code
stringlengths 86
54.5k
| code_codestyle
int64 0
371
| style_context
stringlengths 87
49.2k
| style_context_codestyle
int64 0
349
| label
int64 0
1
|
|---|---|---|---|---|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tf_available, is_torch_available
UpperCamelCase_ = {
"configuration_groupvit": [
"GROUPVIT_PRETRAINED_CONFIG_ARCHIVE_MAP",
"GroupViTConfig",
"GroupViTOnnxConfig",
"GroupViTTextConfig",
"GroupViTVisionConfig",
],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"GROUPVIT_PRETRAINED_MODEL_ARCHIVE_LIST",
"GroupViTModel",
"GroupViTPreTrainedModel",
"GroupViTTextModel",
"GroupViTVisionModel",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"TF_GROUPVIT_PRETRAINED_MODEL_ARCHIVE_LIST",
"TFGroupViTModel",
"TFGroupViTPreTrainedModel",
"TFGroupViTTextModel",
"TFGroupViTVisionModel",
]
if TYPE_CHECKING:
from .configuration_groupvit import (
GROUPVIT_PRETRAINED_CONFIG_ARCHIVE_MAP,
GroupViTConfig,
GroupViTOnnxConfig,
GroupViTTextConfig,
GroupViTVisionConfig,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_groupvit import (
GROUPVIT_PRETRAINED_MODEL_ARCHIVE_LIST,
GroupViTModel,
GroupViTPreTrainedModel,
GroupViTTextModel,
GroupViTVisionModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_groupvit import (
TF_GROUPVIT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFGroupViTModel,
TFGroupViTPreTrainedModel,
TFGroupViTTextModel,
TFGroupViTVisionModel,
)
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from ..models.whisper import WhisperForConditionalGeneration, WhisperProcessor
from .base import PipelineTool
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="openai/whisper-base"
UpperCamelCase__ : int =(
"This is a tool that transcribes an audio into text. It takes an input named `audio` and returns the "
"transcribed text."
)
UpperCamelCase__ : Any ="transcriber"
UpperCamelCase__ : Optional[int] =WhisperProcessor
UpperCamelCase__ : List[str] =WhisperForConditionalGeneration
UpperCamelCase__ : List[Any] =["audio"]
UpperCamelCase__ : Union[str, Any] =["text"]
def __a ( self :int , _lowercase :Any) -> Tuple:
return self.pre_processor(_lowercase , return_tensors='''pt''').input_features
def __a ( self :Dict , _lowercase :Tuple) -> Any:
return self.model.generate(inputs=_lowercase)
def __a ( self :int , _lowercase :Union[str, Any]) -> Optional[Any]:
return self.pre_processor.batch_decode(_lowercase , skip_special_tokens=_lowercase)[0]
| 344 | 1 |
class a_ :
def __init__( self :str) -> List[Any]:
UpperCAmelCase_ = 0
UpperCAmelCase_ = 0
UpperCAmelCase_ = {}
def __a ( self :Tuple , _lowercase :Tuple) -> Optional[int]:
if vertex not in self.adjacency:
UpperCAmelCase_ = {}
self.num_vertices += 1
def __a ( self :List[Any] , _lowercase :Optional[int] , _lowercase :List[str] , _lowercase :Union[str, Any]) -> Any:
self.add_vertex(_lowercase)
self.add_vertex(_lowercase)
if head == tail:
return
UpperCAmelCase_ = weight
UpperCAmelCase_ = weight
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_edges()
for edge in edges:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = edge
edges.remove((tail, head, weight))
for i in range(len(_lowercase)):
UpperCAmelCase_ = list(edges[i])
edges.sort(key=lambda _lowercase: e[2])
for i in range(len(_lowercase) - 1):
if edges[i][2] >= edges[i + 1][2]:
UpperCAmelCase_ = edges[i][2] + 1
for edge in edges:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = edge
UpperCAmelCase_ = weight
UpperCAmelCase_ = weight
def __str__( self :int) -> List[Any]:
UpperCAmelCase_ = ''''''
for tail in self.adjacency:
for head in self.adjacency[tail]:
UpperCAmelCase_ = self.adjacency[head][tail]
string += f"{head} -> {tail} == {weight}\n"
return string.rstrip('''\n''')
def __a ( self :int) -> str:
UpperCAmelCase_ = []
for tail in self.adjacency:
for head in self.adjacency[tail]:
output.append((tail, head, self.adjacency[head][tail]))
return output
def __a ( self :Union[str, Any]) -> List[str]:
return self.adjacency.keys()
@staticmethod
def __a ( _lowercase :Union[str, Any]=None , _lowercase :int=None) -> str:
UpperCAmelCase_ = Graph()
if vertices is None:
UpperCAmelCase_ = []
if edges is None:
UpperCAmelCase_ = []
for vertex in vertices:
g.add_vertex(_lowercase)
for edge in edges:
g.add_edge(*_lowercase)
return g
class a_ :
def __init__( self :List[str]) -> Union[str, Any]:
UpperCAmelCase_ = {}
UpperCAmelCase_ = {}
def __len__( self :str) -> Union[str, Any]:
return len(self.parent)
def __a ( self :List[Any] , _lowercase :Dict) -> List[str]:
if item in self.parent:
return self.find(_lowercase)
UpperCAmelCase_ = item
UpperCAmelCase_ = 0
return item
def __a ( self :Optional[int] , _lowercase :Optional[int]) -> Tuple:
if item not in self.parent:
return self.make_set(_lowercase)
if item != self.parent[item]:
UpperCAmelCase_ = self.find(self.parent[item])
return self.parent[item]
def __a ( self :List[str] , _lowercase :Optional[int] , _lowercase :str) -> Dict:
UpperCAmelCase_ = self.find(_lowercase)
UpperCAmelCase_ = self.find(_lowercase)
if roota == roota:
return roota
if self.rank[roota] > self.rank[roota]:
UpperCAmelCase_ = roota
return roota
if self.rank[roota] < self.rank[roota]:
UpperCAmelCase_ = roota
return roota
if self.rank[roota] == self.rank[roota]:
self.rank[roota] += 1
UpperCAmelCase_ = roota
return roota
return None
@staticmethod
def __a ( _lowercase :Union[str, Any]) -> int:
UpperCAmelCase_ = graph.num_vertices
UpperCAmelCase_ = Graph.UnionFind()
UpperCAmelCase_ = []
while num_components > 1:
UpperCAmelCase_ = {}
for vertex in graph.get_vertices():
UpperCAmelCase_ = -1
UpperCAmelCase_ = graph.get_edges()
for edge in edges:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = edge
edges.remove((tail, head, weight))
for edge in edges:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = edge
UpperCAmelCase_ = union_find.find(_lowercase)
UpperCAmelCase_ = union_find.find(_lowercase)
if seta != seta:
if cheap_edge[seta] == -1 or cheap_edge[seta][2] > weight:
UpperCAmelCase_ = [head, tail, weight]
if cheap_edge[seta] == -1 or cheap_edge[seta][2] > weight:
UpperCAmelCase_ = [head, tail, weight]
for vertex in cheap_edge:
if cheap_edge[vertex] != -1:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = cheap_edge[vertex]
if union_find.find(_lowercase) != union_find.find(_lowercase):
union_find.union(_lowercase , _lowercase)
mst_edges.append(cheap_edge[vertex])
UpperCAmelCase_ = num_components - 1
UpperCAmelCase_ = Graph.build(edges=_lowercase)
return mst
| 344 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
UpperCamelCase_ = {"configuration_opt": ["OPT_PRETRAINED_CONFIG_ARCHIVE_MAP", "OPTConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"OPT_PRETRAINED_MODEL_ARCHIVE_LIST",
"OPTForCausalLM",
"OPTModel",
"OPTPreTrainedModel",
"OPTForSequenceClassification",
"OPTForQuestionAnswering",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["TFOPTForCausalLM", "TFOPTModel", "TFOPTPreTrainedModel"]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxOPTForCausalLM",
"FlaxOPTModel",
"FlaxOPTPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_opt import OPT_PRETRAINED_CONFIG_ARCHIVE_MAP, OPTConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_opt import (
OPT_PRETRAINED_MODEL_ARCHIVE_LIST,
OPTForCausalLM,
OPTForQuestionAnswering,
OPTForSequenceClassification,
OPTModel,
OPTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_opt import TFOPTForCausalLM, TFOPTModel, TFOPTPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_opt import FlaxOPTForCausalLM, FlaxOPTModel, FlaxOPTPreTrainedModel
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 | 1 |
UpperCamelCase_ = {}
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
if late == 3 or absent == 2:
return 0
# if we have no days left, and have not failed any other rules,
# we have a prize string
if days == 0:
return 1
# No easy solution, so now we need to do the recursive calculation
# First, check if the combination is already in the cache, and
# if yes, return the stored value from there since we already
# know the number of possible prize strings from this point on
UpperCAmelCase_ = (days, absent, late)
if key in cache:
return cache[key]
# now we calculate the three possible ways that can unfold from
# this point on, depending on our attendance today
# 1) if we are late (but not absent), the "absent" counter stays as
# it is, but the "late" counter increases by one
UpperCAmelCase_ = _calculate(days - 1 , __UpperCAmelCase , late + 1 )
# 2) if we are absent, the "absent" counter increases by 1, and the
# "late" counter resets to 0
UpperCAmelCase_ = _calculate(days - 1 , absent + 1 , 0 )
# 3) if we are on time, this resets the "late" counter and keeps the
# absent counter
UpperCAmelCase_ = _calculate(days - 1 , __UpperCAmelCase , 0 )
UpperCAmelCase_ = state_late + state_absent + state_ontime
UpperCAmelCase_ = prizestrings
return prizestrings
def A ( __UpperCAmelCase = 30 ) -> int:
'''simple docstring'''
return _calculate(__UpperCAmelCase , absent=0 , late=0 )
if __name__ == "__main__":
print(solution())
| 344 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
UpperCamelCase_ = "▁"
UpperCamelCase_ = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : str =BigBirdTokenizer
UpperCamelCase__ : Tuple =BigBirdTokenizerFast
UpperCamelCase__ : Union[str, Any] =True
UpperCamelCase__ : List[str] =True
def __a ( self :Any) -> List[str]:
super().setUp()
UpperCAmelCase_ = self.tokenizer_class(_lowercase , keep_accents=_lowercase)
tokenizer.save_pretrained(self.tmpdirname)
def __a ( self :Optional[int]) -> str:
UpperCAmelCase_ = '''<s>'''
UpperCAmelCase_ = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_lowercase) , _lowercase)
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_lowercase) , _lowercase)
def __a ( self :str) -> str:
UpperCAmelCase_ = list(self.get_tokenizer().get_vocab().keys())
self.assertEqual(vocab_keys[0] , '''<unk>''')
self.assertEqual(vocab_keys[1] , '''<s>''')
self.assertEqual(vocab_keys[-1] , '''[MASK]''')
self.assertEqual(len(_lowercase) , 1004)
def __a ( self :List[str]) -> int:
self.assertEqual(self.get_tokenizer().vocab_size , 1000)
def __a ( self :Tuple) -> int:
if not self.test_rust_tokenizer:
return
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = '''I was born in 92000, and this is falsé.'''
UpperCAmelCase_ = tokenizer.tokenize(_lowercase)
UpperCAmelCase_ = rust_tokenizer.tokenize(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = tokenizer.encode(_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :Optional[Any]) -> List[str]:
UpperCAmelCase_ = BigBirdTokenizer(_lowercase , keep_accents=_lowercase)
UpperCAmelCase_ = tokenizer.tokenize('''This is a test''')
self.assertListEqual(_lowercase , ['''▁This''', '''▁is''', '''▁a''', '''▁t''', '''est'''])
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_lowercase) , [285, 46, 10, 170, 382] , )
UpperCAmelCase_ = tokenizer.tokenize('''I was born in 92000, and this is falsé.''')
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''9''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''é''',
'''.''',
] , )
UpperCAmelCase_ = tokenizer.convert_tokens_to_ids(_lowercase)
self.assertListEqual(
_lowercase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
UpperCAmelCase_ = tokenizer.convert_ids_to_tokens(_lowercase)
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''<unk>''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''<unk>''',
'''.''',
] , )
@cached_property
def __a ( self :Any) -> List[Any]:
return BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
@slow
def __a ( self :int) -> List[Any]:
UpperCAmelCase_ = '''Hello World!'''
UpperCAmelCase_ = [65, 18536, 2260, 101, 66]
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@slow
def __a ( self :int) -> Any:
UpperCAmelCase_ = (
'''This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) " [ ] ! : - . Also we will'''
''' add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth'''
)
# fmt: off
UpperCAmelCase_ = [65, 871, 419, 358, 946, 991, 2521, 452, 358, 1357, 387, 7751, 3536, 112, 985, 456, 126, 865, 938, 5400, 5734, 458, 1368, 467, 786, 2462, 5246, 1159, 633, 865, 4519, 457, 582, 852, 2557, 427, 916, 508, 405, 34324, 497, 391, 408, 11342, 1244, 385, 100, 938, 985, 456, 574, 362, 12597, 3200, 3129, 1172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@require_torch
@slow
def __a ( self :Dict) -> Union[str, Any]:
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
UpperCAmelCase_ = list(self.big_tokenizer.get_vocab().keys())[:10]
UpperCAmelCase_ = ''' '''.join(_lowercase)
UpperCAmelCase_ = self.big_tokenizer.encode_plus(_lowercase , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = self.big_tokenizer.batch_encode_plus(
[sequence + ''' ''' + sequence] , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = BigBirdConfig(attention_type='''original_full''')
UpperCAmelCase_ = BigBirdModel(_lowercase)
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_lowercase)
model(**_lowercase)
@slow
def __a ( self :Optional[int]) -> Any:
UpperCAmelCase_ = BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
UpperCAmelCase_ = tokenizer.decode(tokenizer('''Paris is the [MASK].''').input_ids)
self.assertTrue(decoded_text == '''[CLS] Paris is the[MASK].[SEP]''')
@slow
def __a ( self :Dict) -> List[str]:
# fmt: off
UpperCAmelCase_ = {'''input_ids''': [[65, 39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114, 66], [65, 448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], '''attention_mask''': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_lowercase , model_name='''google/bigbird-roberta-base''' , revision='''215c99f1600e06f83acce68422f2035b2b5c3510''' , )
| 344 | 1 |
from typing import Any, Dict, Optional
import torch
import torch.nn.functional as F
from torch import nn
from ..utils import maybe_allow_in_graph
from .activations import get_activation
from .attention_processor import Attention
from .embeddings import CombinedTimestepLabelEmbeddings
@maybe_allow_in_graph
class a_ ( nn.Module ):
def __init__( self :str , _lowercase :int , _lowercase :int , _lowercase :int , _lowercase :Optional[int]=0.0 , _lowercase :Optional[int] = None , _lowercase :str = "geglu" , _lowercase :Optional[int] = None , _lowercase :bool = False , _lowercase :bool = False , _lowercase :bool = False , _lowercase :bool = False , _lowercase :bool = True , _lowercase :str = "layer_norm" , _lowercase :bool = False , ) -> str:
super().__init__()
UpperCAmelCase_ = only_cross_attention
UpperCAmelCase_ = (num_embeds_ada_norm is not None) and norm_type == '''ada_norm_zero'''
UpperCAmelCase_ = (num_embeds_ada_norm is not None) and norm_type == '''ada_norm'''
if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None:
raise ValueError(
f"`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to"
f" define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}.")
# Define 3 blocks. Each block has its own normalization layer.
# 1. Self-Attn
if self.use_ada_layer_norm:
UpperCAmelCase_ = AdaLayerNorm(_lowercase , _lowercase)
elif self.use_ada_layer_norm_zero:
UpperCAmelCase_ = AdaLayerNormZero(_lowercase , _lowercase)
else:
UpperCAmelCase_ = nn.LayerNorm(_lowercase , elementwise_affine=_lowercase)
UpperCAmelCase_ = Attention(
query_dim=_lowercase , heads=_lowercase , dim_head=_lowercase , dropout=_lowercase , bias=_lowercase , cross_attention_dim=cross_attention_dim if only_cross_attention else None , upcast_attention=_lowercase , )
# 2. Cross-Attn
if cross_attention_dim is not None or double_self_attention:
# We currently only use AdaLayerNormZero for self attention where there will only be one attention block.
# I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during
# the second cross attention block.
UpperCAmelCase_ = (
AdaLayerNorm(_lowercase , _lowercase)
if self.use_ada_layer_norm
else nn.LayerNorm(_lowercase , elementwise_affine=_lowercase)
)
UpperCAmelCase_ = Attention(
query_dim=_lowercase , cross_attention_dim=cross_attention_dim if not double_self_attention else None , heads=_lowercase , dim_head=_lowercase , dropout=_lowercase , bias=_lowercase , upcast_attention=_lowercase , ) # is self-attn if encoder_hidden_states is none
else:
UpperCAmelCase_ = None
UpperCAmelCase_ = None
# 3. Feed-forward
UpperCAmelCase_ = nn.LayerNorm(_lowercase , elementwise_affine=_lowercase)
UpperCAmelCase_ = FeedForward(_lowercase , dropout=_lowercase , activation_fn=_lowercase , final_dropout=_lowercase)
# let chunk size default to None
UpperCAmelCase_ = None
UpperCAmelCase_ = 0
def __a ( self :Tuple , _lowercase :Optional[int] , _lowercase :int) -> Dict:
# Sets chunk feed-forward
UpperCAmelCase_ = chunk_size
UpperCAmelCase_ = dim
def __a ( self :Dict , _lowercase :torch.FloatTensor , _lowercase :Optional[torch.FloatTensor] = None , _lowercase :Optional[torch.FloatTensor] = None , _lowercase :Optional[torch.FloatTensor] = None , _lowercase :Optional[torch.LongTensor] = None , _lowercase :Dict[str, Any] = None , _lowercase :Optional[torch.LongTensor] = None , ) -> List[Any]:
# Notice that normalization is always applied before the real computation in the following blocks.
# 1. Self-Attention
if self.use_ada_layer_norm:
UpperCAmelCase_ = self.norma(_lowercase , _lowercase)
elif self.use_ada_layer_norm_zero:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = self.norma(
_lowercase , _lowercase , _lowercase , hidden_dtype=hidden_states.dtype)
else:
UpperCAmelCase_ = self.norma(_lowercase)
UpperCAmelCase_ = cross_attention_kwargs if cross_attention_kwargs is not None else {}
UpperCAmelCase_ = self.attna(
_lowercase , encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None , attention_mask=_lowercase , **_lowercase , )
if self.use_ada_layer_norm_zero:
UpperCAmelCase_ = gate_msa.unsqueeze(1) * attn_output
UpperCAmelCase_ = attn_output + hidden_states
# 2. Cross-Attention
if self.attna is not None:
UpperCAmelCase_ = (
self.norma(_lowercase , _lowercase) if self.use_ada_layer_norm else self.norma(_lowercase)
)
UpperCAmelCase_ = self.attna(
_lowercase , encoder_hidden_states=_lowercase , attention_mask=_lowercase , **_lowercase , )
UpperCAmelCase_ = attn_output + hidden_states
# 3. Feed-forward
UpperCAmelCase_ = self.norma(_lowercase)
if self.use_ada_layer_norm_zero:
UpperCAmelCase_ = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None]
if self._chunk_size is not None:
# "feed_forward_chunk_size" can be used to save memory
if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0:
raise ValueError(
f"`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`.")
