in_ptr0, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 64
rnumel = 4
RBLOCK: tl.constexpr = 4
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 + (x0 + (64*r1)), xmask, other=0.0)
tmp1 = r1
tmp2 = tmp1.to(tl.int16)
tmp3 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp4 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5, tmp6, = triton_helpers.sort_with_index(tmp3, tmp4, None, 1, stable=False, descending=True)
tmp7 = tmp5.to(tl.float32)
tmp8 = tl.broadcast_to(tmp7, [XBLOCK, RBLOCK])
tmp9, = tl.associative_scan((tmp8,), 1, _triton_helper_fn_add0)
tl.store(out_ptr0 + (x0 + (64*r1)), tmp5, xmask)
tl.store(out_ptr1 + (x0 + (64*r1)), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/rc/crclwvykuaxrcwig3dfr3ygs4skmnzpu4fkszzcyzp6jebodtdn6.py
# Topologically Sorted Source Nodes: [mul, input_cumsum, support, sum_1], Original ATen: [aten.mul, aten.sub, aten.gt, aten.sum]
# Source node to ATen node mapping:
# input_cumsum => sub_1
# mul => mul_1
# sum_1 => sum_1
# support => gt
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %getitem_2), kwargs = {})
# %sub_1 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%cumsum, 1), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Tensor](args = (%mul_1, %sub_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%gt, [0]), kwargs = {})
triton_poi_fused_gt_mul_sub_sum_2 = async_compile.triton('triton_poi_fused_gt_mul_sub_sum_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: '*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_gt_mul_sub_sum_2', '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_gt_mul_sub_sum_2(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
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp3 = tl.load(in_ptr1 + (x0), xmask)
tmp7 = tl.load(in_ptr0 + (64 + x0), xmask)
tmp10 = tl.load(in_ptr1 + (64 + x0), xmask)
tmp15 = tl.load(in_ptr0 + (128 + x0), xmask)
tmp18 = tl.load(in_ptr1 + (128 + x0), xmask)
tmp23 = tl.load(in_ptr0 + (192 + x0), xmask)
tmp26 = tl.load(in_ptr1 + (192 + x0), xmask)
tmp1 = 1.0
tmp2 = tmp1 * tmp0
tmp4 = tmp3 - tmp1
tmp5 = tmp2 > tmp4
tmp6 = tmp5.to(tl.int64)
tmp8 = 2.0
tmp9 = tmp8 * tmp7
tmp11 = tmp10 - tmp1
tmp12 = tmp9 > tmp11
tmp13 = tmp12.to(tl.int64)
tmp14 = tmp6 + tmp13
tmp16 = 3.0
tmp17 = tmp16 * tmp15
tmp19 = tmp18 - tmp1
tmp20 = tmp17 > tmp19
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp14 + tmp21
tmp24 = 4.0
tmp25 = tmp24 * tmp23
tmp27 = tmp26 - tmp1
tmp28 = tmp25 > tmp27
tmp29 = tmp28.to(tl.int64)
tmp30 = tmp22 + tmp29
tl.store(out_ptr0 + (x0), tmp30, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/6b/c6bfbjhvimrn26eikgp6xm64w3s7epyszepdqbzzk274qgul6f3n.py
# Topologically Sorted Source Nodes: [input_cumsum, sub_1, tau, tau_1, sub_2, output], Original ATen: [aten.sub, aten.gather, aten.div, aten.clamp]
# Source node to ATen node mapping:
# input_cumsum => sub_1
# output => clamp_min
# sub_1 => sub_2
# sub_2 => sub_3
# tau => gather
# tau_1 => div
# Graph fragment:
# %sub_1 : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%cumsum, 1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%unsqueeze, 1), kwargs = {})
# %gather : [num_users=1] = call_function[target=torch.ops.aten.gather.default](args = (%sub_1, 0, %sub_2), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%gather, %unsqueeze), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %div), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sub_3, 0), kwargs = {})
# %copy_ : [num_users=0] = call_function[target=torch.ops.aten.copy_.default](args = (%arg0_1, %sub), kwargs = {})
triton_poi_fused_clamp_div_gather_sub_3 = async_compile.triton('triton_poi_fused_clamp_div_gather_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=[256],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i64', 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_clamp_div_gather_sub_3', 'mutated_arg_names': ['out_ptr1'], '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_clamp_div_gather_sub_3(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
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.full([1], 1, tl.int64)
tmp3 = tmp1 - tmp2
tmp4 = tl.full([XBLOCK], 4, tl.int32)
tmp5 = tmp3 + tmp4
tmp6 = tmp3 < 0
tmp7 = tl.where(tmp6, tmp5, tmp3)
tl.device_assert(((0 <= tmp7) & (tmp7 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp7 < 4")
tmp9 = tl.load(in_ptr2 + (x0 + (64*tmp7)), xmask)
tmp10 = 1.0
tmp11 = tmp9 - tmp10
tmp12 = tmp1.to(tl.float32)
tmp13 = tmp11 / tmp12
tmp14 = tmp0 - tmp13
tmp15 = 0.0
tmp16 = triton_helpers.maximum(tmp14, tmp15)
tl.store(out_ptr0 + (x2), tmp16, xmask)
tl.store(out_ptr1 + (x2), 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: [max_1, input_1], Original ATen: [aten.max, aten.sub]
stream0 = get_raw_stream(0)
triton_poi_fused_max_sub_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
buf1 = 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.float32)
# Topologically Sorted Source Nodes: [sort, cumsum], Original ATen: [aten.sort, aten.cumsum]
triton_per_fused_cumsum_sort_1.run(buf0, buf1, buf3, 64, 4, grid=grid(64), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.int64)
# Topologically Sorted Source Nodes: [mul, input_cumsum, support, sum_1], Original ATen: [aten.mul, aten.sub, aten.gt, aten.sum]
triton_poi_fused_gt_mul_sub_sum_2.run(buf1, buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [input_cumsum, sub_1, tau, tau_1, sub_2, output], Original ATen: [aten.sub, aten.gather, aten.div, aten.clamp]
triton_poi_fused_clamp_div_gather_sub_3.run(buf0, buf4, buf3, buf5, arg0_1, 256, grid=grid(256), stream=stream0)
del arg0_1
del buf0
del buf3
del buf4
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)
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.autograd import Function
import torch
import torch.nn.init
import torch.nn as nn
def _make_ix_like(input, dim=0):
d = input.size(dim)
rho = torch.arange(1, d + 1, device=input.device, dtype=input.dtype)
view = [1] * input.dim()
view[0] = -1
return rho.view(view).transpose(0, dim)
def _threshold_and_support(input, dim=0):
"""Sparsemax building block: compute the threshold
Args:
input: any dimension
dim: dimension along which to apply the sparsemax
Returns:
the threshold value
"""
input_srt, _ = torch.sort(input, descending=True, dim=dim)
input_cumsum = input_srt.cumsum(dim) - 1
rhos = _make_ix_like(input, dim)
support = rhos * input_srt > input_cumsum
support_size = support.sum(dim=dim).unsqueeze(dim)
tau = input_cumsum.gather(dim, support_size - 1)
tau /= support_size
return tau, support_size
def sparsemax(input, dim=0):
return SparsemaxFunction.apply(input, dim)
class SparsemaxFunction(Function):
@staticmethod
def forward(ctx, input, dim=0):
"""sparsemax: normalizing sparse transform (a la softmax)
Parameters:
input (Tensor): any shape
dim: dimension along which to apply sparsemax
Returns:
output (Tensor): same shape as input
"""
ctx.dim = dim
max_val, _ = input.max(dim=dim, keepdim=True)
input -= max_val
tau, supp_size = _threshold_and_support(input, dim=dim)
output = torch.clamp(input - tau, min=0)
ctx.save_for_backward(supp_size, output)
return output
@staticmethod
def backward(ctx, grad_output):
supp_size, output = ctx.saved_tensors
dim = ctx.dim
grad_input = grad_output.clone()
grad_input[output == 0] = 0
v_hat = grad_input.sum(dim=dim) / supp_size.squeeze()
v_hat = v_hat.unsqueeze(dim)
grad_input = torch.where(output != 0, grad_input - v_hat, grad_input)
return grad_input, None
class Sparsemax(nn.Module):
def __init__(self, dim=0):
self.dim = dim
super(Sparsemax, self).__init__()
def forward(self, input):
return sparsemax(input, 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
from torch.autograd import Function
import torch.nn.init
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_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
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
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def _triton_helper_fn_add0(arg0_0, arg1_0):
tmp0 = arg0_0 + arg1_0
return tmp0
@triton.jit
def triton_per_fused_cumsum_sort_1(in_ptr0, out_ptr0, out_ptr1, xnumel,
rnumel, XBLOCK: tl.constexpr):
xnumel = 64
RBLOCK: tl.constexpr = 4
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 + (x0 + 64 * r1), xmask, other=0.0)
tmp1 = r1
tmp2 = tmp1.to(tl.int16)
tmp3 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp4 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5, _tmp6 = triton_helpers.sort_with_index(tmp3, tmp4, None, 1,
stable=False, descending=True)
tmp7 = tmp5.to(tl.float32)
tmp8 = tl.broadcast_to(tmp7, [XBLOCK, RBLOCK])
tmp9, = tl.associative_scan((tmp8,), 1, _triton_helper_fn_add0)
tl.store(out_ptr0 + (x0 + 64 * r1), tmp5, xmask)
tl.store(out_ptr1 + (x0 + 64 * r1), tmp9, xmask)
@triton.jit
def triton_poi_fused_gt_mul_sub_sum_2(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
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp3 = tl.load(in_ptr1 + x0, xmask)
tmp7 = tl.load(in_ptr0 + (64 + x0), xmask)
tmp10 = tl.load(in_ptr1 + (64 + x0), xmask)
tmp15 = tl.load(in_ptr0 + (128 + x0), xmask)
tmp18 = tl.load(in_ptr1 + (128 + x0), xmask)
tmp23 = tl.load(in_ptr0 + (192 + x0), xmask)
tmp26 = tl.load(in_ptr1 + (192 + x0), xmask)
tmp1 = 1.0
tmp2 = tmp1 * tmp0
tmp4 = tmp3 - tmp1
tmp5 = tmp2 > tmp4
tmp6 = tmp5.to(tl.int64)
tmp8 = 2.0
tmp9 = tmp8 * tmp7
tmp11 = tmp10 - tmp1
tmp12 = tmp9 > tmp11
tmp13 = tmp12.to(tl.int64)
tmp14 = tmp6 + tmp13
tmp16 = 3.0
tmp17 = tmp16 * tmp15
tmp19 = tmp18 - tmp1
tmp20 = tmp17 > tmp19
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp14 + tmp21
tmp24 = 4.0
tmp25 = tmp24 * tmp23
tmp27 = tmp26 - tmp1
tmp28 = tmp25 > tmp27
tmp29 = tmp28.to(tl.int64)
tmp30 = tmp22 + tmp29
tl.store(out_ptr0 + x0, tmp30, xmask)
@triton.jit
def triton_poi_fused_clamp_div_gather_sub_3(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
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.full([1], 1, tl.int64)
tmp3 = tmp1 - tmp2
tmp4 = tl.full([XBLOCK], 4, tl.int32)
tmp5 = tmp3 + tmp4
tmp6 = tmp3 < 0
tmp7 = tl.where(tmp6, tmp5, tmp3)
tl.device_assert((0 <= tmp7) & (tmp7 < 4) | ~xmask,
'index out of bounds: 0 <= tmp7 < 4')
tmp9 = tl.load(in_ptr2 + (x0 + 64 * tmp7), xmask)
tmp10 = 1.0
tmp11 = tmp9 - tmp10
tmp12 = tmp1.to(tl.float32)
tmp13 = tmp11 / tmp12
tmp14 = tmp0 - tmp13
tmp15 = 0.0
tmp16 = triton_helpers.maximum(tmp14, tmp15)
tl.store(out_ptr0 + x2, tmp16, xmask)
tl.store(out_ptr1 + x2, 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_max_sub_0[grid(256)](arg0_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf1 = 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.float32)
triton_per_fused_cumsum_sort_1[grid(64)](buf0, buf1, buf3, 64, 4,
XBLOCK=32, num_warps=2, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.int64)
triton_poi_fused_gt_mul_sub_sum_2[grid(64)](buf1, buf3, buf4, 64,
XBLOCK=64, num_warps=1, num_stages=1)
buf5 = buf1
del buf1
triton_poi_fused_clamp_div_gather_sub_3[grid(256)](buf0, buf4, buf3,
buf5, arg0_1, 256, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
del buf0
del buf3
del buf4
return buf5,
def _make_ix_like(input, dim=0):
d = input.size(dim)
rho = torch.arange(1, d + 1, device=input.device, dtype=input.dtype)
view = [1] * input.dim()
view[0] = -1
return rho.view(view).transpose(0, dim)
def _threshold_and_support(input, dim=0):
"""Sparsemax building block: compute the threshold
Args:
input: any dimension
dim: dimension along which to apply the sparsemax
Returns:
the threshold value
"""
input_srt, _ = torch.sort(input, descending=True, dim=dim)
input_cumsum = input_srt.cumsum(dim) - 1
rhos = _make_ix_like(input, dim)
support = rhos * input_srt > input_cumsum
support_size = support.sum(dim=dim).unsqueeze(dim)
tau = input_cumsum.gather(dim, support_size - 1)
tau /= support_size
return tau, support_size
def sparsemax(input, dim=0):
return SparsemaxFunction.apply(input, dim)
class SparsemaxFunction(Function):
@staticmethod
def forward(ctx, input, dim=0):
"""sparsemax: normalizing sparse transform (a la softmax)
Parameters:
input (Tensor): any shape
dim: dimension along which to apply sparsemax
Returns:
output (Tensor): same shape as input
"""
ctx.dim = dim
max_val, _ = input.max(dim=dim, keepdim=True)
input -= max_val
tau, supp_size = _threshold_and_support(input, dim=dim)
output = torch.clamp(input - tau, min=0)
ctx.save_for_backward(supp_size, output)
return output
@staticmethod
def backward(ctx, grad_output):
supp_size, output = ctx.saved_tensors
dim = ctx.dim
grad_input = grad_output.clone()
grad_input[output == 0] = 0
v_hat = grad_input.sum(dim=dim) / supp_size.squeeze()
v_hat = v_hat.unsqueeze(dim)
grad_input = torch.where(output != 0, grad_input - v_hat, grad_input)
return grad_input, None
class SparsemaxNew(nn.Module):
def __init__(self, dim=0):
self.dim = dim
super(SparsemaxNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| gililior/qasrl-modeling | Sparsemax | false | 6,750 | [
"MIT"
] | 1 | 2f9684536f6d5f0283b0e4b90a911ea12fa72f72 | https://github.com/gililior/qasrl-modeling/tree/2f9684536f6d5f0283b0e4b90a911ea12fa72f72 | from torch.autograd import Function
import torch
import torch.nn.init
import torch.nn as nn
def _make_ix_like(input, dim=0):
d = input.size(dim)
rho = torch.arange(1, d + 1, device=input.device, dtype=input.dtype)
view = [1] * input.dim()
view[0] = -1
return rho.view(view).transpose(0, dim)
def _threshold_and_support(input, dim=0):
"""Sparsemax building block: compute the threshold
Args:
input: any dimension
dim: dimension along which to apply the sparsemax
Returns:
the threshold value
"""
input_srt, _ = torch.sort(input, descending=True, dim=dim)
input_cumsum = input_srt.cumsum(dim) - 1
rhos = _make_ix_like(input, dim)
support = rhos * input_srt > input_cumsum
support_size = support.sum(dim=dim).unsqueeze(dim)
tau = input_cumsum.gather(dim, support_size - 1)
tau /= support_size
return tau, support_size
def sparsemax(input, dim=0):
return SparsemaxFunction.apply(input, dim)
class SparsemaxFunction(Function):
@staticmethod
def forward(ctx, input, dim=0):
"""sparsemax: normalizing sparse transform (a la softmax)
Parameters:
input (Tensor): any shape
dim: dimension along which to apply sparsemax
Returns:
output (Tensor): same shape as input
"""
ctx.dim = dim
max_val, _ = input.max(dim=dim, keepdim=True)
input -= max_val
tau, supp_size = _threshold_and_support(input, dim=dim)
output = torch.clamp(input - tau, min=0)
ctx.save_for_backward(supp_size, output)
return output
@staticmethod
def backward(ctx, grad_output):
supp_size, output = ctx.saved_tensors
dim = ctx.dim
grad_input = grad_output.clone()
grad_input[output == 0] = 0
v_hat = grad_input.sum(dim=dim) / supp_size.squeeze()
v_hat = v_hat.unsqueeze(dim)
grad_input = torch.where(output != 0, grad_input - v_hat, grad_input)
return grad_input, None
class Model(nn.Module):
def __init__(self, dim=0):
self.dim = dim
super().__init__()
def forward(self, input):
return sparsemax(input, self.dim)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
ActorNetwork | # 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_4/inductor_cache/r3/cr3febcwm3t44fuoitsx3ou2p6xg4sk4f7unagmmrvffasxf47te.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_2 : [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_4/inductor_cache/rm/crmfikkxblrhxfynyknfm2x3wwcwtibkjkkbyhzwmxqi4kmwkosl.py
# Topologically Sorted Source Nodes: [log_std_1], Original ATen: [aten.clamp, aten.ge, aten.le, aten.logical_and]
# Source node to ATen node mapping:
# log_std_1 => clamp_max, clamp_min
# Graph fragment:
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%view_7, -20), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 2), kwargs = {})
# %ge : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%view_7, -20), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%view_7, 2), kwargs = {})
# %logical_and : [num_users=1] = call_function[target=torch.ops.aten.logical_and.default](args = (%ge, %le), kwargs = {})
triton_poi_fused_clamp_ge_le_logical_and_1 = async_compile.triton('triton_poi_fused_clamp_ge_le_logical_and_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_clamp_ge_le_logical_and_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_clamp_ge_le_logical_and_1(in_ptr0, in_ptr1, 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 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = -20.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 2.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 >= tmp3
tmp8 = tmp2 <= tmp5
tmp9 = tmp7 & tmp8
tl.store(out_ptr0 + (x2), tmp6, xmask)
tl.store(out_ptr1 + (x2), 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, 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, 4), (4, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (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
buf9 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 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, buf9, 256, grid=grid(256), stream=stream0)
del primals_2
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
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 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, buf8, 256, grid=grid(256), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, reinterpret_tensor(buf3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [log_std_1], Original ATen: [aten.clamp, aten.ge, aten.le, aten.logical_and]
triton_poi_fused_clamp_ge_le_logical_and_1.run(buf5, primals_9, buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf5
del primals_9
return (reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0), buf6, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(buf3, (64, 4), (4, 1), 0), buf7, primals_8, primals_6, buf8, primals_4, 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), (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, 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, 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
import torch.nn.functional as F
import torch.utils.data
class ActorNetwork(nn.Module):
def __init__(self, state_size, action_size, hidden_size, init_w=0.003,
log_std_min=-20, log_std_max=2):
super(ActorNetwork, self).__init__()
self.log_std_min = log_std_min
self.log_std_max = log_std_max
self.linear1 = nn.Linear(state_size, hidden_size)
self.linear2 = nn.Linear(hidden_size, hidden_size)
self.mean_linear = nn.Linear(hidden_size, action_size)
self.mean_linear.weight.data.uniform_(-init_w, init_w)
self.mean_linear.bias.data.uniform_(-init_w, init_w)
self.log_std_linear = nn.Linear(hidden_size, action_size)
self.log_std_linear.weight.data.uniform_(-init_w, init_w)
self.log_std_linear.bias.data.uniform_(-init_w, init_w)
def forward(self, state):
x = F.relu(self.linear1(state))
x = F.relu(self.linear2(x))
mean = self.mean_linear(x)
log_std = self.log_std_linear(x)
log_std = torch.clamp(log_std, self.log_std_min, self.log_std_max)
return mean, log_std
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'state_size': 4, 'action_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 import triton_helpers
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_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_clamp_ge_le_logical_and_1(in_ptr0, in_ptr1, 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 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = -20.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 2.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 >= tmp3
tmp8 = tmp2 <= tmp5
tmp9 = tmp7 & tmp8
tl.store(out_ptr0 + x2, tmp6, xmask)
tl.store(out_ptr1 + x2, 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, 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, 4), (4, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (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
buf9 = 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, buf9, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
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
buf8 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_0[grid(256)](buf3,
primals_5, buf8, 256, 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, 4), (
4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf4)
del primals_7
buf5 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf3, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf5)
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_clamp_ge_le_logical_and_1[grid(256)](buf5,
primals_9, buf6, buf7, 256, XBLOCK=128, num_warps=4, num_stages=1)
del buf5
del primals_9
return reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0
), buf6, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(
buf3, (64, 4), (4, 1), 0
), buf7, primals_8, primals_6, buf8, primals_4, buf9
class ActorNetworkNew(nn.Module):
def __init__(self, state_size, action_size, hidden_size, init_w=0.003,
log_std_min=-20, log_std_max=2):
super(ActorNetworkNew, self).__init__()
self.log_std_min = log_std_min
self.log_std_max = log_std_max
self.linear1 = nn.Linear(state_size, hidden_size)
self.linear2 = nn.Linear(hidden_size, hidden_size)
self.mean_linear = nn.Linear(hidden_size, action_size)
self.mean_linear.weight.data.uniform_(-init_w, init_w)
self.mean_linear.bias.data.uniform_(-init_w, init_w)
self.log_std_linear = nn.Linear(hidden_size, action_size)
self.log_std_linear.weight.data.uniform_(-init_w, init_w)
self.log_std_linear.bias.data.uniform_(-init_w, init_w)
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.mean_linear.weight
primals_7 = self.mean_linear.bias
primals_8 = self.log_std_linear.weight
primals_9 = self.log_std_linear.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]
| godnpeter/DMC_Clustering_PICA | ActorNetwork | false | 6,751 | [
"BSD-2-Clause"
] | 1 | 1b3e14dd4034f3941af1caa06c1d4b6f9d606408 | https://github.com/godnpeter/DMC_Clustering_PICA/tree/1b3e14dd4034f3941af1caa06c1d4b6f9d606408 | import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.data
class Model(nn.Module):
def __init__(self, state_size, action_size, hidden_size, init_w=0.003,
log_std_min=-20, log_std_max=2):
super().__init__()
self.log_std_min = log_std_min
self.log_std_max = log_std_max
self.linear1 = nn.Linear(state_size, hidden_size)
self.linear2 = nn.Linear(hidden_size, hidden_size)
self.mean_linear = nn.Linear(hidden_size, action_size)
self.mean_linear.weight.data.uniform_(-init_w, init_w)
self.mean_linear.bias.data.uniform_(-init_w, init_w)
self.log_std_linear = nn.Linear(hidden_size, action_size)
self.log_std_linear.weight.data.uniform_(-init_w, init_w)
self.log_std_linear.bias.data.uniform_(-init_w, init_w)
def forward(self, state):
x = F.relu(self.linear1(state))
x = F.relu(self.linear2(x))
mean = self.mean_linear(x)
log_std = self.log_std_linear(x)
log_std = torch.clamp(log_std, self.log_std_min, self.log_std_max)
return mean, log_std
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 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_4/inductor_cache/lj/clj3mscf4tqz7wlbw6be5v4cyjekfwhnnfmmz6ngsoi4fvpbcy27.py
# Topologically Sorted Source Nodes: [conv2d], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d => 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_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=[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_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 = 144000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 3600) % 10
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_4/inductor_cache/k6/ck6akob3bt2selhn3bjjcpzl2kogpysu4jqltfzvzfkqrhve6olr.py
# Topologically Sorted Source Nodes: [max_pool2d, x], 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 => 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_1 = async_compile.triton('triton_poi_fused_max_pool2d_with_indices_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=[65536],
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_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_max_pool2d_with_indices_relu_1(in_ptr0, out_ptr0, out_ptr1, xnumel, XBLOCK : tl.constexpr):
xnumel = 36000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 30
x3 = (xindex // 30)
x2 = (xindex // 9000)
x4 = xindex % 9000
x5 = xindex
tmp0 = tl.load(in_ptr0 + ((2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (60 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (61 + (2*x0) + (120*x3)), xmask, eviction_policy='evict_last')
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 + (x4 + (9088*x2)), tmp15, xmask)
tl.store(out_ptr1 + (x5), tmp18, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/aw/cawzukwabtkzagdqmtxehml5xtoaxgeoa4r6pkdmncwtapog2lg5.py
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# conv2d_1 => 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_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=[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_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 = 54080
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 676) % 20
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_4/inductor_cache/cj/ccj36xmuetqae6fng7fnwocfjjaiawuogu35vnhfu2gzmre6iebz.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 = (%getitem_2,), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_relu_threshold_backward_3 = async_compile.triton('triton_poi_fused_relu_threshold_backward_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=[2048],
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_relu_threshold_backward_3', '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_relu_threshold_backward_3(in_out_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 1280
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.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = 0.0
tmp4 = tmp2 <= tmp3
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
tl.store(out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/iy/ciysfxgagy5psk3wb5kiw3rln5gle6xf5w5fao4t6e2qonr525fe.py
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_3 => relu_2
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_7), kwargs = {})
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_relu_4 = async_compile.triton('triton_poi_fused_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=[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), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_relu_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_relu_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 150
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_4/inductor_cache/i3/ci3gjxynws4b6snihkmy4mi7mtg3xhsmd2rvhtmsnpenwlazlbbl.py
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_softmax => amax, exp, log, sub, sub_1, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%addmm_1, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%addmm_1, %amax), kwargs = {})
# %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 = {})
triton_per_fused__log_softmax_5 = async_compile.triton('triton_per_fused__log_softmax_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=[4, 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), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__log_softmax_5', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 1, '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__log_softmax_5(in_ptr0, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
rnumel = 10
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
tmp0 = tl.load(in_ptr0 + (r1 + (10*x0)), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, float("-inf"))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(rmask & xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tl_math.log(tmp10)
tmp12 = tmp5 - tmp11
tl.store(out_ptr2 + (r1 + (10*x0)), tmp12, rmask & 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, (10, 1, 5, 5), (25, 25, 5, 1))
assert_size_stride(primals_2, (10, ), (1, ))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_4, (20, 10, 5, 5), (250, 25, 5, 1))
assert_size_stride(primals_5, (20, ), (1, ))
assert_size_stride(primals_6, (150, 320), (320, 1))
assert_size_stride(primals_7, (150, ), (1, ))
assert_size_stride(primals_8, (10, 150), (150, 1))
assert_size_stride(primals_9, (10, ), (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=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 10, 60, 60), (36000, 3600, 60, 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, 144000, grid=grid(144000), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((4, 10, 30, 30), (9088, 900, 30, 1), torch.int8)
buf3 = empty_strided_cuda((4, 10, 30, 30), (9000, 900, 30, 1), torch.float32)
# Topologically Sorted Source Nodes: [max_pool2d, x], Original ATen: [aten.max_pool2d_with_indices, aten.relu]
triton_poi_fused_max_pool2d_with_indices_relu_1.run(buf1, buf2, buf3, 36000, grid=grid(36000), stream=stream0)
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 20, 26, 26), (13520, 676, 26, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_2.run(buf5, primals_5, 54080, grid=grid(54080), stream=stream0)
del primals_5
# Topologically Sorted Source Nodes: [adaptive_max_pool2d], Original ATen: [aten.adaptive_max_pool2d]
buf6 = torch.ops.aten.adaptive_max_pool2d.default(buf5, [4, 4])
buf7 = buf6[0]
buf8 = buf6[1]
del buf6
buf9 = buf7; del buf7 # reuse
buf16 = empty_strided_cuda((4, 20, 4, 4), (320, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_3.run(buf9, buf16, 1280, grid=grid(1280), stream=stream0)
buf10 = empty_strided_cuda((4, 150), (150, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf9, (4, 320), (320, 1), 0), reinterpret_tensor(primals_6, (320, 150), (1, 320), 0), out=buf10)
buf11 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
triton_poi_fused_relu_4.run(buf11, primals_7, 600, grid=grid(600), stream=stream0)
del primals_7
buf12 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, buf11, reinterpret_tensor(primals_8, (150, 10), (1, 150), 0), alpha=1, beta=1, out=buf12)
del primals_9
buf15 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
triton_per_fused__log_softmax_5.run(buf12, buf15, 4, 10, grid=grid(4), stream=stream0)
del buf12
return (buf15, primals_1, primals_3, primals_4, buf1, buf2, buf3, buf5, buf8, reinterpret_tensor(buf9, (4, 320), (320, 1), 0), buf11, buf15, primals_8, primals_6, buf16, )
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((10, 1, 5, 5), (25, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((10, ), (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((20, 10, 5, 5), (250, 25, 5, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((20, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((150, 320), (320, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((150, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((10, 150), (150, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((10, ), (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.functional as F
import torch.nn as nn
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 150)
self.fc2 = nn.Linear(150, 10)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.adaptive_max_pool2d(self.conv2_drop(self.conv2(x)), 4))
x = x.view(-1, 320)
x = F.relu(self.fc1(x))
x = F.dropout(x, training=self.training)
x = self.fc2(x)
return F.log_softmax(x, dim=-1)
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 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 = 144000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 3600 % 10
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_max_pool2d_with_indices_relu_1(in_ptr0, out_ptr0,
out_ptr1, xnumel, XBLOCK: tl.constexpr):
xnumel = 36000
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 30
x3 = xindex // 30
x2 = xindex // 9000
x4 = xindex % 9000
x5 = xindex
tmp0 = tl.load(in_ptr0 + (2 * x0 + 120 * x3), xmask, eviction_policy=
'evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (60 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (61 + 2 * x0 + 120 * x3), xmask,
eviction_policy='evict_last')
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 + (x4 + 9088 * x2), tmp15, xmask)
tl.store(out_ptr1 + x5, tmp18, xmask)
@triton.jit
def triton_poi_fused_convolution_2(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 54080
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 676 % 20
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_relu_threshold_backward_3(in_out_ptr0, out_ptr0,
xnumel, XBLOCK: tl.constexpr):
xnumel = 1280
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.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = 0.0
tmp4 = tmp2 <= tmp3
tl.store(in_out_ptr0 + x0, tmp2, xmask)
tl.store(out_ptr0 + x0, tmp4, xmask)
@triton.jit
def triton_poi_fused_relu_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 150
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_per_fused__log_softmax_5(in_ptr0, out_ptr2, xnumel, rnumel,
XBLOCK: tl.constexpr):
xnumel = 4
rnumel = 10
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
tmp0 = tl.load(in_ptr0 + (r1 + 10 * x0), rmask & xmask, other=0.0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(rmask & xmask, tmp1, float('-inf'))
tmp4 = triton_helpers.max2(tmp3, 1)[:, None]
tmp5 = tmp0 - tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.where(rmask & xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tl_math.log(tmp10)
tmp12 = tmp5 - tmp11
tl.store(out_ptr2 + (r1 + 10 * x0), tmp12, rmask & 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, (10, 1, 5, 5), (25, 25, 5, 1))
assert_size_stride(primals_2, (10,), (1,))
assert_size_stride(primals_3, (4, 1, 64, 64), (4096, 4096, 64, 1))
assert_size_stride(primals_4, (20, 10, 5, 5), (250, 25, 5, 1))
assert_size_stride(primals_5, (20,), (1,))
assert_size_stride(primals_6, (150, 320), (320, 1))
assert_size_stride(primals_7, (150,), (1,))
assert_size_stride(primals_8, (10, 150), (150, 1))
assert_size_stride(primals_9, (10,), (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, 10, 60, 60), (36000, 3600, 60, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(144000)](buf1, primals_2,
144000, XBLOCK=512, num_warps=8, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((4, 10, 30, 30), (9088, 900, 30, 1),
torch.int8)
buf3 = empty_strided_cuda((4, 10, 30, 30), (9000, 900, 30, 1),
torch.float32)
triton_poi_fused_max_pool2d_with_indices_relu_1[grid(36000)](buf1,
buf2, buf3, 36000, XBLOCK=256, num_warps=4, num_stages=1)
buf4 = extern_kernels.convolution(buf3, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 20, 26, 26), (13520, 676, 26, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_2[grid(54080)](buf5, primals_5, 54080,
XBLOCK=512, num_warps=4, num_stages=1)
del primals_5
buf6 = torch.ops.aten.adaptive_max_pool2d.default(buf5, [4, 4])
buf7 = buf6[0]
buf8 = buf6[1]
del buf6
buf9 = buf7
del buf7
buf16 = empty_strided_cuda((4, 20, 4, 4), (320, 16, 4, 1), torch.bool)
triton_poi_fused_relu_threshold_backward_3[grid(1280)](buf9, buf16,
1280, XBLOCK=256, num_warps=4, num_stages=1)
buf10 = empty_strided_cuda((4, 150), (150, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf9, (4, 320), (320, 1), 0),
reinterpret_tensor(primals_6, (320, 150), (1, 320), 0), out=buf10)
buf11 = buf10
del buf10
triton_poi_fused_relu_4[grid(600)](buf11, primals_7, 600, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_7
buf12 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
extern_kernels.addmm(primals_9, buf11, reinterpret_tensor(primals_8,
(150, 10), (1, 150), 0), alpha=1, beta=1, out=buf12)
del primals_9
buf15 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
triton_per_fused__log_softmax_5[grid(4)](buf12, buf15, 4, 10,
XBLOCK=1, num_warps=2, num_stages=1)
del buf12
return (buf15, primals_1, primals_3, primals_4, buf1, buf2, buf3, buf5,
buf8, reinterpret_tensor(buf9, (4, 320), (320, 1), 0), buf11, buf15,
primals_8, primals_6, buf16)
class NetNew(nn.Module):
def __init__(self):
super(NetNew, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 150)
self.fc2 = nn.Linear(150, 10)
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.fc1.weight
primals_7 = self.fc1.bias
primals_8 = self.fc2.weight
primals_9 = self.fc2.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]
| generall/Torchlite | Net | false | 6,752 | [
"MIT"
] | 1 | 2eb3e2a20b7619bd58b0b0fca120e2aefca0e79a | https://github.com/generall/Torchlite/tree/2eb3e2a20b7619bd58b0b0fca120e2aefca0e79a | import torch
import torch.nn.functional as F
import torch.nn as nn
class Model(nn.Module):
def __init__(self):
super().__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 150)
self.fc2 = nn.Linear(150, 10)
def forward(self, x):
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.adaptive_max_pool2d(self.conv2_drop(self.conv2(x)), 4))
x = x.view(-1, 320)
x = F.relu(self.fc1(x))
x = F.dropout(x, training=self.training)
x = self.fc2(x)
return F.log_softmax(x, dim=-1)
def get_inputs():
return [torch.rand([4, 1, 64, 64])]
def get_init_inputs():
return []
|
ClassificationSubNet | # 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_4/inductor_cache/au/caug6nesiygukdpkrndsclkfho3dygoeotjtbnihl4wlyyiuddug.py
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.convolution, aten.relu]
# Source node to ATen node mapping:
# x => convolution
# x_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], [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=[16384],
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 = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 16) % 256
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_4/inductor_cache/ea/ceaqvixbhpmlhiwlwyhwmhtmtxjnyelwwysomzo6d2utzobloyaf.py
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_8 => convolution_4
# Graph fragment:
# %convolution_4 : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%relu_3, %primals_10, %primals_11, [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=[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_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 = 2304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 36
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_4/inductor_cache/l5/cl5gfsdiy7m43uji57rivrvbxfytjyjnnf7nhqksnm6ozowl733t.py
# Topologically Sorted Source Nodes: [reshape], Original ATen: [aten.clone, aten._unsafe_view]
# Source node to ATen node mapping:
# reshape => clone, view
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute,), kwargs = {memory_format: torch.contiguous_format})
# %view : [num_users=1] = call_function[target=torch.ops.aten.reshape.default](args = (%clone, [4, 144, 4]), kwargs = {})
triton_poi_fused__unsafe_view_clone_2 = async_compile.triton('triton_poi_fused__unsafe_view_clone_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, 64], 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__unsafe_view_clone_2', '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__unsafe_view_clone_2(in_out_ptr0, in_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 64
xnumel = 36
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 % 16
y1 = (yindex // 16)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (16*x2) + (576*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.sigmoid(tmp0)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x2 + (36*y3)), tmp1, 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 = args
args.clear()
assert_size_stride(primals_1, (256, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (256, ), (1, ))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_5, (256, ), (1, ))
assert_size_stride(primals_6, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_7, (256, ), (1, ))
assert_size_stride(primals_8, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_9, (256, ), (1, ))
assert_size_stride(primals_10, (36, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_11, (36, ), (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, 256, 4, 4), (4096, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.convolution, aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_0.run(buf1, primals_2, 16384, grid=grid(16384), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [x_2], 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, 256, 4, 4), (4096, 16, 4, 1))
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [x_2, x_3], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_0.run(buf3, primals_5, 16384, grid=grid(16384), stream=stream0)
del primals_5
# Topologically Sorted Source Nodes: [x_4], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 256, 4, 4), (4096, 16, 4, 1))
buf5 = buf4; del buf4 # reuse
# Topologically Sorted Source Nodes: [x_4, x_5], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_0.run(buf5, primals_7, 16384, grid=grid(16384), stream=stream0)
del primals_7
# Topologically Sorted Source Nodes: [x_6], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 256, 4, 4), (4096, 16, 4, 1))
buf7 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [x_6, x_7], Original ATen: [aten.convolution, aten.relu]
triton_poi_fused_convolution_relu_0.run(buf7, primals_9, 16384, grid=grid(16384), stream=stream0)
del primals_9
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.convolution]
buf8 = extern_kernels.convolution(buf7, primals_10, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 36, 4, 4), (576, 16, 4, 1))
buf9 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [x_8], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf9, primals_11, 2304, grid=grid(2304), stream=stream0)
del primals_11
buf10 = empty_strided_cuda((4, 4, 4, 36), (576, 144, 36, 1), torch.float32)
buf11 = reinterpret_tensor(buf10, (4, 144, 4), (576, 4, 1), 0); del buf10 # reuse
# Topologically Sorted Source Nodes: [reshape], Original ATen: [aten.clone, aten._unsafe_view]
triton_poi_fused__unsafe_view_clone_2.run(buf11, buf9, 64, 36, grid=grid(64, 36), stream=stream0)
return (buf11, primals_1, primals_3, primals_4, primals_6, primals_8, primals_10, buf1, buf3, buf5, buf7, 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((256, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((256, ), (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((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((256, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((256, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((36, 256, 3, 3), (2304, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((36, ), (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 numpy as np
import torch.nn as nn
class ClassificationSubNet(nn.Module):
def __init__(self, in_channels, num_classes, num_anchors=9):
super().__init__()
self.num_classes = num_classes
self.conv2d_1 = nn.Conv2d(in_channels, 256, 3, padding=1)
nn.init.normal_(self.conv2d_1.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_1.bias.data)
self.conv2d_2 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_2.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_2.bias.data)
self.conv2d_3 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_3.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_3.bias.data)
self.conv2d_4 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_4.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_4.bias.data)
self.conv2d_5 = nn.Conv2d(256, num_classes * num_anchors, 3, padding=1)
nn.init.normal_(self.conv2d_5.weight.data, std=0.01)
nn.init.constant_(self.conv2d_5.bias.data, val=np.log(1 / 99))
def forward(self, x):
x = self.conv2d_1(x)
x = nn.functional.relu(x)
x = self.conv2d_2(x)
x = nn.functional.relu(x)
x = self.conv2d_3(x)
x = nn.functional.relu(x)
x = self.conv2d_4(x)
x = nn.functional.relu(x)
x = self.conv2d_5(x)
x = torch.sigmoid(x)
x = x.permute(0, 2, 3, 1)
return x.reshape(x.size(0), -1, self.num_classes)
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
from torch._inductor.runtime import triton_helpers
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_convolution_relu_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 // 16 % 256
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_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 2304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 36
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__unsafe_view_clone_2(in_out_ptr0, in_ptr0, ynumel,
xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 64
xnumel = 36
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 % 16
y1 = yindex // 16
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 16 * x2 + 576 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.sigmoid(tmp0)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x2 + 36 * y3), tmp1, 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) = args
args.clear()
assert_size_stride(primals_1, (256, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_2, (256,), (1,))
assert_size_stride(primals_3, (4, 4, 4, 4), (64, 16, 4, 1))
assert_size_stride(primals_4, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_5, (256,), (1,))
assert_size_stride(primals_6, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_7, (256,), (1,))
assert_size_stride(primals_8, (256, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_9, (256,), (1,))
assert_size_stride(primals_10, (36, 256, 3, 3), (2304, 9, 3, 1))
assert_size_stride(primals_11, (36,), (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, 256, 4, 4), (4096, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_relu_0[grid(16384)](buf1, primals_2,
16384, 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, 256, 4, 4), (4096, 16, 4, 1))
buf3 = buf2
del buf2
triton_poi_fused_convolution_relu_0[grid(16384)](buf3, primals_5,
16384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf4 = extern_kernels.convolution(buf3, primals_6, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 256, 4, 4), (4096, 16, 4, 1))
buf5 = buf4
del buf4
triton_poi_fused_convolution_relu_0[grid(16384)](buf5, primals_7,
16384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_7
buf6 = extern_kernels.convolution(buf5, primals_8, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 256, 4, 4), (4096, 16, 4, 1))
buf7 = buf6
del buf6
triton_poi_fused_convolution_relu_0[grid(16384)](buf7, primals_9,
16384, XBLOCK=128, num_warps=4, num_stages=1)
del primals_9
buf8 = extern_kernels.convolution(buf7, primals_10, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf8, (4, 36, 4, 4), (576, 16, 4, 1))
buf9 = buf8
del buf8
triton_poi_fused_convolution_1[grid(2304)](buf9, primals_11, 2304,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_11
buf10 = empty_strided_cuda((4, 4, 4, 36), (576, 144, 36, 1), torch.
float32)
buf11 = reinterpret_tensor(buf10, (4, 144, 4), (576, 4, 1), 0)
del buf10
triton_poi_fused__unsafe_view_clone_2[grid(64, 36)](buf11, buf9, 64,
36, XBLOCK=32, YBLOCK=32, num_warps=4, num_stages=1)
return (buf11, primals_1, primals_3, primals_4, primals_6, primals_8,
primals_10, buf1, buf3, buf5, buf7, buf9)
class ClassificationSubNetNew(nn.Module):
def __init__(self, in_channels, num_classes, num_anchors=9):
super().__init__()
self.num_classes = num_classes
self.conv2d_1 = nn.Conv2d(in_channels, 256, 3, padding=1)
nn.init.normal_(self.conv2d_1.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_1.bias.data)
self.conv2d_2 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_2.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_2.bias.data)
self.conv2d_3 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_3.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_3.bias.data)
self.conv2d_4 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_4.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_4.bias.data)
self.conv2d_5 = nn.Conv2d(256, num_classes * num_anchors, 3, padding=1)
nn.init.normal_(self.conv2d_5.weight.data, std=0.01)
nn.init.constant_(self.conv2d_5.bias.data, val=np.log(1 / 99))
def forward(self, input_0):
primals_1 = self.conv2d_1.weight
primals_2 = self.conv2d_1.bias
primals_4 = self.conv2d_2.weight
primals_5 = self.conv2d_2.bias
primals_6 = self.conv2d_3.weight
primals_7 = self.conv2d_3.bias
primals_8 = self.conv2d_4.weight
primals_9 = self.conv2d_4.bias
primals_10 = self.conv2d_5.weight
primals_11 = self.conv2d_5.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])
return output[0]
| geez0219/ARC | ClassificationSubNet | false | 6,753 | [
"Apache-2.0"
] | 1 | f2176f0d442d4a2d6028f0770b1efc1a9ae982b8 | https://github.com/geez0219/ARC/tree/f2176f0d442d4a2d6028f0770b1efc1a9ae982b8 | import torch
import numpy as np
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_channels, num_classes, num_anchors=9):
super().__init__()
self.num_classes = num_classes
self.conv2d_1 = nn.Conv2d(in_channels, 256, 3, padding=1)
nn.init.normal_(self.conv2d_1.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_1.bias.data)
self.conv2d_2 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_2.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_2.bias.data)
self.conv2d_3 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_3.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_3.bias.data)
self.conv2d_4 = nn.Conv2d(256, 256, 3, padding=1)
nn.init.normal_(self.conv2d_4.weight.data, std=0.01)
nn.init.zeros_(self.conv2d_4.bias.data)
self.conv2d_5 = nn.Conv2d(256, num_classes * num_anchors, 3, padding=1)
nn.init.normal_(self.conv2d_5.weight.data, std=0.01)
nn.init.constant_(self.conv2d_5.bias.data, val=np.log(1 / 99))
def forward(self, x):
x = self.conv2d_1(x)
x = nn.functional.relu(x)
x = self.conv2d_2(x)
x = nn.functional.relu(x)
x = self.conv2d_3(x)
x = nn.functional.relu(x)
x = self.conv2d_4(x)
x = nn.functional.relu(x)
x = self.conv2d_5(x)
x = torch.sigmoid(x)
x = x.permute(0, 2, 3, 1)
return x.reshape(x.size(0), -1, self.num_classes)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ZeroConv1d | # 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_4/inductor_cache/lt/cltvmqefu5tohqzespi5glwmktjvffxn3pkosfvulgdrvu3j2atr.py
# Topologically Sorted Source Nodes: [out, mul, exp, out_1], Original ATen: [aten.convolution, aten.mul, aten.exp]
# Source node to ATen node mapping:
# exp => exp
# mul => mul
# out => convolution
# out_1 => mul_1
# Graph fragment:
# %convolution : [num_users=2] = call_function[target=torch.ops.aten.convolution.default](args = (%unsqueeze, %primals_1, %primals_2, [1], [0], [1], False, [0], 1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_4, 3), 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 = (%squeeze, %exp), kwargs = {})
triton_poi_fused_convolution_exp_mul_0 = async_compile.triton('triton_poi_fused_convolution_exp_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: '*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_convolution_exp_mul_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_convolution_exp_mul_0(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
x1 = (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 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = 3.0
tmp5 = tmp3 * tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp2 * tmp6
tl.store(in_out_ptr0 + (x2), tmp2, xmask)
tl.store(out_ptr0 + (x2), tmp7, 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, 1), (4, 1, 1))
assert_size_stride(primals_2, (4, ), (1, ))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (1, 4, 1), (4, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf0 = extern_kernels.convolution(reinterpret_tensor(primals_3, (1, 4, 4), (16, 4, 1), 0), primals_1, stride=(1,), padding=(0,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf0, (1, 4, 4), (16, 4, 1))
buf1 = buf0; del buf0 # reuse
buf2 = empty_strided_cuda((1, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out, mul, exp, out_1], Original ATen: [aten.convolution, aten.mul, aten.exp]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_exp_mul_0.run(buf1, primals_2, primals_4, buf2, 16, grid=grid(16), stream=stream0)
del primals_2
return (buf2, primals_1, primals_4, reinterpret_tensor(primals_3, (1, 4, 4), (16, 4, 1), 0), 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, 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, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, 4, 1), (4, 1, 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 ZeroConv1d(nn.Module):
def __init__(self, in_channel, out_channel):
super().__init__()
self.conv = nn.Conv1d(in_channel, out_channel, 1, padding=0)
self.conv.weight.data.zero_()
self.conv.bias.data.zero_()
self.scale = nn.Parameter(torch.zeros(1, out_channel, 1))
def forward(self, x):
out = self.conv(x)
out = out * torch.exp(self.scale * 3)
return out
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [[], {'in_channel': 4, 'out_channel': 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_exp_mul_0(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
x1 = 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 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp4 = 3.0
tmp5 = tmp3 * tmp4
tmp6 = tl_math.exp(tmp5)
tmp7 = tmp2 * tmp6
tl.store(in_out_ptr0 + x2, tmp2, xmask)
tl.store(out_ptr0 + x2, tmp7, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = 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, 1))
assert_size_stride(primals_4, (1, 4, 1), (4, 1, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = extern_kernels.convolution(reinterpret_tensor(primals_3, (1,
4, 4), (16, 4, 1), 0), primals_1, stride=(1,), padding=(0,),
dilation=(1,), transposed=False, output_padding=(0,), groups=1,
bias=None)
assert_size_stride(buf0, (1, 4, 4), (16, 4, 1))
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((1, 4, 4), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_convolution_exp_mul_0[grid(16)](buf1, primals_2,
primals_4, buf2, 16, XBLOCK=16, num_warps=1, num_stages=1)
del primals_2
return buf2, primals_1, primals_4, reinterpret_tensor(primals_3, (1, 4,
4), (16, 4, 1), 0), buf1
class ZeroConv1dNew(nn.Module):
def __init__(self, in_channel, out_channel):
super().__init__()
self.conv = nn.Conv1d(in_channel, out_channel, 1, padding=0)
self.conv.weight.data.zero_()
self.conv.bias.data.zero_()
self.scale = nn.Parameter(torch.zeros(1, out_channel, 1))
def forward(self, input_0):
primals_4 = self.scale
primals_1 = self.conv.weight
primals_2 = self.conv.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| gorinars/VQ-VAE-Speech | ZeroConv1d | false | 6,754 | [
"MIT"
] | 1 | 60398f03eb129195bce402a423ace8cca8995f3c | https://github.com/gorinars/VQ-VAE-Speech/tree/60398f03eb129195bce402a423ace8cca8995f3c | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_channel, out_channel):
super().__init__()
self.conv = nn.Conv1d(in_channel, out_channel, 1, padding=0)
self.conv.weight.data.zero_()
self.conv.bias.data.zero_()
self.scale = nn.Parameter(torch.zeros(1, out_channel, 1))
def forward(self, x):
out = self.conv(x)
out = out * torch.exp(self.scale * 3)
return out
def get_inputs():
return [torch.rand([4, 4])]
def get_init_inputs():
return [4, 4]
|
Conv | # 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_4/inductor_cache/6p/c6p2synorvg3uvlsfefdr424l5lgj7xyt4ykg732glc2kxjbetn6.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], 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, 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), 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 = 12
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
tmp0 = tl.load(in_ptr0 + (r1 + (12*x0)), rmask & 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(rmask & 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 + (12*x0)), tmp9, rmask & xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/6x/c6xaexatzxr4vnzybtgtblji5ltakjr2daw3uf75u4az2rkqkvj3.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], [2], [1], False, [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=[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_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 = 96
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 6) % 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, 1, 1), (1, 1, 1))
assert_size_stride(primals_2, (4, 4, 3), (12, 3, 1))
assert_size_stride(primals_3, (4, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1), (1, 4, 4), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 1, 1), (1, 1, 1), 0); del buf0 # reuse
buf2 = empty_strided_cuda((4, 4, 3), (12, 3, 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, 12, grid=grid(4), stream=stream0)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(primals_4, buf2, stride=(1,), padding=(2,), dilation=(1,), transposed=False, output_padding=(0,), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 6), (24, 6, 1))
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf4, primals_3, 96, grid=grid(96), stream=stream0)
del primals_3
return (reinterpret_tensor(buf4, (4, 4, 4), (24, 6, 1), 0), 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), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((4, 4, 3), (12, 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, 4), (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
class Conv(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size=3, dilation=1,
causal=True):
super(Conv, self).__init__()
self.causal = causal
if self.causal:
self.padding = dilation * (kernel_size - 1)
else:
self.padding = dilation * (kernel_size - 1) // 2
self.conv = nn.Conv1d(in_channels, out_channels, kernel_size,
dilation=dilation, padding=self.padding)
self.conv = nn.utils.weight_norm(self.conv)
nn.init.kaiming_normal_(self.conv.weight)
def forward(self, tensor):
out = self.conv(tensor)
if self.causal and self.padding != 0:
out = out[:, :, :-self.padding]
return out
def get_inputs():
return [torch.rand([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
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
rnumel = 12
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
tmp0 = tl.load(in_ptr0 + (r1 + 12 * x0), rmask & 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(rmask & 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 + 12 * x0), tmp9, rmask & xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 96
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 6 % 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, 1, 1), (1, 1, 1))
assert_size_stride(primals_2, (4, 4, 3), (12, 3, 1))
assert_size_stride(primals_3, (4,), (1,))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 1, 1), (1, 4, 4), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 1, 1), (1, 1, 1), 0)
del buf0
buf2 = empty_strided_cuda((4, 4, 3), (12, 3, 1), torch.float32)
get_raw_stream(0)
triton_per_fused__weight_norm_interface_0[grid(4)](buf1, primals_2,
primals_1, buf2, 4, 12, XBLOCK=1, num_warps=2, num_stages=1)
buf3 = extern_kernels.convolution(primals_4, buf2, stride=(1,),
padding=(2,), dilation=(1,), transposed=False, output_padding=(
0,), groups=1, bias=None)
assert_size_stride(buf3, (4, 4, 6), (24, 6, 1))
buf4 = buf3
del buf3
triton_poi_fused_convolution_1[grid(96)](buf4, primals_3, 96,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_3
return reinterpret_tensor(buf4, (4, 4, 4), (24, 6, 1), 0
), buf2, primals_1, primals_2, primals_4, buf1, buf2
class ConvNew(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size=3, dilation=1,
causal=True):
super(ConvNew, self).__init__()
self.causal = causal
if self.causal:
self.padding = dilation * (kernel_size - 1)
else:
self.padding = dilation * (kernel_size - 1) // 2
self.conv = nn.Conv1d(in_channels, out_channels, kernel_size,
dilation=dilation, padding=self.padding)
self.conv = nn.utils.weight_norm(self.conv)
nn.init.kaiming_normal_(self.conv.weight)
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]
| gorinars/VQ-VAE-Speech | Conv | false | 6,755 | [
"MIT"
] | 1 | 60398f03eb129195bce402a423ace8cca8995f3c | https://github.com/gorinars/VQ-VAE-Speech/tree/60398f03eb129195bce402a423ace8cca8995f3c | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size=3, dilation=1,
causal=True):
super().__init__()
self.causal = causal
if self.causal:
self.padding = dilation * (kernel_size - 1)
else:
self.padding = dilation * (kernel_size - 1) // 2
self.conv = nn.Conv1d(in_channels, out_channels, kernel_size,
dilation=dilation, padding=self.padding)
self.conv = nn.utils.weight_norm(self.conv)
nn.init.kaiming_normal_(self.conv.weight)
def forward(self, tensor):
out = self.conv(tensor)
if self.causal and self.padding != 0:
out = out[:, :, :-self.padding]
return out
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
HingeLoss | # 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_4/inductor_cache/kg/ckgjqsaicrgf6lzcklj5dopswm4cuvv23skrpdqbl2wjnu2wrlru.py
# Topologically Sorted Source Nodes: [mul, output, setitem, mean], Original ATen: [aten.mul, aten.rsub, aten.lift_fresh, aten.index_put, aten.mean]
# Source node to ATen node mapping:
# mean => mean
# mul => mul
# output => sub
# setitem => full_default, index_put
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %mul), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0.0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cpu, pin_memory: False})
# %index_put : [num_users=1] = call_function[target=torch.ops.aten.index_put_.default](args = (%sub, [%le], %full_default), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%index_put,), kwargs = {})
triton_per_fused_index_put_lift_fresh_mean_mul_rsub_0 = async_compile.triton('triton_per_fused_index_put_lift_fresh_mean_mul_rsub_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_index_put_lift_fresh_mean_mul_rsub_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_index_put_lift_fresh_mean_mul_rsub_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 = 1.0
tmp4 = tmp3 - tmp2
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tmp7 = tl.where(tmp6, tmp5, tmp4)
tmp8 = tl.broadcast_to(tmp7, [RBLOCK])
tmp10 = triton_helpers.promote_to_tensor(tl.sum(tmp8, 0))
tmp11 = 256.0
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)
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [mul, output, setitem, mean], Original ATen: [aten.mul, aten.rsub, aten.lift_fresh, aten.index_put, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_index_put_lift_fresh_mean_mul_rsub_0.run(buf2, arg0_1, arg1_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_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)
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.utils.data
from torch import nn
import torch
import torch.nn.parallel
import torch.optim
class HingeLoss(nn.Module):
def __init__(self):
super(HingeLoss, self).__init__()
self.margin = 1.0
def hinge_loss(self, input, target):
output = self.margin - input.mul(target)
output[output.le(0)] = 0
return output.mean()
def forward(self, input, target):
return self.hinge_loss(input, target)
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
import torch.utils.data
from torch import nn
import torch
import torch.nn.parallel
import torch.optim
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_index_put_lift_fresh_mean_mul_rsub_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 = 1.0
tmp4 = tmp3 - tmp2
tmp5 = 0.0
tmp6 = tmp4 <= tmp5
tmp7 = tl.where(tmp6, tmp5, tmp4)
tmp8 = tl.broadcast_to(tmp7, [RBLOCK])
tmp10 = triton_helpers.promote_to_tensor(tl.sum(tmp8, 0))
tmp11 = 256.0
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)
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
get_raw_stream(0)
triton_per_fused_index_put_lift_fresh_mean_mul_rsub_0[grid(1)](buf2,
arg0_1, arg1_1, 1, 256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf2,
class HingeLossNew(nn.Module):
def __init__(self):
super(HingeLossNew, self).__init__()
self.margin = 1.0
def hinge_loss(self, input, target):
output = self.margin - input.mul(target)
output[output.le(0)] = 0
return output.mean()
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| graphbuilder/BNN | HingeLoss | false | 6,756 | [
"MIT"
] | 1 | d99eb5c7ef19f8b0c14a135d40a489f154a3c894 | https://github.com/graphbuilder/BNN/tree/d99eb5c7ef19f8b0c14a135d40a489f154a3c894 | import torch
import torch.utils.data
from torch import nn
import torch
import torch.nn.parallel
import torch.optim
class Model(nn.Module):
def __init__(self):
super().__init__()
self.margin = 1.0
def hinge_loss(self, input, target):
output = self.margin - input.mul(target)
output[output.le(0)] = 0
return output.mean()
def forward(self, input, target):
return self.hinge_loss(input, target)
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
ContrastiveLoss | # 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_4/inductor_cache/nh/cnh6iem5ybdb3twbxsrnswdgv7527qrbvucjhftq2p4nhirlo2lo.py
# Topologically Sorted Source Nodes: [diff, pow_1, dist_sq], Original ATen: [aten.sub, aten.pow, aten.sum]
# Source node to ATen node mapping:
# diff => sub
# dist_sq => sum_1
# pow_1 => pow_1
# 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, 2), kwargs = {})
# %sum_1 : [num_users=2] = 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: '*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 = 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)
tmp4 = tl.load(in_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp5 = tl.load(in_ptr1 + (16 + x0 + (64*x1)), xmask)
tmp9 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp10 = tl.load(in_ptr1 + (32 + x0 + (64*x1)), xmask)
tmp14 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp15 = tl.load(in_ptr1 + (48 + x0 + (64*x1)), xmask)
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_4/inductor_cache/vk/cvk36xlkxbjdnwotu3k672dihslsze6ft5fd7mbghpobl4jcjwkc.py
# Topologically Sorted Source Nodes: [mul, sub_2, dist, mdist, dist_1, pow_2, mul_1, loss, sum_2, truediv, loss_1], Original ATen: [aten.mul, aten.rsub, aten.sqrt, aten.clamp, aten.pow, aten.add, aten.sum, aten.div]
# Source node to ATen node mapping:
# dist => sqrt
# dist_1 => clamp_min
# loss => add
# loss_1 => div_1
# mdist => sub_1
# mul => mul
# mul_1 => mul_1
# pow_2 => pow_2
# sub_2 => sub_2
# sum_2 => sum_2
# truediv => div
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg2_1, %sum_1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %arg2_1), kwargs = {})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%sum_1,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %sqrt), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%sub_1, 0.0), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%clamp_min, 2), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_2, %pow_2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %mul_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%add,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sum_2, 2.0), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%div, 4), kwargs = {})
triton_per_fused_add_clamp_div_mul_pow_rsub_sqrt_sum_1 = async_compile.triton('triton_per_fused_add_clamp_div_mul_pow_rsub_sqrt_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, 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_clamp_div_mul_pow_rsub_sqrt_sum_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_add_clamp_div_mul_pow_rsub_sqrt_sum_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)
r2 = rindex
r0 = rindex % 64
tmp0 = tl.load(in_ptr0 + (r2), None)
tmp1 = tl.load(in_ptr1 + (r0), None, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp3 - tmp0
tmp5 = libdevice.sqrt(tmp1)
tmp6 = tmp3 - tmp5
tmp7 = 0.0
tmp8 = triton_helpers.maximum(tmp6, tmp7)
tmp9 = tmp8 * tmp8
tmp10 = tmp4 * tmp9
tmp11 = tmp2 + tmp10
tmp12 = tl.broadcast_to(tmp11, [RBLOCK])
tmp14 = triton_helpers.promote_to_tensor(tl.sum(tmp12, 0))
tmp15 = 0.5
tmp16 = tmp14 * tmp15
tmp17 = 0.25
tmp18 = tmp16 * tmp17
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp18, 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), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [diff, pow_1, dist_sq], Original ATen: [aten.sub, aten.pow, aten.sum]
stream0 = get_raw_stream(0)
triton_poi_fused_pow_sub_sum_0.run(arg0_1, arg1_1, buf0, 64, grid=grid(64), stream=stream0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [mul, sub_2, dist, mdist, dist_1, pow_2, mul_1, loss, sum_2, truediv, loss_1], Original ATen: [aten.mul, aten.rsub, aten.sqrt, aten.clamp, aten.pow, aten.add, aten.sum, aten.div]
triton_per_fused_add_clamp_div_mul_pow_rsub_sqrt_sum_1.run(buf2, arg2_1, buf0, 1, 256, grid=grid(1), stream=stream0)
del arg2_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)
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.utils.data
class ContrastiveLoss(torch.nn.Module):
"""
Contrastive loss function.
"""
def __init__(self, margin=1.0):
super(ContrastiveLoss, self).__init__()
self.margin = margin
def forward(self, x0, x1, y):
diff = x0 - x1
dist_sq = torch.sum(torch.pow(diff, 2), 1)
dist = torch.sqrt(dist_sq)
mdist = self.margin - dist
dist = torch.clamp(mdist, min=0.0)
loss = y * dist_sq + (1 - y) * torch.pow(dist, 2)
loss = torch.sum(loss) / 2.0 / x0.size()[0]
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 [[], {}]
| 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
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_pow_sub_sum_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)
tmp4 = tl.load(in_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp5 = tl.load(in_ptr1 + (16 + x0 + 64 * x1), xmask)
tmp9 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp10 = tl.load(in_ptr1 + (32 + x0 + 64 * x1), xmask)
tmp14 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp15 = tl.load(in_ptr1 + (48 + x0 + 64 * x1), xmask)
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_per_fused_add_clamp_div_mul_pow_rsub_sqrt_sum_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)
r2 = rindex
r0 = rindex % 64
tmp0 = tl.load(in_ptr0 + r2, None)
tmp1 = tl.load(in_ptr1 + r0, None, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp3 - tmp0
tmp5 = libdevice.sqrt(tmp1)
tmp6 = tmp3 - tmp5
tmp7 = 0.0
tmp8 = triton_helpers.maximum(tmp6, tmp7)
tmp9 = tmp8 * tmp8
tmp10 = tmp4 * tmp9
tmp11 = tmp2 + tmp10
tmp12 = tl.broadcast_to(tmp11, [RBLOCK])
tmp14 = triton_helpers.promote_to_tensor(tl.sum(tmp12, 0))
tmp15 = 0.5
tmp16 = tmp14 * tmp15
tmp17 = 0.25
tmp18 = tmp16 * tmp17
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp18, 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), (16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_pow_sub_sum_0[grid(64)](arg0_1, arg1_1, buf0, 64,
XBLOCK=64, num_warps=1, num_stages=1)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
triton_per_fused_add_clamp_div_mul_pow_rsub_sqrt_sum_1[grid(1)](buf2,
arg2_1, buf0, 1, 256, num_warps=2, num_stages=1)
del arg2_1
del buf0
return buf2,
class ContrastiveLossNew(torch.nn.Module):
"""
Contrastive loss function.
"""
def __init__(self, margin=1.0):
super(ContrastiveLossNew, self).__init__()
self.margin = margin
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]
| guruprasaad123/all_dl_projects | ContrastiveLoss | false | 6,757 | [
"Apache-2.0"
] | 1 | 04c869f7f001ef94c467740260663d91a34815e0 | https://github.com/guruprasaad123/all_dl_projects/tree/04c869f7f001ef94c467740260663d91a34815e0 | import torch
import torch.utils.data
class Model(torch.nn.Module):
"""
Contrastive loss function.