UpperCAmelCase_ = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size
UpperCAmelCase_ = torch.cat(
[self.ff(_lowercase) for hid_slice in norm_hidden_states.chunk(_lowercase , dim=self._chunk_dim)] , dim=self._chunk_dim , )
else:
UpperCAmelCase_ = self.ff(_lowercase)
if self.use_ada_layer_norm_zero:
UpperCAmelCase_ = gate_mlp.unsqueeze(1) * ff_output
UpperCAmelCase_ = ff_output + hidden_states
return hidden_states
class a_ ( nn.Module ):
def __init__( self :List[str] , _lowercase :int , _lowercase :Optional[int] = None , _lowercase :int = 4 , _lowercase :float = 0.0 , _lowercase :str = "geglu" , _lowercase :bool = False , ) -> int:
super().__init__()
UpperCAmelCase_ = int(dim * mult)
UpperCAmelCase_ = dim_out if dim_out is not None else dim
if activation_fn == "gelu":
UpperCAmelCase_ = GELU(_lowercase , _lowercase)
if activation_fn == "gelu-approximate":
UpperCAmelCase_ = GELU(_lowercase , _lowercase , approximate='''tanh''')
elif activation_fn == "geglu":
UpperCAmelCase_ = GEGLU(_lowercase , _lowercase)
elif activation_fn == "geglu-approximate":
UpperCAmelCase_ = ApproximateGELU(_lowercase , _lowercase)
UpperCAmelCase_ = nn.ModuleList([])
# project in
self.net.append(_lowercase)
# project dropout
self.net.append(nn.Dropout(_lowercase))
# project out
self.net.append(nn.Linear(_lowercase , _lowercase))
# FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout
if final_dropout:
self.net.append(nn.Dropout(_lowercase))
def __a ( self :Union[str, Any] , _lowercase :Dict) -> Optional[Any]:
for module in self.net:
UpperCAmelCase_ = module(_lowercase)
return hidden_states
class a_ ( nn.Module ):
def __init__( self :Any , _lowercase :int , _lowercase :int , _lowercase :str = "none") -> int:
super().__init__()
UpperCAmelCase_ = nn.Linear(_lowercase , _lowercase)
UpperCAmelCase_ = approximate
def __a ( self :Any , _lowercase :int) -> List[Any]:
if gate.device.type != "mps":
return F.gelu(_lowercase , approximate=self.approximate)
# mps: gelu is not implemented for float16
return F.gelu(gate.to(dtype=torch.floataa) , approximate=self.approximate).to(dtype=gate.dtype)
def __a ( self :List[str] , _lowercase :List[Any]) -> Dict:
UpperCAmelCase_ = self.proj(_lowercase)
UpperCAmelCase_ = self.gelu(_lowercase)
return hidden_states
class a_ ( nn.Module ):
def __init__( self :Optional[int] , _lowercase :int , _lowercase :int) -> List[Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(_lowercase , dim_out * 2)
def __a ( self :Tuple , _lowercase :Tuple) -> Tuple:
if gate.device.type != "mps":
return F.gelu(_lowercase)
# mps: gelu is not implemented for float16
return F.gelu(gate.to(dtype=torch.floataa)).to(dtype=gate.dtype)
def __a ( self :str , _lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ , UpperCAmelCase_ = self.proj(_lowercase).chunk(2 , dim=-1)
return hidden_states * self.gelu(_lowercase)
class a_ ( nn.Module ):
def __init__( self :int , _lowercase :int , _lowercase :int) -> Any:
super().__init__()
UpperCAmelCase_ = nn.Linear(_lowercase , _lowercase)
def __a ( self :int , _lowercase :List[str]) -> int:
UpperCAmelCase_ = self.proj(_lowercase)
return x * torch.sigmoid(1.702 * x)
class a_ ( nn.Module ):
def __init__( self :Dict , _lowercase :str , _lowercase :List[str]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Embedding(_lowercase , _lowercase)
UpperCAmelCase_ = nn.SiLU()
UpperCAmelCase_ = nn.Linear(_lowercase , embedding_dim * 2)
UpperCAmelCase_ = nn.LayerNorm(_lowercase , elementwise_affine=_lowercase)
def __a ( self :Dict , _lowercase :Dict , _lowercase :List[Any]) -> str:
UpperCAmelCase_ = self.linear(self.silu(self.emb(_lowercase)))
UpperCAmelCase_ , UpperCAmelCase_ = torch.chunk(_lowercase , 2)
UpperCAmelCase_ = self.norm(_lowercase) * (1 + scale) + shift
return x
class a_ ( nn.Module ):
def __init__( self :Optional[int] , _lowercase :str , _lowercase :Optional[int]) -> Optional[Any]:
super().__init__()
UpperCAmelCase_ = CombinedTimestepLabelEmbeddings(_lowercase , _lowercase)
UpperCAmelCase_ = nn.SiLU()
UpperCAmelCase_ = nn.Linear(_lowercase , 6 * embedding_dim , bias=_lowercase)
UpperCAmelCase_ = nn.LayerNorm(_lowercase , elementwise_affine=_lowercase , eps=1E-6)
def __a ( self :str , _lowercase :Any , _lowercase :List[str] , _lowercase :List[str] , _lowercase :int=None) -> Union[str, Any]:
UpperCAmelCase_ = self.linear(self.silu(self.emb(_lowercase , _lowercase , hidden_dtype=_lowercase)))
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = emb.chunk(6 , dim=1)
UpperCAmelCase_ = self.norm(_lowercase) * (1 + scale_msa[:, None]) + shift_msa[:, None]
return x, gate_msa, shift_mlp, scale_mlp, gate_mlp
class a_ ( nn.Module ):
def __init__( self :Any , _lowercase :int , _lowercase :int , _lowercase :int , _lowercase :Optional[str] = None , _lowercase :float = 1E-5) -> str:
super().__init__()
UpperCAmelCase_ = num_groups
UpperCAmelCase_ = eps
if act_fn is None:
UpperCAmelCase_ = None
else:
UpperCAmelCase_ = get_activation(_lowercase)
UpperCAmelCase_ = nn.Linear(_lowercase , out_dim * 2)
def __a ( self :Optional[int] , _lowercase :str , _lowercase :Dict) -> Tuple:
if self.act:
UpperCAmelCase_ = self.act(_lowercase)
UpperCAmelCase_ = self.linear(_lowercase)
UpperCAmelCase_ = emb[:, :, None, None]
UpperCAmelCase_ , UpperCAmelCase_ = emb.chunk(2 , dim=1)
UpperCAmelCase_ = F.group_norm(_lowercase , self.num_groups , eps=self.eps)
UpperCAmelCase_ = x * (1 + scale) + shift
return x
| 344 |
from __future__ import annotations
import string
from itertools import cycle, product
from pathlib import Path
UpperCamelCase_ = (
string.ascii_letters + string.digits + string.punctuation + string.whitespace
)
UpperCamelCase_ = [ord(letter) for letter in string.ascii_lowercase]
UpperCamelCase_ = {ord(char) for char in VALID_CHARS}
UpperCamelCase_ = ["the", "be", "to", "of", "and", "in", "that", "have"]
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str | None:
'''simple docstring'''
UpperCAmelCase_ = ""
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
for keychar, cipherchar in zip(cycle(__UpperCAmelCase ) , __UpperCAmelCase ):
UpperCAmelCase_ = cipherchar ^ keychar
if decodedchar not in VALID_INTS:
return None
decoded += chr(__UpperCAmelCase )
return decoded
def A ( __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
for key in product(__UpperCAmelCase , repeat=3 ):
UpperCAmelCase_ = try_key(__UpperCAmelCase , __UpperCAmelCase )
if encoded is not None:
possibles.append(__UpperCAmelCase )
return possibles
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
return [possible for possible in possibles if common_word in possible.lower()]
def A ( __UpperCAmelCase = "p059_cipher.txt" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = Path(__UpperCAmelCase ).parent.joinpath(__UpperCAmelCase ).read_text(encoding='''utf-8''' )
UpperCAmelCase_ = [int(__UpperCAmelCase ) for number in data.strip().split(''',''' )]
UpperCAmelCase_ = filter_valid_chars(__UpperCAmelCase )
for common_word in COMMON_WORDS:
UpperCAmelCase_ = filter_common_word(__UpperCAmelCase , __UpperCAmelCase )
if len(__UpperCAmelCase ) == 1:
break
UpperCAmelCase_ = possibles[0]
return sum(ord(__UpperCAmelCase ) for char in decoded_text )
if __name__ == "__main__":
print(f"{solution() = }")
| 344 | 1 |
import json
from typing import Dict, List, Optional, Tuple, Union
from tokenizers import pre_tokenizers, processors
from ...tokenization_utils_base import AddedToken, BatchEncoding, EncodedInput
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import PaddingStrategy, logging
from .tokenization_led import LEDTokenizer
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {"vocab_file": "vocab.json", "merges_file": "merges.txt", "tokenizer_file": "tokenizer.json"}
UpperCamelCase_ = {
"vocab_file": {
"allenai/led-base-16384": "https://huggingface.co/allenai/led-base-16384/resolve/main/vocab.json",
},
"merges_file": {
"allenai/led-base-16384": "https://huggingface.co/allenai/led-base-16384/resolve/main/merges.txt",
},
"tokenizer_file": {
"allenai/led-base-16384": "https://huggingface.co/allenai/led-base-16384/resolve/main/tokenizer.json",
},
}
UpperCamelCase_ = {
"allenai/led-base-16384": 16_384,
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[int] =VOCAB_FILES_NAMES
UpperCamelCase__ : Tuple =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : str =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Optional[Any] =LEDTokenizer
UpperCamelCase__ : List[str] =["input_ids", "attention_mask"]
def __init__( self :Union[str, Any] , _lowercase :Dict=None , _lowercase :Optional[Any]=None , _lowercase :List[Any]=None , _lowercase :Dict="replace" , _lowercase :Optional[int]="<s>" , _lowercase :Any="</s>" , _lowercase :Union[str, Any]="</s>" , _lowercase :List[Any]="<s>" , _lowercase :int="<unk>" , _lowercase :List[str]="<pad>" , _lowercase :int="<mask>" , _lowercase :Tuple=False , _lowercase :List[str]=True , **_lowercase :str , ) -> Dict:
super().__init__(
_lowercase , _lowercase , tokenizer_file=_lowercase , errors=_lowercase , bos_token=_lowercase , eos_token=_lowercase , sep_token=_lowercase , cls_token=_lowercase , unk_token=_lowercase , pad_token=_lowercase , mask_token=_lowercase , add_prefix_space=_lowercase , trim_offsets=_lowercase , **_lowercase , )
UpperCAmelCase_ = json.loads(self.backend_tokenizer.pre_tokenizer.__getstate__())
if pre_tok_state.get('''add_prefix_space''' , _lowercase) != add_prefix_space:
UpperCAmelCase_ = getattr(_lowercase , pre_tok_state.pop('''type'''))
UpperCAmelCase_ = add_prefix_space
UpperCAmelCase_ = pre_tok_class(**_lowercase)
UpperCAmelCase_ = add_prefix_space
# the pre_tokenizer is already updated in the GPT2TokenizerFast `__init__`
UpperCAmelCase_ = '''post_processor'''
UpperCAmelCase_ = getattr(self.backend_tokenizer , _lowercase , _lowercase)
if tokenizer_component_instance:
UpperCAmelCase_ = json.loads(tokenizer_component_instance.__getstate__())
# The lists 'sep' and 'cls' must be cased in tuples for the object `post_processor_class`
if "sep" in state:
UpperCAmelCase_ = tuple(state['''sep'''])
if "cls" in state:
UpperCAmelCase_ = tuple(state['''cls'''])
UpperCAmelCase_ = False
if state.get('''add_prefix_space''' , _lowercase) != add_prefix_space:
UpperCAmelCase_ = add_prefix_space
UpperCAmelCase_ = True
if state.get('''trim_offsets''' , _lowercase) != trim_offsets:
UpperCAmelCase_ = trim_offsets
UpperCAmelCase_ = True
if changes_to_apply:
UpperCAmelCase_ = getattr(_lowercase , state.pop('''type'''))
UpperCAmelCase_ = component_class(**_lowercase)
setattr(self.backend_tokenizer , _lowercase , _lowercase)
@property
# Copied from transformers.models.bart.tokenization_bart_fast.BartTokenizerFast.mask_token with BART->LED
def __a ( self :List[Any]) -> str:
if self._mask_token is None:
if self.verbose:
logger.error('''Using mask_token, but it is not set yet.''')
return None
return str(self._mask_token)
@mask_token.setter
def __a ( self :List[str] , _lowercase :int) -> Tuple:
UpperCAmelCase_ = AddedToken(_lowercase , lstrip=_lowercase , rstrip=_lowercase) if isinstance(_lowercase , _lowercase) else value
UpperCAmelCase_ = value
def __a ( self :Union[str, Any] , *_lowercase :Tuple , **_lowercase :str) -> BatchEncoding:
UpperCAmelCase_ = kwargs.get('''is_split_into_words''' , _lowercase)
if is_split_into_words and not self.add_prefix_space:
raise ValueError(
f"You need to instantiate {self.__class__.__name__} with add_prefix_space=True "
'''to use it with pretokenized inputs.''')
return super()._batch_encode_plus(*_lowercase , **_lowercase)
def __a ( self :int , *_lowercase :Any , **_lowercase :List[str]) -> BatchEncoding:
UpperCAmelCase_ = kwargs.get('''is_split_into_words''' , _lowercase)
if is_split_into_words and not self.add_prefix_space:
raise ValueError(
f"You need to instantiate {self.__class__.__name__} with add_prefix_space=True "
'''to use it with pretokenized inputs.''')
return super()._encode_plus(*_lowercase , **_lowercase)
def __a ( self :Any , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
UpperCAmelCase_ = self._tokenizer.model.save(_lowercase , name=_lowercase)
return tuple(_lowercase)
def __a ( self :str , _lowercase :Optional[int] , _lowercase :Union[str, Any]=None) -> str:
UpperCAmelCase_ = [self.bos_token_id] + token_ids_a + [self.eos_token_id]
if token_ids_a is None:
return output
return output + [self.eos_token_id] + token_ids_a + [self.eos_token_id]
def __a ( self :Union[str, Any] , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep) * [0]
return len(cls + token_ids_a + sep + sep + token_ids_a + sep) * [0]
def __a ( self :Tuple , _lowercase :Union[Dict[str, EncodedInput], BatchEncoding] , _lowercase :Optional[int] = None , _lowercase :PaddingStrategy = PaddingStrategy.DO_NOT_PAD , _lowercase :Optional[int] = None , _lowercase :Optional[bool] = None , ) -> dict:
UpperCAmelCase_ = super()._pad(
encoded_inputs=_lowercase , max_length=_lowercase , padding_strategy=_lowercase , pad_to_multiple_of=_lowercase , return_attention_mask=_lowercase , )
# Load from model defaults
if return_attention_mask is None:
UpperCAmelCase_ = '''attention_mask''' in self.model_input_names
if return_attention_mask and "global_attention_mask" in encoded_inputs:
UpperCAmelCase_ = encoded_inputs[self.model_input_names[0]]
# `global_attention_mask` need to have the same length as other (sequential) inputs.
UpperCAmelCase_ = len(encoded_inputs['''global_attention_mask''']) != len(_lowercase)
if needs_to_be_padded:
UpperCAmelCase_ = len(_lowercase) - len(encoded_inputs['''global_attention_mask'''])
if self.padding_side == "right":
# Use `-1` since `0` in `global_attention_mask` means `local attention` instead of `not to attend`
UpperCAmelCase_ = (
encoded_inputs['''global_attention_mask'''] + [-1] * difference
)
elif self.padding_side == "left":
UpperCAmelCase_ = [-1] * difference + encoded_inputs[
'''global_attention_mask'''
]
else:
raise ValueError('''Invalid padding strategy:''' + str(self.padding_side))
return encoded_inputs
| 344 |
import pytest
UpperCamelCase_ = "__dummy_dataset1__"
UpperCamelCase_ = "\nimport json\nimport os\n\nimport datasets\n\n\nREPO_URL = \"https://huggingface.co/datasets/albertvillanova/tests-raw-jsonl/resolve/main/\"\nURLS = {\"train\": REPO_URL + \"wikiann-bn-train.jsonl\", \"validation\": REPO_URL + \"wikiann-bn-validation.jsonl\"}\n\n\nclass __DummyDataset1__(datasets.GeneratorBasedBuilder):\n\n def _info(self):\n features = datasets.Features(\n {\n \"tokens\": datasets.Sequence(datasets.Value(\"string\")),\n \"ner_tags\": datasets.Sequence(\n datasets.features.ClassLabel(\n names=[\n \"O\",\n \"B-PER\",\n \"I-PER\",\n \"B-ORG\",\n \"I-ORG\",\n \"B-LOC\",\n \"I-LOC\",\n ]\n )\n ),\n \"langs\": datasets.Sequence(datasets.Value(\"string\")),\n \"spans\": datasets.Sequence(datasets.Value(\"string\")),\n }\n )\n return datasets.DatasetInfo(features=features)\n\n def _split_generators(self, dl_manager):\n dl_path = dl_manager.download(URLS)\n return [\n datasets.SplitGenerator(datasets.Split.TRAIN, gen_kwargs={\"filepath\": dl_path[\"train\"]}),\n datasets.SplitGenerator(datasets.Split.VALIDATION, gen_kwargs={\"filepath\": dl_path[\"validation\"]}),\n ]\n\n def _generate_examples(self, filepath):\n with open(filepath, \"r\", encoding=\"utf-8\") as f:\n for i, line in enumerate(f):\n yield i, json.loads(line)\n"
@pytest.fixture
def A ( ) -> str:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_NAME
@pytest.fixture
def A ( ) -> Any:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_CODE
@pytest.fixture
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = dataset_loading_script_name
UpperCAmelCase_ = tmp_path / '''datasets''' / script_name
script_dir.mkdir(parents=__UpperCAmelCase )
UpperCAmelCase_ = script_dir / f"{script_name}.py"
with open(__UpperCAmelCase , '''w''' ) as f:
f.write(__UpperCAmelCase )
return str(__UpperCAmelCase )
| 344 | 1 |
import importlib
import inspect
import json
import os
import re
import shutil
import sys
from pathlib import Path
from typing import Dict, Optional, Union
from urllib import request
from huggingface_hub import HfFolder, cached_download, hf_hub_download, model_info
from packaging import version
from .. import __version__
from . import DIFFUSERS_DYNAMIC_MODULE_NAME, HF_MODULES_CACHE, logging
UpperCamelCase_ = (
"https://raw.githubusercontent.com/huggingface/diffusers/{revision}/examples/community/{pipeline}.py"
)
UpperCamelCase_ = logging.get_logger(__name__) # pylint: disable=invalid-name
def A ( ) -> Any:
'''simple docstring'''
UpperCAmelCase_ = '''https://pypi.org/pypi/diffusers/json'''
UpperCAmelCase_ = json.loads(request.urlopen(__UpperCAmelCase ).read() )['''releases'''].keys()
return sorted(__UpperCAmelCase , key=lambda __UpperCAmelCase : version.Version(__UpperCAmelCase ) )
def A ( ) -> Any:
'''simple docstring'''
if HF_MODULES_CACHE in sys.path:
return
sys.path.append(__UpperCAmelCase )
os.makedirs(__UpperCAmelCase , exist_ok=__UpperCAmelCase )
UpperCAmelCase_ = Path(__UpperCAmelCase ) / '''__init__.py'''
if not init_path.exists():
init_path.touch()
def A ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
init_hf_modules()
UpperCAmelCase_ = Path(__UpperCAmelCase ) / name
# If the parent module does not exist yet, recursively create it.
if not dynamic_module_path.parent.exists():
create_dynamic_module(dynamic_module_path.parent )
os.makedirs(__UpperCAmelCase , exist_ok=__UpperCAmelCase )
UpperCAmelCase_ = dynamic_module_path / '''__init__.py'''
if not init_path.exists():
init_path.touch()
def A ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
with open(__UpperCAmelCase , '''r''' , encoding='''utf-8''' ) as f:
UpperCAmelCase_ = f.read()
# Imports of the form `import .xxx`
UpperCAmelCase_ = re.findall('''^\s*import\s+\.(\S+)\s*$''' , __UpperCAmelCase , flags=re.MULTILINE )
# Imports of the form `from .xxx import yyy`
relative_imports += re.findall('''^\s*from\s+\.(\S+)\s+import''' , __UpperCAmelCase , flags=re.MULTILINE )
# Unique-ify
return list(set(__UpperCAmelCase ) )
def A ( __UpperCAmelCase ) -> str:
'''simple docstring'''
UpperCAmelCase_ = False
UpperCAmelCase_ = [module_file]
UpperCAmelCase_ = []
# Let's recurse through all relative imports
while not no_change:
UpperCAmelCase_ = []
for f in files_to_check:
new_imports.extend(get_relative_imports(__UpperCAmelCase ) )
UpperCAmelCase_ = Path(__UpperCAmelCase ).parent
UpperCAmelCase_ = [str(module_path / m ) for m in new_imports]
UpperCAmelCase_ = [f for f in new_import_files if f not in all_relative_imports]
UpperCAmelCase_ = [f"{f}.py" for f in new_import_files]
UpperCAmelCase_ = len(__UpperCAmelCase ) == 0
all_relative_imports.extend(__UpperCAmelCase )
return all_relative_imports
def A ( __UpperCAmelCase ) -> Dict:
'''simple docstring'''
with open(__UpperCAmelCase , '''r''' , encoding='''utf-8''' ) as f:
UpperCAmelCase_ = f.read()
# Imports of the form `import xxx`
UpperCAmelCase_ = re.findall('''^\s*import\s+(\S+)\s*$''' , __UpperCAmelCase , flags=re.MULTILINE )
# Imports of the form `from xxx import yyy`
imports += re.findall('''^\s*from\s+(\S+)\s+import''' , __UpperCAmelCase , flags=re.MULTILINE )
# Only keep the top-level module
UpperCAmelCase_ = [imp.split('''.''' )[0] for imp in imports if not imp.startswith('''.''' )]
# Unique-ify and test we got them all
UpperCAmelCase_ = list(set(__UpperCAmelCase ) )
UpperCAmelCase_ = []
for imp in imports:
try:
importlib.import_module(__UpperCAmelCase )
except ImportError:
missing_packages.append(__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0:
raise ImportError(
'''This modeling file requires the following packages that were not found in your environment: '''
f"{', '.join(__UpperCAmelCase )}. Run `pip install {' '.join(__UpperCAmelCase )}`" )
return get_relative_imports(__UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> Dict:
'''simple docstring'''
UpperCAmelCase_ = module_path.replace(os.path.sep , '''.''' )
UpperCAmelCase_ = importlib.import_module(__UpperCAmelCase )
if class_name is None:
return find_pipeline_class(__UpperCAmelCase )
return getattr(__UpperCAmelCase , __UpperCAmelCase )
def A ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from ..pipelines import DiffusionPipeline
UpperCAmelCase_ = dict(inspect.getmembers(__UpperCAmelCase , inspect.isclass ) )
UpperCAmelCase_ = None
for cls_name, cls in cls_members.items():
if (
cls_name != DiffusionPipeline.__name__
and issubclass(cls , __UpperCAmelCase )
and cls.__module__.split('''.''' )[0] != "diffusers"
):
if pipeline_class is not None:
raise ValueError(
f"Multiple classes that inherit from {DiffusionPipeline.__name__} have been found:"
f" {pipeline_class.__name__}, and {cls_name}. Please make sure to define only one in"
f" {loaded_module}." )
UpperCAmelCase_ = cls
return pipeline_class
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase = None , __UpperCAmelCase = False , __UpperCAmelCase = False , __UpperCAmelCase = None , __UpperCAmelCase = None , __UpperCAmelCase = None , __UpperCAmelCase = False , ) -> List[Any]:
'''simple docstring'''
UpperCAmelCase_ = str(__UpperCAmelCase )
UpperCAmelCase_ = os.path.join(__UpperCAmelCase , __UpperCAmelCase )
if os.path.isfile(__UpperCAmelCase ):
UpperCAmelCase_ = module_file_or_url
UpperCAmelCase_ = '''local'''
elif pretrained_model_name_or_path.count('''/''' ) == 0:
UpperCAmelCase_ = get_diffusers_versions()
# cut ".dev0"
UpperCAmelCase_ = '''v''' + '''.'''.join(__version__.split('''.''' )[:3] )
# retrieve github version that matches
if revision is None:
UpperCAmelCase_ = latest_version if latest_version[1:] in available_versions else '''main'''
logger.info(f"Defaulting to latest_version: {revision}." )
elif revision in available_versions:
UpperCAmelCase_ = f"v{revision}"
elif revision == "main":
UpperCAmelCase_ = revision
else:
raise ValueError(
f"`custom_revision`: {revision} does not exist. Please make sure to choose one of"
f" {', '.join(available_versions + ['main'] )}." )
# community pipeline on GitHub
UpperCAmelCase_ = COMMUNITY_PIPELINES_URL.format(revision=__UpperCAmelCase , pipeline=__UpperCAmelCase )
try:
UpperCAmelCase_ = cached_download(
__UpperCAmelCase , cache_dir=__UpperCAmelCase , force_download=__UpperCAmelCase , proxies=__UpperCAmelCase , resume_download=__UpperCAmelCase , local_files_only=__UpperCAmelCase , use_auth_token=__UpperCAmelCase , )
UpperCAmelCase_ = '''git'''
UpperCAmelCase_ = pretrained_model_name_or_path + '''.py'''
except EnvironmentError:
logger.error(f"Could not locate the {module_file} inside {pretrained_model_name_or_path}." )
raise
else:
try:
# Load from URL or cache if already cached
UpperCAmelCase_ = hf_hub_download(
__UpperCAmelCase , __UpperCAmelCase , cache_dir=__UpperCAmelCase , force_download=__UpperCAmelCase , proxies=__UpperCAmelCase , resume_download=__UpperCAmelCase , local_files_only=__UpperCAmelCase , use_auth_token=__UpperCAmelCase , )
UpperCAmelCase_ = os.path.join('''local''' , '''--'''.join(pretrained_model_name_or_path.split('''/''' ) ) )
except EnvironmentError:
logger.error(f"Could not locate the {module_file} inside {pretrained_model_name_or_path}." )
raise
# Check we have all the requirements in our environment
UpperCAmelCase_ = check_imports(__UpperCAmelCase )
# Now we move the module inside our cached dynamic modules.
UpperCAmelCase_ = DIFFUSERS_DYNAMIC_MODULE_NAME + os.path.sep + submodule
create_dynamic_module(__UpperCAmelCase )
UpperCAmelCase_ = Path(__UpperCAmelCase ) / full_submodule
if submodule == "local" or submodule == "git":
# We always copy local files (we could hash the file to see if there was a change, and give them the name of
# that hash, to only copy when there is a modification but it seems overkill for now).
# The only reason we do the copy is to avoid putting too many folders in sys.path.
shutil.copy(__UpperCAmelCase , submodule_path / module_file )
for module_needed in modules_needed:
UpperCAmelCase_ = f"{module_needed}.py"
shutil.copy(os.path.join(__UpperCAmelCase , __UpperCAmelCase ) , submodule_path / module_needed )
else:
# Get the commit hash
# TODO: we will get this info in the etag soon, so retrieve it from there and not here.
if isinstance(__UpperCAmelCase , __UpperCAmelCase ):
UpperCAmelCase_ = use_auth_token
elif use_auth_token is True:
UpperCAmelCase_ = HfFolder.get_token()
else:
UpperCAmelCase_ = None
UpperCAmelCase_ = model_info(__UpperCAmelCase , revision=__UpperCAmelCase , token=__UpperCAmelCase ).sha