"""
def __init__(self, margin=1.0):
super().__init__()
self.margin = margin
def forward(self, x0, x1, y):
diff = x0 - x1
dist_sq = torch.sum(torch.pow(diff, 2), 1)
dist = torch.sqrt(dist_sq)
mdist = self.margin - dist
dist = torch.clamp(mdist, min=0.0)
loss = y * dist_sq + (1 - y) * torch.pow(dist, 2)
loss = torch.sum(loss) / 2.0 / x0.size()[0]
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 []
|
AlignQuestionEmbedding | # 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_4/inductor_cache/sb/csbkytvktsi5xvyifabpordrodcrbhhsdje4rbbt3znhexh2yg7g.py
# Topologically Sorted Source Nodes: [ctx__1, qtn__1], Original ATen: [aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# ctx__1 => relu
# qtn__1 => relu_1
# Graph fragment:
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_1,), kwargs = {})
# %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_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=[64],
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_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], '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_relu_threshold_backward_0(in_out_ptr0, in_out_ptr1, 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')
tmp5 = tl.load(in_out_ptr1 + (x2), xmask)
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = tmp5 + tmp1
tmp7 = triton_helpers.maximum(tmp3, tmp6)
tmp8 = 0.0
tmp9 = tmp7 <= tmp8
tl.store(in_out_ptr0 + (x2), tmp4, xmask)
tl.store(in_out_ptr1 + (x2), tmp7, xmask)
tl.store(out_ptr0 + (x2), tmp9, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/pe/cpezdeafoiw3t5lcgpmwmamuvqlbqkktfy4urmf543yhhkjhvbti.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => amax, exp, sub, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_4, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_4, %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_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: '*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_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__softmax_1(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 // 4)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (4*x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + (4*(x0 // 4))), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr1 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (2 + (4*(x0 // 4))), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (3 + (4*(x0 // 4))), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr1 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp1 = 1.0
tmp2 = tmp0 == tmp1
tmp4 = float("-inf")
tmp5 = tl.where(tmp2, tmp4, tmp3)
tmp7 = tmp6 == tmp1
tmp9 = tl.where(tmp7, tmp4, tmp8)
tmp10 = triton_helpers.maximum(tmp5, tmp9)
tmp12 = tmp11 == tmp1
tmp14 = tl.where(tmp12, tmp4, tmp13)
tmp15 = triton_helpers.maximum(tmp10, tmp14)
tmp17 = tmp16 == tmp1
tmp19 = tl.where(tmp17, tmp4, tmp18)
tmp20 = triton_helpers.maximum(tmp15, tmp19)
tmp21 = tmp5 - tmp20
tmp22 = tl_math.exp(tmp21)
tmp23 = tmp9 - tmp20
tmp24 = tl_math.exp(tmp23)
tmp25 = tmp22 + tmp24
tmp26 = tmp14 - tmp20
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp25 + tmp27
tmp29 = tmp19 - tmp20
tmp30 = tl_math.exp(tmp29)
tmp31 = tmp28 + tmp30
tl.store(out_ptr0 + (x0), tmp20, xmask)
tl.store(out_ptr1 + (x0), tmp31, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/r4/cr4ajqhr6ircz7vgn26jwmc627w6pszjuztpk22dschroj4uqoba.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => amax, div, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_4, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_4, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub,), 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=[64],
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_2', '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_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*(x1 // 4))), xmask)
tmp3 = tl.load(in_ptr1 + (x2), xmask)
tmp6 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp1 = 1.0
tmp2 = tmp0 == tmp1
tmp4 = float("-inf")
tmp5 = tl.where(tmp2, tmp4, tmp3)
tmp7 = tmp5 - tmp6
tmp8 = tl_math.exp(tmp7)
tmp10 = tmp8 / 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, 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), (16, 4, 1))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_5, (4, 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: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_4, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf2)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4), (16, 4, 1), 0); del buf0 # reuse
buf3 = reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0); del buf2 # reuse
buf9 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [ctx__1, qtn__1], Original ATen: [aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0.run(buf1, buf3, primals_2, buf9, 64, grid=grid(64), stream=stream0)
del primals_2
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [align_scores], Original ATen: [aten.bmm]
extern_kernels.bmm(buf1, reinterpret_tensor(buf3, (4, 4, 4), (16, 1, 4), 0), out=buf4)
buf5 = empty_strided_cuda((16, 1), (1, 16), torch.float32)
buf6 = empty_strided_cuda((16, 1), (1, 16), torch.float32)
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(primals_5, buf4, buf5, buf6, 16, grid=grid(16), stream=stream0)
buf7 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(primals_5, buf4, buf5, buf6, buf7, 64, grid=grid(64), stream=stream0)
del buf5
del buf6
buf8 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [align_embedding], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf7, (4, 4, 4), (16, 4, 1), 0), primals_4, out=buf8)
del buf7
return (buf8, primals_4, primals_5, reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), buf1, buf4, buf3, 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), (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), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((4, 4), (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.functional as F
from torch import nn
class AlignQuestionEmbedding(nn.Module):
def __init__(self, input_dim):
super().__init__()
self.linear = nn.Linear(input_dim, input_dim)
self.relu = nn.ReLU()
def forward(self, context, question, question_mask):
ctx_ = self.linear(context)
ctx_ = self.relu(ctx_)
qtn_ = self.linear(question)
qtn_ = self.relu(qtn_)
qtn_transpose = qtn_.permute(0, 2, 1)
align_scores = torch.bmm(ctx_, qtn_transpose)
qtn_mask = question_mask.unsqueeze(1).expand(align_scores.size())
align_scores = align_scores.masked_fill(qtn_mask == 1, -float('inf'))
align_scores_flat = align_scores.view(-1, question.size(1))
alpha = F.softmax(align_scores_flat, dim=1)
alpha = alpha.view(-1, context.shape[1], question.shape[1])
align_embedding = torch.bmm(alpha, question)
return align_embedding
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'input_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._inductor.runtime.triton_helpers import 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_relu_threshold_backward_0(in_out_ptr0, in_out_ptr1,
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')
tmp5 = tl.load(in_out_ptr1 + x2, xmask)
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp6 = tmp5 + tmp1
tmp7 = triton_helpers.maximum(tmp3, tmp6)
tmp8 = 0.0
tmp9 = tmp7 <= tmp8
tl.store(in_out_ptr0 + x2, tmp4, xmask)
tl.store(in_out_ptr1 + x2, tmp7, xmask)
tl.store(out_ptr0 + x2, tmp9, xmask)
@triton.jit
def triton_poi_fused__softmax_1(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 // 4), xmask, eviction_policy='evict_last'
)
tmp3 = tl.load(in_ptr1 + 4 * x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + 4 * (x0 // 4)), xmask, eviction_policy=
'evict_last')
tmp8 = tl.load(in_ptr1 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (2 + 4 * (x0 // 4)), xmask, eviction_policy=
'evict_last')
tmp13 = tl.load(in_ptr1 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp16 = tl.load(in_ptr0 + (3 + 4 * (x0 // 4)), xmask, eviction_policy=
'evict_last')
tmp18 = tl.load(in_ptr1 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp1 = 1.0
tmp2 = tmp0 == tmp1
tmp4 = float('-inf')
tmp5 = tl.where(tmp2, tmp4, tmp3)
tmp7 = tmp6 == tmp1
tmp9 = tl.where(tmp7, tmp4, tmp8)
tmp10 = triton_helpers.maximum(tmp5, tmp9)
tmp12 = tmp11 == tmp1
tmp14 = tl.where(tmp12, tmp4, tmp13)
tmp15 = triton_helpers.maximum(tmp10, tmp14)
tmp17 = tmp16 == tmp1
tmp19 = tl.where(tmp17, tmp4, tmp18)
tmp20 = triton_helpers.maximum(tmp15, tmp19)
tmp21 = tmp5 - tmp20
tmp22 = tl_math.exp(tmp21)
tmp23 = tmp9 - tmp20
tmp24 = tl_math.exp(tmp23)
tmp25 = tmp22 + tmp24
tmp26 = tmp14 - tmp20
tmp27 = tl_math.exp(tmp26)
tmp28 = tmp25 + tmp27
tmp29 = tmp19 - tmp20
tmp30 = tl_math.exp(tmp29)
tmp31 = tmp28 + tmp30
tl.store(out_ptr0 + x0, tmp20, xmask)
tl.store(out_ptr1 + x0, tmp31, xmask)
@triton.jit
def triton_poi_fused__softmax_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
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
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * (x1 // 4)), xmask)
tmp3 = tl.load(in_ptr1 + x2, xmask)
tmp6 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp1 = 1.0
tmp2 = tmp0 == tmp1
tmp4 = float('-inf')
tmp5 = tl.where(tmp2, tmp4, tmp3)
tmp7 = tmp5 - tmp6
tmp8 = tl_math.exp(tmp7)
tmp10 = tmp8 / tmp9
tl.store(out_ptr0 + x2, tmp10, 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), (16, 4, 1))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_5, (4, 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_3, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_4, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf2)
del primals_1
buf1 = reinterpret_tensor(buf0, (4, 4, 4), (16, 4, 1), 0)
del buf0
buf3 = reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0)
del buf2
buf9 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_relu_threshold_backward_0[grid(64)](buf1, buf3,
primals_2, buf9, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_2
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(buf1, reinterpret_tensor(buf3, (4, 4, 4), (16, 1,
4), 0), out=buf4)
buf5 = empty_strided_cuda((16, 1), (1, 16), torch.float32)
buf6 = empty_strided_cuda((16, 1), (1, 16), torch.float32)
triton_poi_fused__softmax_1[grid(16)](primals_5, buf4, buf5, buf6,
16, XBLOCK=16, num_warps=1, num_stages=1)
buf7 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
triton_poi_fused__softmax_2[grid(64)](primals_5, buf4, buf5, buf6,
buf7, 64, XBLOCK=64, num_warps=1, num_stages=1)
del buf5
del buf6
buf8 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf7, (4, 4, 4), (16, 4, 1),
0), primals_4, out=buf8)
del buf7
return buf8, primals_4, primals_5, reinterpret_tensor(primals_3, (16, 4
), (4, 1), 0), buf1, buf4, buf3, buf9
class AlignQuestionEmbeddingNew(nn.Module):
def __init__(self, input_dim):
super().__init__()
self.linear = nn.Linear(input_dim, input_dim)
self.relu = nn.ReLU()
def forward(self, input_0, input_1, input_2):
primals_1 = self.linear.weight
primals_2 = self.linear.bias
primals_3 = input_0
primals_4 = input_1
primals_5 = input_2
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| gustavhartz/legal-contract-elements | AlignQuestionEmbedding | false | 6,758 | [
"MIT"
] | 1 | 7a1e1f0024f9d336c7166f51b4325acf03db86a2 | https://github.com/gustavhartz/legal-contract-elements/tree/7a1e1f0024f9d336c7166f51b4325acf03db86a2 | import torch
import torch.nn.functional as F
from torch import nn
class Model(nn.Module):
def __init__(self, input_dim):
super().__init__()
self.linear = nn.Linear(input_dim, input_dim)
self.relu = nn.ReLU()
def forward(self, context, question, question_mask):
ctx_ = self.linear(context)
ctx_ = self.relu(ctx_)
qtn_ = self.linear(question)
qtn_ = self.relu(qtn_)
qtn_transpose = qtn_.permute(0, 2, 1)
align_scores = torch.bmm(ctx_, qtn_transpose)
qtn_mask = question_mask.unsqueeze(1).expand(align_scores.size())
align_scores = align_scores.masked_fill(qtn_mask == 1, -float('inf'))
align_scores_flat = align_scores.view(-1, question.size(1))
alpha = F.softmax(align_scores_flat, dim=1)
alpha = alpha.view(-1, context.shape[1], question.shape[1])
align_embedding = torch.bmm(alpha, question)
return align_embedding
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [4]
|
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_4/inductor_cache/2h/c2hyujbqipl7k6j47bijuxarwsprhgffgwjyb3pnhunnolvethht.py
# Topologically Sorted Source Nodes: [residual, tensor, one, eq, tensor_1], Original ATen: [aten.clone, aten.sign, aten.ones_like, aten.eq, aten.where]
# Source node to ATen node mapping:
# eq => eq
# one => full_default
# residual => clone
# tensor => sign
# tensor_1 => where
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%primals_1,), kwargs = {})
# %sign : [num_users=2] = call_function[target=torch.ops.aten.sign.default](args = (%primals_1,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 1), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %eq : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%sign, 0), kwargs = {})
# %where : [num_users=3] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default, %sign), kwargs = {})
triton_poi_fused_clone_eq_ones_like_sign_where_0 = async_compile.triton('triton_poi_fused_clone_eq_ones_like_sign_where_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_eq_ones_like_sign_where_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_eq_ones_like_sign_where_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 = tl.full([1], 0, tl.int32)
tmp2 = tmp1 < tmp0
tmp3 = tmp2.to(tl.int8)
tmp4 = tmp0 < tmp1
tmp5 = tmp4.to(tl.int8)
tmp6 = tmp3 - tmp5
tmp7 = tmp6.to(tmp0.dtype)
tmp8 = 0.0
tmp9 = tmp7 == tmp8
tmp10 = 1.0
tmp11 = tl.where(tmp9, tmp10, tmp7)
tl.store(out_ptr0 + (x0), tmp0, xmask)
tl.store(out_ptr1 + (x0), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ru/crueu4jwh7dhznjh46ggqi5ktaws4vzu6wmquhabrpdknhgumwzh.py
# Topologically Sorted Source Nodes: [tensor_2, one_1, eq_1, tensor_3], Original ATen: [aten.sign, aten.ones_like, aten.eq, aten.where]
# Source node to ATen node mapping:
# eq_1 => eq_1
# one_1 => full_default_1
# tensor_2 => sign_1
# tensor_3 => where_1
# Graph fragment:
# %sign_1 : [num_users=2] = call_function[target=torch.ops.aten.sign.default](args = (%primals_3,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 3, 3], 1), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %eq_1 : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%sign_1, 0), kwargs = {})
# %where_1 : [num_users=3] = call_function[target=torch.ops.aten.where.self](args = (%eq_1, %full_default_1, %sign_1), kwargs = {})
triton_poi_fused_eq_ones_like_sign_where_1 = async_compile.triton('triton_poi_fused_eq_ones_like_sign_where_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_eq_ones_like_sign_where_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_eq_ones_like_sign_where_1(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 144
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 = tmp1 < tmp0
tmp3 = tmp2.to(tl.int8)
tmp4 = tmp0 < tmp1
tmp5 = tmp4.to(tl.int8)
tmp6 = tmp3 - tmp5
tmp7 = tmp6.to(tmp0.dtype)
tmp8 = 0.0
tmp9 = tmp7 == tmp8
tmp10 = 1.0
tmp11 = tl.where(tmp9, tmp10, tmp7)
tl.store(out_ptr0 + (x0), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/td/ctdf2ohamonauwhkpjgiz662xpkp5ahv4sfrwu2ixlvyc4uewpxq.py
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten._native_batch_norm_legit_no_training, aten.hardtanh]
# Source node to ATen node mapping:
# out_1 => add_1, mul_1, mul_2, sub_2
# out_2 => clamp_max, clamp_min
# Graph fragment:
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%convolution, %unsqueeze_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_2, %unsqueeze_3), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_1, %unsqueeze_5), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %unsqueeze_7), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%add_1, -1.0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 1.0), kwargs = {})
triton_poi_fused__native_batch_norm_legit_no_training_hardtanh_2 = async_compile.triton('triton_poi_fused__native_batch_norm_legit_no_training_hardtanh_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: '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_batch_norm_legit_no_training_hardtanh_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__native_batch_norm_legit_no_training_hardtanh_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
x1 = (xindex // 16) % 4
tmp0 = tl.load(in_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr1 + (x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x1), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr4 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = 1e-05
tmp5 = tmp3 + tmp4
tmp6 = libdevice.sqrt(tmp5)
tmp7 = tl.full([1], 1, tl.int32)
tmp8 = tmp7 / tmp6
tmp9 = 1.0
tmp10 = tmp8 * tmp9
tmp11 = tmp2 * tmp10
tmp13 = tmp11 * tmp12
tmp15 = tmp13 + tmp14
tmp16 = -1.0
tmp17 = triton_helpers.maximum(tmp15, tmp16)
tmp18 = triton_helpers.minimum(tmp17, tmp9)
tl.store(out_ptr0 + (x3), tmp18, 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, 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))
assert_size_stride(primals_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((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [residual, tensor, one, eq, tensor_1], Original ATen: [aten.clone, aten.sign, aten.ones_like, aten.eq, aten.where]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_eq_ones_like_sign_where_0.run(primals_1, buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.float32)
# Topologically Sorted Source Nodes: [tensor_2, one_1, eq_1, tensor_3], Original ATen: [aten.sign, aten.ones_like, aten.eq, aten.where]
triton_poi_fused_eq_ones_like_sign_where_1.run(primals_3, buf2, 144, grid=grid(144), stream=stream0)
del primals_3
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.convolution]
buf3 = extern_kernels.convolution(buf1, buf2, 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 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten._native_batch_norm_legit_no_training, aten.hardtanh]
triton_poi_fused__native_batch_norm_legit_no_training_hardtanh_2.run(buf3, primals_4, primals_5, primals_6, primals_7, buf4, 256, grid=grid(256), stream=stream0)
# Topologically Sorted Source Nodes: [], Original ATen: []
buf5 = torch.ops.aten.set_.source_Tensor(primals_1, buf1)
assert_size_stride(buf5, (4, 4, 4, 4), (64, 16, 4, 1))
del primals_1
# Topologically Sorted Source Nodes: [], Original ATen: []
buf10 = torch.ops.aten.set_.source_Tensor(primals_2, buf2)
assert_size_stride(buf10, (4, 4, 3, 3), (36, 9, 3, 1))
del primals_2
return (buf4, buf0, primals_4, primals_5, primals_6, primals_7, buf1, 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, 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)
primals_4 = rand_strided((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
from torch import nn
import torch
import torch.nn.parallel
import torch.optim
def Binarize(tensor, quant_mode='det'):
if quant_mode == 'det':
tensor = tensor.sign()
zero = torch.zeros_like(tensor)
one = torch.ones_like(tensor)
zero - one
tensor = torch.where(tensor == 0, one, tensor)
return tensor
else:
return tensor.add_(1).div_(2).add_(torch.rand(tensor.size()).add(-0.5)
).clamp_(0, 1).round().mul_(2).add_(-1)
def Binaryconv3x3(in_planes, out_planes, stride=1):
"""3x3 convolution with padding"""
return BinarizeConv2d(in_planes, out_planes, kernel_size=3, stride=
stride, padding=1, bias=False)
class BinarizeConv2d(nn.Conv2d):
def __init__(self, *kargs, **kwargs):
super(BinarizeConv2d, self).__init__(*kargs, **kwargs)
def forward(self, input):
input.data = Binarize(input.data)
if not hasattr(self.weight, 'org'):
self.weight.org = self.weight.data.clone()
self.weight.data = Binarize(self.weight.org)
out = nn.functional.conv2d(input, self.weight, None, self.stride,
self.padding, self.dilation, self.groups)
if self.bias is not None:
self.bias.org = self.bias.data.clone()
out += self.bias.view(1, -1, 1, 1).expand_as(out)
return out
class BasicBlock(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None,
do_bntan=True):
super(BasicBlock, self).__init__()
self.conv1 = Binaryconv3x3(inplanes, planes, stride)
self.bn1 = nn.BatchNorm2d(planes)
self.tanh1 = nn.Hardtanh(inplace=True)
self.conv2 = Binaryconv3x3(planes, planes)
self.tanh2 = nn.Hardtanh(inplace=True)
self.bn2 = nn.BatchNorm2d(planes)
self.downsample = downsample
self.do_bntan = do_bntan
self.stride = stride
def forward(self, x):
residual = x.clone()
out = self.conv1(x)
out = self.bn1(out)
out = self.tanh1(out)
out = self.conv2(out)
if self.downsample is not None:
if residual.data.max() > 1:
pdb.set_trace()
residual = self.downsample(residual)
out += residual
if self.do_bntan:
out = self.bn2(out)
out = self.tanh2(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.utils.data
from torch import nn
import torch
import torch.nn.parallel
import torch.optim
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_eq_ones_like_sign_where_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 = tl.full([1], 0, tl.int32)
tmp2 = tmp1 < tmp0
tmp3 = tmp2.to(tl.int8)
tmp4 = tmp0 < tmp1
tmp5 = tmp4.to(tl.int8)
tmp6 = tmp3 - tmp5
tmp7 = tmp6.to(tmp0.dtype)
tmp8 = 0.0
tmp9 = tmp7 == tmp8
tmp10 = 1.0
tmp11 = tl.where(tmp9, tmp10, tmp7)
tl.store(out_ptr0 + x0, tmp0, xmask)
tl.store(out_ptr1 + x0, tmp11, xmask)
@triton.jit
def triton_poi_fused_eq_ones_like_sign_where_1(in_ptr0, out_ptr0, xnumel,
XBLOCK: tl.constexpr):
xnumel = 144
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 = tmp1 < tmp0
tmp3 = tmp2.to(tl.int8)
tmp4 = tmp0 < tmp1
tmp5 = tmp4.to(tl.int8)
tmp6 = tmp3 - tmp5
tmp7 = tmp6.to(tmp0.dtype)
tmp8 = 0.0
tmp9 = tmp7 == tmp8
tmp10 = 1.0
tmp11 = tl.where(tmp9, tmp10, tmp7)
tl.store(out_ptr0 + x0, tmp11, xmask)
@triton.jit
def triton_poi_fused__native_batch_norm_legit_no_training_hardtanh_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
x1 = xindex // 16 % 4
tmp0 = tl.load(in_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x1, xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr4 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 - tmp1
tmp4 = 1e-05
tmp5 = tmp3 + tmp4
tmp6 = libdevice.sqrt(tmp5)
tmp7 = tl.full([1], 1, tl.int32)
tmp8 = tmp7 / tmp6
tmp9 = 1.0
tmp10 = tmp8 * tmp9
tmp11 = tmp2 * tmp10
tmp13 = tmp11 * tmp12
tmp15 = tmp13 + tmp14
tmp16 = -1.0
tmp17 = triton_helpers.maximum(tmp15, tmp16)
tmp18 = triton_helpers.minimum(tmp17, tmp9)
tl.store(out_ptr0 + x3, tmp18, 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, 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))
assert_size_stride(primals_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((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_eq_ones_like_sign_where_0[grid(256)](primals_1,
buf0, buf1, 256, XBLOCK=128, num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((4, 4, 3, 3), (36, 9, 3, 1), torch.float32)
triton_poi_fused_eq_ones_like_sign_where_1[grid(144)](primals_3,
buf2, 144, XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
buf3 = extern_kernels.convolution(buf1, buf2, 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 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__native_batch_norm_legit_no_training_hardtanh_2[grid
(256)](buf3, primals_4, primals_5, primals_6, primals_7, buf4,
256, XBLOCK=128, num_warps=4, num_stages=1)
buf5 = torch.ops.aten.set_.source_Tensor(primals_1, buf1)
assert_size_stride(buf5, (4, 4, 4, 4), (64, 16, 4, 1))
del primals_1
buf10 = torch.ops.aten.set_.source_Tensor(primals_2, buf2)
assert_size_stride(buf10, (4, 4, 3, 3), (36, 9, 3, 1))
del primals_2
return (buf4, buf0, primals_4, primals_5, primals_6, primals_7, buf1,
buf2, buf3)
def Binarize(tensor, quant_mode='det'):
if quant_mode == 'det':
tensor = tensor.sign()
zero = torch.zeros_like(tensor)
one = torch.ones_like(tensor)
zero - one
tensor = torch.where(tensor == 0, one, tensor)
return tensor
else:
return tensor.add_(1).div_(2).add_(torch.rand(tensor.size()).add(-0.5)
).clamp_(0, 1).round().mul_(2).add_(-1)
def Binaryconv3x3(in_planes, out_planes, stride=1):
"""3x3 convolution with padding"""
return BinarizeConv2d(in_planes, out_planes, kernel_size=3, stride=
stride, padding=1, bias=False)
class BinarizeConv2d(nn.Conv2d):
def __init__(self, *kargs, **kwargs):
super(BinarizeConv2d, self).__init__(*kargs, **kwargs)
def forward(self, input):
input.data = Binarize(input.data)
if not hasattr(self.weight, 'org'):
self.weight.org = self.weight.data.clone()
self.weight.data = Binarize(self.weight.org)
out = nn.functional.conv2d(input, self.weight, None, self.stride,
self.padding, self.dilation, self.groups)
if self.bias is not None:
self.bias.org = self.bias.data.clone()
out += self.bias.view(1, -1, 1, 1).expand_as(out)
return out
class BasicBlockNew(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None,
do_bntan=True):
super(BasicBlockNew, self).__init__()
self.conv1 = Binaryconv3x3(inplanes, planes, stride)
self.bn1 = nn.BatchNorm2d(planes)
self.tanh1 = nn.Hardtanh(inplace=True)
self.conv2 = Binaryconv3x3(planes, planes)
self.tanh2 = nn.Hardtanh(inplace=True)
self.bn2 = nn.BatchNorm2d(planes)
self.downsample = downsample
self.do_bntan = do_bntan
self.stride = stride
def forward(self, input_0):
primals_2 = self.conv1.weight
primals_4 = self.bn1.weight
primals_5 = self.bn1.bias
primals_3 = self.conv2.weight
primals_6 = self.bn2.weight
primals_7 = self.bn2.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| graphbuilder/BNN | BasicBlock | false | 6,759 | [
"MIT"
] | 1 | d99eb5c7ef19f8b0c14a135d40a489f154a3c894 | https://github.com/graphbuilder/BNN/tree/d99eb5c7ef19f8b0c14a135d40a489f154a3c894 | import torch
import torch.utils.data
from torch import nn
import torch
import torch.nn.parallel
import torch.optim
def Binarize(tensor, quant_mode='det'):
if quant_mode == 'det':
tensor = tensor.sign()
zero = torch.zeros_like(tensor)
one = torch.ones_like(tensor)
zero - one
tensor = torch.where(tensor == 0, one, tensor)
return tensor
else:
return tensor.add_(1).div_(2).add_(torch.rand(tensor.size()).add(-0.5)
).clamp_(0, 1).round().mul_(2).add_(-1)
def Binaryconv3x3(in_planes, out_planes, stride=1):
"""3x3 convolution with padding"""
return BinarizeConv2d(in_planes, out_planes, kernel_size=3, stride=
stride, padding=1, bias=False)
class BinarizeConv2d(nn.Conv2d):
def __init__(self, *kargs, **kwargs):
super().__init__(*kargs, **kwargs)
def forward(self, input):
input.data = Binarize(input.data)
if not hasattr(self.weight, 'org'):
self.weight.org = self.weight.data.clone()
self.weight.data = Binarize(self.weight.org)
out = nn.functional.conv2d(input, self.weight, None, self.stride,
self.padding, self.dilation, self.groups)
if self.bias is not None:
self.bias.org = self.bias.data.clone()
out += self.bias.view(1, -1, 1, 1).expand_as(out)
return out
class Model(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None,
do_bntan=True):
super().__init__()
self.conv1 = Binaryconv3x3(inplanes, planes, stride)
self.bn1 = nn.BatchNorm2d(planes)
self.tanh1 = nn.Hardtanh(inplace=True)
self.conv2 = Binaryconv3x3(planes, planes)
self.tanh2 = nn.Hardtanh(inplace=True)
self.bn2 = nn.BatchNorm2d(planes)
self.downsample = downsample
self.do_bntan = do_bntan
self.stride = stride
def forward(self, x):
residual = x.clone()
out = self.conv1(x)
out = self.bn1(out)
out = self.tanh1(out)
out = self.conv2(out)
if self.downsample is not None:
if residual.data.max() > 1:
pdb.set_trace()
residual = self.downsample(residual)
out += residual
if self.do_bntan:
out = self.bn2(out)
out = self.tanh2(out)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
NetVLAD | # 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_4/inductor_cache/qz/cqza6p5fjiie2hfiu5dfjqqugrnzziwuwxzlhzy2aa7khopxjbym.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, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view, %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=[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
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_4/inductor_cache/v4/cv4nyn2kde7dd2c53ddahw4vtxyldln6pqt62jrliqindkf3sj5m.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_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 = 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_4/inductor_cache/ly/clypou3lcvw5t7oqb5mw44imhtc4fpzxz7nwfo5rjbppyuzq66df.py
# Topologically Sorted Source Nodes: [sum_1, residual_3, sum_2, residual_6, sum_3, residual_9, sum_4], Original ATen: [aten.sum, aten.sub]
# Source node to ATen node mapping:
# residual_3 => sub_2
# residual_6 => sub_3
# residual_9 => sub_4
# sum_1 => sum_2
# sum_2 => sum_3
# sum_3 => sum_4
# sum_4 => sum_5
# Graph fragment:
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view_5, [-2, -1]), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%permute, %unsqueeze_4), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view_9, [-2, -1]), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%permute, %unsqueeze_7), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view_13, [-2, -1]), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%permute, %unsqueeze_10), kwargs = {})
# %sum_5 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view_17, [-2, -1]), kwargs = {})
triton_per_fused_sub_sum_2 = async_compile.triton('triton_per_fused_sub_sum_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=[16, 16],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: '*fp32', 9: '*fp32', 10: 'i32', 11: '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), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_sub_sum_2', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 9, '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_sub_sum_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (r2 + (16*x3)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (r2 + (64*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp9 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr2 + (16 + r2 + (64*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp21 = tl.load(in_ptr2 + (32 + r2 + (64*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp27 = tl.load(in_ptr2 + (48 + r2 + (64*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp7 = tl.where(xmask, tmp5, 0)
tmp8 = tl.sum(tmp7, 1)[:, None]
tmp10 = tmp0 - tmp9
tmp12 = tmp0 - tmp11
tmp14 = tmp0 - tmp13
tmp16 = tmp10 * tmp15
tmp17 = tl.broadcast_to(tmp16, [XBLOCK, RBLOCK])
tmp19 = tl.where(xmask, tmp17, 0)
tmp20 = tl.sum(tmp19, 1)[:, None]
tmp22 = tmp12 * tmp21
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.where(xmask, tmp23, 0)
tmp26 = tl.sum(tmp25, 1)[:, None]
tmp28 = tmp14 * tmp27
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp31 = tl.where(xmask, tmp29, 0)
tmp32 = tl.sum(tmp31, 1)[:, None]
tl.store(out_ptr1 + (r2 + (16*x3)), tmp10, xmask)
tl.store(out_ptr2 + (r2 + (16*x3)), tmp12, xmask)
tl.store(out_ptr3 + (r2 + (16*x3)), tmp14, xmask)
tl.store(out_ptr0 + (x3), tmp8, xmask)
tl.store(out_ptr4 + (x3), tmp20, xmask)
tl.store(out_ptr5 + (x3), tmp26, xmask)
tl.store(out_ptr6 + (x3), tmp32, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/po/cpokjqs2yflzfvbevdhfu6nibie5nu4qwfkcfush4rszsdutui66.py
# Topologically Sorted Source Nodes: [vlad, setitem, setitem_1, setitem_2, setitem_3], Original ATen: [aten.zeros, aten.copy]
# Source node to ATen node mapping:
# setitem => copy
# setitem_1 => copy_1
# setitem_2 => copy_2
# setitem_3 => copy_3
# vlad => full
# Graph fragment:
# %full : [num_users=2] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %copy : [num_users=1] = call_function[target=torch.ops.aten.copy.default](args = (%slice_7, %unsqueeze_2), kwargs = {})
# %slice_scatter_default : [num_users=2] = call_function[target=torch.ops.aten.slice_scatter.default](args = (%full, %copy, 1, 0, 1), kwargs = {})
# %copy_1 : [num_users=1] = call_function[target=torch.ops.aten.copy.default](args = (%slice_26, %unsqueeze_5), kwargs = {})
# %slice_scatter_default_1 : [num_users=2] = call_function[target=torch.ops.aten.slice_scatter.default](args = (%slice_scatter_default, %copy_1, 1, 1, 2), kwargs = {})
# %copy_2 : [num_users=1] = call_function[target=torch.ops.aten.copy.default](args = (%slice_45, %unsqueeze_8), kwargs = {})
# %slice_scatter_default_2 : [num_users=2] = call_function[target=torch.ops.aten.slice_scatter.default](args = (%slice_scatter_default_1, %copy_2, 1, 2, 3), kwargs = {})
# %copy_3 : [num_users=1] = call_function[target=torch.ops.aten.copy.default](args = (%slice_64, %unsqueeze_11), kwargs = {})
# %slice_scatter_default_3 : [num_users=3] = call_function[target=torch.ops.aten.slice_scatter.default](args = (%slice_scatter_default_2, %copy_3, 1, 3, 4), kwargs = {})
triton_poi_fused_copy_zeros_3 = async_compile.triton('triton_poi_fused_copy_zeros_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: '*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_copy_zeros_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_copy_zeros_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 64
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 = tl.full([1], 3, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.load(in_ptr0 + (x0 + (4*x2)), tmp2 & xmask, eviction_policy='evict_last', other=0.0)
tmp4 = tl.full([1], 2, tl.int64)
tmp5 = tmp0 >= tmp4
tmp6 = tmp0 < tmp1
tmp7 = tmp5 & tmp6
tmp8 = tl.load(in_ptr1 + (x0 + (4*x2)), tmp7 & xmask, eviction_policy='evict_last', other=0.0)
tmp9 = tl.full([1], 1, tl.int64)
tmp10 = tmp0 >= tmp9
tmp11 = tmp0 < tmp4
tmp12 = tmp10 & tmp11
tmp13 = tl.load(in_ptr2 + (x0 + (4*x2)), tmp12 & xmask, eviction_policy='evict_last', other=0.0)
tmp14 = tmp0 < tmp9
tmp15 = tl.load(in_ptr3 + (x0 + (4*x2)), tmp14 & xmask, eviction_policy='evict_last', other=0.0)
tmp16 = 0.0
tmp17 = tl.where(tmp14, tmp15, tmp16)
tmp18 = tl.where(tmp12, tmp13, tmp17)
tmp19 = tl.where(tmp7, tmp8, tmp18)
tmp20 = tl.where(tmp2, tmp3, tmp19)
tl.store(out_ptr0 + (x3), tmp20, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/u6/cu6itfv2czb4uip3h2swmtkd3snguyjg332f2e7khvzncm35gr25.py
# Topologically Sorted Source Nodes: [vlad_1, vlad_3, layer_norm], Original ATen: [aten.div, aten.linalg_vector_norm, aten.native_layer_norm]
# Source node to ATen node mapping:
# layer_norm => add, add_1, mul_4, mul_5, rsqrt, sub_5, var_mean
# vlad_1 => div_1
# vlad_3 => div_2, pow_3, pow_4, sum_7
# Graph fragment:
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%slice_scatter_default_3, %expand_4), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_19, 2), kwargs = {})
# %sum_7 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_3, [1], True), kwargs = {})
# %pow_4 : [num_users=2] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_7, 0.5), kwargs = {})
# %div_2 : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%view_19, %expand_5), kwargs = {})
# %var_mean : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%div_2, [1]), 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 = {})
# %sub_5 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%div_2, %getitem_1), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_5, %rsqrt), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_4, %primals_4), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_5, %primals_5), kwargs = {})
triton_per_fused_div_linalg_vector_norm_native_layer_norm_4 = async_compile.triton('triton_per_fused_div_linalg_vector_norm_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.persistent_reduction(
size_hints=[4, 16],
reduction_hint=ReductionHint.DEFAULT,
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*fp32', 5: '*fp32', 6: '*fp32', 7: '*fp32', 8: 'i32', 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, 9), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_div_linalg_vector_norm_native_layer_norm_4', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], 'no_x_dim': False, 'num_load': 7, '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_div_linalg_vector_norm_native_layer_norm_4(in_out_ptr0, in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, out_ptr2, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
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
x0 = xindex
r2 = (rindex // 4)
tmp0 = tl.load(in_ptr0 + (r3 + (16*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + ((4*r2) + (16*x0)), xmask, eviction_policy='evict_last', other=0.0)
tmp3 = tl.load(in_ptr0 + (1 + (4*r2) + (16*x0)), xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (2 + (4*r2) + (16*x0)), xmask, eviction_policy='evict_last', other=0.0)
tmp9 = tl.load(in_ptr0 + (3 + (4*r2) + (16*x0)), xmask, eviction_policy='evict_last', other=0.0)
tmp47 = tl.load(in_ptr1 + (r3), None, eviction_policy='evict_last')
tmp49 = tl.load(in_ptr2 + (r3), None, 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
tmp16 = tmp15 * tmp15
tmp17 = tl.broadcast_to(tmp16, [XBLOCK, RBLOCK])
tmp19 = tl.where(xmask, tmp17, 0)
tmp20 = tl.sum(tmp19, 1)[:, None]
tmp21 = libdevice.sqrt(tmp20)
tmp22 = triton_helpers.maximum(tmp21, tmp13)
tmp23 = tmp15 / tmp22
tmp24 = tl.broadcast_to(tmp23, [XBLOCK, RBLOCK])
tmp26 = tl.where(xmask, tmp24, 0)
tmp27 = tl.broadcast_to(tmp24, [XBLOCK, RBLOCK])
tmp29 = tl.where(xmask, tmp27, 0)
tmp30 = tl.sum(tmp29, 1)[:, None]
tmp31 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp32 = tmp31.to(tl.float32)
tmp33 = tmp30 / tmp32
tmp34 = tmp24 - tmp33
tmp35 = tmp34 * tmp34
tmp36 = tl.broadcast_to(tmp35, [XBLOCK, RBLOCK])
tmp38 = tl.where(xmask, tmp36, 0)
tmp39 = tl.sum(tmp38, 1)[:, None]
tmp40 = 16.0
tmp41 = tmp39 / tmp40
tmp42 = 1e-05
tmp43 = tmp41 + tmp42
tmp44 = libdevice.rsqrt(tmp43)
tmp45 = tmp23 - tmp33
tmp46 = tmp45 * tmp44
tmp48 = tmp46 * tmp47
tmp50 = tmp48 + tmp49
tl.store(out_ptr0 + (r3 + (16*x0)), tmp15, xmask)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp21, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x0), tmp44, xmask)
tl.store(out_ptr2 + (r3 + (16*x0)), tmp50, 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, 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, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (16, ), (1, ))
assert_size_stride(primals_5, (16, ), (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, 4, 4, 4), (64, 16, 4, 1))
buf1 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(buf0, buf1, 256, grid=grid(256), stream=stream0)
buf2 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf1, buf2, 256, grid=grid(256), stream=stream0)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf4 = reinterpret_tensor(buf1, (4, 1, 4, 16), (64, 1, 16, 1), 0); del buf1 # reuse
buf6 = empty_strided_cuda((4, 1, 4, 16), (64, 1, 16, 1), torch.float32)
buf8 = empty_strided_cuda((4, 1, 4, 16), (64, 1, 16, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf9 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sum_1, residual_3, sum_2, residual_6, sum_3, residual_9, sum_4], Original ATen: [aten.sum, aten.sub]
triton_per_fused_sub_sum_2.run(primals_1, primals_3, buf2, buf3, buf4, buf6, buf8, buf5, buf7, buf9, 16, 16, grid=grid(16), stream=stream0)
del buf2
buf10 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [vlad, setitem, setitem_1, setitem_2, setitem_3], Original ATen: [aten.zeros, aten.copy]
triton_poi_fused_copy_zeros_3.run(buf9, buf7, buf5, buf3, buf10, 64, grid=grid(64), stream=stream0)
del buf3
del buf5
del buf7
del buf9
buf11 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf12 = empty_strided_cuda((4, 1), (1, 4), torch.float32)
buf13 = reinterpret_tensor(buf12, (4, 1), (1, 1), 0); del buf12 # reuse
buf14 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
buf15 = empty_strided_cuda((4, 1), (1, 4), torch.float32)
buf17 = reinterpret_tensor(buf15, (4, 1), (1, 1), 0); del buf15 # reuse
buf18 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [vlad_1, vlad_3, layer_norm], Original ATen: [aten.div, aten.linalg_vector_norm, aten.native_layer_norm]
triton_per_fused_div_linalg_vector_norm_native_layer_norm_4.run(buf13, buf17, buf10, primals_4, primals_5, buf11, buf14, buf18, 4, 16, grid=grid(4), stream=stream0)
del buf11
del primals_5
return (buf18, primals_1, primals_2, primals_4, buf0, reinterpret_tensor(primals_3, (1, 4), (4, 1), 0), reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0), buf10, buf13, buf14, buf17, )
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, 1, 1), (4, 1, 1, 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((16, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((16, ), (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
import torch.nn.functional as F
class NetVLAD(nn.Module):
def __init__(self, dims, num_clusters, outdims=None):
super(NetVLAD, self).__init__()
self.num_clusters = num_clusters
self.dims = dims
self.centroids = nn.Parameter(torch.randn(num_clusters, dims) /
math.sqrt(self.dims))
self.conv = nn.Conv2d(dims, num_clusters, kernel_size=1, bias=False)
if outdims is not None:
self.outdims = outdims
self.reduction_layer = nn.Linear(self.num_clusters * self.dims,
self.outdims, bias=False)
else:
self.outdims = self.num_clusters * self.dims
self.norm = nn.LayerNorm(self.outdims)
self.reset_parameters()
def reset_parameters(self):
self.conv.weight = nn.Parameter(self.centroids.detach().clone().
unsqueeze(-1).unsqueeze(-1))
if hasattr(self, 'reduction_layer'):
nn.init.normal_(self.reduction_layer.weight, std=1 / math.sqrt(
self.num_clusters * self.dims))
def forward(self, x, mask=None, sample=False):
N, C, T, R = x.shape
soft_assign = self.conv(x).view(N, self.num_clusters, -1)
soft_assign = F.softmax(soft_assign, dim=1).view(N, self.
num_clusters, T, R)
x_flatten = x.view(N, C, -1)
vlad = torch.zeros([N, self.num_clusters, C], dtype=x.dtype, layout
=x.layout, device=x.device)
for cluster in range(self.num_clusters):
residual = x_flatten.unsqueeze(0).permute(1, 0, 2, 3
) - self.centroids[cluster:cluster + 1, :].expand(x_flatten
.size(-1), -1, -1).permute(1, 2, 0).unsqueeze(0)
residual = residual.view(N, C, T, R)
residual *= soft_assign[:, cluster:cluster + 1, :]
if mask is not None:
residual = residual.masked_fill((1 - mask.unsqueeze(1).
unsqueeze(-1)).bool(), 0.0)
vlad[:, cluster:cluster + 1, :] = residual.sum([-2, -1]).unsqueeze(
1)
vlad = F.normalize(vlad, p=2, dim=2)
vlad = vlad.view(x.size(0), -1)
vlad = F.normalize(vlad, p=2, dim=1)
if hasattr(self, 'reduction_layer'):
vlad = self.reduction_layer(vlad)
return self.norm(vlad)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'dims': 4, 'num_clusters': 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 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
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_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 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_per_fused_sub_sum_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0,
out_ptr1, out_ptr2, out_ptr3, out_ptr4, out_ptr5, out_ptr6, 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)
r2 = rindex
x3 = xindex
x0 = xindex % 4
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + (r2 + 16 * x3), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (r2 + 64 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp9 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr2 + (16 + r2 + 64 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp21 = tl.load(in_ptr2 + (32 + r2 + 64 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp27 = tl.load(in_ptr2 + (48 + r2 + 64 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp2 = tmp0 - tmp1
tmp4 = tmp2 * tmp3
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp7 = tl.where(xmask, tmp5, 0)
tmp8 = tl.sum(tmp7, 1)[:, None]
tmp10 = tmp0 - tmp9
tmp12 = tmp0 - tmp11
tmp14 = tmp0 - tmp13
tmp16 = tmp10 * tmp15
tmp17 = tl.broadcast_to(tmp16, [XBLOCK, RBLOCK])
tmp19 = tl.where(xmask, tmp17, 0)
tmp20 = tl.sum(tmp19, 1)[:, None]
tmp22 = tmp12 * tmp21
tmp23 = tl.broadcast_to(tmp22, [XBLOCK, RBLOCK])
tmp25 = tl.where(xmask, tmp23, 0)
tmp26 = tl.sum(tmp25, 1)[:, None]
tmp28 = tmp14 * tmp27
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp31 = tl.where(xmask, tmp29, 0)
tmp32 = tl.sum(tmp31, 1)[:, None]
tl.store(out_ptr1 + (r2 + 16 * x3), tmp10, xmask)
tl.store(out_ptr2 + (r2 + 16 * x3), tmp12, xmask)
tl.store(out_ptr3 + (r2 + 16 * x3), tmp14, xmask)
tl.store(out_ptr0 + x3, tmp8, xmask)
tl.store(out_ptr4 + x3, tmp20, xmask)
tl.store(out_ptr5 + x3, tmp26, xmask)
tl.store(out_ptr6 + x3, tmp32, xmask)
@triton.jit
def triton_poi_fused_copy_zeros_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 64
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 = tl.full([1], 3, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.load(in_ptr0 + (x0 + 4 * x2), tmp2 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp4 = tl.full([1], 2, tl.int64)
tmp5 = tmp0 >= tmp4
tmp6 = tmp0 < tmp1
tmp7 = tmp5 & tmp6
tmp8 = tl.load(in_ptr1 + (x0 + 4 * x2), tmp7 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp9 = tl.full([1], 1, tl.int64)
tmp10 = tmp0 >= tmp9
tmp11 = tmp0 < tmp4
tmp12 = tmp10 & tmp11
tmp13 = tl.load(in_ptr2 + (x0 + 4 * x2), tmp12 & xmask, eviction_policy
='evict_last', other=0.0)
tmp14 = tmp0 < tmp9
tmp15 = tl.load(in_ptr3 + (x0 + 4 * x2), tmp14 & xmask, eviction_policy
='evict_last', other=0.0)
tmp16 = 0.0
tmp17 = tl.where(tmp14, tmp15, tmp16)
tmp18 = tl.where(tmp12, tmp13, tmp17)
tmp19 = tl.where(tmp7, tmp8, tmp18)
tmp20 = tl.where(tmp2, tmp3, tmp19)
tl.store(out_ptr0 + x3, tmp20, xmask)
@triton.jit
def triton_per_fused_div_linalg_vector_norm_native_layer_norm_4(in_out_ptr0,
in_out_ptr1, in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, out_ptr2,
xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
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
x0 = xindex
r2 = rindex // 4
tmp0 = tl.load(in_ptr0 + (r3 + 16 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr0 + (4 * r2 + 16 * x0), xmask, eviction_policy=
'evict_last', other=0.0)
tmp3 = tl.load(in_ptr0 + (1 + 4 * r2 + 16 * x0), xmask, eviction_policy
='evict_last', other=0.0)
tmp6 = tl.load(in_ptr0 + (2 + 4 * r2 + 16 * x0), xmask, eviction_policy
='evict_last', other=0.0)
tmp9 = tl.load(in_ptr0 + (3 + 4 * r2 + 16 * x0), xmask, eviction_policy
='evict_last', other=0.0)
tmp47 = tl.load(in_ptr1 + r3, None, eviction_policy='evict_last')
tmp49 = tl.load(in_ptr2 + r3, None, 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
tmp16 = tmp15 * tmp15
tmp17 = tl.broadcast_to(tmp16, [XBLOCK, RBLOCK])
tmp19 = tl.where(xmask, tmp17, 0)
tmp20 = tl.sum(tmp19, 1)[:, None]
tmp21 = libdevice.sqrt(tmp20)
tmp22 = triton_helpers.maximum(tmp21, tmp13)
tmp23 = tmp15 / tmp22
tmp24 = tl.broadcast_to(tmp23, [XBLOCK, RBLOCK])
tl.where(xmask, tmp24, 0)
tmp27 = tl.broadcast_to(tmp24, [XBLOCK, RBLOCK])
tmp29 = tl.where(xmask, tmp27, 0)
tmp30 = tl.sum(tmp29, 1)[:, None]
tmp31 = tl.full([XBLOCK, 1], 16, tl.int32)
tmp32 = tmp31.to(tl.float32)
tmp33 = tmp30 / tmp32
tmp34 = tmp24 - tmp33
tmp35 = tmp34 * tmp34
tmp36 = tl.broadcast_to(tmp35, [XBLOCK, RBLOCK])
tmp38 = tl.where(xmask, tmp36, 0)
tmp39 = tl.sum(tmp38, 1)[:, None]
tmp40 = 16.0
tmp41 = tmp39 / tmp40
tmp42 = 1e-05
tmp43 = tmp41 + tmp42
tmp44 = libdevice.rsqrt(tmp43)
tmp45 = tmp23 - tmp33
tmp46 = tmp45 * tmp44
tmp48 = tmp46 * tmp47
tmp50 = tmp48 + tmp49
tl.store(out_ptr0 + (r3 + 16 * x0), tmp15, xmask)
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp21, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + x0, tmp44, xmask)
tl.store(out_ptr2 + (r3 + 16 * x0), tmp50, xmask)
tl.store(out_ptr1 + x0, tmp33, 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, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (16,), (1,))
assert_size_stride(primals_5, (16,), (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, 4, 4, 4), (64, 16, 4, 1))
buf1 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(256)](buf0, buf1, 256, XBLOCK=256,
num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((4, 4, 16), (64, 16, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf1, buf2, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf3 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf4 = reinterpret_tensor(buf1, (4, 1, 4, 16), (64, 1, 16, 1), 0)
del buf1
buf6 = empty_strided_cuda((4, 1, 4, 16), (64, 1, 16, 1), torch.float32)
buf8 = empty_strided_cuda((4, 1, 4, 16), (64, 1, 16, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
buf9 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_per_fused_sub_sum_2[grid(16)](primals_1, primals_3, buf2,
buf3, buf4, buf6, buf8, buf5, buf7, buf9, 16, 16, XBLOCK=1,
num_warps=2, num_stages=1)
del buf2
buf10 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_copy_zeros_3[grid(64)](buf9, buf7, buf5, buf3,
buf10, 64, XBLOCK=64, num_warps=1, num_stages=1)
del buf3
del buf5
del buf7
del buf9
buf11 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
buf12 = empty_strided_cuda((4, 1), (1, 4), torch.float32)
buf13 = reinterpret_tensor(buf12, (4, 1), (1, 1), 0)
del buf12
buf14 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
buf15 = empty_strided_cuda((4, 1), (1, 4), torch.float32)
buf17 = reinterpret_tensor(buf15, (4, 1), (1, 1), 0)
del buf15
buf18 = empty_strided_cuda((4, 16), (16, 1), torch.float32)
triton_per_fused_div_linalg_vector_norm_native_layer_norm_4[grid(4)](
buf13, buf17, buf10, primals_4, primals_5, buf11, buf14, buf18,
4, 16, XBLOCK=1, num_warps=2, num_stages=1)
del buf11
del primals_5
return buf18, primals_1, primals_2, primals_4, buf0, reinterpret_tensor(
primals_3, (1, 4), (4, 1), 0), reinterpret_tensor(buf4, (4, 4, 4, 4
), (64, 16, 4, 1), 0), reinterpret_tensor(buf6, (4, 4, 4, 4), (64,
16, 4, 1), 0), reinterpret_tensor(buf8, (4, 4, 4, 4), (64, 16, 4, 1), 0
), buf10, buf13, buf14, buf17
class NetVLADNew(nn.Module):
def __init__(self, dims, num_clusters, outdims=None):
super(NetVLADNew, self).__init__()
self.num_clusters = num_clusters
self.dims = dims
self.centroids = nn.Parameter(torch.randn(num_clusters, dims) /
math.sqrt(self.dims))
self.conv = nn.Conv2d(dims, num_clusters, kernel_size=1, bias=False)
if outdims is not None:
self.outdims = outdims
self.reduction_layer = nn.Linear(self.num_clusters * self.dims,
self.outdims, bias=False)
else:
self.outdims = self.num_clusters * self.dims
self.norm = nn.LayerNorm(self.outdims)
self.reset_parameters()
def reset_parameters(self):
self.conv.weight = nn.Parameter(self.centroids.detach().clone().
unsqueeze(-1).unsqueeze(-1))
if hasattr(self, 'reduction_layer'):
nn.init.normal_(self.reduction_layer.weight, std=1 / math.sqrt(
self.num_clusters * self.dims))
def forward(self, input_0):
primals_3 = self.centroids
primals_2 = self.conv.weight
primals_4 = self.norm.weight
primals_5 = self.norm.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| glee1228/segment_temporal_context_aggregation | NetVLAD | false | 6,760 | [
"Apache-2.0"
] | 1 | e5778f848f1cfd89bd1f77beb5e1b38a66a2f13d | https://github.com/glee1228/segment_temporal_context_aggregation/tree/e5778f848f1cfd89bd1f77beb5e1b38a66a2f13d | import math
import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, dims, num_clusters, outdims=None):
super().__init__()
self.num_clusters = num_clusters
self.dims = dims
self.centroids = nn.Parameter(torch.randn(num_clusters, dims) /
math.sqrt(self.dims))
self.conv = nn.Conv2d(dims, num_clusters, kernel_size=1, bias=False)
if outdims is not None:
self.outdims = outdims
self.reduction_layer = nn.Linear(self.num_clusters * self.dims,
self.outdims, bias=False)
else:
self.outdims = self.num_clusters * self.dims
self.norm = nn.LayerNorm(self.outdims)
self.reset_parameters()
def reset_parameters(self):
self.conv.weight = nn.Parameter(self.centroids.detach().clone().
unsqueeze(-1).unsqueeze(-1))
if hasattr(self, 'reduction_layer'):
nn.init.normal_(self.reduction_layer.weight, std=1 / math.sqrt(
self.num_clusters * self.dims))
def forward(self, x, mask=None, sample=False):
N, C, T, R = x.shape
soft_assign = self.conv(x).view(N, self.num_clusters, -1)
soft_assign = F.softmax(soft_assign, dim=1).view(N, self.
num_clusters, T, R)
x_flatten = x.view(N, C, -1)
vlad = torch.zeros([N, self.num_clusters, C], dtype=x.dtype, layout
=x.layout, device=x.device)
for cluster in range(self.num_clusters):
residual = x_flatten.unsqueeze(0).permute(1, 0, 2, 3
) - self.centroids[cluster:cluster + 1, :].expand(x_flatten
.size(-1), -1, -1).permute(1, 2, 0).unsqueeze(0)
residual = residual.view(N, C, T, R)
residual *= soft_assign[:, cluster:cluster + 1, :]
if mask is not None:
residual = residual.masked_fill((1 - mask.unsqueeze(1).
unsqueeze(-1)).bool(), 0.0)
vlad[:, cluster:cluster + 1, :] = residual.sum([-2, -1]).unsqueeze(
1)
vlad = F.normalize(vlad, p=2, dim=2)
vlad = vlad.view(x.size(0), -1)
vlad = F.normalize(vlad, p=2, dim=1)
if hasattr(self, 'reduction_layer'):
vlad = self.reduction_layer(vlad)
return self.norm(vlad)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Attention | # 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_4/inductor_cache/nc/cncwsucylpsg2zmlivjfxu6vbd64ztxjndlsix2ysjtby3xohgk4.py
# Topologically Sorted Source Nodes: [tanh], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# tanh => 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=[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': ['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 = 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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/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_3, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %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_4/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=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
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_4/inductor_cache/6v/c6vkdknlihstoo2scjcyini5rby7e5pq6gqo7jjynq75otbegpoc.py
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.sum]
# Source node to ATen node mapping:
# mul => mul
# x => sum_2
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div, %primals_3), 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
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_ptr0 + (16 + x0 + (64*x1)), xmask)
tmp4 = tl.load(in_ptr1 + (16 + x0 + (64*x1)), xmask)
tmp7 = tl.load(in_ptr0 + (32 + x0 + (64*x1)), xmask)
tmp8 = tl.load(in_ptr1 + (32 + x0 + (64*x1)), xmask)
tmp11 = tl.load(in_ptr0 + (48 + x0 + (64*x1)), xmask)
tmp12 = tl.load(in_ptr1 + (48 + x0 + (64*x1)), xmask)
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 + (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, 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, ))
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
# Topologically Sorted Source Nodes: [tanh], Original ATen: [aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_tanh_0.run(buf1, primals_2, 256, grid=grid(256), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], 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, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.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: [softmax], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf3
buf5 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, x], Original ATen: [aten.mul, aten.sum]
triton_poi_fused_mul_sum_3.run(buf4, primals_3, buf5, 64, grid=grid(64), stream=stream0)
return (buf5, buf4, 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((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)
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 Attention(nn.Module):
def __init__(self, in_size, hidden_size):
super(Attention, self).__init__()
self.hidden = nn.Linear(in_size, hidden_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.out = nn.Linear(hidden_size, in_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input):
self.alpha = self.softmax(self.out(torch.tanh(self.hidden(input))))
x = torch.sum(self.alpha * input, 1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_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 import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, 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
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
):
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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, 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_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
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_ptr0 + (16 + x0 + 64 * x1), xmask)
tmp4 = tl.load(in_ptr1 + (16 + x0 + 64 * x1), xmask)
tmp7 = tl.load(in_ptr0 + (32 + x0 + 64 * x1), xmask)
tmp8 = tl.load(in_ptr1 + (32 + x0 + 64 * x1), xmask)
tmp11 = tl.load(in_ptr0 + (48 + x0 + 64 * x1), xmask)
tmp12 = tl.load(in_ptr1 + (48 + x0 + 64 * x1), xmask)
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 + 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, 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,))
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
get_raw_stream(0)
triton_poi_fused_tanh_0[grid(256)](buf1, primals_2, 256, 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, 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, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[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_2[grid(256)](buf3, buf4, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf3
buf5 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_mul_sum_3[grid(64)](buf4, primals_3, buf5, 64,
XBLOCK=64, num_warps=1, num_stages=1)
return buf5, buf4, primals_3, buf1, buf4, primals_4
class AttentionNew(nn.Module):
def __init__(self, in_size, hidden_size):
super(AttentionNew, self).__init__()
self.hidden = nn.Linear(in_size, hidden_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.out = nn.Linear(hidden_size, in_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input_0):
primals_1 = self.hidden.weight
primals_2 = self.hidden.bias
primals_4 = self.out.weight
primals_5 = self.out.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| gchrupala/platalea | Attention | false | 6,761 | [
"Apache-2.0"
] | 1 | 65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | https://github.com/gchrupala/platalea/tree/65833307bb6c5ad6cbdd6b17ad8ca59cf51fcd81 | import torch
import torch.nn as nn
import torch.utils.data
class Model(nn.Module):
def __init__(self, in_size, hidden_size):
super().__init__()
self.hidden = nn.Linear(in_size, hidden_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.out = nn.Linear(hidden_size, in_size)
nn.init.orthogonal_(self.hidden.weight.data)
self.softmax = nn.Softmax(dim=1)
def forward(self, input):
self.alpha = self.softmax(self.out(torch.tanh(self.hidden(input))))
x = torch.sum(self.alpha * input, 1)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Auxiliary | # 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_4/inductor_cache/7b/c7bmc2ofqncrkjepoejfp6ttfcmjqhdlv3bcw2toix4mtle2waun.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten._adaptive_avg_pool2d]
# Source node to ATen node mapping:
# x => _adaptive_avg_pool2d
# Graph fragment:
# %_adaptive_avg_pool2d : [num_users=2] = call_function[target=torch.ops.aten._adaptive_avg_pool2d.default](args = (%primals_1, [4, 4]), kwargs = {})
triton_poi_fused__adaptive_avg_pool2d_0 = async_compile.triton('triton_poi_fused__adaptive_avg_pool2d_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, 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, 3), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__adaptive_avg_pool2d_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__adaptive_avg_pool2d_0(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
xnumel = 16
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 % 4
y1 = (yindex // 4)
tmp0 = tl.load(in_ptr0 + (x2 + (16*y3)), xmask & ymask)
tl.store(out_ptr0 + (y0 + (4*x2) + (64*y1)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/o3/co36mw4kjmgyqpgjwrayb4hvdahf5pqlace7gyvxor6pge3bguca.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# x_1 => convolution
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_adaptive_avg_pool2d, %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, 16], 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_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_convolution_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 512
xnumel = 16
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 % 128
y1 = (yindex // 128)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (128*x2) + (2048*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 + (16*y3)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/6t/c6tvczfz3yzaenwtxcn2nlcjymp7bnekhmmimgqv2po5hnx62wl4.py
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_3 => relu
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_5), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_relu_2 = async_compile.triton('triton_poi_fused_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=[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_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_relu_2(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)
x2 = xindex
x0 = xindex % 1024
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')
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, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (128, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (128, ), (1, ))
assert_size_stride(primals_4, (1024, 2048), (2048, 1))
assert_size_stride(primals_5, (1024, ), (1, ))
assert_size_stride(primals_6, (4, 1024), (1024, 1))
assert_size_stride(primals_7, (4, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 1, 16, 4), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten._adaptive_avg_pool2d]
stream0 = get_raw_stream(0)
triton_poi_fused__adaptive_avg_pool2d_0.run(primals_1, buf0, 16, 16, grid=grid(16, 16), 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=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf1, (4, 128, 4, 4), (2048, 1, 512, 128))
buf2 = empty_strided_cuda((4, 128, 4, 4), (2048, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf1, primals_3, buf2, 512, 16, grid=grid(512, 16), stream=stream0)
del buf1
del primals_3
buf3 = empty_strided_cuda((4, 1024), (1024, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf2, (4, 2048), (2048, 1), 0), reinterpret_tensor(primals_4, (2048, 1024), (1, 2048), 0), out=buf3)
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.relu]
triton_poi_fused_relu_2.run(buf4, primals_5, 4096, grid=grid(4096), stream=stream0)
del primals_5
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_5], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf4, reinterpret_tensor(primals_6, (1024, 4), (1, 1024), 0), alpha=1, beta=1, out=buf5)
del primals_7
return (buf5, primals_2, buf0, reinterpret_tensor(buf2, (4, 2048), (2048, 1), 0), buf4, 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, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((128, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((128, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1024, 2048), (2048, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((1024, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((4, 1024), (1024, 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
import torch.nn.functional as F
class Auxiliary(nn.Module):
def __init__(self, input_channels, n_classes):
super(Auxiliary, self).__init__()
self.Conv2 = nn.Conv2d(input_channels, 128, kernel_size=1)
self.FC1 = nn.Linear(2048, 1024)
self.FC2 = nn.Linear(1024, n_classes)
def forward(self, x):
x = F.adaptive_avg_pool2d(x, (4, 4))
x = self.Conv2(x)
x = x.view(x.size(0), -1)
x = F.relu(self.FC1(x), inplace=True)
x = F.dropout(x, 0.7, training=self.training)
x = self.FC2(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_channels': 4, 'n_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
reinterpret_tensor = torch._C._dynamo.guards._reinterpret_tensor
@triton.jit
def triton_poi_fused__adaptive_avg_pool2d_0(in_ptr0, out_ptr0, ynumel,
xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
xnumel = 16
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 % 4
y1 = yindex // 4
tmp0 = tl.load(in_ptr0 + (x2 + 16 * y3), xmask & ymask)
tl.store(out_ptr0 + (y0 + 4 * x2 + 64 * y1), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_convolution_1(in_ptr0, in_ptr1, out_ptr0, ynumel,
xnumel, YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 512
xnumel = 16
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 % 128
y1 = yindex // 128
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 128 * x2 + 2048 * 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 + 16 * y3), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused_relu_2(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 % 1024
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)
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, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (128, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (128,), (1,))
assert_size_stride(primals_4, (1024, 2048), (2048, 1))
assert_size_stride(primals_5, (1024,), (1,))
assert_size_stride(primals_6, (4, 1024), (1024, 1))
assert_size_stride(primals_7, (4,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 4), (64, 1, 16, 4), torch.float32)
get_raw_stream(0)
triton_poi_fused__adaptive_avg_pool2d_0[grid(16, 16)](primals_1,
buf0, 16, 16, XBLOCK=16, YBLOCK=16, 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, 128, 4, 4), (2048, 1, 512, 128))
buf2 = empty_strided_cuda((4, 128, 4, 4), (2048, 16, 4, 1), torch.
float32)
triton_poi_fused_convolution_1[grid(512, 16)](buf1, primals_3, buf2,
512, 16, XBLOCK=16, YBLOCK=16, num_warps=4, num_stages=1)
del buf1
del primals_3
buf3 = empty_strided_cuda((4, 1024), (1024, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf2, (4, 2048), (2048, 1), 0),
reinterpret_tensor(primals_4, (2048, 1024), (1, 2048), 0), out=buf3
)
buf4 = buf3
del buf3
triton_poi_fused_relu_2[grid(4096)](buf4, primals_5, 4096, XBLOCK=
256, num_warps=4, num_stages=1)
del primals_5
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_7, buf4, reinterpret_tensor(primals_6,
(1024, 4), (1, 1024), 0), alpha=1, beta=1, out=buf5)
del primals_7
return buf5, primals_2, buf0, reinterpret_tensor(buf2, (4, 2048), (2048,
1), 0), buf4, primals_6, primals_4
class AuxiliaryNew(nn.Module):
def __init__(self, input_channels, n_classes):
super(AuxiliaryNew, self).__init__()
self.Conv2 = nn.Conv2d(input_channels, 128, kernel_size=1)
self.FC1 = nn.Linear(2048, 1024)
self.FC2 = nn.Linear(1024, n_classes)
def forward(self, input_0):
primals_2 = self.Conv2.weight
primals_3 = self.Conv2.bias
primals_4 = self.FC1.weight
primals_5 = self.FC1.bias
primals_6 = self.FC2.weight
primals_7 = self.FC2.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| gogo5911/PyTorch_Network | Auxiliary | false | 6,762 | [
"MIT"
] | 1 | 396e2ebfe2c7e23143e72972e2fd55613c0098a3 | https://github.com/gogo5911/PyTorch_Network/tree/396e2ebfe2c7e23143e72972e2fd55613c0098a3 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, input_channels, n_classes):
super().__init__()
self.Conv2 = nn.Conv2d(input_channels, 128, kernel_size=1)
self.FC1 = nn.Linear(2048, 1024)
self.FC2 = nn.Linear(1024, n_classes)
def forward(self, x):
x = F.adaptive_avg_pool2d(x, (4, 4))
x = self.Conv2(x)
x = x.view(x.size(0), -1)
x = F.relu(self.FC1(x), inplace=True)
x = F.dropout(x, 0.7, training=self.training)
x = self.FC2(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
Conv2dBlock | # 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_4/inductor_cache/z3/cz3vliqlpgih6ihwoaxl6cmnicfmv2ygutcuphilcsragp3evc57.py
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.convolution, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# x => convolution
# x_1 => relu
# 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 = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_convolution_relu_threshold_backward_0 = async_compile.triton('triton_poi_fused_convolution_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=[16],
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_convolution_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_convolution_relu_threshold_backward_0(in_out_ptr0, 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
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 = 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)
# Topologically Sorted Source Nodes: [x], 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, 4, 1, 1), (4, 1, 1, 1))
buf1 = buf0; del buf0 # reuse
buf2 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.bool)
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.convolution, aten.relu, aten.threshold_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_relu_threshold_backward_0.run(buf1, primals_3, buf2, 16, grid=grid(16), stream=stream0)
del primals_3
return (buf1, primals_1, primals_2, 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, ), (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.utils.data
import torch
from torch import nn
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-05, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = nn.Parameter(torch.Tensor(num_features).uniform_())
self.beta = nn.Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class Conv2dBlock(nn.Module):
def __init__(self, input_dim, output_dim, kernel_size, stride, padding=
0, norm='none', activation='relu', pad_type='zero'):
super(Conv2dBlock, self).__init__()
self.use_bias = True
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
norm_dim = output_dim
if norm == 'batch':
self.norm = nn.BatchNorm2d(norm_dim)
elif norm == 'inst':
self.norm = nn.InstanceNorm2d(norm_dim, track_running_stats=False)
elif norm == 'ln':
self.norm = LayerNorm(norm_dim)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'prelu':
self.activation = nn.PReLU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
self.conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride,
bias=self.use_bias)
def forward(self, x):
x = self.conv(self.pad(x))
if self.norm:
x = self.norm(x)
if self.activation:
x = self.activation(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'output_dim': 4, 'kernel_size': 4,
'stride': 1}]
| 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
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
@triton.jit
def triton_poi_fused_convolution_relu_threshold_backward_0(in_out_ptr0,
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
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 = 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 = 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, 4, 1, 1), (4, 1, 1, 1))
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_convolution_relu_threshold_backward_0[grid(16)](buf1,
primals_3, buf2, 16, XBLOCK=16, num_warps=1, num_stages=1)
del primals_3
return buf1, primals_1, primals_2, buf2
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-05, affine=True):
super(LayerNorm, self).__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = nn.Parameter(torch.Tensor(num_features).uniform_())
self.beta = nn.Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class Conv2dBlockNew(nn.Module):
def __init__(self, input_dim, output_dim, kernel_size, stride, padding=
0, norm='none', activation='relu', pad_type='zero'):
super(Conv2dBlockNew, self).__init__()
self.use_bias = True
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
norm_dim = output_dim
if norm == 'batch':
self.norm = nn.BatchNorm2d(norm_dim)
elif norm == 'inst':
self.norm = nn.InstanceNorm2d(norm_dim, track_running_stats=False)
elif norm == 'ln':
self.norm = LayerNorm(norm_dim)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'prelu':
self.activation = nn.PReLU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
self.conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride,
bias=self.use_bias)
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]
| guyii54/Contrastive-I2I | Conv2dBlock | false | 6,763 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import torch
import torch.utils.data
import torch
from torch import nn
class LayerNorm(nn.Module):
def __init__(self, num_features, eps=1e-05, affine=True):
super().__init__()
self.num_features = num_features
self.affine = affine
self.eps = eps
if self.affine:
self.gamma = nn.Parameter(torch.Tensor(num_features).uniform_())
self.beta = nn.Parameter(torch.zeros(num_features))
def forward(self, x):
shape = [-1] + [1] * (x.dim() - 1)
mean = x.view(x.size(0), -1).mean(1).view(*shape)
std = x.view(x.size(0), -1).std(1).view(*shape)
x = (x - mean) / (std + self.eps)
if self.affine:
shape = [1, -1] + [1] * (x.dim() - 2)
x = x * self.gamma.view(*shape) + self.beta.view(*shape)
return x
class Model(nn.Module):
def __init__(self, input_dim, output_dim, kernel_size, stride, padding=
0, norm='none', activation='relu', pad_type='zero'):
super().__init__()
self.use_bias = True
if pad_type == 'reflect':
self.pad = nn.ReflectionPad2d(padding)
elif pad_type == 'zero':
self.pad = nn.ZeroPad2d(padding)
else:
assert 0, 'Unsupported padding type: {}'.format(pad_type)
norm_dim = output_dim
if norm == 'batch':
self.norm = nn.BatchNorm2d(norm_dim)
elif norm == 'inst':
self.norm = nn.InstanceNorm2d(norm_dim, track_running_stats=False)
elif norm == 'ln':
self.norm = LayerNorm(norm_dim)
elif norm == 'none':
self.norm = None
else:
assert 0, 'Unsupported normalization: {}'.format(norm)
if activation == 'relu':
self.activation = nn.ReLU(inplace=True)
elif activation == 'lrelu':
self.activation = nn.LeakyReLU(0.2, inplace=True)
elif activation == 'prelu':
self.activation = nn.PReLU()
elif activation == 'selu':
self.activation = nn.SELU(inplace=True)
elif activation == 'tanh':
self.activation = nn.Tanh()
elif activation == 'none':
self.activation = None
else:
assert 0, 'Unsupported activation: {}'.format(activation)
self.conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride,
bias=self.use_bias)
def forward(self, x):
x = self.conv(self.pad(x))
if self.norm:
x = self.norm(x)
if self.activation:
x = self.activation(x)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'output_dim': 4, 'kernel_size': 4,
'stride': 1}]
|
GRU | # 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_4/inductor_cache/g7/cg7yygwsbwngseams3ffhkaht5lhhnqbzaonxa56ywtkmfqsj27o.py
# Topologically Sorted Source Nodes: [rz_t, rz_t_1], Original ATen: [aten.add, aten.sigmoid]
# Source node to ATen node mapping:
# rz_t => add
# rz_t_1 => sigmoid
# Graph fragment:
# %add_tensor_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_3, %primals_2), kwargs = {})
# %add_tensor_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_2, %primals_5), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_tensor_3, %add_tensor_2), kwargs = {})
# %sigmoid : [num_users=3] = call_function[target=torch.ops.aten.sigmoid.default](args = (%add,), kwargs = {})
triton_poi_fused_add_sigmoid_0 = async_compile.triton('triton_poi_fused_add_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=[32],
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_sigmoid_0', '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_sigmoid_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, 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
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)
tmp4 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tl.store(in_out_ptr0 + (x2), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/em/cemoh4ty72bxzpnrqfq5y7hodpenyuc2tp54jl5sph77ekgjjtt5.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 = (%slice_2, %primals_6), 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=[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_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_mul_1(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
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (8*x1)), xmask)
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x2), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/pt/cptohbq652zhzjyclhg7kfvcbcwp3u25ro3hcyopkosa6ffv2lv4.py
# Topologically Sorted Source Nodes: [h_h_tm1_1, h_t_above], Original ATen: [aten.add, aten.tanh]
# Source node to ATen node mapping:
# h_h_tm1_1 => add_1
# h_t_above => tanh
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_8), kwargs = {})
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_10), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_tensor_1, %add_tensor), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add_1,), kwargs = {})
triton_poi_fused_add_tanh_2 = async_compile.triton('triton_poi_fused_add_tanh_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: '*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_tanh_2', '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_tanh_2(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, 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')
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
tmp7 = libdevice.tanh(tmp6)
tl.store(in_out_ptr0 + (x2), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ir/cirdpmtn7oh4j2wa3f4ilfab42gx7p3l7oeugkf7pnt4rrukx2f7.py
# Topologically Sorted Source Nodes: [sub, mul_1, mul_2, h_t, mul_3, sub_1, mul_4, h_t_1], Original ATen: [aten.rsub, aten.mul, aten.add]
# Source node to ATen node mapping:
# h_t => add_2
# h_t_1 => add_3
# mul_1 => mul_1
# mul_2 => mul_2
# mul_3 => mul_3
# mul_4 => mul_4
# sub => sub
# sub_1 => sub_1
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %slice_4), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %primals_6), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%slice_4, %tanh), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %mul_2), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%unsqueeze, %add_2), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %unsqueeze), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %primals_6), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_3, %mul_4), kwargs = {})
triton_poi_fused_add_mul_rsub_3 = async_compile.triton('triton_poi_fused_add_mul_rsub_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: '*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_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_add_mul_rsub_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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
x2 = (xindex // 16)
x1 = (xindex // 4) % 4
x3 = xindex % 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (4 + x0 + (8*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + (x3), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + (x3), xmask, eviction_policy='evict_last')
tmp2 = 1.0
tmp3 = tmp2 - tmp1
tmp5 = tmp3 * tmp4
tmp7 = tmp1 * tmp6
tmp8 = tmp5 + tmp7
tmp9 = tmp0 * tmp8
tmp10 = tmp2 - tmp0
tmp11 = tmp10 * tmp4
tmp12 = tmp9 + tmp11
tl.store(out_ptr0 + (x4), tmp12, 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, (8, 4), (4, 1))
assert_size_stride(primals_2, (8, ), (1, ))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (8, 4), (4, 1))
assert_size_stride(primals_5, (8, ), (1, ))
assert_size_stride(primals_6, (4, 4), (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, 1))
assert_size_stride(primals_10, (4, ), (1, ))
assert_size_stride(primals_11, (4, 4), (4, 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: [], Original ATen: []
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 8), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_6, reinterpret_tensor(primals_4, (4, 8), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_3, reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), out=buf2)
del primals_7
buf3 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [rz_t, rz_t_1], Original ATen: [aten.add, aten.sigmoid]
stream0 = get_raw_stream(0)
triton_poi_fused_add_sigmoid_0.run(buf3, primals_2, buf1, primals_5, 32, grid=grid(32), stream=stream0)
del buf1
del primals_2
del primals_5
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
triton_poi_fused_mul_1.run(buf3, primals_6, buf4, 16, grid=grid(16), stream=stream0)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf4, reinterpret_tensor(primals_9, (4, 4), (1, 4), 0), out=buf5)
buf6 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [h_h_tm1_1, h_t_above], Original ATen: [aten.add, aten.tanh]
triton_poi_fused_add_tanh_2.run(buf6, primals_8, buf5, primals_10, 16, grid=grid(16), stream=stream0)
del buf5
del primals_10
del primals_8
buf7 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [sub, mul_1, mul_2, h_t, mul_3, sub_1, mul_4, h_t_1], Original ATen: [aten.rsub, aten.mul, aten.add]
triton_poi_fused_add_mul_rsub_3.run(primals_11, buf3, primals_6, buf6, buf7, 64, grid=grid(64), stream=stream0)
return (buf7, primals_3, primals_6, primals_11, buf3, buf4, buf6, primals_9, )
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((8, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((8, ), (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((8, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((8, ), (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, 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, 1), device='cuda:0', dtype=torch.float32)
primals_10 = rand_strided((4, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_11 = rand_strided((4, 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 as tc
import torch.nn as nn
class Layer_Norm(nn.Module):
def __init__(self, d_hid, eps=0.001):
super(Layer_Norm, self).__init__()
self.eps = eps
self.g = nn.Parameter(tc.ones(d_hid), requires_grad=True)
self.b = nn.Parameter(tc.zeros(d_hid), requires_grad=True)
def forward(self, z):
if z.size(-1) == 1:
return z
mu = tc.mean(z, dim=-1, keepdim=True)
variance = tc.mean(tc.pow(z - mu, 2), dim=-1, keepdim=True)
sigma = tc.sqrt(variance)
z_norm = tc.div(z - mu, sigma + self.eps)
if z_norm.is_cuda:
z_norm = z_norm * self.g.expand_as(z_norm) + self.b.expand_as(
z_norm)
else:
z_norm = z_norm * self.g.expand_as(z_norm) + self.b.expand_as(
z_norm)
return z_norm
class GRU(nn.Module):
"""
Gated Recurrent Unit network with initial state as parameter:
z_t = sigmoid((x_t dot W_xz + b_xz) + (h_{t-1} dot U_hz + b_hz))
r_t = sigmoid((x_t dot W_xr + b_xr) + (h_{t-1} dot U_hr + b_xr))
=> zr_t = sigmoid((x_t dot W_xzr + b_xzr) + (h_{t-1} dot U_hzr + b_hzr))
slice ...
h_above = tanh(x_t dot W_xh + b_xh + (h_{t-1} dot U_hh + b_hh) * r_t)
h_t = (1 - z_t) * h_above + z_t * h_{t-1}
#h_t = (1 - z_t) * h_{t-1} + z_t * h_above
all parameters are initialized in [-0.01, 0.01]
"""
def __init__(self, input_size, hidden_size, enc_hid_size=None, with_ln=
False, prefix='GRU', **kwargs):
super(GRU, self).__init__()
self.enc_hid_size = enc_hid_size
self.hidden_size = hidden_size
self.with_ln = with_ln
self.prefix = prefix
self.xh = nn.Linear(input_size, hidden_size)
self.hh = nn.Linear(hidden_size, hidden_size)
self.xrz = nn.Linear(input_size, 2 * hidden_size)
self.hrz = nn.Linear(hidden_size, 2 * hidden_size)
if self.enc_hid_size is not None:
self.crz = nn.Linear(enc_hid_size, 2 * hidden_size)
self.ch = nn.Linear(enc_hid_size, hidden_size)
if self.with_ln is True:
self.ln0 = Layer_Norm(2 * hidden_size)
self.ln1 = Layer_Norm(2 * hidden_size)
self.ln2 = Layer_Norm(hidden_size)
self.ln3 = Layer_Norm(hidden_size)
if self.enc_hid_size is not None:
self.ln4 = Layer_Norm(2 * hidden_size)
self.ln5 = Layer_Norm(hidden_size)
self.sigmoid = nn.Sigmoid()
self.tanh = nn.Tanh()
"""
x_t: input at time t
x_m: mask of x_t
h_tm1: previous state
"""
def forward(self, x_t, x_m, h_tm1, attend=None):
x_rz_t, h_rz_tm1, x_h_t = self.xrz(x_t), self.hrz(h_tm1), self.xh(x_t)
if self.with_ln is not True:
if self.enc_hid_size is None:
rz_t = x_rz_t + h_rz_tm1
else:
a_rz_t, a_h_t = self.crz(attend), self.ch(attend)
rz_t = x_rz_t + h_rz_tm1 + a_rz_t
else:
x_rz_t, h_rz_tm1, x_h_t = self.ln0(x_rz_t), self.ln1(h_rz_tm1
), self.ln2(x_h_t)
if self.enc_hid_size is None:
rz_t = x_rz_t + h_rz_tm1
else:
a_rz_t, a_h_t = self.crz(attend), self.ch(attend)
a_rz_t, a_h_t = self.ln4(a_rz_t), self.ln5(a_h_t)
rz_t = x_rz_t + h_rz_tm1 + a_rz_t
assert rz_t.dim() == 2
rz_t = self.sigmoid(rz_t)
r_t, z_t = rz_t[:, :self.hidden_size], rz_t[:, self.hidden_size:]
h_h_tm1 = self.hh(r_t * h_tm1)
if self.with_ln:
h_h_tm1 = self.ln3(h_h_tm1)
if self.enc_hid_size is None:
h_h_tm1 = x_h_t + h_h_tm1
else:
h_h_tm1 = x_h_t + h_h_tm1 + a_h_t
h_t_above = self.tanh(h_h_tm1)
h_t = (1.0 - z_t) * h_tm1 + z_t * h_t_above
if x_m is not None:
h_t = x_m[:, None] * h_t + (1.0 - x_m[:, None]) * h_tm1
return h_t
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([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 as tc
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_sigmoid_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2,
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
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)
tmp4 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = tl.sigmoid(tmp6)
tl.store(in_out_ptr0 + x2, tmp7, xmask)
@triton.jit
def triton_poi_fused_mul_1(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
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 8 * x1), xmask)
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + x2, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_tanh_2(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2,
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')
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
tmp7 = libdevice.tanh(tmp6)
tl.store(in_out_ptr0 + x2, tmp7, xmask)
@triton.jit
def triton_poi_fused_add_mul_rsub_3(in_ptr0, in_ptr1, in_ptr2, in_ptr3,
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
x2 = xindex // 16
x1 = xindex // 4 % 4
x3 = xindex % 16
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp1 = tl.load(in_ptr1 + (4 + x0 + 8 * x1), xmask, eviction_policy=
'evict_last')
tmp4 = tl.load(in_ptr2 + x3, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr3 + x3, xmask, eviction_policy='evict_last')
tmp2 = 1.0
tmp3 = tmp2 - tmp1
tmp5 = tmp3 * tmp4
tmp7 = tmp1 * tmp6
tmp8 = tmp5 + tmp7
tmp9 = tmp0 * tmp8
tmp10 = tmp2 - tmp0
tmp11 = tmp10 * tmp4
tmp12 = tmp9 + tmp11
tl.store(out_ptr0 + x4, tmp12, 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, (8, 4), (4, 1))
assert_size_stride(primals_2, (8,), (1,))
assert_size_stride(primals_3, (4, 4), (4, 1))
assert_size_stride(primals_4, (8, 4), (4, 1))
assert_size_stride(primals_5, (8,), (1,))
assert_size_stride(primals_6, (4, 4), (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, 1))
assert_size_stride(primals_10, (4,), (1,))
assert_size_stride(primals_11, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
extern_kernels.mm(primals_3, reinterpret_tensor(primals_1, (4, 8),
(1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((4, 8), (8, 1), torch.float32)
extern_kernels.mm(primals_6, reinterpret_tensor(primals_4, (4, 8),
(1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(primals_3, reinterpret_tensor(primals_7, (4, 4),
(1, 4), 0), out=buf2)
del primals_7
buf3 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_add_sigmoid_0[grid(32)](buf3, primals_2, buf1,
primals_5, 32, XBLOCK=32, num_warps=1, num_stages=1)
del buf1
del primals_2
del primals_5
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_mul_1[grid(16)](buf3, primals_6, buf4, 16, XBLOCK=
16, num_warps=1, num_stages=1)
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.mm(buf4, reinterpret_tensor(primals_9, (4, 4), (1, 4
), 0), out=buf5)
buf6 = buf2
del buf2
triton_poi_fused_add_tanh_2[grid(16)](buf6, primals_8, buf5,
primals_10, 16, XBLOCK=16, num_warps=1, num_stages=1)
del buf5
del primals_10
del primals_8
buf7 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_mul_rsub_3[grid(64)](primals_11, buf3,
primals_6, buf6, buf7, 64, XBLOCK=64, num_warps=1, num_stages=1)
return buf7, primals_3, primals_6, primals_11, buf3, buf4, buf6, primals_9
class Layer_Norm(nn.Module):
def __init__(self, d_hid, eps=0.001):
super(Layer_Norm, self).__init__()
self.eps = eps
self.g = nn.Parameter(tc.ones(d_hid), requires_grad=True)
self.b = nn.Parameter(tc.zeros(d_hid), requires_grad=True)
def forward(self, z):
if z.size(-1) == 1:
return z
mu = tc.mean(z, dim=-1, keepdim=True)
variance = tc.mean(tc.pow(z - mu, 2), dim=-1, keepdim=True)
sigma = tc.sqrt(variance)
z_norm = tc.div(z - mu, sigma + self.eps)
if z_norm.is_cuda:
z_norm = z_norm * self.g.expand_as(z_norm) + self.b.expand_as(
z_norm)
else:
z_norm = z_norm * self.g.expand_as(z_norm) + self.b.expand_as(
z_norm)
return z_norm
class GRUNew(nn.Module):
"""
Gated Recurrent Unit network with initial state as parameter:
z_t = sigmoid((x_t dot W_xz + b_xz) + (h_{t-1} dot U_hz + b_hz))
r_t = sigmoid((x_t dot W_xr + b_xr) + (h_{t-1} dot U_hr + b_xr))
=> zr_t = sigmoid((x_t dot W_xzr + b_xzr) + (h_{t-1} dot U_hzr + b_hzr))
slice ...
h_above = tanh(x_t dot W_xh + b_xh + (h_{t-1} dot U_hh + b_hh) * r_t)
h_t = (1 - z_t) * h_above + z_t * h_{t-1}
#h_t = (1 - z_t) * h_{t-1} + z_t * h_above
all parameters are initialized in [-0.01, 0.01]
"""
def __init__(self, input_size, hidden_size, enc_hid_size=None, with_ln=
False, prefix='GRU', **kwargs):
super(GRUNew, self).__init__()
self.enc_hid_size = enc_hid_size
self.hidden_size = hidden_size
self.with_ln = with_ln
self.prefix = prefix
self.xh = nn.Linear(input_size, hidden_size)
self.hh = nn.Linear(hidden_size, hidden_size)
self.xrz = nn.Linear(input_size, 2 * hidden_size)
self.hrz = nn.Linear(hidden_size, 2 * hidden_size)
if self.enc_hid_size is not None:
self.crz = nn.Linear(enc_hid_size, 2 * hidden_size)
self.ch = nn.Linear(enc_hid_size, hidden_size)
if self.with_ln is True:
self.ln0 = Layer_Norm(2 * hidden_size)
self.ln1 = Layer_Norm(2 * hidden_size)
self.ln2 = Layer_Norm(hidden_size)
self.ln3 = Layer_Norm(hidden_size)
if self.enc_hid_size is not None:
self.ln4 = Layer_Norm(2 * hidden_size)
self.ln5 = Layer_Norm(hidden_size)
self.sigmoid = nn.Sigmoid()
self.tanh = nn.Tanh()
"""
x_t: input at time t
x_m: mask of x_t
h_tm1: previous state
"""
def forward(self, input_0, input_1, input_2):
primals_3 = self.xh.weight
primals_8 = self.xh.bias
primals_6 = self.hh.weight
primals_10 = self.hh.bias
primals_1 = self.xrz.weight
primals_2 = self.xrz.bias
primals_4 = self.hrz.weight
primals_5 = self.hrz.bias
primals_7 = input_0
primals_9 = input_1
primals_11 = 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]
| gushu333/DA4NMT | GRU | false | 6,764 | [
"Apache-2.0"
] | 1 | dba52a3d3784cd795b6f9aaf655b63475a848798 | https://github.com/gushu333/DA4NMT/tree/dba52a3d3784cd795b6f9aaf655b63475a848798 | import torch
import torch as tc
import torch.nn as nn
class Layer_Norm(nn.Module):
def __init__(self, d_hid, eps=0.001):
super().__init__()
self.eps = eps
self.g = nn.Parameter(tc.ones(d_hid), requires_grad=True)
self.b = nn.Parameter(tc.zeros(d_hid), requires_grad=True)
def forward(self, z):
if z.size(-1) == 1:
return z
mu = tc.mean(z, dim=-1, keepdim=True)
variance = tc.mean(tc.pow(z - mu, 2), dim=-1, keepdim=True)
sigma = tc.sqrt(variance)
z_norm = tc.div(z - mu, sigma + self.eps)
if z_norm.is_cuda:
z_norm = z_norm * self.g.expand_as(z_norm) + self.b.expand_as(
z_norm)
else:
z_norm = z_norm * self.g.expand_as(z_norm) + self.b.expand_as(
z_norm)
return z_norm
class Model(nn.Module):
"""
Gated Recurrent Unit network with initial state as parameter:
z_t = sigmoid((x_t dot W_xz + b_xz) + (h_{t-1} dot U_hz + b_hz))
r_t = sigmoid((x_t dot W_xr + b_xr) + (h_{t-1} dot U_hr + b_xr))
=> zr_t = sigmoid((x_t dot W_xzr + b_xzr) + (h_{t-1} dot U_hzr + b_hzr))
slice ...