# The module file will end up being placed in a subfolder with the git hash of the repo. This way we get the
# benefit of versioning.
UpperCAmelCase_ = submodule_path / commit_hash
UpperCAmelCase_ = full_submodule + os.path.sep + commit_hash
create_dynamic_module(__UpperCAmelCase )
if not (submodule_path / module_file).exists():
shutil.copy(__UpperCAmelCase , submodule_path / module_file )
# Make sure we also have every file with relative
for module_needed in modules_needed:
if not (submodule_path / module_needed).exists():
get_cached_module_file(
__UpperCAmelCase , f"{module_needed}.py" , cache_dir=__UpperCAmelCase , force_download=__UpperCAmelCase , resume_download=__UpperCAmelCase , proxies=__UpperCAmelCase , use_auth_token=__UpperCAmelCase , revision=__UpperCAmelCase , local_files_only=__UpperCAmelCase , )
return os.path.join(__UpperCAmelCase , __UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase = None , __UpperCAmelCase = None , __UpperCAmelCase = False , __UpperCAmelCase = False , __UpperCAmelCase = None , __UpperCAmelCase = None , __UpperCAmelCase = None , __UpperCAmelCase = False , **__UpperCAmelCase , ) -> int:
'''simple docstring'''
UpperCAmelCase_ = get_cached_module_file(
__UpperCAmelCase , __UpperCAmelCase , cache_dir=__UpperCAmelCase , force_download=__UpperCAmelCase , resume_download=__UpperCAmelCase , proxies=__UpperCAmelCase , use_auth_token=__UpperCAmelCase , revision=__UpperCAmelCase , local_files_only=__UpperCAmelCase , )
return get_class_in_module(__UpperCAmelCase , final_module.replace('''.py''' , '''''' ) )
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"s-JoL/Open-Llama-V1": "https://huggingface.co/s-JoL/Open-Llama-V1/blob/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="open-llama"
def __init__( self :Union[str, Any] , _lowercase :List[Any]=100000 , _lowercase :Dict=4096 , _lowercase :List[Any]=11008 , _lowercase :Optional[int]=32 , _lowercase :Union[str, Any]=32 , _lowercase :List[str]="silu" , _lowercase :Union[str, Any]=2048 , _lowercase :Any=0.02 , _lowercase :Optional[Any]=1E-6 , _lowercase :str=True , _lowercase :str=0 , _lowercase :Any=1 , _lowercase :Optional[Any]=2 , _lowercase :str=False , _lowercase :Dict=True , _lowercase :Optional[Any]=0.1 , _lowercase :Tuple=0.1 , _lowercase :Dict=True , _lowercase :List[Any]=True , _lowercase :Dict=None , **_lowercase :Optional[int] , ) -> List[Any]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = rms_norm_eps
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = kwargs.pop(
'''use_memorry_efficient_attention''' , _lowercase)
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_dropout_prob
UpperCAmelCase_ = use_stable_embedding
UpperCAmelCase_ = shared_input_output_embedding
UpperCAmelCase_ = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , tie_word_embeddings=_lowercase , **_lowercase , )
def __a ( self :int) -> str:
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _lowercase) or len(self.rope_scaling) != 2:
raise ValueError(
'''`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, '''
f"got {self.rope_scaling}")
UpperCAmelCase_ = self.rope_scaling.get('''type''' , _lowercase)
UpperCAmelCase_ = self.rope_scaling.get('''factor''' , _lowercase)
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}")
if rope_scaling_factor is None or not isinstance(_lowercase , _lowercase) or rope_scaling_factor <= 1.0:
raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
| 344 | 1 |
import importlib
import sys
from argparse import REMAINDER, ArgumentParser
from pathlib import Path
import torch_xla.distributed.xla_multiprocessing as xmp
def A ( ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = ArgumentParser(
description=(
'''PyTorch TPU distributed training launch helper utility that will spawn up multiple distributed processes'''
) )
# Optional arguments for the launch helper
parser.add_argument('''--num_cores''' , type=__UpperCAmelCase , default=1 , help='''Number of TPU cores to use (1 or 8).''' )
# positional
parser.add_argument(
'''training_script''' , type=__UpperCAmelCase , help=(
'''The full path to the single TPU training '''
'''program/script to be launched in parallel, '''
'''followed by all the arguments for the '''
'''training script'''
) , )
# rest from the training program
parser.add_argument('''training_script_args''' , nargs=__UpperCAmelCase )
return parser.parse_args()
def A ( ) -> Dict:
'''simple docstring'''
UpperCAmelCase_ = parse_args()
# Import training_script as a module.
UpperCAmelCase_ = Path(args.training_script )
sys.path.append(str(script_fpath.parent.resolve() ) )
UpperCAmelCase_ = script_fpath.stem
UpperCAmelCase_ = importlib.import_module(__UpperCAmelCase )
# Patch sys.argv
UpperCAmelCase_ = [args.training_script] + args.training_script_args + ['''--tpu_num_cores''', str(args.num_cores )]
xmp.spawn(mod._mp_fn , args=() , nprocs=args.num_cores )
if __name__ == "__main__":
main()
| 344 |
import tempfile
import torch
from diffusers import (
DEISMultistepScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
UniPCMultistepScheduler,
)
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] =(DPMSolverSinglestepScheduler,)
UpperCamelCase__ : Tuple =(("num_inference_steps", 25),)
def __a ( self :List[Any] , **_lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
'''solver_order''': 2,
'''prediction_type''': '''epsilon''',
'''thresholding''': False,
'''sample_max_value''': 1.0,
'''algorithm_type''': '''dpmsolver++''',
'''solver_type''': '''midpoint''',
'''lambda_min_clipped''': -float('''inf'''),
'''variance_type''': None,
}
config.update(**_lowercase)
return config
def __a ( self :Union[str, Any] , _lowercase :List[Any]=0 , **_lowercase :Optional[int]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ , UpperCAmelCase_ = sample, sample
for t in range(_lowercase , time_step + scheduler.config.solver_order + 1):
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Union[str, Any]) -> List[Any]:
pass
def __a ( self :Optional[Any] , _lowercase :str=0 , **_lowercase :Union[str, Any]) -> Dict:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Dict , _lowercase :Union[str, Any]=None , **_lowercase :List[Any]) -> int:
if scheduler is None:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :int) -> Tuple:
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = 50
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
# make sure that the first t is uneven
for i, t in enumerate(scheduler.timesteps[3:]):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_574) < 1E-3
def __a ( self :List[Any]) -> List[Any]:
for timesteps in [25, 50, 100, 999, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :int) -> Optional[Any]:
# make sure that iterating over schedulers with same config names gives same results
# for defaults
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
UpperCAmelCase_ = DEISMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = UniPCMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverSinglestepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Tuple) -> int:
self.check_over_configs(thresholding=_lowercase)
for order in [1, 2, 3]:
for solver_type in ["midpoint", "heun"]:
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
thresholding=_lowercase , prediction_type=_lowercase , sample_max_value=_lowercase , algorithm_type='''dpmsolver++''' , solver_order=_lowercase , solver_type=_lowercase , )
def __a ( self :List[Any]) -> Any:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Optional[int]:
for algorithm_type in ["dpmsolver", "dpmsolver++"]:
for solver_type in ["midpoint", "heun"]:
for order in [1, 2, 3]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
UpperCAmelCase_ = self.full_loop(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
assert not torch.isnan(_lowercase).any(), "Samples have nan numbers"
def __a ( self :Tuple) -> int:
self.check_over_configs(lower_order_final=_lowercase)
self.check_over_configs(lower_order_final=_lowercase)
def __a ( self :Tuple) -> Optional[Any]:
self.check_over_configs(lambda_min_clipped=-float('''inf'''))
self.check_over_configs(lambda_min_clipped=-5.1)
def __a ( self :Any) -> List[str]:
self.check_over_configs(variance_type=_lowercase)
self.check_over_configs(variance_type='''learned_range''')
def __a ( self :Any) -> Dict:
for num_inference_steps in [1, 2, 3, 5, 10, 50, 100, 999, 1000]:
self.check_over_forward(num_inference_steps=_lowercase , time_step=0)
def __a ( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Any) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop(use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_248) < 1E-3
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.1_453) < 1E-3
def __a ( self :List[Any]) -> Dict:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''' , use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.0_649) < 1E-3
def __a ( self :Any) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(thresholding=_lowercase , dynamic_thresholding_ratio=0)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter.half()
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
assert sample.dtype == torch.floataa
| 344 | 1 |
import argparse
import json
import os
import evaluate
import torch
from datasets import load_dataset
from torch.optim import AdamW
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed
from accelerate import Accelerator, DistributedType
from accelerate.utils.deepspeed import DummyOptim, DummyScheduler
UpperCamelCase_ = 16
UpperCamelCase_ = 32
def A ( __UpperCAmelCase , __UpperCAmelCase = 16 , __UpperCAmelCase = "bert-base-cased" ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = AutoTokenizer.from_pretrained(__UpperCAmelCase )
UpperCAmelCase_ = load_dataset('''glue''' , '''mrpc''' )
def tokenize_function(__UpperCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
UpperCAmelCase_ = tokenizer(examples['''sentence1'''] , examples['''sentence2'''] , truncation=__UpperCAmelCase , max_length=__UpperCAmelCase )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
UpperCAmelCase_ = datasets.map(
__UpperCAmelCase , batched=__UpperCAmelCase , remove_columns=['''idx''', '''sentence1''', '''sentence2'''] , load_from_cache_file=__UpperCAmelCase )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
UpperCAmelCase_ = tokenized_datasets.rename_column('''label''' , '''labels''' )
def collate_fn(__UpperCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
if accelerator.distributed_type == DistributedType.TPU:
return tokenizer.pad(__UpperCAmelCase , padding='''max_length''' , max_length=128 , return_tensors='''pt''' )
return tokenizer.pad(__UpperCAmelCase , padding='''longest''' , return_tensors='''pt''' )
# Instantiate dataloaders.
UpperCAmelCase_ = DataLoader(
tokenized_datasets['''train'''] , shuffle=__UpperCAmelCase , collate_fn=__UpperCAmelCase , batch_size=__UpperCAmelCase )
UpperCAmelCase_ = DataLoader(
tokenized_datasets['''validation'''] , shuffle=__UpperCAmelCase , collate_fn=__UpperCAmelCase , batch_size=__UpperCAmelCase )
return train_dataloader, eval_dataloader
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = Accelerator()
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
UpperCAmelCase_ = config['''lr''']
UpperCAmelCase_ = int(config['''num_epochs'''] )
UpperCAmelCase_ = int(config['''seed'''] )
UpperCAmelCase_ = int(config['''batch_size'''] )
UpperCAmelCase_ = args.model_name_or_path
set_seed(__UpperCAmelCase )
UpperCAmelCase_ , UpperCAmelCase_ = get_dataloaders(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
UpperCAmelCase_ = AutoModelForSequenceClassification.from_pretrained(__UpperCAmelCase , return_dict=__UpperCAmelCase )
# Instantiate optimizer
UpperCAmelCase_ = (
AdamW
if accelerator.state.deepspeed_plugin is None
or '''optimizer''' not in accelerator.state.deepspeed_plugin.deepspeed_config
else DummyOptim
)
UpperCAmelCase_ = optimizer_cls(params=model.parameters() , lr=__UpperCAmelCase )
if accelerator.state.deepspeed_plugin is not None:
UpperCAmelCase_ = accelerator.state.deepspeed_plugin.deepspeed_config[
'''gradient_accumulation_steps'''
]
else:
UpperCAmelCase_ = 1
UpperCAmelCase_ = (len(__UpperCAmelCase ) * num_epochs) // gradient_accumulation_steps
# Instantiate scheduler
if (
accelerator.state.deepspeed_plugin is None
or "scheduler" not in accelerator.state.deepspeed_plugin.deepspeed_config
):
UpperCAmelCase_ = get_linear_schedule_with_warmup(
optimizer=__UpperCAmelCase , num_warmup_steps=0 , num_training_steps=__UpperCAmelCase , )
else:
UpperCAmelCase_ = DummyScheduler(__UpperCAmelCase , total_num_steps=__UpperCAmelCase , warmup_num_steps=0 )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = accelerator.prepare(
__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
# We need to keep track of how many total steps we have iterated over
UpperCAmelCase_ = 0
# We also need to keep track of the stating epoch so files are named properly
UpperCAmelCase_ = 0
# Now we train the model
UpperCAmelCase_ = evaluate.load('''glue''' , '''mrpc''' )
UpperCAmelCase_ = 0
UpperCAmelCase_ = {}
for epoch in range(__UpperCAmelCase , __UpperCAmelCase ):
model.train()
for step, batch in enumerate(__UpperCAmelCase ):
UpperCAmelCase_ = model(**__UpperCAmelCase )
UpperCAmelCase_ = outputs.loss
UpperCAmelCase_ = loss / gradient_accumulation_steps
accelerator.backward(__UpperCAmelCase )
if step % gradient_accumulation_steps == 0:
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
overall_step += 1
model.eval()
UpperCAmelCase_ = 0
for step, batch in enumerate(__UpperCAmelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
with torch.no_grad():
UpperCAmelCase_ = model(**__UpperCAmelCase )
UpperCAmelCase_ = outputs.logits.argmax(dim=-1 )
# It is slightly faster to call this once, than multiple times
UpperCAmelCase_ , UpperCAmelCase_ = accelerator.gather(
(predictions, batch['''labels''']) ) # If we are in a multiprocess environment, the last batch has duplicates
if accelerator.use_distributed:
if step == len(__UpperCAmelCase ) - 1:
UpperCAmelCase_ = predictions[: len(eval_dataloader.dataset ) - samples_seen]
UpperCAmelCase_ = references[: len(eval_dataloader.dataset ) - samples_seen]
else:
samples_seen += references.shape[0]
metric.add_batch(
predictions=__UpperCAmelCase , references=__UpperCAmelCase , )
UpperCAmelCase_ = metric.compute()
# Use accelerator.print to print only on the main process.
accelerator.print(f"epoch {epoch}:" , __UpperCAmelCase )
UpperCAmelCase_ = eval_metric['''accuracy''']
if best_performance < eval_metric["accuracy"]:
UpperCAmelCase_ = eval_metric['''accuracy''']
if args.performance_lower_bound is not None:
assert (
args.performance_lower_bound <= best_performance
), f"Best performance metric {best_performance} is lower than the lower bound {args.performance_lower_bound}"
accelerator.wait_for_everyone()
if accelerator.is_main_process:
with open(os.path.join(args.output_dir , '''all_results.json''' ) , '''w''' ) as f:
json.dump(__UpperCAmelCase , __UpperCAmelCase )
def A ( ) -> Any:
'''simple docstring'''
UpperCAmelCase_ = argparse.ArgumentParser(description='''Simple example of training script tracking peak GPU memory usage.''' )
parser.add_argument(
'''--model_name_or_path''' , type=__UpperCAmelCase , default='''bert-base-cased''' , help='''Path to pretrained model or model identifier from huggingface.co/models.''' , required=__UpperCAmelCase , )
parser.add_argument(
'''--output_dir''' , type=__UpperCAmelCase , default='''.''' , help='''Optional save directory where all checkpoint folders will be stored. Default is the current working directory.''' , )
parser.add_argument(
'''--performance_lower_bound''' , type=__UpperCAmelCase , default=__UpperCAmelCase , help='''Optional lower bound for the performance metric. If set, the training will throw error when the performance metric drops below this value.''' , )
parser.add_argument(
'''--num_epochs''' , type=__UpperCAmelCase , default=3 , help='''Number of train epochs.''' , )
UpperCAmelCase_ = parser.parse_args()
UpperCAmelCase_ = {'''lr''': 2e-5, '''num_epochs''': args.num_epochs, '''seed''': 42, '''batch_size''': 16}
training_function(__UpperCAmelCase , __UpperCAmelCase )
if __name__ == "__main__":
main()
| 344 |
import inspect
import unittest
import torch
import torch.nn as nn
from accelerate.hooks import (
AlignDevicesHook,
ModelHook,
SequentialHook,
add_hook_to_module,
attach_align_device_hook,
remove_hook_from_module,
remove_hook_from_submodules,
)
from accelerate.test_utils import require_multi_gpu
class a_ ( nn.Module ):
def __init__( self :Optional[Any]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(3 , 4)
UpperCAmelCase_ = nn.BatchNormad(4)
UpperCAmelCase_ = nn.Linear(4 , 5)
def __a ( self :Dict , _lowercase :int) -> str:
return self.lineara(self.batchnorm(self.lineara(_lowercase)))
class a_ ( _snake_case ):
def __a ( self :Tuple , _lowercase :Optional[int] , *_lowercase :Union[str, Any] , **_lowercase :Any) -> Optional[Any]:
return (args[0] + 1,) + args[1:], kwargs
class a_ ( _snake_case ):
def __a ( self :Union[str, Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
return output + 1
class a_ ( unittest.TestCase ):
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
self.assertEqual(test_model._hf_hook , _lowercase)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[Any]) -> Any:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
add_hook_to_module(_lowercase , _lowercase , append=_lowercase)
self.assertEqual(isinstance(test_model._hf_hook , _lowercase) , _lowercase)
self.assertEqual(len(test_model._hf_hook.hooks) , 2)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[int]) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(x + 1)
UpperCAmelCase_ = test_model(x + 2)
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PreForwardHook() , PreForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , _lowercase , atol=1E-5)
def __a ( self :List[str]) -> int:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PostForwardHook() , PostForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , output + 2 , atol=1E-5)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1))
self.assertTrue(outputa.requires_grad)
UpperCAmelCase_ = True
UpperCAmelCase_ = test_model(_lowercase)
self.assertFalse(outputa.requires_grad)
@require_multi_gpu
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.batchnorm , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=1))
self.assertEqual(model.lineara.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.running_mean.device , torch.device(0))
self.assertEqual(model.lineara.weight.device , torch.device(1))
# We can still make a forward pass. The input does not need to be on any particular device
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(1))
# We can add a general hook to put back output on same device as input.
add_hook_to_module(_lowercase , AlignDevicesHook(io_same_device=_lowercase))
UpperCAmelCase_ = torch.randn(2 , 3).to(0)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(0))
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = {'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''', '''offload''': True}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(hook_kwargs['''execution_device'''])
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
UpperCAmelCase_ = {
'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''',
'''offload''': True,
'''offload_buffers''': True,
}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase)
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase , offload_buffers=_lowercase)
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict())
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict() , offload_buffers=_lowercase , )
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
| 344 | 1 |
from __future__ import annotations
from collections.abc import Sequence
from typing import Literal
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str | Literal[False]:
'''simple docstring'''
UpperCAmelCase_ = list(__UpperCAmelCase )
UpperCAmelCase_ = list(__UpperCAmelCase )
UpperCAmelCase_ = 0
for i in range(len(__UpperCAmelCase ) ):
if lista[i] != lista[i]:
count += 1
UpperCAmelCase_ = '''_'''
if count > 1:
return False
else:
return "".join(__UpperCAmelCase )
def A ( __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
while True:
UpperCAmelCase_ = ['''$'''] * len(__UpperCAmelCase )
UpperCAmelCase_ = []
for i in range(len(__UpperCAmelCase ) ):
for j in range(i + 1 , len(__UpperCAmelCase ) ):
UpperCAmelCase_ = compare_string(binary[i] , binary[j] )
if k is False:
UpperCAmelCase_ = '''*'''
UpperCAmelCase_ = '''*'''
temp.append('''X''' )
for i in range(len(__UpperCAmelCase ) ):
if checka[i] == "$":
pi.append(binary[i] )
if len(__UpperCAmelCase ) == 0:
return pi
UpperCAmelCase_ = list(set(__UpperCAmelCase ) )
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
for minterm in minterms:
UpperCAmelCase_ = ''''''
for _ in range(__UpperCAmelCase ):
UpperCAmelCase_ = str(minterm % 2 ) + string
minterm //= 2
temp.append(__UpperCAmelCase )
return temp
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> bool:
'''simple docstring'''
UpperCAmelCase_ = list(__UpperCAmelCase )
UpperCAmelCase_ = list(__UpperCAmelCase )
UpperCAmelCase_ = 0
for i in range(len(__UpperCAmelCase ) ):
if lista[i] != lista[i]:
count_n += 1
return count_n == count
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = [0] * len(__UpperCAmelCase )
for i in range(len(chart[0] ) ):
UpperCAmelCase_ = 0
UpperCAmelCase_ = -1
for j in range(len(__UpperCAmelCase ) ):
if chart[j][i] == 1:
count += 1
UpperCAmelCase_ = j
if count == 1:
UpperCAmelCase_ = 1
for i in range(len(__UpperCAmelCase ) ):
if select[i] == 1:
for j in range(len(chart[0] ) ):
if chart[i][j] == 1:
for k in range(len(__UpperCAmelCase ) ):
UpperCAmelCase_ = 0
temp.append(prime_implicants[i] )
while True:
UpperCAmelCase_ = 0
UpperCAmelCase_ = -1
UpperCAmelCase_ = 0
for i in range(len(__UpperCAmelCase ) ):
UpperCAmelCase_ = chart[i].count(1 )
if count_n > max_n:
UpperCAmelCase_ = count_n
UpperCAmelCase_ = i
if max_n == 0:
return temp
temp.append(prime_implicants[rem] )
for i in range(len(chart[0] ) ):
if chart[rem][i] == 1:
for j in range(len(__UpperCAmelCase ) ):
UpperCAmelCase_ = 0
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[list[int]]:
'''simple docstring'''
UpperCAmelCase_ = [[0 for x in range(len(__UpperCAmelCase ) )] for x in range(len(__UpperCAmelCase ) )]
for i in range(len(__UpperCAmelCase ) ):
UpperCAmelCase_ = prime_implicants[i].count('''_''' )
for j in range(len(__UpperCAmelCase ) ):
if is_for_table(prime_implicants[i] , binary[j] , __UpperCAmelCase ):
UpperCAmelCase_ = 1
return chart
def A ( ) -> None:
'''simple docstring'''
UpperCAmelCase_ = int(input('''Enter the no. of variables\n''' ) )
UpperCAmelCase_ = [
float(__UpperCAmelCase )
for x in input(
'''Enter the decimal representation of Minterms \'Spaces Separated\'\n''' ).split()
]
UpperCAmelCase_ = decimal_to_binary(__UpperCAmelCase , __UpperCAmelCase )
UpperCAmelCase_ = check(__UpperCAmelCase )
print('''Prime Implicants are:''' )
print(__UpperCAmelCase )
UpperCAmelCase_ = prime_implicant_chart(__UpperCAmelCase , __UpperCAmelCase )
UpperCAmelCase_ = selection(__UpperCAmelCase , __UpperCAmelCase )
print('''Essential Prime Implicants are:''' )
print(__UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 344 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import (
AutoProcessor,
BertTokenizerFast,
BlipImageProcessor,
GPTaTokenizer,
InstructBlipProcessor,
PreTrainedTokenizerFast,
)
@require_vision
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = BlipImageProcessor()
UpperCAmelCase_ = GPTaTokenizer.from_pretrained('''hf-internal-testing/tiny-random-GPT2Model''')
UpperCAmelCase_ = BertTokenizerFast.from_pretrained('''hf-internal-testing/tiny-random-bert''')
UpperCAmelCase_ = InstructBlipProcessor(_lowercase , _lowercase , _lowercase)
processor.save_pretrained(self.tmpdirname)
def __a ( self :List[Any] , **_lowercase :Dict) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).tokenizer
def __a ( self :Optional[Any] , **_lowercase :Optional[Any]) -> Optional[int]:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).image_processor
def __a ( self :Dict , **_lowercase :Tuple) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).qformer_tokenizer
def __a ( self :Optional[int]) -> str:
shutil.rmtree(self.tmpdirname)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta)]
UpperCAmelCase_ = [Image.fromarray(np.moveaxis(_lowercase , 0 , -1)) for x in image_inputs]
return image_inputs
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() , qformer_tokenizer=self.get_qformer_tokenizer() , )
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = self.get_image_processor(do_normalize=_lowercase , padding_value=1.0)
UpperCAmelCase_ = InstructBlipProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowercase , padding_value=1.0)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
self.assertIsInstance(processor.tokenizer , _lowercase)
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string())
self.assertIsInstance(processor.image_processor , _lowercase)
self.assertIsInstance(processor.qformer_tokenizer , _lowercase)
def __a ( self :Dict) -> Any:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = image_processor(_lowercase , return_tensors='''np''')
UpperCAmelCase_ = processor(images=_lowercase , return_tensors='''np''')
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1E-2)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = processor(text=_lowercase)
UpperCAmelCase_ = tokenizer(_lowercase , return_token_type_ids=_lowercase)
UpperCAmelCase_ = qformer_tokenizer(_lowercase , return_token_type_ids=_lowercase)
for key in encoded_tokens.keys():
self.assertListEqual(encoded_tokens[key] , encoded_processor[key])
for key in encoded_tokens_qformer.keys():
self.assertListEqual(encoded_tokens_qformer[key] , encoded_processor['''qformer_''' + key])
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
# test if it raises when no input is passed
with pytest.raises(_lowercase):
processor()
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
UpperCAmelCase_ = processor.batch_decode(_lowercase)
UpperCAmelCase_ = tokenizer.batch_decode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :str) -> int:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
| 344 | 1 |
import os
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/check_doctest_list.py
UpperCamelCase_ = "."
if __name__ == "__main__":
UpperCamelCase_ = os.path.join(REPO_PATH, "utils/documentation_tests.txt")
UpperCamelCase_ = []
UpperCamelCase_ = []
with open(doctest_file_path) as fp:
for line in fp:
UpperCamelCase_ = line.strip()
UpperCamelCase_ = os.path.join(REPO_PATH, line)
if not (os.path.isfile(path) or os.path.isdir(path)):
non_existent_paths.append(line)
all_paths.append(path)
if len(non_existent_paths) > 0:
UpperCamelCase_ = "\n".join(non_existent_paths)
raise ValueError(f"`utils/documentation_tests.txt` contains non-existent paths:\n{non_existent_paths}")
if all_paths != sorted(all_paths):
raise ValueError("Files in `utils/documentation_tests.txt` are not in alphabetical order.")
| 344 |
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"facebook/levit-128S": "https://huggingface.co/facebook/levit-128S/resolve/main/config.json",
# See all LeViT models at https://huggingface.co/models?filter=levit
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[int] ="levit"
def __init__( self :List[str] , _lowercase :List[Any]=224 , _lowercase :str=3 , _lowercase :Optional[int]=3 , _lowercase :str=2 , _lowercase :List[Any]=1 , _lowercase :str=16 , _lowercase :Dict=[128, 256, 384] , _lowercase :Union[str, Any]=[4, 8, 12] , _lowercase :Tuple=[4, 4, 4] , _lowercase :Dict=[16, 16, 16] , _lowercase :Any=0 , _lowercase :Dict=[2, 2, 2] , _lowercase :Any=[2, 2, 2] , _lowercase :Tuple=0.02 , **_lowercase :Union[str, Any] , ) -> Optional[Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = kernel_size
UpperCAmelCase_ = stride
UpperCAmelCase_ = padding
UpperCAmelCase_ = hidden_sizes
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = depths
UpperCAmelCase_ = key_dim
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = attention_ratio
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = [
['''Subsample''', key_dim[0], hidden_sizes[0] // key_dim[0], 4, 2, 2],
['''Subsample''', key_dim[0], hidden_sizes[1] // key_dim[0], 4, 2, 2],
]
class a_ ( _snake_case ):
UpperCamelCase__ : Union[str, Any] =version.parse("1.11" )
@property
def __a ( self :Any) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
])
@property
def __a ( self :List[Any]) -> float:
return 1E-4
| 344 | 1 |
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def A ( __UpperCAmelCase ) -> Dict[str, torch.Tensor]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for rt in rc.restypes:
UpperCAmelCase_ = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
UpperCAmelCase_ = {name: i for i, name in enumerate(__UpperCAmelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.floataa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = protein['''aatype'''].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the corresponding mask
UpperCAmelCase_ = torch.zeros([21, 37] , dtype=torch.floataa , device=protein['''aatype'''].device )
for restype, restype_letter in enumerate(rc.restypes ):
UpperCAmelCase_ = rc.restype_atoa[restype_letter]
UpperCAmelCase_ = rc.residue_atoms[restype_name]
for atom_name in atom_names:
UpperCAmelCase_ = rc.atom_order[atom_name]
UpperCAmelCase_ = 1
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
return protein
def A ( __UpperCAmelCase ) -> Dict[str, np.ndarray]:
'''simple docstring'''
UpperCAmelCase_ = tree_map(lambda __UpperCAmelCase : torch.tensor(__UpperCAmelCase , device=batch['''aatype'''].device ) , __UpperCAmelCase , np.ndarray )
UpperCAmelCase_ = tensor_tree_map(lambda __UpperCAmelCase : np.array(__UpperCAmelCase ) , make_atomaa_masks(__UpperCAmelCase ) )
return out
| 344 |
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=None , __UpperCAmelCase="no" , __UpperCAmelCase="29500" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = False
UpperCAmelCase_ = False
if any(key.startswith('''KAGGLE''' ) for key in os.environ.keys() ):
UpperCAmelCase_ = True
elif "IPython" in sys.modules:
UpperCAmelCase_ = '''google.colab''' in str(sys.modules['''IPython'''].get_ipython() )
try:
UpperCAmelCase_ = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
f"Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}." )
if (in_colab or in_kaggle) and (os.environ.get('''TPU_NAME''' , __UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '''
'''your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if num_processes is None:
UpperCAmelCase_ = 8
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''TPU''' )
print(f"Launching a training on {num_processes} TPU cores." )
xmp.spawn(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on one CPU.''' )
function(*__UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'''You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.''' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '''
'''inside your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if torch.cuda.is_initialized():
raise ValueError(
'''To launch a multi-GPU training from your notebook, you need to avoid running any instruction '''
'''using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '''
'''function.''' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port=__UpperCAmelCase , mixed_precision=__UpperCAmelCase ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''MULTI_GPU''' )
print(f"Launching training on {num_processes} GPUs." )
try:
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'''CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '''
'''This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '''
'''Please review your imports and test them when running the `notebook_launcher()` to identify '''
'''which one is problematic.''' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
UpperCAmelCase_ = '''1'''
print('''Launching training on MPS.''' )
elif torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on CPU.''' )
function(*__UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=2 ) -> Optional[Any]:
'''simple docstring'''
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port='''29500''' , accelerate_mixed_precision='''no''' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='''yes''' , ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , debug=__UpperCAmelCase )
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
| 344 | 1 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
UpperCamelCase_ = {"configuration_yolos": ["YOLOS_PRETRAINED_CONFIG_ARCHIVE_MAP", "YolosConfig", "YolosOnnxConfig"]}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["YolosFeatureExtractor"]
UpperCamelCase_ = ["YolosImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"YOLOS_PRETRAINED_MODEL_ARCHIVE_LIST",
"YolosForObjectDetection",
"YolosModel",
"YolosPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_yolos import YOLOS_PRETRAINED_CONFIG_ARCHIVE_MAP, YolosConfig, YolosOnnxConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_yolos import YolosFeatureExtractor
from .image_processing_yolos import YolosImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_yolos import (
YOLOS_PRETRAINED_MODEL_ARCHIVE_LIST,
YolosForObjectDetection,
YolosModel,
YolosPreTrainedModel,
)
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
import math
import os
import re
import sys
import unittest
from pathlib import Path
from typing import Tuple
from unittest.mock import patch
from parameterized import parameterized
from transformers.testing_utils import (
CaptureStderr,
ExtendSysPath,
TestCasePlus,
execute_subprocess_async,
get_gpu_count,
get_torch_dist_unique_port,
require_apex,
require_bitsandbytes,
require_fairscale,
require_torch,
require_torch_gpu,
require_torch_multi_gpu,
require_torch_non_multi_gpu,
slow,
)
from transformers.trainer_callback import TrainerState
from transformers.trainer_utils import set_seed
UpperCamelCase_ = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f"{bindir}/../../examples/pytorch/translation"):
from run_translation import main # noqa
set_seed(42)
UpperCamelCase_ = "sshleifer/student_marian_en_ro_6_1"
UpperCamelCase_ = "sshleifer/tiny-mbart"
@require_torch
class a_ ( _snake_case ):
def __a ( self :str , _lowercase :Any=False , _lowercase :Tuple=None , _lowercase :Dict=True , _lowercase :Tuple=True , _lowercase :List[Any]=True , _lowercase :List[str]=True , ) -> int:
UpperCAmelCase_ = self.run_trainer(
eval_steps=1 , max_len=12 , model_name=_lowercase , num_train_epochs=1 , distributed=_lowercase , extra_args_str=_lowercase , predict_with_generate=_lowercase , do_train=_lowercase , do_eval=_lowercase , do_predict=_lowercase , )
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
if not do_eval:
return
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
if predict_with_generate:
assert "eval_bleu" in first_step_stats
UpperCAmelCase_ = eval_metrics[-1]
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
assert not math.isnan(float(last_step_stats['''eval_loss'''])), "eval_loss must not be `nan`"
@require_torch_non_multi_gpu
def __a ( self :Dict) -> str:
self.run_seqaseq_quick()
@require_torch_multi_gpu
def __a ( self :Any) -> int:
self.run_seqaseq_quick(distributed=_lowercase)
@require_torch_multi_gpu
def __a ( self :int) -> Any:
self.run_seqaseq_quick(distributed=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> List[str]:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple --fp16''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Union[str, Any]) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :int) -> Any:
self.run_seqaseq_quick(
distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=_lowercase)
@require_apex
@require_torch_gpu
def __a ( self :Tuple) -> str:
# XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same
# program and it breaks other tests that run from the same pytest worker, therefore until this is
# sorted out it must be run only in an external program, that is distributed=True in this
# test and only under one or more gpus - if we want cpu will need to make a special test
#
# specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via
# 2nd main() call it botches the future eval.
#
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
# test 2nd time - was getting eval_loss': nan'
# to reproduce the problem set distributed=False
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
@parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''])
@require_torch_multi_gpu
def __a ( self :str , _lowercase :Any) -> List[str]:
# as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout
UpperCAmelCase_ = {
# test with the default log_level - should be info and thus log info once
'''base''': {'''extra_args_str''': '''''', '''n_matches''': 1},
# test with low log_level and log_level_replica - should be noisy on all processes
# now the info string should appear twice on 2 processes
'''low''': {'''extra_args_str''': '''--log_level debug --log_level_replica debug''', '''n_matches''': 2},
# test with high log_level and low log_level_replica
# now the info string should appear once only on the replica
'''high''': {'''extra_args_str''': '''--log_level error --log_level_replica debug''', '''n_matches''': 1},
# test with high log_level and log_level_replica - should be quiet on all processes
'''mixed''': {'''extra_args_str''': '''--log_level error --log_level_replica error''', '''n_matches''': 0},
}
UpperCAmelCase_ = experiments[experiment_id]
UpperCAmelCase_ = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False}
UpperCAmelCase_ = '''Running training'''
with CaptureStderr() as cl:
self.run_seqaseq_quick(**_lowercase , extra_args_str=data['''extra_args_str'''])
UpperCAmelCase_ = len(re.findall(_lowercase , cl.err))
self.assertEqual(_lowercase , data['''n_matches'''])
@slow
def __a ( self :Any) -> Dict:
UpperCAmelCase_ = self.run_trainer(
eval_steps=2 , max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=10 , distributed=_lowercase , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
UpperCAmelCase_ = eval_metrics[-1]
assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing"
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
# test if do_predict saves generations and metrics
UpperCAmelCase_ = os.listdir(_lowercase)
UpperCAmelCase_ = {os.path.basename(_lowercase) for p in contents}
assert "generated_predictions.txt" in contents
assert "predict_results.json" in contents
@slow
@require_bitsandbytes
def __a ( self :List[str]) -> str:
from transformers.training_args import OptimizerNames
def train_and_return_metrics(_lowercase :str) -> Tuple[int, float]:
UpperCAmelCase_ = '''--skip_memory_metrics 0'''
UpperCAmelCase_ = self.run_trainer(
max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=1 , optim=_lowercase , distributed=_lowercase , extra_args_str=_lowercase , do_eval=_lowercase , do_predict=_lowercase , n_gpus_to_use=1 , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(Path(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20)
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20)
UpperCAmelCase_ = logs[0]['''train_loss''']
return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value)
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value)
UpperCAmelCase_ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_peak_mem_orig + gpu_alloc_mem_orig
UpperCAmelCase_ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_total_mem_orig - gpu_total_mem_bnb
# sshleifer/student_marian_en_ro_6_1 has 54M parameter, 29M of which is `nn.Embedding` which
# doesn't get quantized and remains in fp32. Therefore we only have 25M parameters quantized
# in 2 bytes and the diff in optim memory usage is derived as so:
#
# - normal 25*8=~200MB (8 bytes per param)
# - bnb 25*2= ~50MB (2 bytes per param)
#
# Thus we should expect ~150MB total memory saved.
#
# Peak memory should be the same - the total should be different by about that same margin
#
# After leaving a small margin to accommodate for differences between gpus let's check
# that we have at least 120MB in savings
UpperCAmelCase_ = 120
# uncomment the following if this test starts failing - requires py38 for a new print feature
# gpu_peak_mem_diff = gpu_peak_mem_orig - gpu_peak_mem_bnb
# print(f"{gpu_alloc_mem_orig=}MB {gpu_peak_mem_orig=}MB {gpu_alloc_mem_orig+gpu_peak_mem_orig=}MB")
# print(f" {gpu_alloc_mem_bnb=}MB {gpu_peak_mem_bnb=}MB {gpu_alloc_mem_bnb+gpu_peak_mem_bnb=}MB")
# print(f"{gpu_alloc_mem_diff=}MB")
# print(f"{gpu_peak_mem_diff=}MB")
# print(f"{gpu_total_mem_orig=}MB, {gpu_total_mem_bnb=}MB")
# print(f"{gpu_total_mem_diff=}MB, {gpu_total_mem_diff=}MB")
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less alloc gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_alloc_mem_diff}MB, with gpu_alloc_mem_orig={gpu_alloc_mem_orig}MB and"
f" gpu_alloc_mem_bnb={gpu_alloc_mem_bnb}MB" , )
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less total gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_total_mem_diff}MB, with gpu_total_mem_orig={gpu_total_mem_orig}MB and"
f" gpu_total_mem_bnb={gpu_total_mem_bnb}MB" , )
self.assertEqual(
_lowercase , _lowercase , f"loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}")
def __a ( self :Any , _lowercase :int , _lowercase :str , _lowercase :int , _lowercase :float = 3E-3 , _lowercase :str = "adafactor" , _lowercase :bool = False , _lowercase :str = None , _lowercase :int = 0 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :int = None , ) -> List[Any]:
UpperCAmelCase_ = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro'''
UpperCAmelCase_ = self.get_auto_remove_tmp_dir()
UpperCAmelCase_ = f"\n --model_name_or_path {model_name}\n --train_file {data_dir}/train.json\n --validation_file {data_dir}/val.json\n --test_file {data_dir}/test.json\n --output_dir {output_dir}\n --overwrite_output_dir\n --max_train_samples 8\n --max_source_length {max_len}\n --max_target_length {max_len}\n --do_train\n --num_train_epochs {str(_lowercase)}\n --per_device_train_batch_size 4\n --learning_rate {learning_rate}\n --warmup_steps 8\n --logging_steps 0\n --logging_strategy no\n --save_steps {str(_lowercase)}\n --group_by_length\n --label_smoothing_factor 0.1\n --target_lang ro_RO\n --source_lang en_XX\n ".split()
UpperCAmelCase_ = f"\n --do_eval\n --per_device_eval_batch_size 4\n --max_eval_samples 8\n --val_max_target_length {max_len}\n --evaluation_strategy steps\n --eval_steps {str(_lowercase)}\n ".split()
UpperCAmelCase_ = '''
--do_predict
'''.split()
UpperCAmelCase_ = []
if do_train:
args += args_train
if do_eval:
args += args_eval
if do_predict:
args += args_predict
if predict_with_generate:
args += "--predict_with_generate".split()
if do_train:
if optim == "adafactor":
args += "--adafactor".split()
else:
args += f"--optim {optim}".split()
if extra_args_str is not None:
args += extra_args_str.split()
if distributed:
if n_gpus_to_use is None:
UpperCAmelCase_ = get_gpu_count()
UpperCAmelCase_ = get_torch_dist_unique_port()
UpperCAmelCase_ = f"\n -m torch.distributed.run\n --nproc_per_node={n_gpus_to_use}\n --master_port={master_port}\n {self.examples_dir_str}/pytorch/translation/run_translation.py\n ".split()
UpperCAmelCase_ = [sys.executable] + distributed_args + args
# keep for quick debug
# print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die
execute_subprocess_async(_lowercase , env=self.get_env())
else:
UpperCAmelCase_ = ['''run_translation.py'''] + args
with patch.object(_lowercase , '''argv''' , _lowercase):
main()
return output_dir
| 344 | 1 |
import datasets
from .evaluate import evaluate
UpperCamelCase_ = "\\n@inproceedings{Rajpurkar2016SQuAD10,\n title={SQuAD: 100, 000+ Questions for Machine Comprehension of Text},\n author={Pranav Rajpurkar and Jian Zhang and Konstantin Lopyrev and Percy Liang},\n booktitle={EMNLP},\n year={2016}\n}\n"
UpperCamelCase_ = "\nThis metric wrap the official scoring script for version 1 of the Stanford Question Answering Dataset (SQuAD).\n\nStanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by\ncrowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span,\nfrom the corresponding reading passage, or the question might be unanswerable.\n"
UpperCamelCase_ = "\nComputes SQuAD scores (F1 and EM).\nArgs:\n predictions: List of question-answers dictionaries with the following key-values:\n - 'id': id of the question-answer pair as given in the references (see below)\n - 'prediction_text': the text of the answer\n references: List of question-answers dictionaries with the following key-values:\n - 'id': id of the question-answer pair (see above),\n - 'answers': a Dict in the SQuAD dataset format\n {\n 'text': list of possible texts for the answer, as a list of strings\n 'answer_start': list of start positions for the answer, as a list of ints\n }\n Note that answer_start values are not taken into account to compute the metric.\nReturns:\n 'exact_match': Exact match (the normalized answer exactly match the gold answer)\n 'f1': The F-score of predicted tokens versus the gold answer\nExamples:\n\n >>> predictions = [{'prediction_text': '1976', 'id': '56e10a3be3433e1400422b22'}]\n >>> references = [{'answers': {'answer_start': [97], 'text': ['1976']}, 'id': '56e10a3be3433e1400422b22'}]\n >>> squad_metric = datasets.load_metric(\"squad\")\n >>> results = squad_metric.compute(predictions=predictions, references=references)\n >>> print(results)\n {'exact_match': 100.0, 'f1': 100.0}\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class a_ ( datasets.Metric ):
def __a ( self :Optional[int]) -> Optional[int]:
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
'''predictions''': {'''id''': datasets.Value('''string'''), '''prediction_text''': datasets.Value('''string''')},
'''references''': {
'''id''': datasets.Value('''string'''),
'''answers''': datasets.features.Sequence(
{
'''text''': datasets.Value('''string'''),
'''answer_start''': datasets.Value('''int32'''),
}),
},
}) , codebase_urls=['''https://rajpurkar.github.io/SQuAD-explorer/'''] , reference_urls=['''https://rajpurkar.github.io/SQuAD-explorer/'''] , )
def __a ( self :str , _lowercase :Dict , _lowercase :Optional[int]) -> Dict:
UpperCAmelCase_ = {prediction['''id''']: prediction['''prediction_text'''] for prediction in predictions}
UpperCAmelCase_ = [
{
'''paragraphs''': [
{
'''qas''': [
{
'''answers''': [{'''text''': answer_text} for answer_text in ref['''answers''']['''text''']],
'''id''': ref['''id'''],
}
for ref in references
]
}
]
}
]
UpperCAmelCase_ = evaluate(dataset=_lowercase , predictions=_lowercase)
return score
| 344 |
import functools
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = len(__UpperCAmelCase )
UpperCAmelCase_ = len(__UpperCAmelCase )
@functools.cache
def min_distance(__UpperCAmelCase , __UpperCAmelCase ) -> int:
# if first word index is overflow - delete all from the second word
if indexa >= len_worda:
return len_worda - indexa
# if second word index is overflow - delete all from the first word
if indexa >= len_worda:
return len_worda - indexa
UpperCAmelCase_ = int(worda[indexa] != worda[indexa] ) # current letters not identical
return min(
1 + min_distance(indexa + 1 , __UpperCAmelCase ) , 1 + min_distance(__UpperCAmelCase , indexa + 1 ) , diff + min_distance(indexa + 1 , indexa + 1 ) , )
return min_distance(0 , 0 )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 | 1 |
import unittest
from transformers import AlbertConfig, is_torch_available
from transformers.models.auto import get_values
from transformers.testing_utils import require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
MODEL_FOR_PRETRAINING_MAPPING,
AlbertForMaskedLM,
AlbertForMultipleChoice,
AlbertForPreTraining,
AlbertForQuestionAnswering,
AlbertForSequenceClassification,
AlbertForTokenClassification,
AlbertModel,
)
from transformers.models.albert.modeling_albert import ALBERT_PRETRAINED_MODEL_ARCHIVE_LIST
class a_ :
def __init__( self :Tuple , _lowercase :Dict , _lowercase :Union[str, Any]=13 , _lowercase :Optional[Any]=7 , _lowercase :List[Any]=True , _lowercase :Dict=True , _lowercase :Optional[Any]=True , _lowercase :str=True , _lowercase :str=99 , _lowercase :Optional[Any]=16 , _lowercase :List[Any]=36 , _lowercase :Dict=6 , _lowercase :Tuple=6 , _lowercase :Optional[int]=6 , _lowercase :int=37 , _lowercase :str="gelu" , _lowercase :List[Any]=0.1 , _lowercase :int=0.1 , _lowercase :Any=512 , _lowercase :str=16 , _lowercase :Optional[Any]=2 , _lowercase :str=0.02 , _lowercase :Tuple=3 , _lowercase :List[Any]=4 , _lowercase :str=None , ) -> List[str]:
UpperCAmelCase_ = parent
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = seq_length
UpperCAmelCase_ = is_training
UpperCAmelCase_ = use_input_mask
UpperCAmelCase_ = use_token_type_ids
UpperCAmelCase_ = use_labels
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = embedding_size
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_hidden_groups
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = type_vocab_size