h_above = tanh(x_t dot W_xh + b_xh + (h_{t-1} dot U_hh + b_hh) * r_t)
h_t = (1 - z_t) * h_above + z_t * h_{t-1}
#h_t = (1 - z_t) * h_{t-1} + z_t * h_above
all parameters are initialized in [-0.01, 0.01]
"""
def __init__(self, input_size, hidden_size, enc_hid_size=None, with_ln=
False, prefix='GRU', **kwargs):
super().__init__()
self.enc_hid_size = enc_hid_size
self.hidden_size = hidden_size
self.with_ln = with_ln
self.prefix = prefix
self.xh = nn.Linear(input_size, hidden_size)
self.hh = nn.Linear(hidden_size, hidden_size)
self.xrz = nn.Linear(input_size, 2 * hidden_size)
self.hrz = nn.Linear(hidden_size, 2 * hidden_size)
if self.enc_hid_size is not None:
self.crz = nn.Linear(enc_hid_size, 2 * hidden_size)
self.ch = nn.Linear(enc_hid_size, hidden_size)
if self.with_ln is True:
self.ln0 = Layer_Norm(2 * hidden_size)
self.ln1 = Layer_Norm(2 * hidden_size)
self.ln2 = Layer_Norm(hidden_size)
self.ln3 = Layer_Norm(hidden_size)
if self.enc_hid_size is not None:
self.ln4 = Layer_Norm(2 * hidden_size)
self.ln5 = Layer_Norm(hidden_size)
self.sigmoid = nn.Sigmoid()
self.tanh = nn.Tanh()
"""
x_t: input at time t
x_m: mask of x_t
h_tm1: previous state
"""
def forward(self, x_t, x_m, h_tm1, attend=None):
x_rz_t, h_rz_tm1, x_h_t = self.xrz(x_t), self.hrz(h_tm1), self.xh(x_t)
if self.with_ln is not True:
if self.enc_hid_size is None:
rz_t = x_rz_t + h_rz_tm1
else:
a_rz_t, a_h_t = self.crz(attend), self.ch(attend)
rz_t = x_rz_t + h_rz_tm1 + a_rz_t
else:
x_rz_t, h_rz_tm1, x_h_t = self.ln0(x_rz_t), self.ln1(h_rz_tm1
), self.ln2(x_h_t)
if self.enc_hid_size is None:
rz_t = x_rz_t + h_rz_tm1
else:
a_rz_t, a_h_t = self.crz(attend), self.ch(attend)
a_rz_t, a_h_t = self.ln4(a_rz_t), self.ln5(a_h_t)
rz_t = x_rz_t + h_rz_tm1 + a_rz_t
assert rz_t.dim() == 2
rz_t = self.sigmoid(rz_t)
r_t, z_t = rz_t[:, :self.hidden_size], rz_t[:, self.hidden_size:]
h_h_tm1 = self.hh(r_t * h_tm1)
if self.with_ln:
h_h_tm1 = self.ln3(h_h_tm1)
if self.enc_hid_size is None:
h_h_tm1 = x_h_t + h_h_tm1
else:
h_h_tm1 = x_h_t + h_h_tm1 + a_h_t
h_t_above = self.tanh(h_h_tm1)
h_t = (1.0 - z_t) * h_tm1 + z_t * h_t_above
if x_m is not None:
h_t = x_m[:, None] * h_t + (1.0 - x_m[:, N
# ... truncated (>4000 chars) for memory efficiency |
Normalize | # 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_4/inductor_cache/6o/c6ojhuuvyz43ogs3olib4lkalnymxhhnoqfyye667gq5hhhh5aqy.py
# Topologically Sorted Source Nodes: [pow_1, sum_1, norm, add, out], Original ATen: [aten.pow, aten.sum, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# norm => pow_2
# out => div
# pow_1 => pow_1
# sum_1 => sum_1
# 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 = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sum_1, 0.5), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_2, 1e-07), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%arg0_1, %add), kwargs = {})
triton_poi_fused_add_div_pow_sum_0 = async_compile.triton('triton_poi_fused_add_div_pow_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_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_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 = libdevice.sqrt(tmp11)
tmp13 = 1e-07
tmp14 = tmp12 + 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, norm, add, out], Original ATen: [aten.pow, aten.sum, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_pow_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.utils.data
import torch
from torch import nn
class Normalize(nn.Module):
def __init__(self, power=2):
super(Normalize, self).__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
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.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
@triton.jit
def triton_poi_fused_add_div_pow_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 = libdevice.sqrt(tmp11)
tmp13 = 1e-07
tmp14 = tmp12 + 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_sum_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class NormalizeNew(nn.Module):
def __init__(self, power=2):
super(NormalizeNew, self).__init__()
self.power = power
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| guyii54/Contrastive-I2I | Normalize | false | 6,765 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import torch
import torch.utils.data
import torch
from torch import nn
class Model(nn.Module):
def __init__(self, power=2):
super().__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
GroupedChannelNorm | # 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_4/inductor_cache/iw/ciwneupqau4kn2jj7timx3ubvnhllst5w6ftifgvb35ru3sw54dr.py
# Topologically Sorted Source Nodes: [mean, sub, std, add, x_norm], Original ATen: [aten.mean, aten.sub, aten.std, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# mean => mean
# std => sqrt, var
# sub => sub
# x_norm => div
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%view, [2], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view, %mean), kwargs = {})
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%view, [2]), kwargs = {correction: 1.0, keepdim: True})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%var,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sqrt, 1e-07), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %add), kwargs = {})
triton_poi_fused_add_div_mean_std_sub_0 = async_compile.triton('triton_poi_fused_add_div_mean_std_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_add_div_mean_std_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_add_div_mean_std_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
tmp11 = tmp1 - tmp9
tmp12 = tmp11 * tmp11
tmp13 = tmp2 - tmp9
tmp14 = tmp13 * tmp13
tmp15 = tmp12 + tmp14
tmp16 = tmp4 - tmp9
tmp17 = tmp16 * tmp16
tmp18 = tmp15 + tmp17
tmp19 = tmp6 - tmp9
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = 3.0
tmp23 = tmp21 / tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = 1e-07
tmp26 = tmp24 + tmp25
tmp27 = tmp10 / tmp26
tl.store(out_ptr0 + (x3), tmp27, 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, 4), (64, 1, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean, sub, std, add, x_norm], Original ATen: [aten.mean, aten.sub, aten.std, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_mean_std_sub_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), (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.utils.data
import torch
from torch import nn
class GroupedChannelNorm(nn.Module):
def __init__(self, num_groups):
super().__init__()
self.num_groups = num_groups
def forward(self, x):
shape = list(x.shape)
new_shape = [shape[0], self.num_groups, shape[1] // self.num_groups
] + shape[2:]
x = x.view(*new_shape)
mean = x.mean(dim=2, keepdim=True)
std = x.std(dim=2, keepdim=True)
x_norm = (x - mean) / (std + 1e-07)
return x_norm.view(*shape)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_groups': 1}]
| 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.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_add_div_mean_std_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
tmp11 = tmp1 - tmp9
tmp12 = tmp11 * tmp11
tmp13 = tmp2 - tmp9
tmp14 = tmp13 * tmp13
tmp15 = tmp12 + tmp14
tmp16 = tmp4 - tmp9
tmp17 = tmp16 * tmp16
tmp18 = tmp15 + tmp17
tmp19 = tmp6 - tmp9
tmp20 = tmp19 * tmp19
tmp21 = tmp18 + tmp20
tmp22 = 3.0
tmp23 = tmp21 / tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = 1e-07
tmp26 = tmp24 + tmp25
tmp27 = tmp10 / tmp26
tl.store(out_ptr0 + x3, tmp27, 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, 4), (64, 1, 16, 4, 1), torch
.float32)
get_raw_stream(0)
triton_poi_fused_add_div_mean_std_sub_0[grid(256)](arg0_1, buf0,
256, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0),
class GroupedChannelNormNew(nn.Module):
def __init__(self, num_groups):
super().__init__()
self.num_groups = num_groups
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| guyii54/Contrastive-I2I | GroupedChannelNorm | false | 6,766 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import torch
import torch.utils.data
import torch
from torch import nn
class Model(nn.Module):
def __init__(self, num_groups):
super().__init__()
self.num_groups = num_groups
def forward(self, x):
shape = list(x.shape)
new_shape = [shape[0], self.num_groups, shape[1] // self.num_groups
] + shape[2:]
x = x.view(*new_shape)
mean = x.mean(dim=2, keepdim=True)
std = x.std(dim=2, keepdim=True)
x_norm = (x - mean) / (std + 1e-07)
return x_norm.view(*shape)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [1]
|
ReshapeF | # 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_4/inductor_cache/ef/cef2l3ajjzplmrqfzg25ehcs22rwnbbsdt4r6sobviedge5satl2.py
# Topologically Sorted Source Nodes: [pow_1, sum_1, norm, add, out], Original ATen: [aten.pow, aten.sum, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# norm => pow_2
# out => div
# pow_1 => pow_1
# sum_1 => sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view, 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 = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_2, 1e-07), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%view, %add), kwargs = {})
triton_poi_fused_add_div_pow_sum_0 = async_compile.triton('triton_poi_fused_add_div_pow_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, 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_add_div_pow_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_sum_0(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
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + ((16*x1) + (64*(y0 // 16)) + (y0 % 16)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + ((64*(y0 // 16)) + (y0 % 16)), ymask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + (64*(y0 // 16)) + (y0 % 16)), ymask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (32 + (64*(y0 // 16)) + (y0 % 16)), ymask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (48 + (64*(y0 // 16)) + (y0 % 16)), 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-07
tmp14 = tmp12 + tmp13
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x1 + (4*y0)), tmp15, xmask & ymask)
''', 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((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [pow_1, sum_1, norm, add, out], Original ATen: [aten.pow, aten.sum, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_pow_sum_0.run(arg0_1, buf0, 64, 4, grid=grid(64, 4), 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.utils.data
import torch
from torch import nn
class Normalize(nn.Module):
def __init__(self, power=2):
super(Normalize, self).__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
class ReshapeF(nn.Module):
def __init__(self):
super(ReshapeF, self).__init__()
model = [nn.AdaptiveAvgPool2d(4)]
self.model = nn.Sequential(*model)
self.l2norm = Normalize(2)
def forward(self, x):
x = self.model(x)
x_reshape = x.permute(0, 2, 3, 1).flatten(0, 2)
return self.l2norm(x_reshape)
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.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
@triton.jit
def triton_poi_fused_add_div_pow_sum_0(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
x1 = xindex
y0 = yindex
tmp0 = tl.load(in_ptr0 + (16 * x1 + 64 * (y0 // 16) + y0 % 16), xmask &
ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (64 * (y0 // 16) + y0 % 16), ymask,
eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + 64 * (y0 // 16) + y0 % 16), ymask,
eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (32 + 64 * (y0 // 16) + y0 % 16), ymask,
eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (48 + 64 * (y0 // 16) + y0 % 16), 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-07
tmp14 = tmp12 + tmp13
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + (x1 + 4 * y0), tmp15, xmask & ymask)
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((64, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_div_pow_sum_0[grid(64, 4)](arg0_1, buf0, 64, 4,
XBLOCK=4, YBLOCK=32, num_warps=4, num_stages=1)
del arg0_1
return buf0,
class Normalize(nn.Module):
def __init__(self, power=2):
super(Normalize, self).__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
class ReshapeFNew(nn.Module):
def __init__(self):
super(ReshapeFNew, self).__init__()
model = [nn.AdaptiveAvgPool2d(4)]
self.model = nn.Sequential(*model)
self.l2norm = Normalize(2)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| guyii54/Contrastive-I2I | ReshapeF | false | 6,767 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import torch
import torch.utils.data
import torch
from torch import nn
class Normalize(nn.Module):
def __init__(self, power=2):
super().__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
class Model(nn.Module):
def __init__(self):
super().__init__()
model = [nn.AdaptiveAvgPool2d(4)]
self.model = nn.Sequential(*model)
self.l2norm = Normalize(2)
def forward(self, x):
x = self.model(x)
x_reshape = x.permute(0, 2, 3, 1).flatten(0, 2)
return self.l2norm(x_reshape)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
SCConv_Layer | # 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_4/inductor_cache/bz/cbzaxiqw5wvhy6st3pbxw7pxkl56seuelga6fjbxl6vmuse54t5v.py
# Topologically Sorted Source Nodes: [stack], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# stack => cat
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_3, %primals_10, %primals_20, %primals_20, %primals_10],), kwargs = {})
triton_poi_fused_stack_0 = async_compile.triton('triton_poi_fused_stack_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: '*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_stack_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_stack_0(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
x1 = (xindex // 4)
x0 = xindex % 4
x2 = 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)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + (4*((-4) + x1))), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (x0 + (4*((-8) + x1))), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr2 + (x0 + (4*((-12) + x1))), tmp19 & xmask, other=0.0)
tmp21 = tmp0 >= tmp17
tmp22 = tl.full([1], 20, tl.int64)
tmp23 = tmp0 < tmp22
tmp24 = tl.load(in_ptr1 + (x0 + (4*((-16) + x1))), tmp21 & xmask, other=0.0)
tmp25 = tl.where(tmp19, tmp20, tmp24)
tmp26 = tl.where(tmp14, tmp15, tmp25)
tmp27 = tl.where(tmp9, tmp10, tmp26)
tmp28 = tl.where(tmp4, tmp5, tmp27)
tl.store(out_ptr0 + (x2), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/d5/cd5avuhhjsvef6q2smzw7obxl67nlcfoq74e3w3gucxesteng7az.py
# Topologically Sorted Source Nodes: [stack_1], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# stack_1 => cat_1
# Graph fragment:
# %cat_1 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_1, %primals_8, %primals_22, %primals_18, %primals_15],), kwargs = {})
triton_poi_fused_stack_1 = async_compile.triton('triton_poi_fused_stack_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: '*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_stack_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_stack_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 80
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 4)
x0 = xindex % 4
x2 = 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)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + (4*((-4) + x1))), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (x0 + (4*((-8) + x1))), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr3 + (x0 + (4*((-12) + x1))), tmp19 & xmask, other=0.0)
tmp21 = tmp0 >= tmp17
tmp22 = tl.full([1], 20, tl.int64)
tmp23 = tmp0 < tmp22
tmp24 = tl.load(in_ptr4 + (x0 + (4*((-16) + x1))), tmp21 & xmask, other=0.0)
tmp25 = tl.where(tmp19, tmp20, tmp24)
tmp26 = tl.where(tmp14, tmp15, tmp25)
tmp27 = tl.where(tmp9, tmp10, tmp26)
tmp28 = tl.where(tmp4, tmp5, tmp27)
tl.store(out_ptr0 + (x2), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/n6/cn6uzhsj4p6y54icvbu7phnuxo4m7pqo3cp3getxhi5452wtzpm5.py
# Topologically Sorted Source Nodes: [stack_2], Original ATen: [aten.stack]
# Source node to ATen node mapping:
# stack_2 => cat_2
# Graph fragment:
# %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%primals_2, %primals_9, %primals_23, %primals_19, %primals_16],), kwargs = {})
triton_poi_fused_stack_2 = async_compile.triton('triton_poi_fused_stack_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=[32],
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), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_stack_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_stack_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 20
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = 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 + (x0), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + ((-4) + x0), tmp9 & xmask, eviction_policy='evict_last', other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + ((-8) + x0), tmp14 & xmask, eviction_policy='evict_last', other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr3 + ((-12) + x0), tmp19 & xmask, eviction_policy='evict_last', other=0.0)
tmp21 = tmp0 >= tmp17
tmp22 = tl.full([1], 20, tl.int64)
tmp23 = tmp0 < tmp22
tmp24 = tl.load(in_ptr4 + ((-16) + x0), tmp21 & xmask, eviction_policy='evict_last', other=0.0)
tmp25 = tl.where(tmp19, tmp20, tmp24)
tmp26 = tl.where(tmp14, tmp15, tmp25)
tmp27 = tl.where(tmp9, tmp10, tmp26)
tmp28 = tl.where(tmp4, tmp5, tmp27)
tl.store(out_ptr0 + (x0), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/t5/ct5rx6v7dkrmt2gksbc2ocdzby4i5tmmwzmibw4x3uf6hyqdv6ko.py
# Topologically Sorted Source Nodes: [n2n_1], Original ATen: [aten.zeros]
# Source node to ATen node mapping:
# n2n_1 => full
# Graph fragment:
# %full : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
triton_poi_fused_zeros_3 = async_compile.triton('triton_poi_fused_zeros_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: '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_zeros_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_zeros_3(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 = 0.0
tl.store(out_ptr0 + (x0), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ev/cevk5nxjhvrl3qc7otvasr4bby7uoclz6akzfj5d3wlqvdd3qpfd.py
# Topologically Sorted Source Nodes: [baddbmm], Original ATen: [aten.baddbmm]
# Source node to ATen node mapping:
# baddbmm => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%unsqueeze, %bmm), kwargs = {})
triton_poi_fused_baddbmm_4 = async_compile.triton('triton_poi_fused_baddbmm_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=[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_baddbmm_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_baddbmm_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 80
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = (xindex // 16)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_out_ptr0 + (x3), xmask)
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + (x3), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/fk/cfkvvmpcu6hrqaxgkxb6lcfetkmrqsw2bz4vnk6fzsje7lnhr26h.py
# Topologically Sorted Source Nodes: [add, relu, X0], Original ATen: [aten.add, aten.relu, aten.mul, aten.threshold_backward]
# Source node to ATen node mapping:
# X0 => mul
# add => add_1
# relu => relu
# Graph fragment:
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%permute_4, %permute_13), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_1,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%relu, 0.5), kwargs = {})
# %le_2 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_mul_relu_threshold_backward_5 = async_compile.triton('triton_poi_fused_add_mul_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=[16],
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_mul_relu_threshold_backward_5', '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_mul_relu_threshold_backward_5(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 + (x0), xmask)
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.5
tmp6 = tmp4 * tmp5
tmp7 = 0.0
tmp8 = tmp4 <= tmp7
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/hd/chdfete5abdgxaowd5v7sftqfrdzmyxqkhnycwe6ny3eileuwdnt.py
# Topologically Sorted Source Nodes: [add_1, add_2, relu_1, X1], Original ATen: [aten.add, aten.relu, aten.mul, aten.threshold_backward]
# Source node to ATen node mapping:
# X1 => mul_1
# add_1 => add_2
# add_2 => add_3
# relu_1 => relu_1
# Graph fragment:
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%permute_18, %permute_9), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_2, %permute_30), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_3,), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%relu_1, 0.3333333333333333), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_add_mul_relu_threshold_backward_6 = async_compile.triton('triton_poi_fused_add_mul_relu_threshold_backward_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: '*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_mul_relu_threshold_backward_6', '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_relu_threshold_backward_6(in_ptr0, in_ptr1, in_ptr2, 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 + (x0), xmask)
tmp3 = tl.load(in_ptr2 + (x0), xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.full([1], 0, tl.int32)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = 0.3333333333333333
tmp8 = tmp6 * tmp7
tmp9 = 0.0
tmp10 = tmp6 <= tmp9
tl.store(out_ptr0 + (x0), tmp8, xmask)
tl.store(out_ptr1 + (x0), 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, 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 = 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, 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, ), (1, ))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4, 4), (4, 1))
assert_size_stride(primals_13, (4, ), (1, ))
assert_size_stride(primals_14, (4, 4), (4, 1))
assert_size_stride(primals_15, (4, 4), (4, 1))
assert_size_stride(primals_16, (4, ), (1, ))
assert_size_stride(primals_17, (4, 4), (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, 1))
assert_size_stride(primals_21, (4, 4), (4, 1))
assert_size_stride(primals_22, (4, 4), (4, 1))
assert_size_stride(primals_23, (4, ), (1, ))
assert_size_stride(primals_24, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((20, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [stack], Original ATen: [aten.stack]
stream0 = get_raw_stream(0)
triton_poi_fused_stack_0.run(primals_3, primals_10, primals_20, buf0, 80, grid=grid(80), stream=stream0)
del primals_20
buf1 = empty_strided_cuda((20, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [stack_1], Original ATen: [aten.stack]
triton_poi_fused_stack_1.run(primals_1, primals_8, primals_22, primals_18, primals_15, buf1, 80, grid=grid(80), stream=stream0)
del primals_1
del primals_15
del primals_18
del primals_22
del primals_8
buf2 = empty_strided_cuda((20, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [stack_2], Original ATen: [aten.stack]
triton_poi_fused_stack_2.run(primals_2, primals_9, primals_23, primals_19, primals_16, buf2, 20, grid=grid(20), stream=stream0)
del primals_16
del primals_19
del primals_2
del primals_23
del primals_9
buf3 = empty_strided_cuda((5, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [baddbmm], Original ATen: [aten.baddbmm]
extern_kernels.bmm(reinterpret_tensor(buf0, (5, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf1, (5, 4, 4), (16, 1, 4), 0), out=buf3)
del buf1
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [n2n_1], Original ATen: [aten.zeros]
triton_poi_fused_zeros_3.run(buf4, 16, grid=grid(16), stream=stream0)
buf5 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [baddbmm], Original ATen: [aten.baddbmm]
triton_poi_fused_baddbmm_4.run(buf5, buf2, 80, grid=grid(80), stream=stream0)
del buf2
# Topologically Sorted Source Nodes: [n2n_1], Original ATen: [aten._sparse_addmm]
buf6 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1, 4), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), beta=0)
buf7 = buf6
del buf6
buf8 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [n2e], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_6, primals_3, reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf8)
del primals_5
del primals_6
# Topologically Sorted Source Nodes: [n2e_1], Original ATen: [aten._sparse_addmm]
buf9 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf8, (4, 4), (1, 4), 0), reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), beta=0)
buf10 = buf9
del buf9
# Topologically Sorted Source Nodes: [e2n_1], Original ATen: [aten._sparse_addmm]
buf11 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1, 4), 16), reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), beta=0)
buf12 = buf11
del buf11
buf13 = buf8; del buf8 # reuse
# Topologically Sorted Source Nodes: [e2e], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_13, primals_10, reinterpret_tensor(primals_12, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf13)
del primals_12
del primals_13
# Topologically Sorted Source Nodes: [e2e_1], Original ATen: [aten._sparse_addmm]
buf14 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf13, (4, 4), (1, 4), 0), reinterpret_tensor(primals_14, (4, 4), (1, 4), 0), beta=0)
del buf13
buf15 = buf14
del buf14
# Topologically Sorted Source Nodes: [e2t_1], Original ATen: [aten._sparse_addmm]
buf16 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1, 4), 64), reinterpret_tensor(primals_17, (4, 4), (1, 4), 0), beta=0)
buf17 = buf16
del buf16
# Topologically Sorted Source Nodes: [t2t_1], Original ATen: [aten._sparse_addmm]
buf18 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1, 4), 48), reinterpret_tensor(primals_21, (4, 4), (1, 4), 0), beta=0)
buf19 = buf18
del buf18
# Topologically Sorted Source Nodes: [t2e_1], Original ATen: [aten._sparse_addmm]
buf20 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1, 4), 32), reinterpret_tensor(primals_24, (4, 4), (1, 4), 0), beta=0)
del buf5
buf21 = buf20
del buf20
buf22 = reinterpret_tensor(buf4, (4, 4), (1, 4), 0); del buf4 # reuse
buf27 = empty_strided_cuda((4, 4), (1, 4), torch.bool)
# Topologically Sorted Source Nodes: [add, relu, X0], Original ATen: [aten.add, aten.relu, aten.mul, aten.threshold_backward]
triton_poi_fused_add_mul_relu_threshold_backward_5.run(buf7, buf12, buf22, buf27, 16, grid=grid(16), stream=stream0)
del buf12
buf23 = reinterpret_tensor(buf7, (4, 4), (1, 4), 0); del buf7 # reuse
buf26 = empty_strided_cuda((4, 4), (1, 4), torch.bool)
# Topologically Sorted Source Nodes: [add_1, add_2, relu_1, X1], Original ATen: [aten.add, aten.relu, aten.mul, aten.threshold_backward]
triton_poi_fused_add_mul_relu_threshold_backward_6.run(buf15, buf10, buf21, buf23, buf26, 16, grid=grid(16), stream=stream0)
del buf10
del buf15
buf24 = reinterpret_tensor(buf21, (4, 4), (1, 4), 0); del buf21 # reuse
buf25 = empty_strided_cuda((4, 4), (1, 4), torch.bool)
# Topologically Sorted Source Nodes: [add_3, relu_2, X2], Original ATen: [aten.add, aten.relu, aten.mul, aten.threshold_backward]
triton_poi_fused_add_mul_relu_threshold_backward_5.run(buf19, buf17, buf24, buf25, 16, grid=grid(16), stream=stream0)
del buf17
del buf19
return (buf22, buf23, buf24, primals_3, primals_10, buf25, buf26, buf27, primals_24, primals_21, primals_17, primals_14, primals_11, primals_7, primals_4, reinterpret_tensor(buf0, (5, 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, 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, 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, ), (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, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((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, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_12 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_13 = rand_strided((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, 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, 4), (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, 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, 4), (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, 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, 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])
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 SCConv_Layer(nn.Module):
def __init__(self, num_node_feats, num_edge_feats, num_triangle_feats,
output_size, bias=True, f=F.relu):
super().__init__()
self.n2n_weights = nn.Linear(num_node_feats, output_size, bias=bias)
self.n2e_weights = nn.Linear(num_node_feats, output_size, bias=bias)
self.e2e_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.e2n_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.e2t_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.t2e_weights = nn.Linear(num_triangle_feats, output_size, bias=bias
)
self.t2t_weights = nn.Linear(num_triangle_feats, output_size, bias=bias
)
self.w = f
def forward(self, X0, X1, X2, L0, L1, L2, B2D3, D2B1TD1inv, D1invB1,
B2TD2inv):
n2n = self.n2n_weights(X0)
n2n = torch.sparse.mm(L0, n2n)
n2e = self.n2e_weights(X0)
n2e = torch.sparse.mm(D2B1TD1inv, n2e)
e2n = self.e2n_weights(X1)
e2n = torch.sparse.mm(D1invB1, e2n)
e2e = self.e2e_weights(X1)
e2e = torch.sparse.mm(L1, e2e)
e2t = self.e2t_weights(X1)
e2t = torch.sparse.mm(B2TD2inv, e2t)
t2t = self.t2t_weights(X2)
t2t = torch.sparse.mm(L2, t2t)
t2e = self.t2e_weights(X2)
t2e = torch.sparse.mm(B2D3, t2e)
X0 = 1 / 2.0 * self.w(n2n + e2n)
X1 = 1 / 3.0 * self.w(e2e + n2e + t2e)
X2 = 1 / 2.0 * self.w(t2t + e2t)
return X0, X1, X2
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4]),
torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4]), torch.
rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4,
4])]
def get_init_inputs():
return [[], {'num_node_feats': 4, 'num_edge_feats': 4,
'num_triangle_feats': 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
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_stack_0(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
x1 = xindex // 4
x0 = xindex % 4
x2 = 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), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + 4 * (-4 + x1)), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (x0 + 4 * (-8 + x1)), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr2 + (x0 + 4 * (-12 + x1)), tmp19 & xmask, other=0.0)
tmp21 = tmp0 >= tmp17
tl.full([1], 20, tl.int64)
tmp24 = tl.load(in_ptr1 + (x0 + 4 * (-16 + x1)), tmp21 & xmask, other=0.0)
tmp25 = tl.where(tmp19, tmp20, tmp24)
tmp26 = tl.where(tmp14, tmp15, tmp25)
tmp27 = tl.where(tmp9, tmp10, tmp26)
tmp28 = tl.where(tmp4, tmp5, tmp27)
tl.store(out_ptr0 + x2, tmp28, xmask)
@triton.jit
def triton_poi_fused_stack_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 80
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 4
x0 = xindex % 4
x2 = 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), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + 4 * (-4 + x1)), tmp9 & xmask, other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (x0 + 4 * (-8 + x1)), tmp14 & xmask, other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr3 + (x0 + 4 * (-12 + x1)), tmp19 & xmask, other=0.0)
tmp21 = tmp0 >= tmp17
tl.full([1], 20, tl.int64)
tmp24 = tl.load(in_ptr4 + (x0 + 4 * (-16 + x1)), tmp21 & xmask, other=0.0)
tmp25 = tl.where(tmp19, tmp20, tmp24)
tmp26 = tl.where(tmp14, tmp15, tmp25)
tmp27 = tl.where(tmp9, tmp10, tmp26)
tmp28 = tl.where(tmp4, tmp5, tmp27)
tl.store(out_ptr0 + x2, tmp28, xmask)
@triton.jit
def triton_poi_fused_stack_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 20
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex
tmp0 = x0
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + x0, tmp4 & xmask, eviction_policy='evict_last',
other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 8, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (-4 + x0), tmp9 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 12, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tmp11 & tmp13
tmp15 = tl.load(in_ptr2 + (-8 + x0), tmp14 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp16 = tmp0 >= tmp12
tmp17 = tl.full([1], 16, tl.int64)
tmp18 = tmp0 < tmp17
tmp19 = tmp16 & tmp18
tmp20 = tl.load(in_ptr3 + (-12 + x0), tmp19 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp21 = tmp0 >= tmp17
tl.full([1], 20, tl.int64)
tmp24 = tl.load(in_ptr4 + (-16 + x0), tmp21 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp25 = tl.where(tmp19, tmp20, tmp24)
tmp26 = tl.where(tmp14, tmp15, tmp25)
tmp27 = tl.where(tmp9, tmp10, tmp26)
tmp28 = tl.where(tmp4, tmp5, tmp27)
tl.store(out_ptr0 + x0, tmp28, xmask)
@triton.jit
def triton_poi_fused_zeros_3(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 = 0.0
tl.store(out_ptr0 + x0, tmp0, xmask)
@triton.jit
def triton_poi_fused_baddbmm_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 80
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 4
x2 = xindex // 16
x3 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp1 = tl.load(in_out_ptr0 + x3, xmask)
tmp2 = tmp0 + tmp1
tl.store(in_out_ptr0 + x3, tmp2, xmask)
@triton.jit
def triton_poi_fused_add_mul_relu_threshold_backward_5(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 + x0, xmask)
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tmp5 = 0.5
tmp6 = tmp4 * tmp5
tmp7 = 0.0
tmp8 = tmp4 <= tmp7
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp8, xmask)
@triton.jit
def triton_poi_fused_add_mul_relu_threshold_backward_6(in_ptr0, in_ptr1,
in_ptr2, 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 + x0, xmask)
tmp3 = tl.load(in_ptr2 + x0, xmask)
tmp2 = tmp0 + tmp1
tmp4 = tmp2 + tmp3
tmp5 = tl.full([1], 0, tl.int32)
tmp6 = triton_helpers.maximum(tmp5, tmp4)
tmp7 = 0.3333333333333333
tmp8 = tmp6 * tmp7
tmp9 = 0.0
tmp10 = tmp6 <= tmp9
tl.store(out_ptr0 + x0, tmp8, xmask)
tl.store(out_ptr1 + x0, 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, 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) = 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, 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,), (1,))
assert_size_stride(primals_7, (4, 4), (4, 1))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (4, 4), (4, 1))
assert_size_stride(primals_11, (4, 4), (4, 1))
assert_size_stride(primals_12, (4, 4), (4, 1))
assert_size_stride(primals_13, (4,), (1,))
assert_size_stride(primals_14, (4, 4), (4, 1))
assert_size_stride(primals_15, (4, 4), (4, 1))
assert_size_stride(primals_16, (4,), (1,))
assert_size_stride(primals_17, (4, 4), (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, 1))
assert_size_stride(primals_21, (4, 4), (4, 1))
assert_size_stride(primals_22, (4, 4), (4, 1))
assert_size_stride(primals_23, (4,), (1,))
assert_size_stride(primals_24, (4, 4), (4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((20, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_stack_0[grid(80)](primals_3, primals_10,
primals_20, buf0, 80, XBLOCK=128, num_warps=4, num_stages=1)
del primals_20
buf1 = empty_strided_cuda((20, 4), (4, 1), torch.float32)
triton_poi_fused_stack_1[grid(80)](primals_1, primals_8, primals_22,
primals_18, primals_15, buf1, 80, XBLOCK=128, num_warps=4,
num_stages=1)
del primals_1
del primals_15
del primals_18
del primals_22
del primals_8
buf2 = empty_strided_cuda((20,), (1,), torch.float32)
triton_poi_fused_stack_2[grid(20)](primals_2, primals_9, primals_23,
primals_19, primals_16, buf2, 20, XBLOCK=32, num_warps=1,
num_stages=1)
del primals_16
del primals_19
del primals_2
del primals_23
del primals_9
buf3 = empty_strided_cuda((5, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(buf0, (5, 4, 4), (16, 4, 1),
0), reinterpret_tensor(buf1, (5, 4, 4), (16, 1, 4), 0), out=buf3)
del buf1
buf4 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_zeros_3[grid(16)](buf4, 16, XBLOCK=16, num_warps=1,
num_stages=1)
buf5 = buf3
del buf3
triton_poi_fused_baddbmm_4[grid(80)](buf5, buf2, 80, XBLOCK=128,
num_warps=4, num_stages=1)
del buf2
buf6 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4,
(4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1, 4), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), beta=0)
buf7 = buf6
del buf6
buf8 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_6, primals_3, reinterpret_tensor(
primals_5, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf8)
del primals_5
del primals_6
buf9 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(buf4,
(4, 4), (1, 4), 0), reinterpret_tensor(buf8, (4, 4), (1, 4), 0),
reinterpret_tensor(primals_7, (4, 4), (1, 4), 0), beta=0)
buf10 = buf9
del buf9
buf11 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(
buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1,
4), 16), reinterpret_tensor(primals_11, (4, 4), (1, 4), 0), beta=0)
buf12 = buf11
del buf11
buf13 = buf8
del buf8
extern_kernels.addmm(primals_13, primals_10, reinterpret_tensor(
primals_12, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf13)
del primals_12
del primals_13
buf14 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(
buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf13, (4, 4), (1,
4), 0), reinterpret_tensor(primals_14, (4, 4), (1, 4), 0), beta=0)
del buf13
buf15 = buf14
del buf14
buf16 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(
buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1,
4), 64), reinterpret_tensor(primals_17, (4, 4), (1, 4), 0), beta=0)
buf17 = buf16
del buf16
buf18 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(
buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1,
4), 48), reinterpret_tensor(primals_21, (4, 4), (1, 4), 0), beta=0)
buf19 = buf18
del buf18
buf20 = torch.ops.aten._sparse_addmm.default(reinterpret_tensor(
buf4, (4, 4), (1, 4), 0), reinterpret_tensor(buf5, (4, 4), (1,
4), 32), reinterpret_tensor(primals_24, (4, 4), (1, 4), 0), beta=0)
del buf5
buf21 = buf20
del buf20
buf22 = reinterpret_tensor(buf4, (4, 4), (1, 4), 0)
del buf4
buf27 = empty_strided_cuda((4, 4), (1, 4), torch.bool)
triton_poi_fused_add_mul_relu_threshold_backward_5[grid(16)](buf7,
buf12, buf22, buf27, 16, XBLOCK=16, num_warps=1, num_stages=1)
del buf12
buf23 = reinterpret_tensor(buf7, (4, 4), (1, 4), 0)
del buf7
buf26 = empty_strided_cuda((4, 4), (1, 4), torch.bool)
triton_poi_fused_add_mul_relu_threshold_backward_6[grid(16)](buf15,
buf10, buf21, buf23, buf26, 16, XBLOCK=16, num_warps=1,
num_stages=1)
del buf10
del buf15
buf24 = reinterpret_tensor(buf21, (4, 4), (1, 4), 0)
del buf21
buf25 = empty_strided_cuda((4, 4), (1, 4), torch.bool)
triton_poi_fused_add_mul_relu_threshold_backward_5[grid(16)](buf19,
buf17, buf24, buf25, 16, XBLOCK=16, num_warps=1, num_stages=1)
del buf17
del buf19
return (buf22, buf23, buf24, primals_3, primals_10, buf25, buf26, buf27,
primals_24, primals_21, primals_17, primals_14, primals_11,
primals_7, primals_4, reinterpret_tensor(buf0, (5, 4, 4), (16, 1, 4
), 0))
class SCConv_LayerNew(nn.Module):
def __init__(self, num_node_feats, num_edge_feats, num_triangle_feats,
output_size, bias=True, f=F.relu):
super().__init__()
self.n2n_weights = nn.Linear(num_node_feats, output_size, bias=bias)
self.n2e_weights = nn.Linear(num_node_feats, output_size, bias=bias)
self.e2e_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.e2n_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.e2t_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.t2e_weights = nn.Linear(num_triangle_feats, output_size, bias=bias
)
self.t2t_weights = nn.Linear(num_triangle_feats, output_size, bias=bias
)
self.w = f
def forward(self, input_0, input_1, input_2, input_3, input_4, input_5,
input_6, input_7, input_8, input_9):
primals_1 = self.n2n_weights.weight
primals_2 = self.n2n_weights.bias
primals_3 = self.n2e_weights.weight
primals_6 = self.n2e_weights.bias
primals_4 = self.e2e_weights.weight
primals_9 = self.e2e_weights.bias
primals_5 = self.e2n_weights.weight
primals_13 = self.e2n_weights.bias
primals_7 = self.e2t_weights.weight
primals_16 = self.e2t_weights.bias
primals_8 = self.t2e_weights.weight
primals_19 = self.t2e_weights.bias
primals_10 = self.t2t_weights.weight
primals_23 = self.t2t_weights.bias
primals_11 = input_0
primals_12 = input_1
primals_14 = input_2
primals_15 = input_3
primals_17 = input_4
primals_18 = input_5
primals_20 = input_6
primals_21 = input_7
primals_22 = input_8
primals_24 = input_9
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])
return output[0], output[1], output[2]
| ggoh29/Simplicial-neural-network-benchmark | SCConv_Layer | false | 6,768 | [
"MIT"
] | 1 | 9a12bcd054251790d85e3971f5473dcffaa5664b | https://github.com/ggoh29/Simplicial-neural-network-benchmark/tree/9a12bcd054251790d85e3971f5473dcffaa5664b | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, num_node_feats, num_edge_feats, num_triangle_feats,
output_size, bias=True, f=F.relu):
super().__init__()
self.n2n_weights = nn.Linear(num_node_feats, output_size, bias=bias)
self.n2e_weights = nn.Linear(num_node_feats, output_size, bias=bias)
self.e2e_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.e2n_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.e2t_weights = nn.Linear(num_edge_feats, output_size, bias=bias)
self.t2e_weights = nn.Linear(num_triangle_feats, output_size, bias=bias
)
self.t2t_weights = nn.Linear(num_triangle_feats, output_size, bias=bias
)
self.w = f
def forward(self, X0, X1, X2, L0, L1, L2, B2D3, D2B1TD1inv, D1invB1,
B2TD2inv):
n2n = self.n2n_weights(X0)
n2n = torch.sparse.mm(L0, n2n)
n2e = self.n2e_weights(X0)
n2e = torch.sparse.mm(D2B1TD1inv, n2e)
e2n = self.e2n_weights(X1)
e2n = torch.sparse.mm(D1invB1, e2n)
e2e = self.e2e_weights(X1)
e2e = torch.sparse.mm(L1, e2e)
e2t = self.e2t_weights(X1)
e2t = torch.sparse.mm(B2TD2inv, e2t)
t2t = self.t2t_weights(X2)
t2t = torch.sparse.mm(L2, t2t)
t2e = self.t2e_weights(X2)
t2e = torch.sparse.mm(B2D3, t2e)
X0 = 1 / 2.0 * self.w(n2n + e2n)
X1 = 1 / 3.0 * self.w(e2e + n2e + t2e)
X2 = 1 / 2.0 * self.w(t2t + e2t)
return X0, X1, X2
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4]),
torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4]), torch.
rand([4, 4]), torch.rand([4, 4]), torch.rand([4, 4]), torch.rand([4,
4])]
def get_init_inputs():
return [[], {'num_node_feats': 4, 'num_edge_feats': 4,
'num_triangle_feats': 4, 'output_size': 4}]
|
FusedLeakyReLU | # 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_4/inductor_cache/cx/ccxnongcgusvhvf5whrpjbz4ddnlsjovjzjngbpjfxvhdw7ihrhu.py
# Topologically Sorted Source Nodes: [add, leaky_relu, out], Original ATen: [aten.add, aten.leaky_relu, aten.mul]
# Source node to ATen node mapping:
# add => add
# leaky_relu => gt, mul, where
# out => mul_1
# Graph fragment:
# %add : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_2, %primals_1), kwargs = {})
# %gt : [num_users=2] = call_function[target=torch.ops.aten.gt.Scalar](args = (%add, 0), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add, 0.2), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %add, %mul), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%where, 1.4142135623730951), kwargs = {})
triton_poi_fused_add_leaky_relu_mul_0 = async_compile.triton('triton_poi_fused_add_leaky_relu_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: '*i1', 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_leaky_relu_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_add_leaky_relu_mul_0(in_ptr0, in_ptr1, 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 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.2
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tmp8 = 1.4142135623730951
tmp9 = tmp7 * tmp8
tl.store(out_ptr0 + (x3), tmp4, xmask)
tl.store(out_ptr1 + (x3), tmp9, 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, (1, 4, 1, 1), (4, 1, 1, 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.bool)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, leaky_relu, out], Original ATen: [aten.add, aten.leaky_relu, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_add_leaky_relu_mul_0.run(primals_2, primals_1, buf0, buf1, 256, grid=grid(256), stream=stream0)
del primals_1
del primals_2
return (buf1, 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((1, 4, 1, 1), (4, 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)
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.utils.data
import torch
from torch import nn
import torch.nn.functional as F
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class FusedLeakyReLU(nn.Module):
def __init__(self, channel, negative_slope=0.2, scale=2 ** 0.5):
super().__init__()
self.bias = nn.Parameter(torch.zeros(1, channel, 1, 1))
self.negative_slope = negative_slope
self.scale = scale
def forward(self, input):
out = fused_leaky_relu(input, self.bias, self.negative_slope, self.
scale)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'channel': 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.utils.data
import torch
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
@triton.jit
def triton_poi_fused_add_leaky_relu_mul_0(in_ptr0, in_ptr1, 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 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = tmp2 > tmp3
tmp5 = 0.2
tmp6 = tmp2 * tmp5
tmp7 = tl.where(tmp4, tmp2, tmp6)
tmp8 = 1.4142135623730951
tmp9 = tmp7 * tmp8
tl.store(out_ptr0 + x3, tmp4, xmask)
tl.store(out_ptr1 + x3, tmp9, xmask)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (1, 4, 1, 1), (4, 1, 1, 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.bool)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_add_leaky_relu_mul_0[grid(256)](primals_2,
primals_1, buf0, buf1, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
del primals_2
return buf1, buf0
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class FusedLeakyReLUNew(nn.Module):
def __init__(self, channel, negative_slope=0.2, scale=2 ** 0.5):
super().__init__()
self.bias = nn.Parameter(torch.zeros(1, channel, 1, 1))
self.negative_slope = negative_slope
self.scale = scale
def forward(self, input_0):
primals_1 = self.bias
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| guyii54/Contrastive-I2I | FusedLeakyReLU | false | 6,769 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import torch
import torch.utils.data
import torch
from torch import nn
import torch.nn.functional as F
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Model(nn.Module):
def __init__(self, channel, negative_slope=0.2, scale=2 ** 0.5):
super().__init__()
self.bias = nn.Parameter(torch.zeros(1, channel, 1, 1))
self.negative_slope = negative_slope
self.scale = scale
def forward(self, input):
out = fused_leaky_relu(input, self.bias, self.negative_slope, self.
scale)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
DCRBranch | # 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_4/inductor_cache/3a/c3ahpcqb72o7gdbgt53cgnedr7p2tje6sfperji7zazze5egjmmz.py
# Topologically Sorted Source Nodes: [prelu], Original ATen: [aten._prelu_kernel]
# Source node to ATen node mapping:
# prelu => gt, mul, 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_2, %view_1), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %view_1, %mul), kwargs = {})
triton_poi_fused__prelu_kernel_0 = async_compile.triton('triton_poi_fused__prelu_kernel_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__prelu_kernel_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__prelu_kernel_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)
tmp3 = tl.load(in_ptr1 + (0))
tmp4 = tl.broadcast_to(tmp3, [XBLOCK])
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp5 = tmp4 * tmp0
tmp6 = tl.where(tmp2, tmp0, tmp5)
tl.store(out_ptr0 + (x0), 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, 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, (1, ), (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((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [representation], 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, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prelu], Original ATen: [aten._prelu_kernel]
stream0 = get_raw_stream(0)
triton_poi_fused__prelu_kernel_0.run(buf0, primals_4, buf1, 256, grid=grid(256), stream=stream0)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [classification], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_6, reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del primals_6
return (reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0), primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(buf1, (64, 4), (4, 1), 0), primals_5, )
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((1, ), (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
import torch.nn as nn
import torch.utils.data
import torch.optim
import torch.utils.data.distributed
class DCRBranch(nn.Module):
"""Branch Network for DCR"""
def __init__(self, num_classes, in_channels, mid_channels,
normalized_embeddings=False):
super().__init__()
self.normalized_embeddings = normalized_embeddings
self.fc1 = nn.Linear(in_channels, mid_channels)
self.fc2 = nn.Linear(mid_channels, num_classes)
self.relu = nn.PReLU()
def forward(self, img):
if self.normalized_embeddings:
representation = torch.nn.functional.normalize(self.fc1(img))
else:
representation = self.fc1(img)
classification = self.fc2(self.relu(representation))
return representation, classification
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_classes': 4, 'in_channels': 4, 'mid_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
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__prelu_kernel_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)
tmp3 = tl.load(in_ptr1 + 0)
tmp4 = tl.broadcast_to(tmp3, [XBLOCK])
tmp1 = 0.0
tmp2 = tmp0 > tmp1
tmp5 = tmp4 * tmp0
tmp6 = tl.where(tmp2, tmp0, tmp5)
tl.store(out_ptr0 + x0, 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, 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, (1,), (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((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, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__prelu_kernel_0[grid(256)](buf0, primals_4, buf1,
256, XBLOCK=128, num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_6, reinterpret_tensor(buf1, (64, 4), (
4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf2)
del primals_6
return reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(buf2, (4, 4, 4, 4), (64, 16, 4, 1), 0
), primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf0, (4, 4, 4, 4), (64, 16, 4, 1), 0
), reinterpret_tensor(buf1, (64, 4), (4, 1), 0), primals_5
class DCRBranchNew(nn.Module):
"""Branch Network for DCR"""
def __init__(self, num_classes, in_channels, mid_channels,
normalized_embeddings=False):
super().__init__()
self.normalized_embeddings = normalized_embeddings
self.fc1 = nn.Linear(in_channels, mid_channels)
self.fc2 = nn.Linear(mid_channels, num_classes)
self.relu = nn.PReLU()
def forward(self, input_0):
primals_1 = self.fc1.weight
primals_2 = self.fc1.bias
primals_5 = self.fc2.weight
primals_6 = self.fc2.bias
primals_4 = self.relu.weight
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6])
return output[0], output[1]
| gyfastas/CS7319E1G16 | DCRBranch | false | 6,770 | [
"MIT"
] | 1 | 03126af04766abcb269d0c8db481c96c856d21ef | https://github.com/gyfastas/CS7319E1G16/tree/03126af04766abcb269d0c8db481c96c856d21ef | import torch
import torch.nn as nn
import torch.utils.data
import torch.optim
import torch.utils.data.distributed
class Model(nn.Module):
"""Branch Network for DCR"""
def __init__(self, num_classes, in_channels, mid_channels,
normalized_embeddings=False):
super().__init__()
self.normalized_embeddings = normalized_embeddings
self.fc1 = nn.Linear(in_channels, mid_channels)
self.fc2 = nn.Linear(mid_channels, num_classes)
self.relu = nn.PReLU()
def forward(self, img):
if self.normalized_embeddings:
representation = torch.nn.functional.normalize(self.fc1(img))
else:
representation = self.fc1(img)
classification = self.fc2(self.relu(representation))
return representation, classification
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
LinearAttentionLayer | # 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_4/inductor_cache/vl/cvle5osmyghpxk2pin4sxeeclrahm5putsp4ne6dffhyumijovq5.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_4, 1), 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=[64],
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 = 64
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_4/inductor_cache/v4/cv4lkzbqtc5n5iyaraeuywvfvxyls3nutfo4p2rrccgn23z7ktmz.py
# Topologically Sorted Source Nodes: [attn_scores_2, alpha], Original ATen: [aten.masked_fill, aten._softmax]
# Source node to ATen node mapping:
# alpha => amax, exp, sub, sum_1
# attn_scores_2 => full_default, where
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -inf), 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, %view_1), 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_masked_fill_1 = async_compile.triton('triton_poi_fused__softmax_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=[16],
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_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_masked_fill_1(in_ptr0, in_ptr1, in_ptr2, 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 % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (16*x1)), xmask).to(tl.int1)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp7 = tl.load(in_ptr0 + (4 + x0 + (16*x1)), xmask).to(tl.int1)
tmp8 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (8 + x0 + (16*x1)), xmask).to(tl.int1)
tmp13 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (12 + x0 + (16*x1)), xmask).to(tl.int1)
tmp18 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp4 = tmp1 + tmp3
tmp5 = float("-inf")
tmp6 = tl.where(tmp0, tmp5, tmp4)
tmp9 = tmp8 + tmp3
tmp10 = tl.where(tmp7, tmp5, tmp9)
tmp11 = triton_helpers.maximum(tmp6, tmp10)
tmp14 = tmp13 + tmp3
tmp15 = tl.where(tmp12, tmp5, tmp14)
tmp16 = triton_helpers.maximum(tmp11, tmp15)
tmp19 = tmp18 + tmp3
tmp20 = tl.where(tmp17, tmp5, tmp19)
tmp21 = triton_helpers.maximum(tmp16, tmp20)
tmp22 = tmp6 - tmp21
tmp23 = tl_math.exp(tmp22)
tmp24 = tmp10 - tmp21
tmp25 = tl_math.exp(tmp24)
tmp26 = tmp23 + tmp25
tmp27 = tmp15 - tmp21
tmp28 = tl_math.exp(tmp27)
tmp29 = tmp26 + tmp28
tmp30 = tmp20 - tmp21
tmp31 = tl_math.exp(tmp30)
tmp32 = tmp29 + tmp31
tl.store(out_ptr0 + (x2), tmp21, xmask)
tl.store(out_ptr1 + (x2), tmp32, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/vp/cvpqglhop6ek6br4n522zary4qd3tyiabrl72yef4qqw6kjjumgz.py
# Topologically Sorted Source Nodes: [attn_scores_2, alpha], Original ATen: [aten.masked_fill, aten._softmax]
# Source node to ATen node mapping:
# alpha => div, exp, sub
# attn_scores_2 => full_default, where
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -inf), 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, %view_1), 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 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), kwargs = {})
triton_poi_fused__softmax_masked_fill_2 = async_compile.triton('triton_poi_fused__softmax_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=[64],
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_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_masked_fill_2(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
x3 = xindex
x4 = xindex % 16
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 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp7 = tl.load(in_ptr3 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr4 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp4 = tmp1 + tmp3
tmp5 = float("-inf")
tmp6 = tl.where(tmp0, tmp5, tmp4)
tmp8 = tmp6 - tmp7
tmp9 = tl_math.exp(tmp8)
tmp11 = tmp9 / tmp10
tl.store(out_ptr0 + (x3), tmp11, 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), (16, 4, 1))
assert_size_stride(primals_2, (1, 4), (4, 1))
assert_size_stride(primals_3, (1, ), (1, ))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 1), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((4, 4, 4), (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_4, buf1, 64, grid=grid(64), stream=stream0)
del primals_4
buf2 = empty_strided_cuda((4, 1, 4), (4, 16, 1), torch.float32)
buf3 = empty_strided_cuda((4, 1, 4), (4, 16, 1), torch.float32)
# Topologically Sorted Source Nodes: [attn_scores_2, alpha], Original ATen: [aten.masked_fill, aten._softmax]
triton_poi_fused__softmax_masked_fill_1.run(buf1, buf0, primals_3, buf2, buf3, 16, grid=grid(16), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [attn_scores_2, alpha], Original ATen: [aten.masked_fill, aten._softmax]
triton_poi_fused__softmax_masked_fill_2.run(buf1, buf0, primals_3, buf2, buf3, buf4, 64, grid=grid(64), stream=stream0)
del buf0
del buf2
del buf3
del primals_3
return (buf4, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), 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, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((4, 4, 4), (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.functional as F
from torch import nn
class LinearAttentionLayer(nn.Module):
def __init__(self, input_dim):
super().__init__()
self.linear = nn.Linear(input_dim, 1)
def forward(self, question, question_mask):
qtn = question.view(-1, question.shape[-1])
attn_scores = self.linear(qtn)
attn_scores = attn_scores.view(question.shape[0], question.shape[1])
attn_scores = attn_scores.masked_fill(question_mask == 1, -float('inf')
)
alpha = F.softmax(attn_scores, dim=1)
return alpha
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'input_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._inductor.runtime.triton_helpers import 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_eq_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
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__softmax_masked_fill_1(in_ptr0, in_ptr1, in_ptr2,
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 % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 16 * x1), xmask).to(tl.int1)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr2 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp7 = tl.load(in_ptr0 + (4 + x0 + 16 * x1), xmask).to(tl.int1)
tmp8 = tl.load(in_ptr1 + (4 + x0), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (8 + x0 + 16 * x1), xmask).to(tl.int1)
tmp13 = tl.load(in_ptr1 + (8 + x0), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (12 + x0 + 16 * x1), xmask).to(tl.int1)
tmp18 = tl.load(in_ptr1 + (12 + x0), xmask, eviction_policy='evict_last')
tmp4 = tmp1 + tmp3
tmp5 = float('-inf')
tmp6 = tl.where(tmp0, tmp5, tmp4)
tmp9 = tmp8 + tmp3
tmp10 = tl.where(tmp7, tmp5, tmp9)
tmp11 = triton_helpers.maximum(tmp6, tmp10)
tmp14 = tmp13 + tmp3
tmp15 = tl.where(tmp12, tmp5, tmp14)
tmp16 = triton_helpers.maximum(tmp11, tmp15)
tmp19 = tmp18 + tmp3
tmp20 = tl.where(tmp17, tmp5, tmp19)
tmp21 = triton_helpers.maximum(tmp16, tmp20)
tmp22 = tmp6 - tmp21
tmp23 = tl_math.exp(tmp22)
tmp24 = tmp10 - tmp21
tmp25 = tl_math.exp(tmp24)
tmp26 = tmp23 + tmp25
tmp27 = tmp15 - tmp21
tmp28 = tl_math.exp(tmp27)
tmp29 = tmp26 + tmp28
tmp30 = tmp20 - tmp21
tmp31 = tl_math.exp(tmp30)
tmp32 = tmp29 + tmp31
tl.store(out_ptr0 + x2, tmp21, xmask)
tl.store(out_ptr1 + x2, tmp32, xmask)
@triton.jit
def triton_poi_fused__softmax_masked_fill_2(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
x3 = xindex
x4 = xindex % 16
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 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp7 = tl.load(in_ptr3 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp10 = tl.load(in_ptr4 + (x0 + 4 * x2), xmask, eviction_policy=
'evict_last')
tmp4 = tmp1 + tmp3
tmp5 = float('-inf')
tmp6 = tl.where(tmp0, tmp5, tmp4)
tmp8 = tmp6 - tmp7
tmp9 = tl_math.exp(tmp8)
tmp11 = tmp9 / tmp10
tl.store(out_ptr0 + x3, tmp11, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_2, (1, 4), (4, 1))
assert_size_stride(primals_3, (1,), (1,))
assert_size_stride(primals_4, (4, 4, 4), (16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((16, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 1), (1, 4), 0), out=buf0)
del primals_2
buf1 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_eq_0[grid(64)](primals_4, buf1, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del primals_4
buf2 = empty_strided_cuda((4, 1, 4), (4, 16, 1), torch.float32)
buf3 = empty_strided_cuda((4, 1, 4), (4, 16, 1), torch.float32)
triton_poi_fused__softmax_masked_fill_1[grid(16)](buf1, buf0,
primals_3, buf2, buf3, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__softmax_masked_fill_2[grid(64)](buf1, buf0,
primals_3, buf2, buf3, buf4, 64, XBLOCK=64, num_warps=1,
num_stages=1)
del buf0
del buf2
del buf3
del primals_3
return buf4, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), buf1, buf4
class LinearAttentionLayerNew(nn.Module):
def __init__(self, input_dim):
super().__init__()
self.linear = nn.Linear(input_dim, 1)
def forward(self, input_0, input_1):
primals_2 = self.linear.weight
primals_3 = self.linear.bias
primals_1 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| gustavhartz/legal-contract-elements | LinearAttentionLayer | false | 6,771 | [
"MIT"
] | 1 | 7a1e1f0024f9d336c7166f51b4325acf03db86a2 | https://github.com/gustavhartz/legal-contract-elements/tree/7a1e1f0024f9d336c7166f51b4325acf03db86a2 | import torch
import torch.nn.functional as F
from torch import nn
class Model(nn.Module):
def __init__(self, input_dim):
super().__init__()
self.linear = nn.Linear(input_dim, 1)
def forward(self, question, question_mask):
qtn = question.view(-1, question.shape[-1])
attn_scores = self.linear(qtn)
attn_scores = attn_scores.view(question.shape[0], question.shape[1])
attn_scores = attn_scores.masked_fill(question_mask == 1, -float('inf')
)
alpha = F.softmax(attn_scores, dim=1)
return alpha
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [4]
|
DownsampleA | # 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_4/inductor_cache/rs/crszklwi4mjsrfrzdtdvnsclt3tdjqb2b2lc3coqw5z6rswsrxqf.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 = ([%avg_pool2d, %mul], 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': 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, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = (xindex // 4) % 8
x0 = xindex % 2
x1 = (xindex // 2) % 2
x3 = (xindex // 32)
x4 = 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 + ((2*x0) + (8*x1) + (16*x2) + (64*x3)), tmp4 & xmask, eviction_policy='evict_last', other=0.0)
tmp6 = 1.0
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_ptr0 + ((2*x0) + (8*x1) + (16*((-4) + x2)) + (64*x3)), tmp10 & xmask, eviction_policy='evict_last', other=0.0)
tmp14 = tmp13 * tmp6
tmp15 = 0.0
tmp16 = tmp14 * tmp15
tmp17 = tl.full(tmp16.shape, 0.0, tmp16.dtype)
tmp18 = tl.where(tmp10, tmp16, tmp17)
tmp19 = tl.where(tmp4, tmp9, tmp18)
tl.store(out_ptr0 + (x4), tmp19, 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, 8, 2, 2), (32, 4, 2, 1), torch.float32)
# Topologically Sorted Source Nodes: [cat], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(arg0_1, buf0, 128, grid=grid(128), 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 DownsampleA(nn.Module):
def __init__(self, nIn, nOut, stride):
super(DownsampleA, self).__init__()
assert stride == 2
self.avg = nn.AvgPool2d(kernel_size=1, stride=stride)
def forward(self, x):
x = self.avg(x)
return torch.cat((x, x.mul(0)), 1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'nIn': 4, 'nOut': 4, 'stride': 2}]
| 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 = 128
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex // 4 % 8
x0 = xindex % 2
x1 = xindex // 2 % 2
x3 = xindex // 32
x4 = xindex
tmp0 = x2
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 4, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (2 * x0 + 8 * x1 + 16 * x2 + 64 * x3), tmp4 &
xmask, eviction_policy='evict_last', other=0.0)
tmp6 = 1.0
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_ptr0 + (2 * x0 + 8 * x1 + 16 * (-4 + x2) + 64 * x3),
tmp10 & xmask, eviction_policy='evict_last', other=0.0)
tmp14 = tmp13 * tmp6
tmp15 = 0.0
tmp16 = tmp14 * tmp15
tmp17 = tl.full(tmp16.shape, 0.0, tmp16.dtype)
tmp18 = tl.where(tmp10, tmp16, tmp17)
tmp19 = tl.where(tmp4, tmp9, tmp18)
tl.store(out_ptr0 + x4, tmp19, 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, 8, 2, 2), (32, 4, 2, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(128)](arg0_1, buf0, 128, XBLOCK=128,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class DownsampleANew(nn.Module):
def __init__(self, nIn, nOut, stride):
super(DownsampleANew, self).__init__()
assert stride == 2
self.avg = nn.AvgPool2d(kernel_size=1, stride=stride)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hamedomidvar/associativeconv | DownsampleA | false | 6,772 | [
"MIT"
] | 1 | 9930915abd3625871354df676865fc44eb92abf3 | https://github.com/hamedomidvar/associativeconv/tree/9930915abd3625871354df676865fc44eb92abf3 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, nIn, nOut, stride):
super().__init__()
assert stride == 2
self.avg = nn.AvgPool2d(kernel_size=1, stride=stride)
def forward(self, x):
x = self.avg(x)
return torch.cat((x, x.mul(0)), 1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 2]
|
Reverse | # 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_4/inductor_cache/x4/cx4malpk6rykdkpiqfs2akjvd7bgxbjcexzrzsq7e6sqcyz2ww4v.py
# Topologically Sorted Source Nodes: [flip], Original ATen: [aten.flip]
# Source node to ATen node mapping:
# flip => rev
# Graph fragment:
# %rev : [num_users=1] = call_function[target=torch.ops.prims.rev.default](args = (%arg0_1, [1]), kwargs = {})
triton_poi_fused_flip_0 = async_compile.triton('triton_poi_fused_flip_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_flip_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_flip_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 + (48 + x0 + ((-16)*x1) + (64*x2)), 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: [flip], Original ATen: [aten.flip]
stream0 = get_raw_stream(0)
triton_poi_fused_flip_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
class Reverse(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, audio):
return torch.flip(audio, dims=[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
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_flip_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 + (48 + x0 + -16 * x1 + 64 * x2), 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_flip_0[grid(256)](arg0_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class ReverseNew(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| h0ngwen/torchaudio-augmentations | Reverse | false | 6,773 | [
"MIT"
] | 1 | d044f9d020e12032ab9280acf5f34a337e72d212 | https://github.com/h0ngwen/torchaudio-augmentations/tree/d044f9d020e12032ab9280acf5f34a337e72d212 | import torch
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, audio):
return torch.flip(audio, dims=[1])
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
PolarityInversion | # 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_4/inductor_cache/hk/chkwhavqfzukyktba7dvmj6ss52pfnxbzhqfb7gpkrye7sko7lrp.py
# Topologically Sorted Source Nodes: [audio], Original ATen: [aten.neg]
# Source node to ATen node mapping:
# audio => neg
# Graph fragment:
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%arg0_1,), kwargs = {})
triton_poi_fused_neg_0 = async_compile.triton('triton_poi_fused_neg_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_neg_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_neg_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
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: [audio], Original ATen: [aten.neg]
stream0 = get_raw_stream(0)
triton_poi_fused_neg_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
class PolarityInversion(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, audio):
audio = torch.neg(audio)
return audio
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
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_neg_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
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_neg_0[grid(256)](arg0_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class PolarityInversionNew(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| h0ngwen/torchaudio-augmentations | PolarityInversion | false | 6,774 | [
"MIT"
] | 1 | d044f9d020e12032ab9280acf5f34a337e72d212 | https://github.com/h0ngwen/torchaudio-augmentations/tree/d044f9d020e12032ab9280acf5f34a337e72d212 | import torch
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
def forward(self, audio):
audio = torch.neg(audio)
return audio
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
ToRGB | # 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_4/inductor_cache/wi/cwiyl3lwwtancorrifw77xt3aqb4lermdintht45zvkj3bg54nbl.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_3, 0.5), 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 = 0.5
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/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_4, 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')
# kernel path: runs/run_shard_4/inductor_cache/on/conl6eemb3vyjzkllydlouehrcxphkzifo5kmslz6fgiz6ixsw5h.py
# Topologically Sorted Source Nodes: [mul_2, weight], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul_2 => mul_2
# weight => mul_3
# Graph fragment:
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_5, 0.5), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_2, %view), 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=[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_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 = 48
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex % 12
x0 = xindex % 4
x2 = (xindex // 12)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (x3), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x4), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/go/cgoav6av4bzem4wmdmkiowlmjpeiubwc67bqu6es4aivwlfpxzhh.py
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.add]
# Source node to ATen node mapping:
# out_3 => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_3, %primals_6), kwargs = {})
triton_poi_fused_add_3 = async_compile.triton('triton_poi_fused_add_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_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_add_3(in_out_ptr0, in_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 3
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 = 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, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (1, 3, 4, 1, 1), (12, 4, 1, 1, 1))
assert_size_stride(primals_6, (1, 3, 1, 1), (3, 1, 1, 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_3, buf0, 16, grid=grid(16), stream=stream0)
del primals_3
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_4, buf1, 4, grid=grid(4), stream=stream0)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1, out], Original ATen: [aten.mul, aten.addmm]
extern_kernels.addmm(buf1, primals_2, reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf0
del buf1
buf3 = empty_strided_cuda((4, 3, 4, 1, 1), (12, 4, 1, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_2, weight], Original ATen: [aten.mul]
triton_poi_fused_mul_2.run(primals_5, buf2, buf3, 48, grid=grid(48), stream=stream0)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf3, (12, 4, 1, 1), (4, 1, 0, 0), 0), stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=4, bias=None)
assert_size_stride(buf4, (1, 12, 4, 4), (192, 16, 4, 1))
buf5 = reinterpret_tensor(buf4, (4, 3, 4, 4), (48, 16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.add]
triton_poi_fused_add_3.run(buf5, primals_6, 192, grid=grid(192), stream=stream0)
del primals_6
return (buf5, primals_2, primals_5, buf2, reinterpret_tensor(buf3, (12, 4, 1, 1), (4, 1, 1, 1), 0), reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 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
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, 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((1, 3, 4, 1, 1), (12, 4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((1, 3, 1, 1), (3, 1, 1, 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 math
import torch
import torch.utils.data
import torch
from torch import nn
import torch.nn.functional as F
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if len(k.shape) == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, minor, in_h, in_w = input.shape
kernel_h, kernel_w = kernel.shape
out = input.view(-1, minor, in_h, 1, in_w, 1)
out = F.pad(out, [0, up_x - 1, 0, 0, 0, up_y - 1, 0, 0])
out = out.view(-1, minor, in_h * up_y, in_w * up_x)
out = F.pad(out, [max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(
pad_y1, 0)])
out = out[:, :, max(-pad_y0, 0):out.shape[2] - max(-pad_y1, 0), max(-
pad_x0, 0):out.shape[3] - max(-pad_x1, 0)]
out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x +
pad_x0 + pad_x1])
w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
out = F.conv2d(out, w)
out = out.reshape(-1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h +
1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
return upfirdn2d_native(input, kernel, up, up, down, down, pad[0], pad[
1], pad[0], pad[1])
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Upsample(nn.Module):
def __init__(self, kernel, factor=2):
super().__init__()
self.factor = factor
kernel = make_kernel(kernel) * factor ** 2
self.register_buffer('kernel', kernel)
p = kernel.shape[0] - factor
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2
self.pad = pad0, pad1
def forward(self, input):
out = upfirdn2d(input, self.kernel, up=self.factor, down=1, pad=
self.pad)
return out
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, input):
out = upfirdn2d(input, self.kernel, pad=self.pad)
return out
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = math.sqrt(1) / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, input):
if self.activation:
out = F.linear(input, self.weight * self.scale)
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
out = F.linear(input, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class ModulatedConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=[1,
3, 3, 1]):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2 + 1
self.blur = Blur(blur_kernel, pad=(pad0, pad1), upsample_factor
=factor)
if downsample:
factor = 2
p = len(blur_kernel) - factor + (kernel_size - 1)
pad0 = (p + 1) // 2
pad1 = p // 2
self.blur = Blur(blur_kernel, pad=(pad0, pad1))
fan_in = in_channel * kernel_size ** 2
self.scale = math.sqrt(1) / math.sqrt(fan_in)
self.padding = kernel_size // 2
self.weight = nn.Parameter(torch.randn(1, out_channel, in_channel,
kernel_size, kernel_size))
if style_dim is not None and style_dim > 0:
self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)
self.demodulate = demodulate
def __repr__(self):
return (
f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, upsample={self.upsample}, downsample={self.downsample})'
)
def forward(self, input, style):
batch, in_channel, height, width = input.shape
if style is not None:
style = self.modulation(style).view(batch, 1, in_channel, 1, 1)
else:
style = torch.ones(batch, 1, in_channel, 1, 1)
weight = self.scale * self.weight * style
if self.demodulate:
demod = torch.rsqrt(weight.pow(2).sum([2, 3, 4]) + 1e-08)
weight = weight * demod.view(batch, self.out_channel, 1, 1, 1)
weight = weight.view(batch * self.out_channel, in_channel, self.
kernel_size, self.kernel_size)
if self.upsample:
input = input.view(1, batch * in_channel, height, width)
weight = weight.view(batch, self.out_channel, in_channel, self.
kernel_size, self.kernel_size)
weight = weight.transpose(1, 2).reshape(batch * in_channel,
self.out_channel, self.kernel_size, self.kernel_size)
out = F.conv_transpose2d(input, weight, padding=0, stride=2,
groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
out = self.blur(out)
elif self.downsample:
input = self.blur(input)
_, _, height, width = input.shape
input = input.view(1, batch * in_channel, height, width)
out = F.conv2d(input, weight, padding=0, stride=2, groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
else:
input = input.view(1, batch * in_channel, height, width)
out = F.conv2d(input, weight, padding=self.padding, groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
return out
class ToRGB(nn.Module):
def __init__(self, in_channel, style_dim, upsample=True, blur_kernel=[1,
3, 3, 1]):
super().__init__()
if upsample:
self.upsample = Upsample(blur_kernel)
self.conv = ModulatedConv2d(in_channel, 3, 1, style_dim, demodulate
=False)
self.bias = nn.Parameter(torch.zeros(1, 3, 1, 1))
def forward(self, input, style, skip=None):
out = self.conv(input, style)
out = out + self.bias
if skip is not None:
skip = self.upsample(skip)
out = out + skip
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'in_channel': 4, 'style_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 math
import torch.utils.data
import torch
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_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 = 0.5
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)
@triton.jit
def triton_poi_fused_mul_2(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 48
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex % 12
x0 = xindex % 4
x2 = xindex // 12
x4 = xindex
tmp0 = tl.load(in_ptr0 + x3, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp1 = 0.5
tmp2 = tmp0 * tmp1
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + x4, tmp4, xmask)
@triton.jit
def triton_poi_fused_add_3(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 3
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 = 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, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (1, 3, 4, 1, 1), (12, 4, 1, 1, 1))
assert_size_stride(primals_6, (1, 3, 1, 1), (3, 1, 1, 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_3, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_3
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
triton_poi_fused_mul_1[grid(4)](primals_4, buf1, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(buf1, primals_2, reinterpret_tensor(buf0, (4,
4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf0
del buf1
buf3 = empty_strided_cuda((4, 3, 4, 1, 1), (12, 4, 1, 1, 1), torch.
float32)
triton_poi_fused_mul_2[grid(48)](primals_5, buf2, buf3, 48, XBLOCK=
64, num_warps=1, num_stages=1)
buf4 = extern_kernels.convolution(reinterpret_tensor(primals_1, (1,
16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf3, (12, 4,
1, 1), (4, 1, 0, 0), 0), stride=(1, 1), padding=(0, 0),
dilation=(1, 1), transposed=False, output_padding=(0, 0),
groups=4, bias=None)
assert_size_stride(buf4, (1, 12, 4, 4), (192, 16, 4, 1))
buf5 = reinterpret_tensor(buf4, (4, 3, 4, 4), (48, 16, 4, 1), 0)
del buf4
triton_poi_fused_add_3[grid(192)](buf5, primals_6, 192, XBLOCK=256,
num_warps=4, num_stages=1)
del primals_6
return buf5, primals_2, primals_5, buf2, reinterpret_tensor(buf3, (12,
4, 1, 1), (4, 1, 1, 1), 0), reinterpret_tensor(primals_1, (1, 16, 4,
4), (256, 16, 4, 1), 0)
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if len(k.shape) == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, minor, in_h, in_w = input.shape
kernel_h, kernel_w = kernel.shape
out = input.view(-1, minor, in_h, 1, in_w, 1)
out = F.pad(out, [0, up_x - 1, 0, 0, 0, up_y - 1, 0, 0])
out = out.view(-1, minor, in_h * up_y, in_w * up_x)
out = F.pad(out, [max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(
pad_y1, 0)])
out = out[:, :, max(-pad_y0, 0):out.shape[2] - max(-pad_y1, 0), max(-
pad_x0, 0):out.shape[3] - max(-pad_x1, 0)]
out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x +
pad_x0 + pad_x1])
w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
out = F.conv2d(out, w)
out = out.reshape(-1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h +
1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
return upfirdn2d_native(input, kernel, up, up, down, down, pad[0], pad[
1], pad[0], pad[1])
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Upsample(nn.Module):
def __init__(self, kernel, factor=2):
super().__init__()
self.factor = factor
kernel = make_kernel(kernel) * factor ** 2
self.register_buffer('kernel', kernel)
p = kernel.shape[0] - factor
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2
self.pad = pad0, pad1
def forward(self, input):
out = upfirdn2d(input, self.kernel, up=self.factor, down=1, pad=
self.pad)
return out
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, input):
out = upfirdn2d(input, self.kernel, pad=self.pad)
return out
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = math.sqrt(1) / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, input):
if self.activation:
out = F.linear(input, self.weight * self.scale)
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
out = F.linear(input, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class ModulatedConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=[1,
3, 3, 1]):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2 + 1
self.blur = Blur(blur_kernel, pad=(pad0, pad1), upsample_factor
=factor)
if downsample:
factor = 2
p = len(blur_kernel) - factor + (kernel_size - 1)
pad0 = (p + 1) // 2
pad1 = p // 2
self.blur = Blur(blur_kernel, pad=(pad0, pad1))
fan_in = in_channel * kernel_size ** 2
self.scale = math.sqrt(1) / math.sqrt(fan_in)
self.padding = kernel_size // 2
self.weight = nn.Parameter(torch.randn(1, out_channel, in_channel,
kernel_size, kernel_size))
if style_dim is not None and style_dim > 0:
self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)
self.demodulate = demodulate
def __repr__(self):
return (
f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, upsample={self.upsample}, downsample={self.downsample})'
)
def forward(self, input, style):
batch, in_channel, height, width = input.shape
if style is not None:
style = self.modulation(style).view(batch, 1, in_channel, 1, 1)
else:
style = torch.ones(batch, 1, in_channel, 1, 1)
weight = self.scale * self.weight * style
if self.demodulate:
demod = torch.rsqrt(weight.pow(2).sum([2, 3, 4]) + 1e-08)
weight = weight * demod.view(batch, self.out_channel, 1, 1, 1)
weight = weight.view(batch * self.out_channel, in_channel, self.
kernel_size, self.kernel_size)
if self.upsample:
input = input.view(1, batch * in_channel, height, width)
weight = weight.view(batch, self.out_channel, in_channel, self.
kernel_size, self.kernel_size)
weight = weight.transpose(1, 2).reshape(batch * in_channel,
self.out_channel, self.kernel_size, self.kernel_size)
out = F.conv_transpose2d(input, weight, padding=0, stride=2,
groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
out = self.blur(out)
elif self.downsample:
input = self.blur(input)
_, _, height, width = input.shape
input = input.view(1, batch * in_channel, height, width)
out = F.conv2d(input, weight, padding=0, stride=2, groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
else:
input = input.view(1, batch * in_channel, height, width)
out = F.conv2d(input, weight, padding=self.padding, groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
return out
class ToRGBNew(nn.Module):
def __init__(self, in_channel, style_dim, upsample=True, blur_kernel=[1,
3, 3, 1]):
super().__init__()
if upsample:
self.upsample = Upsample(blur_kernel)
self.conv = ModulatedConv2d(in_channel, 3, 1, style_dim, demodulate
=False)
self.bias = nn.Parameter(torch.zeros(1, 3, 1, 1))
def forward(self, input_0, input_1):
primals_6 = self.bias
primals_5 = self.conv.weight
primals_2 = self.conv.modulation.weight
primals_4 = self.conv.modulation.bias
primals_1 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6])
return output[0]
| guyii54/Contrastive-I2I | ToRGB | false | 6,775 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import math
import torch
import torch.utils.data
import torch
from torch import nn
import torch.nn.functional as F
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if len(k.shape) == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, minor, in_h, in_w = input.shape
kernel_h, kernel_w = kernel.shape
out = input.view(-1, minor, in_h, 1, in_w, 1)
out = F.pad(out, [0, up_x - 1, 0, 0, 0, up_y - 1, 0, 0])
out = out.view(-1, minor, in_h * up_y, in_w * up_x)
out = F.pad(out, [max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(
pad_y1, 0)])
out = out[:, :, max(-pad_y0, 0):out.shape[2] - max(-pad_y1, 0), max(-
pad_x0, 0):out.shape[3] - max(-pad_x1, 0)]
out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x +
pad_x0 + pad_x1])
w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
out = F.conv2d(out, w)
out = out.reshape(-1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h +
1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
return upfirdn2d_native(input, kernel, up, up, down, down, pad[0], pad[
1], pad[0], pad[1])
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Upsample(nn.Module):
def __init__(self, kernel, factor=2):
super().__init__()
self.factor = factor
kernel = make_kernel(kernel) * factor ** 2
self.register_buffer('kernel', kernel)
p = kernel.shape[0] - factor
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2
self.pad = pad0, pad1
def forward(self, input):
out = upfirdn2d(input, self.kernel, up=self.factor, down=1, pad=
self.pad)
return out
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, input):
out = upfirdn2d(input, self.kernel, pad=self.pad)
return out
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = math.sqrt(1) / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, input):
if self.activation:
out = F.linear(input, self.weight * self.scale)
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
out = F.linear(input, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class ModulatedConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=[1,
3, 3, 1]):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
pad0 = (p + 1) // 2 + factor - 1
pad1 = p /
# ... truncated (>4000 chars) for memory efficiency |
ResidualMLP | # 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_4/inductor_cache/et/cetivjyg7wwtjye65v5bugkwgj3h3fpz4jwyoilyug24te5isjr4.py
# Topologically Sorted Source Nodes: [h], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# h => relu
# Graph fragment:
# %relu : [num_users=3] = call_function[target=torch.ops.aten.relu.default](args = (%view_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=[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_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 = 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)
tl.store(in_out_ptr0 + (x2), tmp4, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/os/cos2tl65abyu25ojd6pqrrbzdv7c4iv3spmuwxywbi5nn4czneqi.py
# Topologically Sorted Source Nodes: [relu_1, h_1], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# h_1 => add
# relu_1 => relu_1
# Graph fragment:
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_3,), 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 = {})
# %le_2 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_1 = async_compile.triton('triton_poi_fused_add_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=[4096],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*fp32', 2: '*fp32', 3: '*fp32', 4: '*i1', 5: '*i1', 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_relu_threshold_backward_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_relu_threshold_backward_1(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, out_ptr2, 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_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr1 + (x2), None)
tmp2 = tl.load(in_ptr2 + (x0), None, 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
tmp9 = tmp0 <= tmp7
tl.store(out_ptr0 + (x2), tmp6, None)
tl.store(out_ptr1 + (x2), tmp8, None)
tl.store(out_ptr2 + (x2), tmp9, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/lq/clqx5kc5xyrdeenypwimwq4lw6kxkdogo65qhezv536tsnduv5wo.py
# Topologically Sorted Source Nodes: [relu_2, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
# Source node to ATen node mapping:
# h_2 => add_1
# relu_2 => relu_2
# Graph fragment:
# %relu_2 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%view_5,), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %relu_2), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_2, 0), kwargs = {})
triton_poi_fused_add_relu_threshold_backward_2 = async_compile.triton('triton_poi_fused_add_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=[4096],
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_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_relu_threshold_backward_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0, out_ptr1, 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_ptr0 + (x2), None)
tmp1 = tl.load(in_ptr1 + (x2), None)
tmp2 = tl.load(in_ptr2 + (x0), None, 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 + (x2), tmp6, None)
tl.store(out_ptr1 + (x2), tmp8, 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, (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, (64, 64), (64, 1))
assert_size_stride(primals_7, (64, ), (1, ))
assert_size_stride(primals_8, (4, 64), (64, 1))
assert_size_stride(primals_9, (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
# Topologically Sorted Source Nodes: [h], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_2, 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 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.float32)
buf8 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool)
buf9 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu_1, h_1], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_1.run(buf1, buf2, primals_5, buf3, buf8, buf9, 4096, grid=grid(4096), stream=stream0)
del primals_5
buf4 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf3, (64, 64), (64, 1), 0), reinterpret_tensor(primals_6, (64, 64), (1, 64), 0), out=buf4)
buf5 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.float32)
buf7 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool)
# Topologically Sorted Source Nodes: [relu_2, h_2], Original ATen: [aten.relu, aten.add, aten.threshold_backward]
triton_poi_fused_add_relu_threshold_backward_2.run(buf3, buf4, primals_7, buf5, buf7, 4096, grid=grid(4096), stream=stream0)
del buf4
del primals_7
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf5, (64, 64), (64, 1), 0), reinterpret_tensor(primals_8, (64, 4), (1, 64), 0), alpha=1, beta=1, out=buf6)
del primals_9
return (reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0), 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), reinterpret_tensor(buf5, (64, 64), (64, 1), 0), primals_8, buf7, primals_6, buf8, primals_4, 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((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((64, 64), (64, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((64, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 64), (64, 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 ResidualMLP(nn.Module):
def __init__(self, input_dim, target_dim, hidden_dim=64):
super(ResidualMLP, self).__init__()
self.linear1 = nn.Linear(input_dim, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, hidden_dim)
self.linear4 = nn.Linear(hidden_dim, target_dim)
def forward(self, h):
h = self.linear1(h).relu()
h = h + self.linear2(h).relu()
h = h + self.linear3(h).relu()
return self.linear4(h)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_dim': 4, 'target_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
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_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 % 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_add_relu_threshold_backward_1(in_ptr0, in_ptr1,
in_ptr2, out_ptr0, out_ptr1, out_ptr2, 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_ptr0 + x2, None)
tmp1 = tl.load(in_ptr1 + x2, None)
tmp2 = tl.load(in_ptr2 + x0, None, 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
tmp9 = tmp0 <= tmp7
tl.store(out_ptr0 + x2, tmp6, None)
tl.store(out_ptr1 + x2, tmp8, None)
tl.store(out_ptr2 + x2, tmp9, None)
@triton.jit
def triton_poi_fused_add_relu_threshold_backward_2(in_ptr0, in_ptr1,
in_ptr2, 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)
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + x2, None)
tmp1 = tl.load(in_ptr1 + x2, None)
tmp2 = tl.load(in_ptr2 + x0, None, 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 + x2, tmp6, None)
tl.store(out_ptr1 + x2, tmp8, 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, (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, (64, 64), (64, 1))
assert_size_stride(primals_7, (64,), (1,))
assert_size_stride(primals_8, (4, 64), (64, 1))
assert_size_stride(primals_9, (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
get_raw_stream(0)
triton_poi_fused_relu_0[grid(4096)](buf1, primals_2, 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 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.
float32)
buf8 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool
)
buf9 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool
)
triton_poi_fused_add_relu_threshold_backward_1[grid(4096)](buf1,
buf2, primals_5, buf3, buf8, buf9, 4096, XBLOCK=128, num_warps=
4, num_stages=1)
del primals_5
buf4 = buf2
del buf2
extern_kernels.mm(reinterpret_tensor(buf3, (64, 64), (64, 1), 0),
reinterpret_tensor(primals_6, (64, 64), (1, 64), 0), out=buf4)
buf5 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.
float32)
buf7 = empty_strided_cuda((4, 4, 4, 64), (1024, 256, 64, 1), torch.bool
)
triton_poi_fused_add_relu_threshold_backward_2[grid(4096)](buf3,
buf4, primals_7, buf5, buf7, 4096, XBLOCK=256, num_warps=4,
num_stages=1)
del buf4
del primals_7
buf6 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_9, reinterpret_tensor(buf5, (64, 64),
(64, 1), 0), reinterpret_tensor(primals_8, (64, 4), (1, 64), 0),
alpha=1, beta=1, out=buf6)
del primals_9
return reinterpret_tensor(buf6, (4, 4, 4, 4), (64, 16, 4, 1), 0
), 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), reinterpret_tensor(buf5, (64, 64), (64,
1), 0), primals_8, buf7, primals_6, buf8, primals_4, buf9
class ResidualMLPNew(nn.Module):
def __init__(self, input_dim, target_dim, hidden_dim=64):
super(ResidualMLPNew, self).__init__()
self.linear1 = nn.Linear(input_dim, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, hidden_dim)
self.linear4 = nn.Linear(hidden_dim, target_dim)
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.linear3.weight
primals_7 = self.linear3.bias
primals_8 = self.linear4.weight
primals_9 = self.linear4.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]
| greydanus/piecewise_node | ResidualMLP | false | 6,776 | [
"Apache-2.0"
] | 1 | 9d218d4ec1bab486ae954ad2e84732a5f952770f | https://github.com/greydanus/piecewise_node/tree/9d218d4ec1bab486ae954ad2e84732a5f952770f | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, input_dim, target_dim, hidden_dim=64):
super().__init__()
self.linear1 = nn.Linear(input_dim, hidden_dim)
self.linear2 = nn.Linear(hidden_dim, hidden_dim)
self.linear3 = nn.Linear(hidden_dim, hidden_dim)
self.linear4 = nn.Linear(hidden_dim, target_dim)
def forward(self, h):
h = self.linear1(h).relu()
h = h + self.linear2(h).relu()
h = h + self.linear3(h).relu()
return self.linear4(h)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ModulatedConv2d | # 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_4/inductor_cache/wi/cwiyl3lwwtancorrifw77xt3aqb4lermdintht45zvkj3bg54nbl.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_3, 0.5), 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 = 0.5
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/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_4, 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')
# kernel path: runs/run_shard_4/inductor_cache/ri/criuvsdl3sferb4bb6ci5zaps3wys7xxcpybz7vfo2ba4q7cuq6c.py
# Topologically Sorted Source Nodes: [mul_2, weight, pow_1, sum_1, add, demod, weight_1], Original ATen: [aten.mul, aten.pow, aten.sum, aten.add, aten.rsqrt]
# Source node to ATen node mapping:
# add => add
# demod => rsqrt
# mul_2 => mul_2
# pow_1 => pow_1
# sum_1 => sum_1
# weight => mul_3
# weight_1 => mul_4
# Graph fragment:
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_5, 0.125), kwargs = {})
# %mul_3 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_2, %view), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%mul_3, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [2, 3, 4]), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, 1e-08), kwargs = {})
# %rsqrt : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add,), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul_3, %view_1), kwargs = {})
triton_per_fused_add_mul_pow_rsqrt_sum_2 = async_compile.triton('triton_per_fused_add_mul_pow_rsqrt_sum_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=[16, 64],
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_add_mul_pow_rsqrt_sum_2', '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_add_mul_pow_rsqrt_sum_2(in_out_ptr0, in_ptr0, in_ptr1, out_ptr0, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 16
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)
r5 = rindex
x0 = xindex % 4
r3 = (rindex // 16)
x1 = (xindex // 4)
x4 = xindex
tmp0 = tl.load(in_ptr0 + (r5 + (64*x0)), xmask, eviction_policy='evict_last', other=0.0)
tmp3 = tl.load(in_ptr1 + (r3 + (4*x1)), xmask, eviction_policy='evict_last', other=0.0)
tmp1 = 0.125
tmp2 = tmp0 * tmp1
tmp4 = tmp2 * tmp3
tmp5 = tmp4 * tmp4
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = 1e-08
tmp11 = tmp9 + tmp10
tmp12 = libdevice.rsqrt(tmp11)
tmp13 = tmp4 * tmp12
tl.debug_barrier()
tl.store(in_out_ptr0 + (x4), tmp12, xmask)
tl.store(out_ptr0 + (r5 + (64*x4)), 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 = 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, 4), (4, 1))
assert_size_stride(primals_4, (4, ), (1, ))
assert_size_stride(primals_5, (1, 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, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_3, buf0, 16, grid=grid(16), stream=stream0)
del primals_3
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_4, buf1, 4, grid=grid(4), stream=stream0)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_1, out], Original ATen: [aten.mul, aten.addmm]
extern_kernels.addmm(buf1, primals_2, reinterpret_tensor(buf0, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf1
buf3 = buf0; del buf0 # reuse
buf4 = buf3; del buf3 # reuse
buf5 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul_2, weight, pow_1, sum_1, add, demod, weight_1], Original ATen: [aten.mul, aten.pow, aten.sum, aten.add, aten.rsqrt]
triton_per_fused_add_mul_pow_rsqrt_sum_2.run(buf4, primals_5, buf2, buf5, 16, 64, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4, 4, 4), (64, 16, 4, 1), 0), stride=(1, 1), padding=(2, 2), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=4, bias=None)
assert_size_stride(buf6, (1, 16, 5, 5), (400, 25, 5, 1))
return (reinterpret_tensor(buf6, (4, 4, 5, 5), (100, 25, 5, 1), 0), primals_2, primals_5, buf2, buf4, reinterpret_tensor(buf5, (16, 4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_1, (1, 16, 4, 4), (256, 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
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, 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((1, 4, 4, 4, 4), (256, 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.utils.data
import torch
from torch import nn
import torch.nn.functional as F
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if len(k.shape) == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, minor, in_h, in_w = input.shape
kernel_h, kernel_w = kernel.shape
out = input.view(-1, minor, in_h, 1, in_w, 1)
out = F.pad(out, [0, up_x - 1, 0, 0, 0, up_y - 1, 0, 0])
out = out.view(-1, minor, in_h * up_y, in_w * up_x)
out = F.pad(out, [max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(
pad_y1, 0)])
out = out[:, :, max(-pad_y0, 0):out.shape[2] - max(-pad_y1, 0), max(-
pad_x0, 0):out.shape[3] - max(-pad_x1, 0)]
out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x +
pad_x0 + pad_x1])
w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
out = F.conv2d(out, w)
out = out.reshape(-1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h +
1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
return upfirdn2d_native(input, kernel, up, up, down, down, pad[0], pad[
1], pad[0], pad[1])
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, input):
out = upfirdn2d(input, self.kernel, pad=self.pad)
return out
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = math.sqrt(1) / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, input):
if self.activation:
out = F.linear(input, self.weight * self.scale)
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
out = F.linear(input, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class ModulatedConv2d(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=[1,
3, 3, 1]):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2 + 1
self.blur = Blur(blur_kernel, pad=(pad0, pad1), upsample_factor
=factor)
if downsample:
factor = 2
p = len(blur_kernel) - factor + (kernel_size - 1)
pad0 = (p + 1) // 2
pad1 = p // 2
self.blur = Blur(blur_kernel, pad=(pad0, pad1))
fan_in = in_channel * kernel_size ** 2
self.scale = math.sqrt(1) / math.sqrt(fan_in)
self.padding = kernel_size // 2
self.weight = nn.Parameter(torch.randn(1, out_channel, in_channel,
kernel_size, kernel_size))
if style_dim is not None and style_dim > 0:
self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)
self.demodulate = demodulate
def __repr__(self):
return (
f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, upsample={self.upsample}, downsample={self.downsample})'
)
def forward(self, input, style):
batch, in_channel, height, width = input.shape
if style is not None:
style = self.modulation(style).view(batch, 1, in_channel, 1, 1)
else:
style = torch.ones(batch, 1, in_channel, 1, 1)
weight = self.scale * self.weight * style
if self.demodulate:
demod = torch.rsqrt(weight.pow(2).sum([2, 3, 4]) + 1e-08)
weight = weight * demod.view(batch, self.out_channel, 1, 1, 1)
weight = weight.view(batch * self.out_channel, in_channel, self.
kernel_size, self.kernel_size)
if self.upsample:
input = input.view(1, batch * in_channel, height, width)
weight = weight.view(batch, self.out_channel, in_channel, self.
kernel_size, self.kernel_size)
weight = weight.transpose(1, 2).reshape(batch * in_channel,
self.out_channel, self.kernel_size, self.kernel_size)
out = F.conv_transpose2d(input, weight, padding=0, stride=2,
groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
out = self.blur(out)
elif self.downsample:
input = self.blur(input)
_, _, height, width = input.shape
input = input.view(1, batch * in_channel, height, width)
out = F.conv2d(input, weight, padding=0, stride=2, groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
else:
input = input.view(1, batch * in_channel, height, width)
out = F.conv2d(input, weight, padding=self.padding, groups=batch)
_, _, height, width = out.shape
out = out.view(batch, self.out_channel, height, width)
return out
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4])]
def get_init_inputs():
return [[], {'in_channel': 4, 'out_channel': 4, 'kernel_size': 4,
'style_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 math
import torch.utils.data
import torch
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_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 = 0.5
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)
@triton.jit
def triton_per_fused_add_mul_pow_rsqrt_sum_2(in_out_ptr0, in_ptr0, in_ptr1,
out_ptr0, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 16
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)
r5 = rindex
x0 = xindex % 4
r3 = rindex // 16
x1 = xindex // 4
x4 = xindex
tmp0 = tl.load(in_ptr0 + (r5 + 64 * x0), xmask, eviction_policy=
'evict_last', other=0.0)
tmp3 = tl.load(in_ptr1 + (r3 + 4 * x1), xmask, eviction_policy=
'evict_last', other=0.0)
tmp1 = 0.125
tmp2 = tmp0 * tmp1
tmp4 = tmp2 * tmp3
tmp5 = tmp4 * tmp4
tmp6 = tl.broadcast_to(tmp5, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = 1e-08
tmp11 = tmp9 + tmp10
tmp12 = libdevice.rsqrt(tmp11)
tmp13 = tmp4 * tmp12
tl.debug_barrier()
tl.store(in_out_ptr0 + x4, tmp12, xmask)
tl.store(out_ptr0 + (r5 + 64 * x4), tmp13, 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, 4), (4, 1))
assert_size_stride(primals_4, (4,), (1,))
assert_size_stride(primals_5, (1, 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, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(16)](primals_3, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_3
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
triton_poi_fused_mul_1[grid(4)](primals_4, buf1, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_4
buf2 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
extern_kernels.addmm(buf1, primals_2, reinterpret_tensor(buf0, (4,
4), (1, 4), 0), alpha=1, beta=1, out=buf2)
del buf1
buf3 = buf0
del buf0
buf4 = buf3
del buf3
buf5 = empty_strided_cuda((4, 4, 4, 4, 4), (256, 64, 16, 4, 1),
torch.float32)
triton_per_fused_add_mul_pow_rsqrt_sum_2[grid(16)](buf4, primals_5,
buf2, buf5, 16, 64, XBLOCK=8, num_warps=4, num_stages=1)
buf6 = extern_kernels.convolution(reinterpret_tensor(primals_1, (1,
16, 4, 4), (256, 16, 4, 1), 0), reinterpret_tensor(buf5, (16, 4,
4, 4), (64, 16, 4, 1), 0), stride=(1, 1), padding=(2, 2),
dilation=(1, 1), transposed=False, output_padding=(0, 0),
groups=4, bias=None)
assert_size_stride(buf6, (1, 16, 5, 5), (400, 25, 5, 1))
return reinterpret_tensor(buf6, (4, 4, 5, 5), (100, 25, 5, 1), 0
), primals_2, primals_5, buf2, buf4, reinterpret_tensor(buf5, (16,
4, 4, 4), (64, 16, 4, 1), 0), reinterpret_tensor(primals_1, (1, 16,
4, 4), (256, 16, 4, 1), 0)
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if len(k.shape) == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, minor, in_h, in_w = input.shape
kernel_h, kernel_w = kernel.shape
out = input.view(-1, minor, in_h, 1, in_w, 1)
out = F.pad(out, [0, up_x - 1, 0, 0, 0, up_y - 1, 0, 0])
out = out.view(-1, minor, in_h * up_y, in_w * up_x)
out = F.pad(out, [max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(
pad_y1, 0)])
out = out[:, :, max(-pad_y0, 0):out.shape[2] - max(-pad_y1, 0), max(-
pad_x0, 0):out.shape[3] - max(-pad_x1, 0)]
out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x +
pad_x0 + pad_x1])
w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
out = F.conv2d(out, w)
out = out.reshape(-1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h +
1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
return upfirdn2d_native(input, kernel, up, up, down, down, pad[0], pad[
1], pad[0], pad[1])
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, input):
out = upfirdn2d(input, self.kernel, pad=self.pad)
return out
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = math.sqrt(1) / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, input):
if self.activation:
out = F.linear(input, self.weight * self.scale)
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
out = F.linear(input, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class ModulatedConv2dNew(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=[1,
3, 3, 1]):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2 + 1
self.blur = Blur(blur_kernel, pad=(pad0, pad1), upsample_factor
=factor)
if downsample:
factor = 2
p = len(blur_kernel) - factor + (kernel_size - 1)
pad0 = (p + 1) // 2
pad1 = p // 2
self.blur = Blur(blur_kernel, pad=(pad0, pad1))
fan_in = in_channel * kernel_size ** 2
self.scale = math.sqrt(1) / math.sqrt(fan_in)
self.padding = kernel_size // 2
self.weight = nn.Parameter(torch.randn(1, out_channel, in_channel,
kernel_size, kernel_size))
if style_dim is not None and style_dim > 0:
self.modulation = EqualLinear(style_dim, in_channel, bias_init=1)
self.demodulate = demodulate
def __repr__(self):
return (
f'{self.__class__.__name__}({self.in_channel}, {self.out_channel}, {self.kernel_size}, upsample={self.upsample}, downsample={self.downsample})'
)
def forward(self, input_0, input_1):
primals_5 = self.weight
primals_2 = self.modulation.weight
primals_4 = self.modulation.bias
primals_1 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| guyii54/Contrastive-I2I | ModulatedConv2d | false | 6,777 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import math
import torch
import torch.utils.data
import torch
from torch import nn
import torch.nn.functional as F
def make_kernel(k):
k = torch.tensor(k, dtype=torch.float32)
if len(k.shape) == 1:
k = k[None, :] * k[:, None]
k /= k.sum()
return k
def upfirdn2d_native(input, kernel, up_x, up_y, down_x, down_y, pad_x0,
pad_x1, pad_y0, pad_y1):
_, minor, in_h, in_w = input.shape
kernel_h, kernel_w = kernel.shape
out = input.view(-1, minor, in_h, 1, in_w, 1)
out = F.pad(out, [0, up_x - 1, 0, 0, 0, up_y - 1, 0, 0])
out = out.view(-1, minor, in_h * up_y, in_w * up_x)
out = F.pad(out, [max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(
pad_y1, 0)])
out = out[:, :, max(-pad_y0, 0):out.shape[2] - max(-pad_y1, 0), max(-
pad_x0, 0):out.shape[3] - max(-pad_x1, 0)]
out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x +
pad_x0 + pad_x1])
w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
out = F.conv2d(out, w)
out = out.reshape(-1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h +
1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1)
return out[:, :, ::down_y, ::down_x]
def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
return upfirdn2d_native(input, kernel, up, up, down, down, pad[0], pad[
1], pad[0], pad[1])
def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5):
return F.leaky_relu(input + bias, negative_slope) * scale
class Blur(nn.Module):
def __init__(self, kernel, pad, upsample_factor=1):
super().__init__()
kernel = make_kernel(kernel)
if upsample_factor > 1:
kernel = kernel * upsample_factor ** 2
self.register_buffer('kernel', kernel)
self.pad = pad
def forward(self, input):
out = upfirdn2d(input, self.kernel, pad=self.pad)
return out
class EqualLinear(nn.Module):
def __init__(self, in_dim, out_dim, bias=True, bias_init=0, lr_mul=1,
activation=None):
super().__init__()
self.weight = nn.Parameter(torch.randn(out_dim, in_dim).div_(lr_mul))
if bias:
self.bias = nn.Parameter(torch.zeros(out_dim).fill_(bias_init))
else:
self.bias = None
self.activation = activation
self.scale = math.sqrt(1) / math.sqrt(in_dim) * lr_mul
self.lr_mul = lr_mul
def forward(self, input):
if self.activation:
out = F.linear(input, self.weight * self.scale)
out = fused_leaky_relu(out, self.bias * self.lr_mul)
else:
out = F.linear(input, self.weight * self.scale, bias=self.bias *
self.lr_mul)
return out
def __repr__(self):
return (
f'{self.__class__.__name__}({self.weight.shape[1]}, {self.weight.shape[0]})'
)
class Model(nn.Module):
def __init__(self, in_channel, out_channel, kernel_size, style_dim,
demodulate=True, upsample=False, downsample=False, blur_kernel=[1,
3, 3, 1]):
super().__init__()
self.eps = 1e-08
self.kernel_size = kernel_size
self.in_channel = in_channel
self.out_channel = out_channel
self.upsample = upsample
self.downsample = downsample
if upsample:
factor = 2
p = len(blur_kernel) - factor - (kernel_size - 1)
pad0 = (p + 1) // 2 + factor - 1
pad1 = p // 2 + 1
self.blur = Blur(blur_kernel, pad=(pad0, pad1), upsample_factor
=factor)
if downsample:
factor = 2
p = len(blur_kernel) - factor + (kernel_size - 1)
pad0 = (p + 1) // 2
pad1 = p // 2
self.blur = Blur(blur_kernel, pad=(pad0, pad1))
fan_in = in_channel * kernel_size ** 2
self.scale = math.sqrt(1) / math.sqrt(fan_in)
self.padding = kernel_size // 2
self.weight = nn.Parameter(
# ... truncated (>4000 chars) for memory efficiency |
Multiply | # 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_4/inductor_cache/e4/ce4h2z2vthjeez5cqvrjtnep5b7t7jnbbufxqxbrvyhke72ruy4y.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 = (%arg0_1, 4), 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 = 4.0
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, = 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], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_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
from abc import ABC
class BaseOperator(ABC):
"""
Abstract class defining the basic structure for operator implementations in Hummingbird.