UpperCAmelCase_ = type_sequence_label_size
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = num_labels
UpperCAmelCase_ = num_choices
UpperCAmelCase_ = scope
def __a ( self :int) -> Optional[Any]:
UpperCAmelCase_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size)
UpperCAmelCase_ = None
if self.use_input_mask:
UpperCAmelCase_ = random_attention_mask([self.batch_size, self.seq_length])
UpperCAmelCase_ = None
if self.use_token_type_ids:
UpperCAmelCase_ = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size)
UpperCAmelCase_ = None
UpperCAmelCase_ = None
UpperCAmelCase_ = None
if self.use_labels:
UpperCAmelCase_ = ids_tensor([self.batch_size] , self.type_sequence_label_size)
UpperCAmelCase_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels)
UpperCAmelCase_ = ids_tensor([self.batch_size] , self.num_choices)
UpperCAmelCase_ = self.get_config()
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def __a ( self :int) -> Optional[Any]:
return AlbertConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , initializer_range=self.initializer_range , num_hidden_groups=self.num_hidden_groups , )
def __a ( self :Optional[int] , _lowercase :Tuple , _lowercase :List[Any] , _lowercase :List[Any] , _lowercase :Tuple , _lowercase :Optional[int] , _lowercase :Dict , _lowercase :List[Any]) -> List[Any]:
UpperCAmelCase_ = AlbertModel(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase)
UpperCAmelCase_ = model(_lowercase , token_type_ids=_lowercase)
UpperCAmelCase_ = model(_lowercase)
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size))
self.parent.assertEqual(result.pooler_output.shape , (self.batch_size, self.hidden_size))
def __a ( self :Any , _lowercase :str , _lowercase :str , _lowercase :str , _lowercase :List[str] , _lowercase :str , _lowercase :Any , _lowercase :Optional[int]) -> Any:
UpperCAmelCase_ = AlbertForPreTraining(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(
_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase , labels=_lowercase , sentence_order_label=_lowercase , )
self.parent.assertEqual(result.prediction_logits.shape , (self.batch_size, self.seq_length, self.vocab_size))
self.parent.assertEqual(result.sop_logits.shape , (self.batch_size, config.num_labels))
def __a ( self :Union[str, Any] , _lowercase :List[Any] , _lowercase :Union[str, Any] , _lowercase :str , _lowercase :Dict , _lowercase :int , _lowercase :Tuple , _lowercase :Dict) -> Any:
UpperCAmelCase_ = AlbertForMaskedLM(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase , labels=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size))
def __a ( self :Dict , _lowercase :Union[str, Any] , _lowercase :Dict , _lowercase :List[Any] , _lowercase :List[str] , _lowercase :int , _lowercase :int , _lowercase :str) -> str:
UpperCAmelCase_ = AlbertForQuestionAnswering(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(
_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase , start_positions=_lowercase , end_positions=_lowercase , )
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length))
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length))
def __a ( self :Tuple , _lowercase :List[str] , _lowercase :Any , _lowercase :Dict , _lowercase :List[Any] , _lowercase :Optional[Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
UpperCAmelCase_ = self.num_labels
UpperCAmelCase_ = AlbertForSequenceClassification(_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase , labels=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels))
def __a ( self :Union[str, Any] , _lowercase :Optional[int] , _lowercase :str , _lowercase :Optional[int] , _lowercase :int , _lowercase :List[Any] , _lowercase :List[Any] , _lowercase :List[str]) -> Any:
UpperCAmelCase_ = self.num_labels
UpperCAmelCase_ = AlbertForTokenClassification(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase , labels=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels))
def __a ( self :Union[str, Any] , _lowercase :Optional[Any] , _lowercase :Tuple , _lowercase :List[str] , _lowercase :Union[str, Any] , _lowercase :List[str] , _lowercase :int , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = self.num_choices
UpperCAmelCase_ = AlbertForMultipleChoice(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = input_ids.unsqueeze(1).expand(-1 , self.num_choices , -1).contiguous()
UpperCAmelCase_ = token_type_ids.unsqueeze(1).expand(-1 , self.num_choices , -1).contiguous()
UpperCAmelCase_ = input_mask.unsqueeze(1).expand(-1 , self.num_choices , -1).contiguous()
UpperCAmelCase_ = model(
_lowercase , attention_mask=_lowercase , token_type_ids=_lowercase , labels=_lowercase , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices))
def __a ( self :Any) -> Tuple:
UpperCAmelCase_ = self.prepare_config_and_inputs()
(
(
UpperCAmelCase_
) , (
UpperCAmelCase_
) , (
UpperCAmelCase_
) , (
UpperCAmelCase_
) , (
UpperCAmelCase_
) , (
UpperCAmelCase_
) , (
UpperCAmelCase_
) ,
) = config_and_inputs
UpperCAmelCase_ = {'''input_ids''': input_ids, '''token_type_ids''': token_type_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a_ ( _snake_case , _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =(
(
AlbertModel,
AlbertForPreTraining,
AlbertForMaskedLM,
AlbertForMultipleChoice,
AlbertForSequenceClassification,
AlbertForTokenClassification,
AlbertForQuestionAnswering,
)
if is_torch_available()
else ()
)
UpperCamelCase__ : Tuple =(
{
"feature-extraction": AlbertModel,
"fill-mask": AlbertForMaskedLM,
"question-answering": AlbertForQuestionAnswering,
"text-classification": AlbertForSequenceClassification,
"token-classification": AlbertForTokenClassification,
"zero-shot": AlbertForSequenceClassification,
}
if is_torch_available()
else {}
)
UpperCamelCase__ : List[str] =True
def __a ( self :Optional[int] , _lowercase :str , _lowercase :Tuple , _lowercase :Dict=False) -> Dict:
UpperCAmelCase_ = super()._prepare_for_class(_lowercase , _lowercase , return_labels=_lowercase)
if return_labels:
if model_class in get_values(_lowercase):
UpperCAmelCase_ = torch.zeros(
(self.model_tester.batch_size, self.model_tester.seq_length) , dtype=torch.long , device=_lowercase)
UpperCAmelCase_ = torch.zeros(
self.model_tester.batch_size , dtype=torch.long , device=_lowercase)
return inputs_dict
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = AlbertModelTester(self)
UpperCAmelCase_ = ConfigTester(self , config_class=_lowercase , hidden_size=37)
def __a ( self :int) -> Optional[Any]:
self.config_tester.run_common_tests()
def __a ( self :Optional[Any]) -> Optional[Any]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*_lowercase)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_pretraining(*_lowercase)
def __a ( self :Dict) -> List[str]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*_lowercase)
def __a ( self :Optional[int]) -> Tuple:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_multiple_choice(*_lowercase)
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*_lowercase)
def __a ( self :List[str]) -> Union[str, Any]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*_lowercase)
def __a ( self :Optional[int]) -> int:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
for type in ["absolute", "relative_key", "relative_key_query"]:
UpperCAmelCase_ = type
self.model_tester.create_and_check_model(*_lowercase)
@slow
def __a ( self :Optional[Any]) -> Dict:
for model_name in ALBERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
UpperCAmelCase_ = AlbertModel.from_pretrained(_lowercase)
self.assertIsNotNone(_lowercase)
@require_torch
class a_ ( unittest.TestCase ):
@slow
def __a ( self :str) -> List[str]:
UpperCAmelCase_ = AlbertModel.from_pretrained('''albert-base-v2''')
UpperCAmelCase_ = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]])
UpperCAmelCase_ = torch.tensor([[0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])
with torch.no_grad():
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase)[0]
UpperCAmelCase_ = torch.Size((1, 11, 768))
self.assertEqual(output.shape , _lowercase)
UpperCAmelCase_ = torch.tensor(
[[[-0.6_513, 1.5_035, -0.2_766], [-0.6_515, 1.5_046, -0.2_780], [-0.6_512, 1.5_049, -0.2_784]]])
self.assertTrue(torch.allclose(output[:, 1:4, 1:4] , _lowercase , atol=1E-4))
| 344 |
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import SPIECE_UNDERLINE, logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {"vocab_file": "spiece.model"}
UpperCamelCase_ = {
"vocab_file": {
"xlnet-base-cased": "https://huggingface.co/xlnet-base-cased/resolve/main/spiece.model",
"xlnet-large-cased": "https://huggingface.co/xlnet-large-cased/resolve/main/spiece.model",
}
}
UpperCamelCase_ = {
"xlnet-base-cased": None,
"xlnet-large-cased": None,
}
# Segments (not really needed)
UpperCamelCase_ = 0
UpperCamelCase_ = 1
UpperCamelCase_ = 2
UpperCamelCase_ = 3
UpperCamelCase_ = 4
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =VOCAB_FILES_NAMES
UpperCamelCase__ : Optional[Any] =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : Tuple =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Any ="left"
def __init__( self :Optional[int] , _lowercase :Union[str, Any] , _lowercase :Union[str, Any]=False , _lowercase :Optional[int]=True , _lowercase :Union[str, Any]=False , _lowercase :Tuple="<s>" , _lowercase :Any="</s>" , _lowercase :Dict="<unk>" , _lowercase :str="<sep>" , _lowercase :Tuple="<pad>" , _lowercase :Any="<cls>" , _lowercase :List[str]="<mask>" , _lowercase :Union[str, Any]=["<eop>", "<eod>"] , _lowercase :Optional[Dict[str, Any]] = None , **_lowercase :Union[str, Any] , ) -> None:
# Mask token behave like a normal word, i.e. include the space before it
UpperCAmelCase_ = AddedToken(_lowercase , lstrip=_lowercase , rstrip=_lowercase) if isinstance(_lowercase , _lowercase) else mask_token
UpperCAmelCase_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_lowercase , remove_space=_lowercase , keep_accents=_lowercase , bos_token=_lowercase , eos_token=_lowercase , unk_token=_lowercase , sep_token=_lowercase , pad_token=_lowercase , cls_token=_lowercase , mask_token=_lowercase , additional_special_tokens=_lowercase , sp_model_kwargs=self.sp_model_kwargs , **_lowercase , )
UpperCAmelCase_ = 3
UpperCAmelCase_ = do_lower_case
UpperCAmelCase_ = remove_space
UpperCAmelCase_ = keep_accents
UpperCAmelCase_ = vocab_file
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(_lowercase)
@property
def __a ( self :int) -> List[Any]:
return len(self.sp_model)
def __a ( self :Optional[int]) -> List[Any]:
UpperCAmelCase_ = {self.convert_ids_to_tokens(_lowercase): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def __getstate__( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.__dict__.copy()
UpperCAmelCase_ = None
return state
def __setstate__( self :Optional[Any] , _lowercase :Optional[Any]) -> List[Any]:
UpperCAmelCase_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs'''):
UpperCAmelCase_ = {}
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(self.vocab_file)
def __a ( self :List[str] , _lowercase :Tuple) -> Optional[int]:
if self.remove_space:
UpperCAmelCase_ = ''' '''.join(inputs.strip().split())
else:
UpperCAmelCase_ = inputs
UpperCAmelCase_ = outputs.replace('''``''' , '''"''').replace('''\'\'''' , '''"''')
if not self.keep_accents:
UpperCAmelCase_ = unicodedata.normalize('''NFKD''' , _lowercase)
UpperCAmelCase_ = ''''''.join([c for c in outputs if not unicodedata.combining(_lowercase)])
if self.do_lower_case:
UpperCAmelCase_ = outputs.lower()
return outputs
def __a ( self :str , _lowercase :str) -> List[str]:
UpperCAmelCase_ = self.preprocess_text(_lowercase)
UpperCAmelCase_ = self.sp_model.encode(_lowercase , out_type=_lowercase)
UpperCAmelCase_ = []
for piece in pieces:
if len(_lowercase) > 1 and piece[-1] == str(''',''') and piece[-2].isdigit():
UpperCAmelCase_ = self.sp_model.EncodeAsPieces(piece[:-1].replace(_lowercase , ''''''))
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0]) == 1:
UpperCAmelCase_ = cur_pieces[1:]
else:
UpperCAmelCase_ = cur_pieces[0][1:]
cur_pieces.append(piece[-1])
new_pieces.extend(_lowercase)
else:
new_pieces.append(_lowercase)
return new_pieces
def __a ( self :Optional[Any] , _lowercase :Union[str, Any]) -> Tuple:
return self.sp_model.PieceToId(_lowercase)
def __a ( self :Optional[int] , _lowercase :Optional[Any]) -> List[str]:
return self.sp_model.IdToPiece(_lowercase)
def __a ( self :List[Any] , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = ''''''.join(_lowercase).replace(_lowercase , ''' ''').strip()
return out_string
def __a ( self :Union[str, Any] , _lowercase :List[int] , _lowercase :bool = False , _lowercase :bool = None , _lowercase :bool = True , **_lowercase :Tuple , ) -> str:
UpperCAmelCase_ = kwargs.pop('''use_source_tokenizer''' , _lowercase)
UpperCAmelCase_ = self.convert_ids_to_tokens(_lowercase , skip_special_tokens=_lowercase)
# To avoid mixing byte-level and unicode for byte-level BPT
# we need to build string separately for added tokens and byte-level tokens
# cf. https://github.com/huggingface/transformers/issues/1133
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for token in filtered_tokens:
if skip_special_tokens and token in self.all_special_ids:
continue
if token in self.added_tokens_encoder:
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
UpperCAmelCase_ = []
sub_texts.append(_lowercase)
else:
current_sub_text.append(_lowercase)
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
# Mimic the behavior of the Rust tokenizer:
# By default, there are no spaces between special tokens
UpperCAmelCase_ = ''''''.join(_lowercase)
UpperCAmelCase_ = (
clean_up_tokenization_spaces
if clean_up_tokenization_spaces is not None
else self.clean_up_tokenization_spaces
)
if clean_up_tokenization_spaces:
UpperCAmelCase_ = self.clean_up_tokenization(_lowercase)
return clean_text
else:
return text
def __a ( self :str , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [self.cls_token_id]
if token_ids_a is None:
return token_ids_a + sep + cls
return token_ids_a + sep + token_ids_a + sep + cls
def __a ( self :Dict , _lowercase :List[int] , _lowercase :Optional[List[int]] = None , _lowercase :bool = False) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowercase , token_ids_a=_lowercase , already_has_special_tokens=_lowercase)
if token_ids_a is not None:
return ([0] * len(_lowercase)) + [1] + ([0] * len(_lowercase)) + [1, 1]
return ([0] * len(_lowercase)) + [1, 1]
def __a ( self :Optional[int] , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [2]
if token_ids_a is None:
return len(token_ids_a + sep) * [0] + cls_segment_id
return len(token_ids_a + sep) * [0] + len(token_ids_a + sep) * [1] + cls_segment_id
def __a ( self :str , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
if not os.path.isdir(_lowercase):
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
return
UpperCAmelCase_ = os.path.join(
_lowercase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''])
if os.path.abspath(self.vocab_file) != os.path.abspath(_lowercase) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , _lowercase)
elif not os.path.isfile(self.vocab_file):
with open(_lowercase , '''wb''') as fi:
UpperCAmelCase_ = self.sp_model.serialized_model_proto()
fi.write(_lowercase)
return (out_vocab_file,)
| 344 | 1 |
from maths.prime_check import is_prime
def A ( __UpperCAmelCase ) -> int:
'''simple docstring'''
if not isinstance(__UpperCAmelCase , __UpperCAmelCase ):
UpperCAmelCase_ = f"Input value of [number={number}] must be an integer"
raise TypeError(__UpperCAmelCase )
if is_prime(__UpperCAmelCase ) and is_prime(number + 2 ):
return number + 2
else:
return -1
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
UpperCamelCase_ = logging.get_logger(__name__)
class a_ ( _snake_case , _snake_case ):
UpperCamelCase__ : Union[str, Any] ="maskformer-swin"
UpperCamelCase__ : List[str] ={
"num_attention_heads": "num_heads",
"num_hidden_layers": "num_layers",
}
def __init__( self :Union[str, Any] , _lowercase :Optional[int]=224 , _lowercase :List[str]=4 , _lowercase :Tuple=3 , _lowercase :List[Any]=96 , _lowercase :Any=[2, 2, 6, 2] , _lowercase :int=[3, 6, 12, 24] , _lowercase :List[Any]=7 , _lowercase :Dict=4.0 , _lowercase :Any=True , _lowercase :int=0.0 , _lowercase :List[Any]=0.0 , _lowercase :Tuple=0.1 , _lowercase :str="gelu" , _lowercase :Union[str, Any]=False , _lowercase :Tuple=0.02 , _lowercase :List[str]=1E-5 , _lowercase :List[str]=None , _lowercase :Any=None , **_lowercase :Any , ) -> Union[str, Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = embed_dim
UpperCAmelCase_ = depths
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = num_heads
UpperCAmelCase_ = window_size
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = qkv_bias
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = use_absolute_embeddings
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = initializer_range
# we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
UpperCAmelCase_ = int(embed_dim * 2 ** (len(_lowercase) - 1))
UpperCAmelCase_ = ['''stem'''] + [f"stage{idx}" for idx in range(1 , len(_lowercase) + 1)]
UpperCAmelCase_ , UpperCAmelCase_ = get_aligned_output_features_output_indices(
out_features=_lowercase , out_indices=_lowercase , stage_names=self.stage_names)
| 344 | 1 |
import gc
import unittest
import numpy as np
import torch
from diffusers import StableDiffusionKDiffusionPipeline
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
enable_full_determinism()
@slow
@require_torch_gpu
class a_ ( unittest.TestCase ):
def __a ( self :int) -> int:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def __a ( self :Optional[int]) -> Union[str, Any]:
UpperCAmelCase_ = StableDiffusionKDiffusionPipeline.from_pretrained('''CompVis/stable-diffusion-v1-4''')
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
sd_pipe.set_scheduler('''sample_euler''')
UpperCAmelCase_ = '''A painting of a squirrel eating a burger'''
UpperCAmelCase_ = torch.manual_seed(0)
UpperCAmelCase_ = sd_pipe([prompt] , generator=_lowercase , guidance_scale=9.0 , num_inference_steps=20 , output_type='''np''')
UpperCAmelCase_ = output.images
UpperCAmelCase_ = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
UpperCAmelCase_ = np.array([0.0_447, 0.0_492, 0.0_468, 0.0_408, 0.0_383, 0.0_408, 0.0_354, 0.0_380, 0.0_339])
assert np.abs(image_slice.flatten() - expected_slice).max() < 1E-2
def __a ( self :Union[str, Any]) -> Union[str, Any]:
UpperCAmelCase_ = StableDiffusionKDiffusionPipeline.from_pretrained('''stabilityai/stable-diffusion-2-1-base''')
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
sd_pipe.set_scheduler('''sample_euler''')
UpperCAmelCase_ = '''A painting of a squirrel eating a burger'''
UpperCAmelCase_ = torch.manual_seed(0)
UpperCAmelCase_ = sd_pipe([prompt] , generator=_lowercase , guidance_scale=9.0 , num_inference_steps=20 , output_type='''np''')
UpperCAmelCase_ = output.images
UpperCAmelCase_ = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
UpperCAmelCase_ = np.array([0.1_237, 0.1_320, 0.1_438, 0.1_359, 0.1_390, 0.1_132, 0.1_277, 0.1_175, 0.1_112])
assert np.abs(image_slice.flatten() - expected_slice).max() < 5E-1
def __a ( self :List[str]) -> str:
UpperCAmelCase_ = StableDiffusionKDiffusionPipeline.from_pretrained('''stabilityai/stable-diffusion-2-1-base''')
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
sd_pipe.set_scheduler('''sample_dpmpp_2m''')
UpperCAmelCase_ = '''A painting of a squirrel eating a burger'''
UpperCAmelCase_ = torch.manual_seed(0)
UpperCAmelCase_ = sd_pipe(
[prompt] , generator=_lowercase , guidance_scale=7.5 , num_inference_steps=15 , output_type='''np''' , use_karras_sigmas=_lowercase , )
UpperCAmelCase_ = output.images
UpperCAmelCase_ = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
UpperCAmelCase_ = np.array(
[0.11_381_689, 0.12_112_921, 0.1_389_457, 0.12_549_606, 0.1_244_964, 0.10_831_517, 0.11_562_866, 0.10_867_816, 0.10_499_048])
assert np.abs(image_slice.flatten() - expected_slice).max() < 1E-2
| 344 |
import os
from argparse import ArgumentParser, Namespace
from ..data import SingleSentenceClassificationProcessor as Processor
from ..pipelines import TextClassificationPipeline
from ..utils import is_tf_available, is_torch_available, logging
from . import BaseTransformersCLICommand
if not is_tf_available() and not is_torch_available():
raise RuntimeError("At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training")
# TF training parameters
UpperCamelCase_ = False
UpperCamelCase_ = False
def A ( __UpperCAmelCase ) -> Any:
'''simple docstring'''
return TrainCommand(__UpperCAmelCase )
class a_ ( _snake_case ):
@staticmethod
def __a ( _lowercase :ArgumentParser) -> List[Any]:
UpperCAmelCase_ = parser.add_parser('''train''' , help='''CLI tool to train a model on a task.''')
train_parser.add_argument(
'''--train_data''' , type=_lowercase , required=_lowercase , help='''path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.''' , )
train_parser.add_argument(
'''--column_label''' , type=_lowercase , default=0 , help='''Column of the dataset csv file with example labels.''')
train_parser.add_argument(
'''--column_text''' , type=_lowercase , default=1 , help='''Column of the dataset csv file with example texts.''')
train_parser.add_argument(
'''--column_id''' , type=_lowercase , default=2 , help='''Column of the dataset csv file with example ids.''')
train_parser.add_argument(
'''--skip_first_row''' , action='''store_true''' , help='''Skip the first row of the csv file (headers).''')
train_parser.add_argument('''--validation_data''' , type=_lowercase , default='''''' , help='''path to validation dataset.''')
train_parser.add_argument(
'''--validation_split''' , type=_lowercase , default=0.1 , help='''if validation dataset is not provided, fraction of train dataset to use as validation dataset.''' , )
train_parser.add_argument('''--output''' , type=_lowercase , default='''./''' , help='''path to saved the trained model.''')
train_parser.add_argument(
'''--task''' , type=_lowercase , default='''text_classification''' , help='''Task to train the model on.''')
train_parser.add_argument(
'''--model''' , type=_lowercase , default='''bert-base-uncased''' , help='''Model\'s name or path to stored model.''')
train_parser.add_argument('''--train_batch_size''' , type=_lowercase , default=32 , help='''Batch size for training.''')
train_parser.add_argument('''--valid_batch_size''' , type=_lowercase , default=64 , help='''Batch size for validation.''')
train_parser.add_argument('''--learning_rate''' , type=_lowercase , default=3E-5 , help='''Learning rate.''')
train_parser.add_argument('''--adam_epsilon''' , type=_lowercase , default=1E-0_8 , help='''Epsilon for Adam optimizer.''')