"""
def __init__(self, regression=False, classification=False, transformer=
False, anomaly_detection=False, **kwargs):
"""
Args:
regression: Whether operator is a regression model.
classification: Whether the operator is a classification model.
transformer: Whether the operator is a feature transformer.
anomaly_detection: Whether the operator is an anomaly detection model.
kwargs: Other keyword arguments.
"""
super().__init__()
self.regression = regression
self.classification = classification
self.transformer = transformer
self.anomaly_detection = anomaly_detection
class Multiply(BaseOperator, torch.nn.Module):
"""
Module used to multiply features in a pipeline by a score.
"""
def __init__(self, score):
super(Multiply, self).__init__()
self.score = score
def forward(self, x):
return x * self.score
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'score': 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 abc import ABC
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 = 4.0
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + 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)
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)](arg0_1, buf0, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
class BaseOperator(ABC):
"""
Abstract class defining the basic structure for operator implementations in Hummingbird.
"""
def __init__(self, regression=False, classification=False, transformer=
False, anomaly_detection=False, **kwargs):
"""
Args:
regression: Whether operator is a regression model.
classification: Whether the operator is a classification model.
transformer: Whether the operator is a feature transformer.
anomaly_detection: Whether the operator is an anomaly detection model.
kwargs: Other keyword arguments.
"""
super().__init__()
self.regression = regression
self.classification = classification
self.transformer = transformer
self.anomaly_detection = anomaly_detection
class MultiplyNew(BaseOperator, torch.nn.Module):
"""
Module used to multiply features in a pipeline by a score.
"""
def __init__(self, score):
super(MultiplyNew, self).__init__()
self.score = score
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hannahaih/hummingbird | Multiply | false | 6,778 | [
"MIT"
] | 1 | b8ec670b3c90ec7e87d3ae4a2b268075bd5eae65 | https://github.com/hannahaih/hummingbird/tree/b8ec670b3c90ec7e87d3ae4a2b268075bd5eae65 | import torch
from abc import ABC
class BaseOperator(ABC):
"""
Abstract class defining the basic structure for operator implementations in Hummingbird.
"""
def __init__(self, regression=False, classification=False, transformer=
False, anomaly_detection=False, **kwargs):
"""
Args:
regression: Whether operator is a regression model.
classification: Whether the operator is a classification model.
transformer: Whether the operator is a feature transformer.
anomaly_detection: Whether the operator is an anomaly detection model.
kwargs: Other keyword arguments.
"""
super().__init__()
self.regression = regression
self.classification = classification
self.transformer = transformer
self.anomaly_detection = anomaly_detection
class Model(BaseOperator, torch.nn.Module):
"""
Module used to multiply features in a pipeline by a score.
"""
def __init__(self, score):
super().__init__()
self.score = score
def forward(self, x):
return x * self.score
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
PoolingF | # 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_4/inductor_cache/ku/ckuooqx6q27vkk7tqqmkpsydmkdv3enwjhqctsyozm4owufb7yb6.py
# Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.adaptive_max_pool2d]
# Source node to ATen node mapping:
# input_1 => adaptive_max_pool2d
# Graph fragment:
# %adaptive_max_pool2d : [num_users=1] = call_function[target=torch.ops.aten.adaptive_max_pool2d.default](args = (%arg0_1, [1, 1]), kwargs = {})
triton_poi_fused_adaptive_max_pool2d_0 = async_compile.triton('triton_poi_fused_adaptive_max_pool2d_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_adaptive_max_pool2d_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_adaptive_max_pool2d_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 + (16*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (16*x0)), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (2 + (16*x0)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (3 + (16*x0)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (4 + (16*x0)), xmask, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr0 + (5 + (16*x0)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (6 + (16*x0)), xmask, eviction_policy='evict_last')
tmp13 = tl.load(in_ptr0 + (7 + (16*x0)), xmask, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr0 + (8 + (16*x0)), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (9 + (16*x0)), xmask, eviction_policy='evict_last')
tmp19 = tl.load(in_ptr0 + (10 + (16*x0)), xmask, eviction_policy='evict_last')
tmp21 = tl.load(in_ptr0 + (11 + (16*x0)), xmask, eviction_policy='evict_last')
tmp23 = tl.load(in_ptr0 + (12 + (16*x0)), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr0 + (13 + (16*x0)), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr0 + (14 + (16*x0)), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr0 + (15 + (16*x0)), 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 + (x0), tmp30, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/oy/coyfug4lezxxlpdfds2pbopjbi4465nns5jojyud3gmda7742ba5.py
# Topologically Sorted Source Nodes: [pow_1, sum_1, norm, add, out], Original ATen: [aten.pow, aten.sum, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# norm => pow_2
# out => div
# pow_1 => pow_1
# sum_1 => sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%getitem, 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 = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%pow_2, 1e-07), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%getitem, %add), kwargs = {})
triton_poi_fused_add_div_pow_sum_1 = async_compile.triton('triton_poi_fused_add_div_pow_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=[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_add_div_pow_sum_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_add_div_pow_sum_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')
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-07
tmp14 = 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):
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, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [input_1], Original ATen: [aten.adaptive_max_pool2d]
stream0 = get_raw_stream(0)
triton_poi_fused_adaptive_max_pool2d_0.run(arg0_1, buf0, 16, grid=grid(16), stream=stream0)
del arg0_1
buf1 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [pow_1, sum_1, norm, add, out], Original ATen: [aten.pow, aten.sum, aten.add, aten.div]
triton_poi_fused_add_div_pow_sum_1.run(buf0, buf1, 16, grid=grid(16), stream=stream0)
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, 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.utils.data
import torch
from torch import nn
class Normalize(nn.Module):
def __init__(self, power=2):
super(Normalize, self).__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
class PoolingF(nn.Module):
def __init__(self):
super(PoolingF, self).__init__()
model = [nn.AdaptiveMaxPool2d(1)]
self.model = nn.Sequential(*model)
self.l2norm = Normalize(2)
def forward(self, x):
return self.l2norm(self.model(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
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
@triton.jit
def triton_poi_fused_adaptive_max_pool2d_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 + 16 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp3 = tl.load(in_ptr0 + (2 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp5 = tl.load(in_ptr0 + (3 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp7 = tl.load(in_ptr0 + (4 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp9 = tl.load(in_ptr0 + (5 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp11 = tl.load(in_ptr0 + (6 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp13 = tl.load(in_ptr0 + (7 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp15 = tl.load(in_ptr0 + (8 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp17 = tl.load(in_ptr0 + (9 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp19 = tl.load(in_ptr0 + (10 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp21 = tl.load(in_ptr0 + (11 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp23 = tl.load(in_ptr0 + (12 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp25 = tl.load(in_ptr0 + (13 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp27 = tl.load(in_ptr0 + (14 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp29 = tl.load(in_ptr0 + (15 + 16 * x0), 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 + x0, tmp30, xmask)
@triton.jit
def triton_poi_fused_add_div_pow_sum_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')
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-07
tmp14 = tmp12 + tmp13
tmp15 = tmp0 / tmp14
tl.store(out_ptr0 + x2, 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, 1, 1), (4, 1, 16, 16), torch.float32)
get_raw_stream(0)
triton_poi_fused_adaptive_max_pool2d_0[grid(16)](arg0_1, buf0, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del arg0_1
buf1 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 1, 1), torch.float32)
triton_poi_fused_add_div_pow_sum_1[grid(16)](buf0, buf1, 16, XBLOCK
=16, num_warps=1, num_stages=1)
del buf0
return buf1,
class Normalize(nn.Module):
def __init__(self, power=2):
super(Normalize, self).__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
class PoolingFNew(nn.Module):
def __init__(self):
super(PoolingFNew, self).__init__()
model = [nn.AdaptiveMaxPool2d(1)]
self.model = nn.Sequential(*model)
self.l2norm = Normalize(2)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| guyii54/Contrastive-I2I | PoolingF | false | 6,779 | [
"BSD-3-Clause"
] | 1 | e73daa0f9d3770c2280a304c39678d5b22440647 | https://github.com/guyii54/Contrastive-I2I/tree/e73daa0f9d3770c2280a304c39678d5b22440647 | import torch
import torch.utils.data
import torch
from torch import nn
class Normalize(nn.Module):
def __init__(self, power=2):
super().__init__()
self.power = power
def forward(self, x):
norm = x.pow(self.power).sum(1, keepdim=True).pow(1.0 / self.power)
out = x.div(norm + 1e-07)
return out
class Model(nn.Module):
def __init__(self):
super().__init__()
model = [nn.AdaptiveMaxPool2d(1)]
self.model = nn.Sequential(*model)
self.l2norm = Normalize(2)
def forward(self, x):
return self.l2norm(self.model(x))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
HierarchicalPolicy | # 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_4/inductor_cache/md/cmd3ewacyhu5w5hausgbjbmtnt5rr66cgczh4ibdypq7dz6p4v7g.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 : [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')
# kernel path: runs/run_shard_4/inductor_cache/hz/chz2sqsqk26mwhf2dxhgh44jfpu2er5yqjftwkzfav5ctqtx5e7f.py
# Topologically Sorted Source Nodes: [prob], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# prob => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_3, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %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
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_4/inductor_cache/3f/c3fx6bzkalkw7u7askqdnz4rzlcoyqiec4r434sjc5x3axxgkrmr.py
# Topologically Sorted Source Nodes: [prob], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# prob => 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
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, (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, ))
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
buf5 = empty_strided_cuda((4, 4, 4, 128), (2048, 512, 128, 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, buf5, 8192, grid=grid(8192), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [q], 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((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [prob], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.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: [prob], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf3
return (buf4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(buf1, (64, 128), (128, 1), 0), buf4, primals_4, 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((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)
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)
| from _paritybench_helpers import _mock_config
import torch
import torch.nn as nn
import torch.nn.functional as f
class HierarchicalPolicy(nn.Module):
def __init__(self, args):
super(HierarchicalPolicy, self).__init__()
self.fc_1 = nn.Linear(args.state_shape, 128)
self.fc_2 = nn.Linear(128, args.noise_dim)
def forward(self, state):
x = f.relu(self.fc_1(state))
q = self.fc_2(x)
prob = f.softmax(q, dim=-1)
return prob
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'args': _mock_config(state_shape=4, noise_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._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 % 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)
@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)
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, 4), (64, 16, 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_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
buf5 = 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, buf5, 8192, 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, 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((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[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_2[grid(256)](buf3, buf4, 256, XBLOCK=256,
num_warps=4, num_stages=1)
del buf3
return buf4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), reinterpret_tensor(buf1, (64, 128), (128, 1), 0
), buf4, primals_4, buf5
class HierarchicalPolicyNew(nn.Module):
def __init__(self, args):
super(HierarchicalPolicyNew, self).__init__()
self.fc_1 = nn.Linear(args.state_shape, 128)
self.fc_2 = nn.Linear(128, args.noise_dim)
def forward(self, input_0):
primals_1 = self.fc_1.weight
primals_2 = self.fc_1.bias
primals_4 = self.fc_2.weight
primals_5 = self.fc_2.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| hanhanAnderson/LSF-SAC | HierarchicalPolicy | false | 6,780 | [
"MIT"
] | 1 | 3e2daf0da23b0ea08e92948c87f7e583f3fb1ed9 | https://github.com/hanhanAnderson/LSF-SAC/tree/3e2daf0da23b0ea08e92948c87f7e583f3fb1ed9 | from _paritybench_helpers import _mock_config
import torch
import torch.nn as nn
import torch.nn.functional as f
class Model(nn.Module):
def __init__(self, args):
super().__init__()
self.fc_1 = nn.Linear(args.state_shape, 128)
self.fc_2 = nn.Linear(128, args.noise_dim)
def forward(self, state):
x = f.relu(self.fc_1(state))
q = self.fc_2(x)
prob = f.softmax(q, dim=-1)
return prob
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
NumericLabelEncoder | # 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_4/inductor_cache/ob/cobmpcdvzzwgkde2zq2sq22d4cwqk3fg7cfsx34h6d2vcadz4yty.py
# Topologically Sorted Source Nodes: [eq, int_1, argmax], Original ATen: [aten.eq, aten._to_copy, aten.argmax]
# Source node to ATen node mapping:
# argmax => argmax
# eq => eq
# int_1 => convert_element_type
# Graph fragment:
# %eq : [num_users=1] = call_function[target=torch.ops.aten.eq.Tensor](args = (%view, %arg1_1), kwargs = {})
# %convert_element_type : [num_users=1] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%eq, torch.int32), kwargs = {})
# %argmax : [num_users=1] = call_function[target=torch.ops.aten.argmax.default](args = (%convert_element_type, 1), kwargs = {})
triton_poi_fused__to_copy_argmax_eq_0 = async_compile.triton('triton_poi_fused__to_copy_argmax_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: '*i32', 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__to_copy_argmax_eq_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__to_copy_argmax_eq_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 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp6 = tl.load(in_ptr1 + (1))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp20 = tl.load(in_ptr1 + (2))
tmp21 = tl.broadcast_to(tmp20, [XBLOCK])
tmp33 = tl.load(in_ptr1 + (3))
tmp34 = tl.broadcast_to(tmp33, [XBLOCK])
tmp3 = tmp2.to(tl.float32)
tmp4 = tmp0 == tmp3
tmp5 = tmp4.to(tl.int32)
tmp8 = tmp7.to(tl.float32)
tmp9 = tmp0 == tmp8
tmp10 = tmp9.to(tl.int32)
tmp11 = tmp5 > tmp10
tmp12 = tmp5 == tmp10
tmp13 = tl.full([1], 0, tl.int64)
tmp14 = tl.full([1], 1, tl.int64)
tmp15 = tmp13 < tmp14
tmp16 = tmp12 & tmp15
tmp17 = tmp11 | tmp16
tmp18 = tl.where(tmp17, tmp5, tmp10)
tmp19 = tl.where(tmp17, tmp13, tmp14)
tmp22 = tmp21.to(tl.float32)
tmp23 = tmp0 == tmp22
tmp24 = tmp23.to(tl.int32)
tmp25 = tmp18 > tmp24
tmp26 = tmp18 == tmp24
tmp27 = tl.full([1], 2, tl.int64)
tmp28 = tmp19 < tmp27
tmp29 = tmp26 & tmp28
tmp30 = tmp25 | tmp29
tmp31 = tl.where(tmp30, tmp18, tmp24)
tmp32 = tl.where(tmp30, tmp19, tmp27)
tmp35 = tmp34.to(tl.float32)
tmp36 = tmp0 == tmp35
tmp37 = tmp36.to(tl.int32)
tmp38 = tmp31 > tmp37
tmp39 = tmp31 == tmp37
tmp40 = tl.full([1], 3, tl.int64)
tmp41 = tmp32 < tmp40
tmp42 = tmp39 & tmp41
tmp43 = tmp38 | tmp42
tmp44 = tl.where(tmp43, tmp31, tmp37)
tmp45 = tl.where(tmp43, tmp32, tmp40)
tl.store(out_ptr0 + (x0), tmp45, 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, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((256, ), (1, ), torch.int64)
# Topologically Sorted Source Nodes: [eq, int_1, argmax], Original ATen: [aten.eq, aten._to_copy, aten.argmax]
stream0 = get_raw_stream(0)
triton_poi_fused__to_copy_argmax_eq_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, ), (1, ), device='cuda:0', dtype=torch.int32)
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 abc import ABC
class BaseOperator(ABC):
"""
Abstract class defining the basic structure for operator implementations in Hummingbird.
"""
def __init__(self, regression=False, classification=False, transformer=
False, anomaly_detection=False, **kwargs):
"""
Args:
regression: Whether operator is a regression model.
classification: Whether the operator is a classification model.
transformer: Whether the operator is a feature transformer.
anomaly_detection: Whether the operator is an anomaly detection model.
kwargs: Other keyword arguments.
"""
super().__init__()
self.regression = regression
self.classification = classification
self.transformer = transformer
self.anomaly_detection = anomaly_detection
class NumericLabelEncoder(BaseOperator, torch.nn.Module):
def __init__(self, classes, device):
super(NumericLabelEncoder, self).__init__(transformer=True)
self.regression = False
self.check_tensor = torch.nn.Parameter(torch.IntTensor(classes),
requires_grad=False)
def forward(self, x):
x = x.view(-1, 1)
return torch.argmax(torch.eq(x, self.check_tensor).int(), dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'classes': 4, 'device': 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 abc import ABC
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_argmax_eq_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 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp6 = tl.load(in_ptr1 + 1)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp20 = tl.load(in_ptr1 + 2)
tmp21 = tl.broadcast_to(tmp20, [XBLOCK])
tmp33 = tl.load(in_ptr1 + 3)
tmp34 = tl.broadcast_to(tmp33, [XBLOCK])
tmp3 = tmp2.to(tl.float32)
tmp4 = tmp0 == tmp3
tmp5 = tmp4.to(tl.int32)
tmp8 = tmp7.to(tl.float32)
tmp9 = tmp0 == tmp8
tmp10 = tmp9.to(tl.int32)
tmp11 = tmp5 > tmp10
tmp12 = tmp5 == tmp10
tmp13 = tl.full([1], 0, tl.int64)
tmp14 = tl.full([1], 1, tl.int64)
tmp15 = tmp13 < tmp14
tmp16 = tmp12 & tmp15
tmp17 = tmp11 | tmp16
tmp18 = tl.where(tmp17, tmp5, tmp10)
tmp19 = tl.where(tmp17, tmp13, tmp14)
tmp22 = tmp21.to(tl.float32)
tmp23 = tmp0 == tmp22
tmp24 = tmp23.to(tl.int32)
tmp25 = tmp18 > tmp24
tmp26 = tmp18 == tmp24
tmp27 = tl.full([1], 2, tl.int64)
tmp28 = tmp19 < tmp27
tmp29 = tmp26 & tmp28
tmp30 = tmp25 | tmp29
tmp31 = tl.where(tmp30, tmp18, tmp24)
tmp32 = tl.where(tmp30, tmp19, tmp27)
tmp35 = tmp34.to(tl.float32)
tmp36 = tmp0 == tmp35
tmp37 = tmp36.to(tl.int32)
tmp38 = tmp31 > tmp37
tmp39 = tmp31 == tmp37
tmp40 = tl.full([1], 3, tl.int64)
tmp41 = tmp32 < tmp40
tmp42 = tmp39 & tmp41
tmp43 = tmp38 | tmp42
tl.where(tmp43, tmp31, tmp37)
tmp45 = tl.where(tmp43, tmp32, tmp40)
tl.store(out_ptr0 + x0, tmp45, 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,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((256,), (1,), torch.int64)
get_raw_stream(0)
triton_poi_fused__to_copy_argmax_eq_0[grid(256)](arg0_1, arg1_1,
buf0, 256, XBLOCK=256, num_warps=4, num_stages=1)
del arg0_1
del arg1_1
return buf0,
class BaseOperator(ABC):
"""
Abstract class defining the basic structure for operator implementations in Hummingbird.
"""
def __init__(self, regression=False, classification=False, transformer=
False, anomaly_detection=False, **kwargs):
"""
Args:
regression: Whether operator is a regression model.
classification: Whether the operator is a classification model.
transformer: Whether the operator is a feature transformer.
anomaly_detection: Whether the operator is an anomaly detection model.
kwargs: Other keyword arguments.
"""
super().__init__()
self.regression = regression
self.classification = classification
self.transformer = transformer
self.anomaly_detection = anomaly_detection
class NumericLabelEncoderNew(BaseOperator, torch.nn.Module):
def __init__(self, classes, device):
super(NumericLabelEncoderNew, self).__init__(transformer=True)
self.regression = False
self.check_tensor = torch.nn.Parameter(torch.IntTensor(classes),
requires_grad=False)
def forward(self, input_0):
arg1_1 = self.check_tensor
arg0_1 = input_0
output = call([arg0_1, arg1_1])
return output[0]
| hannahaih/hummingbird | NumericLabelEncoder | false | 6,781 | [
"MIT"
] | 1 | b8ec670b3c90ec7e87d3ae4a2b268075bd5eae65 | https://github.com/hannahaih/hummingbird/tree/b8ec670b3c90ec7e87d3ae4a2b268075bd5eae65 | import torch
from abc import ABC
class BaseOperator(ABC):
"""
Abstract class defining the basic structure for operator implementations in Hummingbird.
"""
def __init__(self, regression=False, classification=False, transformer=
False, anomaly_detection=False, **kwargs):
"""
Args:
regression: Whether operator is a regression model.
classification: Whether the operator is a classification model.
transformer: Whether the operator is a feature transformer.
anomaly_detection: Whether the operator is an anomaly detection model.
kwargs: Other keyword arguments.
"""
super().__init__()
self.regression = regression
self.classification = classification
self.transformer = transformer
self.anomaly_detection = anomaly_detection
class Model(BaseOperator, torch.nn.Module):
def __init__(self, classes, device):
super().__init__(transformer=True)
self.regression = False
self.check_tensor = torch.nn.Parameter(torch.IntTensor(classes),
requires_grad=False)
def forward(self, x):
x = x.view(-1, 1)
return torch.argmax(torch.eq(x, self.check_tensor).int(), dim=1)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 0]
|
PositionalEncoding | # 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_4/inductor_cache/i5/ci5qztmjm5trfof5nnr4q6zfdgt5kz3i6kpes2s3qrkgziadh2ag.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.add]
# Source node to ATen node mapping:
# x => add
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%primals_1, %unsqueeze), 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
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):
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, (5, 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: [x], Original ATen: [aten.add]
stream0 = get_raw_stream(0)
triton_poi_fused_add_0.run(primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
del primals_1
del primals_2
return (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((5, 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 torch
from torch import nn
class PositionalEncoding(nn.Module):
def __init__(self, patch_num, d_model, dropout=0.1):
super(PositionalEncoding, self).__init__()
self.pe = nn.Parameter(torch.rand(patch_num + 1, d_model))
self.add_positional_encoding = lambda x: x + self.pe[:x.size(1)
].unsqueeze(0)
self.dropout = nn.Dropout(p=dropout)
def forward(self, x):
x = self.add_positional_encoding(x)
return self.dropout(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'patch_num': 4, 'd_model': 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 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_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
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):
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, (5, 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_add_0[grid(256)](primals_1, primals_2, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
del primals_2
return buf0,
class PositionalEncodingNew(nn.Module):
def __init__(self, patch_num, d_model, dropout=0.1):
super(PositionalEncodingNew, self).__init__()
self.pe = nn.Parameter(torch.rand(patch_num + 1, d_model))
self.add_positional_encoding = lambda x: x + self.pe[:x.size(1)
].unsqueeze(0)
self.dropout = nn.Dropout(p=dropout)
def forward(self, input_0):
primals_2 = self.pe
primals_1 = input_0
output = call([primals_1, primals_2])
return output[0]
| hankyul2/ImageClassification | PositionalEncoding | false | 6,782 | [
"Apache-2.0"
] | 1 | c4df6bf3dc1ee804f9885d586aa581ebb4d7ca05 | https://github.com/hankyul2/ImageClassification/tree/c4df6bf3dc1ee804f9885d586aa581ebb4d7ca05 | import torch
from torch import nn
class Model(nn.Module):
def __init__(self, patch_num, d_model, dropout=0.1):
super().__init__()
self.pe = nn.Parameter(torch.rand(patch_num + 1, d_model))
self.add_positional_encoding = lambda x: x + self.pe[:x.size(1)
].unsqueeze(0)
self.dropout = nn.Dropout(p=dropout)
def forward(self, x):
x = self.add_positional_encoding(x)
return self.dropout(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
StdConv | # 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_4/inductor_cache/ho/chomvcb3zp3abdoxfii63z2qmwawz6tyihk3k3kabbhwqodpe5ub.py
# Topologically Sorted Source Nodes: [mean, sub, std, add, truediv], Original ATen: [aten.mean, aten.sub, aten.std, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# mean => mean
# std => sqrt, var
# sub => sub
# truediv => div
# Graph fragment:
# %mean : [num_users=2] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [1, 2, 3], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %mean), kwargs = {})
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%primals_1, [1, 2, 3]), kwargs = {correction: 1.0, keepdim: True})
# %sqrt : [num_users=2] = call_function[target=torch.ops.aten.sqrt.default](args = (%var,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sqrt, 1e-06), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %add), kwargs = {})
triton_per_fused_add_div_mean_std_sub_0 = async_compile.triton('triton_per_fused_add_div_mean_std_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=[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_add_div_mean_std_sub_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], '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_add_div_mean_std_sub_0(in_out_ptr0, in_out_ptr1, in_ptr0, 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)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp6 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp1 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 64.0
tmp20 = tmp4 / tmp19
tmp21 = 63.0
tmp22 = tmp18 / tmp21
tmp23 = libdevice.sqrt(tmp22)
tmp24 = tmp0 - tmp20
tmp25 = 1e-06
tmp26 = tmp23 + tmp25
tmp27 = tmp24 / tmp26
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp20, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x0), tmp23, xmask)
tl.store(out_ptr0 + (r1 + (64*x0)), tmp27, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/tc/ctcagp37ljugm52zu6ckorigrppqo67voefe2f2odg5r6hyllhyu.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_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 = 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, 1), (1, 4, 4, 4), torch.float32)
buf3 = 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
buf5 = reinterpret_tensor(buf3, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf3 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mean, sub, std, add, truediv], Original ATen: [aten.mean, aten.sub, aten.std, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_per_fused_add_div_mean_std_sub_0.run(buf1, buf5, primals_1, buf6, 4, 64, grid=grid(4), stream=stream0)
# 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_1.run(buf8, primals_2, 16, grid=grid(16), stream=stream0)
del primals_2
return (buf8, primals_1, primals_3, buf1, 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
from torch import nn
class StdConv(nn.Conv2d):
def forward(self, x):
return self._conv_forward(x, self.standarize(self.weight), self.bias)
def standarize(self, x):
return (x - x.mean(dim=(1, 2, 3), keepdim=True)) / (x.std(dim=(1, 2,
3), keepdim=True) + 1e-06)
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
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_per_fused_add_div_mean_std_sub_0(in_out_ptr0, in_out_ptr1,
in_ptr0, 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)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp6 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp8 = tl.where(xmask, tmp6, 0)
tmp9 = tl.sum(tmp8, 1)[:, None]
tmp10 = tl.full([XBLOCK, 1], 64, tl.int32)
tmp11 = tmp10.to(tl.float32)
tmp12 = tmp9 / tmp11
tmp13 = tmp1 - tmp12
tmp14 = tmp13 * tmp13
tmp15 = tl.broadcast_to(tmp14, [XBLOCK, RBLOCK])
tmp17 = tl.where(xmask, tmp15, 0)
tmp18 = tl.sum(tmp17, 1)[:, None]
tmp19 = 64.0
tmp20 = tmp4 / tmp19
tmp21 = 63.0
tmp22 = tmp18 / tmp21
tmp23 = libdevice.sqrt(tmp22)
tmp24 = tmp0 - tmp20
tmp25 = 1e-06
tmp26 = tmp23 + tmp25
tmp27 = tmp24 / tmp26
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp20, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + x0, tmp23, xmask)
tl.store(out_ptr0 + (r1 + 64 * x0), tmp27, 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 = 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, 1), (1, 4, 4, 4), torch.float32)
buf3 = 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
buf5 = reinterpret_tensor(buf3, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf3
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused_add_div_mean_std_sub_0[grid(4)](buf1, buf5,
primals_1, buf6, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
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_1[grid(16)](buf8, primals_2, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_2
return buf8, primals_1, primals_3, buf1, buf5, buf6
class StdConvNew(nn.Conv2d):
def standarize(self, x):
return (x - x.mean(dim=(1, 2, 3), keepdim=True)) / (x.std(dim=(1, 2,
3), keepdim=True) + 1e-06)
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]
| hankyul2/ImageClassification | StdConv | false | 6,783 | [
"Apache-2.0"
] | 1 | c4df6bf3dc1ee804f9885d586aa581ebb4d7ca05 | https://github.com/hankyul2/ImageClassification/tree/c4df6bf3dc1ee804f9885d586aa581ebb4d7ca05 | import torch
from torch import nn
class Model(nn.Conv2d):
def forward(self, x):
return self._conv_forward(x, self.standarize(self.weight), self.bias)
def standarize(self, x):
return (x - x.mean(dim=(1, 2, 3), keepdim=True)) / (x.std(dim=(1, 2,
3), keepdim=True) + 1e-06)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
rec_attention | # 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_4/inductor_cache/2h/c2hhglkzlzg2xzmumfmt4aaa7rs7o47r76zo3zxuztunan3wht6k.py
# Topologically Sorted Source Nodes: [result_2], Original ATen: [aten.cat]
# Source node to ATen node mapping:
# result_2 => cat_2
# Graph fragment:
# %cat_2 : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%cat_1, %unsqueeze_3],), 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: '*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_cat_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_cat_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
x1 = (xindex // 4)
x0 = xindex % 4
x2 = xindex
tmp0 = x1
tmp1 = tl.full([1], 0, tl.int64)
tmp2 = tmp0 >= tmp1
tmp3 = tl.full([1], 3, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.full([1], 2, tl.int64)
tmp6 = tmp0 < tmp5
tmp7 = tmp6 & tmp4
tmp8 = tl.full([1], 1, tl.int64)
tmp9 = tmp0 < tmp8
tmp10 = tmp9 & tmp7
tmp11 = tl.load(in_ptr0 + (x0), tmp10 & xmask, eviction_policy='evict_last', other=0.0)
tmp12 = tmp0 >= tmp8
tmp13 = tmp12 & tmp7
tmp14 = tl.load(in_ptr1 + (x0), tmp13 & xmask, eviction_policy='evict_last', other=0.0)
tmp15 = tl.where(tmp9, tmp11, tmp14)
tmp16 = tl.full(tmp15.shape, 0.0, tmp15.dtype)
tmp17 = tl.where(tmp7, tmp15, tmp16)
tmp18 = tmp0 >= tmp5
tmp19 = tmp18 & tmp4
tmp20 = tl.load(in_ptr2 + (x0), tmp19 & xmask, eviction_policy='evict_last', other=0.0)
tmp21 = tl.where(tmp6, tmp17, tmp20)
tmp22 = tl.full(tmp21.shape, 0.0, tmp21.dtype)
tmp23 = tl.where(tmp4, tmp21, tmp22)
tmp24 = tmp0 >= tmp3
tmp25 = tl.full([1], 4, tl.int64)
tmp26 = tmp0 < tmp25
tmp27 = tl.load(in_ptr3 + (x0), tmp24 & xmask, eviction_policy='evict_last', other=0.0)
tmp28 = tl.where(tmp4, tmp23, tmp27)
tl.store(out_ptr0 + (x2), tmp28, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ts/ctscnzvbagjv4t25zui245b3recij5udu7nvujnr5rixcyo7elc6.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%squeeze, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%squeeze, %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=[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 = 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_4/inductor_cache/k6/ck6fz3qsfeqgn5jtm4ugikmu7cwvvlq3jpttijbb5kdniicwtyz6.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => 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=[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_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 = 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')
async_compile.wait(globals())
del async_compile
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 1), (1, 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((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [op], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 0), primals_1, out=buf0)
buf1 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [op_2], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 16), primals_1, out=buf1)
buf2 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [op_4], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 32), primals_1, out=buf2)
buf3 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [op_6], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 48), primals_1, out=buf3)
del primals_1
buf4 = empty_strided_cuda((4, 4, 1), (4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [result_2], Original ATen: [aten.cat]
stream0 = get_raw_stream(0)
triton_poi_fused_cat_0.run(buf0, buf1, buf2, buf3, buf4, 16, grid=grid(16), stream=stream0)
del buf0
del buf1
del buf2
del buf3
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf4, buf5, 16, grid=grid(16), stream=stream0)
buf6 = reinterpret_tensor(buf4, (4, 4), (4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf5, buf6, 16, grid=grid(16), stream=stream0)
buf7 = reinterpret_tensor(buf5, (4, 1, 4), (4, 4, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [repres], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf6, (4, 1, 4), (4, 4, 1), 0), primals_2, out=buf7)
return (buf7, buf6, buf6, reinterpret_tensor(primals_2, (4, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 48), reinterpret_tensor(primals_2, (4, 4), (1, 4), 32), reinterpret_tensor(primals_2, (4, 4), (1, 4), 16), reinterpret_tensor(primals_2, (4, 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, 1), (1, 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)
| from _paritybench_helpers import _mock_config
import torch
import torch.utils.data
import torch.nn as nn
def batch_product(iput, mat2):
result = None
for i in range(iput.size()[0]):
op = torch.mm(iput[i], mat2)
op = op.unsqueeze(0)
if result is None:
result = op
else:
result = torch.cat((result, op), 0)
return result.squeeze(2)
class rec_attention(nn.Module):
def __init__(self, hm, args):
super(rec_attention, self).__init__()
self.num_directions = 2 if args.bidirectional else 1
if hm is False:
self.bin_rep_size = args.bin_rnn_size * self.num_directions
else:
self.bin_rep_size = args.bin_rnn_size
self.bin_context_vector = nn.Parameter(torch.Tensor(self.
bin_rep_size, 1), requires_grad=True)
self.softmax = nn.Softmax(dim=1)
self.bin_context_vector.data.uniform_(-0.1, 0.1)
def forward(self, iput):
alpha = self.softmax(batch_product(iput, self.bin_context_vector))
[batch_size, source_length, _bin_rep_size2] = iput.size()
repres = torch.bmm(alpha.unsqueeze(2).view(batch_size, -1,
source_length), iput)
return repres, alpha
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'hm': 4, 'args': _mock_config(bidirectional=4,
bin_rnn_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.utils.data
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, 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
x1 = xindex // 4
x0 = xindex % 4
x2 = xindex
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 3, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.full([1], 2, tl.int64)
tmp6 = tmp0 < tmp5
tmp7 = tmp6 & tmp4
tmp8 = tl.full([1], 1, tl.int64)
tmp9 = tmp0 < tmp8
tmp10 = tmp9 & tmp7
tmp11 = tl.load(in_ptr0 + x0, tmp10 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp12 = tmp0 >= tmp8
tmp13 = tmp12 & tmp7
tmp14 = tl.load(in_ptr1 + x0, tmp13 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp15 = tl.where(tmp9, tmp11, tmp14)
tmp16 = tl.full(tmp15.shape, 0.0, tmp15.dtype)
tmp17 = tl.where(tmp7, tmp15, tmp16)
tmp18 = tmp0 >= tmp5
tmp19 = tmp18 & tmp4
tmp20 = tl.load(in_ptr2 + x0, tmp19 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp21 = tl.where(tmp6, tmp17, tmp20)
tmp22 = tl.full(tmp21.shape, 0.0, tmp21.dtype)
tmp23 = tl.where(tmp4, tmp21, tmp22)
tmp24 = tmp0 >= tmp3
tl.full([1], 4, tl.int64)
tmp27 = tl.load(in_ptr3 + x0, tmp24 & xmask, eviction_policy=
'evict_last', other=0.0)
tmp28 = tl.where(tmp4, tmp23, tmp27)
tl.store(out_ptr0 + x2, tmp28, 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 = 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 = 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)
def call(args):
primals_1, primals_2 = args
args.clear()
assert_size_stride(primals_1, (4, 1), (1, 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((4, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 0),
primals_1, out=buf0)
buf1 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 16),
primals_1, out=buf1)
buf2 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 32),
primals_1, out=buf2)
buf3 = empty_strided_cuda((4, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (4, 4), (4, 1), 48),
primals_1, out=buf3)
del primals_1
buf4 = empty_strided_cuda((4, 4, 1), (4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_cat_0[grid(16)](buf0, buf1, buf2, buf3, buf4, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del buf0
del buf1
del buf2
del buf3
buf5 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(16)](buf4, buf5, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf6 = reinterpret_tensor(buf4, (4, 4), (4, 1), 0)
del buf4
triton_poi_fused__softmax_2[grid(16)](buf5, buf6, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf7 = reinterpret_tensor(buf5, (4, 1, 4), (4, 4, 1), 0)
del buf5
extern_kernels.bmm(reinterpret_tensor(buf6, (4, 1, 4), (4, 4, 1), 0
), primals_2, out=buf7)
return buf7, buf6, buf6, reinterpret_tensor(primals_2, (4, 4, 4), (16,
1, 4), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 48
), reinterpret_tensor(primals_2, (4, 4), (1, 4), 32
), reinterpret_tensor(primals_2, (4, 4), (1, 4), 16
), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0)
def batch_product(iput, mat2):
result = None
for i in range(iput.size()[0]):
op = torch.mm(iput[i], mat2)
op = op.unsqueeze(0)
if result is None:
result = op
else:
result = torch.cat((result, op), 0)
return result.squeeze(2)
class rec_attentionNew(nn.Module):
def __init__(self, hm, args):
super(rec_attentionNew, self).__init__()
self.num_directions = 2 if args.bidirectional else 1
if hm is False:
self.bin_rep_size = args.bin_rnn_size * self.num_directions
else:
self.bin_rep_size = args.bin_rnn_size
self.bin_context_vector = nn.Parameter(torch.Tensor(self.
bin_rep_size, 1), requires_grad=True)
self.softmax = nn.Softmax(dim=1)
self.bin_context_vector.data.uniform_(-0.1, 0.1)
def forward(self, input_0):
primals_1 = self.bin_context_vector
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0], output[1]
| gzerveas/TransformChrome | rec_attention | false | 6,784 | [
"MIT"
] | 1 | ab1046009aff2ec863aa65223dcfcd750d41ab86 | https://github.com/gzerveas/TransformChrome/tree/ab1046009aff2ec863aa65223dcfcd750d41ab86 | from _paritybench_helpers import _mock_config
import torch
import torch.utils.data
import torch.nn as nn
def batch_product(iput, mat2):
result = None
for i in range(iput.size()[0]):
op = torch.mm(iput[i], mat2)
op = op.unsqueeze(0)
if result is None:
result = op
else:
result = torch.cat((result, op), 0)
return result.squeeze(2)
class Model(nn.Module):
def __init__(self, hm, args):
super().__init__()
self.num_directions = 2 if args.bidirectional else 1
if hm is False:
self.bin_rep_size = args.bin_rnn_size * self.num_directions
else:
self.bin_rep_size = args.bin_rnn_size
self.bin_context_vector = nn.Parameter(torch.Tensor(self.
bin_rep_size, 1), requires_grad=True)
self.softmax = nn.Softmax(dim=1)
self.bin_context_vector.data.uniform_(-0.1, 0.1)
def forward(self, iput):
alpha = self.softmax(batch_product(iput, self.bin_context_vector))
[batch_size, source_length, _bin_rep_size2] = iput.size()
repres = torch.bmm(alpha.unsqueeze(2).view(batch_size, -1,
source_length), iput)
return repres, alpha
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'hm': 4, 'args': _mock_config(bidirectional=4,
bin_rnn_size=4)}]
|
ConvertPointsToHomogeneous | # 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_4/inductor_cache/2o/c2ouwmbjvteb4nvr7vyvyz6jp657ga66uiadoicxwun5d3s4kog7.py
# Topologically Sorted Source Nodes: [pad], Original ATen: [aten.constant_pad_nd]
# Source node to ATen node mapping:
# pad => constant_pad_nd
# Graph fragment:
# %constant_pad_nd : [num_users=1] = call_function[target=torch.ops.aten.constant_pad_nd.default](args = (%arg0_1, [0, 1], 1.0), 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=[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_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 = 320
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 5
x1 = (xindex // 5)
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 4, tl.int64)
tmp2 = tmp0 < tmp1
tmp3 = tl.load(in_ptr0 + (x0 + (4*x1)), tmp2 & xmask, other=1.0)
tl.store(out_ptr0 + (x2), tmp3, 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, 5), (80, 20, 5, 1), torch.float32)
# Topologically Sorted Source Nodes: [pad], Original ATen: [aten.constant_pad_nd]
stream0 = get_raw_stream(0)
triton_poi_fused_constant_pad_nd_0.run(arg0_1, buf0, 320, grid=grid(320), 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
def convert_points_to_homogeneous(points):
"""Function that converts points from Euclidean to homogeneous space.
See :class:`~torchgeometry.ConvertPointsToHomogeneous` for details.
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> output = tgm.convert_points_to_homogeneous(input) # BxNx4
"""
if not torch.is_tensor(points):
raise TypeError('Input type is not a torch.Tensor. Got {}'.format(
type(points)))
if len(points.shape) < 2:
raise ValueError('Input must be at least a 2D tensor. Got {}'.
format(points.shape))
return nn.functional.pad(points, (0, 1), 'constant', 1.0)
class ConvertPointsToHomogeneous(nn.Module):
"""Creates a transformation to convert points from Euclidean to
homogeneous space.
Args:
points (Tensor): tensor of N-dimensional points.
Returns:
Tensor: tensor of N+1-dimensional points.
Shape:
- Input: :math:`(B, D, N)` or :math:`(D, N)`
- Output: :math:`(B, D, N + 1)` or :math:`(D, N + 1)`
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> transform = tgm.ConvertPointsToHomogeneous()
>>> output = transform(input) # BxNx4
"""
def __init__(self):
super(ConvertPointsToHomogeneous, self).__init__()
def forward(self, input):
return convert_points_to_homogeneous(input)
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_constant_pad_nd_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 320
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 5
x1 = xindex // 5
x2 = xindex
tmp0 = x0
tmp1 = tl.full([1], 4, tl.int64)
tmp2 = tmp0 < tmp1
tmp3 = tl.load(in_ptr0 + (x0 + 4 * x1), tmp2 & xmask, other=1.0)
tl.store(out_ptr0 + x2, tmp3, 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, 5), (80, 20, 5, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_constant_pad_nd_0[grid(320)](arg0_1, buf0, 320,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf0,
def convert_points_to_homogeneous(points):
"""Function that converts points from Euclidean to homogeneous space.
See :class:`~torchgeometry.ConvertPointsToHomogeneous` for details.
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> output = tgm.convert_points_to_homogeneous(input) # BxNx4
"""
if not torch.is_tensor(points):
raise TypeError('Input type is not a torch.Tensor. Got {}'.format(
type(points)))
if len(points.shape) < 2:
raise ValueError('Input must be at least a 2D tensor. Got {}'.
format(points.shape))
return nn.functional.pad(points, (0, 1), 'constant', 1.0)
class ConvertPointsToHomogeneousNew(nn.Module):
"""Creates a transformation to convert points from Euclidean to
homogeneous space.
Args:
points (Tensor): tensor of N-dimensional points.
Returns:
Tensor: tensor of N+1-dimensional points.
Shape:
- Input: :math:`(B, D, N)` or :math:`(D, N)`
- Output: :math:`(B, D, N + 1)` or :math:`(D, N + 1)`
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> transform = tgm.ConvertPointsToHomogeneous()
>>> output = transform(input) # BxNx4
"""
def __init__(self):
super(ConvertPointsToHomogeneousNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hassony2/frankmocap | ConvertPointsToHomogeneous | false | 6,785 | [
"BSD-3-Clause"
] | 1 | 50aae41d9b41d2f344ae1709bbf1b25974209fa9 | https://github.com/hassony2/frankmocap/tree/50aae41d9b41d2f344ae1709bbf1b25974209fa9 | import torch
import torch.nn as nn
def convert_points_to_homogeneous(points):
"""Function that converts points from Euclidean to homogeneous space.
See :class:`~torchgeometry.ConvertPointsToHomogeneous` for details.
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> output = tgm.convert_points_to_homogeneous(input) # BxNx4
"""
if not torch.is_tensor(points):
raise TypeError('Input type is not a torch.Tensor. Got {}'.format(
type(points)))
if len(points.shape) < 2:
raise ValueError('Input must be at least a 2D tensor. Got {}'.
format(points.shape))
return nn.functional.pad(points, (0, 1), 'constant', 1.0)
class Model(nn.Module):
"""Creates a transformation to convert points from Euclidean to
homogeneous space.
Args:
points (Tensor): tensor of N-dimensional points.
Returns:
Tensor: tensor of N+1-dimensional points.
Shape:
- Input: :math:`(B, D, N)` or :math:`(D, N)`
- Output: :math:`(B, D, N + 1)` or :math:`(D, N + 1)`
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> transform = tgm.ConvertPointsToHomogeneous()
>>> output = transform(input) # BxNx4
"""
def __init__(self):
super().__init__()
def forward(self, input):
return convert_points_to_homogeneous(input)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Hidden2Discrete | # 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_4/inductor_cache/xe/cxeq77gbpevhf6jov7fs3c25pvswzi43xn2bxfthg2nvsuurswra.py
# Topologically Sorted Source Nodes: [log_qy], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_qy => amax, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_2, [1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_2, %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=[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__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 = 1024
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_4/inductor_cache/nj/cnjj3kjcokm5rrbv6azeg2i2dkelsepqzurxngdhwjbc5vp6wfpj.py
# Topologically Sorted Source Nodes: [log_qy], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_qy => exp, log, sub_1, sum_1
# 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 = {})
triton_poi_fused__log_softmax_1 = async_compile.triton('triton_poi_fused__log_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=[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__log_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__log_softmax_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
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 = tl_math.exp(tmp1)
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp2 + tmp4
tmp7 = tl_math.exp(tmp6)
tmp8 = tmp5 + tmp7
tmp10 = tl_math.exp(tmp9)
tmp11 = tmp8 + tmp10
tmp12 = tl_math.log(tmp11)
tmp13 = tmp0 - 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 = 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))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
# Topologically Sorted Source Nodes: [logits], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 16), (1, 4), 0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((256, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_qy], Original ATen: [aten._log_softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax_0.run(buf0, buf1, 1024, grid=grid(1024), stream=stream0)
buf2 = empty_strided_cuda((256, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_qy], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_1.run(buf1, buf2, 1024, grid=grid(1024), stream=stream0)
del buf1
return (reinterpret_tensor(buf0, (256, 4), (4, 1), 0), buf2, reinterpret_tensor(primals_3, (64, 4), (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
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)
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
import torch.nn.init
class Hidden2Discrete(nn.Module):
def __init__(self, input_size, y_size, k_size, is_lstm=False, has_bias=True
):
super(Hidden2Discrete, self).__init__()
self.y_size = y_size
self.k_size = k_size
latent_size = self.k_size * self.y_size
if is_lstm:
self.p_h = nn.Linear(input_size, latent_size, bias=has_bias)
self.p_c = nn.Linear(input_size, latent_size, bias=has_bias)
else:
self.p_h = nn.Linear(input_size, latent_size, bias=has_bias)
self.is_lstm = is_lstm
def forward(self, inputs):
"""
:param inputs: batch_size x input_size
:return:
"""
if self.is_lstm:
h, c = inputs
if h.dim() == 3:
h = h.squeeze(0)
c = c.squeeze(0)
logits = self.p_h(h) + self.p_c(c)
else:
logits = self.p_h(inputs)
logits = logits.view(-1, self.k_size)
log_qy = F.log_softmax(logits, dim=1)
return logits, log_qy
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'input_size': 4, 'y_size': 4, 'k_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
import torch.nn.init
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__log_softmax_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
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_poi_fused__log_softmax_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
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 = tl_math.exp(tmp1)
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp2 + tmp4
tmp7 = tl_math.exp(tmp6)
tmp8 = tmp5 + tmp7
tmp10 = tl_math.exp(tmp9)
tmp11 = tmp8 + tmp10
tmp12 = tl_math.log(tmp11)
tmp13 = tmp0 - tmp12
tl.store(out_ptr0 + x2, tmp13, xmask)
def call(args):
primals_1, primals_2, primals_3 = 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))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((64, 16), (16, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 16), (1, 4),
0), alpha=1, beta=1, out=buf0)
del primals_1
del primals_2
buf1 = empty_strided_cuda((256, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax_0[grid(1024)](buf0, buf1, 1024,
XBLOCK=128, num_warps=4, num_stages=1)
buf2 = empty_strided_cuda((256, 4), (4, 1), torch.float32)
triton_poi_fused__log_softmax_1[grid(1024)](buf1, buf2, 1024,
XBLOCK=128, num_warps=4, num_stages=1)
del buf1
return reinterpret_tensor(buf0, (256, 4), (4, 1), 0
), buf2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), buf2
class Hidden2DiscreteNew(nn.Module):
def __init__(self, input_size, y_size, k_size, is_lstm=False, has_bias=True
):
super(Hidden2DiscreteNew, self).__init__()
self.y_size = y_size
self.k_size = k_size
latent_size = self.k_size * self.y_size
if is_lstm:
self.p_h = nn.Linear(input_size, latent_size, bias=has_bias)
self.p_c = nn.Linear(input_size, latent_size, bias=has_bias)
else:
self.p_h = nn.Linear(input_size, latent_size, bias=has_bias)
self.is_lstm = is_lstm
def forward(self, input_0):
primals_1 = self.p_h.weight
primals_2 = self.p_h.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0], output[1]
| haojiepan1/CrossWOZ | Hidden2Discrete | false | 6,786 | [
"Apache-2.0"
] | 1 | 6d7b4c4cfb73a528b76074764687906abecc90b6 | https://github.com/haojiepan1/CrossWOZ/tree/6d7b4c4cfb73a528b76074764687906abecc90b6 | import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.nn.init
class Model(nn.Module):
def __init__(self, input_size, y_size, k_size, is_lstm=False, has_bias=True
):
super().__init__()
self.y_size = y_size
self.k_size = k_size
latent_size = self.k_size * self.y_size
if is_lstm:
self.p_h = nn.Linear(input_size, latent_size, bias=has_bias)
self.p_c = nn.Linear(input_size, latent_size, bias=has_bias)
else:
self.p_h = nn.Linear(input_size, latent_size, bias=has_bias)
self.is_lstm = is_lstm
def forward(self, inputs):
"""
:param inputs: batch_size x input_size
:return:
"""
if self.is_lstm:
h, c = inputs
if h.dim() == 3:
h = h.squeeze(0)
c = c.squeeze(0)
logits = self.p_h(h) + self.p_c(c)
else:
logits = self.p_h(inputs)
logits = logits.view(-1, self.k_size)
log_qy = F.log_softmax(logits, dim=1)
return logits, log_qy
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
ConvertPointsFromHomogeneous | # 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_4/inductor_cache/m7/cm7u2xidzfzafcfgds4g3sogyf6hfbowvehkuqc2rzfo2k3uufot.py
# Topologically Sorted Source Nodes: [truediv], Original ATen: [aten.div]
# Source node to ATen node mapping:
# truediv => div
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%slice_1, %slice_2), 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': 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_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 3
x1 = (xindex // 3)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x1)), xmask)
tmp1 = tl.load(in_ptr0 + (3 + (4*x1)), 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, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 3), (48, 12, 3, 1), torch.float32)
# Topologically Sorted Source Nodes: [truediv], Original ATen: [aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_div_0.run(arg0_1, buf0, 192, grid=grid(192), 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
def convert_points_from_homogeneous(points):
"""Function that converts points from homogeneous to Euclidean space.
See :class:`~torchgeometry.ConvertPointsFromHomogeneous` for details.
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> output = tgm.convert_points_from_homogeneous(input) # BxNx2
"""
if not torch.is_tensor(points):
raise TypeError('Input type is not a torch.Tensor. Got {}'.format(
type(points)))
if len(points.shape) < 2:
raise ValueError('Input must be at least a 2D tensor. Got {}'.
format(points.shape))
return points[..., :-1] / points[..., -1:]
class ConvertPointsFromHomogeneous(nn.Module):
"""Creates a transformation that converts points from homogeneous to
Euclidean space.
Args:
points (Tensor): tensor of N-dimensional points.
Returns:
Tensor: tensor of N-1-dimensional points.
Shape:
- Input: :math:`(B, D, N)` or :math:`(D, N)`
- Output: :math:`(B, D, N + 1)` or :math:`(D, N + 1)`
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> transform = tgm.ConvertPointsFromHomogeneous()
>>> output = transform(input) # BxNx2
"""
def __init__(self):
super(ConvertPointsFromHomogeneous, self).__init__()
def forward(self, input):
return convert_points_from_homogeneous(input)
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_div_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 3
x1 = xindex // 3
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x1), xmask)
tmp1 = tl.load(in_ptr0 + (3 + 4 * x1), 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, 4), (64, 16, 4, 1))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 4, 4, 3), (48, 12, 3, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_div_0[grid(192)](arg0_1, buf0, 192, XBLOCK=256,
num_warps=4, num_stages=1)
del arg0_1
return buf0,
def convert_points_from_homogeneous(points):
"""Function that converts points from homogeneous to Euclidean space.
See :class:`~torchgeometry.ConvertPointsFromHomogeneous` for details.
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> output = tgm.convert_points_from_homogeneous(input) # BxNx2
"""
if not torch.is_tensor(points):
raise TypeError('Input type is not a torch.Tensor. Got {}'.format(
type(points)))
if len(points.shape) < 2:
raise ValueError('Input must be at least a 2D tensor. Got {}'.
format(points.shape))
return points[..., :-1] / points[..., -1:]
class ConvertPointsFromHomogeneousNew(nn.Module):
"""Creates a transformation that converts points from homogeneous to
Euclidean space.
Args:
points (Tensor): tensor of N-dimensional points.
Returns:
Tensor: tensor of N-1-dimensional points.
Shape:
- Input: :math:`(B, D, N)` or :math:`(D, N)`
- Output: :math:`(B, D, N + 1)` or :math:`(D, N + 1)`
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> transform = tgm.ConvertPointsFromHomogeneous()
>>> output = transform(input) # BxNx2
"""
def __init__(self):
super(ConvertPointsFromHomogeneousNew, self).__init__()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hassony2/frankmocap | ConvertPointsFromHomogeneous | false | 6,787 | [
"BSD-3-Clause"
] | 1 | 50aae41d9b41d2f344ae1709bbf1b25974209fa9 | https://github.com/hassony2/frankmocap/tree/50aae41d9b41d2f344ae1709bbf1b25974209fa9 | import torch
import torch.nn as nn
def convert_points_from_homogeneous(points):
"""Function that converts points from homogeneous to Euclidean space.
See :class:`~torchgeometry.ConvertPointsFromHomogeneous` for details.
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> output = tgm.convert_points_from_homogeneous(input) # BxNx2
"""
if not torch.is_tensor(points):
raise TypeError('Input type is not a torch.Tensor. Got {}'.format(
type(points)))
if len(points.shape) < 2:
raise ValueError('Input must be at least a 2D tensor. Got {}'.
format(points.shape))
return points[..., :-1] / points[..., -1:]
class Model(nn.Module):
"""Creates a transformation that converts points from homogeneous to
Euclidean space.
Args:
points (Tensor): tensor of N-dimensional points.
Returns:
Tensor: tensor of N-1-dimensional points.
Shape:
- Input: :math:`(B, D, N)` or :math:`(D, N)`
- Output: :math:`(B, D, N + 1)` or :math:`(D, N + 1)`
Examples::
>>> input = torch.rand(2, 4, 3) # BxNx3
>>> transform = tgm.ConvertPointsFromHomogeneous()
>>> output = transform(input) # BxNx2
"""
def __init__(self):
super().__init__()
def forward(self, input):
return convert_points_from_homogeneous(input)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
GCN_conv | # 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_4/inductor_cache/3v/c3v7n6hzyrv5pn6uojl3hf6tko347a672spakigdzmqm7ebd4zwl.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_4,), 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: '*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_relu_threshold_backward_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_relu_threshold_backward_0(in_out_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.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = 0.0
tmp4 = tmp2 <= tmp3
tl.store(in_out_ptr0 + (x0), tmp2, xmask)
tl.store(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, (4, 4), (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: [x_1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(primals_1, (64, 4), (4, 1), 0), primals_2, out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(primals_3, (16, 4, 4), (16, 4, 1), 0), reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0), out=buf1)
del buf0
buf2 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf1 # reuse
buf3 = 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(buf2, buf3, 256, grid=grid(256), stream=stream0)
return (buf2, buf3, reinterpret_tensor(primals_3, (16, 4, 4), (16, 1, 4), 0), reinterpret_tensor(primals_1, (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, 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, 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 math
import torch
import torch.utils.data
import torch.nn as nn
import torch.nn.functional as F
from torch.nn.parameter import Parameter
class GCN_conv(nn.Module):
def __init__(self, in_ft, out_ft, bias=False, dropout=0.0, activation=F
.relu):
super(GCN_conv, self).__init__()
self.weight = Parameter(torch.Tensor(in_ft, out_ft))
if bias:
self.bias = Parameter(torch.Tensor(out_ft))
else:
self.register_parameter('bias', None)
self.reset_parameters()
self.dropout = nn.Dropout(p=dropout)
self.activation = activation
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.weight.size(1))
self.weight.data.uniform_(-stdv, stdv)
if self.bias is not None:
self.bias.data.uniform_(-stdv, stdv)
def forward(self, x: 'torch.Tensor', G: 'torch.Tensor'):
x = self.dropout(x)
x = x.matmul(self.weight)
if self.bias is not None:
x = x + self.bias
x = G.matmul(x)
return self.activation(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_ft': 4, 'out_ft': 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 math
import torch.utils.data
import torch.nn as nn
import torch.nn.functional as F
from torch.nn.parameter import Parameter
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, 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.full([1], 0, tl.int32)
tmp2 = triton_helpers.maximum(tmp1, tmp0)
tmp3 = 0.0
tmp4 = tmp2 <= tmp3
tl.store(in_out_ptr0 + x0, tmp2, xmask)
tl.store(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, (4, 4), (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_1, (64, 4), (4, 1), 0),
primals_2, out=buf0)
del primals_2
buf1 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
extern_kernels.bmm(reinterpret_tensor(primals_3, (16, 4, 4), (16, 4,
1), 0), reinterpret_tensor(buf0, (16, 4, 4), (16, 4, 1), 0),
out=buf1)
del buf0
buf2 = reinterpret_tensor(buf1, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf1
buf3 = 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)](buf2, buf3,
256, XBLOCK=128, num_warps=4, num_stages=1)
return buf2, buf3, reinterpret_tensor(primals_3, (16, 4, 4), (16, 1, 4), 0
), reinterpret_tensor(primals_1, (4, 64), (1, 4), 0)
class GCN_convNew(nn.Module):
def __init__(self, in_ft, out_ft, bias=False, dropout=0.0, activation=F
.relu):
super(GCN_convNew, self).__init__()
self.weight = Parameter(torch.Tensor(in_ft, out_ft))
if bias:
self.bias = Parameter(torch.Tensor(out_ft))
else:
self.register_parameter('bias', None)
self.reset_parameters()
self.dropout = nn.Dropout(p=dropout)
self.activation = activation
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.weight.size(1))
self.weight.data.uniform_(-stdv, stdv)
if self.bias is not None:
self.bias.data.uniform_(-stdv, stdv)
def forward(self, input_0, input_1):
primals_2 = self.weight
primals_1 = input_0
primals_3 = input_1
output = call([primals_1, primals_2, primals_3])
return output[0]
| haoyfan/Dual-SVDAE | GCN_conv | false | 6,788 | [
"MIT"
] | 1 | 1fcb61960606d743438f33b740cb434dbfcfd727 | https://github.com/haoyfan/Dual-SVDAE/tree/1fcb61960606d743438f33b740cb434dbfcfd727 | import math
import torch
import torch.utils.data
import torch.nn as nn
import torch.nn.functional as F
from torch.nn.parameter import Parameter
class Model(nn.Module):
def __init__(self, in_ft, out_ft, bias=False, dropout=0.0, activation=F
.relu):
super().__init__()
self.weight = Parameter(torch.Tensor(in_ft, out_ft))
if bias:
self.bias = Parameter(torch.Tensor(out_ft))
else:
self.register_parameter('bias', None)
self.reset_parameters()
self.dropout = nn.Dropout(p=dropout)
self.activation = activation
def reset_parameters(self):
stdv = 1.0 / math.sqrt(self.weight.size(1))
self.weight.data.uniform_(-stdv, stdv)
if self.bias is not None:
self.bias.data.uniform_(-stdv, stdv)
def forward(self, x: 'torch.Tensor', G: 'torch.Tensor'):
x = self.dropout(x)
x = x.matmul(self.weight)
if self.bias is not None:
x = x + self.bias
x = G.matmul(x)
return self.activation(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
SelfAttn | # 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_4/inductor_cache/ud/cudbnoor4dpxxdmeqssolgpxcs76nh6jnfj2xf7auyk3sidsohnf.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_1, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_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=[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_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 = 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_4/inductor_cache/sx/csxi2lkjrgsyc43k5vauxutocx5sik4gnrglbjfjmvytt3alfq7w.py
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# alpha => 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_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 = 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_4/inductor_cache/wn/cwnhqhm2qqdlayd3loxd7a3oc7htjhgtsaapgavqmkseyqwaftly.py
# Topologically Sorted Source Nodes: [alpha, mul, summary], Original ATen: [aten._softmax, aten.mul, aten.sum]
# Source node to ATen node mapping:
# alpha => div, sum_1
# mul => mul
# summary => sum_2
# 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 = {})
# %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__softmax_mul_sum_2 = async_compile.triton('triton_poi_fused__softmax_mul_sum_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: '*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__softmax_mul_sum_2', '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__softmax_mul_sum_2(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 = args
args.clear()
assert_size_stride(primals_1, (1, 4), (4, 1))
assert_size_stride(primals_2, (1, ), (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)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf1)
del primals_1
del primals_2
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(buf1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [alpha], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf2, buf3, 64, grid=grid(64), stream=stream0)
buf4 = reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [alpha, mul, summary], Original ATen: [aten._softmax, aten.mul, aten.sum]
triton_poi_fused__softmax_mul_sum_2.run(primals_4, buf3, buf4, 64, grid=grid(64), stream=stream0)
del buf3
return (buf4, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0), 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((1, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((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, 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
import torch.nn.functional as F
import torch.nn.init
import torch as th
class SelfAttn(nn.Module):
def __init__(self, hidden_size):
super(SelfAttn, self).__init__()
self.query = nn.Linear(hidden_size, 1)
def forward(self, keys, values, attn_mask=None):
"""
:param attn_inputs: batch_size x time_len x hidden_size
:param attn_mask: batch_size x time_len
:return: summary state
"""
alpha = F.softmax(self.query(keys), dim=1)
if attn_mask is not None:
alpha = alpha * attn_mask.unsqueeze(2)
alpha = alpha / th.sum(alpha, dim=1, keepdim=True)
summary = th.sum(values * alpha, dim=1)
return summary
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'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 import triton_helpers
from torch._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.nn.init
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 = 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_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 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_poi_fused__softmax_mul_sum_2(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 = args
args.clear()
assert_size_stride(primals_1, (1, 4), (4, 1))
assert_size_stride(primals_2, (1,), (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)
buf1 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.addmm(primals_2, reinterpret_tensor(primals_3, (64,
4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 1), (1, 4), 0
), alpha=1, beta=1, out=buf1)
del primals_1
del primals_2
buf2 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(64)](buf1, buf2, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_1[grid(64)](buf2, buf3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf4 = reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0)
del buf2
triton_poi_fused__softmax_mul_sum_2[grid(64)](primals_4, buf3, buf4,
64, XBLOCK=64, num_warps=1, num_stages=1)
del buf3
return buf4, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf1
class SelfAttnNew(nn.Module):
def __init__(self, hidden_size):
super(SelfAttnNew, self).__init__()
self.query = nn.Linear(hidden_size, 1)
def forward(self, input_0, input_1):
primals_1 = self.query.weight
primals_2 = self.query.bias
primals_3 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| haojiepan1/CrossWOZ | SelfAttn | false | 6,789 | [
"Apache-2.0"
] | 1 | 6d7b4c4cfb73a528b76074764687906abecc90b6 | https://github.com/haojiepan1/CrossWOZ/tree/6d7b4c4cfb73a528b76074764687906abecc90b6 | import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.nn.init
import torch as th
class Model(nn.Module):
def __init__(self, hidden_size):
super().__init__()
self.query = nn.Linear(hidden_size, 1)
def forward(self, keys, values, attn_mask=None):
"""
:param attn_inputs: batch_size x time_len x hidden_size
:param attn_mask: batch_size x time_len
:return: summary state
"""
alpha = F.softmax(self.query(keys), dim=1)
if attn_mask is not None:
alpha = alpha * attn_mask.unsqueeze(2)
alpha = alpha / th.sum(alpha, dim=1, keepdim=True)
summary = th.sum(values * alpha, dim=1)
return summary
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
NormKLLoss | # 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_4/inductor_cache/2z/c2z4eec2npk4p6knlspn7qephjsfp5ynyiqqo7utfl3ifeidiszq.py
# Topologically Sorted Source Nodes: [sub, loss, sub_1, pow_1, exp, div, loss_1, exp_1, exp_2, div_1, loss_2, sum_1, kl_loss, avg_kl_loss], Original ATen: [aten.sub, aten.add, aten.pow, aten.exp, aten.div, aten.sum, aten.mul, aten.mean]
# Source node to ATen node mapping:
# avg_kl_loss => mean
# div => div
# div_1 => div_1
# exp => exp
# exp_1 => exp_1
# exp_2 => exp_2
# kl_loss => mul
# loss => add
# loss_1 => sub_2
# loss_2 => sub_3
# pow_1 => pow_1
# sub => sub
# sub_1 => sub_1
# sum_1 => sum_1
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub, 1.0), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%arg2_1, %arg3_1), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%sub_1, 2), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%arg1_1,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%pow_1, %exp), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%add, %div), kwargs = {})
# %exp_1 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%arg0_1,), kwargs = {})
# %exp_2 : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%arg1_1,), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_1, %exp_2), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub_2, %div_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%sub_3, [1]), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, -0.5), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%mul,), kwargs = {})
triton_per_fused_add_div_exp_mean_mul_pow_sub_sum_0 = async_compile.triton('triton_per_fused_add_div_exp_mean_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, 64],
reduction_hint=ReductionHint.INNER,
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': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 6), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_exp_mean_mul_pow_sub_sum_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 16, '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_div_exp_mean_mul_pow_sub_sum_0(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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)
r2 = rindex
tmp0 = tl.load(in_ptr0 + (r0 + (64*r1)), None)
tmp1 = tl.load(in_ptr1 + (r0 + (64*r1)), None)
tmp5 = tl.load(in_ptr2 + (r0 + (64*r1)), None)
tmp6 = tl.load(in_ptr3 + (r0 + (64*r1)), None)
tmp15 = tl.load(in_ptr0 + (16 + r0 + (64*r1)), None)
tmp16 = tl.load(in_ptr1 + (16 + r0 + (64*r1)), None)
tmp19 = tl.load(in_ptr2 + (16 + r0 + (64*r1)), None)
tmp20 = tl.load(in_ptr3 + (16 + r0 + (64*r1)), None)
tmp30 = tl.load(in_ptr0 + (32 + r0 + (64*r1)), None)
tmp31 = tl.load(in_ptr1 + (32 + r0 + (64*r1)), None)
tmp34 = tl.load(in_ptr2 + (32 + r0 + (64*r1)), None)
tmp35 = tl.load(in_ptr3 + (32 + r0 + (64*r1)), None)
tmp45 = tl.load(in_ptr0 + (48 + r0 + (64*r1)), None)
tmp46 = tl.load(in_ptr1 + (48 + r0 + (64*r1)), None)
tmp49 = tl.load(in_ptr2 + (48 + r0 + (64*r1)), None)
tmp50 = tl.load(in_ptr3 + (48 + r0 + (64*r1)), None)
tmp2 = tmp0 - tmp1
tmp3 = 1.0
tmp4 = tmp2 + tmp3
tmp7 = tmp5 - tmp6
tmp8 = tmp7 * tmp7
tmp9 = tl_math.exp(tmp1)
tmp10 = tmp8 / tmp9
tmp11 = tmp4 - tmp10
tmp12 = tl_math.exp(tmp0)
tmp13 = tmp12 / tmp9
tmp14 = tmp11 - tmp13
tmp17 = tmp15 - tmp16
tmp18 = tmp17 + tmp3
tmp21 = tmp19 - tmp20
tmp22 = tmp21 * tmp21
tmp23 = tl_math.exp(tmp16)
tmp24 = tmp22 / tmp23
tmp25 = tmp18 - tmp24
tmp26 = tl_math.exp(tmp15)
tmp27 = tmp26 / tmp23
tmp28 = tmp25 - tmp27
tmp29 = tmp14 + tmp28
tmp32 = tmp30 - tmp31
tmp33 = tmp32 + tmp3
tmp36 = tmp34 - tmp35
tmp37 = tmp36 * tmp36
tmp38 = tl_math.exp(tmp31)
tmp39 = tmp37 / tmp38
tmp40 = tmp33 - tmp39
tmp41 = tl_math.exp(tmp30)
tmp42 = tmp41 / tmp38
tmp43 = tmp40 - tmp42
tmp44 = tmp29 + tmp43
tmp47 = tmp45 - tmp46
tmp48 = tmp47 + tmp3
tmp51 = tmp49 - tmp50
tmp52 = tmp51 * tmp51
tmp53 = tl_math.exp(tmp46)
tmp54 = tmp52 / tmp53
tmp55 = tmp48 - tmp54
tmp56 = tl_math.exp(tmp45)
tmp57 = tmp56 / tmp53
tmp58 = tmp55 - tmp57
tmp59 = tmp44 + tmp58
tmp60 = -0.5
tmp61 = tmp59 * tmp60
tmp62 = tl.broadcast_to(tmp61, [XBLOCK, RBLOCK])
tmp64 = tl.sum(tmp62, 1)[:, None]
tmp65 = 64.0
tmp66 = tmp64 / tmp65
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp66, None)
''', 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)
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [sub, loss, sub_1, pow_1, exp, div, loss_1, exp_1, exp_2, div_1, loss_2, sum_1, kl_loss, avg_kl_loss], Original ATen: [aten.sub, aten.add, aten.pow, aten.exp, aten.div, aten.sum, aten.mul, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_add_div_exp_mean_mul_pow_sub_sum_0.run(buf2, arg0_1, arg1_1, arg2_1, arg3_1, 1, 64, grid=grid(1), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
del arg3_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)
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.init
import torch as th
from torch.nn.modules.loss import _Loss
class NormKLLoss(_Loss):
def __init__(self, unit_average=False):
super(NormKLLoss, self).__init__()
self.unit_average = unit_average
def forward(self, recog_mu, recog_logvar, prior_mu, prior_logvar):
loss = 1.0 + (recog_logvar - prior_logvar)
loss -= th.div(th.pow(prior_mu - recog_mu, 2), th.exp(prior_logvar))
loss -= th.div(th.exp(recog_logvar), th.exp(prior_logvar))
if self.unit_average:
kl_loss = -0.5 * th.mean(loss, dim=1)
else:
kl_loss = -0.5 * th.sum(loss, dim=1)
avg_kl_loss = th.mean(kl_loss)
return avg_kl_loss
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 [[], {}]
| 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.init
from torch.nn.modules.loss import _Loss
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_div_exp_mean_mul_pow_sub_sum_0(in_out_ptr0,
in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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)
tmp1 = tl.load(in_ptr1 + (r0 + 64 * r1), None)
tmp5 = tl.load(in_ptr2 + (r0 + 64 * r1), None)
tmp6 = tl.load(in_ptr3 + (r0 + 64 * r1), None)
tmp15 = tl.load(in_ptr0 + (16 + r0 + 64 * r1), None)
tmp16 = tl.load(in_ptr1 + (16 + r0 + 64 * r1), None)
tmp19 = tl.load(in_ptr2 + (16 + r0 + 64 * r1), None)
tmp20 = tl.load(in_ptr3 + (16 + r0 + 64 * r1), None)
tmp30 = tl.load(in_ptr0 + (32 + r0 + 64 * r1), None)
tmp31 = tl.load(in_ptr1 + (32 + r0 + 64 * r1), None)
tmp34 = tl.load(in_ptr2 + (32 + r0 + 64 * r1), None)
tmp35 = tl.load(in_ptr3 + (32 + r0 + 64 * r1), None)
tmp45 = tl.load(in_ptr0 + (48 + r0 + 64 * r1), None)
tmp46 = tl.load(in_ptr1 + (48 + r0 + 64 * r1), None)
tmp49 = tl.load(in_ptr2 + (48 + r0 + 64 * r1), None)
tmp50 = tl.load(in_ptr3 + (48 + r0 + 64 * r1), None)
tmp2 = tmp0 - tmp1
tmp3 = 1.0
tmp4 = tmp2 + tmp3
tmp7 = tmp5 - tmp6
tmp8 = tmp7 * tmp7
tmp9 = tl_math.exp(tmp1)
tmp10 = tmp8 / tmp9
tmp11 = tmp4 - tmp10
tmp12 = tl_math.exp(tmp0)
tmp13 = tmp12 / tmp9
tmp14 = tmp11 - tmp13
tmp17 = tmp15 - tmp16
tmp18 = tmp17 + tmp3
tmp21 = tmp19 - tmp20
tmp22 = tmp21 * tmp21
tmp23 = tl_math.exp(tmp16)
tmp24 = tmp22 / tmp23
tmp25 = tmp18 - tmp24
tmp26 = tl_math.exp(tmp15)
tmp27 = tmp26 / tmp23
tmp28 = tmp25 - tmp27
tmp29 = tmp14 + tmp28
tmp32 = tmp30 - tmp31
tmp33 = tmp32 + tmp3
tmp36 = tmp34 - tmp35
tmp37 = tmp36 * tmp36
tmp38 = tl_math.exp(tmp31)
tmp39 = tmp37 / tmp38
tmp40 = tmp33 - tmp39
tmp41 = tl_math.exp(tmp30)
tmp42 = tmp41 / tmp38
tmp43 = tmp40 - tmp42
tmp44 = tmp29 + tmp43
tmp47 = tmp45 - tmp46
tmp48 = tmp47 + tmp3
tmp51 = tmp49 - tmp50
tmp52 = tmp51 * tmp51
tmp53 = tl_math.exp(tmp46)
tmp54 = tmp52 / tmp53
tmp55 = tmp48 - tmp54
tmp56 = tl_math.exp(tmp45)
tmp57 = tmp56 / tmp53
tmp58 = tmp55 - tmp57
tmp59 = tmp44 + tmp58
tmp60 = -0.5
tmp61 = tmp59 * tmp60
tmp62 = tl.broadcast_to(tmp61, [XBLOCK, RBLOCK])
tmp64 = tl.sum(tmp62, 1)[:, None]
tmp65 = 64.0
tmp66 = tmp64 / tmp65
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp66, None)
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)
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
get_raw_stream(0)
triton_per_fused_add_div_exp_mean_mul_pow_sub_sum_0[grid(1)](buf2,
arg0_1, arg1_1, arg2_1, arg3_1, 1, 64, XBLOCK=1, num_warps=2,
num_stages=1)
del arg0_1
del arg1_1
del arg2_1
del arg3_1
return buf2,
class NormKLLossNew(_Loss):
def __init__(self, unit_average=False):
super(NormKLLossNew, self).__init__()
self.unit_average = unit_average
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]
| haojiepan1/CrossWOZ | NormKLLoss | false | 6,790 | [
"Apache-2.0"
] | 1 | 6d7b4c4cfb73a528b76074764687906abecc90b6 | https://github.com/haojiepan1/CrossWOZ/tree/6d7b4c4cfb73a528b76074764687906abecc90b6 | import torch
import torch.nn.init
import torch as th
from torch.nn.modules.loss import _Loss
class Model(_Loss):
def __init__(self, unit_average=False):
super().__init__()
self.unit_average = unit_average
def forward(self, recog_mu, recog_logvar, prior_mu, prior_logvar):
loss = 1.0 + (recog_logvar - prior_logvar)
loss -= th.div(th.pow(prior_mu - recog_mu, 2), th.exp(prior_logvar))
loss -= th.div(th.exp(recog_logvar), th.exp(prior_logvar))
if self.unit_average:
kl_loss = -0.5 * th.mean(loss, dim=1)
else:
kl_loss = -0.5 * th.sum(loss, dim=1)
avg_kl_loss = th.mean(kl_loss)
return avg_kl_loss
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 []
|
Zeronet | # 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_4/inductor_cache/2v/c2vwatafc7bysodsvfwld7cmvue26dh5jpkxqo6qevfrerixs7yp.py
# Topologically Sorted Source Nodes: [zeros_like], Original ATen: [aten.zeros_like]
# Source node to ATen node mapping:
# zeros_like => full_default
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
triton_poi_fused_zeros_like_0 = async_compile.triton('triton_poi_fused_zeros_like_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: '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_zeros_like_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_zeros_like_0(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 = 0.0
tl.store(out_ptr0 + (x0), 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: [zeros_like], Original ATen: [aten.zeros_like]
stream0 = get_raw_stream(0)
triton_poi_fused_zeros_like_0.run(buf0, 256, grid=grid(256), stream=stream0)
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 Zeronet(nn.Module):
def forward(self, x):
"""
Return a zero-out copy of x
:param x: torch.Tensor
:return: x*0, type torch.Tensor
"""
return torch.zeros_like(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
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_zeros_like_0(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 = 0.0
tl.store(out_ptr0 + x0, 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_zeros_like_0[grid(256)](buf0, 256, XBLOCK=128,
num_warps=4, num_stages=1)
return buf0,
class ZeronetNew(nn.Module):
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hedixia/xhd_source | Zeronet | false | 6,791 | [
"MIT"
] | 1 | cb176bceb5f5349d68206aaf60014e251de36300 | https://github.com/hedixia/xhd_source/tree/cb176bceb5f5349d68206aaf60014e251de36300 | import torch
import torch.nn as nn
class Model(nn.Module):
def forward(self, x):
"""
Return a zero-out copy of x
:param x: torch.Tensor
:return: x*0, type torch.Tensor
"""
return torch.zeros_like(x)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
BinaryDiceLoss | # 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_4/inductor_cache/s7/cs7jv7spsytbq3ouvdhla2tcr7wzgoznysid6m7rapuqn7g7cc3h.py
# Topologically Sorted Source Nodes: [mul, sum_1, sum_2, sum_3], Original ATen: [aten.mul, aten.sum]
# Source node to ATen node mapping:
# mul => mul
# sum_1 => sum_1
# sum_2 => sum_2
# sum_3 => sum_3
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %view_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view, [1]), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view_1, [1]), 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=[4, 64],
reduction_hint=ReductionHint.INNER,
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_mul_sum_0', 'mutated_arg_names': [], 'no_x_dim': False, 'num_load': 2, '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_mul_sum_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, out_ptr2, 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 + (r1 + (64*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.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp9 = tl.where(xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp13 = tl.where(xmask, tmp11, 0)
tmp14 = tl.sum(tmp13, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp10, xmask)
tl.store(out_ptr2 + (x0), tmp14, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/fo/cfogfxs73ofhe7t2r3fk5wthaeggkah4nq5qphtpn62jmraqsxsg.py
# Topologically Sorted Source Nodes: [mul_1, num, mul_2, add_1, den, truediv, loss, loss_1, mul_3], Original ATen: [aten.mul, aten.add, aten.div, aten.rsub, aten.mean]
# Source node to ATen node mapping:
# add_1 => add_1
# den => add_2
# loss => sub
# loss_1 => mean
# mul_1 => mul_1
# mul_2 => mul_2
# mul_3 => mul_3
# num => add
# truediv => div
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, 2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, 1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_2, 1.0), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %sum_3), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, 1), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %add_2), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sub, [0]), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, 1.0), kwargs = {})
triton_per_fused_add_div_mean_mul_rsub_1 = async_compile.triton('triton_per_fused_add_div_mean_mul_rsub_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, 4],
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': {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_add_div_mean_mul_rsub_1', 'mutated_arg_names': ['in_out_ptr0'], '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_add_div_mean_mul_rsub_1(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 + (r0), None)
tmp5 = tl.load(in_ptr1 + (r0), None)
tmp7 = tl.load(in_ptr2 + (r0), None)
tmp1 = 2.0
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 + tmp3
tmp6 = tmp5 * tmp3
tmp8 = tmp6 + tmp7
tmp9 = tmp8 + tmp3
tmp10 = tmp4 / tmp9
tmp11 = tmp3 - tmp10
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp16 * tmp3
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp17, 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, ), (1, ), torch.float32)
buf1 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf2 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul, sum_1, sum_2, sum_3], Original ATen: [aten.mul, aten.sum]
stream0 = get_raw_stream(0)
triton_per_fused_mul_sum_0.run(arg0_1, arg1_1, buf0, buf1, buf2, 4, 64, grid=grid(4), stream=stream0)
del arg0_1
del arg1_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [mul_1, num, mul_2, add_1, den, truediv, loss, loss_1, mul_3], Original ATen: [aten.mul, aten.add, aten.div, aten.rsub, aten.mean]
triton_per_fused_add_div_mean_mul_rsub_1.run(buf4, buf0, buf1, buf2, 1, 4, grid=grid(1), stream=stream0)
del buf0
del buf1
del buf2
return (buf4, )
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
class BinaryDiceLoss(nn.Module):
"""二分类版本的Dice Loss"""
def __init__(self, smooth: 'int'=1, exponent: 'int'=1, reduction: 'str'
='mean', loss_weight: 'float'=1.0, balance_weight: 'float'=1.0,
activation: 'bool'=False) ->None:
super(BinaryDiceLoss, self).__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.loss_weight = loss_weight
self.activation = activation
self.balanced_weight = balance_weight
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
num = torch.sum(torch.mul(pred, target), dim=1) * 2 + self.smooth
den = self.balanced_weight * torch.sum(pred.pow(self.exponent), dim=1
) + torch.sum(target.pow(self.exponent), dim=1) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * 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 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_per_fused_mul_sum_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
out_ptr2, 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 + (r1 + 64 * 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.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp9 = tl.where(xmask, tmp7, 0)
tmp10 = tl.sum(tmp9, 1)[:, None]
tmp11 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp13 = tl.where(xmask, tmp11, 0)
tmp14 = tl.sum(tmp13, 1)[:, None]
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp10, xmask)
tl.store(out_ptr2 + x0, tmp14, xmask)
@triton.jit
def triton_per_fused_add_div_mean_mul_rsub_1(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 + r0, None)
tmp5 = tl.load(in_ptr1 + r0, None)
tmp7 = tl.load(in_ptr2 + r0, None)
tmp1 = 2.0
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 + tmp3
tmp6 = tmp5 * tmp3
tmp8 = tmp6 + tmp7
tmp9 = tmp8 + tmp3
tmp10 = tmp4 / tmp9
tmp11 = tmp3 - tmp10
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp15 = 4.0
tmp16 = tmp14 / tmp15
tmp17 = tmp16 * tmp3
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp17, 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,), (1,), torch.float32)
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
buf2 = empty_strided_cuda((4,), (1,), torch.float32)
get_raw_stream(0)
triton_per_fused_mul_sum_0[grid(4)](arg0_1, arg1_1, buf0, buf1,
buf2, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf4 = buf3
del buf3
triton_per_fused_add_div_mean_mul_rsub_1[grid(1)](buf4, buf0, buf1,
buf2, 1, 4, XBLOCK=1, num_warps=2, num_stages=1)
del buf0
del buf1
del buf2
return buf4,
class BinaryDiceLossNew(nn.Module):
"""二分类版本的Dice Loss"""
def __init__(self, smooth: 'int'=1, exponent: 'int'=1, reduction: 'str'
='mean', loss_weight: 'float'=1.0, balance_weight: 'float'=1.0,
activation: 'bool'=False) ->None:
super(BinaryDiceLossNew, self).__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.loss_weight = loss_weight
self.activation = activation
self.balanced_weight = balance_weight
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| hehaoming/RSI-ChangeDetection | BinaryDiceLoss | false | 6,792 | [
"MIT"
] | 1 | f24a1d79c03fb9fefc49bc91bc94b3c120992496 | https://github.com/hehaoming/RSI-ChangeDetection/tree/f24a1d79c03fb9fefc49bc91bc94b3c120992496 | import torch
import torch.nn as nn
class Model(nn.Module):
"""二分类版本的Dice Loss"""
def __init__(self, smooth: 'int'=1, exponent: 'int'=1, reduction: 'str'
='mean', loss_weight: 'float'=1.0, balance_weight: 'float'=1.0,
activation: 'bool'=False) ->None:
super().__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.loss_weight = loss_weight
self.activation = activation
self.balanced_weight = balance_weight
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
num = torch.sum(torch.mul(pred, target), dim=1) * 2 + self.smooth
den = self.balanced_weight * torch.sum(pred.pow(self.exponent), dim=1
) + torch.sum(target.pow(self.exponent), dim=1) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
EqualizedLinear | # 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_4/inductor_cache/cv/ccvbgakrb4cp474imlxmtcx4i3upgifiohrjchmepetym6r7zjiq.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, 0.007071067811865476), 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 = 0.007071067811865476
tmp2 = tmp0 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/fj/cfjnczsousahbedhpdxgwg3r2czwvtdvud7lnsgyonowh3unwpdt.py
# Topologically Sorted Source Nodes: [bias], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# bias => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, 0.01), 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 = 0.01
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, ), (1, ))
assert_size_stride(primals_2, (4, 4), (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_1], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(primals_2, buf0, 16, grid=grid(16), stream=stream0)
del primals_2
buf1 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [bias], Original ATen: [aten.mul]
triton_poi_fused_mul_1.run(primals_1, buf1, 4, grid=grid(4), stream=stream0)
del primals_1
buf2 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [bias, 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, ), (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, 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 EqualizedLinear(nn.Module):
def __init__(self, input_size, output_size, gain=2 ** 0.5, lrmul=0.01):
super().__init__()
he_std = gain * input_size ** -0.5
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
def forward(self, x):
bias = self.bias
bias = bias * self.b_mul
return F.linear(x, self.weight * self.w_mul, bias)
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 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 = 0.007071067811865476
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 = 0.01
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,), (1,))
assert_size_stride(primals_2, (4, 4), (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_2, buf0, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del primals_2
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
triton_poi_fused_mul_1[grid(4)](primals_1, buf1, 4, XBLOCK=4,
num_warps=1, num_stages=1)
del primals_1
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 EqualizedLinearNew(nn.Module):
def __init__(self, input_size, output_size, gain=2 ** 0.5, lrmul=0.01):
super().__init__()
he_std = gain * input_size ** -0.5
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
def forward(self, input_0):
primals_2 = self.weight
primals_1 = self.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| hejj16/Landscape-StyleGAN | EqualizedLinear | false | 6,793 | [
"MIT"
] | 1 | a93cd32b588ab21da9d7589e705ca6f09db18408 | https://github.com/hejj16/Landscape-StyleGAN/tree/a93cd32b588ab21da9d7589e705ca6f09db18408 | import torch
from torch import nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self, input_size, output_size, gain=2 ** 0.5, lrmul=0.01):
super().__init__()
he_std = gain * input_size ** -0.5
init_std = 1.0 / lrmul
self.w_mul = he_std * lrmul
self.weight = torch.nn.Parameter(torch.randn(output_size,
input_size) * init_std)
self.bias = torch.nn.Parameter(torch.zeros(output_size))
self.b_mul = lrmul
def forward(self, x):
bias = self.bias
bias = bias * self.b_mul
return F.linear(x, self.weight * self.w_mul, bias)
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
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_4/inductor_cache/4u/c4ue3hy6x2lp63muanaoiwi6daoemjiri2lnpo3layzgnwdaon7e.py
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_softmax => amax, exp, log, sub, sub_1, sum_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%addmm, [1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%addmm, %amax), kwargs = {})
# %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 = {})
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=[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), 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': 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__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 12
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = (xindex // 3)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr0 + (3*x1), xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + (3*x1)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + (3*x1)), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp0 - tmp5
tmp7 = tmp1 - tmp5
tmp8 = tl_math.exp(tmp7)
tmp9 = tmp2 - tmp5
tmp10 = tl_math.exp(tmp9)
tmp11 = tmp8 + tmp10
tmp12 = tmp4 - tmp5
tmp13 = tl_math.exp(tmp12)
tmp14 = tmp11 + tmp13
tmp15 = tl_math.log(tmp14)
tmp16 = tmp6 - tmp15
tl.store(out_ptr0 + (x2), tmp16, 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, 121, 121), (58564, 14641, 121, 1))
assert_size_stride(primals_2, (3, 900), (900, 1))
assert_size_stride(primals_3, (3, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.avg_pool2d]
buf0 = torch.ops.aten.avg_pool2d.default(primals_1, [8, 8], [8, 8], [0, 0], False, True, None)
del primals_1
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 3), (3, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_3, reinterpret_tensor(buf1, (4, 900), (900, 1), 0), reinterpret_tensor(primals_2, (900, 3), (1, 900), 0), alpha=1, beta=1, out=buf2)
del primals_2
del primals_3
buf3 = empty_strided_cuda((4, 3), (3, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_softmax], Original ATen: [aten._log_softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__log_softmax_0.run(buf2, buf3, 12, grid=grid(12), stream=stream0)
del buf2
return (buf3, reinterpret_tensor(buf1, (4, 900), (900, 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, 121, 121), (58564, 14641, 121, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((3, 900), (900, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((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.nn.functional as F
class Classifier(nn.Module):
def __init__(self):
super(Classifier, self).__init__()
self.fc1 = nn.Linear(900, 3)
def forward(self, x):
x = F.avg_pool2d(x, 8)
x = x.view(-1, 900)
x = self.fc1(x)
return F.log_softmax(x, dim=1)
def get_inputs():
return [torch.rand([4, 4, 121, 121])]
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__log_softmax_0(in_ptr0, out_ptr0, xnumel, XBLOCK: tl.
constexpr):
xnumel = 12
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x1 = xindex // 3
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr0 + 3 * x1, xmask, eviction_policy='evict_last')
tmp2 = tl.load(in_ptr0 + (1 + 3 * x1), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr0 + (2 + 3 * x1), xmask, eviction_policy='evict_last')
tmp3 = triton_helpers.maximum(tmp1, tmp2)
tmp5 = triton_helpers.maximum(tmp3, tmp4)
tmp6 = tmp0 - tmp5
tmp7 = tmp1 - tmp5
tmp8 = tl_math.exp(tmp7)
tmp9 = tmp2 - tmp5
tmp10 = tl_math.exp(tmp9)
tmp11 = tmp8 + tmp10
tmp12 = tmp4 - tmp5
tmp13 = tl_math.exp(tmp12)
tmp14 = tmp11 + tmp13
tmp15 = tl_math.log(tmp14)
tmp16 = tmp6 - tmp15
tl.store(out_ptr0 + x2, tmp16, xmask)
def call(args):
primals_1, primals_2, primals_3 = args
args.clear()
assert_size_stride(primals_1, (4, 4, 121, 121), (58564, 14641, 121, 1))
assert_size_stride(primals_2, (3, 900), (900, 1))
assert_size_stride(primals_3, (3,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = torch.ops.aten.avg_pool2d.default(primals_1, [8, 8], [8, 8],
[0, 0], False, True, None)
del primals_1
buf1 = buf0
del buf0
buf2 = empty_strided_cuda((4, 3), (3, 1), torch.float32)
extern_kernels.addmm(primals_3, reinterpret_tensor(buf1, (4, 900),
(900, 1), 0), reinterpret_tensor(primals_2, (900, 3), (1, 900),
0), alpha=1, beta=1, out=buf2)
del primals_2
del primals_3
buf3 = empty_strided_cuda((4, 3), (3, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__log_softmax_0[grid(12)](buf2, buf3, 12, XBLOCK=16,
num_warps=1, num_stages=1)
del buf2
return buf3, reinterpret_tensor(buf1, (4, 900), (900, 1), 0), buf3
class ClassifierNew(nn.Module):
def __init__(self):
super(ClassifierNew, self).__init__()
self.fc1 = nn.Linear(900, 3)
def forward(self, input_0):
primals_2 = self.fc1.weight
primals_3 = self.fc1.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| helinwang/pytorch-semseg | Classifier | false | 6,794 | [
"MIT"
] | 1 | 117e5fb8afbad87d6968de1683867854ddec5885 | https://github.com/helinwang/pytorch-semseg/tree/117e5fb8afbad87d6968de1683867854ddec5885 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super().__init__()
self.fc1 = nn.Linear(900, 3)
def forward(self, x):
x = F.avg_pool2d(x, 8)
x = x.view(-1, 900)
x = self.fc1(x)
return F.log_softmax(x, dim=1)
def get_inputs():
return [torch.rand([4, 4, 121, 121])]
def get_init_inputs():
return []
|
ClassificationLogSoftmax | # 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_4/inductor_cache/yy/cyya3js6wt64vdji3sfisvrqyfvqxwkwqq5mzg5bqjl2crzjs4t3.py
# Topologically Sorted Source Nodes: [output_states], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# output_states => clone
# Graph fragment:
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%select,), 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 % 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')
# kernel path: runs/run_shard_4/inductor_cache/2g/c2gw7362i2a6wsfdx2sxyywx4o6ronjg6goebvdn44w6gpjsxpbc.py
# Topologically Sorted Source Nodes: [output_states, output_states_1], Original ATen: [aten.add, aten.tanh]
# Source node to ATen node mapping:
# output_states => add
# output_states_1 => tanh
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %primals_3), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add,), kwargs = {})
triton_poi_fused_add_tanh_1 = async_compile.triton('triton_poi_fused_add_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=[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_add_tanh_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_add_tanh_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
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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/4n/c4nc446yqbrtfyayl4mzt4mxucu6lyinpbl5i77rrpgokfkjfnsn.py
# Topologically Sorted Source Nodes: [log_probs], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_probs => amax, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_3, [-1], True), kwargs = {})
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %amax), kwargs = {})
triton_poi_fused__log_softmax_2 = async_compile.triton('triton_poi_fused__log_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__log_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__log_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
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/e5/ce5suhs2e2ygw6kp4ycmsbsq4xfgw573srqfqshl4crsnmymkvfl.py
# Topologically Sorted Source Nodes: [log_probs], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# log_probs => exp, log, sub_1, sum_1
# 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 = {})
triton_poi_fused__log_softmax_3 = async_compile.triton('triton_poi_fused__log_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__log_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__log_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')
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 = tl_math.exp(tmp1)
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp2 + tmp4
tmp7 = tl_math.exp(tmp6)
tmp8 = tmp5 + tmp7
tmp10 = tl_math.exp(tmp9)
tmp11 = tmp8 + tmp10
tmp12 = tl_math.log(tmp11)
tmp13 = tmp0 - 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 = 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((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output_states], 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_states], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf1)
del primals_2
buf2 = reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [output_states, output_states_1], Original ATen: [aten.add, aten.tanh]
triton_poi_fused_add_tanh_1.run(buf2, primals_3, 64, grid=grid(64), stream=stream0)
del primals_3
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_5, reinterpret_tensor(buf2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf3)
del primals_5
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [log_probs], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_2.run(buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = reinterpret_tensor(buf3, (4, 4, 4), (16, 4, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [log_probs], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_3.run(buf4, buf5, 64, grid=grid(64), stream=stream0)
del buf4
return (buf5, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), buf2, buf5, 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 ClassificationLogSoftmax(nn.Module):
"""
Classifier on top of the hidden representation of the first token, which
is usually [CLS] token in BERT-like architectures.
"""
def __init__(self, hidden_size, num_classes):
super().__init__()
self.dense1 = nn.Linear(hidden_size, hidden_size)
self.dense2 = nn.Linear(hidden_size, num_classes)
def forward(self, hidden_states):
output_states = self.dense1(hidden_states[:, 0])
output_states = torch.tanh(output_states)
output_states = self.dense2(output_states).float()
log_probs = torch.log_softmax(output_states, dim=-1)
return log_probs
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'hidden_size': 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
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_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)
@triton.jit
def triton_poi_fused_add_tanh_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
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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, xmask)
@triton.jit
def triton_poi_fused__log_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
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused__log_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')
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 = tl_math.exp(tmp1)
tmp4 = tl_math.exp(tmp3)
tmp5 = tmp2 + tmp4
tmp7 = tl_math.exp(tmp6)
tmp8 = tmp5 + tmp7
tmp10 = tl_math.exp(tmp9)
tmp11 = tmp8 + tmp10
tmp12 = tl_math.log(tmp11)
tmp13 = tmp0 - tmp12
tl.store(out_ptr0 + x2, tmp13, 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((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.mm(reinterpret_tensor(buf0, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_2, (4, 4), (1, 4), 0), out=buf1)
del primals_2
buf2 = reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0)
del buf1
triton_poi_fused_add_tanh_1[grid(64)](buf2, primals_3, 64, XBLOCK=
64, num_warps=1, num_stages=1)
del primals_3
buf3 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.addmm(primals_5, reinterpret_tensor(buf2, (16, 4), (
4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf3)
del primals_5
buf4 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused__log_softmax_2[grid(64)](buf3, buf4, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf5 = reinterpret_tensor(buf3, (4, 4, 4), (16, 4, 1), 0)
del buf3
triton_poi_fused__log_softmax_3[grid(64)](buf4, buf5, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del buf4
return buf5, reinterpret_tensor(buf0, (16, 4), (4, 1), 0
), buf2, buf5, primals_4
class ClassificationLogSoftmaxNew(nn.Module):
"""
Classifier on top of the hidden representation of the first token, which
is usually [CLS] token in BERT-like architectures.
"""
def __init__(self, hidden_size, num_classes):
super().__init__()
self.dense1 = nn.Linear(hidden_size, hidden_size)
self.dense2 = nn.Linear(hidden_size, num_classes)
def forward(self, input_0):
primals_2 = self.dense1.weight
primals_3 = self.dense1.bias
primals_4 = self.dense2.weight
primals_5 = self.dense2.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| harisankarh/NeMo | ClassificationLogSoftmax | false | 6,795 | [
"Apache-2.0"
] | 1 | 27bfb1aed24a786626e1c27c37417ebcd226ca8a | https://github.com/harisankarh/NeMo/tree/27bfb1aed24a786626e1c27c37417ebcd226ca8a | import torch
import torch.nn as nn
class Model(nn.Module):
"""
Classifier on top of the hidden representation of the first token, which
is usually [CLS] token in BERT-like architectures.
"""
def __init__(self, hidden_size, num_classes):
super().__init__()
self.dense1 = nn.Linear(hidden_size, hidden_size)
self.dense2 = nn.Linear(hidden_size, num_classes)
def forward(self, hidden_states):
output_states = self.dense1(hidden_states[:, 0])
output_states = torch.tanh(output_states)
output_states = self.dense2(output_states).float()
log_probs = torch.log_softmax(output_states, dim=-1)
return log_probs
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
BalancedBinaryCrossEntropy | # 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_4/inductor_cache/js/cjs76msm7qdpszudivunn4n5y2nu4rb4bk4jjbzuyofbjobushkf.py
# Topologically Sorted Source Nodes: [loss, loss_1], Original ATen: [aten.binary_cross_entropy, aten.mean]
# Source node to ATen node mapping:
# loss => full_default, full_default_1, log, log1p, maximum, maximum_1, mul, mul_1, neg, sub, sub_1
# loss_1 => mean
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_1, 1), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%view,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%neg,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log1p, %full_default), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %maximum), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%view,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum_1 : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log, %full_default_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %maximum_1), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul, %mul_1), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sub_1, [1]), kwargs = {})
triton_per_fused_binary_cross_entropy_mean_0 = async_compile.triton('triton_per_fused_binary_cross_entropy_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=[4, 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': {}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 4), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_binary_cross_entropy_mean_0', '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_binary_cross_entropy_mean_0(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)
tmp3 = tl.load(in_ptr1 + (r1 + (64*x0)), xmask, other=0.0)
tmp1 = 1.0
tmp2 = tmp0 - tmp1
tmp4 = -tmp3
tmp5 = libdevice.log1p(tmp4)
tmp6 = -100.0
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp2 * tmp7
tmp9 = tl_math.log(tmp3)
tmp10 = triton_helpers.maximum(tmp9, tmp6)
tmp11 = tmp0 * tmp10
tmp12 = tmp8 - tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp16, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/vs/cvsu42fy57afybbyvyjgvezv4vqbc4sjnt5buhwf67ysl6y3ogvq.py
# Topologically Sorted Source Nodes: [loss, loss_1, loss_2, mul], Original ATen: [aten.binary_cross_entropy, aten.mean, aten.mul]
# Source node to ATen node mapping:
# loss => full_default, full_default_1, log, log1p, maximum, maximum_1, mul, mul_1, neg, sub, sub_1
# loss_1 => mean
# loss_2 => mean_1
# mul => mul_2
# Graph fragment:
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_1, 1), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%view,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%neg,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log1p, %full_default), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub, %maximum), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%view,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum_1 : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log, %full_default_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_1, %maximum_1), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul, %mul_1), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sub_1, [1]), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mean, [0]), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean_1, 1.0), kwargs = {})
triton_per_fused_binary_cross_entropy_mean_mul_1 = async_compile.triton('triton_per_fused_binary_cross_entropy_mean_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.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_binary_cross_entropy_mean_mul_1', '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_binary_cross_entropy_mean_mul_1(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 = 64.0
tmp2 = tmp0 / tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.sum(tmp3, 1)[:, None]
tmp6 = 4.0
tmp7 = tmp5 / tmp6
tmp8 = 1.0
tmp9 = tmp7 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp9, 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, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [loss, loss_1], Original ATen: [aten.binary_cross_entropy, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_binary_cross_entropy_mean_0.run(arg1_1, arg0_1, buf0, 4, 64, grid=grid(4), stream=stream0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [loss, loss_1, loss_2, mul], Original ATen: [aten.binary_cross_entropy, aten.mean, aten.mul]
triton_per_fused_binary_cross_entropy_mean_mul_1.run(buf2, buf0, 1, 4, grid=grid(1), stream=stream0)
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
from typing import Any
import torch.nn.functional as F
class BalancedBinaryCrossEntropy(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False) ->None:
super(BalancedBinaryCrossEntropy, self).__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.activation = activation
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
if self.class_weight is not None:
class_weight = self.class_weight[0] * target + self.class_weight[1
] * (1 - target)
self.class_weight = class_weight.clone().detach().requires_grad_(
False)
loss = F.binary_cross_entropy(pred, target, reduction='none',
weight=self.class_weight)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * 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
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import torch.nn as nn
from typing import Any
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_binary_cross_entropy_mean_0(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)
tmp3 = tl.load(in_ptr1 + (r1 + 64 * x0), xmask, other=0.0)
tmp1 = 1.0
tmp2 = tmp0 - tmp1
tmp4 = -tmp3
tmp5 = libdevice.log1p(tmp4)
tmp6 = -100.0
tmp7 = triton_helpers.maximum(tmp5, tmp6)
tmp8 = tmp2 * tmp7
tmp9 = tl_math.log(tmp3)
tmp10 = triton_helpers.maximum(tmp9, tmp6)
tmp11 = tmp0 * tmp10
tmp12 = tmp8 - tmp11
tmp13 = tl.broadcast_to(tmp12, [XBLOCK, RBLOCK])
tmp15 = tl.where(xmask, tmp13, 0)
tmp16 = tl.sum(tmp15, 1)[:, None]
tl.store(out_ptr0 + x0, tmp16, xmask)
@triton.jit
def triton_per_fused_binary_cross_entropy_mean_mul_1(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 = 64.0
tmp2 = tmp0 / tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.sum(tmp3, 1)[:, None]
tmp6 = 4.0
tmp7 = tmp5 / tmp6
tmp8 = 1.0
tmp9 = tmp7 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp9, 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,), (1,), torch.float32)
get_raw_stream(0)
triton_per_fused_binary_cross_entropy_mean_0[grid(4)](arg1_1,
arg0_1, buf0, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((), (), torch.float32)
buf2 = buf1
del buf1
triton_per_fused_binary_cross_entropy_mean_mul_1[grid(1)](buf2,
buf0, 1, 4, XBLOCK=1, num_warps=2, num_stages=1)
del buf0
return buf2,
class BalancedBinaryCrossEntropyNew(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False) ->None:
super(BalancedBinaryCrossEntropyNew, self).__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.activation = activation
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| hehaoming/RSI-ChangeDetection | BalancedBinaryCrossEntropy | false | 6,796 | [
"MIT"
] | 1 | f24a1d79c03fb9fefc49bc91bc94b3c120992496 | https://github.com/hehaoming/RSI-ChangeDetection/tree/f24a1d79c03fb9fefc49bc91bc94b3c120992496 | import torch
import torch.nn as nn
from typing import Any
import torch.nn.functional as F
class Model(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False) ->None:
super().__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.activation = activation
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
if self.class_weight is not None:
class_weight = self.class_weight[0] * target + self.class_weight[1
] * (1 - target)
self.class_weight = class_weight.clone().detach().requires_grad_(
False)
loss = F.binary_cross_entropy(pred, target, reduction='none',
weight=self.class_weight)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
SeperableConv | # 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_4/inductor_cache/xl/cxlq33nvqoe4qybjyopeuxrddy5uzid5fs5kzfhjq367a6fsdj3s.py
# Topologically Sorted Source Nodes: [conv2d, x], Original ATen: [aten.convolution, aten.hardtanh, aten.hardtanh_backward]
# Source node to ATen node mapping:
# conv2d => convolution
# x => clamp_max, clamp_min
# Graph fragment:
# %convolution : [num_users=3] = call_function[target=torch.ops.aten.convolution.default](args = (%primals_3, %primals_1, %primals_2, [1, 1], [1, 1], [1, 1], False, [0, 0], 4), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%convolution, 0), kwargs = {})
# %clamp_max : [num_users=2] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 6), kwargs = {})
# %le_1 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%convolution, 0), kwargs = {})
# %ge_1 : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%convolution, 6), kwargs = {})
# %bitwise_or_1 : [num_users=1] = call_function[target=torch.ops.aten.bitwise_or.Tensor](args = (%le_1, %ge_1), kwargs = {})
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0 = async_compile.triton('triton_poi_fused_convolution_hardtanh_hardtanh_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: '*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_convolution_hardtanh_hardtanh_backward_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_convolution_hardtanh_hardtanh_backward_0(in_ptr0, in_ptr1, 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 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + (x3), tmp6, xmask)
tl.store(out_ptr1 + (x3), 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 = args
args.clear()
assert_size_stride(primals_1, (4, 1, 3, 3), (9, 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, 1, 1), (4, 1, 1, 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=4, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [conv2d, x], Original ATen: [aten.convolution, aten.hardtanh, aten.hardtanh_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0.run(buf0, primals_2, buf1, buf5, 256, grid=grid(256), stream=stream0)
del primals_2
# Topologically Sorted Source Nodes: [conv2d_1], Original ATen: [aten.convolution]
buf2 = extern_kernels.convolution(buf1, 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, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf0; del buf0 # reuse
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [conv2d_1, x_1], Original ATen: [aten.convolution, aten.hardtanh, aten.hardtanh_backward]
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0.run(buf2, primals_5, buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf2
del primals_5
return (buf3, primals_1, primals_3, primals_4, buf1, buf4, 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, 1, 3, 3), (9, 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, 1, 1), (4, 1, 1, 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.nn.functional as F
def _get_padding(kernel_size, stride, dilation):
padding = (stride - 1 + dilation * (kernel_size - 1)) // 2
return padding
class SeperableConv(nn.Module):
def __init__(self, inp, outp, k=3, stride=1, dilation=1):
super(SeperableConv, self).__init__()
self.depthwise = nn.Conv2d(inp, inp, k, stride, padding=
_get_padding(k, stride, dilation), dilation=dilation, groups=inp)
self.pointwise = nn.Conv2d(inp, outp, 1, 1)
def forward(self, x):
x = F.relu6(self.depthwise(x))
x = F.relu6(self.pointwise(x))
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'inp': 4, 'outp': 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_hardtanh_hardtanh_backward_0(in_ptr0,
in_ptr1, 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 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + x3, tmp6, xmask)
tl.store(out_ptr1 + x3, tmp9, xmask)
def call(args):
primals_1, primals_2, primals_3, primals_4, primals_5 = args
args.clear()
assert_size_stride(primals_1, (4, 1, 3, 3), (9, 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, 1, 1), (4, 1, 1, 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=4, bias=None)
assert_size_stride(buf0, (4, 4, 4, 4), (64, 16, 4, 1))
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf5 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0[grid(256)](
buf0, primals_2, buf1, buf5, 256, XBLOCK=256, num_warps=4,
num_stages=1)
del primals_2
buf2 = extern_kernels.convolution(buf1, 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, 4, 4, 4), (64, 16, 4, 1))
buf3 = buf0
del buf0
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0[grid(256)](
buf2, primals_5, buf3, buf4, 256, XBLOCK=256, num_warps=4,
num_stages=1)
del buf2
del primals_5
return buf3, primals_1, primals_3, primals_4, buf1, buf4, buf5
def _get_padding(kernel_size, stride, dilation):
padding = (stride - 1 + dilation * (kernel_size - 1)) // 2
return padding
class SeperableConvNew(nn.Module):
def __init__(self, inp, outp, k=3, stride=1, dilation=1):
super(SeperableConvNew, self).__init__()
self.depthwise = nn.Conv2d(inp, inp, k, stride, padding=
_get_padding(k, stride, dilation), dilation=dilation, groups=inp)
self.pointwise = nn.Conv2d(inp, outp, 1, 1)
def forward(self, input_0):
primals_1 = self.depthwise.weight
primals_2 = self.depthwise.bias
primals_4 = self.pointwise.weight
primals_5 = self.pointwise.bias
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4, primals_5])
return output[0]
| henningpohl/body-based-ar | SeperableConv | false | 6,797 | [
"MIT"
] | 1 | dc7d5d6eaf8dd4427de0f2b1cfdcc415cbfffdfb | https://github.com/henningpohl/body-based-ar/tree/dc7d5d6eaf8dd4427de0f2b1cfdcc415cbfffdfb | import torch
import torch.nn as nn
import torch.nn.functional as F
def _get_padding(kernel_size, stride, dilation):
padding = (stride - 1 + dilation * (kernel_size - 1)) // 2
return padding
class Model(nn.Module):
def __init__(self, inp, outp, k=3, stride=1, dilation=1):
super().__init__()
self.depthwise = nn.Conv2d(inp, inp, k, stride, padding=
_get_padding(k, stride, dilation), dilation=dilation, groups=inp)
self.pointwise = nn.Conv2d(inp, outp, 1, 1)
def forward(self, x):
x = F.relu6(self.depthwise(x))
x = F.relu6(self.pointwise(x))
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
HybridLoss | # 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_4/inductor_cache/re/creemvwl4ubcqbc4txngtzzz2zy4nmf5lvb7d67nqi7drv5f64gn.py
# Topologically Sorted Source Nodes: [mul, sum_1, sum_2, sum_3, loss_2, loss_3], Original ATen: [aten.mul, aten.sum, aten.binary_cross_entropy, aten.mean]
# Source node to ATen node mapping:
# loss_2 => full_default, full_default_1, log, log1p, maximum, maximum_1, mul_4, mul_5, neg, sub_1, sub_2
# loss_3 => mean_1
# mul => mul
# sum_1 => sum_1
# sum_2 => sum_2
# sum_3 => sum_3
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %view_1), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view, [1]), kwargs = {})
# %sum_3 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%view_1, [1]), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, 1), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%view_2,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%neg,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log1p, %full_default), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %maximum), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%view_2,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum_1 : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log, %full_default_1), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, %maximum_1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_4, %mul_5), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sub_2, [1]), kwargs = {})
triton_per_fused_binary_cross_entropy_mean_mul_sum_0 = async_compile.triton('triton_per_fused_binary_cross_entropy_mean_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=[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_binary_cross_entropy_mean_mul_sum_0', '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_binary_cross_entropy_mean_mul_sum_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, out_ptr2, out_ptr3, 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 + (r1 + (64*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 = 1.0
tmp8 = tmp1 - tmp7
tmp9 = -tmp0
tmp10 = libdevice.log1p(tmp9)
tmp11 = -100.0
tmp12 = triton_helpers.maximum(tmp10, tmp11)
tmp13 = tmp8 * tmp12
tmp14 = tl_math.log(tmp0)
tmp15 = triton_helpers.maximum(tmp14, tmp11)
tmp16 = tmp1 * tmp15
tmp17 = tmp13 - tmp16
tmp18 = tl.broadcast_to(tmp17, [XBLOCK, RBLOCK])
tmp20 = tl.where(xmask, tmp18, 0)
tmp21 = tl.sum(tmp20, 1)[:, None]
tmp22 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp24 = tl.where(xmask, tmp22, 0)
tmp25 = tl.sum(tmp24, 1)[:, None]
tmp26 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp28 = tl.where(xmask, tmp26, 0)
tmp29 = tl.sum(tmp28, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp21, xmask)
tl.store(out_ptr2 + (x0), tmp25, xmask)
tl.store(out_ptr3 + (x0), tmp29, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/gm/cgmuhi5xfdt7pzqlxvusrp7du6o4qzodttnjk2erqfzdx4bdygpn.py
# Topologically Sorted Source Nodes: [mul_1, num, mul_2, add_1, den, truediv, loss, loss_1, mul_3, loss_2, loss_3, loss_4, mul_4, add_3], Original ATen: [aten.mul, aten.add, aten.div, aten.rsub, aten.mean, aten.binary_cross_entropy]
# Source node to ATen node mapping:
# add_1 => add_1
# add_3 => add_3
# den => add_2
# loss => sub
# loss_1 => mean
# loss_2 => full_default, full_default_1, log, log1p, maximum, maximum_1, mul_4, mul_5, neg, sub_1, sub_2
# loss_3 => mean_1
# loss_4 => mean_2
# mul_1 => mul_1
# mul_2 => mul_2
# mul_3 => mul_3
# mul_4 => mul_6
# num => add
# truediv => div
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, 2), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, 1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_2, 1.0), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, %sum_3), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add_1, 1), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %add_2), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %div), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sub, [0]), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, 1.0), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, 1), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%view_2,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%neg,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log1p, %full_default), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %maximum), kwargs = {})
# %log : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%view_2,), kwargs = {})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], -100), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %maximum_1 : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%log, %full_default_1), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_3, %maximum_1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_4, %mul_5), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%sub_2, [1]), kwargs = {})
# %mean_2 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mean_1, [0]), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean_2, 1.0), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_3, %mul_6), kwargs = {})
triton_per_fused_add_binary_cross_entropy_div_mean_mul_rsub_1 = async_compile.triton('triton_per_fused_add_binary_cross_entropy_div_mean_mul_rsub_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, 4],
reduction_hint=ReductionHint.INNER,
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': {5: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4), equal_to_1=(5,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_binary_cross_entropy_div_mean_mul_rsub_1', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 4, '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_add_binary_cross_entropy_div_mean_mul_rsub_1(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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)
tmp5 = tl.load(in_ptr1 + (r0), None)
tmp7 = tl.load(in_ptr2 + (r0), None)
tmp15 = tl.load(in_ptr3 + (r0), None)
tmp1 = 2.0
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 + tmp3
tmp6 = tmp5 * tmp3
tmp8 = tmp6 + tmp7
tmp9 = tmp8 + tmp3
tmp10 = tmp4 / tmp9
tmp11 = tmp3 - tmp10
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp16 = 64.0
tmp17 = tmp15 / tmp16
tmp18 = tl.broadcast_to(tmp17, [XBLOCK, RBLOCK])
tmp20 = tl.sum(tmp18, 1)[:, None]
tmp21 = 4.0
tmp22 = tmp14 / tmp21
tmp23 = tmp22 * tmp3
tmp24 = tmp20 / tmp21
tmp25 = tmp24 * tmp3
tmp26 = tmp23 + tmp25
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp26, 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, ), (1, ), torch.float32)
buf4 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf1 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf2 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul, sum_1, sum_2, sum_3, loss_2, loss_3], Original ATen: [aten.mul, aten.sum, aten.binary_cross_entropy, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_binary_cross_entropy_mean_mul_sum_0.run(arg0_1, arg1_1, buf0, buf4, buf1, buf2, 4, 64, grid=grid(4), stream=stream0)
del arg0_1
del arg1_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf6 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [mul_1, num, mul_2, add_1, den, truediv, loss, loss_1, mul_3, loss_2, loss_3, loss_4, mul_4, add_3], Original ATen: [aten.mul, aten.add, aten.div, aten.rsub, aten.mean, aten.binary_cross_entropy]
triton_per_fused_add_binary_cross_entropy_div_mean_mul_rsub_1.run(buf6, buf0, buf1, buf2, buf4, 1, 4, grid=grid(1), stream=stream0)
del buf0
del buf1
del buf2
del buf4
return (buf6, )
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
from typing import Any
import torch.nn.functional as F
class BalancedBinaryCrossEntropy(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False) ->None:
super(BalancedBinaryCrossEntropy, self).__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.activation = activation
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
if self.class_weight is not None:
class_weight = self.class_weight[0] * target + self.class_weight[1
] * (1 - target)
self.class_weight = class_weight.clone().detach().requires_grad_(
False)
loss = F.binary_cross_entropy(pred, target, reduction='none',
weight=self.class_weight)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
class BinaryDiceLoss(nn.Module):
"""二分类版本的Dice Loss"""
def __init__(self, smooth: 'int'=1, exponent: 'int'=1, reduction: 'str'
='mean', loss_weight: 'float'=1.0, balance_weight: 'float'=1.0,
activation: 'bool'=False) ->None:
super(BinaryDiceLoss, self).__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.loss_weight = loss_weight
self.activation = activation
self.balanced_weight = balance_weight
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
num = torch.sum(torch.mul(pred, target), dim=1) * 2 + self.smooth
den = self.balanced_weight * torch.sum(pred.pow(self.exponent), dim=1
) + torch.sum(target.pow(self.exponent), dim=1) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
class HybridLoss(nn.Module):
def __init__(self):
super(HybridLoss, self).__init__()
self.dice = BinaryDiceLoss()
self.bce = BalancedBinaryCrossEntropy()
def forward(self, pred, target):
return self.dice(pred, target) + self.bce(pred, target)
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 libdevice, math as tl_math
import torch.nn as nn
from typing import Any
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_per_fused_binary_cross_entropy_mean_mul_sum_0(in_ptr0, in_ptr1,
out_ptr0, out_ptr1, out_ptr2, out_ptr3, 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 + (r1 + 64 * 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 = 1.0
tmp8 = tmp1 - tmp7
tmp9 = -tmp0
tmp10 = libdevice.log1p(tmp9)
tmp11 = -100.0
tmp12 = triton_helpers.maximum(tmp10, tmp11)
tmp13 = tmp8 * tmp12
tmp14 = tl_math.log(tmp0)
tmp15 = triton_helpers.maximum(tmp14, tmp11)
tmp16 = tmp1 * tmp15
tmp17 = tmp13 - tmp16
tmp18 = tl.broadcast_to(tmp17, [XBLOCK, RBLOCK])
tmp20 = tl.where(xmask, tmp18, 0)
tmp21 = tl.sum(tmp20, 1)[:, None]
tmp22 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp24 = tl.where(xmask, tmp22, 0)
tmp25 = tl.sum(tmp24, 1)[:, None]
tmp26 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp28 = tl.where(xmask, tmp26, 0)
tmp29 = tl.sum(tmp28, 1)[:, None]
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp21, xmask)
tl.store(out_ptr2 + x0, tmp25, xmask)
tl.store(out_ptr3 + x0, tmp29, xmask)
@triton.jit
def triton_per_fused_add_binary_cross_entropy_div_mean_mul_rsub_1(in_out_ptr0,
in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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)
tmp5 = tl.load(in_ptr1 + r0, None)
tmp7 = tl.load(in_ptr2 + r0, None)
tmp15 = tl.load(in_ptr3 + r0, None)
tmp1 = 2.0
tmp2 = tmp0 * tmp1
tmp3 = 1.0
tmp4 = tmp2 + tmp3
tmp6 = tmp5 * tmp3
tmp8 = tmp6 + tmp7
tmp9 = tmp8 + tmp3
tmp10 = tmp4 / tmp9
tmp11 = tmp3 - tmp10
tmp12 = tl.broadcast_to(tmp11, [XBLOCK, RBLOCK])
tmp14 = tl.sum(tmp12, 1)[:, None]
tmp16 = 64.0
tmp17 = tmp15 / tmp16
tmp18 = tl.broadcast_to(tmp17, [XBLOCK, RBLOCK])
tmp20 = tl.sum(tmp18, 1)[:, None]
tmp21 = 4.0
tmp22 = tmp14 / tmp21
tmp23 = tmp22 * tmp3
tmp24 = tmp20 / tmp21
tmp25 = tmp24 * tmp3
tmp26 = tmp23 + tmp25
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp26, 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,), (1,), torch.float32)
buf4 = empty_strided_cuda((4,), (1,), torch.float32)
buf1 = empty_strided_cuda((4,), (1,), torch.float32)
buf2 = empty_strided_cuda((4,), (1,), torch.float32)
get_raw_stream(0)
triton_per_fused_binary_cross_entropy_mean_mul_sum_0[grid(4)](arg0_1,
arg1_1, buf0, buf4, buf1, buf2, 4, 64, XBLOCK=1, num_warps=2,
num_stages=1)
del arg0_1
del arg1_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf6 = buf3
del buf3
triton_per_fused_add_binary_cross_entropy_div_mean_mul_rsub_1[grid(1)](
buf6, buf0, buf1, buf2, buf4, 1, 4, XBLOCK=1, num_warps=2,
num_stages=1)
del buf0
del buf1
del buf2
del buf4
return buf6,
class BalancedBinaryCrossEntropy(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False) ->None:
super(BalancedBinaryCrossEntropy, self).__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.activation = activation
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
if self.class_weight is not None:
class_weight = self.class_weight[0] * target + self.class_weight[1
] * (1 - target)
self.class_weight = class_weight.clone().detach().requires_grad_(
False)
loss = F.binary_cross_entropy(pred, target, reduction='none',
weight=self.class_weight)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
class BinaryDiceLoss(nn.Module):
"""二分类版本的Dice Loss"""
def __init__(self, smooth: 'int'=1, exponent: 'int'=1, reduction: 'str'
='mean', loss_weight: 'float'=1.0, balance_weight: 'float'=1.0,
activation: 'bool'=False) ->None:
super(BinaryDiceLoss, self).__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.loss_weight = loss_weight
self.activation = activation
self.balanced_weight = balance_weight
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
num = torch.sum(torch.mul(pred, target), dim=1) * 2 + self.smooth
den = self.balanced_weight * torch.sum(pred.pow(self.exponent), dim=1
) + torch.sum(target.pow(self.exponent), dim=1) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
class HybridLossNew(nn.Module):
def __init__(self):
super(HybridLossNew, self).__init__()
self.dice = BinaryDiceLoss()
self.bce = BalancedBinaryCrossEntropy()
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| hehaoming/RSI-ChangeDetection | HybridLoss | false | 6,798 | [
"MIT"
] | 1 | f24a1d79c03fb9fefc49bc91bc94b3c120992496 | https://github.com/hehaoming/RSI-ChangeDetection/tree/f24a1d79c03fb9fefc49bc91bc94b3c120992496 | import torch
import torch.nn as nn
from typing import Any
import torch.nn.functional as F
class BalancedBinaryCrossEntropy(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False) ->None:
super().__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.activation = activation
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
if self.class_weight is not None:
class_weight = self.class_weight[0] * target + self.class_weight[1
] * (1 - target)
self.class_weight = class_weight.clone().detach().requires_grad_(
False)
loss = F.binary_cross_entropy(pred, target, reduction='none',
weight=self.class_weight)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
class BinaryDiceLoss(nn.Module):
"""二分类版本的Dice Loss"""
def __init__(self, smooth: 'int'=1, exponent: 'int'=1, reduction: 'str'
='mean', loss_weight: 'float'=1.0, balance_weight: 'float'=1.0,
activation: 'bool'=False) ->None:
super().__init__()
self.smooth = smooth
self.exponent = exponent
self.reduction = reduction
self.loss_weight = loss_weight
self.activation = activation
self.balanced_weight = balance_weight
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
if self.activation:
pred = torch.sigmoid(pred)
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
num = torch.sum(torch.mul(pred, target), dim=1) * 2 + self.smooth
den = self.balanced_weight * torch.sum(pred.pow(self.exponent), dim=1
) + torch.sum(target.pow(self.exponent), dim=1) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
class Model(nn.Module):
def __init__(self):
super().__init__()
self.dice = BinaryDiceLoss()
self.bce = BalancedBinaryCrossEntropy()
def forward(self, pred, target):
return self.dice(pred, target) + self.bce(pred, target)
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
BalancedBinaryCrossEntropyWithLogits | # 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_4/inductor_cache/mk/cmkn6hrs27ozkjedibujstlag2ncutxymi2fgzsty7vxyuu44ovn.py
# Topologically Sorted Source Nodes: [sum_1, sub, sum_2], Original ATen: [aten.sum, aten.rsub]
# Source node to ATen node mapping:
# sub => sub
# sum_1 => sum_1
# sum_2 => sum_2
# Graph fragment:
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%view_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1.0, %view_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%sub,), kwargs = {})
triton_per_fused_rsub_sum_0 = async_compile.triton('triton_per_fused_rsub_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_rsub_sum_0', 'mutated_arg_names': [], 'no_x_dim': True, 'num_load': 1, '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_rsub_sum_0(in_ptr0, out_ptr0, out_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.broadcast_to(tmp0, [RBLOCK])
tmp3 = triton_helpers.promote_to_tensor(tl.sum(tmp1, 0))
tmp4 = 1.0
tmp5 = tmp4 - tmp0
tmp6 = tl.broadcast_to(tmp5, [RBLOCK])
tmp8 = triton_helpers.promote_to_tensor(tl.sum(tmp6, 0))
tl.store(out_ptr0 + (tl.full([1], 0, tl.int32)), tmp3, None)
tl.store(out_ptr1 + (tl.full([1], 0, tl.int32)), tmp8, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/7y/c7ydqqyty3copabgwbvtpc3aqysnn2njfmhuqlbqkqmy3nbx66qd.py
# Topologically Sorted Source Nodes: [mul, count_pos, count_neg, add_1, beta_back, binary_cross_entropy_with_logits, beta, loss, loss_1], Original ATen: [aten.mul, aten.add, aten.div, aten.binary_cross_entropy_with_logits, aten.mean]
# Source node to ATen node mapping:
# add_1 => add_1
# beta => div
# beta_back => div_1
# binary_cross_entropy_with_logits => abs_1, add_2, exp, full_default, log1p, minimum, mul_2, mul_3, mul_4, neg, sub_1, sub_2, sub_3, sub_4
# count_neg => mul_1
# count_pos => add
# loss => mul_5
# loss_1 => mean
# mul => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, 1.0), kwargs = {})
# %add : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 1e-10), kwargs = {})
# %mul_1 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_2, 1.0), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %mul_1), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %add_1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %view_1), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_2, %view), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_1, %add), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%div, 1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %view_1), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, 1), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %minimum : [num_users=1] = call_function[target=torch.ops.aten.minimum.default](args = (%full_default, %view), kwargs = {})
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%view,), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%abs_1,), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%neg,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%exp,), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%minimum, %log1p), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_2, %sub_3), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_3, %mul_4), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div_1, %sub_4), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mul_5, [1]), kwargs = {})
triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_1 = async_compile.triton('triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_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.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: '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_add_binary_cross_entropy_with_logits_div_mean_mul_1', '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_add_binary_cross_entropy_with_logits_div_mean_mul_1(in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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 + (0))
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp6 = tl.load(in_ptr1 + (0))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp11 = tl.load(in_ptr2 + (r1 + (64*x0)), xmask, other=0.0)
tmp13 = tl.load(in_ptr3 + (r1 + (64*x0)), xmask, other=0.0)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 1e-10
tmp5 = tmp3 + tmp4
tmp8 = tmp7 * tmp2
tmp9 = tmp5 + tmp8
tmp10 = tmp5 / tmp9
tmp12 = tmp2 - tmp11
tmp14 = tmp12 * tmp13
tmp15 = tmp8 / tmp5
tmp16 = tmp15 - tmp2
tmp17 = tmp16 * tmp11
tmp18 = tmp17 + tmp2
tmp19 = 0.0
tmp20 = triton_helpers.minimum(tmp19, tmp13)
tmp21 = tl_math.abs(tmp13)
tmp22 = -tmp21
tmp23 = tl_math.exp(tmp22)
tmp24 = libdevice.log1p(tmp23)
tmp25 = tmp20 - tmp24
tmp26 = tmp18 * tmp25
tmp27 = tmp14 - tmp26
tmp28 = tmp10 * tmp27
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp31 = tl.where(xmask, tmp29, 0)
tmp32 = tl.sum(tmp31, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp32, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/5f/c5fgmprlaqevsbmjkhre66xxcyxngyuo5dbl4rjphrwknwcbikik.py
# Topologically Sorted Source Nodes: [mul, count_pos, count_neg, add_1, beta_back, binary_cross_entropy_with_logits, beta, loss, loss_1, loss_2, mul_3], Original ATen: [aten.mul, aten.add, aten.div, aten.binary_cross_entropy_with_logits, aten.mean]
# Source node to ATen node mapping:
# add_1 => add_1
# beta => div
# beta_back => div_1
# binary_cross_entropy_with_logits => abs_1, add_2, exp, full_default, log1p, minimum, mul_2, mul_3, mul_4, neg, sub_1, sub_2, sub_3, sub_4
# count_neg => mul_1
# count_pos => add
# loss => mul_5
# loss_1 => mean
# loss_2 => mean_1
# mul => mul
# mul_3 => mul_6
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_1, 1.0), kwargs = {})
# %add : [num_users=3] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 1e-10), kwargs = {})
# %mul_1 : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sum_2, 1.0), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %mul_1), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %add_1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %view_1), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_2, %view), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul_1, %add), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%div, 1), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %view_1), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_2, 1), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %minimum : [num_users=1] = call_function[target=torch.ops.aten.minimum.default](args = (%full_default, %view), kwargs = {})
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%view,), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%abs_1,), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%neg,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%exp,), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%minimum, %log1p), kwargs = {})
# %mul_4 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_2, %sub_3), kwargs = {})
# %sub_4 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul_3, %mul_4), kwargs = {})
# %mul_5 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%div_1, %sub_4), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mul_5, [1]), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%mean, [0]), kwargs = {})
# %mul_6 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean_1, 1.0), kwargs = {})
triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_2 = async_compile.triton('triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_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.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_add_binary_cross_entropy_with_logits_div_mean_mul_2', '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_add_binary_cross_entropy_with_logits_div_mean_mul_2(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 = 64.0
tmp2 = tmp0 / tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.sum(tmp3, 1)[:, None]
tmp6 = 4.0
tmp7 = tmp5 / tmp6
tmp8 = 1.0
tmp9 = tmp7 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp9, 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 = empty_strided_cuda((), (), torch.float32)
# Topologically Sorted Source Nodes: [sum_1, sub, sum_2], Original ATen: [aten.sum, aten.rsub]
stream0 = get_raw_stream(0)
triton_per_fused_rsub_sum_0.run(arg1_1, buf0, buf1, 1, 256, grid=grid(1), stream=stream0)
buf2 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul, count_pos, count_neg, add_1, beta_back, binary_cross_entropy_with_logits, beta, loss, loss_1], Original ATen: [aten.mul, aten.add, aten.div, aten.binary_cross_entropy_with_logits, aten.mean]
triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_1.run(buf0, buf1, arg1_1, arg0_1, buf2, 4, 64, grid=grid(4), stream=stream0)
del arg0_1
del arg1_1
del buf0
buf3 = buf1; del buf1 # reuse
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [mul, count_pos, count_neg, add_1, beta_back, binary_cross_entropy_with_logits, beta, loss, loss_1, loss_2, mul_3], Original ATen: [aten.mul, aten.add, aten.div, aten.binary_cross_entropy_with_logits, aten.mean]
triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_2.run(buf4, buf2, 1, 4, grid=grid(1), stream=stream0)
del buf2
return (buf4, )
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
from typing import Any
class BalancedBinaryCrossEntropyWithLogits(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False, eposion:
'float'=1e-10) ->None:
super(BalancedBinaryCrossEntropyWithLogits, self).__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.eps = eposion
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
count_pos = torch.sum(target) * 1.0 + self.eps
count_neg = torch.sum(1.0 - target) * 1.0
beta = count_neg / count_pos
beta_back = count_pos / (count_pos + count_neg)
bce1 = nn.BCEWithLogitsLoss(pos_weight=beta, reduction='none')
loss = beta_back * bce1(pred, target)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * 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
from torch._inductor.runtime import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
import torch.nn as nn
from typing import Any
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_rsub_sum_0(in_ptr0, out_ptr0, out_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.broadcast_to(tmp0, [RBLOCK])
tmp3 = triton_helpers.promote_to_tensor(tl.sum(tmp1, 0))
tmp4 = 1.0
tmp5 = tmp4 - tmp0
tmp6 = tl.broadcast_to(tmp5, [RBLOCK])
tmp8 = triton_helpers.promote_to_tensor(tl.sum(tmp6, 0))
tl.store(out_ptr0 + tl.full([1], 0, tl.int32), tmp3, None)
tl.store(out_ptr1 + tl.full([1], 0, tl.int32), tmp8, None)
@triton.jit
def triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_1(
in_ptr0, in_ptr1, in_ptr2, in_ptr3, 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 + 0)
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp6 = tl.load(in_ptr1 + 0)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp11 = tl.load(in_ptr2 + (r1 + 64 * x0), xmask, other=0.0)
tmp13 = tl.load(in_ptr3 + (r1 + 64 * x0), xmask, other=0.0)
tmp2 = 1.0
tmp3 = tmp1 * tmp2
tmp4 = 1e-10
tmp5 = tmp3 + tmp4
tmp8 = tmp7 * tmp2
tmp9 = tmp5 + tmp8
tmp10 = tmp5 / tmp9
tmp12 = tmp2 - tmp11
tmp14 = tmp12 * tmp13
tmp15 = tmp8 / tmp5
tmp16 = tmp15 - tmp2
tmp17 = tmp16 * tmp11
tmp18 = tmp17 + tmp2
tmp19 = 0.0
tmp20 = triton_helpers.minimum(tmp19, tmp13)
tmp21 = tl_math.abs(tmp13)
tmp22 = -tmp21
tmp23 = tl_math.exp(tmp22)
tmp24 = libdevice.log1p(tmp23)
tmp25 = tmp20 - tmp24
tmp26 = tmp18 * tmp25
tmp27 = tmp14 - tmp26
tmp28 = tmp10 * tmp27
tmp29 = tl.broadcast_to(tmp28, [XBLOCK, RBLOCK])
tmp31 = tl.where(xmask, tmp29, 0)
tmp32 = tl.sum(tmp31, 1)[:, None]
tl.store(out_ptr0 + x0, tmp32, xmask)
@triton.jit
def triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_2(
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 = 64.0
tmp2 = tmp0 / tmp1
tmp3 = tl.broadcast_to(tmp2, [XBLOCK, RBLOCK])
tmp5 = tl.sum(tmp3, 1)[:, None]
tmp6 = 4.0
tmp7 = tmp5 / tmp6
tmp8 = 1.0
tmp9 = tmp7 * tmp8
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp9, 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 = empty_strided_cuda((), (), torch.float32)
get_raw_stream(0)
triton_per_fused_rsub_sum_0[grid(1)](arg1_1, buf0, buf1, 1, 256,
num_warps=2, num_stages=1)
buf2 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_1[
grid(4)](buf0, buf1, arg1_1, arg0_1, buf2, 4, 64, XBLOCK=1,
num_warps=2, num_stages=1)
del arg0_1
del arg1_1
del buf0
buf3 = buf1
del buf1
buf4 = buf3
del buf3
triton_per_fused_add_binary_cross_entropy_with_logits_div_mean_mul_2[
grid(1)](buf4, buf2, 1, 4, XBLOCK=1, num_warps=2, num_stages=1)
del buf2
return buf4,
class BalancedBinaryCrossEntropyWithLogitsNew(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False, eposion:
'float'=1e-10) ->None:
super(BalancedBinaryCrossEntropyWithLogitsNew, self).__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.eps = eposion
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| hehaoming/RSI-ChangeDetection | BalancedBinaryCrossEntropyWithLogits | false | 6,799 | [
"MIT"
] | 1 | f24a1d79c03fb9fefc49bc91bc94b3c120992496 | https://github.com/hehaoming/RSI-ChangeDetection/tree/f24a1d79c03fb9fefc49bc91bc94b3c120992496 | import torch
import torch.nn as nn
from typing import Any
class Model(nn.Module):
"""二分类加权交叉熵"""
def __init__(self, reduction: 'str'='mean', class_weight: 'Any'=None,
loss_weight: 'float'=1.0, activation: 'bool'=False, eposion:
'float'=1e-10) ->None:
super().__init__()
self.reduction = reduction
self.class_weight = class_weight
self.loss_weight = loss_weight
self.eps = eposion
def forward(self, pred, target):
assert self.reduction in ['sum', 'mean'
], "reduction is in ['sum', 'mean']"
assert pred.size(0) == target.size(0)
pred = pred.view(pred.size(0), -1)
target = target.view(target.size(0), -1)
count_pos = torch.sum(target) * 1.0 + self.eps
count_neg = torch.sum(1.0 - target) * 1.0
beta = count_neg / count_pos
beta_back = count_pos / (count_pos + count_neg)
bce1 = nn.BCEWithLogitsLoss(pos_weight=beta, reduction='none')
loss = beta_back * bce1(pred, target)
loss = loss.mean(dim=1)
if self.reduction == 'mean':
loss = loss.mean(dim=0)
else:
loss = loss.sum(dim=0)
return self.loss_weight * loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
InputConv | # 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_4/inductor_cache/xl/cxlq33nvqoe4qybjyopeuxrddy5uzid5fs5kzfhjq367a6fsdj3s.py
# Topologically Sorted Source Nodes: [conv2d, relu6], Original ATen: [aten.convolution, aten.hardtanh, aten.hardtanh_backward]
# Source node to ATen node mapping:
# conv2d => convolution
# relu6 => clamp_max, clamp_min
# Graph fragment:
# %convolution : [num_users=3] = 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 = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%convolution, 0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 6), kwargs = {})
# %le : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%convolution, 0), kwargs = {})
# %ge : [num_users=1] = call_function[target=torch.ops.aten.ge.Scalar](args = (%convolution, 6), kwargs = {})
# %bitwise_or : [num_users=1] = call_function[target=torch.ops.aten.bitwise_or.Tensor](args = (%le, %ge), kwargs = {})
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0 = async_compile.triton('triton_poi_fused_convolution_hardtanh_hardtanh_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: '*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_convolution_hardtanh_hardtanh_backward_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_convolution_hardtanh_hardtanh_backward_0(in_ptr0, in_ptr1, 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 + (x1), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + (x3), tmp6, xmask)
tl.store(out_ptr1 + (x3), tmp9, 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=(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((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [conv2d, relu6], Original ATen: [aten.convolution, aten.hardtanh, aten.hardtanh_backward]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0.run(buf0, primals_2, buf1, buf2, 256, grid=grid(256), stream=stream0)
del buf0
del primals_2
return (buf1, primals_1, primals_3, 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, 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 torch.nn as nn
import torch.nn.functional as F
def _get_padding(kernel_size, stride, dilation):
padding = (stride - 1 + dilation * (kernel_size - 1)) // 2
return padding
class InputConv(nn.Module):
def __init__(self, inp, outp, k=3, stride=1, dilation=1):
super(InputConv, self).__init__()
self.conv = nn.Conv2d(inp, outp, k, stride, padding=_get_padding(k,
stride, dilation), dilation=dilation)
def forward(self, x):
return F.relu6(self.conv(x))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'inp': 4, 'outp': 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_hardtanh_hardtanh_backward_0(in_ptr0,
in_ptr1, 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 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 0.0
tmp4 = triton_helpers.maximum(tmp2, tmp3)
tmp5 = 6.0
tmp6 = triton_helpers.minimum(tmp4, tmp5)
tmp7 = tmp2 <= tmp3
tmp8 = tmp2 >= tmp5
tmp9 = tmp7 | tmp8
tl.store(out_ptr0 + x3, tmp6, xmask)
tl.store(out_ptr1 + x3, tmp9, 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=(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((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_convolution_hardtanh_hardtanh_backward_0[grid(256)](
buf0, primals_2, buf1, buf2, 256, XBLOCK=256, num_warps=4,
num_stages=1)
del buf0
del primals_2
return buf1, primals_1, primals_3, buf2
def _get_padding(kernel_size, stride, dilation):
padding = (stride - 1 + dilation * (kernel_size - 1)) // 2
return padding
class InputConvNew(nn.Module):
def __init__(self, inp, outp, k=3, stride=1, dilation=1):
super(InputConvNew, self).__init__()
self.conv = nn.Conv2d(inp, outp, k, stride, padding=_get_padding(k,
stride, dilation), dilation=dilation)
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]
| henningpohl/body-based-ar | InputConv | false | 6,800 | [
"MIT"
] | 1 | dc7d5d6eaf8dd4427de0f2b1cfdcc415cbfffdfb | https://github.com/henningpohl/body-based-ar/tree/dc7d5d6eaf8dd4427de0f2b1cfdcc415cbfffdfb | import torch
import torch.nn as nn
import torch.nn.functional as F
def _get_padding(kernel_size, stride, dilation):
padding = (stride - 1 + dilation * (kernel_size - 1)) // 2
return padding
class Model(nn.Module):
def __init__(self, inp, outp, k=3, stride=1, dilation=1):
super().__init__()
self.conv = nn.Conv2d(inp, outp, k, stride, padding=_get_padding(k,
stride, dilation), dilation=dilation)
def forward(self, x):
return F.relu6(self.conv(x))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
ActQuant_PACT | # 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_4/inductor_cache/b4/cb44ft5rf5ygi33pb7o3ytmoov6fgeuzto5iydqu7nh52h2hqu53.py
# Topologically Sorted Source Nodes: [abs_1, abs_2, sub, abs_3, sub_1, add, out], Original ATen: [aten.abs, aten.sub, aten.add, aten.mul]
# Source node to ATen node mapping:
# abs_1 => abs_1
# abs_2 => abs_2
# abs_3 => abs_3
# add => add
# out => mul
# sub => sub
# sub_1 => sub_1
# Graph fragment:
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%primals_1,), kwargs = {})
# %abs_2 : [num_users=2] = call_function[target=torch.ops.aten.abs.default](args = (%primals_2,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %abs_2), kwargs = {})
# %abs_3 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%sub,), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%abs_1, %abs_3), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sub_1, %abs_2), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add, 0.5), kwargs = {})
triton_poi_fused_abs_add_mul_sub_0 = async_compile.triton('triton_poi_fused_abs_add_mul_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: '*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_abs_add_mul_sub_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_abs_add_mul_sub_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)
tmp2 = tl.load(in_ptr1 + (0))
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp1 = tl_math.abs(tmp0)
tmp4 = tl_math.abs(tmp3)
tmp5 = tmp0 - tmp4
tmp6 = tl_math.abs(tmp5)
tmp7 = tmp1 - tmp6
tmp8 = tmp7 + tmp4
tmp9 = 0.5
tmp10 = tmp8 * tmp9
tl.store(out_ptr0 + (x0), 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, 4), (64, 16, 4, 1))
assert_size_stride(primals_2, (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: [abs_1, abs_2, sub, abs_3, sub_1, add, out], Original ATen: [aten.abs, aten.sub, aten.add, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_abs_add_mul_sub_0.run(primals_1, primals_2, buf0, 256, grid=grid(256), stream=stream0)
return (buf0, 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, 4, 4, 4), (64, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((1, ), (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 uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class ActQuant_PACT(nn.Module):
def __init__(self, act_bit=4, scale_coef=1.0):
super(ActQuant_PACT, self).__init__()
self.act_bit = act_bit
self.scale_coef = nn.Parameter(torch.ones(1) * scale_coef)
self.uniform_q = uniform_quantize(k=act_bit)
def forward(self, x):
if self.act_bit == 32:
out = 0.5 * (x.abs() - (x - self.scale_coef.abs()).abs() + self
.scale_coef.abs()) / self.scale_coef.abs()
else:
out = 0.5 * (x.abs() - (x - self.scale_coef.abs()).abs() + self
.scale_coef.abs())
activation_q = self.uniform_q(out / self.scale_coef)
return activation_q
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_abs_add_mul_sub_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)
tmp2 = tl.load(in_ptr1 + 0)
tmp3 = tl.broadcast_to(tmp2, [XBLOCK])
tmp1 = tl_math.abs(tmp0)
tmp4 = tl_math.abs(tmp3)
tmp5 = tmp0 - tmp4
tmp6 = tl_math.abs(tmp5)
tmp7 = tmp1 - tmp6
tmp8 = tmp7 + tmp4
tmp9 = 0.5
tmp10 = tmp8 * tmp9
tl.store(out_ptr0 + x0, tmp10, 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, (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_abs_add_mul_sub_0[grid(256)](primals_1, primals_2,
buf0, 256, XBLOCK=128, num_warps=4, num_stages=1)
return buf0, primals_1, primals_2
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class ActQuant_PACTNew(nn.Module):
def __init__(self, act_bit=4, scale_coef=1.0):
super(ActQuant_PACTNew, self).__init__()
self.act_bit = act_bit
self.scale_coef = nn.Parameter(torch.ones(1) * scale_coef)
self.uniform_q = uniform_quantize(k=act_bit)
def forward(self, input_0):
primals_2 = self.scale_coef
primals_1 = input_0
output = call([primals_1, primals_2])
return output[0]
| heymesut/SJTU_microe | ActQuant_PACT | false | 6,801 | [
"BSD-3-Clause"
] | 1 | 7a862d03b4d8fe4c8608173a16082f44001f3f13 | https://github.com/heymesut/SJTU_microe/tree/7a862d03b4d8fe4c8608173a16082f44001f3f13 | import torch
import torch.nn as nn
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class Model(nn.Module):
def __init__(self, act_bit=4, scale_coef=1.0):
super().__init__()
self.act_bit = act_bit
self.scale_coef = nn.Parameter(torch.ones(1) * scale_coef)
self.uniform_q = uniform_quantize(k=act_bit)
def forward(self, x):
if self.act_bit == 32:
out = 0.5 * (x.abs() - (x - self.scale_coef.abs()).abs() + self
.scale_coef.abs()) / self.scale_coef.abs()
else:
out = 0.5 * (x.abs() - (x - self.scale_coef.abs()).abs() + self
.scale_coef.abs())
activation_q = self.uniform_q(out / self.scale_coef)
return activation_q
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Mish | # 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_4/inductor_cache/nf/cnfomfgkrddid5xfz2abcevz3avk7wtz7mughpoinpbdg3kjz27l.py
# Topologically Sorted Source Nodes: [softplus, tanh, mul_], Original ATen: [aten.softplus, aten.tanh, aten.mul]
# Source node to ATen node mapping:
# mul_ => mul
# softplus => exp, gt, log1p, where
# tanh => tanh
# Graph fragment:
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Scalar](args = (%arg0_1, 20), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%arg0_1,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%exp,), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%gt, %arg0_1, %log1p), kwargs = {})
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%where,), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %tanh), kwargs = {})
# %copy_ : [num_users=1] = call_function[target=torch.ops.aten.copy_.default](args = (%arg0_1, %mul), kwargs = {})
triton_poi_fused_mul_softplus_tanh_0 = async_compile.triton('triton_poi_fused_mul_softplus_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_mul_softplus_tanh_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_softplus_tanh_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 = 20.0
tmp2 = tmp0 > tmp1
tmp3 = tl_math.exp(tmp0)
tmp4 = libdevice.log1p(tmp3)
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp6 = libdevice.tanh(tmp5)
tmp7 = tmp0 * tmp6
tl.store(out_ptr1 + (x0), tmp7, 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: [softplus, tanh, mul_], Original ATen: [aten.softplus, aten.tanh, aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_softplus_tanh_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
import torch.nn.functional as F
class Mish(nn.Module):
def forward(self, x):
return x.mul_(F.softplus(x).tanh())
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, math as tl_math
import torch.nn as nn
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
@triton.jit
def triton_poi_fused_mul_softplus_tanh_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 = 20.0
tmp2 = tmp0 > tmp1
tmp3 = tl_math.exp(tmp0)
tmp4 = libdevice.log1p(tmp3)
tmp5 = tl.where(tmp2, tmp0, tmp4)
tmp6 = libdevice.tanh(tmp5)
tmp7 = tmp0 * tmp6
tl.store(out_ptr1 + x0, tmp7, 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_softplus_tanh_0[grid(256)](arg0_1, arg0_1, 256,
XBLOCK=128, num_warps=4, num_stages=1)
return arg0_1,
class MishNew(nn.Module):
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| heymesut/SJTU_microe | Mish | false | 6,802 | [
"BSD-3-Clause"
] | 1 | 7a862d03b4d8fe4c8608173a16082f44001f3f13 | https://github.com/heymesut/SJTU_microe/tree/7a862d03b4d8fe4c8608173a16082f44001f3f13 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def forward(self, x):
return x.mul_(F.softplus(x).tanh())
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
activation_quantize_fn | # AOT ID: ['1_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_4/inductor_cache/ey/ceyfjb5l4jwqlb5hnlsrrauzvcyyuprkp5vh3odjdm5fjdhyuxlp.py
# Topologically Sorted Source Nodes: [mul, round_1, out], Original ATen: [aten.mul, aten.round, aten.div]
# Source node to ATen node mapping:
# mul => mul
# out => div
# round_1 => round_1
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 15.0), kwargs = {})
# %round_1 : [num_users=1] = call_function[target=torch.ops.aten.round.default](args = (%mul,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%round_1, 15.0), kwargs = {})
triton_poi_fused_div_mul_round_0 = async_compile.triton('triton_poi_fused_div_mul_round_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_mul_round_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_div_mul_round_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 = 15.0
tmp2 = tmp0 * tmp1
tmp3 = libdevice.nearbyint(tmp2)
tmp4 = 0.06666666666666667
tmp5 = tmp3 * tmp4
tl.store(out_ptr0 + (x0), tmp5, 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, round_1, out], Original ATen: [aten.mul, aten.round, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_div_mul_round_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
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class activation_quantize_fn(nn.Module):
def __init__(self, a_bit):
super(activation_quantize_fn, self).__init__()
assert a_bit <= 8 or a_bit == 32
self.a_bit = a_bit
self.uniform_q = uniform_quantize(k=a_bit)
def forward(self, x):
if self.a_bit == 32:
activation_q = torch.clamp(x, 0, 6)
else:
activation_q = self.uniform_q(torch.clamp(x, 0, 1))
return activation_q
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'a_bit': 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
@triton.jit
def triton_poi_fused_div_mul_round_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 = 15.0
tmp2 = tmp0 * tmp1
tmp3 = libdevice.nearbyint(tmp2)
tmp4 = 0.06666666666666667
tmp5 = tmp3 * tmp4
tl.store(out_ptr0 + x0, tmp5, 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_div_mul_round_0[grid(256)](arg0_1, buf0, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
return buf0,
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class activation_quantize_fnNew(nn.Module):
def __init__(self, a_bit):
super(activation_quantize_fnNew, self).__init__()
assert a_bit <= 8 or a_bit == 32
self.a_bit = a_bit
self.uniform_q = uniform_quantize(k=a_bit)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| heymesut/SJTU_microe | activation_quantize_fn | false | 6,803 | [
"BSD-3-Clause"
] | 1 | 7a862d03b4d8fe4c8608173a16082f44001f3f13 | https://github.com/heymesut/SJTU_microe/tree/7a862d03b4d8fe4c8608173a16082f44001f3f13 | import torch
import torch.nn as nn
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class Model(nn.Module):
def __init__(self, a_bit):
super().__init__()
assert a_bit <= 8 or a_bit == 32
self.a_bit = a_bit
self.uniform_q = uniform_quantize(k=a_bit)
def forward(self, x):
if self.a_bit == 32:
activation_q = torch.clamp(x, 0, 6)
else:
activation_q = self.uniform_q(torch.clamp(x, 0, 1))
return activation_q
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
weightedFeatureFusion | # 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_4/inductor_cache/fs/cfs4xkoaiu25wxx5ko6j355loos24sadbddfu2644hsgmchy36go.py
# Topologically Sorted Source Nodes: [x, x_1], Original ATen: [aten.add]
# Source node to ATen node mapping:
# x => add
# x_1 => add_1
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, %select), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%add, %select_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
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (256 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tmp2 + tmp1
tl.store(out_ptr0 + (x2), tmp3, 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, (5, 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: [x, x_1], 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((5, 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
class weightedFeatureFusion(nn.Module):
def __init__(self, layers, weight=False):
super(weightedFeatureFusion, self).__init__()
self.layers = layers
self.weight = weight
self.n = len(layers) + 1
if weight:
self.w = torch.nn.Parameter(torch.zeros(self.n))
def forward(self, x, outputs):
if self.weight:
w = torch.sigmoid(self.w) * (2 / self.n)
x = x * w[0]
nc = x.shape[1]
for i in range(self.n - 1):
a = outputs[self.layers[i]] * w[i + 1] if self.weight else outputs[
self.layers[i]]
ac = a.shape[1]
dc = nc - ac
if dc > 0:
x[:, :ac] = x[:, :ac] + a
elif dc < 0:
x = x + a[:, :nc]
else:
x = x + a
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([5, 4, 4, 4])]
def get_init_inputs():
return [[], {'layers': [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 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_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
x2 = xindex
x0 = xindex % 64
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + (256 + x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tmp2 + tmp1
tl.store(out_ptr0 + x2, tmp3, 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, (5, 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 weightedFeatureFusionNew(nn.Module):
def __init__(self, layers, weight=False):
super(weightedFeatureFusionNew, self).__init__()
self.layers = layers
self.weight = weight
self.n = len(layers) + 1
if weight:
self.w = torch.nn.Parameter(torch.zeros(self.n))
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| heymesut/SJTU_microe | weightedFeatureFusion | false | 6,804 | [
"BSD-3-Clause"
] | 1 | 7a862d03b4d8fe4c8608173a16082f44001f3f13 | https://github.com/heymesut/SJTU_microe/tree/7a862d03b4d8fe4c8608173a16082f44001f3f13 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, layers, weight=False):
super().__init__()
self.layers = layers
self.weight = weight
self.n = len(layers) + 1
if weight:
self.w = torch.nn.Parameter(torch.zeros(self.n))
def forward(self, x, outputs):
if self.weight:
w = torch.sigmoid(self.w) * (2 / self.n)
x = x * w[0]
nc = x.shape[1]
for i in range(self.n - 1):
a = outputs[self.layers[i]] * w[i + 1] if self.weight else outputs[
self.layers[i]]
ac = a.shape[1]
dc = nc - ac
if dc > 0:
x[:, :ac] = x[:, :ac] + a
elif dc < 0:
x = x + a[:, :nc]
else:
x = x + a
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([5, 4, 4, 4])]
def get_init_inputs():
return []
|
ClipGlobalAvgPool2d | # 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_4/inductor_cache/4j/c4jgwgjmi42wee5hywcb44njpoactuw45h5mhl5ptm3heseytfxr.py
# Topologically Sorted Source Nodes: [mean, x_1], Original ATen: [aten.mean, aten.clamp]
# Source node to ATen node mapping:
# mean => mean
# x_1 => clamp_max, clamp_min
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%view, [-1]), kwargs = {})
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%view_1, 0.0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min, 1.0), kwargs = {})
triton_per_fused_clamp_mean_0 = async_compile.triton('triton_per_fused_clamp_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_clamp_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_clamp_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
tmp7 = 0.0
tmp8 = triton_helpers.maximum(tmp6, tmp7)
tmp9 = 1.0
tmp10 = triton_helpers.minimum(tmp8, tmp9)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp10, 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, 1), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [mean, x_1], Original ATen: [aten.mean, aten.clamp]
stream0 = get_raw_stream(0)
triton_per_fused_clamp_mean_0.run(buf1, arg0_1, 16, 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
from torch import nn
class FastGlobalAvgPool2d(nn.Module):
def __init__(self, flatten=False):
super(FastGlobalAvgPool2d, self).__init__()
self.flatten = flatten
def forward(self, x):
if self.flatten:
in_size = x.size()
return x.view((in_size[0], in_size[1], -1)).mean(dim=2)
else:
return x.view(x.size(0), x.size(1), -1).mean(-1).view(x.size(0),
x.size(1), 1, 1)
class ClipGlobalAvgPool2d(nn.Module):
def __init__(self):
super().__init__()
self.avgpool = FastGlobalAvgPool2d()
def forward(self, x):
x = self.avgpool(x)
x = torch.clamp(x, min=0.0, max=1.0)
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
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_per_fused_clamp_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
tmp7 = 0.0
tmp8 = triton_helpers.maximum(tmp6, tmp7)
tmp9 = 1.0
tmp10 = triton_helpers.minimum(tmp8, tmp9)
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp10, 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, 1), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf0
get_raw_stream(0)
triton_per_fused_clamp_mean_0[grid(16)](buf1, arg0_1, 16, 16,
XBLOCK=8, num_warps=2, num_stages=1)
del arg0_1
return buf1,
class FastGlobalAvgPool2d(nn.Module):
def __init__(self, flatten=False):
super(FastGlobalAvgPool2d, self).__init__()
self.flatten = flatten
def forward(self, x):
if self.flatten:
in_size = x.size()
return x.view((in_size[0], in_size[1], -1)).mean(dim=2)
else:
return x.view(x.size(0), x.size(1), -1).mean(-1).view(x.size(0),
x.size(1), 1, 1)
class ClipGlobalAvgPool2dNew(nn.Module):
def __init__(self):
super().__init__()
self.avgpool = FastGlobalAvgPool2d()
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hfyer/NAIC2020_ReID_R1 | ClipGlobalAvgPool2d | false | 6,805 | [
"Apache-2.0"
] | 1 | 240f0c9f65e482e6b0090f01d9f9e3373a337033 | https://github.com/hfyer/NAIC2020_ReID_R1/tree/240f0c9f65e482e6b0090f01d9f9e3373a337033 | import torch
from torch import nn
class FastGlobalAvgPool2d(nn.Module):
def __init__(self, flatten=False):
super().__init__()
self.flatten = flatten
def forward(self, x):
if self.flatten:
in_size = x.size()
return x.view((in_size[0], in_size[1], -1)).mean(dim=2)
else:
return x.view(x.size(0), x.size(1), -1).mean(-1).view(x.size(0),
x.size(1), 1, 1)
class Model(nn.Module):
def __init__(self):
super().__init__()
self.avgpool = FastGlobalAvgPool2d()
def forward(self, x):
x = self.avgpool(x)
x = torch.clamp(x, min=0.0, max=1.0)
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
GeneralizedMeanPooling | # 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_4/inductor_cache/xs/cxsstbq3wbw7zaisbdiefnedbcsbrsh6zg7qkoguxoj7qqghnei2.py
# Topologically Sorted Source Nodes: [clamp, x, adaptive_avg_pool2d, pow_2], Original ATen: [aten.clamp, aten.pow, aten.mean]
# Source node to ATen node mapping:
# adaptive_avg_pool2d => mean
# clamp => clamp_min
# pow_2 => pow_2
# x => pow_1
# Graph fragment:
# %clamp_min : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%arg0_1, 1e-06), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%clamp_min, 4.0), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_1, [-1, -2], True), kwargs = {})
# %pow_2 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%mean, 0.25), kwargs = {})
triton_per_fused_clamp_mean_pow_0 = async_compile.triton('triton_per_fused_clamp_mean_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.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_clamp_mean_pow_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_clamp_mean_pow_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 = 1e-06
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp3 = tmp2 * tmp2
tmp4 = tmp3 * tmp3
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp7 = tl.where(xmask, tmp5, 0)
tmp8 = tl.sum(tmp7, 1)[:, None]
tmp9 = 16.0
tmp10 = tmp8 / tmp9
tmp11 = 0.25
tmp12 = libdevice.pow(tmp10, tmp11)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp12, 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, 1, 1), (4, 1, 16, 16), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [clamp, x, adaptive_avg_pool2d, pow_2], Original ATen: [aten.clamp, aten.pow, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_clamp_mean_pow_0.run(buf1, arg0_1, 16, 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
from torch import nn
class GeneralizedMeanPooling(nn.Module):
"""Applies a 2D power-average adaptive pooling over an input signal composed of several input planes.