train_parser.set_defaults(func=_lowercase)
def __init__( self :Union[str, Any] , _lowercase :Namespace) -> Union[str, Any]:
UpperCAmelCase_ = logging.get_logger('''transformers-cli/training''')
UpperCAmelCase_ = '''tf''' if is_tf_available() else '''torch'''
os.makedirs(args.output , exist_ok=_lowercase)
UpperCAmelCase_ = args.output
UpperCAmelCase_ = args.column_label
UpperCAmelCase_ = args.column_text
UpperCAmelCase_ = args.column_id
self.logger.info(f"Loading {args.task} pipeline for {args.model}")
if args.task == "text_classification":
UpperCAmelCase_ = TextClassificationPipeline.from_pretrained(args.model)
elif args.task == "token_classification":
raise NotImplementedError
elif args.task == "question_answering":
raise NotImplementedError
self.logger.info(f"Loading dataset from {args.train_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.train_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = None
if args.validation_data:
self.logger.info(f"Loading validation dataset from {args.validation_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.validation_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = args.validation_split
UpperCAmelCase_ = args.train_batch_size
UpperCAmelCase_ = args.valid_batch_size
UpperCAmelCase_ = args.learning_rate
UpperCAmelCase_ = args.adam_epsilon
def __a ( self :int) -> Tuple:
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def __a ( self :Optional[Any]) -> Any:
raise NotImplementedError
def __a ( self :int) -> Optional[Any]:
self.pipeline.fit(
self.train_dataset , validation_data=self.valid_dataset , validation_split=self.validation_split , learning_rate=self.learning_rate , adam_epsilon=self.adam_epsilon , train_batch_size=self.train_batch_size , valid_batch_size=self.valid_batch_size , )
# Save trained pipeline
self.pipeline.save_pretrained(self.output)
| 344 | 1 |
import warnings
from functools import wraps
from typing import Callable
def A ( __UpperCAmelCase ) -> Callable:
'''simple docstring'''
@wraps(__UpperCAmelCase )
def _inner_fn(*__UpperCAmelCase , **__UpperCAmelCase ):
warnings.warn(
(f"'{fn.__name__}' is experimental and might be subject to breaking changes in the future.") , __UpperCAmelCase , )
return fn(*__UpperCAmelCase , **__UpperCAmelCase )
return _inner_fn
| 344 |
import unittest
from typing import Dict, List, Optional, Union
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import BridgeTowerImageProcessor
class a_ ( unittest.TestCase ):
def __init__( self :Tuple , _lowercase :List[Any] , _lowercase :bool = True , _lowercase :Dict[str, int] = None , _lowercase :int = 32 , _lowercase :bool = True , _lowercase :Union[int, float] = 1 / 255 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :Optional[Union[float, List[float]]] = [0.48_145_466, 0.4_578_275, 0.40_821_073] , _lowercase :Optional[Union[float, List[float]]] = [0.26_862_954, 0.26_130_258, 0.27_577_711] , _lowercase :bool = True , _lowercase :List[Any]=7 , _lowercase :Dict=30 , _lowercase :Optional[int]=400 , _lowercase :Any=3 , ) -> Any:
UpperCAmelCase_ = parent
UpperCAmelCase_ = do_resize
UpperCAmelCase_ = size if size is not None else {'''shortest_edge''': 288}
UpperCAmelCase_ = size_divisor
UpperCAmelCase_ = do_rescale
UpperCAmelCase_ = rescale_factor
UpperCAmelCase_ = do_normalize
UpperCAmelCase_ = do_center_crop
UpperCAmelCase_ = image_mean
UpperCAmelCase_ = image_std
UpperCAmelCase_ = do_pad
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = min_resolution
UpperCAmelCase_ = max_resolution
def __a ( self :str) -> Tuple:
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"size_divisor": self.size_divisor,
}
def __a ( self :List[Any] , _lowercase :Tuple , _lowercase :List[str]=False) -> int:
if not batched:
UpperCAmelCase_ = self.size['''shortest_edge''']
UpperCAmelCase_ = image_inputs[0]
if isinstance(_lowercase , Image.Image):
UpperCAmelCase_ , UpperCAmelCase_ = image.size
else:
UpperCAmelCase_ , UpperCAmelCase_ = image.shape[1], image.shape[2]
UpperCAmelCase_ = size / min(_lowercase , _lowercase)
if h < w:
UpperCAmelCase_ , UpperCAmelCase_ = size, scale * w
else:
UpperCAmelCase_ , UpperCAmelCase_ = scale * h, size
UpperCAmelCase_ = int((1333 / 800) * size)
if max(_lowercase , _lowercase) > max_size:
UpperCAmelCase_ = max_size / max(_lowercase , _lowercase)
UpperCAmelCase_ = newh * scale
UpperCAmelCase_ = neww * scale
UpperCAmelCase_ , UpperCAmelCase_ = int(newh + 0.5), int(neww + 0.5)
UpperCAmelCase_ , UpperCAmelCase_ = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
UpperCAmelCase_ = []
for image in image_inputs:
UpperCAmelCase_ , UpperCAmelCase_ = self.get_expected_values([image])
expected_values.append((expected_height, expected_width))
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[0])[0]
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[1])[1]
return expected_height, expected_width
@require_torch
@require_vision
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =BridgeTowerImageProcessor if is_vision_available() else None
def __a ( self :int) -> Dict:
UpperCAmelCase_ = BridgeTowerImageProcessingTester(self)
@property
def __a ( self :Dict) -> Any:
return self.image_processor_tester.prepare_image_processor_dict()
def __a ( self :Dict) -> Tuple:
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(_lowercase , '''image_mean'''))
self.assertTrue(hasattr(_lowercase , '''image_std'''))
self.assertTrue(hasattr(_lowercase , '''do_normalize'''))
self.assertTrue(hasattr(_lowercase , '''do_resize'''))
self.assertTrue(hasattr(_lowercase , '''size'''))
self.assertTrue(hasattr(_lowercase , '''size_divisor'''))
def __a ( self :Union[str, Any]) -> Tuple:
pass
def __a ( self :List[str]) -> Tuple:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , Image.Image)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :Union[str, Any]) -> Optional[int]:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , numpify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , np.ndarray)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :str) -> int:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , torchify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , torch.Tensor)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
| 344 | 1 |
from typing import Optional, Tuple, Union
import flax
import flax.linen as nn
import jax
import jax.numpy as jnp
from flax.core.frozen_dict import FrozenDict
from ..configuration_utils import ConfigMixin, flax_register_to_config
from ..utils import BaseOutput
from .embeddings_flax import FlaxTimestepEmbedding, FlaxTimesteps
from .modeling_flax_utils import FlaxModelMixin
from .unet_ad_blocks_flax import (
FlaxCrossAttnDownBlockaD,
FlaxCrossAttnUpBlockaD,
FlaxDownBlockaD,
FlaxUNetMidBlockaDCrossAttn,
FlaxUpBlockaD,
)
@flax.struct.dataclass
class a_ ( _snake_case ):
UpperCamelCase__ : jnp.ndarray
@flax_register_to_config
class a_ ( nn.Module , _snake_case , _snake_case ):
UpperCamelCase__ : int =32
UpperCamelCase__ : int =4
UpperCamelCase__ : int =4
UpperCamelCase__ : Tuple[str] =(
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
UpperCamelCase__ : Tuple[str] =("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
UpperCamelCase__ : Union[bool, Tuple[bool]] =False
UpperCamelCase__ : Tuple[int] =(3_20, 6_40, 12_80, 12_80)
UpperCamelCase__ : int =2
UpperCamelCase__ : Union[int, Tuple[int]] =8
UpperCamelCase__ : Optional[Union[int, Tuple[int]]] =None
UpperCamelCase__ : int =12_80
UpperCamelCase__ : float =0.0
UpperCamelCase__ : bool =False
UpperCamelCase__ : jnp.dtype =jnp.floataa
UpperCamelCase__ : bool =True
UpperCamelCase__ : int =0
UpperCamelCase__ : bool =False
def __a ( self :int , _lowercase :jax.random.KeyArray) -> FrozenDict:
# init input tensors
UpperCAmelCase_ = (1, self.in_channels, self.sample_size, self.sample_size)
UpperCAmelCase_ = jnp.zeros(_lowercase , dtype=jnp.floataa)
UpperCAmelCase_ = jnp.ones((1,) , dtype=jnp.intaa)
UpperCAmelCase_ = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa)
UpperCAmelCase_ , UpperCAmelCase_ = jax.random.split(_lowercase)
UpperCAmelCase_ = {'''params''': params_rng, '''dropout''': dropout_rng}
return self.init(_lowercase , _lowercase , _lowercase , _lowercase)["params"]
def __a ( self :Optional[int]) -> str:
UpperCAmelCase_ = self.block_out_channels
UpperCAmelCase_ = block_out_channels[0] * 4
if self.num_attention_heads is not None:
raise ValueError(
'''At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19.''')
# If `num_attention_heads` is not defined (which is the case for most models)
# it will default to `attention_head_dim`. This looks weird upon first reading it and it is.
# The reason for this behavior is to correct for incorrectly named variables that were introduced
# when this library was created. The incorrect naming was only discovered much later in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131
# Changing `attention_head_dim` to `num_attention_heads` for 40,000+ configurations is too backwards breaking
# which is why we correct for the naming here.
UpperCAmelCase_ = self.num_attention_heads or self.attention_head_dim
# input
UpperCAmelCase_ = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
UpperCAmelCase_ = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift)
UpperCAmelCase_ = FlaxTimestepEmbedding(_lowercase , dtype=self.dtype)
UpperCAmelCase_ = self.only_cross_attention
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = (only_cross_attention,) * len(self.down_block_types)
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = (num_attention_heads,) * len(self.down_block_types)
# down
UpperCAmelCase_ = []
UpperCAmelCase_ = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types):
UpperCAmelCase_ = output_channel
UpperCAmelCase_ = block_out_channels[i]
UpperCAmelCase_ = i == len(_lowercase) - 1
if down_block_type == "CrossAttnDownBlock2D":
UpperCAmelCase_ = FlaxCrossAttnDownBlockaD(
in_channels=_lowercase , out_channels=_lowercase , dropout=self.dropout , num_layers=self.layers_per_block , num_attention_heads=num_attention_heads[i] , add_downsample=not is_final_block , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
UpperCAmelCase_ = FlaxDownBlockaD(
in_channels=_lowercase , out_channels=_lowercase , dropout=self.dropout , num_layers=self.layers_per_block , add_downsample=not is_final_block , dtype=self.dtype , )
down_blocks.append(_lowercase)
UpperCAmelCase_ = down_blocks
# mid
UpperCAmelCase_ = FlaxUNetMidBlockaDCrossAttn(
in_channels=block_out_channels[-1] , dropout=self.dropout , num_attention_heads=num_attention_heads[-1] , use_linear_projection=self.use_linear_projection , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
# up
UpperCAmelCase_ = []
UpperCAmelCase_ = list(reversed(_lowercase))
UpperCAmelCase_ = list(reversed(_lowercase))
UpperCAmelCase_ = list(reversed(_lowercase))
UpperCAmelCase_ = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types):
UpperCAmelCase_ = output_channel
UpperCAmelCase_ = reversed_block_out_channels[i]
UpperCAmelCase_ = reversed_block_out_channels[min(i + 1 , len(_lowercase) - 1)]
UpperCAmelCase_ = i == len(_lowercase) - 1
if up_block_type == "CrossAttnUpBlock2D":
UpperCAmelCase_ = FlaxCrossAttnUpBlockaD(
in_channels=_lowercase , out_channels=_lowercase , prev_output_channel=_lowercase , num_layers=self.layers_per_block + 1 , num_attention_heads=reversed_num_attention_heads[i] , add_upsample=not is_final_block , dropout=self.dropout , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
UpperCAmelCase_ = FlaxUpBlockaD(
in_channels=_lowercase , out_channels=_lowercase , prev_output_channel=_lowercase , num_layers=self.layers_per_block + 1 , add_upsample=not is_final_block , dropout=self.dropout , dtype=self.dtype , )
up_blocks.append(_lowercase)
UpperCAmelCase_ = output_channel
UpperCAmelCase_ = up_blocks
# out
UpperCAmelCase_ = nn.GroupNorm(num_groups=32 , epsilon=1E-5)
UpperCAmelCase_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self :Any , _lowercase :Optional[int] , _lowercase :Optional[int] , _lowercase :Tuple , _lowercase :List[str]=None , _lowercase :Union[str, Any]=None , _lowercase :bool = True , _lowercase :bool = False , ) -> Union[FlaxUNetaDConditionOutput, Tuple]:
# 1. time
if not isinstance(_lowercase , jnp.ndarray):
UpperCAmelCase_ = jnp.array([timesteps] , dtype=jnp.intaa)
elif isinstance(_lowercase , jnp.ndarray) and len(timesteps.shape) == 0:
UpperCAmelCase_ = timesteps.astype(dtype=jnp.floataa)
UpperCAmelCase_ = jnp.expand_dims(_lowercase , 0)
UpperCAmelCase_ = self.time_proj(_lowercase)
UpperCAmelCase_ = self.time_embedding(_lowercase)
# 2. pre-process
UpperCAmelCase_ = jnp.transpose(_lowercase , (0, 2, 3, 1))
UpperCAmelCase_ = self.conv_in(_lowercase)
# 3. down
UpperCAmelCase_ = (sample,)
for down_block in self.down_blocks:
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ , UpperCAmelCase_ = down_block(_lowercase , _lowercase , _lowercase , deterministic=not train)
else:
UpperCAmelCase_ , UpperCAmelCase_ = down_block(_lowercase , _lowercase , deterministic=not train)
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
UpperCAmelCase_ = ()
for down_block_res_sample, down_block_additional_residual in zip(
_lowercase , _lowercase):
down_block_res_sample += down_block_additional_residual
new_down_block_res_samples += (down_block_res_sample,)
UpperCAmelCase_ = new_down_block_res_samples
# 4. mid
UpperCAmelCase_ = self.mid_block(_lowercase , _lowercase , _lowercase , deterministic=not train)
if mid_block_additional_residual is not None:
sample += mid_block_additional_residual
# 5. up
for up_block in self.up_blocks:
UpperCAmelCase_ = down_block_res_samples[-(self.layers_per_block + 1) :]
UpperCAmelCase_ = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = up_block(
_lowercase , temb=_lowercase , encoder_hidden_states=_lowercase , res_hidden_states_tuple=_lowercase , deterministic=not train , )
else:
UpperCAmelCase_ = up_block(_lowercase , temb=_lowercase , res_hidden_states_tuple=_lowercase , deterministic=not train)
# 6. post-process
UpperCAmelCase_ = self.conv_norm_out(_lowercase)
UpperCAmelCase_ = nn.silu(_lowercase)
UpperCAmelCase_ = self.conv_out(_lowercase)
UpperCAmelCase_ = jnp.transpose(_lowercase , (0, 3, 1, 2))
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=_lowercase)
| 344 |
def A ( __UpperCAmelCase = 100_0000 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = [i - 1 for i in range(limit + 1 )]
for i in range(2 , limit + 1 ):
if phi[i] == i - 1:
for j in range(2 * i , limit + 1 , __UpperCAmelCase ):
phi[j] -= phi[j] // i
return sum(phi[2 : limit + 1] )
if __name__ == "__main__":
print(solution())
| 344 | 1 |
import sys
def A ( __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = len(__UpperCAmelCase )
UpperCAmelCase_ = [[0 for x in range(__UpperCAmelCase )] for x in range(__UpperCAmelCase )]
UpperCAmelCase_ = [[0 for x in range(__UpperCAmelCase )] for x in range(__UpperCAmelCase )]
for chain_length in range(2 , __UpperCAmelCase ):
for a in range(1 , n - chain_length + 1 ):
UpperCAmelCase_ = a + chain_length - 1
UpperCAmelCase_ = sys.maxsize
for c in range(__UpperCAmelCase , __UpperCAmelCase ):
UpperCAmelCase_ = (
matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b]
)
if cost < matrix[a][b]:
UpperCAmelCase_ = cost
UpperCAmelCase_ = c
return matrix, sol
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Any:
'''simple docstring'''
if i == j:
print('''A''' + str(__UpperCAmelCase ) , end=''' ''' )
else:
print('''(''' , end=''' ''' )
print_optiomal_solution(__UpperCAmelCase , __UpperCAmelCase , optimal_solution[i][j] )
print_optiomal_solution(__UpperCAmelCase , optimal_solution[i][j] + 1 , __UpperCAmelCase )
print(''')''' , end=''' ''' )
def A ( ) -> Tuple:
'''simple docstring'''
UpperCAmelCase_ = [30, 35, 15, 5, 10, 20, 25]
UpperCAmelCase_ = len(__UpperCAmelCase )
# Size of matrix created from above array will be
# 30*35 35*15 15*5 5*10 10*20 20*25
UpperCAmelCase_ , UpperCAmelCase_ = matrix_chain_order(__UpperCAmelCase )
print('''No. of Operation required: ''' + str(matrix[1][n - 1] ) )
print_optiomal_solution(__UpperCAmelCase , 1 , n - 1 )
if __name__ == "__main__":
main()
| 344 |
import tempfile
import torch
from diffusers import PNDMScheduler
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =(PNDMScheduler,)
UpperCamelCase__ : Optional[Any] =(("num_inference_steps", 50),)
def __a ( self :Union[str, Any] , **_lowercase :Any) -> Union[str, Any]:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
}
config.update(**_lowercase)
return config
def __a ( self :str , _lowercase :List[Any]=0 , **_lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Any) -> Optional[Any]:
pass
def __a ( self :str , _lowercase :int=0 , **_lowercase :Union[str, Any]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :int , **_lowercase :str) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.prk_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
for i, t in enumerate(scheduler.plms_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :Union[str, Any]) -> int:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
if num_inference_steps is not None and hasattr(_lowercase , '''set_timesteps'''):
scheduler.set_timesteps(_lowercase)
elif num_inference_steps is not None and not hasattr(_lowercase , '''set_timesteps'''):
UpperCAmelCase_ = num_inference_steps
# copy over dummy past residuals (must be done after set_timesteps)
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
def __a ( self :Any) -> Dict:
for timesteps in [100, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :List[Any]) -> Any:
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(steps_offset=1)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(10)
assert torch.equal(
scheduler.timesteps , torch.LongTensor(
[901, 851, 851, 801, 801, 751, 751, 701, 701, 651, 651, 601, 601, 501, 401, 301, 201, 101, 1]) , )
def __a ( self :Optional[int]) -> str:
for beta_start, beta_end in zip([0.0_001, 0.001] , [0.002, 0.02]):
self.check_over_configs(beta_start=_lowercase , beta_end=_lowercase)
def __a ( self :Any) -> List[str]:
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=_lowercase)
def __a ( self :List[Any]) -> Dict:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Tuple:
for t in [1, 5, 10]:
self.check_over_forward(time_step=_lowercase)
def __a ( self :Tuple) -> Dict:
for t, num_inference_steps in zip([1, 5, 10] , [10, 50, 100]):
self.check_over_forward(num_inference_steps=_lowercase)
def __a ( self :str) -> List[Any]:
# earlier version of set_timesteps() caused an error indexing alpha's with inference steps as power of 3
UpperCAmelCase_ = 27
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# before power of 3 fix, would error on first step, so we only need to do two
for i, t in enumerate(scheduler.prk_timesteps[:2]):
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
def __a ( self :List[str]) -> int:
with self.assertRaises(_lowercase):
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.step_plms(self.dummy_sample , 1 , self.dummy_sample).prev_sample
def __a ( self :List[str]) -> Dict:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 198.1_318) < 1E-2
assert abs(result_mean.item() - 0.2_580) < 1E-3
def __a ( self :Any) -> Tuple:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 67.3_986) < 1E-2
assert abs(result_mean.item() - 0.0_878) < 1E-3
def __a ( self :int) -> Any:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 230.0_399) < 1E-2
assert abs(result_mean.item() - 0.2_995) < 1E-3
def __a ( self :Any) -> Dict:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 186.9_482) < 1E-2
assert abs(result_mean.item() - 0.2_434) < 1E-3
| 344 | 1 |
import argparse
import torch
from diffusers.pipelines.stable_diffusion.convert_from_ckpt import download_from_original_stable_diffusion_ckpt
if __name__ == "__main__":
UpperCamelCase_ = argparse.ArgumentParser()
parser.add_argument(
"--checkpoint_path", default=None, type=str, required=True, help="Path to the checkpoint to convert."
)
# !wget https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml
parser.add_argument(
"--original_config_file",
default=None,
type=str,
help="The YAML config file corresponding to the original architecture.",
)
parser.add_argument(
"--num_in_channels",
default=None,
type=int,
help="The number of input channels. If `None` number of input channels will be automatically inferred.",
)
parser.add_argument(
"--scheduler_type",
default="pndm",
type=str,
help="Type of scheduler to use. Should be one of ['pndm', 'lms', 'ddim', 'euler', 'euler-ancestral', 'dpm']",
)
parser.add_argument(
"--pipeline_type",
default=None,
type=str,
help=(
"The pipeline type. One of 'FrozenOpenCLIPEmbedder', 'FrozenCLIPEmbedder', 'PaintByExample'"
". If `None` pipeline will be automatically inferred."
),
)
parser.add_argument(
"--image_size",
default=None,
type=int,
help=(
"The image size that the model was trained on. Use 512 for Stable Diffusion v1.X and Stable Siffusion v2"
" Base. Use 768 for Stable Diffusion v2."
),
)
parser.add_argument(
"--prediction_type",
default=None,
type=str,
help=(
"The prediction type that the model was trained on. Use 'epsilon' for Stable Diffusion v1.X and Stable"
" Diffusion v2 Base. Use 'v_prediction' for Stable Diffusion v2."
),
)
parser.add_argument(
"--extract_ema",
action="store_true",
help=(
"Only relevant for checkpoints that have both EMA and non-EMA weights. Whether to extract the EMA weights"
" or not. Defaults to `False`. Add `--extract_ema` to extract the EMA weights. EMA weights usually yield"
" higher quality images for inference. Non-EMA weights are usually better to continue fine-tuning."
),
)
parser.add_argument(
"--upcast_attention",
action="store_true",
help=(
"Whether the attention computation should always be upcasted. This is necessary when running stable"
" diffusion 2.1."
),
)
parser.add_argument(
"--from_safetensors",
action="store_true",
help="If `--checkpoint_path` is in `safetensors` format, load checkpoint with safetensors instead of PyTorch.",
)
parser.add_argument(
"--to_safetensors",
action="store_true",
help="Whether to store pipeline in safetensors format or not.",
)
parser.add_argument("--dump_path", default=None, type=str, required=True, help="Path to the output model.")
parser.add_argument("--device", type=str, help="Device to use (e.g. cpu, cuda:0, cuda:1, etc.)")
parser.add_argument(
"--stable_unclip",
type=str,
default=None,
required=False,
help="Set if this is a stable unCLIP model. One of 'txt2img' or 'img2img'.",
)
parser.add_argument(
"--stable_unclip_prior",
type=str,
default=None,
required=False,
help="Set if this is a stable unCLIP txt2img model. Selects which prior to use. If `--stable_unclip` is set to `txt2img`, the karlo prior (https://huggingface.co/kakaobrain/karlo-v1-alpha/tree/main/prior) is selected by default.",
)
parser.add_argument(
"--clip_stats_path",
type=str,
help="Path to the clip stats file. Only required if the stable unclip model's config specifies `model.params.noise_aug_config.params.clip_stats_path`.",
required=False,
)
parser.add_argument(
"--controlnet", action="store_true", default=None, help="Set flag if this is a controlnet checkpoint."
)
parser.add_argument("--half", action="store_true", help="Save weights in half precision.")
parser.add_argument(
"--vae_path",
type=str,
default=None,
required=False,
help="Set to a path, hub id to an already converted vae to not convert it again.",
)
UpperCamelCase_ = parser.parse_args()
UpperCamelCase_ = download_from_original_stable_diffusion_ckpt(
checkpoint_path=args.checkpoint_path,
original_config_file=args.original_config_file,
image_size=args.image_size,
prediction_type=args.prediction_type,
model_type=args.pipeline_type,
extract_ema=args.extract_ema,
scheduler_type=args.scheduler_type,
num_in_channels=args.num_in_channels,
upcast_attention=args.upcast_attention,
from_safetensors=args.from_safetensors,
device=args.device,
stable_unclip=args.stable_unclip,
stable_unclip_prior=args.stable_unclip_prior,
clip_stats_path=args.clip_stats_path,
controlnet=args.controlnet,
vae_path=args.vae_path,
)
if args.half:
pipe.to(torch_dtype=torch.floataa)
if args.controlnet:
# only save the controlnet model
pipe.controlnet.save_pretrained(args.dump_path, safe_serialization=args.to_safetensors)
else:
pipe.save_pretrained(args.dump_path, safe_serialization=args.to_safetensors)
| 344 |
import os
from dataclasses import dataclass, field
from io import BytesIO
from typing import TYPE_CHECKING, Any, ClassVar, Dict, Optional, Union
import numpy as np
import pyarrow as pa
from .. import config
from ..download.streaming_download_manager import xopen, xsplitext
from ..table import array_cast
from ..utils.py_utils import no_op_if_value_is_null, string_to_dict
if TYPE_CHECKING:
from .features import FeatureType
UpperCamelCase_ , UpperCamelCase_ , UpperCamelCase_ = False, False, False
@dataclass
class a_ :
UpperCamelCase__ : Optional[int] =None
UpperCamelCase__ : bool =True
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[str] =None
# Automatically constructed
UpperCamelCase__ : ClassVar[str] ="dict"
UpperCamelCase__ : ClassVar[Any] =pa.struct({"bytes": pa.binary(), "path": pa.string()} )
UpperCamelCase__ : str =field(default="Audio" , init=_snake_case , repr=_snake_case )
def __call__( self :List[Any]) -> List[Any]:
return self.pa_type
def __a ( self :Any , _lowercase :Union[str, bytes, dict]) -> dict:
try:
import soundfile as sf # soundfile is a dependency of librosa, needed to decode audio files.
except ImportError as err:
raise ImportError('''To support encoding audio data, please install \'soundfile\'.''') from err
if isinstance(_lowercase , _lowercase):
return {"bytes": None, "path": value}
elif isinstance(_lowercase , _lowercase):
return {"bytes": value, "path": None}
elif "array" in value:
# convert the audio array to wav bytes
UpperCAmelCase_ = BytesIO()
sf.write(_lowercase , value['''array'''] , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
elif value.get('''path''') is not None and os.path.isfile(value['''path''']):
# we set "bytes": None to not duplicate the data if they're already available locally
if value["path"].endswith('''pcm'''):
# "PCM" only has raw audio bytes
if value.get('''sampling_rate''') is None:
# At least, If you want to convert "PCM-byte" to "WAV-byte", you have to know sampling rate
raise KeyError('''To use PCM files, please specify a \'sampling_rate\' in Audio object''')
if value.get('''bytes'''):
# If we already had PCM-byte, we don`t have to make "read file, make bytes" (just use it!)
UpperCAmelCase_ = np.frombuffer(value['''bytes'''] , dtype=np.intaa).astype(np.floataa) / 32767
else:
UpperCAmelCase_ = np.memmap(value['''path'''] , dtype='''h''' , mode='''r''').astype(np.floataa) / 32767
UpperCAmelCase_ = BytesIO(bytes())
sf.write(_lowercase , _lowercase , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
else:
return {"bytes": None, "path": value.get('''path''')}
elif value.get('''bytes''') is not None or value.get('''path''') is not None:
# store the audio bytes, and path is used to infer the audio format using the file extension
return {"bytes": value.get('''bytes'''), "path": value.get('''path''')}
else:
raise ValueError(
f"An audio sample should have one of 'path' or 'bytes' but they are missing or None in {value}.")
def __a ( self :Dict , _lowercase :dict , _lowercase :Optional[Dict[str, Union[str, bool, None]]] = None) -> dict:
if not self.decode:
raise RuntimeError('''Decoding is disabled for this feature. Please use Audio(decode=True) instead.''')
UpperCAmelCase_ , UpperCAmelCase_ = (value['''path'''], BytesIO(value['''bytes'''])) if value['''bytes'''] is not None else (value['''path'''], None)
if path is None and file is None:
raise ValueError(f"An audio sample should have one of 'path' or 'bytes' but both are None in {value}.")
try:
import librosa
import soundfile as sf
except ImportError as err:
raise ImportError('''To support decoding audio files, please install \'librosa\' and \'soundfile\'.''') from err
UpperCAmelCase_ = xsplitext(_lowercase)[1][1:].lower() if path is not None else None
if not config.IS_OPUS_SUPPORTED and audio_format == "opus":
raise RuntimeError(
'''Decoding \'opus\' files requires system library \'libsndfile\'>=1.0.31, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
elif not config.IS_MP3_SUPPORTED and audio_format == "mp3":
raise RuntimeError(
'''Decoding \'mp3\' files requires system library \'libsndfile\'>=1.1.0, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
if file is None:
UpperCAmelCase_ = token_per_repo_id or {}
UpperCAmelCase_ = path.split('''::''')[-1]
try:
UpperCAmelCase_ = string_to_dict(_lowercase , config.HUB_DATASETS_URL)['''repo_id''']
UpperCAmelCase_ = token_per_repo_id[repo_id]
except (ValueError, KeyError):
UpperCAmelCase_ = None
with xopen(_lowercase , '''rb''' , use_auth_token=_lowercase) as f:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
else:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
UpperCAmelCase_ = array.T
if self.mono:
UpperCAmelCase_ = librosa.to_mono(_lowercase)
if self.sampling_rate and self.sampling_rate != sampling_rate:
UpperCAmelCase_ = librosa.resample(_lowercase , orig_sr=_lowercase , target_sr=self.sampling_rate)
UpperCAmelCase_ = self.sampling_rate
return {"path": path, "array": array, "sampling_rate": sampling_rate}
def __a ( self :Union[str, Any]) -> Union["FeatureType", Dict[str, "FeatureType"]]:
from .features import Value
if self.decode:
raise ValueError('''Cannot flatten a decoded Audio feature.''')
return {
"bytes": Value('''binary'''),
"path": Value('''string'''),
}
def __a ( self :int , _lowercase :Union[pa.StringArray, pa.StructArray]) -> pa.StructArray:
if pa.types.is_string(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, storage] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_binary(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([storage, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_struct(storage.type) and storage.type.get_all_field_indices('''array'''):
UpperCAmelCase_ = pa.array([Audio().encode_example(_lowercase) if x is not None else None for x in storage.to_pylist()])
elif pa.types.is_struct(storage.type):
if storage.type.get_field_index('''bytes''') >= 0:
UpperCAmelCase_ = storage.field('''bytes''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
if storage.type.get_field_index('''path''') >= 0:
UpperCAmelCase_ = storage.field('''path''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
return array_cast(_lowercase , self.pa_type)
def __a ( self :Any , _lowercase :pa.StructArray) -> pa.StructArray:
@no_op_if_value_is_null
def path_to_bytes(_lowercase :Tuple):
with xopen(_lowercase , '''rb''') as f:
UpperCAmelCase_ = f.read()
return bytes_
UpperCAmelCase_ = pa.array(
[
(path_to_bytes(x['''path''']) if x['''bytes'''] is None else x['''bytes''']) if x is not None else None
for x in storage.to_pylist()
] , type=pa.binary() , )
UpperCAmelCase_ = pa.array(
[os.path.basename(_lowercase) if path is not None else None for path in storage.field('''path''').to_pylist()] , type=pa.string() , )
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
| 344 | 1 |
import fire
from utils import calculate_rouge, save_json
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase=None , **__UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
UpperCAmelCase_ = [x.strip() for x in open(__UpperCAmelCase ).readlines()]
UpperCAmelCase_ = [x.strip() for x in open(__UpperCAmelCase ).readlines()][: len(__UpperCAmelCase )]
UpperCAmelCase_ = calculate_rouge(__UpperCAmelCase , __UpperCAmelCase , **__UpperCAmelCase )
if save_path is not None:
save_json(__UpperCAmelCase , __UpperCAmelCase , indent=__UpperCAmelCase )
return metrics # these print nicely
if __name__ == "__main__":
fire.Fire(calculate_rouge_path)
| 344 |
from ..utils import DummyObject, requires_backends
class a_ ( metaclass=_snake_case ):
UpperCamelCase__ : Any =["torch", "scipy"]
def __init__( self :List[str] , *_lowercase :List[str] , **_lowercase :Union[str, Any]) -> List[Any]:
requires_backends(self , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Dict , *_lowercase :Any , **_lowercase :Dict) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Optional[Any] , *_lowercase :str , **_lowercase :Optional[Any]) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
| 344 | 1 |
import unittest
from accelerate import debug_launcher
from accelerate.test_utils import require_cpu, test_ops, test_script
@require_cpu
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> Union[str, Any]:
debug_launcher(test_script.main)
def __a ( self :List[str]) -> str:
debug_launcher(test_ops.main)
| 344 |
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def A ( __UpperCAmelCase ) -> Dict[str, torch.Tensor]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for rt in rc.restypes:
UpperCAmelCase_ = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
UpperCAmelCase_ = {name: i for i, name in enumerate(__UpperCAmelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.floataa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = protein['''aatype'''].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the corresponding mask
UpperCAmelCase_ = torch.zeros([21, 37] , dtype=torch.floataa , device=protein['''aatype'''].device )
for restype, restype_letter in enumerate(rc.restypes ):
UpperCAmelCase_ = rc.restype_atoa[restype_letter]
UpperCAmelCase_ = rc.residue_atoms[restype_name]
for atom_name in atom_names:
UpperCAmelCase_ = rc.atom_order[atom_name]
UpperCAmelCase_ = 1
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
return protein
def A ( __UpperCAmelCase ) -> Dict[str, np.ndarray]:
'''simple docstring'''
UpperCAmelCase_ = tree_map(lambda __UpperCAmelCase : torch.tensor(__UpperCAmelCase , device=batch['''aatype'''].device ) , __UpperCAmelCase , np.ndarray )
UpperCAmelCase_ = tensor_tree_map(lambda __UpperCAmelCase : np.array(__UpperCAmelCase ) , make_atomaa_masks(__UpperCAmelCase ) )
return out
| 344 | 1 |
from typing import TYPE_CHECKING
from ..utils import _LazyModule
UpperCamelCase_ = {
"config": [
"EXTERNAL_DATA_FORMAT_SIZE_LIMIT",
"OnnxConfig",
"OnnxConfigWithPast",
"OnnxSeq2SeqConfigWithPast",
"PatchingSpec",
],
"convert": ["export", "validate_model_outputs"],
"features": ["FeaturesManager"],
"utils": ["ParameterFormat", "compute_serialized_parameters_size"],
}
if TYPE_CHECKING:
from .config import (
EXTERNAL_DATA_FORMAT_SIZE_LIMIT,
OnnxConfig,
OnnxConfigWithPast,
OnnxSeqaSeqConfigWithPast,
PatchingSpec,
)
from .convert import export, validate_model_outputs
from .features import FeaturesManager
from .utils import ParameterFormat, compute_serialized_parameters_size
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from ..models.whisper import WhisperForConditionalGeneration, WhisperProcessor
from .base import PipelineTool
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="openai/whisper-base"
UpperCamelCase__ : int =(
"This is a tool that transcribes an audio into text. It takes an input named `audio` and returns the "
"transcribed text."
)
UpperCamelCase__ : Any ="transcriber"
UpperCamelCase__ : Optional[int] =WhisperProcessor
UpperCamelCase__ : List[str] =WhisperForConditionalGeneration
UpperCamelCase__ : List[Any] =["audio"]
UpperCamelCase__ : Union[str, Any] =["text"]
def __a ( self :int , _lowercase :Any) -> Tuple:
return self.pre_processor(_lowercase , return_tensors='''pt''').input_features
def __a ( self :Dict , _lowercase :Tuple) -> Any:
return self.model.generate(inputs=_lowercase)
def __a ( self :int , _lowercase :Union[str, Any]) -> Optional[Any]:
return self.pre_processor.batch_decode(_lowercase , skip_special_tokens=_lowercase)[0]
| 344 | 1 |
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"microsoft/beit-base-patch16-224-pt22k": (
"https://huggingface.co/microsoft/beit-base-patch16-224-pt22k/resolve/main/config.json"
),
# See all BEiT models at https://huggingface.co/models?filter=beit
}
class a_ ( _snake_case ):
UpperCamelCase__ : Tuple ="beit"
def __init__( self :Tuple , _lowercase :Dict=8192 , _lowercase :List[Any]=768 , _lowercase :Tuple=12 , _lowercase :Optional[int]=12 , _lowercase :int=3072 , _lowercase :int="gelu" , _lowercase :Union[str, Any]=0.0 , _lowercase :List[str]=0.0 , _lowercase :Dict=0.02 , _lowercase :str=1E-1_2 , _lowercase :List[str]=224 , _lowercase :Any=16 , _lowercase :Optional[int]=3 , _lowercase :List[Any]=False , _lowercase :int=False , _lowercase :Tuple=False , _lowercase :int=False , _lowercase :str=0.1 , _lowercase :List[Any]=0.1 , _lowercase :Optional[Any]=True , _lowercase :Optional[Any]=[3, 5, 7, 11] , _lowercase :Union[str, Any]=[1, 2, 3, 6] , _lowercase :List[Any]=True , _lowercase :Optional[Any]=0.4 , _lowercase :Dict=256 , _lowercase :Optional[Any]=1 , _lowercase :Optional[int]=False , _lowercase :Optional[Any]=255 , **_lowercase :List[str] , ) -> Any:
super().__init__(**_lowercase)
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = image_size
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = use_mask_token
UpperCAmelCase_ = use_absolute_position_embeddings
UpperCAmelCase_ = use_relative_position_bias
UpperCAmelCase_ = use_shared_relative_position_bias
UpperCAmelCase_ = layer_scale_init_value
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = use_mean_pooling
# decode head attributes (semantic segmentation)
UpperCAmelCase_ = out_indices
UpperCAmelCase_ = pool_scales
# auxiliary head attributes (semantic segmentation)
UpperCAmelCase_ = use_auxiliary_head
UpperCAmelCase_ = auxiliary_loss_weight
UpperCAmelCase_ = auxiliary_channels
UpperCAmelCase_ = auxiliary_num_convs
UpperCAmelCase_ = auxiliary_concat_input
UpperCAmelCase_ = semantic_loss_ignore_index
class a_ ( _snake_case ):
UpperCamelCase__ : Dict =version.parse("1.11" )
@property
def __a ( self :Optional[Any]) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
])
@property
def __a ( self :Optional[int]) -> float:
return 1E-4
| 344 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
UpperCamelCase_ = {"configuration_opt": ["OPT_PRETRAINED_CONFIG_ARCHIVE_MAP", "OPTConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"OPT_PRETRAINED_MODEL_ARCHIVE_LIST",
"OPTForCausalLM",
"OPTModel",
"OPTPreTrainedModel",
"OPTForSequenceClassification",
"OPTForQuestionAnswering",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["TFOPTForCausalLM", "TFOPTModel", "TFOPTPreTrainedModel"]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxOPTForCausalLM",
"FlaxOPTModel",
"FlaxOPTPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_opt import OPT_PRETRAINED_CONFIG_ARCHIVE_MAP, OPTConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_opt import (
OPT_PRETRAINED_MODEL_ARCHIVE_LIST,
OPTForCausalLM,
OPTForQuestionAnswering,
OPTForSequenceClassification,
OPTModel,
OPTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_opt import TFOPTForCausalLM, TFOPTModel, TFOPTPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_opt import FlaxOPTForCausalLM, FlaxOPTModel, FlaxOPTPreTrainedModel
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 | 1 |
import itertools
from dataclasses import dataclass
from typing import Optional
import pandas as pd
import pyarrow as pa
import datasets
from datasets.table import table_cast
@dataclass
class a_ ( datasets.BuilderConfig ):
UpperCamelCase__ : Optional[datasets.Features] =None
class a_ ( datasets.ArrowBasedBuilder ):
UpperCamelCase__ : Optional[int] =PandasConfig
def __a ( self :int) -> int:
return datasets.DatasetInfo(features=self.config.features)
def __a ( self :List[str] , _lowercase :Dict) -> Tuple:
if not self.config.data_files:
raise ValueError(f"At least one data file must be specified, but got data_files={self.config.data_files}")
UpperCAmelCase_ = dl_manager.download_and_extract(self.config.data_files)
if isinstance(_lowercase , (str, list, tuple)):
UpperCAmelCase_ = data_files
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = [files]
# Use `dl_manager.iter_files` to skip hidden files in an extracted archive
UpperCAmelCase_ = [dl_manager.iter_files(_lowercase) for file in files]
return [datasets.SplitGenerator(name=datasets.Split.TRAIN , gen_kwargs={'''files''': files})]
UpperCAmelCase_ = []
for split_name, files in data_files.items():
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = [files]
# Use `dl_manager.iter_files` to skip hidden files in an extracted archive
UpperCAmelCase_ = [dl_manager.iter_files(_lowercase) for file in files]
splits.append(datasets.SplitGenerator(name=_lowercase , gen_kwargs={'''files''': files}))
return splits
def __a ( self :Optional[Any] , _lowercase :pa.Table) -> pa.Table:
if self.config.features is not None:
# more expensive cast to support nested features with keys in a different order
# allows str <-> int/float or str to Audio for example
UpperCAmelCase_ = table_cast(_lowercase , self.config.features.arrow_schema)
return pa_table
def __a ( self :List[Any] , _lowercase :Union[str, Any]) -> Any:
for i, file in enumerate(itertools.chain.from_iterable(_lowercase)):
with open(_lowercase , '''rb''') as f:
UpperCAmelCase_ = pa.Table.from_pandas(pd.read_pickle(_lowercase))
yield i, self._cast_table(_lowercase)
| 344 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
UpperCamelCase_ = "▁"
UpperCamelCase_ = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : str =BigBirdTokenizer
UpperCamelCase__ : Tuple =BigBirdTokenizerFast
UpperCamelCase__ : Union[str, Any] =True
UpperCamelCase__ : List[str] =True
def __a ( self :Any) -> List[str]:
super().setUp()
UpperCAmelCase_ = self.tokenizer_class(_lowercase , keep_accents=_lowercase)
tokenizer.save_pretrained(self.tmpdirname)
def __a ( self :Optional[int]) -> str:
UpperCAmelCase_ = '''<s>'''
UpperCAmelCase_ = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_lowercase) , _lowercase)
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_lowercase) , _lowercase)
def __a ( self :str) -> str:
UpperCAmelCase_ = list(self.get_tokenizer().get_vocab().keys())
self.assertEqual(vocab_keys[0] , '''<unk>''')
self.assertEqual(vocab_keys[1] , '''<s>''')
self.assertEqual(vocab_keys[-1] , '''[MASK]''')
self.assertEqual(len(_lowercase) , 1004)
def __a ( self :List[str]) -> int:
self.assertEqual(self.get_tokenizer().vocab_size , 1000)
def __a ( self :Tuple) -> int:
if not self.test_rust_tokenizer:
return
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = '''I was born in 92000, and this is falsé.'''
UpperCAmelCase_ = tokenizer.tokenize(_lowercase)
UpperCAmelCase_ = rust_tokenizer.tokenize(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = tokenizer.encode(_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :Optional[Any]) -> List[str]:
UpperCAmelCase_ = BigBirdTokenizer(_lowercase , keep_accents=_lowercase)
UpperCAmelCase_ = tokenizer.tokenize('''This is a test''')
self.assertListEqual(_lowercase , ['''▁This''', '''▁is''', '''▁a''', '''▁t''', '''est'''])
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_lowercase) , [285, 46, 10, 170, 382] , )
UpperCAmelCase_ = tokenizer.tokenize('''I was born in 92000, and this is falsé.''')
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''9''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''é''',
'''.''',
] , )
UpperCAmelCase_ = tokenizer.convert_tokens_to_ids(_lowercase)
self.assertListEqual(
_lowercase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
UpperCAmelCase_ = tokenizer.convert_ids_to_tokens(_lowercase)
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''<unk>''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''<unk>''',
'''.''',
] , )
@cached_property
def __a ( self :Any) -> List[Any]:
return BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
@slow
def __a ( self :int) -> List[Any]:
UpperCAmelCase_ = '''Hello World!'''