The function computed is: :math:`f(X) = pow(sum(pow(X, p)), 1/p)`
- At p = infinity, one gets Max Pooling
- At p = 1, one gets Average Pooling
The output is of size H x W, for any input size.
The number of output features is equal to the number of input planes.
Args:
output_size: the target output size of the image of the form H x W.
Can be a tuple (H, W) or a single H for a square image H x H
H and W can be either a ``int``, or ``None`` which means the size will
be the same as that of the input.
"""
def __init__(self, norm, output_size=1, eps=1e-06):
super(GeneralizedMeanPooling, self).__init__()
assert norm > 0
self.p = float(norm)
self.output_size = output_size
self.eps = eps
def forward(self, x):
x = x.clamp(min=self.eps).pow(self.p)
return torch.nn.functional.adaptive_avg_pool2d(x, self.output_size
).pow(1.0 / self.p)
def __repr__(self):
return self.__class__.__name__ + '(' + str(self.p
) + ', ' + 'output_size=' + str(self.output_size) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'norm': 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
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_per_fused_clamp_mean_pow_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 = 1e-06
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp3 = tmp2 * tmp2
tmp4 = tmp3 * tmp3
tmp5 = tl.broadcast_to(tmp4, [XBLOCK, RBLOCK])
tmp7 = tl.where(xmask, tmp5, 0)
tmp8 = tl.sum(tmp7, 1)[:, None]
tmp9 = 16.0
tmp10 = tmp8 / tmp9
tmp11 = 0.25
tmp12 = libdevice.pow(tmp10, tmp11)
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp12, 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, 1, 1), (4, 1, 16, 16), torch.float32)
buf1 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf0
get_raw_stream(0)
triton_per_fused_clamp_mean_pow_0[grid(16)](buf1, arg0_1, 16, 16,
XBLOCK=1, num_warps=2, num_stages=1)
del arg0_1
return buf1,
class GeneralizedMeanPoolingNew(nn.Module):
"""Applies a 2D power-average adaptive pooling over an input signal composed of several input planes.
The function computed is: :math:`f(X) = pow(sum(pow(X, p)), 1/p)`
- At p = infinity, one gets Max Pooling
- At p = 1, one gets Average Pooling
The output is of size H x W, for any input size.
The number of output features is equal to the number of input planes.
Args:
output_size: the target output size of the image of the form H x W.
Can be a tuple (H, W) or a single H for a square image H x H
H and W can be either a ``int``, or ``None`` which means the size will
be the same as that of the input.
"""
def __init__(self, norm, output_size=1, eps=1e-06):
super(GeneralizedMeanPoolingNew, self).__init__()
assert norm > 0
self.p = float(norm)
self.output_size = output_size
self.eps = eps
def __repr__(self):
return self.__class__.__name__ + '(' + str(self.p
) + ', ' + 'output_size=' + str(self.output_size) + ')'
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hfyer/NAIC2020_ReID_R1 | GeneralizedMeanPooling | false | 6,806 | [
"Apache-2.0"
] | 1 | 240f0c9f65e482e6b0090f01d9f9e3373a337033 | https://github.com/hfyer/NAIC2020_ReID_R1/tree/240f0c9f65e482e6b0090f01d9f9e3373a337033 | import torch
from torch import nn
class Model(nn.Module):
"""Applies a 2D power-average adaptive pooling over an input signal composed of several input planes.
The function computed is: :math:`f(X) = pow(sum(pow(X, p)), 1/p)`
- At p = infinity, one gets Max Pooling
- At p = 1, one gets Average Pooling
The output is of size H x W, for any input size.
The number of output features is equal to the number of input planes.
Args:
output_size: the target output size of the image of the form H x W.
Can be a tuple (H, W) or a single H for a square image H x H
H and W can be either a ``int``, or ``None`` which means the size will
be the same as that of the input.
"""
def __init__(self, norm, output_size=1, eps=1e-06):
super().__init__()
assert norm > 0
self.p = float(norm)
self.output_size = output_size
self.eps = eps
def forward(self, x):
x = x.clamp(min=self.eps).pow(self.p)
return torch.nn.functional.adaptive_avg_pool2d(x, self.output_size
).pow(1.0 / self.p)
def __repr__(self):
return self.__class__.__name__ + '(' + str(self.p
) + ', ' + 'output_size=' + str(self.output_size) + ')'
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
weight_quantize_fn | # 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_4/inductor_cache/bg/cbgdkimkstvz2qtck2yyugmlvmqfkfdg4psycwh5p5vsxuely43k.py
# Topologically Sorted Source Nodes: [weight, abs_1, max_1, weight_1], Original ATen: [aten.tanh, aten.abs, aten.max, aten.div]
# Source node to ATen node mapping:
# abs_1 => abs_1
# max_1 => max_1
# weight => tanh
# weight_1 => div
# Graph fragment:
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%arg0_1,), kwargs = {})
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%tanh,), kwargs = {})
# %max_1 : [num_users=1] = call_function[target=torch.ops.aten.max.default](args = (%abs_1,), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%tanh, %max_1), kwargs = {})
triton_per_fused_abs_div_max_tanh_0 = async_compile.triton('triton_per_fused_abs_div_max_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.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_div_max_tanh_0', 'mutated_arg_names': [], '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_div_max_tanh_0(in_ptr0, out_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 = libdevice.tanh(tmp0)
tmp2 = tl_math.abs(tmp1)
tmp3 = tl.broadcast_to(tmp2, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp3, 0))
tmp6 = tmp1 / tmp5
tl.store(out_ptr1 + (tl.broadcast_to(r0, [RBLOCK])), tmp6, 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)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [weight, abs_1, max_1, weight_1], Original ATen: [aten.tanh, aten.abs, aten.max, aten.div]
stream0 = get_raw_stream(0)
triton_per_fused_abs_div_max_tanh_0.run(arg0_1, buf1, 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
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class weight_quantize_fn(nn.Module):
def __init__(self, w_bit):
super(weight_quantize_fn, self).__init__()
assert w_bit <= 8 or w_bit == 32
self.w_bit = w_bit
self.uniform_q = uniform_quantize(k=w_bit - 1)
def forward(self, x):
if self.w_bit == 32:
weight = torch.tanh(x)
weight_q = weight / torch.max(torch.abs(weight))
elif self.w_bit == 1:
E = torch.mean(torch.abs(x)).detach()
weight_q = (self.uniform_q(x / E) + 1) / 2 * E
else:
weight = torch.tanh(x)
weight = weight / torch.max(torch.abs(weight))
weight_q = self.uniform_q(weight)
return weight_q
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'w_bit': 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, 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_per_fused_abs_div_max_tanh_0(in_ptr0, out_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 = libdevice.tanh(tmp0)
tmp2 = tl_math.abs(tmp1)
tmp3 = tl.broadcast_to(tmp2, [RBLOCK])
tmp5 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp3, 0))
tmp6 = tmp1 / tmp5
tl.store(out_ptr1 + tl.broadcast_to(r0, [RBLOCK]), tmp6, 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)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused_abs_div_max_tanh_0[grid(1)](arg0_1, buf1, 1, 256,
num_warps=2, num_stages=1)
del arg0_1
return buf1,
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class weight_quantize_fnNew(nn.Module):
def __init__(self, w_bit):
super(weight_quantize_fnNew, self).__init__()
assert w_bit <= 8 or w_bit == 32
self.w_bit = w_bit
self.uniform_q = uniform_quantize(k=w_bit - 1)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| heymesut/SJTU_microe | weight_quantize_fn | false | 6,807 | [
"BSD-3-Clause"
] | 1 | 7a862d03b4d8fe4c8608173a16082f44001f3f13 | https://github.com/heymesut/SJTU_microe/tree/7a862d03b4d8fe4c8608173a16082f44001f3f13 | import torch
import torch.nn as nn
def uniform_quantize(k):
class qfn(torch.autograd.Function):
@staticmethod
def forward(ctx, input):
if k == 32:
out = input
elif k == 1:
out = torch.sign(input)
else:
n = float(2 ** k - 1)
out = torch.round(input * n) / n
return out
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output.clone()
return grad_input
return qfn().apply
class Model(nn.Module):
def __init__(self, w_bit):
super().__init__()
assert w_bit <= 8 or w_bit == 32
self.w_bit = w_bit
self.uniform_q = uniform_quantize(k=w_bit - 1)
def forward(self, x):
if self.w_bit == 32:
weight = torch.tanh(x)
weight_q = weight / torch.max(torch.abs(weight))
elif self.w_bit == 1:
E = torch.mean(torch.abs(x)).detach()
weight_q = (self.uniform_q(x / E) + 1) / 2 * E
else:
weight = torch.tanh(x)
weight = weight / torch.max(torch.abs(weight))
weight_q = self.uniform_q(weight)
return weight_q
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
TLU | # 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_4/inductor_cache/cx/ccxf66rqhne57tzktyn4px2euzht4rw75vuy3fmua4ausvcns3e5.py
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.maximum, aten.eq, aten.gt]
# Source node to ATen node mapping:
# max_1 => maximum
# Graph fragment:
# %maximum : [num_users=1] = call_function[target=torch.ops.aten.maximum.default](args = (%primals_2, %view), kwargs = {})
# %eq : [num_users=1] = call_function[target=torch.ops.aten.eq.Tensor](args = (%primals_2, %view), kwargs = {})
# %gt : [num_users=1] = call_function[target=torch.ops.aten.gt.Tensor](args = (%primals_2, %view), kwargs = {})
triton_poi_fused_eq_gt_maximum_0 = async_compile.triton('triton_poi_fused_eq_gt_maximum_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: '*i1', 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_eq_gt_maximum_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_eq_gt_maximum_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1, out_ptr2, 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')
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp3 = tmp0 == tmp1
tmp4 = tmp0 > tmp1
tl.store(out_ptr0 + (x3), tmp2, xmask)
tl.store(out_ptr1 + (x3), tmp3, xmask)
tl.store(out_ptr2 + (x3), tmp4, 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)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.maximum, aten.eq, aten.gt]
stream0 = get_raw_stream(0)
triton_poi_fused_eq_gt_maximum_0.run(primals_2, primals_1, buf0, buf1, buf2, 256, grid=grid(256), stream=stream0)
del primals_1
del primals_2
return (buf0, 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, ), 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
from torch import nn
from torch.nn import Parameter
from torch.nn.parameter import Parameter
class TLU(nn.Module):
def __init__(self, num_features):
"""max(y, tau) = max(y - tau, 0) + tau = ReLU(y - tau) + tau"""
super(TLU, self).__init__()
self.num_features = num_features
self.tau = Parameter(torch.Tensor(num_features))
self.reset_parameters()
def reset_parameters(self):
nn.init.zeros_(self.tau)
def extra_repr(self):
return 'num_features={num_features}'.format(**self.__dict__)
def forward(self, x):
return torch.max(x, self.tau.view(1, self.num_features, 1, 1))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'num_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
from torch._inductor.runtime import triton_helpers
from torch import nn
from torch.nn import Parameter
from torch.nn.parameter 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_eq_gt_maximum_0(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
out_ptr2, 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')
tmp2 = triton_helpers.maximum(tmp0, tmp1)
tmp3 = tmp0 == tmp1
tmp4 = tmp0 > tmp1
tl.store(out_ptr0 + x3, tmp2, xmask)
tl.store(out_ptr1 + x3, tmp3, xmask)
tl.store(out_ptr2 + x3, tmp4, 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)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.bool)
get_raw_stream(0)
triton_poi_fused_eq_gt_maximum_0[grid(256)](primals_2, primals_1,
buf0, buf1, buf2, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_1
del primals_2
return buf0, buf1, buf2
class TLUNew(nn.Module):
def __init__(self, num_features):
"""max(y, tau) = max(y - tau, 0) + tau = ReLU(y - tau) + tau"""
super(TLUNew, self).__init__()
self.num_features = num_features
self.tau = Parameter(torch.Tensor(num_features))
self.reset_parameters()
def reset_parameters(self):
nn.init.zeros_(self.tau)
def extra_repr(self):
return 'num_features={num_features}'.format(**self.__dict__)
def forward(self, input_0):
primals_1 = self.tau
primals_2 = input_0
output = call([primals_1, primals_2])
return output[0]
| hfyer/NAIC2020_ReID_R1 | TLU | false | 6,808 | [
"Apache-2.0"
] | 1 | 240f0c9f65e482e6b0090f01d9f9e3373a337033 | https://github.com/hfyer/NAIC2020_ReID_R1/tree/240f0c9f65e482e6b0090f01d9f9e3373a337033 | import torch
from torch import nn
from torch.nn import Parameter
from torch.nn.parameter import Parameter
class Model(nn.Module):
def __init__(self, num_features):
"""max(y, tau) = max(y - tau, 0) + tau = ReLU(y - tau) + tau"""
super().__init__()
self.num_features = num_features
self.tau = Parameter(torch.Tensor(num_features))
self.reset_parameters()
def reset_parameters(self):
nn.init.zeros_(self.tau)
def extra_repr(self):
return 'num_features={num_features}'.format(**self.__dict__)
def forward(self, x):
return torch.max(x, self.tau.view(1, self.num_features, 1, 1))
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
YOLOLayer | # 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_4/inductor_cache/m5/cm5m3z3xhum74gicxtirsr2czpw5lxk6lux37hbq7jatahh7wss4.py
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# output => convolution
# Graph fragment:
# %convolution : [num_users=1] = 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=[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_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 = 2304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 36
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, (36, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_2, (36, ), (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: [output], 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, 36, 4, 4), (576, 16, 4, 1))
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.convolution]
stream0 = get_raw_stream(0)
triton_poi_fused_convolution_0.run(buf1, primals_2, 2304, grid=grid(2304), stream=stream0)
del primals_2
return (reinterpret_tensor(buf1, (4, 4, 4, 4, 9), (576, 144, 4, 1, 16), 0), 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((36, 4, 1, 1), (4, 1, 1, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((36, ), (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
import torch.nn as nn
class YOLOLayer(nn.Module):
"""
Detection Layer
"""
def __init__(self, in_ch, n_anchors, n_classes):
super(YOLOLayer, self).__init__()
self.n_anchors = n_anchors
self.n_classes = n_classes
self.conv = nn.Conv2d(in_channels=in_ch, out_channels=self.
n_anchors * (self.n_classes + 5), kernel_size=1, stride=1,
padding=0)
def forward(self, x, targets=None):
output = self.conv(x)
batchsize = output.shape[0]
fsize = output.shape[2]
n_ch = 5 + self.n_classes
output = output.view(batchsize, self.n_anchors, n_ch, fsize, fsize)
output = output.permute(0, 1, 3, 4, 2)
output[..., np.r_[:2, 4:n_ch]] = torch.sigmoid(output[..., np.r_[:2,
4:n_ch]])
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_ch': 4, 'n_anchors': 4, 'n_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
assert_size_stride = torch._C._dynamo.guards.assert_size_stride
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 = 2304
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 36
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, (36, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_2, (36,), (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=(1,
1), padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf0, (4, 36, 4, 4), (576, 16, 4, 1))
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_convolution_0[grid(2304)](buf1, primals_2, 2304,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_2
return reinterpret_tensor(buf1, (4, 4, 4, 4, 9), (576, 144, 4, 1, 16), 0
), primals_1, primals_3
class YOLOLayerNew(nn.Module):
"""
Detection Layer
"""
def __init__(self, in_ch, n_anchors, n_classes):
super(YOLOLayerNew, self).__init__()
self.n_anchors = n_anchors
self.n_classes = n_classes
self.conv = nn.Conv2d(in_channels=in_ch, out_channels=self.
n_anchors * (self.n_classes + 5), kernel_size=1, stride=1,
padding=0)
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]
| hiroki-kawauchi/SHAPObjectDetection | YOLOLayer | false | 6,809 | [
"MIT"
] | 1 | 3667d026949137cf710fc627672809c8564f5c6f | https://github.com/hiroki-kawauchi/SHAPObjectDetection/tree/3667d026949137cf710fc627672809c8564f5c6f | import torch
import numpy as np
import torch.nn as nn
class Model(nn.Module):
"""
Detection Layer
"""
def __init__(self, in_ch, n_anchors, n_classes):
super().__init__()
self.n_anchors = n_anchors
self.n_classes = n_classes
self.conv = nn.Conv2d(in_channels=in_ch, out_channels=self.
n_anchors * (self.n_classes + 5), kernel_size=1, stride=1,
padding=0)
def forward(self, x, targets=None):
output = self.conv(x)
batchsize = output.shape[0]
fsize = output.shape[2]
n_ch = 5 + self.n_classes
output = output.view(batchsize, self.n_anchors, n_ch, fsize, fsize)
output = output.permute(0, 1, 3, 4, 2)
output[..., np.r_[:2, 4:n_ch]] = torch.sigmoid(output[..., np.r_[:2,
4:n_ch]])
return output
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
AdaptiveAvgMaxPool2d | # 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_4/inductor_cache/gi/cgit46xvjyubowyg3p4y4wbubxiu32bmfjnskwxwflyyb3xn43bd.py
# Topologically Sorted Source Nodes: [max_feat, mean, feat], Original ATen: [aten.adaptive_max_pool2d, aten.mean, aten.add]
# Source node to ATen node mapping:
# feat => add
# max_feat => adaptive_max_pool2d
# mean => mean
# Graph fragment:
# %adaptive_max_pool2d : [num_users=1] = call_function[target=torch.ops.aten.adaptive_max_pool2d.default](args = (%arg0_1, [1, 1]), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%view, [-1]), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %getitem), kwargs = {})
triton_per_fused_adaptive_max_pool2d_add_mean_0 = async_compile.triton('triton_per_fused_adaptive_max_pool2d_add_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_adaptive_max_pool2d_add_mean_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': False, 'num_load': 17, '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_adaptive_max_pool2d_add_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)
tmp5 = tl.load(in_ptr0 + (16*x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + (16*x0)), xmask, eviction_policy='evict_last')
tmp8 = tl.load(in_ptr0 + (2 + (16*x0)), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr0 + (3 + (16*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (4 + (16*x0)), xmask, eviction_policy='evict_last')
tmp14 = tl.load(in_ptr0 + (5 + (16*x0)), xmask, eviction_policy='evict_last')
tmp16 = tl.load(in_ptr0 + (6 + (16*x0)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr0 + (7 + (16*x0)), xmask, eviction_policy='evict_last')
tmp20 = tl.load(in_ptr0 + (8 + (16*x0)), xmask, eviction_policy='evict_last')
tmp22 = tl.load(in_ptr0 + (9 + (16*x0)), xmask, eviction_policy='evict_last')
tmp24 = tl.load(in_ptr0 + (10 + (16*x0)), xmask, eviction_policy='evict_last')
tmp26 = tl.load(in_ptr0 + (11 + (16*x0)), xmask, eviction_policy='evict_last')
tmp28 = tl.load(in_ptr0 + (12 + (16*x0)), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr0 + (13 + (16*x0)), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr0 + (14 + (16*x0)), xmask, eviction_policy='evict_last')
tmp34 = tl.load(in_ptr0 + (15 + (16*x0)), xmask, eviction_policy='evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp7 = triton_helpers.maximum(tmp6, tmp5)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp11 = triton_helpers.maximum(tmp10, tmp9)
tmp13 = triton_helpers.maximum(tmp12, tmp11)
tmp15 = triton_helpers.maximum(tmp14, tmp13)
tmp17 = triton_helpers.maximum(tmp16, tmp15)
tmp19 = triton_helpers.maximum(tmp18, tmp17)
tmp21 = triton_helpers.maximum(tmp20, tmp19)
tmp23 = triton_helpers.maximum(tmp22, tmp21)
tmp25 = triton_helpers.maximum(tmp24, tmp23)
tmp27 = triton_helpers.maximum(tmp26, tmp25)
tmp29 = triton_helpers.maximum(tmp28, tmp27)
tmp31 = triton_helpers.maximum(tmp30, tmp29)
tmp33 = triton_helpers.maximum(tmp32, tmp31)
tmp35 = triton_helpers.maximum(tmp34, tmp33)
tmp36 = 16.0
tmp37 = tmp4 / tmp36
tmp38 = tmp37 + tmp35
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp38, 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, 1, 1), (4, 1, 16, 16), torch.float32)
buf2 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [max_feat, mean, feat], Original ATen: [aten.adaptive_max_pool2d, aten.mean, aten.add]
stream0 = get_raw_stream(0)
triton_per_fused_adaptive_max_pool2d_add_mean_0.run(buf2, arg0_1, 16, 16, grid=grid(16), 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
from torch import nn
class FastGlobalAvgPool2d(nn.Module):
def __init__(self, flatten=False):
super(FastGlobalAvgPool2d, self).__init__()
self.flatten = flatten
def forward(self, x):
if self.flatten:
in_size = x.size()
return x.view((in_size[0], in_size[1], -1)).mean(dim=2)
else:
return x.view(x.size(0), x.size(1), -1).mean(-1).view(x.size(0),
x.size(1), 1, 1)
class AdaptiveAvgMaxPool2d(nn.Module):
def __init__(self):
super(AdaptiveAvgMaxPool2d, self).__init__()
self.gap = FastGlobalAvgPool2d()
self.gmp = nn.AdaptiveMaxPool2d(1)
def forward(self, x):
avg_feat = self.gap(x)
max_feat = self.gmp(x)
feat = avg_feat + max_feat
return feat
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 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_per_fused_adaptive_max_pool2d_add_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)
tmp5 = tl.load(in_ptr0 + 16 * x0, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp8 = tl.load(in_ptr0 + (2 + 16 * x0), xmask, eviction_policy='evict_last'
)
tmp10 = tl.load(in_ptr0 + (3 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp12 = tl.load(in_ptr0 + (4 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp14 = tl.load(in_ptr0 + (5 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp16 = tl.load(in_ptr0 + (6 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp18 = tl.load(in_ptr0 + (7 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp20 = tl.load(in_ptr0 + (8 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp22 = tl.load(in_ptr0 + (9 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp24 = tl.load(in_ptr0 + (10 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp26 = tl.load(in_ptr0 + (11 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp28 = tl.load(in_ptr0 + (12 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp30 = tl.load(in_ptr0 + (13 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp32 = tl.load(in_ptr0 + (14 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp34 = tl.load(in_ptr0 + (15 + 16 * x0), xmask, eviction_policy=
'evict_last')
tmp1 = tl.broadcast_to(tmp0, [XBLOCK, RBLOCK])
tmp3 = tl.where(xmask, tmp1, 0)
tmp4 = tl.sum(tmp3, 1)[:, None]
tmp7 = triton_helpers.maximum(tmp6, tmp5)
tmp9 = triton_helpers.maximum(tmp8, tmp7)
tmp11 = triton_helpers.maximum(tmp10, tmp9)
tmp13 = triton_helpers.maximum(tmp12, tmp11)
tmp15 = triton_helpers.maximum(tmp14, tmp13)
tmp17 = triton_helpers.maximum(tmp16, tmp15)
tmp19 = triton_helpers.maximum(tmp18, tmp17)
tmp21 = triton_helpers.maximum(tmp20, tmp19)
tmp23 = triton_helpers.maximum(tmp22, tmp21)
tmp25 = triton_helpers.maximum(tmp24, tmp23)
tmp27 = triton_helpers.maximum(tmp26, tmp25)
tmp29 = triton_helpers.maximum(tmp28, tmp27)
tmp31 = triton_helpers.maximum(tmp30, tmp29)
tmp33 = triton_helpers.maximum(tmp32, tmp31)
tmp35 = triton_helpers.maximum(tmp34, tmp33)
tmp36 = 16.0
tmp37 = tmp4 / tmp36
tmp38 = tmp37 + tmp35
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp38, 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, 1, 1), (4, 1, 16, 16), torch.float32)
buf2 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf0
get_raw_stream(0)
triton_per_fused_adaptive_max_pool2d_add_mean_0[grid(16)](buf2,
arg0_1, 16, 16, XBLOCK=8, num_warps=2, num_stages=1)
del arg0_1
return buf2,
class FastGlobalAvgPool2d(nn.Module):
def __init__(self, flatten=False):
super(FastGlobalAvgPool2d, self).__init__()
self.flatten = flatten
def forward(self, x):
if self.flatten:
in_size = x.size()
return x.view((in_size[0], in_size[1], -1)).mean(dim=2)
else:
return x.view(x.size(0), x.size(1), -1).mean(-1).view(x.size(0),
x.size(1), 1, 1)
class AdaptiveAvgMaxPool2dNew(nn.Module):
def __init__(self):
super(AdaptiveAvgMaxPool2dNew, self).__init__()
self.gap = FastGlobalAvgPool2d()
self.gmp = nn.AdaptiveMaxPool2d(1)
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hfyer/NAIC2020_ReID_R1 | AdaptiveAvgMaxPool2d | false | 6,810 | [
"Apache-2.0"
] | 1 | 240f0c9f65e482e6b0090f01d9f9e3373a337033 | https://github.com/hfyer/NAIC2020_ReID_R1/tree/240f0c9f65e482e6b0090f01d9f9e3373a337033 | import torch
from torch import nn
class FastGlobalAvgPool2d(nn.Module):
def __init__(self, flatten=False):
super().__init__()
self.flatten = flatten
def forward(self, x):
if self.flatten:
in_size = x.size()
return x.view((in_size[0], in_size[1], -1)).mean(dim=2)
else:
return x.view(x.size(0), x.size(1), -1).mean(-1).view(x.size(0),
x.size(1), 1, 1)
class Model(nn.Module):
def __init__(self):
super().__init__()
self.gap = FastGlobalAvgPool2d()
self.gmp = nn.AdaptiveMaxPool2d(1)
def forward(self, x):
avg_feat = self.gap(x)
max_feat = self.gmp(x)
feat = avg_feat + max_feat
return feat
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
KeyValueAttention | # 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_4/inductor_cache/ao/caoovxtqrx42gvkmjirowqmmbh6kppvfh5ebrzzv4kzkgwm2umii.py
# Topologically Sorted Source Nodes: [k], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# k => 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=[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_4/inductor_cache/z7/cz7feaqdg3gmfplkn7nqvmafptfffgz3wm2tf4wbh3ip3wywrdsl.py
# Topologically Sorted Source Nodes: [add, h], Original ATen: [aten.add, aten.tanh]
# Source node to ATen node mapping:
# add => add
# h => tanh
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_1, %expand), kwargs = {})
# %tanh : [num_users=2] = call_function[target=torch.ops.aten.tanh.default](args = (%add,), kwargs = {})
triton_poi_fused_add_tanh_1 = async_compile.triton('triton_poi_fused_add_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=[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_tanh_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_tanh_1(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
x3 = xindex
x0 = xindex % 4
x2 = (xindex // 16)
tmp0 = tl.load(in_out_ptr0 + (x3), xmask)
tmp1 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x0 + (4*x2)), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = libdevice.tanh(tmp6)
tl.store(in_out_ptr0 + (x3), tmp7, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/2f/c2fecjxc5dsfckn3atawgehw7voreh2onnyz6mfji7zpuuxjss6k.py
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
# Source node to ATen node mapping:
# max_1 => getitem_1
# Graph fragment:
# %getitem_1 : [num_users=1] = call_function[target=operator.getitem](args = (%max_1, 1), 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=[4],
filename=__file__,
triton_meta={'signature': {0: '*fp32', 1: '*i64', 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_max_2', '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_2(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 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp32 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
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)
tl.store(out_ptr0 + (x0), tmp46, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/gt/cgtkiticmc7vvz2aygwwz7nyratwybz4zmm4u2hakcxej5vuhoyy.py
# Topologically Sorted Source Nodes: [logit_1], Original ATen: [aten.sub]
# Source node to ATen node mapping:
# logit_1 => sub
# Graph fragment:
# %sub : [num_users=2] = call_function[target=torch.ops.aten.sub.Tensor](args = (%squeeze, %expand_1), kwargs = {})
triton_poi_fused_sub_3 = async_compile.triton('triton_poi_fused_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=[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_sub_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_sub_3(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
tl.store(out_ptr0 + (x2), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/nl/cnla2cqa5fw6bip3brpt6zibngvgv3feeidyi5rvukf2am4vwmmr.py
# Topologically Sorted Source Nodes: [p], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# p => amax, exp, sub_1
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%sub, [1], True), kwargs = {})
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sub, %amax), kwargs = {})
# %exp : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_1,), kwargs = {})
triton_poi_fused__softmax_4 = async_compile.triton('triton_poi_fused__softmax_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=[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_4', '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_4(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_4/inductor_cache/qf/cqfbrg74qey6mqge6e7bnjziqndekcustllat4jl4solktk6wuem.py
# Topologically Sorted Source Nodes: [p], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# p => 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_5 = async_compile.triton('triton_poi_fused__softmax_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=[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_5', '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_5(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_4/inductor_cache/rz/crzpfhrntno5tklcxc2yic5zlon4af35hbgkpfwnp66i6nu3tdkz.py
# Topologically Sorted Source Nodes: [v, mul, h_1], Original ATen: [aten.clone, aten.mul, aten.sum]
# Source node to ATen node mapping:
# h_1 => sum_2
# mul => mul
# v => clone_1
# Graph fragment:
# %clone_1 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_1,), kwargs = {memory_format: torch.contiguous_format})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%clone_1, %expand_2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1], True), kwargs = {})
triton_poi_fused_clone_mul_sum_6 = async_compile.triton('triton_poi_fused_clone_mul_sum_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_clone_mul_sum_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_clone_mul_sum_6(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 + (4*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x2), xmask)
tmp4 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (32 + x2), xmask)
tmp8 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (48 + x2), xmask)
tmp12 = tl.load(in_ptr1 + (3 + (4*x1)), 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 + (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, primals_6, primals_7, primals_8, primals_9 = 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, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (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, ), (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: [k], 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: [], Original ATen: []
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, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_7, reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf2)
del primals_5
buf3 = reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [add, h], Original ATen: [aten.add, aten.tanh]
triton_poi_fused_add_tanh_1.run(buf3, primals_4, buf2, primals_6, 64, grid=grid(64), stream=stream0)
del primals_4
del primals_6
buf5 = reinterpret_tensor(buf2, (16, 1), (1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [y_1], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 1), (1, 4), 0), alpha=1, beta=1, out=buf5)
del primals_9
buf6 = empty_strided_cuda((4, 1), (1, 1), torch.int64)
# Topologically Sorted Source Nodes: [max_1], Original ATen: [aten.max]
triton_poi_fused_max_2.run(buf5, buf6, 4, grid=grid(4), stream=stream0)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [logit_1], Original ATen: [aten.sub]
triton_poi_fused_sub_3.run(buf5, buf7, 16, grid=grid(16), stream=stream0)
buf8 = reinterpret_tensor(buf5, (4, 4), (4, 1), 0); del buf5 # reuse
# Topologically Sorted Source Nodes: [p], Original ATen: [aten._softmax]
triton_poi_fused__softmax_4.run(buf7, buf8, 16, grid=grid(16), stream=stream0)
buf9 = buf7; del buf7 # reuse
# Topologically Sorted Source Nodes: [p], Original ATen: [aten._softmax]
triton_poi_fused__softmax_5.run(buf8, buf9, 16, grid=grid(16), stream=stream0)
buf10 = reinterpret_tensor(buf8, (4, 1, 4), (4, 4, 1), 0); del buf8 # reuse
# Topologically Sorted Source Nodes: [v, mul, h_1], Original ATen: [aten.clone, aten.mul, aten.sum]
triton_poi_fused_clone_mul_sum_6.run(primals_2, buf9, buf10, 16, grid=grid(16), stream=stream0)
return (reinterpret_tensor(buf10, (1, 4, 4), (4, 4, 1), 0), buf9, primals_2, primals_7, reinterpret_tensor(buf0, (16, 4), (4, 1), 0), buf3, buf6, buf9, primals_8, )
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, ), (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)
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, ), (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
from torch.autograd import Variable
import torch.nn.init
class KeyValueAttention(nn.Module):
def __init__(self, query_size, key_size, value_size, hid_size, init_range):
super(KeyValueAttention, self).__init__()
self.key2hid = nn.Linear(key_size, hid_size)
self.query2hid = nn.Linear(query_size, hid_size)
self.hid2output = nn.Linear(hid_size, 1)
self.key2hid.weight.data.uniform_(-init_range, init_range)
self.key2hid.bias.data.fill_(0)
self.query2hid.weight.data.uniform_(-init_range, init_range)
self.query2hid.bias.data.fill_(0)
self.hid2output.weight.data.uniform_(-init_range, init_range)
self.hid2output.bias.data.fill_(0)
def _bottle(self, linear, x):
y = linear(x.view(-1, x.size(-1)))
return y.view(x.size(0), x.size(1), -1)
def forward_attn(self, h):
logit = self.attn(h.view(-1, h.size(2))).view(h.size(0), h.size(1))
return logit
def forward(self, q, k, v, mask=None):
k = k.transpose(0, 1).contiguous()
v = v.transpose(0, 1).contiguous()
h_k = self._bottle(self.key2hid, k)
h_q = self.query2hid(q)
h = F.tanh(h_k + h_q.unsqueeze(1).expand_as(h_k))
logit = self._bottle(self.hid2output, h).squeeze(2)
logit = logit.sub(logit.max(1, keepdim=True)[0].expand_as(logit))
if mask is not None:
logit = torch.add(logit, Variable(mask))
p = F.softmax(logit)
w = p.unsqueeze(2).expand_as(v)
h = torch.sum(torch.mul(v, w), 1, keepdim=True)
h = h.transpose(0, 1).contiguous()
return h, p
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'query_size': 4, 'key_size': 4, 'value_size': 4,
'hid_size': 4, 'init_range': 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
import torch.nn.init
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_add_tanh_1(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
x3 = xindex
x0 = xindex % 4
x2 = xindex // 16
tmp0 = tl.load(in_out_ptr0 + x3, xmask)
tmp1 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (x0 + 4 * x2), xmask, eviction_policy='evict_last'
)
tmp4 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp5 = tmp3 + tmp4
tmp6 = tmp2 + tmp5
tmp7 = libdevice.tanh(tmp6)
tl.store(in_out_ptr0 + x3, tmp7, xmask)
@triton.jit
def triton_poi_fused_max_2(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 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp17 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp32 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
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)
tl.store(out_ptr0 + x0, tmp46, xmask)
@triton.jit
def triton_poi_fused_sub_3(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
tl.store(out_ptr0 + x2, tmp8, xmask)
@triton.jit
def triton_poi_fused__softmax_4(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_5(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_poi_fused_clone_mul_sum_6(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 + 4 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr0 + (16 + x2), xmask)
tmp4 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr0 + (32 + x2), xmask)
tmp8 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp11 = tl.load(in_ptr0 + (48 + x2), xmask)
tmp12 = tl.load(in_ptr1 + (3 + 4 * x1), 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 + x2, tmp14, 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), (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,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (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,), (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.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, 1), torch.float32)
extern_kernels.mm(primals_7, reinterpret_tensor(primals_5, (4, 4),
(1, 4), 0), out=buf2)
del primals_5
buf3 = reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0)
del buf1
triton_poi_fused_add_tanh_1[grid(64)](buf3, primals_4, buf2,
primals_6, 64, XBLOCK=64, num_warps=1, num_stages=1)
del primals_4
del primals_6
buf5 = reinterpret_tensor(buf2, (16, 1), (1, 1), 0)
del buf2
extern_kernels.addmm(primals_9, reinterpret_tensor(buf3, (16, 4), (
4, 1), 0), reinterpret_tensor(primals_8, (4, 1), (1, 4), 0),
alpha=1, beta=1, out=buf5)
del primals_9
buf6 = empty_strided_cuda((4, 1), (1, 1), torch.int64)
triton_poi_fused_max_2[grid(4)](buf5, buf6, 4, XBLOCK=4, num_warps=
1, num_stages=1)
buf7 = empty_strided_cuda((4, 4), (4, 1), torch.float32)
triton_poi_fused_sub_3[grid(16)](buf5, buf7, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf8 = reinterpret_tensor(buf5, (4, 4), (4, 1), 0)
del buf5
triton_poi_fused__softmax_4[grid(16)](buf7, buf8, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf9 = buf7
del buf7
triton_poi_fused__softmax_5[grid(16)](buf8, buf9, 16, XBLOCK=16,
num_warps=1, num_stages=1)
buf10 = reinterpret_tensor(buf8, (4, 1, 4), (4, 4, 1), 0)
del buf8
triton_poi_fused_clone_mul_sum_6[grid(16)](primals_2, buf9, buf10,
16, XBLOCK=16, num_warps=1, num_stages=1)
return reinterpret_tensor(buf10, (1, 4, 4), (4, 4, 1), 0
), buf9, primals_2, primals_7, reinterpret_tensor(buf0, (16, 4), (4,
1), 0), buf3, buf6, buf9, primals_8
class KeyValueAttentionNew(nn.Module):
def __init__(self, query_size, key_size, value_size, hid_size, init_range):
super(KeyValueAttentionNew, self).__init__()
self.key2hid = nn.Linear(key_size, hid_size)
self.query2hid = nn.Linear(query_size, hid_size)
self.hid2output = nn.Linear(hid_size, 1)
self.key2hid.weight.data.uniform_(-init_range, init_range)
self.key2hid.bias.data.fill_(0)
self.query2hid.weight.data.uniform_(-init_range, init_range)
self.query2hid.bias.data.fill_(0)
self.hid2output.weight.data.uniform_(-init_range, init_range)
self.hid2output.bias.data.fill_(0)
def _bottle(self, linear, x):
y = linear(x.view(-1, x.size(-1)))
return y.view(x.size(0), x.size(1), -1)
def forward_attn(self, h):
logit = self.attn(h.view(-1, h.size(2))).view(h.size(0), h.size(1))
return logit
def forward(self, input_0, input_1, input_2):
primals_3 = self.key2hid.weight
primals_4 = self.key2hid.bias
primals_5 = self.query2hid.weight
primals_6 = self.query2hid.bias
primals_8 = self.hid2output.weight
primals_9 = self.hid2output.bias
primals_7 = input_0
primals_1 = input_1
primals_2 = input_2
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]
| haojiepan1/CrossWOZ | KeyValueAttention | false | 6,811 | [
"Apache-2.0"
] | 1 | 6d7b4c4cfb73a528b76074764687906abecc90b6 | https://github.com/haojiepan1/CrossWOZ/tree/6d7b4c4cfb73a528b76074764687906abecc90b6 | import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
import torch.nn.init
class Model(nn.Module):
def __init__(self, query_size, key_size, value_size, hid_size, init_range):
super().__init__()
self.key2hid = nn.Linear(key_size, hid_size)
self.query2hid = nn.Linear(query_size, hid_size)
self.hid2output = nn.Linear(hid_size, 1)
self.key2hid.weight.data.uniform_(-init_range, init_range)
self.key2hid.bias.data.fill_(0)
self.query2hid.weight.data.uniform_(-init_range, init_range)
self.query2hid.bias.data.fill_(0)
self.hid2output.weight.data.uniform_(-init_range, init_range)
self.hid2output.bias.data.fill_(0)
def _bottle(self, linear, x):
y = linear(x.view(-1, x.size(-1)))
return y.view(x.size(0), x.size(1), -1)
def forward_attn(self, h):
logit = self.attn(h.view(-1, h.size(2))).view(h.size(0), h.size(1))
return logit
def forward(self, q, k, v, mask=None):
k = k.transpose(0, 1).contiguous()
v = v.transpose(0, 1).contiguous()
h_k = self._bottle(self.key2hid, k)
h_q = self.query2hid(q)
h = F.tanh(h_k + h_q.unsqueeze(1).expand_as(h_k))
logit = self._bottle(self.hid2output, h).squeeze(2)
logit = logit.sub(logit.max(1, keepdim=True)[0].expand_as(logit))
if mask is not None:
logit = torch.add(logit, Variable(mask))
p = F.softmax(logit)
w = p.unsqueeze(2).expand_as(v)
h = torch.sum(torch.mul(v, w), 1, keepdim=True)
h = h.transpose(0, 1).contiguous()
return h, p
def get_inputs():
return [torch.rand([4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'query_size': 4, 'key_size': 4, 'value_size': 4,
'hid_size': 4, 'init_range': 4}]
|
LandmarksLoss | # 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_4/inductor_cache/c7/cc7suqkqnonshdhuex2a5c5ohbijd6mcn2sk2ght6w3pguvyu6do.py
# Topologically Sorted Source Nodes: [mul, mul_1], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# mul => mul
# mul_1 => mul_1
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, %arg1_1), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%arg2_1, %arg1_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=[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_mul_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_mul_0(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
x0 = xindex
tmp0 = tl.load(in_ptr0 + (x0), xmask)
tmp1 = tl.load(in_ptr1 + (x0), xmask)
tmp3 = tl.load(in_ptr2 + (x0), xmask)
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tl.store(out_ptr0 + (x0), tmp2, xmask)
tl.store(out_ptr1 + (x0), tmp4, 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, 4), (64, 16, 4, 1), torch.float32)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [mul, mul_1], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(arg0_1, arg1_1, arg2_1, buf0, buf1, 256, grid=grid(256), stream=stream0)
del arg0_1
del arg1_1
del arg2_1
return (buf0, 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 numpy as np
import torch.nn as nn
import torch.utils.data
class WingLoss(nn.Module):
def __init__(self, w=10, e=2):
super(WingLoss, self).__init__()
self.w = w
self.e = e
self.C = self.w - self.w * np.log(1 + self.w / self.e)
def forward(self, x, t, sigma=1):
weight = torch.ones_like(t)
weight[torch.where(t == -1)] = 0
diff = weight * (x - t)
abs_diff = diff.abs()
flag = (abs_diff.data < self.w).float()
y = flag * self.w * torch.log(1 + abs_diff / self.e) + (1 - flag) * (
abs_diff - self.C)
return y.sum()
class LandmarksLoss(nn.Module):
def __init__(self, alpha=1.0):
super(LandmarksLoss, self).__init__()
self.loss_fcn = WingLoss()
self.alpha = alpha
def forward(self, pred, truel, mask):
loss = self.loss_fcn(pred * mask, truel * mask)
return loss / (torch.sum(mask) + 1e-13)
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
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
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, 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
x0 = xindex
tmp0 = tl.load(in_ptr0 + x0, xmask)
tmp1 = tl.load(in_ptr1 + x0, xmask)
tmp3 = tl.load(in_ptr2 + x0, xmask)
tmp2 = tmp0 * tmp1
tmp4 = tmp3 * tmp1
tl.store(out_ptr0 + x0, tmp2, xmask)
tl.store(out_ptr1 + x0, tmp4, 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, 4), (64, 16, 4, 1), torch.float32)
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(256)](arg0_1, arg1_1, arg2_1, buf0,
buf1, 256, XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
del arg1_1
del arg2_1
return buf0, buf1
class WingLoss(nn.Module):
def __init__(self, w=10, e=2):
super(WingLoss, self).__init__()
self.w = w
self.e = e
self.C = self.w - self.w * np.log(1 + self.w / self.e)
def forward(self, x, t, sigma=1):
weight = torch.ones_like(t)
weight[torch.where(t == -1)] = 0
diff = weight * (x - t)
abs_diff = diff.abs()
flag = (abs_diff.data < self.w).float()
y = flag * self.w * torch.log(1 + abs_diff / self.e) + (1 - flag) * (
abs_diff - self.C)
return y.sum()
class LandmarksLossNew(nn.Module):
def __init__(self, alpha=1.0):
super(LandmarksLossNew, self).__init__()
self.loss_fcn = WingLoss()
self.alpha = alpha
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]
| homomorfism/wise-programming | LandmarksLoss | false | 6,812 | [
"MIT"
] | 1 | e0589e8900237ddc9c3abf54c85be532cacf2d33 | https://github.com/homomorfism/wise-programming/tree/e0589e8900237ddc9c3abf54c85be532cacf2d33 | import torch
import numpy as np
import torch.nn as nn
import torch.utils.data
class WingLoss(nn.Module):
def __init__(self, w=10, e=2):
super().__init__()
self.w = w
self.e = e
self.C = self.w - self.w * np.log(1 + self.w / self.e)
def forward(self, x, t, sigma=1):
weight = torch.ones_like(t)
weight[torch.where(t == -1)] = 0
diff = weight * (x - t)
abs_diff = diff.abs()
flag = (abs_diff.data < self.w).float()
y = flag * self.w * torch.log(1 + abs_diff / self.e) + (1 - flag) * (
abs_diff - self.C)
return y.sum()
class Model(nn.Module):
def __init__(self, alpha=1.0):
super().__init__()
self.loss_fcn = WingLoss()
self.alpha = alpha
def forward(self, pred, truel, mask):
loss = self.loss_fcn(pred * mask, truel * mask)
return loss / (torch.sum(mask) + 1e-13)
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 []
|
Decoder1 | # 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_4/inductor_cache/32/c32omun6n6qdhq2gfcl5hdhgrjaib6kycryu7grqnm3rz3eayvzl.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.reflection_pad2d]
# Source node to ATen node mapping:
# out => _unsafe_index, _unsafe_index_1
# Graph fragment:
# %_unsafe_index : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%primals_1, [None, None, %sub_1, None]), kwargs = {})
# %_unsafe_index_1 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index, [None, None, None, %sub_1]), kwargs = {})
triton_poi_fused_reflection_pad2d_0 = async_compile.triton('triton_poi_fused_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=[16384],
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_reflection_pad2d_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_reflection_pad2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 9216
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 6
x1 = (xindex // 6) % 6
x2 = (xindex // 36)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (15 + ((-1)*(tl_math.abs((-3) + (tl_math.abs((-1) + x0))))) + ((-4)*(tl_math.abs((-3) + (tl_math.abs((-1) + x1))))) + (16*x2)), xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/er/ceran7cepmoot3unpkyiv6ws2wivez2mb54b3w5bjh4gxygcg23y.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 = (%_unsafe_index_1, %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 = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 16) % 3
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, 64, 4, 4), (1024, 16, 4, 1))
assert_size_stride(primals_2, (3, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_3, (3, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 64, 6, 6), (2304, 36, 6, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.reflection_pad2d]
stream0 = get_raw_stream(0)
triton_poi_fused_reflection_pad2d_0.run(primals_1, buf0, 9216, grid=grid(9216), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [out_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, 3, 4, 4), (48, 16, 4, 1))
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [out_1], Original ATen: [aten.convolution]
triton_poi_fused_convolution_1.run(buf2, primals_3, 192, grid=grid(192), 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, 64, 4, 4), (1024, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((3, 64, 3, 3), (576, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((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
class Decoder1(nn.Module):
def __init__(self):
super(Decoder1, self).__init__()
self.reflecPad2 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv3 = nn.Conv2d(64, 3, 3, 1, 0)
def forward(self, x):
out = self.reflecPad2(x)
out = self.conv3(out)
return out
def get_inputs():
return [torch.rand([4, 64, 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.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_reflection_pad2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK:
tl.constexpr):
xnumel = 9216
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 6
x1 = xindex // 6 % 6
x2 = xindex // 36
x3 = xindex
tmp0 = tl.load(in_ptr0 + (15 + -1 * tl_math.abs(-3 + tl_math.abs(-1 +
x0)) + -4 * tl_math.abs(-3 + tl_math.abs(-1 + x1)) + 16 * x2),
xmask, eviction_policy='evict_last')
tl.store(out_ptr0 + x3, tmp0, xmask)
@triton.jit
def triton_poi_fused_convolution_1(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 192
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 16 % 3
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, 64, 4, 4), (1024, 16, 4, 1))
assert_size_stride(primals_2, (3, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_3, (3,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 64, 6, 6), (2304, 36, 6, 1), torch.
float32)
get_raw_stream(0)
triton_poi_fused_reflection_pad2d_0[grid(9216)](primals_1, buf0,
9216, 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, 3, 4, 4), (48, 16, 4, 1))
buf2 = buf1
del buf1
triton_poi_fused_convolution_1[grid(192)](buf2, primals_3, 192,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_3
return buf2, primals_2, buf0
class Decoder1New(nn.Module):
def __init__(self):
super(Decoder1New, self).__init__()
self.reflecPad2 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv3 = nn.Conv2d(64, 3, 3, 1, 0)
def forward(self, input_0):
primals_2 = self.conv3.weight
primals_3 = self.conv3.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3])
return output[0]
| hologerry/wct_experiment | Decoder1 | false | 6,814 | [
"MIT"
] | 1 | 890d885561dc8df8c4ae732aebd902aa838257e6 | https://github.com/hologerry/wct_experiment/tree/890d885561dc8df8c4ae732aebd902aa838257e6 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self):
super().__init__()
self.reflecPad2 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv3 = nn.Conv2d(64, 3, 3, 1, 0)
def forward(self, x):
out = self.reflecPad2(x)
out = self.conv3(out)
return out
def get_inputs():
return [torch.rand([4, 64, 4, 4])]
def get_init_inputs():
return []
|
QuantMeasure | # 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_4/inductor_cache/2q/c2qfftq24ton4cb76uuegdwiejatkiqpir6ng7evufq2rm2ldwbg.py
# Topologically Sorted Source Nodes: [min_value, neg, add_, max_value, sub, scale, scale_1, div_, output_1, clamp_, output_2, add__2, mul_, output_3], Original ATen: [aten.min, aten.neg, aten.add, aten.max, aten.sub, aten.div, aten.clamp, aten.round, aten.mul]
# Source node to ATen node mapping:
# add_ => add
# add__2 => add_2
# clamp_ => clamp_max, clamp_min_1
# div_ => div_1
# max_value => max_1
# min_value => min_1
# mul_ => mul
# neg => neg
# output_1 => add_1
# output_2 => round_1
# output_3 => add_3
# scale => div
# scale_1 => clamp_min
# sub => sub
# Graph fragment:
# %min_1 : [num_users=3] = call_function[target=torch.ops.aten.min.default](args = (%arg0_1,), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%min_1,), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%arg0_1, %neg), kwargs = {})
# %max_1 : [num_users=1] = call_function[target=torch.ops.aten.max.default](args = (%arg0_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%max_1, %min_1), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, 255.0), kwargs = {})
# %clamp_min : [num_users=2] = call_function[target=torch.ops.aten.clamp_min.default](args = (%div, 1e-08), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%add, %clamp_min), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div_1, 0.0), kwargs = {})
# %clamp_min_1 : [num_users=1] = call_function[target=torch.ops.aten.clamp_min.default](args = (%add_1, 0.0), kwargs = {})
# %clamp_max : [num_users=1] = call_function[target=torch.ops.aten.clamp_max.default](args = (%clamp_min_1, 255.0), kwargs = {})
# %round_1 : [num_users=1] = call_function[target=torch.ops.aten.round.default](args = (%clamp_max,), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%round_1, -0.0), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_2, %clamp_min), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, %min_1), kwargs = {})
triton_per_fused_add_clamp_div_max_min_mul_neg_round_sub_0 = async_compile.triton('triton_per_fused_add_clamp_div_max_min_mul_neg_round_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_add_clamp_div_max_min_mul_neg_round_sub_0', 'mutated_arg_names': [], 'no_x_dim': True, 'num_load': 1, '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_add_clamp_div_max_min_mul_neg_round_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 = triton_helpers.promote_to_tensor(triton_helpers.min2(tmp1, 0))
tmp5 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp1, 0))
tmp6 = -tmp3
tmp7 = tmp0 + tmp6
tmp8 = tmp5 - tmp3
tmp9 = 0.00392156862745098
tmp10 = tmp8 * tmp9
tmp11 = 1e-08
tmp12 = triton_helpers.maximum(tmp10, tmp11)
tmp13 = tmp7 / tmp12
tmp14 = 0.0
tmp15 = tmp13 + tmp14
tmp16 = triton_helpers.maximum(tmp15, tmp14)
tmp17 = 255.0
tmp18 = triton_helpers.minimum(tmp16, tmp17)
tmp19 = libdevice.nearbyint(tmp18)
tmp20 = -0.0
tmp21 = tmp19 + tmp20
tmp22 = tmp21 * tmp12
tmp23 = tmp22 + tmp3
tl.store(out_ptr2 + (tl.broadcast_to(r0, [RBLOCK])), tmp23, 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)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [min_value, neg, add_, max_value, sub, scale, scale_1, div_, output_1, clamp_, output_2, add__2, mul_, output_3], Original ATen: [aten.min, aten.neg, aten.add, aten.max, aten.sub, aten.div, aten.clamp, aten.round, aten.mul]
stream0 = get_raw_stream(0)
triton_per_fused_add_clamp_div_max_min_mul_neg_round_sub_0.run(arg0_1, buf2, 1, 256, grid=grid(1), 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
from torch import nn
from torch.autograd.function import InplaceFunction
def quantize(x, num_bits=8, min_value=None, max_value=None, num_chunks=None,
stochastic=False, inplace=False, quantize=False, layer_num=-1, multi=
False, index=[], is_act=False):
return UniformQuantize().apply(x, num_bits, min_value, max_value,
stochastic, inplace, num_chunks, False, quantize, layer_num, multi,
index, is_act)
class UniformQuantize(InplaceFunction):
@classmethod
def forward(cls, ctx, input, num_bits=8, min_value=None, max_value=None,
stochastic=False, inplace=False, num_chunks=None, out_half=False,
quantize=False, layer_num=-1, multi=False, index=[], is_act=False):
if is_act:
multi = False
num_chunks = num_chunks = input.shape[0
] if num_chunks is None else num_chunks
if min_value is None or max_value is None:
B = input.shape[0]
y = input.view(B // num_chunks, -1)
if min_value is None:
min_value = y.min(-1)[0].mean(-1)
if max_value is None:
max_value = y.max(-1)[0].mean(-1)
ctx.inplace = inplace
ctx.num_bits = num_bits
ctx.min_value = min_value
ctx.max_value = max_value
ctx.stochastic = stochastic
if ctx.inplace:
ctx.mark_dirty(input)
output = input
else:
output = input.clone()
if multi:
bit_max = 8
for i in range(bit_max):
if len(index[layer_num][i]) == 0:
continue
else:
idx = index[layer_num][i]
min_value = output[idx].min()
max_value = output[idx].max()
qmin = 0.0
qmax = 2.0 ** (1 + i) - 1.0
scale = (max_value - min_value) / (qmax - qmin)
scale = max(scale, 1e-08)
output[idx] = output[idx].add_(-min_value).div_(scale).add_(
qmin)
output[idx] = output[idx].clamp_(qmin, qmax).round_()
output[idx] = output[idx].add_(-qmin).mul_(scale).add_(
min_value)
else:
min_value = output.min()
max_value = output.max()
qmin = 0.0
qmax = 2.0 ** num_bits - 1.0
scale = (max_value - min_value) / (qmax - qmin)
scale = max(scale, 1e-08)
output = output.add_(-min_value).div_(scale).add_(qmin)
output = output.clamp_(qmin, qmax).round_()
output = output.add_(-qmin).mul_(scale).add_(min_value)
return output
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output
return (grad_input, None, None, None, None, None, None, None, None,
None, None, None, None)
class QuantMeasure(nn.Module):
"""docstring for QuantMeasure."""
def __init__(self, num_bits=8, quantize=False, momentum=0.1):
super(QuantMeasure, self).__init__()
self.register_buffer('running_min', torch.zeros(1))
self.register_buffer('running_max', torch.zeros(1))
self.momentum = momentum
self.num_bits = num_bits
self.quantize = quantize
def forward(self, input):
if self.training:
min_value = input.detach().view(input.size(0), -1).min(-1)[0].mean(
)
max_value = input.detach().view(input.size(0), -1).max(-1)[0].mean(
)
self.running_min.mul_(self.momentum).add_(min_value * (1 - self
.momentum))
self.running_max.mul_(self.momentum).add_(max_value * (1 - self
.momentum))
else:
min_value = self.running_min
max_value = self.running_max
return quantize(input, self.num_bits, min_value=float(min_value),
max_value=float(max_value), num_chunks=16, quantize=self.
quantize, is_act=True)
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
from torch import nn
from torch.autograd.function import InplaceFunction
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_clamp_div_max_min_mul_neg_round_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 = triton_helpers.promote_to_tensor(triton_helpers.min2(tmp1, 0))
tmp5 = triton_helpers.promote_to_tensor(triton_helpers.max2(tmp1, 0))
tmp6 = -tmp3
tmp7 = tmp0 + tmp6
tmp8 = tmp5 - tmp3
tmp9 = 0.00392156862745098
tmp10 = tmp8 * tmp9
tmp11 = 1e-08
tmp12 = triton_helpers.maximum(tmp10, tmp11)
tmp13 = tmp7 / tmp12
tmp14 = 0.0
tmp15 = tmp13 + tmp14
tmp16 = triton_helpers.maximum(tmp15, tmp14)
tmp17 = 255.0
tmp18 = triton_helpers.minimum(tmp16, tmp17)
tmp19 = libdevice.nearbyint(tmp18)
tmp20 = -0.0
tmp21 = tmp19 + tmp20
tmp22 = tmp21 * tmp12
tmp23 = tmp22 + tmp3
tl.store(out_ptr2 + tl.broadcast_to(r0, [RBLOCK]), tmp23, 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)
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused_add_clamp_div_max_min_mul_neg_round_sub_0[grid(1)](
arg0_1, buf2, 1, 256, num_warps=2, num_stages=1)
del arg0_1
return buf2,
def quantize(x, num_bits=8, min_value=None, max_value=None, num_chunks=None,
stochastic=False, inplace=False, quantize=False, layer_num=-1, multi=
False, index=[], is_act=False):
return UniformQuantize().apply(x, num_bits, min_value, max_value,
stochastic, inplace, num_chunks, False, quantize, layer_num, multi,
index, is_act)
class UniformQuantize(InplaceFunction):
@classmethod
def forward(cls, ctx, input, num_bits=8, min_value=None, max_value=None,
stochastic=False, inplace=False, num_chunks=None, out_half=False,
quantize=False, layer_num=-1, multi=False, index=[], is_act=False):
if is_act:
multi = False
num_chunks = num_chunks = input.shape[0
] if num_chunks is None else num_chunks
if min_value is None or max_value is None:
B = input.shape[0]
y = input.view(B // num_chunks, -1)
if min_value is None:
min_value = y.min(-1)[0].mean(-1)
if max_value is None:
max_value = y.max(-1)[0].mean(-1)
ctx.inplace = inplace
ctx.num_bits = num_bits
ctx.min_value = min_value
ctx.max_value = max_value
ctx.stochastic = stochastic
if ctx.inplace:
ctx.mark_dirty(input)
output = input
else:
output = input.clone()
if multi:
bit_max = 8
for i in range(bit_max):
if len(index[layer_num][i]) == 0:
continue
else:
idx = index[layer_num][i]
min_value = output[idx].min()
max_value = output[idx].max()
qmin = 0.0
qmax = 2.0 ** (1 + i) - 1.0
scale = (max_value - min_value) / (qmax - qmin)
scale = max(scale, 1e-08)
output[idx] = output[idx].add_(-min_value).div_(scale).add_(
qmin)
output[idx] = output[idx].clamp_(qmin, qmax).round_()
output[idx] = output[idx].add_(-qmin).mul_(scale).add_(
min_value)
else:
min_value = output.min()
max_value = output.max()
qmin = 0.0
qmax = 2.0 ** num_bits - 1.0
scale = (max_value - min_value) / (qmax - qmin)
scale = max(scale, 1e-08)
output = output.add_(-min_value).div_(scale).add_(qmin)
output = output.clamp_(qmin, qmax).round_()
output = output.add_(-qmin).mul_(scale).add_(min_value)
return output
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output
return (grad_input, None, None, None, None, None, None, None, None,
None, None, None, None)
class QuantMeasureNew(nn.Module):
"""docstring for QuantMeasure."""
def __init__(self, num_bits=8, quantize=False, momentum=0.1):
super(QuantMeasureNew, self).__init__()
self.register_buffer('running_min', torch.zeros(1))
self.register_buffer('running_max', torch.zeros(1))
self.momentum = momentum
self.num_bits = num_bits
self.quantize = quantize
def forward(self, input_0):
arg0_1 = input_0
output = call([arg0_1])
return output[0]
| hoseung2/DNAS-Compression | QuantMeasure | false | 6,815 | [
"MIT"
] | 1 | 645407fc572045f33278c935091a07e0ccfce87f | https://github.com/hoseung2/DNAS-Compression/tree/645407fc572045f33278c935091a07e0ccfce87f | import torch
from torch import nn
from torch.autograd.function import InplaceFunction
def quantize(x, num_bits=8, min_value=None, max_value=None, num_chunks=None,
stochastic=False, inplace=False, quantize=False, layer_num=-1, multi=
False, index=[], is_act=False):
return UniformQuantize().apply(x, num_bits, min_value, max_value,
stochastic, inplace, num_chunks, False, quantize, layer_num, multi,
index, is_act)
class UniformQuantize(InplaceFunction):
@classmethod
def forward(cls, ctx, input, num_bits=8, min_value=None, max_value=None,
stochastic=False, inplace=False, num_chunks=None, out_half=False,
quantize=False, layer_num=-1, multi=False, index=[], is_act=False):
if is_act:
multi = False
num_chunks = num_chunks = input.shape[0
] if num_chunks is None else num_chunks
if min_value is None or max_value is None:
B = input.shape[0]
y = input.view(B // num_chunks, -1)
if min_value is None:
min_value = y.min(-1)[0].mean(-1)
if max_value is None:
max_value = y.max(-1)[0].mean(-1)
ctx.inplace = inplace
ctx.num_bits = num_bits
ctx.min_value = min_value
ctx.max_value = max_value
ctx.stochastic = stochastic
if ctx.inplace:
ctx.mark_dirty(input)
output = input
else:
output = input.clone()
if multi:
bit_max = 8
for i in range(bit_max):
if len(index[layer_num][i]) == 0:
continue
else:
idx = index[layer_num][i]
min_value = output[idx].min()
max_value = output[idx].max()
qmin = 0.0
qmax = 2.0 ** (1 + i) - 1.0
scale = (max_value - min_value) / (qmax - qmin)
scale = max(scale, 1e-08)
output[idx] = output[idx].add_(-min_value).div_(scale).add_(
qmin)
output[idx] = output[idx].clamp_(qmin, qmax).round_()
output[idx] = output[idx].add_(-qmin).mul_(scale).add_(
min_value)
else:
min_value = output.min()
max_value = output.max()
qmin = 0.0
qmax = 2.0 ** num_bits - 1.0
scale = (max_value - min_value) / (qmax - qmin)
scale = max(scale, 1e-08)
output = output.add_(-min_value).div_(scale).add_(qmin)
output = output.clamp_(qmin, qmax).round_()
output = output.add_(-qmin).mul_(scale).add_(min_value)
return output
@staticmethod
def backward(ctx, grad_output):
grad_input = grad_output
return (grad_input, None, None, None, None, None, None, None, None,
None, None, None, None)
class Model(nn.Module):
"""docstring for QuantMeasure."""
def __init__(self, num_bits=8, quantize=False, momentum=0.1):
super().__init__()
self.register_buffer('running_min', torch.zeros(1))
self.register_buffer('running_max', torch.zeros(1))
self.momentum = momentum
self.num_bits = num_bits
self.quantize = quantize
def forward(self, input):
if self.training:
min_value = input.detach().view(input.size(0), -1).min(-1)[0].mean(
)
max_value = input.detach().view(input.size(0), -1).max(-1)[0].mean(
)
self.running_min.mul_(self.momentum).add_(min_value * (1 - self
.momentum))
self.running_max.mul_(self.momentum).add_(max_value * (1 - self
.momentum))
else:
min_value = self.running_min
max_value = self.running_max
return quantize(input, self.num_bits, min_value=float(min_value),
max_value=float(max_value), num_chunks=16, quantize=self.
quantize, is_act=
# ... truncated (>4000 chars) for memory efficiency |
SmooothLabelCELoss | # 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_4/inductor_cache/kb/ckb5dxt7bl4nrv3xbxip5db57t6yqtx6yshfe4o5a4q6qxq6kusa.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.div]
# Source node to ATen node mapping:
# out => full_default
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4], 0.02500000037252903), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
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=[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_div_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_div_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
tmp0 = 0.02500000037252903
tl.store(out_ptr0 + (x0), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ov/covxwftg3izjg5actw2wvxpitaiwk6gdgw3gixdqaa2jdyz4zuhd.py
# Topologically Sorted Source Nodes: [out, setitem], Original ATen: [aten.div, aten.lift_fresh, aten.index_put]
# Source node to ATen node mapping:
# out => full_default
# setitem => full_default_1, index_put
# Graph fragment:
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4], 0.02500000037252903), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %full_default_1 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0.925000011920929), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %index_put : [num_users=5] = call_function[target=torch.ops.aten.index_put_.default](args = (%full_default, [%iota_default, %arg1_1], %full_default_1), kwargs = {})
triton_poi_fused_div_index_put_lift_fresh_1 = async_compile.triton('triton_poi_fused_div_index_put_lift_fresh_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: '*i64', 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_div_index_put_lift_fresh_1', 'mutated_arg_names': ['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_div_index_put_lift_fresh_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 = tl.full([XBLOCK], 4, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tl.device_assert(((0 <= tmp4) & (tmp4 < 4)) | ~(xmask), "index out of bounds: 0 <= tmp4 < 4")
tmp6 = 0.925000011920929
tl.store(out_ptr0 + (tmp4 + (4*x0)), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/pv/cpvkbqma42nsdzqpqpkanztnk3kwdfx7iozdpzeynueg2vowikim.py
# Topologically Sorted Source Nodes: [output], Original ATen: [aten._log_softmax]
# Source node to ATen node mapping:
# output => 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_2 = async_compile.triton('triton_poi_fused__log_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__log_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__log_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 = 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_4/inductor_cache/vu/cvud7ahmbxwqwxoogftz5uvfsgc76moedr4ma4x2m7tcie7qhfs3.py
# Topologically Sorted Source Nodes: [loss, output], Original ATen: [aten.xlogy, aten._log_softmax, aten.mul, aten.sub, aten.sum, aten.div]
# Source node to ATen node mapping:
# loss => div_1, eq, full_default_2, full_default_3, isnan, log_1, mul_2, mul_3, sub_2, sum_2, where, where_1
# output => exp, log, sub_1, sum_1
# Graph fragment:
# %isnan : [num_users=1] = call_function[target=torch.ops.aten.isnan.default](args = (%index_put,), kwargs = {})
# %full_default_3 : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], nan), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %eq : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%index_put, 0), kwargs = {})
# %full_default_2 : [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})
# %log_1 : [num_users=1] = call_function[target=torch.ops.aten.log.default](args = (%index_put,), kwargs = {})
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%index_put, %log_1), kwargs = {})
# %where : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%eq, %full_default_2, %mul_3), kwargs = {})
# %where_1 : [num_users=1] = call_function[target=torch.ops.aten.where.self](args = (%isnan, %full_default_3, %where), kwargs = {})
# %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_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%index_put, %sub_1), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%where_1, %mul_2), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.default](args = (%sub_2,), kwargs = {})
# %div_1 : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%sum_2, 4), kwargs = {})
triton_per_fused__log_softmax_div_mul_sub_sum_xlogy_3 = async_compile.triton('triton_per_fused__log_softmax_div_mul_sub_sum_xlogy_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=[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__log_softmax_div_mul_sub_sum_xlogy_3', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, 'num_load': 6, '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_div_mul_sub_sum_xlogy_3(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 % 16
r3 = rindex
r2 = (rindex // 64)
tmp0 = tl.load(in_ptr0 + (r0), None, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + (r3), None)
tmp10 = tl.load(in_ptr1 + (r0 + (64*r2)), None, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (16 + r0 + (64*r2)), None, eviction_policy='evict_last')
tmp15 = tl.load(in_ptr1 + (32 + r0 + (64*r2)), None, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr1 + (48 + r0 + (64*r2)), None, eviction_policy='evict_last')
tmp1 = libdevice.isnan(tmp0).to(tl.int1)
tmp2 = 0.0
tmp3 = tmp0 == tmp2
tmp4 = tl_math.log(tmp0)
tmp5 = tmp0 * tmp4
tmp6 = tl.where(tmp3, tmp2, tmp5)
tmp7 = float("nan")
tmp8 = tl.where(tmp1, tmp7, tmp6)
tmp11 = tl_math.exp(tmp10)
tmp13 = tl_math.exp(tmp12)
tmp14 = tmp11 + tmp13
tmp16 = tl_math.exp(tmp15)
tmp17 = tmp14 + tmp16
tmp19 = tl_math.exp(tmp18)
tmp20 = tmp17 + tmp19
tmp21 = tl_math.log(tmp20)
tmp22 = tmp9 - tmp21
tmp23 = tmp0 * tmp22
tmp24 = tmp8 - tmp23
tmp25 = tl.broadcast_to(tmp24, [RBLOCK])
tmp27 = triton_helpers.promote_to_tensor(tl.sum(tmp25, 0))
tmp28 = 0.25
tmp29 = tmp27 * tmp28
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp29, 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, ), (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: [out], Original ATen: [aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_div_0.run(buf0, 16, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [out, setitem], Original ATen: [aten.div, aten.lift_fresh, aten.index_put]
triton_poi_fused_div_index_put_lift_fresh_1.run(arg1_1, buf0, 4, grid=grid(4), stream=stream0)
del arg1_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [output], Original ATen: [aten._log_softmax]
triton_poi_fused__log_softmax_2.run(arg0_1, buf2, 256, grid=grid(256), stream=stream0)
del arg0_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf4 = buf3; del buf3 # reuse
# Topologically Sorted Source Nodes: [loss, output], Original ATen: [aten.xlogy, aten._log_softmax, aten.mul, aten.sub, aten.sum, aten.div]
triton_per_fused__log_softmax_div_mul_sub_sum_xlogy_3.run(buf4, buf0, buf2, 1, 256, grid=grid(1), stream=stream0)
del buf0
del buf2
return (buf4, )
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, ), (1, ), device='cuda:0', dtype=torch.int64)
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
class SmooothLabelCELoss(nn.Module):
def __init__(self, smooth=0.1, use_uniform=False, reduction='mean'):
super(SmooothLabelCELoss, self).__init__()
self.smooth_coef = smooth
self.smooth_std = 0.5
self.reduction = reduction
self.use_uniform = use_uniform
self.intervals = 5
self._loss = nn.KLDivLoss(reduction='batchmean')
self.norm = nn.LogSoftmax(dim=1)
def get_closest_number(self, y, C, num):
assert num <= C
half = num // 2
if y + half < C:
left = max(y - half, 0)
right = left + num
else:
right = C
left = right - num
return left, right
def center_smooth_label(self, target, C):
B = target.shape[0]
I = torch.empty(size=(B, self.intervals), device=target.device,
dtype=torch.int64)
for b in range(B):
left, right = self.get_closest_number(target[b].item(), C, self
.intervals)
I[b, :] = torch.arange(left, right)
softed = torch.zeros((B, C), dtype=torch.float32, device=target.device)
use_uniform = self.use_uniform
if use_uniform:
softed.scatter_(1, I, self.smooth_coef * 1.0 / self.intervals)
softed[torch.arange(0, B), target
] = 1 - self.smooth_coef + self.smooth_coef * 1.0 / self.intervals
else:
src = torch.exp(-self.smooth_std * (I - target.unsqueeze(dim=1)
) ** 2)
softed.scatter_(1, I, src)
softed = softed / softed.sum(1, keepdim=True)
return softed
def global_smooth_label(self, target, C):
B = target.shape[0]
if C is None:
C = target.max() + 1
out = torch.ones(B, C, device=target.device
) * self.smooth_coef * 1.0 / C
out[torch.arange(0, B), target
] = 1 - self.smooth_coef + self.smooth_coef * 1.0 / C
return out
def forward(self, output, label):
output = self.norm(output)
C = output.shape[1]
soft_label = self.global_smooth_label(label, C)
loss = self._loss(output, soft_label)
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.ones([4], dtype=torch.int64)]
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, 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_div_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
tmp0 = 0.02500000037252903
tl.store(out_ptr0 + x0, tmp0, xmask)
@triton.jit
def triton_poi_fused_div_index_put_lift_fresh_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 = tl.full([XBLOCK], 4, tl.int32)
tmp2 = tmp0 + tmp1
tmp3 = tmp0 < 0
tmp4 = tl.where(tmp3, tmp2, tmp0)
tl.device_assert((0 <= tmp4) & (tmp4 < 4) | ~xmask,
'index out of bounds: 0 <= tmp4 < 4')
tmp6 = 0.925000011920929
tl.store(out_ptr0 + (tmp4 + 4 * x0), tmp6, xmask)
@triton.jit
def triton_poi_fused__log_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 = 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_div_mul_sub_sum_xlogy_3(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 % 16
r3 = rindex
r2 = rindex // 64
tmp0 = tl.load(in_ptr0 + r0, None, eviction_policy='evict_last')
tmp9 = tl.load(in_ptr1 + r3, None)
tmp10 = tl.load(in_ptr1 + (r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp12 = tl.load(in_ptr1 + (16 + r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp15 = tl.load(in_ptr1 + (32 + r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp18 = tl.load(in_ptr1 + (48 + r0 + 64 * r2), None, eviction_policy=
'evict_last')
tmp1 = libdevice.isnan(tmp0).to(tl.int1)
tmp2 = 0.0
tmp3 = tmp0 == tmp2
tmp4 = tl_math.log(tmp0)
tmp5 = tmp0 * tmp4
tmp6 = tl.where(tmp3, tmp2, tmp5)
tmp7 = float('nan')
tmp8 = tl.where(tmp1, tmp7, tmp6)
tmp11 = tl_math.exp(tmp10)
tmp13 = tl_math.exp(tmp12)
tmp14 = tmp11 + tmp13
tmp16 = tl_math.exp(tmp15)
tmp17 = tmp14 + tmp16
tmp19 = tl_math.exp(tmp18)
tmp20 = tmp17 + tmp19
tmp21 = tl_math.log(tmp20)
tmp22 = tmp9 - tmp21
tmp23 = tmp0 * tmp22
tmp24 = tmp8 - tmp23
tmp25 = tl.broadcast_to(tmp24, [RBLOCK])
tmp27 = triton_helpers.promote_to_tensor(tl.sum(tmp25, 0))
tmp28 = 0.25
tmp29 = tmp27 * tmp28
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp29, 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,), (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_div_0[grid(16)](buf0, 16, XBLOCK=16, num_warps=1,
num_stages=1)
triton_poi_fused_div_index_put_lift_fresh_1[grid(4)](arg1_1, buf0,
4, XBLOCK=4, num_warps=1, num_stages=1)
del arg1_1
buf2 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__log_softmax_2[grid(256)](arg0_1, buf2, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del arg0_1
buf3 = empty_strided_cuda((), (), torch.float32)
buf4 = buf3
del buf3
triton_per_fused__log_softmax_div_mul_sub_sum_xlogy_3[grid(1)](buf4,
buf0, buf2, 1, 256, num_warps=2, num_stages=1)
del buf0
del buf2
return buf4,
class SmooothLabelCELossNew(nn.Module):
def __init__(self, smooth=0.1, use_uniform=False, reduction='mean'):
super(SmooothLabelCELossNew, self).__init__()
self.smooth_coef = smooth
self.smooth_std = 0.5
self.reduction = reduction
self.use_uniform = use_uniform
self.intervals = 5
self._loss = nn.KLDivLoss(reduction='batchmean')
self.norm = nn.LogSoftmax(dim=1)
def get_closest_number(self, y, C, num):
assert num <= C
half = num // 2
if y + half < C:
left = max(y - half, 0)
right = left + num
else:
right = C
left = right - num
return left, right
def center_smooth_label(self, target, C):
B = target.shape[0]
I = torch.empty(size=(B, self.intervals), device=target.device,
dtype=torch.int64)
for b in range(B):
left, right = self.get_closest_number(target[b].item(), C, self
.intervals)
I[b, :] = torch.arange(left, right)
softed = torch.zeros((B, C), dtype=torch.float32, device=target.device)
use_uniform = self.use_uniform
if use_uniform:
softed.scatter_(1, I, self.smooth_coef * 1.0 / self.intervals)
softed[torch.arange(0, B), target
] = 1 - self.smooth_coef + self.smooth_coef * 1.0 / self.intervals
else:
src = torch.exp(-self.smooth_std * (I - target.unsqueeze(dim=1)
) ** 2)
softed.scatter_(1, I, src)
softed = softed / softed.sum(1, keepdim=True)
return softed
def global_smooth_label(self, target, C):
B = target.shape[0]
if C is None:
C = target.max() + 1
out = torch.ones(B, C, device=target.device
) * self.smooth_coef * 1.0 / C
out[torch.arange(0, B), target
] = 1 - self.smooth_coef + self.smooth_coef * 1.0 / C
return out
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| hekq/3DFingerPose | SmooothLabelCELoss | false | 6,816 | [
"MIT"
] | 1 | 385c672408e2fd29ed0373a842727c9fcfd0fc59 | https://github.com/hekq/3DFingerPose/tree/385c672408e2fd29ed0373a842727c9fcfd0fc59 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, smooth=0.1, use_uniform=False, reduction='mean'):
super().__init__()
self.smooth_coef = smooth
self.smooth_std = 0.5
self.reduction = reduction
self.use_uniform = use_uniform
self.intervals = 5
self._loss = nn.KLDivLoss(reduction='batchmean')
self.norm = nn.LogSoftmax(dim=1)
def get_closest_number(self, y, C, num):
assert num <= C
half = num // 2
if y + half < C:
left = max(y - half, 0)
right = left + num
else:
right = C
left = right - num
return left, right
def center_smooth_label(self, target, C):
B = target.shape[0]
I = torch.empty(size=(B, self.intervals), device=target.device,
dtype=torch.int64)
for b in range(B):
left, right = self.get_closest_number(target[b].item(), C, self
.intervals)
I[b, :] = torch.arange(left, right)
softed = torch.zeros((B, C), dtype=torch.float32, device=target.device)
use_uniform = self.use_uniform
if use_uniform:
softed.scatter_(1, I, self.smooth_coef * 1.0 / self.intervals)
softed[torch.arange(0, B), target
] = 1 - self.smooth_coef + self.smooth_coef * 1.0 / self.intervals
else:
src = torch.exp(-self.smooth_std * (I - target.unsqueeze(dim=1)
) ** 2)
softed.scatter_(1, I, src)
softed = softed / softed.sum(1, keepdim=True)
return softed
def global_smooth_label(self, target, C):
B = target.shape[0]
if C is None:
C = target.max() + 1
out = torch.ones(B, C, device=target.device
) * self.smooth_coef * 1.0 / C
out[torch.arange(0, B), target
] = 1 - self.smooth_coef + self.smooth_coef * 1.0 / C
return out
def forward(self, output, label):
output = self.norm(output)
C = output.shape[1]
soft_label = self.global_smooth_label(label, C)
loss = self._loss(output, soft_label)
return loss
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.ones([4], dtype=torch.int64)]
def get_init_inputs():
return []
|
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_4/inductor_cache/qz/cqza6p5fjiie2hfiu5dfjqqugrnzziwuwxzlhzy2aa7khopxjbym.py
# Topologically Sorted Source Nodes: [predict], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# predict => 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=[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
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_4/inductor_cache/v4/cv4nyn2kde7dd2c53ddahw4vtxyldln6pqt62jrliqindkf3sj5m.py
# Topologically Sorted Source Nodes: [predict], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# predict => 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=[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 = 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_4/inductor_cache/7e/c7e5xjhhnst5z5rbsmctwrwgiwdqyubsrbgy2ujvvf6nflma2ws7.py
# Topologically Sorted Source Nodes: [mul, sum_1, pow_1, pow_2, add_1, sum_2], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
# Source node to ATen node mapping:
# add_1 => add_1
# mul => mul
# pow_1 => pow_1
# pow_2 => pow_2
# sum_1 => sum_2
# sum_2 => sum_3
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view, %view_1), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [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 = (%view_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 = {})
triton_per_fused_add_mul_pow_sum_2 = async_compile.triton('triton_per_fused_add_mul_pow_sum_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=[4, 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, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mul_pow_sum_2', 'mutated_arg_names': [], '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_add_mul_pow_sum_2(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
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 + (64*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (r1 + (64*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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/bf/cbfvj76s5xemb4bs7kwt3su5gmofkie26hpppsikx7ik7sbfiunm.py
# Topologically Sorted Source Nodes: [mul_3, sum_7, pow_7, pow_8, add_11, sum_8], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
# Source node to ATen node mapping:
# add_11 => add_13
# mul_3 => mul_3
# pow_7 => pow_7
# pow_8 => pow_8
# sum_7 => sum_8
# sum_8 => sum_9
# Graph fragment:
# %mul_3 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_6, %view_7), kwargs = {})
# %sum_8 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_3, [1]), kwargs = {})
# %pow_7 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_6, 2), kwargs = {})
# %pow_8 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_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 = {})
triton_per_fused_add_mul_pow_sum_3 = async_compile.triton('triton_per_fused_add_mul_pow_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=[4, 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, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mul_pow_sum_3', 'mutated_arg_names': [], '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_add_mul_pow_sum_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
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 + (48 + r1 + (64*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (48 + r1 + (64*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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/u6/cu6mzldqxg26sec4gcx3r552m2kvchtopdml3edcvo2in2ct7bjg.py
# Topologically Sorted Source Nodes: [mul_1, sum_3, pow_3, pow_4, add_5, sum_4], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
# Source node to ATen node mapping:
# add_5 => add_5
# mul_1 => mul_1
# pow_3 => pow_3
# pow_4 => pow_4
# sum_3 => sum_4
# sum_4 => sum_5
# Graph fragment:
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_2, %view_3), kwargs = {})
# %sum_4 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_1, [1]), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_2, 2), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_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 = {})
triton_per_fused_add_mul_pow_sum_4 = async_compile.triton('triton_per_fused_add_mul_pow_sum_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.persistent_reduction(
size_hints=[4, 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, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mul_pow_sum_4', 'mutated_arg_names': [], '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_add_mul_pow_sum_4(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
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 + (16 + r1 + (64*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (16 + r1 + (64*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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/6m/c6mgqj2hrkokpnq6dgwmadtxpefexwavsviyw5qzkfpmafiwfc3s.py
# Topologically Sorted Source Nodes: [mul_2, sum_5, pow_5, pow_6, add_8, sum_6], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
# Source node to ATen node mapping:
# add_8 => add_9
# mul_2 => mul_2
# pow_5 => pow_5
# pow_6 => pow_6
# sum_5 => sum_6
# sum_6 => sum_7
# Graph fragment:
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%view_4, %view_5), kwargs = {})
# %sum_6 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul_2, [1]), kwargs = {})
# %pow_5 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_4, 2), kwargs = {})
# %pow_6 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%view_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 = {})
triton_per_fused_add_mul_pow_sum_5 = async_compile.triton('triton_per_fused_add_mul_pow_sum_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=[4, 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, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_mul_pow_sum_5', 'mutated_arg_names': [], '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_add_mul_pow_sum_5(in_ptr0, in_ptr1, out_ptr0, out_ptr1, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
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 + (32 + r1 + (64*x0)), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (32 + r1 + (64*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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + (x0), tmp6, xmask)
tl.store(out_ptr1 + (x0), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/2s/c2so3mbaiq42vb4cvpgytop3adq2nepk75nyco32vv2evdtdbz33.py
# Topologically Sorted Source Nodes: [num, den, truediv, loss, dice_loss, total_loss, num_1, den_1, truediv_1, loss_1, dice_loss_1, total_loss_1, num_2, den_2, truediv_2, loss_2, dice_loss_2, total_loss_2, num_3, den_3, truediv_3, loss_3, dice_loss_3, total_loss_3, truediv_4], Original ATen: [aten.add, aten.div, aten.rsub, aten.mean]
# Source node to ATen node mapping:
# den => add_2
# den_1 => add_6
# den_2 => add_10
# den_3 => add_14
# dice_loss => mean
# dice_loss_1 => mean_1
# dice_loss_2 => mean_2
# dice_loss_3 => mean_3
# loss => sub_1
# loss_1 => sub_2
# loss_2 => sub_3
# loss_3 => sub_4
# num => add
# num_1 => add_4
# num_2 => add_8
# num_3 => add_12
# 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
# truediv_4 => div_5
# Graph fragment:
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_2, 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 = {})
# %add_4 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_4, 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 = {})
# %add_8 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_6, 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 = {})
# %add_12 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_8, 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 = {})
triton_per_fused_add_div_mean_rsub_6 = async_compile.triton('triton_per_fused_add_div_mean_rsub_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.persistent_reduction(
size_hints=[1, 4],
reduction_hint=ReductionHint.INNER,
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', 10: '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': {9: 1}, 'configs': [AttrsDescriptor(divisible_by_16=(0, 1, 2, 3, 4, 5, 6, 7, 8), equal_to_1=(9,))]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused_add_div_mean_rsub_6', '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_add_div_mean_rsub_6(in_out_ptr0, in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, 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)
tmp3 = tl.load(in_ptr1 + (r0), None)
tmp10 = tl.load(in_ptr2 + (r0), None)
tmp12 = tl.load(in_ptr3 + (r0), None)
tmp19 = tl.load(in_ptr4 + (r0), None)
tmp21 = tl.load(in_ptr5 + (r0), None)
tmp28 = tl.load(in_ptr6 + (r0), None)
tmp30 = tl.load(in_ptr7 + (r0), None)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp4 = tmp3 + tmp1
tmp5 = tmp2 / tmp4
tmp6 = tmp1 - tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp11 = tmp10 + tmp1
tmp13 = tmp12 + tmp1
tmp14 = tmp11 / tmp13
tmp15 = tmp1 - tmp14
tmp16 = tl.broadcast_to(tmp15, [XBLOCK, RBLOCK])
tmp18 = tl.sum(tmp16, 1)[:, None]
tmp20 = tmp19 + tmp1
tmp22 = tmp21 + tmp1
tmp23 = tmp20 / tmp22
tmp24 = tmp1 - tmp23
tmp25 = tl.broadcast_to(tmp24, [XBLOCK, RBLOCK])
tmp27 = tl.sum(tmp25, 1)[:, None]
tmp29 = tmp28 + tmp1
tmp31 = tmp30 + tmp1
tmp32 = tmp29 / tmp31
tmp33 = tmp1 - tmp32
tmp34 = tl.broadcast_to(tmp33, [XBLOCK, RBLOCK])
tmp36 = tl.sum(tmp34, 1)[:, None]
tmp37 = 4.0
tmp38 = tmp9 / tmp37
tmp39 = 0.0
tmp40 = tmp38 + tmp39
tmp41 = tmp18 / tmp37
tmp42 = tmp40 + tmp41
tmp43 = tmp27 / tmp37
tmp44 = tmp42 + tmp43
tmp45 = tmp36 / tmp37
tmp46 = tmp44 + tmp45
tmp47 = 0.25
tmp48 = tmp46 * tmp47
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([XBLOCK, 1], 0, tl.int32)), tmp48, 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: [predict], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(arg0_1, buf0, 256, grid=grid(256), stream=stream0)
del arg0_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [predict], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf0, buf1, 256, grid=grid(256), stream=stream0)
del buf0
buf2 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf3 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul, sum_1, pow_1, pow_2, add_1, sum_2], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
triton_per_fused_add_mul_pow_sum_2.run(buf1, arg1_1, buf2, buf3, 4, 16, grid=grid(4), stream=stream0)
buf11 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf12 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul_3, sum_7, pow_7, pow_8, add_11, sum_8], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
triton_per_fused_add_mul_pow_sum_3.run(buf1, arg1_1, buf11, buf12, 4, 16, grid=grid(4), stream=stream0)
buf5 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf6 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul_1, sum_3, pow_3, pow_4, add_5, sum_4], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
triton_per_fused_add_mul_pow_sum_4.run(buf1, arg1_1, buf5, buf6, 4, 16, grid=grid(4), stream=stream0)
buf8 = empty_strided_cuda((4, ), (1, ), torch.float32)
buf9 = empty_strided_cuda((4, ), (1, ), torch.float32)
# Topologically Sorted Source Nodes: [mul_2, sum_5, pow_5, pow_6, add_8, sum_6], Original ATen: [aten.mul, aten.sum, aten.pow, aten.add]
triton_per_fused_add_mul_pow_sum_5.run(buf1, arg1_1, buf8, buf9, 4, 16, grid=grid(4), stream=stream0)
del arg1_1
del buf1
buf10 = empty_strided_cuda((), (), torch.float32)
buf14 = buf10; del buf10 # reuse
# Topologically Sorted Source Nodes: [num, den, truediv, loss, dice_loss, total_loss, num_1, den_1, truediv_1, loss_1, dice_loss_1, total_loss_1, num_2, den_2, truediv_2, loss_2, dice_loss_2, total_loss_2, num_3, den_3, truediv_3, loss_3, dice_loss_3, total_loss_3, truediv_4], Original ATen: [aten.add, aten.div, aten.rsub, aten.mean]
triton_per_fused_add_div_mean_rsub_6.run(buf14, buf2, buf3, buf5, buf6, buf8, buf9, buf11, buf12, 1, 4, grid=grid(1), stream=stream0)
del buf11
del buf12
del buf2
del buf3
del buf5
del buf6
del buf8
del buf9
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, 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.functional as F
class BinaryDiceLoss(nn.Module):
"""Dice loss of binary class
Args:
smooth: A float number to smooth loss, and avoid NaN error, default: 1
p: Denominator value: \\sum{x^p} + \\sum{y^p}, default: 2
predict: A tensor of shape [N, *]
target: A tensor of shape same with predict
reduction: Reduction method to apply, return mean over batch if 'mean',
return sum if 'sum', return a tensor of shape [N,] if 'none'
Returns:
Loss tensor according to arg reduction
Raise:
Exception if unexpected reduction
"""
def __init__(self, smooth=1, p=2, reduction='mean'):
super(BinaryDiceLoss, self).__init__()
self.smooth = smooth
self.p = p
self.reduction = reduction
def forward(self, predict, target):
assert predict.shape[0] == target.shape[0
], "predict & target batch size don't match"
predict = predict.contiguous().view(predict.shape[0], -1)
target = target.contiguous().view(target.shape[0], -1)
num = torch.sum(torch.mul(predict, target), dim=1) + self.smooth
den = torch.sum(predict.pow(self.p) + target.pow(self.p), dim=1
) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
return loss.mean()
elif self.reduction == 'sum':
return loss.sum()
elif self.reduction == 'none':
return loss
else:
raise Exception('Unexpected reduction {}'.format(self.reduction))
class DiceLoss(nn.Module):
"""Dice loss, need one hot encode input
Args:
weight: An array of shape [num_classes,]
ignore_index: class index to ignore
predict: A tensor of shape [N, C, *], should be raw logits
target: A tensor of same shape with predict
other args pass to BinaryDiceLoss
Return:
same as BinaryDiceLoss
"""
def __init__(self, weight=None, ignore_index=None, **kwargs):
super(DiceLoss, self).__init__()
self.kwargs = kwargs
self.weight = weight
self.ignore_index = ignore_index
def forward(self, predict, target):
assert predict.shape == target.shape, 'predict & target shape do not match'
dice = BinaryDiceLoss(**self.kwargs)
total_loss = 0
predict = F.softmax(predict, dim=1)
for i in range(target.shape[1]):
if i != self.ignore_index:
dice_loss = dice(predict[:, i], target[:, i])
if self.weight is not None:
assert self.weight.shape[0] == target.shape[1
], 'Expect weight shape [{}], get[{}]'.format(target
.shape[1], self.weight.shape[0])
dice_loss *= self.weight[i]
total_loss += dice_loss
return total_loss / target.shape[1]
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
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 = 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_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 = tmp1 + tmp2
tmp5 = tmp3 + tmp4
tmp7 = tmp5 + tmp6
tmp8 = tmp0 / tmp7
tl.store(out_ptr0 + x3, tmp8, xmask)
@triton.jit
def triton_per_fused_add_mul_pow_sum_2(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
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 + 64 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (r1 + 64 * 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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp13, xmask)
@triton.jit
def triton_per_fused_add_mul_pow_sum_3(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
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 + (48 + r1 + 64 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (48 + r1 + 64 * 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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp13, xmask)
@triton.jit
def triton_per_fused_add_mul_pow_sum_4(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
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 + (16 + r1 + 64 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (16 + r1 + 64 * 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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp13, xmask)
@triton.jit
def triton_per_fused_add_mul_pow_sum_5(in_ptr0, in_ptr1, out_ptr0, out_ptr1,
xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
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 + (32 + r1 + 64 * x0), xmask, other=0.0)
tmp1 = tl.load(in_ptr1 + (32 + r1 + 64 * 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 = tmp0 * tmp0
tmp8 = tmp1 * tmp1
tmp9 = tmp7 + tmp8
tmp10 = tl.broadcast_to(tmp9, [XBLOCK, RBLOCK])
tmp12 = tl.where(xmask, tmp10, 0)
tmp13 = tl.sum(tmp12, 1)[:, None]
tl.store(out_ptr0 + x0, tmp6, xmask)
tl.store(out_ptr1 + x0, tmp13, xmask)
@triton.jit
def triton_per_fused_add_div_mean_rsub_6(in_out_ptr0, in_ptr0, in_ptr1,
in_ptr2, in_ptr3, in_ptr4, in_ptr5, in_ptr6, in_ptr7, 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)
tmp3 = tl.load(in_ptr1 + r0, None)
tmp10 = tl.load(in_ptr2 + r0, None)
tmp12 = tl.load(in_ptr3 + r0, None)
tmp19 = tl.load(in_ptr4 + r0, None)
tmp21 = tl.load(in_ptr5 + r0, None)
tmp28 = tl.load(in_ptr6 + r0, None)
tmp30 = tl.load(in_ptr7 + r0, None)
tmp1 = 1.0
tmp2 = tmp0 + tmp1
tmp4 = tmp3 + tmp1
tmp5 = tmp2 / tmp4
tmp6 = tmp1 - tmp5
tmp7 = tl.broadcast_to(tmp6, [XBLOCK, RBLOCK])
tmp9 = tl.sum(tmp7, 1)[:, None]
tmp11 = tmp10 + tmp1
tmp13 = tmp12 + tmp1
tmp14 = tmp11 / tmp13
tmp15 = tmp1 - tmp14
tmp16 = tl.broadcast_to(tmp15, [XBLOCK, RBLOCK])
tmp18 = tl.sum(tmp16, 1)[:, None]
tmp20 = tmp19 + tmp1
tmp22 = tmp21 + tmp1
tmp23 = tmp20 / tmp22
tmp24 = tmp1 - tmp23
tmp25 = tl.broadcast_to(tmp24, [XBLOCK, RBLOCK])
tmp27 = tl.sum(tmp25, 1)[:, None]
tmp29 = tmp28 + tmp1
tmp31 = tmp30 + tmp1
tmp32 = tmp29 / tmp31
tmp33 = tmp1 - tmp32
tmp34 = tl.broadcast_to(tmp33, [XBLOCK, RBLOCK])
tmp36 = tl.sum(tmp34, 1)[:, None]
tmp37 = 4.0
tmp38 = tmp9 / tmp37
tmp39 = 0.0
tmp40 = tmp38 + tmp39
tmp41 = tmp18 / tmp37
tmp42 = tmp40 + tmp41
tmp43 = tmp27 / tmp37
tmp44 = tmp42 + tmp43
tmp45 = tmp36 / tmp37
tmp46 = tmp44 + tmp45
tmp47 = 0.25
tmp48 = tmp46 * tmp47
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([XBLOCK, 1], 0, tl.int32), tmp48, 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__softmax_0[grid(256)](arg0_1, buf0, 256, XBLOCK=
256, num_warps=4, num_stages=1)
del arg0_1
buf1 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused__softmax_1[grid(256)](buf0, buf1, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf0
buf2 = empty_strided_cuda((4,), (1,), torch.float32)
buf3 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_mul_pow_sum_2[grid(4)](buf1, arg1_1, buf2,
buf3, 4, 16, XBLOCK=1, num_warps=2, num_stages=1)
buf11 = empty_strided_cuda((4,), (1,), torch.float32)
buf12 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_mul_pow_sum_3[grid(4)](buf1, arg1_1, buf11,
buf12, 4, 16, XBLOCK=1, num_warps=2, num_stages=1)
buf5 = empty_strided_cuda((4,), (1,), torch.float32)
buf6 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_mul_pow_sum_4[grid(4)](buf1, arg1_1, buf5,
buf6, 4, 16, XBLOCK=1, num_warps=2, num_stages=1)
buf8 = empty_strided_cuda((4,), (1,), torch.float32)
buf9 = empty_strided_cuda((4,), (1,), torch.float32)
triton_per_fused_add_mul_pow_sum_5[grid(4)](buf1, arg1_1, buf8,
buf9, 4, 16, XBLOCK=1, num_warps=2, num_stages=1)
del arg1_1
del buf1
buf10 = empty_strided_cuda((), (), torch.float32)
buf14 = buf10
del buf10
triton_per_fused_add_div_mean_rsub_6[grid(1)](buf14, buf2, buf3,
buf5, buf6, buf8, buf9, buf11, buf12, 1, 4, XBLOCK=1, num_warps
=2, num_stages=1)
del buf11
del buf12
del buf2
del buf3
del buf5
del buf6
del buf8
del buf9
return buf14,
class BinaryDiceLoss(nn.Module):
"""Dice loss of binary class
Args:
smooth: A float number to smooth loss, and avoid NaN error, default: 1
p: Denominator value: \\sum{x^p} + \\sum{y^p}, default: 2
predict: A tensor of shape [N, *]
target: A tensor of shape same with predict
reduction: Reduction method to apply, return mean over batch if 'mean',
return sum if 'sum', return a tensor of shape [N,] if 'none'
Returns:
Loss tensor according to arg reduction
Raise:
Exception if unexpected reduction
"""
def __init__(self, smooth=1, p=2, reduction='mean'):
super(BinaryDiceLoss, self).__init__()
self.smooth = smooth
self.p = p
self.reduction = reduction
def forward(self, predict, target):
assert predict.shape[0] == target.shape[0
], "predict & target batch size don't match"
predict = predict.contiguous().view(predict.shape[0], -1)
target = target.contiguous().view(target.shape[0], -1)
num = torch.sum(torch.mul(predict, target), dim=1) + self.smooth
den = torch.sum(predict.pow(self.p) + target.pow(self.p), dim=1
) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
return loss.mean()
elif self.reduction == 'sum':
return loss.sum()
elif self.reduction == 'none':
return loss
else:
raise Exception('Unexpected reduction {}'.format(self.reduction))
class DiceLossNew(nn.Module):
"""Dice loss, need one hot encode input
Args:
weight: An array of shape [num_classes,]
ignore_index: class index to ignore
predict: A tensor of shape [N, C, *], should be raw logits
target: A tensor of same shape with predict
other args pass to BinaryDiceLoss
Return:
same as BinaryDiceLoss
"""
def __init__(self, weight=None, ignore_index=None, **kwargs):
super(DiceLossNew, self).__init__()
self.kwargs = kwargs
self.weight = 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]
| hekq/3DFingerPose | DiceLoss | false | 6,817 | [
"MIT"
] | 1 | 385c672408e2fd29ed0373a842727c9fcfd0fc59 | https://github.com/hekq/3DFingerPose/tree/385c672408e2fd29ed0373a842727c9fcfd0fc59 | import torch
import torch.nn as nn
import torch.nn.functional as F
class BinaryDiceLoss(nn.Module):
"""Dice loss of binary class
Args:
smooth: A float number to smooth loss, and avoid NaN error, default: 1
p: Denominator value: \\sum{x^p} + \\sum{y^p}, default: 2
predict: A tensor of shape [N, *]
target: A tensor of shape same with predict
reduction: Reduction method to apply, return mean over batch if 'mean',
return sum if 'sum', return a tensor of shape [N,] if 'none'
Returns:
Loss tensor according to arg reduction
Raise:
Exception if unexpected reduction
"""
def __init__(self, smooth=1, p=2, reduction='mean'):
super().__init__()
self.smooth = smooth
self.p = p
self.reduction = reduction
def forward(self, predict, target):
assert predict.shape[0] == target.shape[0
], "predict & target batch size don't match"
predict = predict.contiguous().view(predict.shape[0], -1)
target = target.contiguous().view(target.shape[0], -1)
num = torch.sum(torch.mul(predict, target), dim=1) + self.smooth
den = torch.sum(predict.pow(self.p) + target.pow(self.p), dim=1
) + self.smooth
loss = 1 - num / den
if self.reduction == 'mean':
return loss.mean()
elif self.reduction == 'sum':
return loss.sum()
elif self.reduction == 'none':
return loss
else:
raise Exception('Unexpected reduction {}'.format(self.reduction))
class Model(nn.Module):
"""Dice loss, need one hot encode input
Args:
weight: An array of shape [num_classes,]
ignore_index: class index to ignore
predict: A tensor of shape [N, C, *], should be raw logits
target: A tensor of same shape with predict
other args pass to BinaryDiceLoss
Return:
same as BinaryDiceLoss
"""
def __init__(self, weight=None, ignore_index=None, **kwargs):
super().__init__()
self.kwargs = kwargs
self.weight = weight
self.ignore_index = ignore_index
def forward(self, predict, target):
assert predict.shape == target.shape, 'predict & target shape do not match'
dice = BinaryDiceLoss(**self.kwargs)
total_loss = 0
predict = F.softmax(predict, dim=1)
for i in range(target.shape[1]):
if i != self.ignore_index:
dice_loss = dice(predict[:, i], target[:, i])
if self.weight is not None:
assert self.weight.shape[0] == target.shape[1
], 'Expect weight shape [{}], get[{}]'.format(target
.shape[1], self.weight.shape[0])
dice_loss *= self.weight[i]
total_loss += dice_loss
return total_loss / target.shape[1]
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
StdConv2d | # 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_4/inductor_cache/v7/cv7v5z5haeq6vp3kkiznbophyoey5r52utdywxy2lrupnip3tbh6.py
# Topologically Sorted Source Nodes: [v, m, sub, add, sqrt, w], Original ATen: [aten.var, aten.mean, aten.sub, aten.add, aten.sqrt, aten.div]
# Source node to ATen node mapping:
# add => add
# m => mean
# sqrt => sqrt
# sub => sub
# v => var
# w => div
# Graph fragment:
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%primals_1, [1, 2, 3]), kwargs = {correction: 0, keepdim: True})
# %mean : [num_users=2] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [1, 2, 3], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %mean), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%var, 1e-10), kwargs = {})
# %sqrt : [num_users=2] = call_function[target=torch.ops.aten.sqrt.default](args = (%add,), kwargs = {})
# %div : [num_users=2] = call_function[target=torch.ops.aten.div.Tensor](args = (%sub, %sqrt), kwargs = {})
triton_per_fused_add_div_mean_sqrt_sub_var_0 = async_compile.triton('triton_per_fused_add_div_mean_sqrt_sub_var_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_add_div_mean_sqrt_sub_var_0', 'mutated_arg_names': ['in_out_ptr0', 'in_out_ptr1'], '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_add_div_mean_sqrt_sub_var_0(in_out_ptr0, in_out_ptr1, in_ptr0, 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)
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]
tmp18 = tl.sum(tmp3, 1)[:, None]
tmp19 = 64.0
tmp20 = tmp18 / tmp19
tmp21 = tmp16 / tmp19
tmp22 = 1e-10
tmp23 = tmp21 + tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = tmp0 - tmp20
tmp26 = tmp25 / tmp24
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp20, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + (x0), tmp24, xmask)
tl.store(out_ptr0 + (r1 + (64*x0)), tmp26, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/tc/ctcagp37ljugm52zu6ckorigrppqo67voefe2f2odg5r6hyllhyu.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_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 = 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)
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf3 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf4 = reinterpret_tensor(buf3, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf3 # reuse
buf5 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0); del buf1 # reuse
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [v, m, sub, add, sqrt, w], Original ATen: [aten.var, aten.mean, aten.sub, aten.add, aten.sqrt, aten.div]
stream0 = get_raw_stream(0)
triton_per_fused_add_div_mean_sqrt_sub_var_0.run(buf4, buf5, primals_1, buf6, 4, 64, grid=grid(4), stream=stream0)
# 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_1.run(buf8, primals_2, 16, grid=grid(16), stream=stream0)
del primals_2
return (buf8, primals_1, primals_3, buf4, 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.functional as F
class StdConv2d(nn.Conv2d):
def forward(self, x):
w = self.weight
v = torch.var(w, dim=[1, 2, 3], keepdim=True, unbiased=False)
m = torch.mean(w, dim=[1, 2, 3], keepdim=True)
w = (w - m) / torch.sqrt(v + 1e-10)
return F.conv2d(x, w, 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
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_add_div_mean_sqrt_sub_var_0(in_out_ptr0, in_out_ptr1,
in_ptr0, 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)
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]
tmp18 = tl.sum(tmp3, 1)[:, None]
tmp19 = 64.0
tmp20 = tmp18 / tmp19
tmp21 = tmp16 / tmp19
tmp22 = 1e-10
tmp23 = tmp21 + tmp22
tmp24 = libdevice.sqrt(tmp23)
tmp25 = tmp0 - tmp20
tmp26 = tmp25 / tmp24
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp20, xmask)
tl.debug_barrier()
tl.store(in_out_ptr1 + x0, tmp24, xmask)
tl.store(out_ptr0 + (r1 + 64 * x0), tmp26, 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 = 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)
buf1 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf3 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
buf4 = reinterpret_tensor(buf3, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf3
buf5 = reinterpret_tensor(buf1, (4, 1, 1, 1), (1, 1, 1, 1), 0)
del buf1
buf6 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
get_raw_stream(0)
triton_per_fused_add_div_mean_sqrt_sub_var_0[grid(4)](buf4, buf5,
primals_1, buf6, 4, 64, XBLOCK=1, num_warps=2, num_stages=1)
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_1[grid(16)](buf8, primals_2, 16,
XBLOCK=16, num_warps=1, num_stages=1)
del primals_2
return buf8, primals_1, primals_3, buf4, buf5, buf6
class StdConv2dNew(nn.Conv2d):
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]
| hrlblab/Glo-In-One | StdConv2d | false | 6,818 | [
"Apache-2.0"
] | 1 | 7daef49c557bccd6f5c956b88603357346dc78a2 | https://github.com/hrlblab/Glo-In-One/tree/7daef49c557bccd6f5c956b88603357346dc78a2 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Conv2d):
def forward(self, x):
w = self.weight
v = torch.var(w, dim=[1, 2, 3], keepdim=True, unbiased=False)
m = torch.mean(w, dim=[1, 2, 3], keepdim=True)
w = (w - m) / torch.sqrt(v + 1e-10)
return F.conv2d(x, w, 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]
|
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_4/inductor_cache/os/coskposucjuzyvraxc3tqieebquslvmnvai5jjzk6fnbq5ziatlo.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# out => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view, %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=[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_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 = 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)
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 = 0.5
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 = tl_math.exp(tmp14)
tl.store(out_ptr0 + (x2), tmp15, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/kj/ckjtlefzavjukjsytvkak6ek26zmzexpcbnlwelx4k5kascjxlf3.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# out => 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=1] = 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=[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
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), (16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4), (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: [bmm], Original ATen: [aten.bmm]
extern_kernels.bmm(arg1_1, reinterpret_tensor(arg0_1, (4, 4, 4), (16, 1, 4), 0), out=buf0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten._softmax]
stream0 = get_raw_stream(0)
triton_poi_fused__softmax_0.run(buf0, buf1, 64, grid=grid(64), stream=stream0)
buf2 = reinterpret_tensor(buf0, (16, 4), (4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [out], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf1, buf2, 64, grid=grid(64), stream=stream0)
buf3 = reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.bmm]
extern_kernels.bmm(reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0), arg2_1, out=buf3)
del arg2_1
return (buf3, reinterpret_tensor(buf2, (4, 4, 4), (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), (16, 4, 1), device='cuda:0', dtype=torch.float32)
arg1_1 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
arg2_1 = rand_strided((4, 4, 4), (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
from torch.autograd import Variable
import torch.nn as nn
import torch.optim
class Bottle(nn.Module):
""" Perform the reshape routine before and after an operation """
def forward(self, input):
if len(input.size()) <= 2:
return super(Bottle, self).forward(input)
size = input.size()[:2]
out = super(Bottle, self).forward(input.view(size[0] * size[1], -1))
return out.view(size[0], size[1], -1)
class BottleSoftmax(Bottle, nn.Softmax):
""" Perform the reshape routine before and after a softmax operation"""
pass
class ScaledDotProductAttention(nn.Module):
""" Scaled Dot-Product Attention """
def __init__(self, d_model, attn_dropout=0.1):
super(ScaledDotProductAttention, self).__init__()
self.temper = d_model ** 0.5
self.dropout = nn.Dropout(attn_dropout)
self.softmax = BottleSoftmax(dim=1)
def forward(self, q, k, v, attn_mask=None):
"""
:type attn_mask: torch.FloatTensor
:param attn_mask: Mask of the attention.
3D tensor with shape [batch_size, time_step_key, time_step_value]
"""
attn = torch.bmm(q, k.transpose(1, 2)) / self.temper
if attn_mask is not None:
assert attn_mask.size() == attn.size(
), 'Attention mask shape {} mismatch with Attention logit tensor shape {}.'.format(
attn_mask.size(), attn.size())
attn = attn.masked_fill(Variable(attn_mask), -1e+18)
attn = self.softmax(attn)
attn = self.dropout(attn)
output = torch.bmm(attn, v)
return output, attn
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [[], {'d_model': 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
import torch.optim
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 = 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)
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 = 0.5
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 = tl_math.exp(tmp14)
tl.store(out_ptr0 + x2, tmp15, 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
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), (16, 4, 1))
assert_size_stride(arg1_1, (4, 4, 4), (16, 4, 1))
assert_size_stride(arg2_1, (4, 4, 4), (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)
extern_kernels.bmm(arg1_1, reinterpret_tensor(arg0_1, (4, 4, 4), (
16, 1, 4), 0), out=buf0)
del arg0_1
del arg1_1
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused__softmax_0[grid(64)](buf0, buf1, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf2 = reinterpret_tensor(buf0, (16, 4), (4, 1), 0)
del buf0
triton_poi_fused__softmax_1[grid(64)](buf1, buf2, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf3 = reinterpret_tensor(buf1, (4, 4, 4), (16, 4, 1), 0)
del buf1
extern_kernels.bmm(reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1),
0), arg2_1, out=buf3)
del arg2_1
return buf3, reinterpret_tensor(buf2, (4, 4, 4), (16, 4, 1), 0)
class Bottle(nn.Module):
""" Perform the reshape routine before and after an operation """
def forward(self, input):
if len(input.size()) <= 2:
return super(Bottle, self).forward(input)
size = input.size()[:2]
out = super(Bottle, self).forward(input.view(size[0] * size[1], -1))
return out.view(size[0], size[1], -1)
class BottleSoftmax(Bottle, nn.Softmax):
""" Perform the reshape routine before and after a softmax operation"""
pass
class ScaledDotProductAttentionNew(nn.Module):
""" Scaled Dot-Product Attention """
def __init__(self, d_model, attn_dropout=0.1):
super(ScaledDotProductAttentionNew, self).__init__()
self.temper = d_model ** 0.5
self.dropout = nn.Dropout(attn_dropout)
self.softmax = BottleSoftmax(dim=1)
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]
| howardchenhd/Transformer-pytorch | ScaledDotProductAttention | false | 6,819 | [
"MIT"
] | 1 | ae71ed5767272feb7e717be6d5bfce46f80ec57a | https://github.com/howardchenhd/Transformer-pytorch/tree/ae71ed5767272feb7e717be6d5bfce46f80ec57a | import torch
from torch.autograd import Variable
import torch.nn as nn
import torch.optim
class Bottle(nn.Module):
""" Perform the reshape routine before and after an operation """
def forward(self, input):
if len(input.size()) <= 2:
return super(Bottle, self).forward(input)
size = input.size()[:2]
out = super(Bottle, self).forward(input.view(size[0] * size[1], -1))
return out.view(size[0], size[1], -1)
class BottleSoftmax(Bottle, nn.Softmax):
""" Perform the reshape routine before and after a softmax operation"""
pass
class Model(nn.Module):
""" Scaled Dot-Product Attention """
def __init__(self, d_model, attn_dropout=0.1):
super().__init__()
self.temper = d_model ** 0.5
self.dropout = nn.Dropout(attn_dropout)
self.softmax = BottleSoftmax(dim=1)
def forward(self, q, k, v, attn_mask=None):
"""
:type attn_mask: torch.FloatTensor
:param attn_mask: Mask of the attention.
3D tensor with shape [batch_size, time_step_key, time_step_value]
"""
attn = torch.bmm(q, k.transpose(1, 2)) / self.temper
if attn_mask is not None:
assert attn_mask.size() == attn.size(
), 'Attention mask shape {} mismatch with Attention logit tensor shape {}.'.format(
attn_mask.size(), attn.size())
attn = attn.masked_fill(Variable(attn_mask), -1e+18)
attn = self.softmax(attn)
attn = self.dropout(attn)
output = torch.bmm(attn, v)
return output, attn
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]
|
LeNet_300_100 | # 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_4/inductor_cache/gi/cgisfp7cjgya6wz3zy6wgsfrei7pmk7oec6olmpydu3ckl4g7flh.py
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_1 => relu
# Graph fragment:
# %add_tensor_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default_1, %primals_3), 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=[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 = 1200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 300
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_4/inductor_cache/d7/cd7rilnjd42cirsc5dhnnwlficmjz5omrtsdfojgouhplcpynn4n.py
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.relu]
# Source node to ATen node mapping:
# x_2 => relu_1
# Graph fragment:
# %add_tensor : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mm_default, %primals_5), kwargs = {})
# %relu_1 : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%add_tensor,), kwargs = {})
triton_poi_fused_relu_1 = async_compile.triton('triton_poi_fused_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=[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_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_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
x2 = xindex
x0 = xindex % 100
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')
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, 784), (784, 1))
assert_size_stride(primals_2, (300, 784), (784, 1))
assert_size_stride(primals_3, (300, ), (1, ))
assert_size_stride(primals_4, (100, 300), (300, 1))
assert_size_stride(primals_5, (100, ), (1, ))
assert_size_stride(primals_6, (10, 100), (100, 1))
assert_size_stride(primals_7, (10, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 300), (300, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784, 300), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0; del buf0 # reuse
# Topologically Sorted Source Nodes: [x_1], Original ATen: [aten.relu]
stream0 = get_raw_stream(0)
triton_poi_fused_relu_0.run(buf1, primals_3, 1200, grid=grid(1200), stream=stream0)
del primals_3
buf2 = empty_strided_cuda((4, 100), (100, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (300, 100), (1, 300), 0), out=buf2)
buf3 = buf2; del buf2 # reuse
# Topologically Sorted Source Nodes: [x_2], Original ATen: [aten.relu]
triton_poi_fused_relu_1.run(buf3, primals_5, 400, grid=grid(400), stream=stream0)
del primals_5
buf4 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
# Topologically Sorted Source Nodes: [x_3], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_7, buf3, reinterpret_tensor(primals_6, (100, 10), (1, 100), 0), alpha=1, beta=1, out=buf4)
del primals_7
return (buf4, primals_1, buf1, buf3, 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((300, 784), (784, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((300, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((100, 300), (300, 1), device='cuda:0', dtype=torch.float32)
primals_5 = rand_strided((100, ), (1, ), device='cuda:0', dtype=torch.float32)
primals_6 = rand_strided((10, 100), (100, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((10, ), (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.nn.functional as F
class LeNet_300_100(nn.Module):
def __init__(self):
super().__init__()
self.fc1 = nn.Linear(28 * 28, 300)
self.fc2 = nn.Linear(300, 100)
self.fc3 = nn.Linear(100, 10)
self.relu = nn.ReLU()
self.lastbias = 'fc3.bias'
def forward(self, x):
x = x.view(-1, 28 * 28)
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, 784])]
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_relu_0(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl.constexpr
):
xnumel = 1200
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x2 = xindex
x0 = xindex % 300
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_relu_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
x2 = xindex
x0 = xindex % 100
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)
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, 784), (784, 1))
assert_size_stride(primals_2, (300, 784), (784, 1))
assert_size_stride(primals_3, (300,), (1,))
assert_size_stride(primals_4, (100, 300), (300, 1))
assert_size_stride(primals_5, (100,), (1,))
assert_size_stride(primals_6, (10, 100), (100, 1))
assert_size_stride(primals_7, (10,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 300), (300, 1), torch.float32)
extern_kernels.mm(primals_1, reinterpret_tensor(primals_2, (784,
300), (1, 784), 0), out=buf0)
del primals_2
buf1 = buf0
del buf0
get_raw_stream(0)
triton_poi_fused_relu_0[grid(1200)](buf1, primals_3, 1200, XBLOCK=
128, num_warps=4, num_stages=1)
del primals_3
buf2 = empty_strided_cuda((4, 100), (100, 1), torch.float32)
extern_kernels.mm(buf1, reinterpret_tensor(primals_4, (300, 100), (
1, 300), 0), out=buf2)
buf3 = buf2
del buf2
triton_poi_fused_relu_1[grid(400)](buf3, primals_5, 400, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_5
buf4 = empty_strided_cuda((4, 10), (10, 1), torch.float32)
extern_kernels.addmm(primals_7, buf3, reinterpret_tensor(primals_6,
(100, 10), (1, 100), 0), alpha=1, beta=1, out=buf4)
del primals_7
return buf4, primals_1, buf1, buf3, primals_6, primals_4
class LeNet_300_100New(nn.Module):
def __init__(self):
super().__init__()
self.fc1 = nn.Linear(28 * 28, 300)
self.fc2 = nn.Linear(300, 100)
self.fc3 = nn.Linear(100, 10)
self.relu = nn.ReLU()
self.lastbias = 'fc3.bias'
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_6 = self.fc3.weight
primals_7 = self.fc3.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| htt-trangtran/smg | LeNet_300_100 | false | 6,820 | [
"MIT"
] | 1 | b7a49055e7d48ec456bac67ab473db2183d2f597 | https://github.com/htt-trangtran/smg/tree/b7a49055e7d48ec456bac67ab473db2183d2f597 | import torch
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super().__init__()
self.fc1 = nn.Linear(28 * 28, 300)
self.fc2 = nn.Linear(300, 100)
self.fc3 = nn.Linear(100, 10)
self.relu = nn.ReLU()
self.lastbias = 'fc3.bias'
def forward(self, x):
x = x.view(-1, 28 * 28)
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, 784])]
def get_init_inputs():
return []
|
IA_gate | # 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_4/inductor_cache/wt/cwtdijbiacyxk63gj3jzdaisb2mebdlxtfx3oswf3s3prqzcaqh7.py
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mul]
# Source node to ATen node mapping:
# x => mul
# Graph fragment:
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%unsqueeze_1, %primals_4), 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=[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_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 = 4096
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = (xindex // 16)
x4 = xindex % 256
x5 = xindex
tmp0 = tl.load(in_ptr0 + (x3), None, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + (x4), None, eviction_policy='evict_last')
tmp1 = libdevice.tanh(tmp0)
tmp2 = 1.0
tmp3 = tmp1 + tmp2
tmp5 = tmp3 * tmp4
tl.store(out_ptr0 + (x5), tmp5, None)
''', 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, ), (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 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [a], 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, 4, 4, 4), (1024, 256, 64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [x], Original ATen: [aten.mul]
stream0 = get_raw_stream(0)
triton_poi_fused_mul_0.run(buf0, primals_4, buf1, 4096, grid=grid(4096), stream=stream0)
return (buf1, 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, 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
class IA_gate(nn.Module):
def __init__(self, in_dim, out_dim):
super(IA_gate, self).__init__()
self.IA = nn.Linear(in_dim, out_dim)
def forward(self, x, IA_head):
a = self.IA(IA_head)
a = 1.0 + torch.tanh(a)
a = a.unsqueeze(-1).unsqueeze(-1)
x = a * x
return x
def get_inputs():
return [torch.rand([4, 4, 4, 4]), 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._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_mul_0(in_ptr0, in_ptr1, out_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 // 16
x4 = xindex % 256
x5 = xindex
tmp0 = tl.load(in_ptr0 + x3, None, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr1 + x4, None, eviction_policy='evict_last')
tmp1 = libdevice.tanh(tmp0)
tmp2 = 1.0
tmp3 = tmp1 + tmp2
tmp5 = tmp3 * tmp4
tl.store(out_ptr0 + x5, tmp5, None)
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,), (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 = 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, 4, 4, 4), (1024, 256, 64, 16, 4,
1), torch.float32)
get_raw_stream(0)
triton_poi_fused_mul_0[grid(4096)](buf0, primals_4, buf1, 4096,
XBLOCK=128, num_warps=4, num_stages=1)
return buf1, primals_4, reinterpret_tensor(primals_3, (64, 4), (4, 1), 0
), buf0
class IA_gateNew(nn.Module):
def __init__(self, in_dim, out_dim):
super(IA_gateNew, self).__init__()
self.IA = nn.Linear(in_dim, out_dim)
def forward(self, input_0, input_1):
primals_1 = self.IA.weight
primals_2 = self.IA.bias
primals_3 = input_0
primals_4 = input_1
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| huanglf714/COMatchNet | IA_gate | false | 6,821 | [
"Apache-2.0"
] | 1 | 79023f5be65d354eb9bdac026d7e0d73110bc4aa | https://github.com/huanglf714/COMatchNet/tree/79023f5be65d354eb9bdac026d7e0d73110bc4aa | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, in_dim, out_dim):
super().__init__()
self.IA = nn.Linear(in_dim, out_dim)
def forward(self, x, IA_head):
a = self.IA(IA_head)
a = 1.0 + torch.tanh(a)
a = a.unsqueeze(-1).unsqueeze(-1)
x = a * x
return x
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_4/inductor_cache/4m/c4msm3gcnvqm52kgfglaipw57py7fgpg6ss2ephzs6eadc3ycwd6.py
# Topologically Sorted Source Nodes: [out1, out1_1], Original ATen: [aten._native_batch_norm_legit, aten.relu]
# Source node to ATen node mapping:
# out1 => var_mean
# out1_1 => 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})
# %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: '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__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_ptr2, 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)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/s7/cs7bn5vk6z77xfaow55ofvrgwbti6m3wecvgag3dz7w6xz3pjuyc.py
# Topologically Sorted Source Nodes: [out2, out2_1], Original ATen: [aten._native_batch_norm_legit, aten.relu]
# Source node to ATen node mapping:
# out2 => add_1, rsqrt_1, var_mean_1
# out2_1 => relu_1
# Graph fragment:
# %var_mean_1 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_4, [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_5,), kwargs = {})
triton_per_fused__native_batch_norm_legit_relu_1 = async_compile.triton('triton_per_fused__native_batch_norm_legit_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.persistent_reduction(
size_hints=[8, 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, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__native_batch_norm_legit_relu_1', '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_1(in_ptr0, out_ptr0, out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 8
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_4/inductor_cache/lt/cltsrsxjtguuc4qfizqdocxzrpyopl6e6rupmtti5ryt2idude3y.py
# Topologically Sorted Source Nodes: [out3, out3_1], Original ATen: [aten._native_batch_norm_legit, aten.relu]
# Source node to ATen node mapping:
# out3 => add_2, rsqrt_2, var_mean_2
# out3_1 => relu_2
# Graph fragment:
# %var_mean_2 : [num_users=2] = call_function[target=torch.ops.aten.var_mean.correction](args = (%view_9, [0, 2, 3]), kwargs = {correction: 0, keepdim: True})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem_4, 1e-05), kwargs = {})
# %rsqrt_2 : [num_users=2] = call_function[target=torch.ops.aten.rsqrt.default](args = (%add_2,), kwargs = {})
# %relu_2 : [num_users=1] = call_function[target=torch.ops.aten.relu.default](args = (%view_10,), kwargs = {})
triton_per_fused__native_batch_norm_legit_relu_2 = async_compile.triton('triton_per_fused__native_batch_norm_legit_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.persistent_reduction(
size_hints=[4, 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, 5), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_per_fused__native_batch_norm_legit_relu_2', '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_2(in_ptr0, out_ptr0, out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK : tl.constexpr):
xnumel = 4
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_4/inductor_cache/4m/c4mjawqnzpr6o4o442ynjxsffhu5gxk44hpqmwnacvrkkldqyh5n.py
# Topologically Sorted Source Nodes: [out3_3, out3_4], Original ATen: [aten.cat, aten.add]
# Source node to ATen node mapping:
# out3_3 => cat
# out3_4 => add_3
# Graph fragment:
# %cat : [num_users=1] = call_function[target=torch.ops.aten.cat.default](args = ([%convolution, %convolution_1, %convolution_2], 1), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%cat, %primals_1), kwargs = {})
triton_poi_fused_add_cat_3 = async_compile.triton('triton_poi_fused_add_cat_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: '*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_cat_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_add_cat_3(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
x1 = (xindex // 16) % 4
x0 = xindex % 16
x2 = (xindex // 64)
x3 = xindex
tmp17 = tl.load(in_ptr3 + (x3), xmask)
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 + (16*x1) + (32*x2)), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 3, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + (16*x2)), tmp9 & xmask, eviction_policy='evict_last', other=0.0)
tmp11 = tmp0 >= tmp7
tmp12 = tl.full([1], 4, tl.int64)
tmp13 = tmp0 < tmp12
tmp14 = tl.load(in_ptr2 + (x0 + (16*x2)), tmp11 & xmask, eviction_policy='evict_last', other=0.0)
tmp15 = tl.where(tmp9, tmp10, tmp14)
tmp16 = tl.where(tmp4, tmp5, tmp15)
tmp18 = tmp16 + tmp17
tl.store(out_ptr0 + (x3), tmp18, 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, (2, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (1, 2, 3, 3), (18, 9, 3, 1))
assert_size_stride(primals_4, (1, 1, 3, 3), (9, 9, 3, 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: [out1, out1_1], Original ATen: [aten._native_batch_norm_legit, aten.relu]
stream0 = get_raw_stream(0)
triton_per_fused__native_batch_norm_legit_relu_0.run(primals_1, buf3, 16, 16, grid=grid(16), stream=stream0)
# Topologically Sorted Source Nodes: [out1_2], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_2, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 2, 4, 4), (32, 16, 4, 1))
buf5 = empty_strided_cuda((1, 8, 1, 1), (8, 1, 8, 8), torch.float32)
buf9 = empty_strided_cuda((4, 2, 4, 4), (32, 16, 4, 1), torch.float32)
buf8 = empty_strided_cuda((1, 8, 1, 1), (8, 1, 8, 8), torch.float32)
# Topologically Sorted Source Nodes: [out2, out2_1], Original ATen: [aten._native_batch_norm_legit, aten.relu]
triton_per_fused__native_batch_norm_legit_relu_1.run(buf4, buf5, buf9, buf8, 8, 16, grid=grid(8), stream=stream0)
# Topologically Sorted Source Nodes: [out2_2], Original ATen: [aten.convolution]
buf10 = extern_kernels.convolution(buf9, primals_3, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf10, (4, 1, 4, 4), (16, 16, 4, 1))
buf11 = empty_strided_cuda((1, 4, 1, 1), (4, 1, 4, 4), torch.float32)
buf15 = empty_strided_cuda((4, 1, 4, 4), (16, 1, 4, 1), torch.float32)
buf14 = empty_strided_cuda((1, 4, 1, 1), (4, 1, 4, 4), torch.float32)
# Topologically Sorted Source Nodes: [out3, out3_1], Original ATen: [aten._native_batch_norm_legit, aten.relu]
triton_per_fused__native_batch_norm_legit_relu_2.run(buf10, buf11, buf15, buf14, 4, 16, grid=grid(4), stream=stream0)
# Topologically Sorted Source Nodes: [out3_2], Original ATen: [aten.convolution]
buf16 = extern_kernels.convolution(reinterpret_tensor(buf15, (4, 1, 4, 4), (16, 0, 4, 1), 0), primals_4, stride=(1, 1), padding=(1, 1), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf16, (4, 1, 4, 4), (16, 16, 4, 1))
buf17 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [out3_3, out3_4], Original ATen: [aten.cat, aten.add]
triton_poi_fused_add_cat_3.run(buf4, buf10, buf16, primals_1, buf17, 256, grid=grid(256), stream=stream0)
del buf16
del primals_1
return (buf17, primals_2, primals_3, primals_4, buf3, buf4, reinterpret_tensor(buf8, (8, ), (1, ), 0), buf9, buf10, reinterpret_tensor(buf14, (4, ), (1, ), 0), reinterpret_tensor(buf15, (4, 1, 4, 4), (16, 16, 4, 1), 0), reinterpret_tensor(buf11, (1, 4, 1, 1), (4, 1, 1, 1), 0), reinterpret_tensor(buf5, (1, 8, 1, 1), (8, 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((2, 4, 3, 3), (36, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((1, 2, 3, 3), (18, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_4 = rand_strided((1, 1, 3, 3), (9, 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
import torch.nn.functional as F
def conv3x3(in_planes, out_planes, strd=1, padding=1, bias=False):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=strd,
padding=padding, bias=bias)
class ConvBlock(nn.Module):
def __init__(self, in_planes, out_planes, use_instance_norm):
super(ConvBlock, self).__init__()
self.bn1 = nn.InstanceNorm2d(in_planes
) if use_instance_norm else nn.BatchNorm2d(in_planes)
self.conv1 = conv3x3(in_planes, int(out_planes // 2))
self.bn2 = nn.InstanceNorm2d(int(out_planes / 2)
) if use_instance_norm else nn.BatchNorm2d(int(out_planes / 2))
self.conv2 = conv3x3(int(out_planes / 2), int(out_planes / 4))
self.bn3 = nn.InstanceNorm2d(int(out_planes / 4)
) if use_instance_norm else nn.BatchNorm2d(int(out_planes / 4))
self.conv3 = conv3x3(int(out_planes / 4), int(out_planes / 4))
if in_planes != out_planes:
self.downsample = nn.Sequential(nn.InstanceNorm2d(in_planes) if
use_instance_norm else nn.BatchNorm2d(in_planes), nn.ReLU(
True), nn.Conv2d(in_planes, out_planes, kernel_size=1,
stride=1, bias=False))
else:
self.downsample = None
def forward(self, x):
residual = x
out1 = self.bn1(x)
out1 = F.relu(out1, True)
out1 = self.conv1(out1)
out2 = self.bn2(out1)
out2 = F.relu(out2, True)
out2 = self.conv2(out2)
out3 = self.bn3(out2)
out3 = F.relu(out3, True)
out3 = self.conv3(out3)
out3 = torch.cat((out1, out2, out3), 1)
if self.downsample is not None:
residual = self.downsample(residual)
out3 += residual
return out3
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'in_planes': 4, 'out_planes': 4, 'use_instance_norm': 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
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_ptr2,
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)
@triton.jit
def triton_per_fused__native_batch_norm_legit_relu_1(in_ptr0, out_ptr0,
out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 8
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_2(in_ptr0, out_ptr0,
out_ptr2, out_ptr3, xnumel, rnumel, XBLOCK: tl.constexpr):
xnumel = 4
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_poi_fused_add_cat_3(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
x1 = xindex // 16 % 4
x0 = xindex % 16
x2 = xindex // 64
x3 = xindex
tmp17 = tl.load(in_ptr3 + x3, xmask)
tmp0 = x1
tl.full([1], 0, tl.int64)
tmp3 = tl.full([1], 2, tl.int64)
tmp4 = tmp0 < tmp3
tmp5 = tl.load(in_ptr0 + (x0 + 16 * x1 + 32 * x2), tmp4 & xmask, other=0.0)
tmp6 = tmp0 >= tmp3
tmp7 = tl.full([1], 3, tl.int64)
tmp8 = tmp0 < tmp7
tmp9 = tmp6 & tmp8
tmp10 = tl.load(in_ptr1 + (x0 + 16 * x2), tmp9 & xmask, eviction_policy
='evict_last', other=0.0)
tmp11 = tmp0 >= tmp7
tl.full([1], 4, tl.int64)
tmp14 = tl.load(in_ptr2 + (x0 + 16 * x2), tmp11 & xmask,
eviction_policy='evict_last', other=0.0)
tmp15 = tl.where(tmp9, tmp10, tmp14)
tmp16 = tl.where(tmp4, tmp5, tmp15)
tmp18 = tmp16 + tmp17
tl.store(out_ptr0 + x3, tmp18, 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, (2, 4, 3, 3), (36, 9, 3, 1))
assert_size_stride(primals_3, (1, 2, 3, 3), (18, 9, 3, 1))
assert_size_stride(primals_4, (1, 1, 3, 3), (9, 9, 3, 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__native_batch_norm_legit_relu_0[grid(16)](primals_1,
buf3, 16, 16, XBLOCK=8, num_warps=2, num_stages=1)
buf4 = extern_kernels.convolution(buf3, primals_2, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 2, 4, 4), (32, 16, 4, 1))
buf5 = empty_strided_cuda((1, 8, 1, 1), (8, 1, 8, 8), torch.float32)
buf9 = empty_strided_cuda((4, 2, 4, 4), (32, 16, 4, 1), torch.float32)
buf8 = empty_strided_cuda((1, 8, 1, 1), (8, 1, 8, 8), torch.float32)
triton_per_fused__native_batch_norm_legit_relu_1[grid(8)](buf4,
buf5, buf9, buf8, 8, 16, XBLOCK=8, num_warps=2, num_stages=1)
buf10 = extern_kernels.convolution(buf9, primals_3, stride=(1, 1),
padding=(1, 1), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf10, (4, 1, 4, 4), (16, 16, 4, 1))
buf11 = empty_strided_cuda((1, 4, 1, 1), (4, 1, 4, 4), torch.float32)
buf15 = empty_strided_cuda((4, 1, 4, 4), (16, 1, 4, 1), torch.float32)
buf14 = empty_strided_cuda((1, 4, 1, 1), (4, 1, 4, 4), torch.float32)
triton_per_fused__native_batch_norm_legit_relu_2[grid(4)](buf10,
buf11, buf15, buf14, 4, 16, XBLOCK=1, num_warps=2, num_stages=1)
buf16 = extern_kernels.convolution(reinterpret_tensor(buf15, (4, 1,
4, 4), (16, 0, 4, 1), 0), primals_4, stride=(1, 1), padding=(1,
1), dilation=(1, 1), transposed=False, output_padding=(0, 0),
groups=1, bias=None)
assert_size_stride(buf16, (4, 1, 4, 4), (16, 16, 4, 1))
buf17 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_cat_3[grid(256)](buf4, buf10, buf16, primals_1,
buf17, 256, XBLOCK=128, num_warps=4, num_stages=1)
del buf16
del primals_1
return (buf17, primals_2, primals_3, primals_4, buf3, buf4,
reinterpret_tensor(buf8, (8,), (1,), 0), buf9, buf10,
reinterpret_tensor(buf14, (4,), (1,), 0), reinterpret_tensor(buf15,
(4, 1, 4, 4), (16, 16, 4, 1), 0), reinterpret_tensor(buf11, (1, 4,
1, 1), (4, 1, 1, 1), 0), reinterpret_tensor(buf5, (1, 8, 1, 1), (8,
1, 1, 1), 0))
def conv3x3(in_planes, out_planes, strd=1, padding=1, bias=False):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=strd,
padding=padding, bias=bias)
class ConvBlockNew(nn.Module):
def __init__(self, in_planes, out_planes, use_instance_norm):
super(ConvBlockNew, self).__init__()
self.bn1 = nn.InstanceNorm2d(in_planes
) if use_instance_norm else nn.BatchNorm2d(in_planes)
self.conv1 = conv3x3(in_planes, int(out_planes // 2))
self.bn2 = nn.InstanceNorm2d(int(out_planes / 2)
) if use_instance_norm else nn.BatchNorm2d(int(out_planes / 2))
self.conv2 = conv3x3(int(out_planes / 2), int(out_planes / 4))
self.bn3 = nn.InstanceNorm2d(int(out_planes / 4)
) if use_instance_norm else nn.BatchNorm2d(int(out_planes / 4))
self.conv3 = conv3x3(int(out_planes / 4), int(out_planes / 4))
if in_planes != out_planes:
self.downsample = nn.Sequential(nn.InstanceNorm2d(in_planes) if
use_instance_norm else nn.BatchNorm2d(in_planes), nn.ReLU(
True), nn.Conv2d(in_planes, out_planes, kernel_size=1,
stride=1, bias=False))
else:
self.downsample = None
def forward(self, input_0):
primals_2 = self.conv1.weight
primals_3 = self.conv2.weight
primals_4 = self.conv3.weight
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| hhj1897/fan_training | ConvBlock | false | 6,822 | [
"MIT"
] | 1 | 5882f9edf2f1a07c80a6d1f3341a7cf1d348e217 | https://github.com/hhj1897/fan_training/tree/5882f9edf2f1a07c80a6d1f3341a7cf1d348e217 | import torch
import torch.nn as nn
import torch.nn.functional as F
def conv3x3(in_planes, out_planes, strd=1, padding=1, bias=False):
"""3x3 convolution with padding"""
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=strd,
padding=padding, bias=bias)
class Model(nn.Module):
def __init__(self, in_planes, out_planes, use_instance_norm):
super().__init__()
self.bn1 = nn.InstanceNorm2d(in_planes
) if use_instance_norm else nn.BatchNorm2d(in_planes)
self.conv1 = conv3x3(in_planes, int(out_planes // 2))
self.bn2 = nn.InstanceNorm2d(int(out_planes / 2)
) if use_instance_norm else nn.BatchNorm2d(int(out_planes / 2))
self.conv2 = conv3x3(int(out_planes / 2), int(out_planes / 4))
self.bn3 = nn.InstanceNorm2d(int(out_planes / 4)
) if use_instance_norm else nn.BatchNorm2d(int(out_planes / 4))
self.conv3 = conv3x3(int(out_planes / 4), int(out_planes / 4))
if in_planes != out_planes:
self.downsample = nn.Sequential(nn.InstanceNorm2d(in_planes) if
use_instance_norm else nn.BatchNorm2d(in_planes), nn.ReLU(
True), nn.Conv2d(in_planes, out_planes, kernel_size=1,
stride=1, bias=False))
else:
self.downsample = None
def forward(self, x):
residual = x
out1 = self.bn1(x)
out1 = F.relu(out1, True)
out1 = self.conv1(out1)
out2 = self.bn2(out1)
out2 = F.relu(out2, True)
out2 = self.conv2(out2)
out3 = self.bn3(out2)
out3 = F.relu(out3, True)
out3 = self.conv3(out3)
out3 = torch.cat((out1, out2, out3), 1)
if self.downsample is not None:
residual = self.downsample(residual)
out3 += residual
return out3
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4, 4]
|
PositionwiseFeedForward | # 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_4/inductor_cache/df/cdfcie57v6pcdd6oeaz4mvlgksxgyuxzmlv5bklwemyulqhtcxta.py
# Topologically Sorted Source Nodes: [mean, std, sub, mul, add, truediv, add_1], Original ATen: [aten.mean, aten.std, aten.sub, aten.mul, aten.add, aten.div]
# Source node to ATen node mapping:
# add => add
# add_1 => add_1
# mean => mean
# mul => mul
# std => sqrt, var
# sub => sub
# truediv => div
# Graph fragment:
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%primals_1, [-1], True), kwargs = {})
# %var : [num_users=1] = call_function[target=torch.ops.aten.var.correction](args = (%primals_1, [-1]), kwargs = {correction: 1.0, keepdim: True})
# %sqrt : [num_users=1] = call_function[target=torch.ops.aten.sqrt.default](args = (%var,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%primals_1, %mean), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_2, %sub), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sqrt, 1e-06), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%mul, %add), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%div, %primals_3), kwargs = {})
triton_poi_fused_add_div_mean_mul_std_sub_0 = async_compile.triton('triton_poi_fused_add_div_mean_mul_std_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: '*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_mean_mul_std_sub_0', '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_div_mean_mul_std_sub_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
x2 = xindex
x1 = (xindex // 4)
tmp0 = tl.load(in_ptr0 + (x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2), xmask)
tmp2 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp8 = tmp6 + tmp7
tmp9 = 4.0
tmp10 = tmp8 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp0 * tmp11
tmp13 = tmp2 - tmp10
tmp14 = tmp13 * tmp13
tmp15 = tmp3 - tmp10
tmp16 = tmp15 * tmp15
tmp17 = tmp14 + tmp16
tmp18 = tmp5 - tmp10
tmp19 = tmp18 * tmp18
tmp20 = tmp17 + tmp19
tmp21 = tmp7 - tmp10
tmp22 = tmp21 * tmp21
tmp23 = tmp20 + tmp22
tmp24 = 3.0
tmp25 = tmp23 / tmp24
tmp26 = libdevice.sqrt(tmp25)
tmp27 = 1e-06
tmp28 = tmp26 + tmp27
tmp29 = tmp12 / tmp28
tmp31 = tmp29 + tmp30
tl.store(out_ptr0 + (x2), tmp31, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/5j/c5jlcloqxxiimmj2g6mcsdngchc2t4okjea7crlmnmc6tvuksxsd.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=1] = 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 = (%view_6, 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=[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_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 = 256
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x4 = xindex
x0 = xindex % 4
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 + (x4), tmp6, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/em/cem665hvpdhqghs2hb665fzk7y5lc4maw4gcdl6y6fforpi5efqx.py
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.view]
# Source node to ATen node mapping:
# linear_1 => view_7
# Graph fragment:
# %view_7 : [num_users=2] = call_function[target=torch.ops.aten.reshape.default](args = (%view_6, [64, 4]), kwargs = {})
triton_poi_fused_view_2 = async_compile.triton('triton_poi_fused_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=[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_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_view_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
x0 = xindex % 4
x1 = (xindex // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + (4*x1) + (16*((x1 % 4) // 4)) + (64*(((4*((x1 // 4) % 4)) + (x1 % 4)) // 16))), xmask)
tl.store(out_ptr0 + (x2), tmp0, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/5q/c5qmnkuxezgezseizmolw3mx24fyy6xp3cfoz3egpqwcprxgwjre.py
# Topologically Sorted Source Nodes: [add_2], Original ATen: [aten.add]
# Source node to ATen node mapping:
# add_2 => add_2
# Graph fragment:
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_8, %primals_1), kwargs = {})
triton_poi_fused_add_3 = async_compile.triton('triton_poi_fused_add_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: '*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_3', '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_3(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, primals_6, primals_7 = 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, ))
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, ))
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, std, sub, mul, add, truediv, add_1], Original ATen: [aten.mean, aten.std, aten.sub, aten.mul, aten.add, aten.div]
stream0 = get_raw_stream(0)
triton_poi_fused_add_div_mean_mul_std_sub_0.run(primals_2, primals_1, primals_3, buf0, 256, grid=grid(256), stream=stream0)
del primals_2
del primals_3
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(buf0, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
buf2 = 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.bool)
# Topologically Sorted Source Nodes: [relu], Original ATen: [aten.relu, aten.threshold_backward]
triton_poi_fused_relu_threshold_backward_1.run(buf2, primals_5, buf6, 256, grid=grid(256), stream=stream0)
del primals_5
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.view]
triton_poi_fused_view_2.run(buf2, buf3, 256, grid=grid(256), stream=stream0)
buf4 = reinterpret_tensor(buf2, (64, 4), (4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(buf3, reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf4)
buf5 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0); del buf4 # reuse
# Topologically Sorted Source Nodes: [add_2], Original ATen: [aten.add]
triton_poi_fused_add_3.run(buf5, primals_7, primals_1, 256, grid=grid(256), stream=stream0)
del primals_7
return (buf5, primals_1, reinterpret_tensor(buf0, (64, 4), (4, 1), 0), buf3, primals_6, buf6, 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, ), (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)
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.optim
class LayerNorm(nn.Module):
def __init__(self, features, eps=1e-06):
super(LayerNorm, self).__init__()
self.a_2 = nn.Parameter(torch.ones(features))
self.b_2 = nn.Parameter(torch.zeros(features))
self.eps = eps
def forward(self, x):
mean = x.mean(-1, keepdim=True)
std = x.std(-1, keepdim=True)
return self.a_2 * (x - mean) / (std + self.eps) + self.b_2
class PositionwiseFeedForward(nn.Module):
""" A two-layer Feed-Forward-Network with residual layer norm.
Args:
size (int): the size of input for the first-layer of the FFN.
hidden_size (int): the hidden layer size of the second-layer
of the FNN.
dropout (float): dropout probability(0-1.0).
"""
def __init__(self, size, hidden_size, dropout=0.1):
super(PositionwiseFeedForward, self).__init__()
self.w_1 = nn.Linear(size, hidden_size)
self.w_2 = nn.Linear(hidden_size, size)
self.layer_norm = LayerNorm(size)
self.dropout_1 = nn.Dropout(dropout, inplace=True)
self.relu = nn.ReLU(inplace=True)
self.dropout_2 = nn.Dropout(dropout)
def forward(self, x):
inter = self.dropout_1(self.relu(self.w_1(self.layer_norm(x))))
output = self.dropout_2(self.w_2(inter))
return output + x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'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 import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice
import torch.nn as nn
import torch.optim
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_div_mean_mul_std_sub_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
x2 = xindex
x1 = xindex // 4
tmp0 = tl.load(in_ptr0 + x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x2, xmask)
tmp2 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last')
tmp30 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp4 = tmp2 + tmp3
tmp6 = tmp4 + tmp5
tmp8 = tmp6 + tmp7
tmp9 = 4.0
tmp10 = tmp8 / tmp9
tmp11 = tmp1 - tmp10
tmp12 = tmp0 * tmp11
tmp13 = tmp2 - tmp10
tmp14 = tmp13 * tmp13
tmp15 = tmp3 - tmp10
tmp16 = tmp15 * tmp15
tmp17 = tmp14 + tmp16
tmp18 = tmp5 - tmp10
tmp19 = tmp18 * tmp18
tmp20 = tmp17 + tmp19
tmp21 = tmp7 - tmp10
tmp22 = tmp21 * tmp21
tmp23 = tmp20 + tmp22
tmp24 = 3.0
tmp25 = tmp23 / tmp24
tmp26 = libdevice.sqrt(tmp25)
tmp27 = 1e-06
tmp28 = tmp26 + tmp27
tmp29 = tmp12 / tmp28
tmp31 = tmp29 + tmp30
tl.store(out_ptr0 + x2, tmp31, xmask)
@triton.jit
def triton_poi_fused_relu_threshold_backward_1(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
x4 = xindex
x0 = xindex % 4
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 + x4, tmp6, xmask)
@triton.jit
def triton_poi_fused_view_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
x0 = xindex % 4
x1 = xindex // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + (x0 + 4 * x1 + 16 * (x1 % 4 // 4) + 64 * ((4 *
(x1 // 4 % 4) + x1 % 4) // 16)), xmask)
tl.store(out_ptr0 + x2, tmp0, xmask)
@triton.jit
def triton_poi_fused_add_3(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, primals_6,
primals_7) = 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,))
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,))
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_mean_mul_std_sub_0[grid(256)](primals_2,
primals_1, primals_3, buf0, 256, XBLOCK=128, num_warps=4,
num_stages=1)
del primals_2
del primals_3
buf1 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf0, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
buf2 = 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.bool)
triton_poi_fused_relu_threshold_backward_1[grid(256)](buf2,
primals_5, buf6, 256, XBLOCK=128, num_warps=4, num_stages=1)
del primals_5
buf3 = empty_strided_cuda((64, 4), (4, 1), torch.float32)
triton_poi_fused_view_2[grid(256)](buf2, buf3, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf4 = reinterpret_tensor(buf2, (64, 4), (4, 1), 0)
del buf2
extern_kernels.mm(buf3, reinterpret_tensor(primals_6, (4, 4), (1, 4
), 0), out=buf4)
buf5 = reinterpret_tensor(buf4, (4, 4, 4, 4), (64, 16, 4, 1), 0)
del buf4
triton_poi_fused_add_3[grid(256)](buf5, primals_7, primals_1, 256,
XBLOCK=128, num_warps=4, num_stages=1)
del primals_7
return buf5, primals_1, reinterpret_tensor(buf0, (64, 4), (4, 1), 0
), buf3, primals_6, buf6, primals_4
class LayerNorm(nn.Module):
def __init__(self, features, eps=1e-06):
super(LayerNorm, self).__init__()
self.a_2 = nn.Parameter(torch.ones(features))
self.b_2 = nn.Parameter(torch.zeros(features))
self.eps = eps
def forward(self, x):
mean = x.mean(-1, keepdim=True)
std = x.std(-1, keepdim=True)
return self.a_2 * (x - mean) / (std + self.eps) + self.b_2
class PositionwiseFeedForwardNew(nn.Module):
""" A two-layer Feed-Forward-Network with residual layer norm.
Args:
size (int): the size of input for the first-layer of the FFN.
hidden_size (int): the hidden layer size of the second-layer
of the FNN.
dropout (float): dropout probability(0-1.0).
"""
def __init__(self, size, hidden_size, dropout=0.1):
super(PositionwiseFeedForwardNew, self).__init__()
self.w_1 = nn.Linear(size, hidden_size)
self.w_2 = nn.Linear(hidden_size, size)
self.layer_norm = LayerNorm(size)
self.dropout_1 = nn.Dropout(dropout, inplace=True)
self.relu = nn.ReLU(inplace=True)
self.dropout_2 = nn.Dropout(dropout)
def forward(self, input_0):
primals_4 = self.w_1.weight
primals_2 = self.w_1.bias
primals_6 = self.w_2.weight
primals_3 = self.w_2.bias
primals_5 = self.layer_norm.a_2
primals_7 = self.layer_norm.b_2
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| howardchenhd/Transformer-pytorch | PositionwiseFeedForward | false | 6,823 | [
"MIT"
] | 1 | ae71ed5767272feb7e717be6d5bfce46f80ec57a | https://github.com/howardchenhd/Transformer-pytorch/tree/ae71ed5767272feb7e717be6d5bfce46f80ec57a | import torch
import torch.nn as nn
import torch.optim
class LayerNorm(nn.Module):
def __init__(self, features, eps=1e-06):
super().__init__()
self.a_2 = nn.Parameter(torch.ones(features))
self.b_2 = nn.Parameter(torch.zeros(features))
self.eps = eps
def forward(self, x):
mean = x.mean(-1, keepdim=True)
std = x.std(-1, keepdim=True)
return self.a_2 * (x - mean) / (std + self.eps) + self.b_2
class Model(nn.Module):
""" A two-layer Feed-Forward-Network with residual layer norm.
Args:
size (int): the size of input for the first-layer of the FFN.
hidden_size (int): the hidden layer size of the second-layer
of the FNN.
dropout (float): dropout probability(0-1.0).
"""
def __init__(self, size, hidden_size, dropout=0.1):
super().__init__()
self.w_1 = nn.Linear(size, hidden_size)
self.w_2 = nn.Linear(hidden_size, size)
self.layer_norm = LayerNorm(size)
self.dropout_1 = nn.Dropout(dropout, inplace=True)
self.relu = nn.ReLU(inplace=True)
self.dropout_2 = nn.Dropout(dropout)
def forward(self, x):
inter = self.dropout_1(self.relu(self.w_1(self.layer_norm(x))))
output = self.dropout_2(self.w_2(inter))
return output + x
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4, 4]
|
MultiHeadedAttention | # 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_4/inductor_cache/7c/c7c7bmvdtfwg2cjdph3ycnfts3mkxkveriaohpvvm4wxz2v7zwbx.py
# Topologically Sorted Source Nodes: [query_up_1, scores], Original ATen: [aten.div, aten.clone]
# Source node to ATen node mapping:
# query_up_1 => div
# scores => clone
# Graph fragment:
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%permute_5, 1.0), kwargs = {})
# %clone : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_div_0 = async_compile.triton('triton_poi_fused_clone_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=[16, 4], 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_clone_div_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_clone_div_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + (4*y3)), tmp4, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/mh/cmhet4vfl4jlxtge4zzaaa2nugvxpr5f4ge7rs72qwe2aow7vxwy.py
# Topologically Sorted Source Nodes: [scores], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# scores => 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=[16, 4], 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_clone_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_clone_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*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 + (4*y3)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/tt/cttmvktt3m2x2nl56afa7l3abaxt7wlehowakdzngkhgs35f3n7u.py
# Topologically Sorted Source Nodes: [attn], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# attn => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_11, [-1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_11, %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=[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 = 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_4/inductor_cache/ry/cryn7ntc2gpkbfzbre3xh7lffx7zkbskw6oihbzsekkgajmdbki6.py
# Topologically Sorted Source Nodes: [attn], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# 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=3] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp, %sum_1), 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')
# kernel path: runs/run_shard_4/inductor_cache/6t/c6t5a5ere3lqjiu7zh3uu4oxmpdoujdaqqmeunxqapgzo4m74uav.py
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# contiguous => clone_4
# Graph fragment:
# %clone_4 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_4 = async_compile.triton('triton_poi_fused_clone_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=[16, 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_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_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (4*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/je/cjeiy6xodw4t4ib6usignjcyib4nntuiic5meeubxslxfkxcw5wj.py
# Topologically Sorted Source Nodes: [top_attn], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# top_attn => clone_5
# Graph fragment:
# %clone_5 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%select,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_5 = async_compile.triton('triton_poi_fused_clone_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: '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_5', '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_5(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, 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), (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, ), (1, ))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4, ), (1, ))
assert_size_stride(primals_7, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (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: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf0)
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_7, (16, 4), (4, 1), 0), reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf1)
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf2)
del primals_8
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [query_up_1, scores], Original ATen: [aten.div, aten.clone]
stream0 = get_raw_stream(0)
triton_poi_fused_clone_div_0.run(buf2, primals_9, buf3, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_9
buf4 = reinterpret_tensor(buf2, (4, 4, 1, 4), (16, 4, 4, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [scores], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf0, primals_4, buf4, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_4
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [scores], 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: [attn], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.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: [attn], Original ATen: [aten._softmax]
triton_poi_fused__softmax_3.run(buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf6
buf8 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [matmul_1], Original ATen: [aten.clone]
triton_poi_fused_clone_1.run(buf1, primals_6, buf8, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_6
buf9 = reinterpret_tensor(buf1, (16, 4, 1), (4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [matmul_1], 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)
buf10 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [contiguous], Original ATen: [aten.clone]
triton_poi_fused_clone_4.run(buf9, buf10, 16, 4, grid=grid(16, 4), stream=stream0)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [output], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_11, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf11)
del primals_11
buf12 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [top_attn], Original ATen: [aten.clone]
triton_poi_fused_clone_5.run(buf7, buf12, 64, grid=grid(64), stream=stream0)
return (reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0), buf12, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0), reinterpret_tensor(primals_7, (16, 4), (4, 1), 0), reinterpret_tensor(primals_2, (16, 4), (4, 1), 0), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), primals_10, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 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, 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, ), (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)
primals_7 = rand_strided((4, 4, 4), (16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_8 = rand_strided((4, 4), (4, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((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
from torch.autograd import Variable
import torch.nn as nn
import torch.optim
class MultiHeadedAttention(nn.Module):
"""
Multi-Head Attention module from
"Attention is All You Need"
:cite:`DBLP:journals/corr/VaswaniSPUJGKP17`.
Similar to standard `dot` attention but uses
multiple attention distributions simulataneously
to select relevant items.
.. mermaid::
graph BT
A[key]
B[value]
C[query]
O[output]
subgraph Attn
D[Attn 1]
E[Attn 2]
F[Attn N]
end
A --> D
C --> D
A --> E
C --> E
A --> F
C --> F
D --> O
E --> O
F --> O
B --> O
Also includes several additional tricks.
Args:
head_count (int): number of parallel heads
model_dim (int): the dimension of keys/values/queries,
must be divisible by head_count
dropout (float): dropout parameter
"""
def __init__(self, head_count, model_dim, dropout=0.1):
assert model_dim % head_count == 0
self.dim_per_head = model_dim // head_count
self.model_dim = model_dim
super(MultiHeadedAttention, self).__init__()
self.head_count = head_count
self.linear_keys = nn.Linear(model_dim, head_count * self.dim_per_head)
self.linear_values = nn.Linear(model_dim, head_count * self.
dim_per_head)
self.linear_query = nn.Linear(model_dim, head_count * self.dim_per_head
)
self.sm = nn.Softmax(dim=-1)
self.dropout = nn.Dropout(dropout)
self.final_linear = nn.Linear(model_dim, model_dim)
def _split_heads(self, x):
batch_size = x.size(0)
return x.view(batch_size, -1, self.head_count, self.dim_per_head
).transpose(1, 2)
def _combine_heads(self, x):
""":param x: [batch_size * head_count, seq_len, dim_per_head]"""
seq_len = x.size(2)
return x.transpose(1, 2).contiguous().view(-1, seq_len, self.
head_count * self.dim_per_head)
def compute_cache(self, key, value):
key_up = self._split_heads(self.linear_keys(key))
value_up = self._split_heads(self.linear_values(value))
return [key_up, value_up]
def forward(self, key, value, query, mask=None, cache=None):
"""
Compute the context vector and the attention vectors.
Args:
key (`FloatTensor`): set of `key_len`
key vectors `[batch, key_len, dim]`
value (`FloatTensor`): set of `key_len`
value vectors `[batch, key_len, dim]`
query (`FloatTensor`): set of `query_len`
query vectors `[batch, query_len, dim]`
mask: binary mask indicating which keys have
non-zero attention `[batch, query_len, key_len]`
Returns:
(`FloatTensor`, `FloatTensor`) :
* output context vectors `[batch, query_len, dim]`
* one of the attention vectors `[batch, query_len, key_len]`
"""
batch_size = key.size(0)
dim_per_head = self.dim_per_head
head_count = self.head_count
key_len = key.size(1)
query_len = query.size(1)
if cache is not None:
key_up, value_up = cache
else:
key_up = self._split_heads(self.linear_keys(key))
value_up = self._split_heads(self.linear_values(value))
query_up = self._split_heads(self.linear_query(query))
query_up = query_up / math.sqrt(dim_per_head)
scores = torch.matmul(query_up, key_up.transpose(2, 3))
if mask is not None:
mask = mask.unsqueeze(1).expand_as(scores)
scores = scores.masked_fill(Variable(mask), -1e+18)
attn = self.sm(scores)
drop_attn = self.dropout(attn)
context = self._combine_heads(torch.matmul(drop_attn, value_up))
output = self.final_linear(context)
top_attn = attn.view(batch_size, head_count, query_len, key_len)[:,
0, :, :].contiguous()
return output, top_attn
def get_inputs():
return [torch.rand([4, 4, 4]), torch.rand([4, 4, 4]), torch.rand([4, 4, 4])
]
def get_init_inputs():
return [[], {'head_count': 4, 'model_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._inductor.runtime.triton_helpers import math as tl_math
import torch.nn as nn
import torch.optim
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_div_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + 4 * y3), tmp4, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_1(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel,
YBLOCK: tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * 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 + 4 * y3), tmp2, 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)
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 = 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_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 4 * y3), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_5(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, 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), (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,), (1,))
assert_size_stride(primals_5, (4, 4), (4, 1))
assert_size_stride(primals_6, (4,), (1,))
assert_size_stride(primals_7, (4, 4, 4), (16, 4, 1))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (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.mm(reinterpret_tensor(primals_1, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_3, (4, 4), (1, 4), 0), out=buf0)
del primals_3
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_7, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_5, (4, 4), (1, 4), 0), out=buf1)
del primals_5
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_2, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), out=buf2)
del primals_8
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_clone_div_0[grid(16, 4)](buf2, primals_9, buf3, 16,
4, XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_9
buf4 = reinterpret_tensor(buf2, (4, 4, 1, 4), (16, 4, 4, 1), 0)
del buf2
triton_poi_fused_clone_1[grid(16, 4)](buf0, primals_4, buf4, 16, 4,
XBLOCK=2, YBLOCK=16, num_warps=1, num_stages=1)
del primals_4
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_2[grid(256)](buf5, buf6, 256, XBLOCK=128,
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_3[grid(256)](buf6, buf7, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf6
buf8 = reinterpret_tensor(buf0, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf0
triton_poi_fused_clone_1[grid(16, 4)](buf1, primals_6, buf8, 16, 4,
XBLOCK=2, YBLOCK=16, num_warps=1, num_stages=1)
del primals_6
buf9 = reinterpret_tensor(buf1, (16, 4, 1), (4, 1, 1), 0)
del buf1
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 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_4[grid(16, 4)](buf9, buf10, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0)
del buf9
extern_kernels.addmm(primals_11, reinterpret_tensor(buf10, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_10, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf11)
del primals_11
buf12 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_clone_5[grid(64)](buf7, buf12, 64, XBLOCK=64,
num_warps=1, num_stages=1)
return reinterpret_tensor(buf11, (4, 4, 4), (16, 4, 1), 0
), buf12, reinterpret_tensor(primals_1, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_7, (16, 4), (4, 1), 0
), reinterpret_tensor(primals_2, (16, 4), (4, 1), 0
), buf7, reinterpret_tensor(buf10, (16, 4), (4, 1), 0
), primals_10, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0)
class MultiHeadedAttentionNew(nn.Module):
"""
Multi-Head Attention module from
"Attention is All You Need"
:cite:`DBLP:journals/corr/VaswaniSPUJGKP17`.
Similar to standard `dot` attention but uses
multiple attention distributions simulataneously
to select relevant items.
.. mermaid::
graph BT
A[key]
B[value]
C[query]
O[output]
subgraph Attn
D[Attn 1]
E[Attn 2]
F[Attn N]
end
A --> D
C --> D
A --> E
C --> E
A --> F
C --> F
D --> O
E --> O
F --> O
B --> O
Also includes several additional tricks.
Args:
head_count (int): number of parallel heads
model_dim (int): the dimension of keys/values/queries,
must be divisible by head_count
dropout (float): dropout parameter
"""
def __init__(self, head_count, model_dim, dropout=0.1):
assert model_dim % head_count == 0
self.dim_per_head = model_dim // head_count
self.model_dim = model_dim
super(MultiHeadedAttentionNew, self).__init__()
self.head_count = head_count
self.linear_keys = nn.Linear(model_dim, head_count * self.dim_per_head)
self.linear_values = nn.Linear(model_dim, head_count * self.
dim_per_head)
self.linear_query = nn.Linear(model_dim, head_count * self.dim_per_head
)
self.sm = nn.Softmax(dim=-1)
self.dropout = nn.Dropout(dropout)
self.final_linear = nn.Linear(model_dim, model_dim)
def _split_heads(self, x):
batch_size = x.size(0)
return x.view(batch_size, -1, self.head_count, self.dim_per_head
).transpose(1, 2)
def _combine_heads(self, x):
""":param x: [batch_size * head_count, seq_len, dim_per_head]"""
seq_len = x.size(2)
return x.transpose(1, 2).contiguous().view(-1, seq_len, self.
head_count * self.dim_per_head)
def compute_cache(self, key, value):
key_up = self._split_heads(self.linear_keys(key))
value_up = self._split_heads(self.linear_values(value))
return [key_up, value_up]
def forward(self, input_0, input_1, input_2):
primals_3 = self.linear_keys.weight
primals_4 = self.linear_keys.bias
primals_5 = self.linear_values.weight
primals_6 = self.linear_values.bias
primals_8 = self.linear_query.weight
primals_9 = self.linear_query.bias
primals_10 = self.final_linear.weight
primals_11 = self.final_linear.bias
primals_1 = input_0
primals_2 = input_1
primals_7 = 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], output[1]
| howardchenhd/Transformer-pytorch | MultiHeadedAttention | false | 6,824 | [
"MIT"
] | 1 | ae71ed5767272feb7e717be6d5bfce46f80ec57a | https://github.com/howardchenhd/Transformer-pytorch/tree/ae71ed5767272feb7e717be6d5bfce46f80ec57a | import math
import torch
from torch.autograd import Variable
import torch.nn as nn
import torch.optim
class Model(nn.Module):
"""
Multi-Head Attention module from
"Attention is All You Need"
:cite:`DBLP:journals/corr/VaswaniSPUJGKP17`.
Similar to standard `dot` attention but uses
multiple attention distributions simulataneously
to select relevant items.
.. mermaid::
graph BT
A[key]
B[value]
C[query]
O[output]
subgraph Attn
D[Attn 1]
E[Attn 2]
F[Attn N]
end
A --> D
C --> D
A --> E
C --> E
A --> F
C --> F
D --> O
E --> O
F --> O
B --> O
Also includes several additional tricks.
Args:
head_count (int): number of parallel heads
model_dim (int): the dimension of keys/values/queries,
must be divisible by head_count
dropout (float): dropout parameter
"""
def __init__(self, head_count, model_dim, dropout=0.1):
assert model_dim % head_count == 0
self.dim_per_head = model_dim // head_count
self.model_dim = model_dim
super().__init__()
self.head_count = head_count
self.linear_keys = nn.Linear(model_dim, head_count * self.dim_per_head)
self.linear_values = nn.Linear(model_dim, head_count * self.
dim_per_head)
self.linear_query = nn.Linear(model_dim, head_count * self.dim_per_head
)
self.sm = nn.Softmax(dim=-1)
self.dropout = nn.Dropout(dropout)
self.final_linear = nn.Linear(model_dim, model_dim)
def _split_heads(self, x):
batch_size = x.size(0)
return x.view(batch_size, -1, self.head_count, self.dim_per_head
).transpose(1, 2)
def _combine_heads(self, x):
""":param x: [batch_size * head_count, seq_len, dim_per_head]"""
seq_len = x.size(2)
return x.transpose(1, 2).contiguous().view(-1, seq_len, self.
head_count * self.dim_per_head)
def compute_cache(self, key, value):
key_up = self._split_heads(self.linear_keys(key))
value_up = self._split_heads(self.linear_values(value))
return [key_up, value_up]
def forward(self, key, value, query, mask=None, cache=None):
"""
Compute the context vector and the attention vectors.
Args:
key (`FloatTensor`): set of `key_len`
key vectors `[batch, key_len, dim]`
value (`FloatTensor`): set of `key_len`
value vectors `[batch, key_len, dim]`
query (`FloatTensor`): set of `query_len`
query vectors `[batch, query_len, dim]`
mask: binary mask indicating which keys have
non-zero attention `[batch, query_len, key_len]`
Returns:
(`FloatTensor`, `FloatTensor`) :
* output context vectors `[batch, query_len, dim]`
* one of the attention vectors `[batch, query_len, key_len]`
"""
batch_size = key.size(0)
dim_per_head = self.dim_per_head
head_count = self.head_count
key_len = key.size(1)
query_len = query.size(1)
if cache is not None:
key_up, value_up = cache
else:
key_up = self._split_heads(self.linear_keys(key))
value_up = self._split_heads(self.linear_values(value))
query_up = self._split_heads(self.linear_query(query))
query_up = query_up / math.sqrt(dim_per_head)
scores = torch.matmul(query_up, key_up.transpose(2, 3))
if mask is not None:
mask = mask.unsqueeze(1).expand_as(scores)
scores = scores.masked_fill(Variable(mask), -1e+18)
attn = self.sm(scores)
drop_attn = self.dropout(attn)
context = self._combine_heads(torch.matmul(drop_attn, value_up))
output = sel
# ... truncated (>4000 chars) for memory efficiency |
Decoder2 | # 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_4/inductor_cache/qz/cqz2rbvbxkrzxkla4tsvnvl2bfpsyravtu2nuyuef657luu7fvhi.py
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.reflection_pad2d]
# Source node to ATen node mapping:
# out => _unsafe_index, _unsafe_index_1
# Graph fragment:
# %_unsafe_index : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%primals_1, [None, None, %sub_1, None]), kwargs = {})
# %_unsafe_index_1 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index, [None, None, None, %sub_1]), kwargs = {})
triton_poi_fused_reflection_pad2d_0 = async_compile.triton('triton_poi_fused_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=[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_reflection_pad2d_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_reflection_pad2d_0(in_ptr0, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 18432
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x0 = xindex % 6
x1 = (xindex // 6) % 6
x2 = (xindex // 36)
x3 = xindex
tmp0 = tl.load(in_ptr0 + (15 + ((-1)*(tl_math.abs((-3) + (tl_math.abs((-1) + x0))))) + ((-4)*(tl_math.abs((-3) + (tl_math.abs((-1) + x1))))) + (16*x2)), None, eviction_policy='evict_last')
tl.store(out_ptr0 + (x3), tmp0, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/kk/ckkeehad7xvjmxduxmwzjdm4zu3f5inttmd6kj4pju55xhwrnoea.py
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.arange, aten.add, aten.mul, aten._to_copy]
# Source node to ATen node mapping:
# out_3 => add, add_1, convert_element_type, convert_element_type_1, iota_2, mul, mul_1
# Graph fragment:
# %iota_2 : [num_users=1] = call_function[target=torch.ops.prims.iota.default](args = (8,), kwargs = {start: 0, step: 1, dtype: torch.int64, device: cuda:0, requires_grad: False})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%iota_2, 1), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul, 0), kwargs = {})
# %convert_element_type : [num_users=1] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%add, torch.float32), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%convert_element_type, 0.0), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%add_1, 0.5), kwargs = {})
# %convert_element_type_1 : [num_users=3] = call_function[target=torch.ops.prims.convert_element_type.default](args = (%mul_1, torch.int64), kwargs = {})
triton_poi_fused__to_copy_add_arange_mul_1 = async_compile.triton('triton_poi_fused__to_copy_add_arange_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=[8],
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,), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused__to_copy_add_arange_mul_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_add_arange_mul_1(out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 8
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 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tl.store(out_ptr0 + (x0), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/xl/cxlvflabt3qtgizdi4mapr74rn4h32hq5duymp6n4ywbvyv3cceu.py
# Topologically Sorted Source Nodes: [out_1, out_2, out_3, out_4], Original ATen: [aten.convolution, aten.relu, aten._unsafe_index, aten.reflection_pad2d]
# Source node to ATen node mapping:
# out_1 => convolution
# out_2 => relu
# out_3 => _unsafe_index_2
# out_4 => _unsafe_index_3, _unsafe_index_4
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
# %_unsafe_index_2 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%relu, [None, None, %unsqueeze, %convert_element_type_1]), kwargs = {})
# %_unsafe_index_3 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index_2, [None, None, %sub_5, None]), kwargs = {})
# %_unsafe_index_4 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index_3, [None, None, None, %sub_5]), kwargs = {})
triton_poi_fused__unsafe_index_convolution_reflection_pad2d_relu_2 = async_compile.triton('triton_poi_fused__unsafe_index_convolution_reflection_pad2d_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=[32768],
filename=__file__,
triton_meta={'signature': {0: '*i64', 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__unsafe_index_convolution_reflection_pad2d_relu_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__unsafe_index_convolution_reflection_pad2d_relu_2(in_ptr0, in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = (xindex // 10) % 10
x0 = xindex % 10
x4 = (xindex // 100)
x2 = (xindex // 100) % 64
x7 = xindex
tmp0 = tl.load(in_ptr0 + (7 + ((-1)*(tl_math.abs((-7) + (tl_math.abs((-1) + x1)))))), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (7 + ((-1)*(tl_math.abs((-7) + (tl_math.abs((-1) + x0)))))), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr2 + (x2), xmask, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 4, 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_ptr1 + (tmp8 + (4*tmp4) + (16*x4)), xmask, eviction_policy='evict_last')
tmp11 = tmp9 + tmp10
tmp12 = tl.full([1], 0, tl.int32)
tmp13 = triton_helpers.maximum(tmp12, tmp11)
tl.store(out_ptr0 + (x7), tmp13, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/i2/ci2dp6hwk3cwmyt6qycmkdd63v5237ewhy4impa26eqor4clwsfr.py
# Topologically Sorted Source Nodes: [out_5, out_6, out_7], Original ATen: [aten.convolution, aten.relu, aten.reflection_pad2d]
# Source node to ATen node mapping:
# out_5 => convolution_1
# out_6 => relu_1
# out_7 => _unsafe_index_5, _unsafe_index_6
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_4, %primals_4, %primals_5, [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 = {})
# %_unsafe_index_5 : [num_users=1] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%relu_1, [None, None, %sub_5, None]), kwargs = {})
# %_unsafe_index_6 : [num_users=2] = call_function[target=torch.ops.aten._unsafe_index.Tensor](args = (%_unsafe_index_5, [None, None, None, %sub_5]), kwargs = {})
triton_poi_fused_convolution_reflection_pad2d_relu_3 = async_compile.triton('triton_poi_fused_convolution_reflection_pad2d_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=[32768],
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_convolution_reflection_pad2d_relu_3', '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_reflection_pad2d_relu_3(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 10
x1 = (xindex // 10) % 10
x4 = (xindex // 100)
x2 = (xindex // 100) % 64
x5 = xindex
tmp0 = tl.load(in_ptr0 + (63 + ((-1)*(tl_math.abs((-7) + (tl_math.abs((-1) + x0))))) + ((-8)*(tl_math.abs((-7) + (tl_math.abs((-1) + x1))))) + (64*x4)), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (x2), xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(out_ptr0 + (x5), tmp4, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/r6/cr6sve5hpdupp6pjum6t5l4nu5tabvmkgkrspeo4wakvu2tdwpnw.py
# Topologically Sorted Source Nodes: [out_8], Original ATen: [aten.convolution]
# Source node to ATen node mapping:
# out_8 => convolution_2
# Graph fragment:
# %convolution_2 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_6, %primals_6, %primals_7, [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=[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_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 = 768
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = (xindex // 64) % 3
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_4/inductor_cache/vo/cvo263v6l4c2f2fnxsfyok2ve5gfukurbkglyypoajhqnljgqwhm.py
# Topologically Sorted Source Nodes: [out_5, out_6], Original ATen: [aten.convolution, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# out_5 => convolution_1
# out_6 => relu_1
# Graph fragment:
# %convolution_1 : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_4, %primals_4, %primals_5, [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 = {})
# %le_18 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu_1, 0), kwargs = {})
triton_poi_fused_convolution_relu_threshold_backward_5 = async_compile.triton('triton_poi_fused_convolution_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=[16384],
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_convolution_relu_threshold_backward_5', '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_threshold_backward_5(in_ptr0, in_ptr1, out_ptr0, xnumel, XBLOCK : tl.constexpr):
xnumel = 16384
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = tl.full([XBLOCK], True, tl.int1)
x3 = xindex
x1 = (xindex // 64) % 64
tmp0 = tl.load(in_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr1 + (x1), 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(out_ptr0 + (x3), tmp6, None)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/yt/cytsrygmfas2outodyxpichydoq3ospcwkst4js3nduvvj2nmei2.py
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten.convolution, aten.relu, aten.threshold_backward]
# Source node to ATen node mapping:
# out_1 => convolution
# out_2 => relu
# Graph fragment:
# %convolution : [num_users=1] = call_function[target=torch.ops.aten.convolution.default](args = (%_unsafe_index_1, %primals_2, %primals_3, [1, 1], [0, 0], [1, 1], False, [0, 0], 1), kwargs = {})
# %relu : [num_users=2] = call_function[target=torch.ops.aten.relu.default](args = (%convolution,), kwargs = {})
# %le_37 : [num_users=1] = call_function[target=torch.ops.aten.le.Scalar](args = (%relu, 0), kwargs = {})
triton_poi_fused_convolution_relu_threshold_backward_6 = async_compile.triton('triton_poi_fused_convolution_relu_threshold_backward_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=[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_convolution_relu_threshold_backward_6', '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_threshold_backward_6(in_ptr0, in_ptr1, 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)
x3 = xindex
x1 = (xindex // 16) % 64
tmp0 = tl.load(in_ptr0 + (x3), None)
tmp1 = tl.load(in_ptr1 + (x1), 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(out_ptr0 + (x3), 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, primals_6, primals_7 = args
args.clear()
assert_size_stride(primals_1, (4, 128, 4, 4), (2048, 16, 4, 1))
assert_size_stride(primals_2, (64, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_3, (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, (3, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_7, (3, ), (1, ))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 128, 6, 6), (4608, 36, 6, 1), torch.float32)
# Topologically Sorted Source Nodes: [out], Original ATen: [aten.reflection_pad2d]
stream0 = get_raw_stream(0)
triton_poi_fused_reflection_pad2d_0.run(primals_1, buf0, 18432, grid=grid(18432), stream=stream0)
del primals_1
# Topologically Sorted Source Nodes: [out_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, 64, 4, 4), (1024, 16, 4, 1))
buf2 = empty_strided_cuda((8, ), (1, ), torch.int64)
# Topologically Sorted Source Nodes: [out_3], Original ATen: [aten.arange, aten.add, aten.mul, aten._to_copy]
triton_poi_fused__to_copy_add_arange_mul_1.run(buf2, 8, grid=grid(8), stream=stream0)
buf3 = empty_strided_cuda((4, 64, 10, 10), (6400, 100, 10, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_1, out_2, out_3, out_4], Original ATen: [aten.convolution, aten.relu, aten._unsafe_index, aten.reflection_pad2d]
triton_poi_fused__unsafe_index_convolution_reflection_pad2d_relu_2.run(buf2, buf1, primals_3, buf3, 25600, grid=grid(25600), stream=stream0)
# Topologically Sorted Source Nodes: [out_5], Original ATen: [aten.convolution]
buf4 = extern_kernels.convolution(buf3, primals_4, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 64, 8, 8), (4096, 64, 8, 1))
buf5 = empty_strided_cuda((4, 64, 10, 10), (6400, 100, 10, 1), torch.float32)
# Topologically Sorted Source Nodes: [out_5, out_6, out_7], Original ATen: [aten.convolution, aten.relu, aten.reflection_pad2d]
triton_poi_fused_convolution_reflection_pad2d_relu_3.run(buf4, primals_5, buf5, 25600, grid=grid(25600), stream=stream0)
# Topologically Sorted Source Nodes: [out_8], Original ATen: [aten.convolution]
buf6 = extern_kernels.convolution(buf5, primals_6, stride=(1, 1), padding=(0, 0), dilation=(1, 1), transposed=False, output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 3, 8, 8), (192, 64, 8, 1))
buf7 = buf6; del buf6 # reuse
# Topologically Sorted Source Nodes: [out_8], Original ATen: [aten.convolution]
triton_poi_fused_convolution_4.run(buf7, primals_7, 768, grid=grid(768), stream=stream0)
del primals_7
buf8 = empty_strided_cuda((4, 64, 8, 8), (4096, 64, 8, 1), torch.bool)
# Topologically Sorted Source Nodes: [out_5, out_6], Original ATen: [aten.convolution, aten.relu, aten.threshold_backward]
triton_poi_fused_convolution_relu_threshold_backward_5.run(buf4, primals_5, buf8, 16384, grid=grid(16384), stream=stream0)
del buf4
del primals_5
buf9 = empty_strided_cuda((4, 64, 4, 4), (1024, 16, 4, 1), torch.bool)
# Topologically Sorted Source Nodes: [out_1, out_2], Original ATen: [aten.convolution, aten.relu, aten.threshold_backward]
triton_poi_fused_convolution_relu_threshold_backward_6.run(buf1, primals_3, buf9, 4096, grid=grid(4096), stream=stream0)
del buf1
del primals_3
return (buf7, primals_2, primals_4, primals_6, buf0, buf2, 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, 128, 4, 4), (2048, 16, 4, 1), device='cuda:0', dtype=torch.float32)
primals_2 = rand_strided((64, 128, 3, 3), (1152, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_3 = rand_strided((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((3, 64, 3, 3), (576, 9, 3, 1), device='cuda:0', dtype=torch.float32)
primals_7 = rand_strided((3, ), (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
class Decoder2(nn.Module):
def __init__(self):
super(Decoder2, self).__init__()
self.reflecPad5 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv5 = nn.Conv2d(128, 64, 3, 1, 0)
self.relu5 = nn.ReLU(inplace=True)
self.unpool = nn.UpsamplingNearest2d(scale_factor=2)
self.reflecPad6 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv6 = nn.Conv2d(64, 64, 3, 1, 0)
self.relu6 = nn.ReLU(inplace=True)
self.reflecPad7 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv7 = nn.Conv2d(64, 3, 3, 1, 0)
def forward(self, x):
out = self.reflecPad5(x)
out = self.conv5(out)
out = self.relu5(out)
out = self.unpool(out)
out = self.reflecPad6(out)
out = self.conv6(out)
out = self.relu6(out)
out = self.reflecPad7(out)
out = self.conv7(out)
return out
def get_inputs():
return [torch.rand([4, 128, 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
@triton.jit
def triton_poi_fused_reflection_pad2d_0(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)
x0 = xindex % 6
x1 = xindex // 6 % 6
x2 = xindex // 36
x3 = xindex
tmp0 = tl.load(in_ptr0 + (15 + -1 * tl_math.abs(-3 + tl_math.abs(-1 +
x0)) + -4 * tl_math.abs(-3 + tl_math.abs(-1 + x1)) + 16 * x2), None,
eviction_policy='evict_last')
tl.store(out_ptr0 + x3, tmp0, None)
@triton.jit
def triton_poi_fused__to_copy_add_arange_mul_1(out_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 8
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 = 0.5
tmp3 = tmp1 * tmp2
tmp4 = tmp3.to(tl.int32)
tl.store(out_ptr0 + x0, tmp4, xmask)
@triton.jit
def triton_poi_fused__unsafe_index_convolution_reflection_pad2d_relu_2(in_ptr0,
in_ptr1, in_ptr2, out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x1 = xindex // 10 % 10
x0 = xindex % 10
x4 = xindex // 100
x2 = xindex // 100 % 64
x7 = xindex
tmp0 = tl.load(in_ptr0 + (7 + -1 * tl_math.abs(-7 + tl_math.abs(-1 + x1
))), xmask, eviction_policy='evict_last')
tmp5 = tl.load(in_ptr0 + (7 + -1 * tl_math.abs(-7 + tl_math.abs(-1 + x0
))), xmask, eviction_policy='evict_last')
tmp10 = tl.load(in_ptr2 + x2, xmask, eviction_policy='evict_last')
tmp1 = tl.full([XBLOCK], 4, 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_ptr1 + (tmp8 + 4 * tmp4 + 16 * x4), xmask,
eviction_policy='evict_last')
tmp11 = tmp9 + tmp10
tmp12 = tl.full([1], 0, tl.int32)
tmp13 = triton_helpers.maximum(tmp12, tmp11)
tl.store(out_ptr0 + x7, tmp13, xmask)
@triton.jit
def triton_poi_fused_convolution_reflection_pad2d_relu_3(in_ptr0, in_ptr1,
out_ptr0, xnumel, XBLOCK: tl.constexpr):
xnumel = 25600
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x0 = xindex % 10
x1 = xindex // 10 % 10
x4 = xindex // 100
x2 = xindex // 100 % 64
x5 = xindex
tmp0 = tl.load(in_ptr0 + (63 + -1 * tl_math.abs(-7 + tl_math.abs(-1 +
x0)) + -8 * tl_math.abs(-7 + tl_math.abs(-1 + x1)) + 64 * x4),
xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + x2, xmask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = tl.full([1], 0, tl.int32)
tmp4 = triton_helpers.maximum(tmp3, tmp2)
tl.store(out_ptr0 + x5, tmp4, xmask)
@triton.jit
def triton_poi_fused_convolution_4(in_out_ptr0, in_ptr0, xnumel, XBLOCK: tl
.constexpr):
xnumel = 768
xoffset = tl.program_id(0) * XBLOCK
xindex = xoffset + tl.arange(0, XBLOCK)[:]
xmask = xindex < xnumel
x3 = xindex
x1 = xindex // 64 % 3
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_convolution_relu_threshold_backward_5(in_ptr0, in_ptr1,
out_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 // 64 % 64
tmp0 = tl.load(in_ptr0 + x3, None)
tmp1 = tl.load(in_ptr1 + x1, 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(out_ptr0 + x3, tmp6, None)
@triton.jit
def triton_poi_fused_convolution_relu_threshold_backward_6(in_ptr0, in_ptr1,
out_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_ptr0 + x3, None)
tmp1 = tl.load(in_ptr1 + x1, 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(out_ptr0 + x3, tmp6, None)
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, 128, 4, 4), (2048, 16, 4, 1))
assert_size_stride(primals_2, (64, 128, 3, 3), (1152, 9, 3, 1))
assert_size_stride(primals_3, (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, (3, 64, 3, 3), (576, 9, 3, 1))
assert_size_stride(primals_7, (3,), (1,))
with torch.cuda._DeviceGuard(0):
torch.cuda.set_device(0)
buf0 = empty_strided_cuda((4, 128, 6, 6), (4608, 36, 6, 1), torch.
float32)
get_raw_stream(0)
triton_poi_fused_reflection_pad2d_0[grid(18432)](primals_1, buf0,
18432, 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, 64, 4, 4), (1024, 16, 4, 1))
buf2 = empty_strided_cuda((8,), (1,), torch.int64)
triton_poi_fused__to_copy_add_arange_mul_1[grid(8)](buf2, 8, XBLOCK
=8, num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((4, 64, 10, 10), (6400, 100, 10, 1),
torch.float32)
triton_poi_fused__unsafe_index_convolution_reflection_pad2d_relu_2[grid
(25600)](buf2, buf1, primals_3, buf3, 25600, XBLOCK=128,
num_warps=4, num_stages=1)
buf4 = extern_kernels.convolution(buf3, primals_4, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf4, (4, 64, 8, 8), (4096, 64, 8, 1))
buf5 = empty_strided_cuda((4, 64, 10, 10), (6400, 100, 10, 1),
torch.float32)
triton_poi_fused_convolution_reflection_pad2d_relu_3[grid(25600)](buf4,
primals_5, buf5, 25600, XBLOCK=128, num_warps=4, num_stages=1)
buf6 = extern_kernels.convolution(buf5, primals_6, stride=(1, 1),
padding=(0, 0), dilation=(1, 1), transposed=False,
output_padding=(0, 0), groups=1, bias=None)
assert_size_stride(buf6, (4, 3, 8, 8), (192, 64, 8, 1))
buf7 = buf6
del buf6
triton_poi_fused_convolution_4[grid(768)](buf7, primals_7, 768,
XBLOCK=256, num_warps=4, num_stages=1)
del primals_7
buf8 = empty_strided_cuda((4, 64, 8, 8), (4096, 64, 8, 1), torch.bool)
triton_poi_fused_convolution_relu_threshold_backward_5[grid(16384)](
buf4, primals_5, buf8, 16384, XBLOCK=128, num_warps=4, num_stages=1
)
del buf4
del primals_5
buf9 = empty_strided_cuda((4, 64, 4, 4), (1024, 16, 4, 1), torch.bool)
triton_poi_fused_convolution_relu_threshold_backward_6[grid(4096)](buf1
, primals_3, buf9, 4096, XBLOCK=128, num_warps=4, num_stages=1)
del buf1
del primals_3
return (buf7, primals_2, primals_4, primals_6, buf0, buf2, buf3, buf5,
buf8, buf9)
class Decoder2New(nn.Module):
def __init__(self):
super(Decoder2New, self).__init__()
self.reflecPad5 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv5 = nn.Conv2d(128, 64, 3, 1, 0)
self.relu5 = nn.ReLU(inplace=True)
self.unpool = nn.UpsamplingNearest2d(scale_factor=2)
self.reflecPad6 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv6 = nn.Conv2d(64, 64, 3, 1, 0)
self.relu6 = nn.ReLU(inplace=True)
self.reflecPad7 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv7 = nn.Conv2d(64, 3, 3, 1, 0)
def forward(self, input_0):
primals_2 = self.conv5.weight
primals_3 = self.conv5.bias
primals_4 = self.conv6.weight
primals_5 = self.conv6.bias
primals_6 = self.conv7.weight
primals_7 = self.conv7.bias
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4,
primals_5, primals_6, primals_7])
return output[0]
| hologerry/wct_experiment | Decoder2 | false | 6,825 | [
"MIT"
] | 1 | 890d885561dc8df8c4ae732aebd902aa838257e6 | https://github.com/hologerry/wct_experiment/tree/890d885561dc8df8c4ae732aebd902aa838257e6 | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self):
super().__init__()
self.reflecPad5 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv5 = nn.Conv2d(128, 64, 3, 1, 0)
self.relu5 = nn.ReLU(inplace=True)
self.unpool = nn.UpsamplingNearest2d(scale_factor=2)
self.reflecPad6 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv6 = nn.Conv2d(64, 64, 3, 1, 0)
self.relu6 = nn.ReLU(inplace=True)
self.reflecPad7 = nn.ReflectionPad2d((1, 1, 1, 1))
self.conv7 = nn.Conv2d(64, 3, 3, 1, 0)
def forward(self, x):
out = self.reflecPad5(x)
out = self.conv5(out)
out = self.relu5(out)
out = self.unpool(out)
out = self.reflecPad6(out)
out = self.conv6(out)
out = self.relu6(out)
out = self.reflecPad7(out)
out = self.conv7(out)
return out
def get_inputs():
return [torch.rand([4, 128, 4, 4])]
def get_init_inputs():
return []
|
HirarchicalAttention | # 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_4/inductor_cache/nc/cncwsucylpsg2zmlivjfxu6vbd64ztxjndlsix2ysjtby3xohgk4.py
# Topologically Sorted Source Nodes: [u_it], Original ATen: [aten.tanh]
# Source node to ATen node mapping:
# u_it => 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=[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': ['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 = 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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + (x2), tmp3, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/dm/cdmkcxuzpnailvibeivaikqdr4zvashgzwju7qijhq5aizlo3aor.py
# Topologically Sorted Source Nodes: [a_it], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# a_it => amax, exp, sub
# Graph fragment:
# %amax : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_3, [1], True), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_3, %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_4/inductor_cache/kt/cktghousutx6xui2sl2rvevzmb7gkacvfhntjq5n2xzeu7v57oz6.py
# Topologically Sorted Source Nodes: [a_it], Original ATen: [aten._softmax]
# Source node to ATen node mapping:
# a_it => 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=[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_4/inductor_cache/x5/cx5qa4dyw6ydnehpkcpbpdbuv5qdkuwkqx5hgsacgljy6mj5hebt.py
# Topologically Sorted Source Nodes: [a_it, mul, s_i], Original ATen: [aten._softmax, aten.mul, aten.sum]
# Source node to ATen node mapping:
# a_it => div, sum_1
# mul => mul
# s_i => sum_2
# 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 = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_3, %div), kwargs = {})
# %sum_2 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%mul, [1]), kwargs = {})
triton_poi_fused__softmax_mul_sum_3 = async_compile.triton('triton_poi_fused__softmax_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__softmax_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__softmax_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 = 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, (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: [], 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
# Topologically Sorted Source Nodes: [u_it], Original ATen: [aten.tanh]
stream0 = get_raw_stream(0)
triton_poi_fused_tanh_0.run(buf1, primals_2, 256, grid=grid(256), stream=stream0)
del primals_2
buf2 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
# Topologically Sorted Source Nodes: [linear_1], Original ATen: [aten.mm]
extern_kernels.mm(reinterpret_tensor(buf1, (64, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 1), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [a_it], Original ATen: [aten._softmax]
triton_poi_fused__softmax_1.run(buf2, buf3, 64, grid=grid(64), stream=stream0)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
# Topologically Sorted Source Nodes: [a_it], Original ATen: [aten._softmax]
triton_poi_fused__softmax_2.run(buf3, buf4, 64, grid=grid(64), stream=stream0)
buf5 = reinterpret_tensor(buf3, (4, 4, 4), (16, 4, 1), 0); del buf3 # reuse
# Topologically Sorted Source Nodes: [a_it, mul, s_i], Original ATen: [aten._softmax, aten.mul, aten.sum]
triton_poi_fused__softmax_mul_sum_3.run(primals_3, buf4, buf5, 64, grid=grid(64), stream=stream0)
del buf4
return (buf5, primals_3, buf1, buf2, 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, 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((1, 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)
| from torch.nn import Module
import torch
from typing import *
import torch.utils.data
import torch.nn as nn
import torch.onnx.operators
import torch.optim
class HirarchicalAttention(Module):
"""
ref: Hierarchical Attention Networks for Document Classification
"""
def __init__(self, hidden_size: 'int'):
super(HirarchicalAttention, self).__init__()
self.w_linear = nn.Linear(hidden_size, hidden_size)
self.u_w = nn.Linear(hidden_size, 1, bias=False)
def forward(self, input: 'torch.Tensor') ->torch.Tensor:
u_it = torch.tanh(self.w_linear(input))
a_it = torch.softmax(self.u_w(u_it), dim=1)
s_i = (input * a_it).sum(dim=1)
return s_i
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [[], {'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 import triton_helpers
from torch._inductor.runtime.triton_helpers import libdevice, math as tl_math
from torch.nn import Module
from typing import *
import torch.utils.data
import torch.nn as nn
import torch.onnx.operators
import torch.optim
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
):
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 = libdevice.tanh(tmp2)
tl.store(in_out_ptr0 + x2, tmp3, 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__softmax_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 = 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, (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_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
get_raw_stream(0)
triton_poi_fused_tanh_0[grid(256)](buf1, primals_2, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del primals_2
buf2 = empty_strided_cuda((64, 1), (1, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(buf1, (64, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 1), (1, 4), 0), out=buf2)
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_1[grid(64)](buf2, buf3, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf4 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 64), torch.float32)
triton_poi_fused__softmax_2[grid(64)](buf3, buf4, 64, XBLOCK=64,
num_warps=1, num_stages=1)
buf5 = reinterpret_tensor(buf3, (4, 4, 4), (16, 4, 1), 0)
del buf3
triton_poi_fused__softmax_mul_sum_3[grid(64)](primals_3, buf4, buf5,
64, XBLOCK=64, num_warps=1, num_stages=1)
del buf4
return buf5, primals_3, buf1, buf2, primals_4
class HirarchicalAttentionNew(Module):
"""
ref: Hierarchical Attention Networks for Document Classification
"""
def __init__(self, hidden_size: 'int'):
super(HirarchicalAttentionNew, self).__init__()
self.w_linear = nn.Linear(hidden_size, hidden_size)
self.u_w = nn.Linear(hidden_size, 1, bias=False)
def forward(self, input_0):
primals_1 = self.w_linear.weight
primals_2 = self.w_linear.bias
primals_4 = self.u_w.weight
primals_3 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| hrshy0629/naturalcc | HirarchicalAttention | false | 6,826 | [
"MIT"
] | 1 | 9c3329dd8387c8242deb52bf590ebe3ac795f8de | https://github.com/hrshy0629/naturalcc/tree/9c3329dd8387c8242deb52bf590ebe3ac795f8de | from torch.nn import Module
import torch
from typing import *
import torch.utils.data
import torch.nn as nn
import torch.onnx.operators
import torch.optim
class Model(Module):
"""
ref: Hierarchical Attention Networks for Document Classification
"""
def __init__(self, hidden_size: 'int'):
super().__init__()
self.w_linear = nn.Linear(hidden_size, hidden_size)
self.u_w = nn.Linear(hidden_size, 1, bias=False)
def forward(self, input: 'torch.Tensor') ->torch.Tensor:
u_it = torch.tanh(self.w_linear(input))
a_it = torch.softmax(self.u_w(u_it), dim=1)
s_i = (input * a_it).sum(dim=1)
return s_i
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
GCT | # 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_4/inductor_cache/vt/cvtlwsahn5oxfhga3dwjdswt3bhkvbh4zkaufpwgeopelblrwpnw.py
# Topologically Sorted Source Nodes: [pow_1, sum_1, add, pow_2], Original ATen: [aten.pow, aten.sum, aten.add]
# Source node to ATen node mapping:
# add => add
# pow_1 => pow_1
# pow_2 => pow_2
# sum_1 => sum_1
# Graph fragment:
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%primals_1, 2), kwargs = {})
# %sum_1 : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%pow_1, [2, 3], True), kwargs = {})
# %add : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%sum_1, 1e-05), kwargs = {})
# %pow_2 : [num_users=2] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%add, 0.5), kwargs = {})
triton_per_fused_add_pow_sum_0 = async_compile.triton('triton_per_fused_add_pow_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=[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_add_pow_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_add_pow_sum_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 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.sqrt(tmp7)
tl.debug_barrier()
tl.store(in_out_ptr0 + (x0), tmp8, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/bk/cbk3oabqcrqeiz2rnrvk4rxkdmk4ef46ye5mdaspijuw6pzegair.py
# Topologically Sorted Source Nodes: [embedding, pow_3, mean, add_1, pow_4], Original ATen: [aten.mul, aten.pow, aten.mean, aten.add]
# Source node to ATen node mapping:
# add_1 => add_1
# embedding => mul
# mean => mean
# pow_3 => pow_3
# pow_4 => pow_4
# Graph fragment:
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%pow_2, %primals_2), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%mul, 2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_3, [1], True), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, 1e-05), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%add_1, 0.5), kwargs = {})
triton_poi_fused_add_mean_mul_pow_1 = async_compile.triton('triton_poi_fused_add_mean_mul_pow_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: '*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), equal_to_1=())]},
inductor_meta={'autotune_hints': set(), 'kernel_name': 'triton_poi_fused_add_mean_mul_pow_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_mean_mul_pow_1(in_ptr0, in_ptr1, 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 + (4*x0), xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (0))
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp5 = tl.load(in_ptr0 + (1 + (4*x0)), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr1 + (1))
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp11 = tl.load(in_ptr0 + (2 + (4*x0)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr1 + (2))
tmp13 = tl.broadcast_to(tmp12, [XBLOCK])
tmp17 = tl.load(in_ptr0 + (3 + (4*x0)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr1 + (3))
tmp19 = tl.broadcast_to(tmp18, [XBLOCK])
tmp3 = tmp0 * tmp2
tmp4 = tmp3 * tmp3
tmp8 = tmp5 * tmp7
tmp9 = tmp8 * tmp8
tmp10 = tmp4 + tmp9
tmp14 = tmp11 * tmp13
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp20 = tmp17 * tmp19
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp23 = 4.0
tmp24 = tmp22 / tmp23
tmp25 = 1e-05
tmp26 = tmp24 + tmp25
tmp27 = libdevice.sqrt(tmp26)
tl.store(out_ptr0 + (x0), tmp27, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/ni/cnixk7pwsdlgsjxwsmpbrzopu3rzfg26mvtp5cemfxo35lnktysi.py
# Topologically Sorted Source Nodes: [embedding, pow_3, mean, add_1, pow_4, norm, mul_1, add_2, tanh, gate], Original ATen: [aten.mul, aten.pow, aten.mean, aten.add, aten.div, aten.tanh]
# Source node to ATen node mapping:
# add_1 => add_1
# add_2 => add_2
# embedding => mul
# gate => add_3
# mean => mean
# mul_1 => mul_1
# norm => div
# pow_3 => pow_3
# pow_4 => pow_4
# tanh => tanh
# Graph fragment:
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%pow_2, %primals_2), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%mul, 2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_3, [1], True), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, 1e-05), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%add_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_3, %pow_4), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %div), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_4), kwargs = {})
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_2,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%tanh, 1.0), kwargs = {})
triton_poi_fused_add_div_mean_mul_pow_tanh_2 = async_compile.triton('triton_poi_fused_add_div_mean_mul_pow_tanh_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: '*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_div_mean_mul_pow_tanh_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_add_div_mean_mul_pow_tanh_2(in_ptr0, in_ptr1, in_ptr2, in_ptr3, in_ptr4, 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)
tmp1 = tl.load(in_ptr1 + (x0), xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + (x0), xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr3 + (x1), xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + (x0), xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 / tmp4
tmp6 = tmp2 * tmp5
tmp8 = tmp6 + tmp7
tmp9 = libdevice.tanh(tmp8)
tmp10 = 1.0
tmp11 = tmp9 + tmp10
tl.store(out_ptr0 + (x2), tmp11, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/vz/cvzd7uxhjcuuvhocpajasfl7npgg3w4tjtzribzneczbe5yv6xr6.py
# Topologically Sorted Source Nodes: [embedding, pow_3, mean, add_1, pow_4, norm, mul_1, add_2, tanh, gate, mul_2], Original ATen: [aten.mul, aten.pow, aten.mean, aten.add, aten.div, aten.tanh]
# Source node to ATen node mapping:
# add_1 => add_1
# add_2 => add_2
# embedding => mul
# gate => add_3
# mean => mean
# mul_1 => mul_1
# mul_2 => mul_2
# norm => div
# pow_3 => pow_3
# pow_4 => pow_4
# tanh => tanh
# Graph fragment:
# %mul : [num_users=2] = call_function[target=torch.ops.aten.mul.Tensor](args = (%pow_2, %primals_2), kwargs = {})
# %pow_3 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%mul, 2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.dim](args = (%pow_3, [1], True), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mean, 1e-05), kwargs = {})
# %pow_4 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%add_1, 0.5), kwargs = {})
# %div : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%primals_3, %pow_4), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %div), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_4), kwargs = {})
# %tanh : [num_users=1] = call_function[target=torch.ops.aten.tanh.default](args = (%add_2,), kwargs = {})
# %add_3 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%tanh, 1.0), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%primals_1, %add_3), kwargs = {})
triton_poi_fused_add_div_mean_mul_pow_tanh_3 = async_compile.triton('triton_poi_fused_add_div_mean_mul_pow_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=[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_div_mean_mul_pow_tanh_3', '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_div_mean_mul_pow_tanh_3(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
x2 = xindex
x1 = (xindex // 16)
tmp0 = tl.load(in_ptr0 + (x2), xmask)
tmp1 = tl.load(in_ptr1 + (x1), 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):
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, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_3, (1, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_4, (1, 4, 1, 1), (4, 1, 1, 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 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [pow_1, sum_1, add, pow_2], Original ATen: [aten.pow, aten.sum, aten.add]
stream0 = get_raw_stream(0)
triton_per_fused_add_pow_sum_0.run(buf1, primals_1, 16, 16, grid=grid(16), stream=stream0)
buf2 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
# Topologically Sorted Source Nodes: [embedding, pow_3, mean, add_1, pow_4], Original ATen: [aten.mul, aten.pow, aten.mean, aten.add]
triton_poi_fused_add_mean_mul_pow_1.run(buf1, primals_2, buf2, 4, grid=grid(4), stream=stream0)
buf3 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
# Topologically Sorted Source Nodes: [embedding, pow_3, mean, add_1, pow_4, norm, mul_1, add_2, tanh, gate], Original ATen: [aten.mul, aten.pow, aten.mean, aten.add, aten.div, aten.tanh]
triton_poi_fused_add_div_mean_mul_pow_tanh_2.run(buf1, primals_2, primals_3, buf2, primals_4, buf3, 16, grid=grid(16), stream=stream0)
del buf2
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [embedding, pow_3, mean, add_1, pow_4, norm, mul_1, add_2, tanh, gate, mul_2], Original ATen: [aten.mul, aten.pow, aten.mean, aten.add, aten.div, aten.tanh]
triton_poi_fused_add_div_mean_mul_pow_tanh_3.run(primals_1, buf3, buf4, 256, grid=grid(256), stream=stream0)
del buf3
return (buf4, primals_1, primals_2, primals_3, primals_4, 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, 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, 4, 1, 1), (4, 1, 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)
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 GCT(nn.Module):
def __init__(self, num_channels, epsilon=1e-05, mode='l2', after_relu=False
):
super(GCT, self).__init__()
self.alpha = nn.Parameter(torch.ones(1, num_channels, 1, 1))
self.gamma = nn.Parameter(torch.zeros(1, num_channels, 1, 1))
self.beta = nn.Parameter(torch.zeros(1, num_channels, 1, 1))
self.epsilon = epsilon
self.mode = mode
self.after_relu = after_relu
def forward(self, x):
if self.mode == 'l2':
embedding = (x.pow(2).sum((2, 3), keepdim=True) + self.epsilon
).pow(0.5) * self.alpha
norm = self.gamma / (embedding.pow(2).mean(dim=1, keepdim=True) +
self.epsilon).pow(0.5)
elif self.mode == 'l1':
if not self.after_relu:
_x = torch.abs(x)
else:
_x = x
embedding = _x.sum((2, 3), keepdim=True) * self.alpha
norm = self.gamma / (torch.abs(embedding).mean(dim=1, keepdim=
True) + self.epsilon)
else:
None
exit()
gate = 1.0 + torch.tanh(embedding * norm + self.beta)
return x * gate
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 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_add_pow_sum_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 = tmp0 * tmp0
tmp2 = tl.broadcast_to(tmp1, [XBLOCK, RBLOCK])
tmp4 = tl.where(xmask, tmp2, 0)
tmp5 = tl.sum(tmp4, 1)[:, None]
tmp6 = 1e-05
tmp7 = tmp5 + tmp6
tmp8 = libdevice.sqrt(tmp7)
tl.debug_barrier()
tl.store(in_out_ptr0 + x0, tmp8, xmask)
@triton.jit
def triton_poi_fused_add_mean_mul_pow_1(in_ptr0, in_ptr1, 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 + 4 * x0, xmask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + 0)
tmp2 = tl.broadcast_to(tmp1, [XBLOCK])
tmp5 = tl.load(in_ptr0 + (1 + 4 * x0), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr1 + 1)
tmp7 = tl.broadcast_to(tmp6, [XBLOCK])
tmp11 = tl.load(in_ptr0 + (2 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp12 = tl.load(in_ptr1 + 2)
tmp13 = tl.broadcast_to(tmp12, [XBLOCK])
tmp17 = tl.load(in_ptr0 + (3 + 4 * x0), xmask, eviction_policy='evict_last'
)
tmp18 = tl.load(in_ptr1 + 3)
tmp19 = tl.broadcast_to(tmp18, [XBLOCK])
tmp3 = tmp0 * tmp2
tmp4 = tmp3 * tmp3
tmp8 = tmp5 * tmp7
tmp9 = tmp8 * tmp8
tmp10 = tmp4 + tmp9
tmp14 = tmp11 * tmp13
tmp15 = tmp14 * tmp14
tmp16 = tmp10 + tmp15
tmp20 = tmp17 * tmp19
tmp21 = tmp20 * tmp20
tmp22 = tmp16 + tmp21
tmp23 = 4.0
tmp24 = tmp22 / tmp23
tmp25 = 1e-05
tmp26 = tmp24 + tmp25
tmp27 = libdevice.sqrt(tmp26)
tl.store(out_ptr0 + x0, tmp27, xmask)
@triton.jit
def triton_poi_fused_add_div_mean_mul_pow_tanh_2(in_ptr0, in_ptr1, in_ptr2,
in_ptr3, in_ptr4, 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)
tmp1 = tl.load(in_ptr1 + x0, xmask, eviction_policy='evict_last')
tmp3 = tl.load(in_ptr2 + x0, xmask, eviction_policy='evict_last')
tmp4 = tl.load(in_ptr3 + x1, xmask, eviction_policy='evict_last')
tmp7 = tl.load(in_ptr4 + x0, xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tmp5 = tmp3 / tmp4
tmp6 = tmp2 * tmp5
tmp8 = tmp6 + tmp7
tmp9 = libdevice.tanh(tmp8)
tmp10 = 1.0
tmp11 = tmp9 + tmp10
tl.store(out_ptr0 + x2, tmp11, xmask)
@triton.jit
def triton_poi_fused_add_div_mean_mul_pow_tanh_3(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
x2 = xindex
x1 = xindex // 16
tmp0 = tl.load(in_ptr0 + x2, xmask)
tmp1 = tl.load(in_ptr1 + x1, xmask, eviction_policy='evict_last')
tmp2 = tmp0 * tmp1
tl.store(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, 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, 4, 1, 1), (4, 1, 1, 1))
assert_size_stride(primals_4, (1, 4, 1, 1), (4, 1, 1, 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 = reinterpret_tensor(buf0, (4, 4, 1, 1), (4, 1, 1, 1), 0)
del buf0
get_raw_stream(0)
triton_per_fused_add_pow_sum_0[grid(16)](buf1, primals_1, 16, 16,
XBLOCK=1, num_warps=2, num_stages=1)
buf2 = empty_strided_cuda((4, 1, 1, 1), (1, 4, 4, 4), torch.float32)
triton_poi_fused_add_mean_mul_pow_1[grid(4)](buf1, primals_2, buf2,
4, XBLOCK=4, num_warps=1, num_stages=1)
buf3 = empty_strided_cuda((4, 4, 1, 1), (4, 1, 16, 16), torch.float32)
triton_poi_fused_add_div_mean_mul_pow_tanh_2[grid(16)](buf1,
primals_2, primals_3, buf2, primals_4, buf3, 16, XBLOCK=16,
num_warps=1, num_stages=1)
del buf2
buf4 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_add_div_mean_mul_pow_tanh_3[grid(256)](primals_1,
buf3, buf4, 256, XBLOCK=256, num_warps=4, num_stages=1)
del buf3
return buf4, primals_1, primals_2, primals_3, primals_4, buf1
class GCTNew(nn.Module):
def __init__(self, num_channels, epsilon=1e-05, mode='l2', after_relu=False
):
super(GCTNew, self).__init__()
self.alpha = nn.Parameter(torch.ones(1, num_channels, 1, 1))
self.gamma = nn.Parameter(torch.zeros(1, num_channels, 1, 1))
self.beta = nn.Parameter(torch.zeros(1, num_channels, 1, 1))
self.epsilon = epsilon
self.mode = mode
self.after_relu = after_relu
def forward(self, input_0):
primals_2 = self.alpha
primals_3 = self.gamma
primals_4 = self.beta
primals_1 = input_0
output = call([primals_1, primals_2, primals_3, primals_4])
return output[0]
| huanglf714/COMatchNet | GCT | false | 6,827 | [
"Apache-2.0"
] | 1 | 79023f5be65d354eb9bdac026d7e0d73110bc4aa | https://github.com/huanglf714/COMatchNet/tree/79023f5be65d354eb9bdac026d7e0d73110bc4aa | import torch
import torch.nn as nn
class Model(nn.Module):
def __init__(self, num_channels, epsilon=1e-05, mode='l2', after_relu=False
):
super().__init__()
self.alpha = nn.Parameter(torch.ones(1, num_channels, 1, 1))
self.gamma = nn.Parameter(torch.zeros(1, num_channels, 1, 1))
self.beta = nn.Parameter(torch.zeros(1, num_channels, 1, 1))
self.epsilon = epsilon
self.mode = mode
self.after_relu = after_relu
def forward(self, x):
if self.mode == 'l2':
embedding = (x.pow(2).sum((2, 3), keepdim=True) + self.epsilon
).pow(0.5) * self.alpha
norm = self.gamma / (embedding.pow(2).mean(dim=1, keepdim=True) +
self.epsilon).pow(0.5)
elif self.mode == 'l1':
if not self.after_relu:
_x = torch.abs(x)
else:
_x = x
embedding = _x.sum((2, 3), keepdim=True) * self.alpha
norm = self.gamma / (torch.abs(embedding).mean(dim=1, keepdim=
True) + self.epsilon)
else:
None
exit()
gate = 1.0 + torch.tanh(embedding * norm + self.beta)
return x * gate
def get_inputs():
return [torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return [4]
|
BertAttention | # 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_4/inductor_cache/x2/cx2hdvwyo7m5jvhhvtugzxqvmy6z4nsfhkkjhvgzbbm3cb6dsum2.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %mul_scalar : [num_users=1] = call_function[target=torch.ops.aten.mul.Scalar](args = (%permute_default, 1.0), kwargs = {})
# %clone_default : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_default,), kwargs = {memory_format: torch.contiguous_format})
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=[16, 4], 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_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_0(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + (y0), ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + (4*y3)), tmp4, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/5j/c5jll3kxtd32cl7pwubrb5oky2mtzckfgip2xbwad7crvvp4zk4r.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %amax_default : [num_users=1] = call_function[target=torch.ops.aten.amax.default](args = (%view_default_2, [-1], True), kwargs = {})
# %sub_tensor : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%view_default_2, %amax_default), kwargs = {})
# %exp_default : [num_users=2] = call_function[target=torch.ops.aten.exp.default](args = (%sub_tensor,), kwargs = {})
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=[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_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_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_4/inductor_cache/kt/cktnex5febczl2ac6zugjmcksgsd5kjdufazv65vtepuwob3cb7a.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %sum_dim_int_list : [num_users=1] = call_function[target=torch.ops.aten.sum.dim_IntList](args = (%exp_default, [-1], True), kwargs = {})
# %div_tensor : [num_users=1] = call_function[target=torch.ops.aten.div.Tensor](args = (%exp_default, %sum_dim_int_list), kwargs = {})
# %eq_scalar : [num_users=1] = call_function[target=torch.ops.aten.eq.Scalar](args = (%view_default_2, -inf), kwargs = {})
# %logical_not_default : [num_users=1] = call_function[target=torch.ops.aten.logical_not.default](args = (%eq_scalar,), kwargs = {})
# %any_dim : [num_users=1] = call_function[target=torch.ops.aten.any.dim](args = (%logical_not_default, -1, True), kwargs = {})
# %logical_not_default_1 : [num_users=1] = call_function[target=torch.ops.aten.logical_not.default](args = (%any_dim,), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([4, 4, 4, 4], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %where_self : [num_users=2] = call_function[target=torch.ops.aten.where.self](args = (%logical_not_default_1, %full_default, %div_tensor), kwargs = {})
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=[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_2', '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_2(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 // 4)
x2 = xindex
tmp0 = tl.load(in_ptr0 + (4*x1), xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp18 = tl.load(in_ptr0 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp25 = tl.load(in_ptr1 + (x2), xmask)
tmp26 = tl.load(in_ptr1 + (4*x1), xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr1 + (1 + (4*x1)), xmask, eviction_policy='evict_last')
tmp29 = tl.load(in_ptr1 + (2 + (4*x1)), xmask, eviction_policy='evict_last')
tmp31 = tl.load(in_ptr1 + (3 + (4*x1)), xmask, eviction_policy='evict_last')
tmp1 = float("-inf")
tmp2 = tmp0 == tmp1
tmp3 = tmp2 == 0
tmp4 = tmp3.to(tl.int64)
tmp5 = (tmp4 != 0)
tmp7 = tmp6 == tmp1
tmp8 = tmp7 == 0
tmp9 = tmp8.to(tl.int64)
tmp10 = (tmp9 != 0)
tmp11 = tmp5 | tmp10
tmp13 = tmp12 == tmp1
tmp14 = tmp13 == 0
tmp15 = tmp14.to(tl.int64)
tmp16 = (tmp15 != 0)
tmp17 = tmp11 | tmp16
tmp19 = tmp18 == tmp1
tmp20 = tmp19 == 0
tmp21 = tmp20.to(tl.int64)
tmp22 = (tmp21 != 0)
tmp23 = tmp17 | tmp22
tmp24 = tmp23 == 0
tmp28 = tmp26 + tmp27
tmp30 = tmp28 + tmp29
tmp32 = tmp30 + tmp31
tmp33 = tmp25 / tmp32
tmp34 = 0.0
tmp35 = tl.where(tmp24, tmp34, tmp33)
tl.store(out_ptr0 + (x2), tmp35, xmask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/vv/cvvnhithjvmvhfjufxwwzclfobkrgbyyteg66hp24r675f7elw4c.py
# Topologically Sorted Source Nodes: [], Original ATen: []
# Source node to ATen node mapping:
# Graph fragment:
# %clone_default_2 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%expand_default_3,), kwargs = {memory_format: torch.contiguous_format})
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=[16, 4], 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_3', '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_3(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*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 + (4*y3)), tmp2, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/6t/c6t5a5ere3lqjiu7zh3uu4oxmpdoujdaqqmeunxqapgzo4m74uav.py
# Topologically Sorted Source Nodes: [context_layer_1], Original ATen: [aten.clone]
# Source node to ATen node mapping:
# context_layer_1 => clone_4
# Graph fragment:
# %clone_4 : [num_users=1] = call_function[target=torch.ops.aten.clone.default](args = (%permute_7,), kwargs = {memory_format: torch.contiguous_format})
triton_poi_fused_clone_4 = async_compile.triton('triton_poi_fused_clone_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=[16, 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_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_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK : tl.constexpr, XBLOCK : tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = (yindex // 4)
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + (4*x2) + (16*y1)), xmask & ymask, eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + (4*y3)), tmp0, xmask & ymask)
''', device_str='cuda')
# kernel path: runs/run_shard_4/inductor_cache/6m/c6mhj5zwirfhy5e4o45uaeov72uwfby4udubpm2fcz42iqvs2g57.py
# Topologically Sorted Source Nodes: [add, hidden_states_2], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# hidden_states_2 => var_mean
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %primals_3), 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_5 = async_compile.triton('triton_poi_fused_add_native_layer_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.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_5', '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_5(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_4/inductor_cache/l3/cl3bnd5pv2p4ydfmlj74bv4mbiwr2ntrdvbubnjubetyhosmxag6.py
# Topologically Sorted Source Nodes: [add, hidden_states_2], Original ATen: [aten.add, aten.native_layer_norm]
# Source node to ATen node mapping:
# add => add
# hidden_states_2 => add_1, add_2, mul, mul_1, rsqrt, sub_1
# Graph fragment:
# %add : [num_users=2] = call_function[target=torch.ops.aten.add.Tensor](args = (%view_17, %primals_3), kwargs = {})
# %add_1 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%getitem, 1.0), 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 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %rsqrt), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mul, %primals_10), kwargs = {})
# %add_2 : [num_users=1] = call_function[target=torch.ops.aten.add.Tensor](args = (%mul_1, %primals_11), kwargs = {})
triton_poi_fused_add_native_layer_norm_6 = async_compile.triton('triton_poi_fused_add_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=[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_6', '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_6(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 = 1.0
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, 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, 1))
assert_size_stride(primals_7, (4, ), (1, ))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4, ), (1, ))
assert_size_stride(primals_10, (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: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0), reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf2)
del primals_6
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
stream0 = get_raw_stream(0)
triton_poi_fused_0.run(buf0, primals_2, buf3, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_2
buf4 = reinterpret_tensor(buf0, (4, 4, 1, 4), (16, 4, 4, 1), 0); del buf0 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_0.run(buf1, primals_5, buf4, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_5
buf5 = empty_strided_cuda((16, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
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: [], Original ATen: []
triton_poi_fused_1.run(buf5, buf6, 256, grid=grid(256), stream=stream0)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_2.run(buf5, buf6, buf7, 256, grid=grid(256), stream=stream0)
del buf5
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0); del buf1 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
triton_poi_fused_3.run(buf2, primals_7, buf8, 16, 4, grid=grid(16, 4), stream=stream0)
del primals_7
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0); del buf2 # reuse
# Topologically Sorted Source Nodes: [], Original ATen: []
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 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
# Topologically Sorted Source Nodes: [context_layer_1], Original ATen: [aten.clone]
triton_poi_fused_clone_4.run(buf9, buf10, 16, 4, grid=grid(16, 4), stream=stream0)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0); del buf9 # reuse
# Topologically Sorted Source Nodes: [hidden_states], Original ATen: [aten.addmm]
extern_kernels.addmm(primals_9, reinterpret_tensor(buf10, (16, 4), (4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0), alpha=1, beta=1, out=buf11)
del primals_9
buf12 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
buf13 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
# Topologically Sorted Source Nodes: [add, hidden_states_2], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_5.run(buf11, primals_3, buf12, buf13, 16, grid=grid(16), stream=stream0)
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
# Topologically Sorted Source Nodes: [add, hidden_states_2], Original ATen: [aten.add, aten.native_layer_norm]
triton_poi_fused_add_native_layer_norm_6.run(buf11, primals_3, buf12, buf13, primals_10, primals_11, buf14, 64, grid=grid(64), stream=stream0)
del buf12
del buf13
del primals_11
return (buf14, primals_3, primals_10, buf7, reinterpret_tensor(buf8, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0), reinterpret_tensor(buf10, (16, 4), (4, 1), 0), buf11, primals_8, )
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, 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, 1), device='cuda:0', dtype=torch.float32)
primals_9 = rand_strided((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)
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)
| from _paritybench_helpers import _mock_config
import math
import torch
import torch.nn as nn
class BertSelfAttention(nn.Module):
def __init__(self, config):
super().__init__()
if (config.hidden_size % config.num_attention_heads != 0 and not
hasattr(config, 'embedding_size')):
raise ValueError(
'The hidden size (%d) is not a multiple of the number of attention heads (%d)'
% (config.hidden_size, config.num_attention_heads))
self.num_attention_heads = config.num_attention_heads
self.attention_head_size = int(config.hidden_size / config.
num_attention_heads)
self.all_head_size = (self.num_attention_heads * self.
attention_head_size)
self.query = nn.Linear(config.hidden_size, self.all_head_size)
self.key = nn.Linear(config.hidden_size, self.all_head_size)
self.value = nn.Linear(config.hidden_size, self.all_head_size)
self.dropout = nn.Dropout(config.attention_probs_dropout_prob)
def transpose_for_scores(self, x):
new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.
attention_head_size)
x = x.view(*new_x_shape)
return x.permute(0, 2, 1, 3)
def forward(self, hidden_states, attention_mask=None, head_mask=None,
encoder_hidden_states=None, encoder_attention_mask=None,
output_attentions=False):
mixed_query_layer = self.query(hidden_states)
if encoder_hidden_states is not None:
mixed_key_layer = self.key(encoder_hidden_states)
mixed_value_layer = self.value(encoder_hidden_states)
attention_mask = encoder_attention_mask
else:
mixed_key_layer = self.key(hidden_states)
mixed_value_layer = self.value(hidden_states)
query_layer = self.transpose_for_scores(mixed_query_layer)
key_layer = self.transpose_for_scores(mixed_key_layer)
value_layer = self.transpose_for_scores(mixed_value_layer)
attention_scores = torch.matmul(query_layer, key_layer.transpose(-1,
-2))
attention_scores = attention_scores / math.sqrt(self.
attention_head_size)
if attention_mask is not None:
attention_scores = attention_scores + attention_mask
attention_probs = nn.Softmax(dim=-1)(attention_scores)
attention_probs = self.dropout(attention_probs)
if head_mask is not None:
attention_probs = attention_probs * head_mask
context_layer = torch.matmul(attention_probs, value_layer)
context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
new_context_layer_shape = context_layer.size()[:-2] + (self.
all_head_size,)
context_layer = context_layer.view(*new_context_layer_shape)
outputs = (context_layer, attention_probs) if output_attentions else (
context_layer,)
return outputs
class BertSelfOutput(nn.Module):
def __init__(self, config):
super().__init__()
self.dense = nn.Linear(config.hidden_size, config.hidden_size)
self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.
layer_norm_eps)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
def forward(self, hidden_states, input_tensor):
hidden_states = self.dense(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.LayerNorm(hidden_states + input_tensor)
return hidden_states
class BertAttention(nn.Module):
def __init__(self, config):
super().__init__()
self.self = BertSelfAttention(config)
self.output = BertSelfOutput(config)
self.pruned_heads = set()
def prune_heads(self, heads):
if len(heads) == 0:
return
heads, index = find_pruneable_heads_and_indices(heads, self.self.
num_attention_heads, self.self.attention_head_size, self.
pruned_heads)
self.self.query = prune_linear_layer(self.self.query, index)
self.self.key = prune_linear_layer(self.self.key, index)
self.self.value = prune_linear_layer(self.self.value, index)
self.output.dense = prune_linear_layer(self.output.dense, index, dim=1)
self.self.num_attention_heads = self.self.num_attention_heads - len(
heads)
self.self.all_head_size = (self.self.attention_head_size * self.
self.num_attention_heads)
self.pruned_heads = self.pruned_heads.union(heads)
def forward(self, hidden_states, attention_mask=None, head_mask=None,
encoder_hidden_states=None, encoder_attention_mask=None,
output_attentions=False):
self_outputs = self.self(hidden_states, attention_mask, head_mask,
encoder_hidden_states, encoder_attention_mask, output_attentions)
attention_output = self.output(self_outputs[0], hidden_states)
outputs = (attention_output,) + self_outputs[1:]
return outputs
def get_inputs():
return [torch.rand([4, 4, 4])]
def get_init_inputs():
return [[], {'config': _mock_config(hidden_size=4, num_attention_heads=
4, attention_probs_dropout_prob=0.5, layer_norm_eps=1,
hidden_dropout_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
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, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK:
tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tmp1 = tl.load(in_ptr1 + y0, ymask, eviction_policy='evict_last')
tmp2 = tmp0 + tmp1
tmp3 = 1.0
tmp4 = tmp2 * tmp3
tl.store(out_ptr0 + (x2 + 4 * y3), tmp4, xmask & ymask)
@triton.jit
def triton_poi_fused_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_2(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 // 4
x2 = xindex
tmp0 = tl.load(in_ptr0 + 4 * x1, xmask, eviction_policy='evict_last')
tmp6 = tl.load(in_ptr0 + (1 + 4 * x1), xmask, eviction_policy='evict_last')
tmp12 = tl.load(in_ptr0 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp18 = tl.load(in_ptr0 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp25 = tl.load(in_ptr1 + x2, xmask)
tmp26 = tl.load(in_ptr1 + 4 * x1, xmask, eviction_policy='evict_last')
tmp27 = tl.load(in_ptr1 + (1 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp29 = tl.load(in_ptr1 + (2 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp31 = tl.load(in_ptr1 + (3 + 4 * x1), xmask, eviction_policy='evict_last'
)
tmp1 = float('-inf')
tmp2 = tmp0 == tmp1
tmp3 = tmp2 == 0
tmp4 = tmp3.to(tl.int64)
tmp5 = tmp4 != 0
tmp7 = tmp6 == tmp1
tmp8 = tmp7 == 0
tmp9 = tmp8.to(tl.int64)
tmp10 = tmp9 != 0
tmp11 = tmp5 | tmp10
tmp13 = tmp12 == tmp1
tmp14 = tmp13 == 0
tmp15 = tmp14.to(tl.int64)
tmp16 = tmp15 != 0
tmp17 = tmp11 | tmp16
tmp19 = tmp18 == tmp1
tmp20 = tmp19 == 0
tmp21 = tmp20.to(tl.int64)
tmp22 = tmp21 != 0
tmp23 = tmp17 | tmp22
tmp24 = tmp23 == 0
tmp28 = tmp26 + tmp27
tmp30 = tmp28 + tmp29
tmp32 = tmp30 + tmp31
tmp33 = tmp25 / tmp32
tmp34 = 0.0
tmp35 = tl.where(tmp24, tmp34, tmp33)
tl.store(out_ptr0 + x2, tmp35, xmask)
@triton.jit
def triton_poi_fused_3(in_ptr0, in_ptr1, out_ptr0, ynumel, xnumel, YBLOCK:
tl.constexpr, XBLOCK: tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * 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 + 4 * y3), tmp2, xmask & ymask)
@triton.jit
def triton_poi_fused_clone_4(in_ptr0, out_ptr0, ynumel, xnumel, YBLOCK: tl.
constexpr, XBLOCK: tl.constexpr):
ynumel = 16
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
x2 = xindex
y0 = yindex % 4
y1 = yindex // 4
y3 = yindex
tmp0 = tl.load(in_ptr0 + (y0 + 4 * x2 + 16 * y1), xmask & ymask,
eviction_policy='evict_last')
tl.store(out_ptr0 + (x2 + 4 * y3), tmp0, xmask & ymask)
@triton.jit
def triton_poi_fused_add_native_layer_norm_5(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_6(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 = 1.0
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, 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, 1))
assert_size_stride(primals_7, (4,), (1,))
assert_size_stride(primals_8, (4, 4), (4, 1))
assert_size_stride(primals_9, (4,), (1,))
assert_size_stride(primals_10, (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.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_1, (4, 4), (1, 4), 0), out=buf0)
del primals_1
buf1 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_4, (4, 4), (1, 4), 0), out=buf1)
del primals_4
buf2 = empty_strided_cuda((16, 4), (4, 1), torch.float32)
extern_kernels.mm(reinterpret_tensor(primals_3, (16, 4), (4, 1), 0),
reinterpret_tensor(primals_6, (4, 4), (1, 4), 0), out=buf2)
del primals_6
buf3 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
get_raw_stream(0)
triton_poi_fused_0[grid(16, 4)](buf0, primals_2, buf3, 16, 4,
XBLOCK=4, YBLOCK=8, num_warps=1, num_stages=1)
del primals_2
buf4 = reinterpret_tensor(buf0, (4, 4, 1, 4), (16, 4, 4, 1), 0)
del buf0
triton_poi_fused_0[grid(16, 4)](buf1, primals_5, buf4, 16, 4,
XBLOCK=4, YBLOCK=8, 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_1[grid(256)](buf5, buf6, 256, XBLOCK=128,
num_warps=4, num_stages=1)
buf7 = empty_strided_cuda((4, 4, 4, 4), (64, 16, 4, 1), torch.float32)
triton_poi_fused_2[grid(256)](buf5, buf6, buf7, 256, XBLOCK=128,
num_warps=4, num_stages=1)
del buf5
del buf6
buf8 = reinterpret_tensor(buf1, (4, 4, 4, 1), (16, 4, 1, 1), 0)
del buf1
triton_poi_fused_3[grid(16, 4)](buf2, primals_7, buf8, 16, 4,
XBLOCK=4, YBLOCK=16, num_warps=1, num_stages=1)
del primals_7
buf9 = reinterpret_tensor(buf2, (16, 4, 1), (4, 1, 1), 0)
del buf2
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 = empty_strided_cuda((4, 4, 4, 1), (16, 4, 1, 1), torch.float32)
triton_poi_fused_clone_4[grid(16, 4)](buf9, buf10, 16, 4, XBLOCK=4,
YBLOCK=16, num_warps=1, num_stages=1)
buf11 = reinterpret_tensor(buf9, (16, 4), (4, 1), 0)
del buf9
extern_kernels.addmm(primals_9, reinterpret_tensor(buf10, (16, 4),
(4, 1), 0), reinterpret_tensor(primals_8, (4, 4), (1, 4), 0),
alpha=1, beta=1, out=buf11)
del primals_9
buf12 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
buf13 = empty_strided_cuda((4, 4, 1), (4, 1, 16), torch.float32)
triton_poi_fused_add_native_layer_norm_5[grid(16)](buf11, primals_3,
buf12, buf13, 16, XBLOCK=16, num_warps=1, num_stages=1)
buf14 = empty_strided_cuda((4, 4, 4), (16, 4, 1), torch.float32)
triton_poi_fused_add_native_layer_norm_6[grid(64)](buf11, primals_3,
buf12, buf13, primals_10, primals_11, buf14, 64, XBLOCK=64,
num_warps=1, num_stages=1)
del buf12
del buf13
del primals_11
return buf14, primals_3, primals_10, buf7, reinterpret_tensor(buf8, (16,
1, 4), (4, 1, 1), 0), reinterpret_tensor(buf3, (16, 1, 4), (4, 1, 1), 0
), reinterpret_tensor(buf4, (16, 4, 1), (4, 1, 4), 0
), reinterpret_tensor(buf10, (16, 4), (4, 1), 0), buf11, primals_8
class BertSelfAttention(nn.Module):
def __init__(self, config):
super().__init__()
if (config.hidden_size % config.num_attention_heads != 0 and not
hasattr(config, 'embedding_size')):
raise ValueError(
'The hidden size (%d) is not a multiple of the number of attention heads (%d)'
% (config.hidden_size, config.num_attention_heads))
self.num_attention_heads = config.num_attention_heads
self.attention_head_size = int(config.hidden_size / config.
num_attention_heads)
self.all_head_size = (self.num_attention_heads * self.
attention_head_size)
self.query = nn.Linear(config.hidden_size, self.all_head_size)
self.key = nn.Linear(config.hidden_size, self.all_head_size)
self.value = nn.Linear(config.hidden_size, self.all_head_size)
self.dropout = nn.Dropout(config.attention_probs_dropout_prob)
def transpose_for_scores(self, x):
new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.
attention_head_size)
x = x.view(*new_x_shape)
return x.permute(0, 2, 1, 3)
def forward(self, hidden_states, attention_mask=None, head_mask=None,
encoder_hidden_states=None, encoder_attention_mask=None,
output_attentions=False):
mixed_query_layer = self.query(hidden_states)
if encoder_hidden_states is not None:
mixed_key_layer = self.key(encoder_hidden_states)
mixed_value_layer = self.value(encoder_hidden_states)
attention_mask = encoder_attention_mask
else:
mixed_key_layer = self.key(hidden_states)
mixed_value_layer = self.value(hidden_states)
query_layer = self.transpose_for_scores(mixed_query_layer)
key_layer = self.transpose_for_scores(mixed_key_layer)
value_layer = self.transpose_for_scores(mixed_value_layer)
attention_scores = torch.matmul(query_layer, key_layer.transpose(-1,
-2))
attention_scores = attention_scores / math.sqrt(self.
attention_head_size)
if attention_mask is not None:
attention_scores = attention_scores + attention_mask
attention_probs = nn.Softmax(dim=-1)(attention_scores)
attention_probs = self.dropout(attention_probs)
if head_mask is not None:
attention_probs = attention_probs * head_mask
context_layer = torch.matmul(attention_probs, value_layer)
context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
new_context_layer_shape = context_layer.size()[:-2] + (self.
all_head_size,)
context_layer = context_layer.view(*new_context_layer_shape)
outputs = (context_layer, attention_probs) if output_attentions else (
context_layer,)
return outputs
class BertSelfOutput(nn.Module):
def __init__(self, config):
super().__init__()
self.dense = nn.Linear(config.hidden_size, config.hidden_size)
self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.
layer_norm_eps)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
def forward(self, hidden_states, input_tensor):
hidden_states = self.dense(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.LayerNorm(hidden_states + input_tensor)
return hidden_states
class BertAttentionNew(nn.Module):
def __init__(self, config):
super().__init__()
self.self = BertSelfAttention(config)
self.output = BertSelfOutput(config)
self.pruned_heads = set()
def prune_heads(self, heads):
if len(heads) == 0:
return
heads, index = find_pruneable_heads_and_indices(heads, self.self.
num_attention_heads, self.self.attention_head_size, self.
pruned_heads)
self.self.query = prune_linear_layer(self.self.query, index)
self.self.key = prune_linear_layer(self.self.key, index)
self.self.value = prune_linear_layer(self.self.value, index)
self.output.dense = prune_linear_layer(self.output.dense, index, dim=1)
self.self.num_attention_heads = self.self.num_attention_heads - len(
heads)
self.self.all_head_size = (self.self.attention_head_size * self.
self.num_attention_heads)
self.pruned_heads = self.pruned_heads.union(heads)
def forward(self, input_0):
primals_1 = self.self.query.weight
primals_2 = self.self.query.bias
primals_4 = self.self.key.weight
primals_5 = self.self.key.bias
primals_6 = self.self.value.weight
primals_7 = self.self.value.bias
primals_8 = self.output.dense.weight
primals_9 = self.output.dense.bias
primals_10 = self.output.LayerNorm.weight
primals_11 = self.output.LayerNorm.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])
return output[0]
| hongyuntw/Col-KBERT | BertAttention | false | 6,828 | [
"MIT"
] | 1 | e77ce2585d228a783bf83cc1de53583aff70f7b4 | https://github.com/hongyuntw/Col-KBERT/tree/e77ce2585d228a783bf83cc1de53583aff70f7b4 | from _paritybench_helpers import _mock_config
import math
import torch
import torch.nn as nn
class BertSelfAttention(nn.Module):
def __init__(self, config):
super().__init__()
if (config.hidden_size % config.num_attention_heads != 0 and not
hasattr(config, 'embedding_size')):
raise ValueError(
'The hidden size (%d) is not a multiple of the number of attention heads (%d)'
% (config.hidden_size, config.num_attention_heads))
self.num_attention_heads = config.num_attention_heads
self.attention_head_size = int(config.hidden_size / config.
num_attention_heads)
self.all_head_size = (self.num_attention_heads * self.
attention_head_size)
self.query = nn.Linear(config.hidden_size, self.all_head_size)
self.key = nn.Linear(config.hidden_size, self.all_head_size)
self.value = nn.Linear(config.hidden_size, self.all_head_size)
self.dropout = nn.Dropout(config.attention_probs_dropout_prob)
def transpose_for_scores(self, x):
new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.
attention_head_size)
x = x.view(*new_x_shape)
return x.permute(0, 2, 1, 3)
def forward(self, hidden_states, attention_mask=None, head_mask=None,
encoder_hidden_states=None, encoder_attention_mask=None,
output_attentions=False):
mixed_query_layer = self.query(hidden_states)
if encoder_hidden_states is not None:
mixed_key_layer = self.key(encoder_hidden_states)
mixed_value_layer = self.value(encoder_hidden_states)
attention_mask = encoder_attention_mask
else:
mixed_key_layer = self.key(hidden_states)
mixed_value_layer = self.value(hidden_states)
query_layer = self.transpose_for_scores(mixed_query_layer)
key_layer = self.transpose_for_scores(mixed_key_layer)
value_layer = self.transpose_for_scores(mixed_value_layer)
attention_scores = torch.matmul(query_layer, key_layer.transpose(-1,
-2))
attention_scores = attention_scores / math.sqrt(self.
attention_head_size)
if attention_mask is not None:
attention_scores = attention_scores + attention_mask
attention_probs = nn.Softmax(dim=-1)(attention_scores)
attention_probs = self.dropout(attention_probs)
if head_mask is not None:
attention_probs = attention_probs * head_mask
context_layer = torch.matmul(attention_probs, value_layer)
context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
new_context_layer_shape = context_layer.size()[:-2] + (self.
all_head_size,)
context_layer = context_layer.view(*new_context_layer_shape)
outputs = (context_layer, attention_probs) if output_attentions else (
context_layer,)
return outputs
class BertSelfOutput(nn.Module):
def __init__(self, config):
super().__init__()
self.dense = nn.Linear(config.hidden_size, config.hidden_size)
self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.
layer_norm_eps)
self.dropout = nn.Dropout(config.hidden_dropout_prob)
def forward(self, hidden_states, input_tensor):
hidden_states = self.dense(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.LayerNorm(hidden_states + input_tensor)
return hidden_states
class Model(nn.Module):
def __init__(self, config):
super().__init__()
self.self = BertSelfAttention(config)
self.output = BertSelfOutput(config)
self.pruned_heads = set()
def prune_heads(self, heads):
if len(heads) == 0:
return
heads, index = find_pruneable_heads_and_indices(heads, self.self.
num_attention_heads, self.self.attention_head_size, self.
# ... truncated (>4000 chars) for memory efficiency |
SimpleGFLLoss | # 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_4/inductor_cache/c5/cc5ww3st2dz7msmmetezkkm7td77m3jj4cdvku2tafehlilcy3dg.py
# Topologically Sorted Source Nodes: [binary_cross_entropy_with_logits, sigmoid, sub, abs_1, loss, loss_1, loss_2, mul_1], Original ATen: [aten.binary_cross_entropy_with_logits, aten.sigmoid, aten.sub, aten.abs, aten.pow, aten.mul, aten.mean]
# Source node to ATen node mapping:
# abs_1 => abs_1
# binary_cross_entropy_with_logits => abs_2, exp, full_default, log1p, mean, minimum, mul, neg, sub_1, sub_2, sub_3
# loss => pow_1
# loss_1 => mul_1
# loss_2 => mean_1
# mul_1 => mul_2
# sigmoid => sigmoid
# sub => sub
# Graph fragment:
# %sub_1 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (1, %arg1_1), kwargs = {})
# %mul : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%sub_1, %arg0_1), kwargs = {})
# %full_default : [num_users=1] = call_function[target=torch.ops.aten.full.default](args = ([], 0), kwargs = {dtype: torch.float32, layout: torch.strided, device: cuda:0, pin_memory: False})
# %minimum : [num_users=1] = call_function[target=torch.ops.aten.minimum.default](args = (%full_default, %arg0_1), kwargs = {})
# %abs_2 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%arg0_1,), kwargs = {})
# %neg : [num_users=1] = call_function[target=torch.ops.aten.neg.default](args = (%abs_2,), kwargs = {})
# %exp : [num_users=1] = call_function[target=torch.ops.aten.exp.default](args = (%neg,), kwargs = {})
# %log1p : [num_users=1] = call_function[target=torch.ops.aten.log1p.default](args = (%exp,), kwargs = {})
# %sub_2 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%minimum, %log1p), kwargs = {})
# %sub_3 : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%mul, %sub_2), kwargs = {})
# %mean : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%sub_3,), kwargs = {})
# %sigmoid : [num_users=1] = call_function[target=torch.ops.aten.sigmoid.default](args = (%arg0_1,), kwargs = {})
# %sub : [num_users=1] = call_function[target=torch.ops.aten.sub.Tensor](args = (%sigmoid, %arg1_1), kwargs = {})
# %abs_1 : [num_users=1] = call_function[target=torch.ops.aten.abs.default](args = (%sub,), kwargs = {})
# %pow_1 : [num_users=1] = call_function[target=torch.ops.aten.pow.Tensor_Scalar](args = (%abs_1, 2.0), kwargs = {})
# %mul_1 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean, %pow_1), kwargs = {})
# %mean_1 : [num_users=1] = call_function[target=torch.ops.aten.mean.default](args = (%mul_1,), kwargs = {})
# %mul_2 : [num_users=1] = call_function[target=torch.ops.aten.mul.Tensor](args = (%mean_1, 1.0), kwargs = {})
triton_per_fused_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_sub_0 = async_compile.triton('triton_per_fused_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_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: '*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_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_sub_0', 'mutated_arg_names': ['in_out_ptr0'], 'no_x_dim': True, '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_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_sub_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 = tmp1 - tmp0
tmp4 = tmp2 * tmp3
tmp5 = 0.0
tmp6 = triton_helpers.minimum(tmp5, tmp3)
tmp7 = tl_math.abs(tmp3)
tmp8 = -tmp7
tmp9 = tl_math.exp(tmp8)
tmp10 = libdevice.log1p(tmp9)
tmp11 = tmp6 - tmp10
tmp12 = tmp4 - tmp11
tmp13 = tl.broadcast_to(tmp12, [RBLOCK])
tmp15 = triton_helpers.promote_to_tensor(tl.sum(tmp13, 0))
tmp16 = 256.0
tmp17 = tmp15 / tmp16
tmp18 = tl.sigmoid(tmp3)
tmp19 = tmp18 - tmp0
tmp20 = tl_math.abs(tmp19)
tmp21 = tmp20 * tmp20
tmp22 = tmp17 * tmp21
tmp23 = tl.broadcast_to(tmp22, [RBLOCK])
tmp25 = triton_helpers.promote_to_tensor(tl.sum(tmp23, 0))
tmp26 = tmp25 / tmp16
tmp27 = tmp26 * tmp1
tl.debug_barrier()
tl.store(in_out_ptr0 + (tl.full([1], 0, tl.int32)), tmp27, 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
buf2 = buf1; del buf1 # reuse
# Topologically Sorted Source Nodes: [binary_cross_entropy_with_logits, sigmoid, sub, abs_1, loss, loss_1, loss_2, mul_1], Original ATen: [aten.binary_cross_entropy_with_logits, aten.sigmoid, aten.sub, aten.abs, aten.pow, aten.mul, aten.mean]
stream0 = get_raw_stream(0)
triton_per_fused_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_sub_0.run(buf2, arg1_1, arg0_1, 1, 256, grid=grid(1), stream=stream0)
del arg0_1
del arg1_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)
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
def simple_gfl(pred, target, beta):
"""Simply add a pow of abs difference in front of BCE"""
assert pred.size() == target.size(
), 'simple GFL assume pred and target to have the same shape'
loss = (pred.sigmoid() - target).abs().pow(beta)
loss = F.binary_cross_entropy_with_logits(pred, target) * loss
return loss
class SimpleGFLLoss(torch.nn.Module):
"""Simple version of Quality Focal Loss from paper `Generalized Focal Loss: Learning
Qualified and Distributed Bounding Boxes for Dense Object Detection
<https://arxiv.org/abs/2006.04388>`_. """
def __init__(self, beta=2.0, loss_weight=1.0):
super(SimpleGFLLoss, self).__init__()
self.beta = beta
self.loss_weight = loss_weight
def forward(self, pred, target, weight=None, avg_factor=None):
loss = simple_gfl(pred, target, self.beta)
if weight is not None:
loss *= weight
if avg_factor is None:
loss = loss.mean()
else:
loss = loss.sum() / avg_factor
return loss * self.loss_weight
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 libdevice, math as tl_math
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_per_fused_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_sub_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 = tmp1 - tmp0
tmp4 = tmp2 * tmp3
tmp5 = 0.0
tmp6 = triton_helpers.minimum(tmp5, tmp3)
tmp7 = tl_math.abs(tmp3)
tmp8 = -tmp7
tmp9 = tl_math.exp(tmp8)
tmp10 = libdevice.log1p(tmp9)
tmp11 = tmp6 - tmp10
tmp12 = tmp4 - tmp11
tmp13 = tl.broadcast_to(tmp12, [RBLOCK])
tmp15 = triton_helpers.promote_to_tensor(tl.sum(tmp13, 0))
tmp16 = 256.0
tmp17 = tmp15 / tmp16
tmp18 = tl.sigmoid(tmp3)
tmp19 = tmp18 - tmp0
tmp20 = tl_math.abs(tmp19)
tmp21 = tmp20 * tmp20
tmp22 = tmp17 * tmp21
tmp23 = tl.broadcast_to(tmp22, [RBLOCK])
tmp25 = triton_helpers.promote_to_tensor(tl.sum(tmp23, 0))
tmp26 = tmp25 / tmp16
tmp27 = tmp26 * tmp1
tl.debug_barrier()
tl.store(in_out_ptr0 + tl.full([1], 0, tl.int32), tmp27, 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
buf2 = buf1
del buf1
get_raw_stream(0)
triton_per_fused_abs_binary_cross_entropy_with_logits_mean_mul_pow_sigmoid_sub_0[
grid(1)](buf2, arg1_1, arg0_1, 1, 256, num_warps=2, num_stages=1)
del arg0_1
del arg1_1
return buf2,
def simple_gfl(pred, target, beta):
"""Simply add a pow of abs difference in front of BCE"""
assert pred.size() == target.size(
), 'simple GFL assume pred and target to have the same shape'
loss = (pred.sigmoid() - target).abs().pow(beta)
loss = F.binary_cross_entropy_with_logits(pred, target) * loss
return loss
class SimpleGFLLossNew(torch.nn.Module):
"""Simple version of Quality Focal Loss from paper `Generalized Focal Loss: Learning
Qualified and Distributed Bounding Boxes for Dense Object Detection
<https://arxiv.org/abs/2006.04388>`_. """
def __init__(self, beta=2.0, loss_weight=1.0):
super(SimpleGFLLossNew, self).__init__()
self.beta = beta
self.loss_weight = loss_weight
def forward(self, input_0, input_1):
arg0_1 = input_0
arg1_1 = input_1
output = call([arg0_1, arg1_1])
return output[0]
| huimlight/OpenMMLab-IoUNet | SimpleGFLLoss | false | 6,829 | [
"Apache-2.0"
] | 1 | 00536bac99f4d3d7ad2682dad44f299f714565b6 | https://github.com/huimlight/OpenMMLab-IoUNet/tree/00536bac99f4d3d7ad2682dad44f299f714565b6 | import torch
import torch.nn.functional as F
def simple_gfl(pred, target, beta):
"""Simply add a pow of abs difference in front of BCE"""
assert pred.size() == target.size(
), 'simple GFL assume pred and target to have the same shape'
loss = (pred.sigmoid() - target).abs().pow(beta)
loss = F.binary_cross_entropy_with_logits(pred, target) * loss
return loss
class Model(torch.nn.Module):
"""Simple version of Quality Focal Loss from paper `Generalized Focal Loss: Learning
Qualified and Distributed Bounding Boxes for Dense Object Detection
<https://arxiv.org/abs/2006.04388>`_. """
def __init__(self, beta=2.0, loss_weight=1.0):
super().__init__()
self.beta = beta
self.loss_weight = loss_weight
def forward(self, pred, target, weight=None, avg_factor=None):
loss = simple_gfl(pred, target, self.beta)
if weight is not None:
loss *= weight
if avg_factor is None:
loss = loss.mean()
else:
loss = loss.sum() / avg_factor
return loss * self.loss_weight
def get_inputs():
return [torch.rand([4, 4, 4, 4]), torch.rand([4, 4, 4, 4])]
def get_init_inputs():
return []
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.