UpperCAmelCase_ = [65, 18536, 2260, 101, 66]
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@slow
def __a ( self :int) -> Any:
UpperCAmelCase_ = (
'''This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) " [ ] ! : - . Also we will'''
''' add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth'''
)
# fmt: off
UpperCAmelCase_ = [65, 871, 419, 358, 946, 991, 2521, 452, 358, 1357, 387, 7751, 3536, 112, 985, 456, 126, 865, 938, 5400, 5734, 458, 1368, 467, 786, 2462, 5246, 1159, 633, 865, 4519, 457, 582, 852, 2557, 427, 916, 508, 405, 34324, 497, 391, 408, 11342, 1244, 385, 100, 938, 985, 456, 574, 362, 12597, 3200, 3129, 1172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@require_torch
@slow
def __a ( self :Dict) -> Union[str, Any]:
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
UpperCAmelCase_ = list(self.big_tokenizer.get_vocab().keys())[:10]
UpperCAmelCase_ = ''' '''.join(_lowercase)
UpperCAmelCase_ = self.big_tokenizer.encode_plus(_lowercase , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = self.big_tokenizer.batch_encode_plus(
[sequence + ''' ''' + sequence] , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = BigBirdConfig(attention_type='''original_full''')
UpperCAmelCase_ = BigBirdModel(_lowercase)
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_lowercase)
model(**_lowercase)
@slow
def __a ( self :Optional[int]) -> Any:
UpperCAmelCase_ = BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
UpperCAmelCase_ = tokenizer.decode(tokenizer('''Paris is the [MASK].''').input_ids)
self.assertTrue(decoded_text == '''[CLS] Paris is the[MASK].[SEP]''')
@slow
def __a ( self :Dict) -> List[str]:
# fmt: off
UpperCAmelCase_ = {'''input_ids''': [[65, 39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114, 66], [65, 448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], '''attention_mask''': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_lowercase , model_name='''google/bigbird-roberta-base''' , revision='''215c99f1600e06f83acce68422f2035b2b5c3510''' , )
| 344 | 1 |
import os
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = "▁"
UpperCamelCase_ = {"vocab_file": "spiece.model"}
UpperCamelCase_ = {
"vocab_file": {
"google/reformer-crime-and-punishment": (
"https://huggingface.co/google/reformer-crime-and-punishment/resolve/main/spiece.model"
)
}
}
UpperCamelCase_ = {
"google/reformer-crime-and-punishment": 524_288,
}
class a_ ( _snake_case ):
UpperCamelCase__ : Tuple =VOCAB_FILES_NAMES
UpperCamelCase__ : Any =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : int =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Any =["input_ids", "attention_mask"]
def __init__( self :List[str] , _lowercase :List[str] , _lowercase :List[Any]="</s>" , _lowercase :List[str]="<unk>" , _lowercase :Dict=[] , _lowercase :Optional[Dict[str, Any]] = None , **_lowercase :Optional[int] , ) -> None:
UpperCAmelCase_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
eos_token=_lowercase , unk_token=_lowercase , additional_special_tokens=_lowercase , sp_model_kwargs=self.sp_model_kwargs , **_lowercase , )
UpperCAmelCase_ = vocab_file
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(_lowercase)
@property
def __a ( self :Any) -> List[Any]:
return self.sp_model.get_piece_size()
def __a ( self :Any) -> Dict[str, int]:
UpperCAmelCase_ = {self.convert_ids_to_tokens(_lowercase): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def __getstate__( self :List[str]) -> int:
UpperCAmelCase_ = self.__dict__.copy()
UpperCAmelCase_ = None
return state
def __setstate__( self :Optional[Any] , _lowercase :Union[str, Any]) -> Dict:
UpperCAmelCase_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs'''):
UpperCAmelCase_ = {}
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(self.vocab_file)
def __a ( self :List[str] , _lowercase :str) -> List[str]:
return self.sp_model.encode(_lowercase , out_type=_lowercase)
def __a ( self :Tuple , _lowercase :Tuple) -> Tuple:
return self.sp_model.piece_to_id(_lowercase)
def __a ( self :Optional[Any] , _lowercase :Tuple) -> List[str]:
if index < self.sp_model.get_piece_size():
UpperCAmelCase_ = self.sp_model.IdToPiece(_lowercase)
return token
def __a ( self :Any , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = []
UpperCAmelCase_ = ''''''
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
out_string += self.sp_model.decode(_lowercase) + token
UpperCAmelCase_ = []
else:
current_sub_tokens.append(_lowercase)
out_string += self.sp_model.decode(_lowercase)
return out_string.strip()
def __a ( self :Optional[Any] , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
if not os.path.isdir(_lowercase):
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
return
UpperCAmelCase_ = os.path.join(
_lowercase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''])
if os.path.abspath(self.vocab_file) != os.path.abspath(_lowercase) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , _lowercase)
elif not os.path.isfile(self.vocab_file):
with open(_lowercase , '''wb''') as fi:
UpperCAmelCase_ = self.sp_model.serialized_model_proto()
fi.write(_lowercase)
return (out_vocab_file,)
| 344 |
from __future__ import annotations
import string
from itertools import cycle, product
from pathlib import Path
UpperCamelCase_ = (
string.ascii_letters + string.digits + string.punctuation + string.whitespace
)
UpperCamelCase_ = [ord(letter) for letter in string.ascii_lowercase]
UpperCamelCase_ = {ord(char) for char in VALID_CHARS}
UpperCamelCase_ = ["the", "be", "to", "of", "and", "in", "that", "have"]
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str | None:
'''simple docstring'''
UpperCAmelCase_ = ""
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
for keychar, cipherchar in zip(cycle(__UpperCAmelCase ) , __UpperCAmelCase ):
UpperCAmelCase_ = cipherchar ^ keychar
if decodedchar not in VALID_INTS:
return None
decoded += chr(__UpperCAmelCase )
return decoded
def A ( __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
for key in product(__UpperCAmelCase , repeat=3 ):
UpperCAmelCase_ = try_key(__UpperCAmelCase , __UpperCAmelCase )
if encoded is not None:
possibles.append(__UpperCAmelCase )
return possibles
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
return [possible for possible in possibles if common_word in possible.lower()]
def A ( __UpperCAmelCase = "p059_cipher.txt" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = Path(__UpperCAmelCase ).parent.joinpath(__UpperCAmelCase ).read_text(encoding='''utf-8''' )
UpperCAmelCase_ = [int(__UpperCAmelCase ) for number in data.strip().split(''',''' )]
UpperCAmelCase_ = filter_valid_chars(__UpperCAmelCase )
for common_word in COMMON_WORDS:
UpperCAmelCase_ = filter_common_word(__UpperCAmelCase , __UpperCAmelCase )
if len(__UpperCAmelCase ) == 1:
break
UpperCAmelCase_ = possibles[0]
return sum(ord(__UpperCAmelCase ) for char in decoded_text )
if __name__ == "__main__":
print(f"{solution() = }")
| 344 | 1 |
import argparse
import json
import requests
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import ConvNextConfig, SegformerImageProcessor, UperNetConfig, UperNetForSemanticSegmentation
def A ( __UpperCAmelCase ) -> Dict:
'''simple docstring'''
UpperCAmelCase_ = 384
if "tiny" in model_name:
UpperCAmelCase_ = [3, 3, 9, 3]
UpperCAmelCase_ = [96, 192, 384, 768]
if "small" in model_name:
UpperCAmelCase_ = [3, 3, 27, 3]
UpperCAmelCase_ = [96, 192, 384, 768]
if "base" in model_name:
UpperCAmelCase_ = [3, 3, 27, 3]
UpperCAmelCase_ = [128, 256, 512, 1024]
UpperCAmelCase_ = 512
if "large" in model_name:
UpperCAmelCase_ = [3, 3, 27, 3]
UpperCAmelCase_ = [192, 384, 768, 1536]
UpperCAmelCase_ = 768
if "xlarge" in model_name:
UpperCAmelCase_ = [3, 3, 27, 3]
UpperCAmelCase_ = [256, 512, 1024, 2048]
UpperCAmelCase_ = 1024
# set label information
UpperCAmelCase_ = 150
UpperCAmelCase_ = '''huggingface/label-files'''
UpperCAmelCase_ = '''ade20k-id2label.json'''
UpperCAmelCase_ = json.load(open(hf_hub_download(__UpperCAmelCase , __UpperCAmelCase , repo_type='''dataset''' ) , '''r''' ) )
UpperCAmelCase_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
UpperCAmelCase_ = {v: k for k, v in idalabel.items()}
UpperCAmelCase_ = ConvNextConfig(
depths=__UpperCAmelCase , hidden_sizes=__UpperCAmelCase , out_features=['''stage1''', '''stage2''', '''stage3''', '''stage4'''] )
UpperCAmelCase_ = UperNetConfig(
backbone_config=__UpperCAmelCase , auxiliary_in_channels=__UpperCAmelCase , num_labels=__UpperCAmelCase , idalabel=__UpperCAmelCase , labelaid=__UpperCAmelCase , )
return config
def A ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
UpperCAmelCase_ = []
# fmt: off
# stem
rename_keys.append(('''backbone.downsample_layers.0.0.weight''', '''backbone.embeddings.patch_embeddings.weight''') )
rename_keys.append(('''backbone.downsample_layers.0.0.bias''', '''backbone.embeddings.patch_embeddings.bias''') )
rename_keys.append(('''backbone.downsample_layers.0.1.weight''', '''backbone.embeddings.layernorm.weight''') )
rename_keys.append(('''backbone.downsample_layers.0.1.bias''', '''backbone.embeddings.layernorm.bias''') )
# stages
for i in range(len(config.backbone_config.depths ) ):
for j in range(config.backbone_config.depths[i] ):
rename_keys.append((f"backbone.stages.{i}.{j}.gamma", f"backbone.encoder.stages.{i}.layers.{j}.layer_scale_parameter") )
rename_keys.append((f"backbone.stages.{i}.{j}.depthwise_conv.weight", f"backbone.encoder.stages.{i}.layers.{j}.dwconv.weight") )
rename_keys.append((f"backbone.stages.{i}.{j}.depthwise_conv.bias", f"backbone.encoder.stages.{i}.layers.{j}.dwconv.bias") )
rename_keys.append((f"backbone.stages.{i}.{j}.norm.weight", f"backbone.encoder.stages.{i}.layers.{j}.layernorm.weight") )
rename_keys.append((f"backbone.stages.{i}.{j}.norm.bias", f"backbone.encoder.stages.{i}.layers.{j}.layernorm.bias") )
rename_keys.append((f"backbone.stages.{i}.{j}.pointwise_conv1.weight", f"backbone.encoder.stages.{i}.layers.{j}.pwconv1.weight") )
rename_keys.append((f"backbone.stages.{i}.{j}.pointwise_conv1.bias", f"backbone.encoder.stages.{i}.layers.{j}.pwconv1.bias") )
rename_keys.append((f"backbone.stages.{i}.{j}.pointwise_conv2.weight", f"backbone.encoder.stages.{i}.layers.{j}.pwconv2.weight") )
rename_keys.append((f"backbone.stages.{i}.{j}.pointwise_conv2.bias", f"backbone.encoder.stages.{i}.layers.{j}.pwconv2.bias") )
if i > 0:
rename_keys.append((f"backbone.downsample_layers.{i}.0.weight", f"backbone.encoder.stages.{i}.downsampling_layer.0.weight") )
rename_keys.append((f"backbone.downsample_layers.{i}.0.bias", f"backbone.encoder.stages.{i}.downsampling_layer.0.bias") )
rename_keys.append((f"backbone.downsample_layers.{i}.1.weight", f"backbone.encoder.stages.{i}.downsampling_layer.1.weight") )
rename_keys.append((f"backbone.downsample_layers.{i}.1.bias", f"backbone.encoder.stages.{i}.downsampling_layer.1.bias") )
rename_keys.append((f"backbone.norm{i}.weight", f"backbone.hidden_states_norms.stage{i+1}.weight") )
rename_keys.append((f"backbone.norm{i}.bias", f"backbone.hidden_states_norms.stage{i+1}.bias") )
# decode head
rename_keys.extend(
[
('''decode_head.conv_seg.weight''', '''decode_head.classifier.weight'''),
('''decode_head.conv_seg.bias''', '''decode_head.classifier.bias'''),
('''auxiliary_head.conv_seg.weight''', '''auxiliary_head.classifier.weight'''),
('''auxiliary_head.conv_seg.bias''', '''auxiliary_head.classifier.bias'''),
] )
# fmt: on
return rename_keys
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Tuple:
'''simple docstring'''
UpperCAmelCase_ = dct.pop(__UpperCAmelCase )
UpperCAmelCase_ = val
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> str:
'''simple docstring'''
UpperCAmelCase_ = {
'''upernet-convnext-tiny''': '''https://download.openmmlab.com/mmsegmentation/v0.5/convnext/upernet_convnext_tiny_fp16_512x512_160k_ade20k/upernet_convnext_tiny_fp16_512x512_160k_ade20k_20220227_124553-cad485de.pth''',
'''upernet-convnext-small''': '''https://download.openmmlab.com/mmsegmentation/v0.5/convnext/upernet_convnext_small_fp16_512x512_160k_ade20k/upernet_convnext_small_fp16_512x512_160k_ade20k_20220227_131208-1b1e394f.pth''',
'''upernet-convnext-base''': '''https://download.openmmlab.com/mmsegmentation/v0.5/convnext/upernet_convnext_base_fp16_512x512_160k_ade20k/upernet_convnext_base_fp16_512x512_160k_ade20k_20220227_181227-02a24fc6.pth''',
'''upernet-convnext-large''': '''https://download.openmmlab.com/mmsegmentation/v0.5/convnext/upernet_convnext_large_fp16_640x640_160k_ade20k/upernet_convnext_large_fp16_640x640_160k_ade20k_20220226_040532-e57aa54d.pth''',
'''upernet-convnext-xlarge''': '''https://download.openmmlab.com/mmsegmentation/v0.5/convnext/upernet_convnext_xlarge_fp16_640x640_160k_ade20k/upernet_convnext_xlarge_fp16_640x640_160k_ade20k_20220226_080344-95fc38c2.pth''',
}
UpperCAmelCase_ = model_name_to_url[model_name]
UpperCAmelCase_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase , map_location='''cpu''' )['''state_dict''']
UpperCAmelCase_ = get_upernet_config(__UpperCAmelCase )
UpperCAmelCase_ = UperNetForSemanticSegmentation(__UpperCAmelCase )
model.eval()
# replace "bn" => "batch_norm"
for key in state_dict.copy().keys():
UpperCAmelCase_ = state_dict.pop(__UpperCAmelCase )
if "bn" in key:
UpperCAmelCase_ = key.replace('''bn''' , '''batch_norm''' )
UpperCAmelCase_ = val
# rename keys
UpperCAmelCase_ = create_rename_keys(__UpperCAmelCase )
for src, dest in rename_keys:
rename_key(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
# verify on image
UpperCAmelCase_ = '''https://huggingface.co/datasets/hf-internal-testing/fixtures_ade20k/resolve/main/ADE_val_00000001.jpg'''
UpperCAmelCase_ = Image.open(requests.get(__UpperCAmelCase , stream=__UpperCAmelCase ).raw ).convert('''RGB''' )
UpperCAmelCase_ = SegformerImageProcessor()
UpperCAmelCase_ = processor(__UpperCAmelCase , return_tensors='''pt''' ).pixel_values
with torch.no_grad():
UpperCAmelCase_ = model(__UpperCAmelCase )
if model_name == "upernet-convnext-tiny":
UpperCAmelCase_ = torch.tensor(
[[-8.8_110, -8.8_110, -8.6_521], [-8.8_110, -8.8_110, -8.6_521], [-8.7_746, -8.7_746, -8.6_130]] )
elif model_name == "upernet-convnext-small":
UpperCAmelCase_ = torch.tensor(
[[-8.8_236, -8.8_236, -8.6_771], [-8.8_236, -8.8_236, -8.6_771], [-8.7_638, -8.7_638, -8.6_240]] )
elif model_name == "upernet-convnext-base":
UpperCAmelCase_ = torch.tensor(
[[-8.8_558, -8.8_558, -8.6_905], [-8.8_558, -8.8_558, -8.6_905], [-8.7_669, -8.7_669, -8.6_021]] )
elif model_name == "upernet-convnext-large":
UpperCAmelCase_ = torch.tensor(
[[-8.6_660, -8.6_660, -8.6_210], [-8.6_660, -8.6_660, -8.6_210], [-8.6_310, -8.6_310, -8.5_964]] )
elif model_name == "upernet-convnext-xlarge":
UpperCAmelCase_ = torch.tensor(
[[-8.4_980, -8.4_980, -8.3_977], [-8.4_980, -8.4_980, -8.3_977], [-8.4_379, -8.4_379, -8.3_412]] )
print('''Logits:''' , outputs.logits[0, 0, :3, :3] )
assert torch.allclose(outputs.logits[0, 0, :3, :3] , __UpperCAmelCase , atol=1e-4 )
print('''Looks ok!''' )
if pytorch_dump_folder_path is not None:
print(f"Saving model {model_name} to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
print(f"Saving processor to {pytorch_dump_folder_path}" )
processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print(f"Pushing model and processor for {model_name} to hub" )
model.push_to_hub(f"openmmlab/{model_name}" )
processor.push_to_hub(f"openmmlab/{model_name}" )
if __name__ == "__main__":
UpperCamelCase_ = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--model_name",
default="upernet-convnext-tiny",
type=str,
choices=[f"upernet-convnext-{size}" for size in ["tiny", "small", "base", "large", "xlarge"]],
help="Name of the ConvNext UperNet model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
)
UpperCamelCase_ = parser.parse_args()
convert_upernet_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 344 |
import pytest
UpperCamelCase_ = "__dummy_dataset1__"
UpperCamelCase_ = "\nimport json\nimport os\n\nimport datasets\n\n\nREPO_URL = \"https://huggingface.co/datasets/albertvillanova/tests-raw-jsonl/resolve/main/\"\nURLS = {\"train\": REPO_URL + \"wikiann-bn-train.jsonl\", \"validation\": REPO_URL + \"wikiann-bn-validation.jsonl\"}\n\n\nclass __DummyDataset1__(datasets.GeneratorBasedBuilder):\n\n def _info(self):\n features = datasets.Features(\n {\n \"tokens\": datasets.Sequence(datasets.Value(\"string\")),\n \"ner_tags\": datasets.Sequence(\n datasets.features.ClassLabel(\n names=[\n \"O\",\n \"B-PER\",\n \"I-PER\",\n \"B-ORG\",\n \"I-ORG\",\n \"B-LOC\",\n \"I-LOC\",\n ]\n )\n ),\n \"langs\": datasets.Sequence(datasets.Value(\"string\")),\n \"spans\": datasets.Sequence(datasets.Value(\"string\")),\n }\n )\n return datasets.DatasetInfo(features=features)\n\n def _split_generators(self, dl_manager):\n dl_path = dl_manager.download(URLS)\n return [\n datasets.SplitGenerator(datasets.Split.TRAIN, gen_kwargs={\"filepath\": dl_path[\"train\"]}),\n datasets.SplitGenerator(datasets.Split.VALIDATION, gen_kwargs={\"filepath\": dl_path[\"validation\"]}),\n ]\n\n def _generate_examples(self, filepath):\n with open(filepath, \"r\", encoding=\"utf-8\") as f:\n for i, line in enumerate(f):\n yield i, json.loads(line)\n"
@pytest.fixture
def A ( ) -> str:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_NAME
@pytest.fixture
def A ( ) -> Any:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_CODE
@pytest.fixture
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = dataset_loading_script_name
UpperCAmelCase_ = tmp_path / '''datasets''' / script_name
script_dir.mkdir(parents=__UpperCAmelCase )
UpperCAmelCase_ = script_dir / f"{script_name}.py"
with open(__UpperCAmelCase , '''w''' ) as f:
f.write(__UpperCAmelCase )
return str(__UpperCAmelCase )
| 344 | 1 |
import os
from dataclasses import dataclass, field
from io import BytesIO
from typing import TYPE_CHECKING, Any, ClassVar, Dict, Optional, Union
import numpy as np
import pyarrow as pa
from .. import config
from ..download.streaming_download_manager import xopen, xsplitext
from ..table import array_cast
from ..utils.py_utils import no_op_if_value_is_null, string_to_dict
if TYPE_CHECKING:
from .features import FeatureType
UpperCamelCase_ , UpperCamelCase_ , UpperCamelCase_ = False, False, False
@dataclass
class a_ :
UpperCamelCase__ : Optional[int] =None
UpperCamelCase__ : bool =True
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[str] =None
# Automatically constructed
UpperCamelCase__ : ClassVar[str] ="dict"
UpperCamelCase__ : ClassVar[Any] =pa.struct({"bytes": pa.binary(), "path": pa.string()} )
UpperCamelCase__ : str =field(default="Audio" , init=_snake_case , repr=_snake_case )
def __call__( self :List[Any]) -> List[Any]:
return self.pa_type
def __a ( self :Any , _lowercase :Union[str, bytes, dict]) -> dict:
try:
import soundfile as sf # soundfile is a dependency of librosa, needed to decode audio files.
except ImportError as err:
raise ImportError('''To support encoding audio data, please install \'soundfile\'.''') from err
if isinstance(_lowercase , _lowercase):
return {"bytes": None, "path": value}
elif isinstance(_lowercase , _lowercase):
return {"bytes": value, "path": None}
elif "array" in value:
# convert the audio array to wav bytes
UpperCAmelCase_ = BytesIO()
sf.write(_lowercase , value['''array'''] , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
elif value.get('''path''') is not None and os.path.isfile(value['''path''']):
# we set "bytes": None to not duplicate the data if they're already available locally
if value["path"].endswith('''pcm'''):
# "PCM" only has raw audio bytes
if value.get('''sampling_rate''') is None:
# At least, If you want to convert "PCM-byte" to "WAV-byte", you have to know sampling rate
raise KeyError('''To use PCM files, please specify a \'sampling_rate\' in Audio object''')
if value.get('''bytes'''):
# If we already had PCM-byte, we don`t have to make "read file, make bytes" (just use it!)
UpperCAmelCase_ = np.frombuffer(value['''bytes'''] , dtype=np.intaa).astype(np.floataa) / 32767
else:
UpperCAmelCase_ = np.memmap(value['''path'''] , dtype='''h''' , mode='''r''').astype(np.floataa) / 32767
UpperCAmelCase_ = BytesIO(bytes())
sf.write(_lowercase , _lowercase , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
else:
return {"bytes": None, "path": value.get('''path''')}
elif value.get('''bytes''') is not None or value.get('''path''') is not None:
# store the audio bytes, and path is used to infer the audio format using the file extension
return {"bytes": value.get('''bytes'''), "path": value.get('''path''')}
else:
raise ValueError(
f"An audio sample should have one of 'path' or 'bytes' but they are missing or None in {value}.")
def __a ( self :Dict , _lowercase :dict , _lowercase :Optional[Dict[str, Union[str, bool, None]]] = None) -> dict:
if not self.decode:
raise RuntimeError('''Decoding is disabled for this feature. Please use Audio(decode=True) instead.''')
UpperCAmelCase_ , UpperCAmelCase_ = (value['''path'''], BytesIO(value['''bytes'''])) if value['''bytes'''] is not None else (value['''path'''], None)
if path is None and file is None:
raise ValueError(f"An audio sample should have one of 'path' or 'bytes' but both are None in {value}.")
try:
import librosa
import soundfile as sf
except ImportError as err:
raise ImportError('''To support decoding audio files, please install \'librosa\' and \'soundfile\'.''') from err
UpperCAmelCase_ = xsplitext(_lowercase)[1][1:].lower() if path is not None else None
if not config.IS_OPUS_SUPPORTED and audio_format == "opus":
raise RuntimeError(
'''Decoding \'opus\' files requires system library \'libsndfile\'>=1.0.31, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
elif not config.IS_MP3_SUPPORTED and audio_format == "mp3":
raise RuntimeError(
'''Decoding \'mp3\' files requires system library \'libsndfile\'>=1.1.0, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
if file is None:
UpperCAmelCase_ = token_per_repo_id or {}
UpperCAmelCase_ = path.split('''::''')[-1]
try:
UpperCAmelCase_ = string_to_dict(_lowercase , config.HUB_DATASETS_URL)['''repo_id''']
UpperCAmelCase_ = token_per_repo_id[repo_id]
except (ValueError, KeyError):
UpperCAmelCase_ = None
with xopen(_lowercase , '''rb''' , use_auth_token=_lowercase) as f:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
else:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
UpperCAmelCase_ = array.T
if self.mono:
UpperCAmelCase_ = librosa.to_mono(_lowercase)
if self.sampling_rate and self.sampling_rate != sampling_rate:
UpperCAmelCase_ = librosa.resample(_lowercase , orig_sr=_lowercase , target_sr=self.sampling_rate)
UpperCAmelCase_ = self.sampling_rate
return {"path": path, "array": array, "sampling_rate": sampling_rate}
def __a ( self :Union[str, Any]) -> Union["FeatureType", Dict[str, "FeatureType"]]:
from .features import Value
if self.decode:
raise ValueError('''Cannot flatten a decoded Audio feature.''')
return {
"bytes": Value('''binary'''),
"path": Value('''string'''),
}
def __a ( self :int , _lowercase :Union[pa.StringArray, pa.StructArray]) -> pa.StructArray:
if pa.types.is_string(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, storage] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_binary(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([storage, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_struct(storage.type) and storage.type.get_all_field_indices('''array'''):
UpperCAmelCase_ = pa.array([Audio().encode_example(_lowercase) if x is not None else None for x in storage.to_pylist()])
elif pa.types.is_struct(storage.type):
if storage.type.get_field_index('''bytes''') >= 0:
UpperCAmelCase_ = storage.field('''bytes''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
if storage.type.get_field_index('''path''') >= 0:
UpperCAmelCase_ = storage.field('''path''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
return array_cast(_lowercase , self.pa_type)
def __a ( self :Any , _lowercase :pa.StructArray) -> pa.StructArray:
@no_op_if_value_is_null
def path_to_bytes(_lowercase :Tuple):
with xopen(_lowercase , '''rb''') as f:
UpperCAmelCase_ = f.read()
return bytes_
UpperCAmelCase_ = pa.array(
[
(path_to_bytes(x['''path''']) if x['''bytes'''] is None else x['''bytes''']) if x is not None else None
for x in storage.to_pylist()
] , type=pa.binary() , )
UpperCAmelCase_ = pa.array(
[os.path.basename(_lowercase) if path is not None else None for path in storage.field('''path''').to_pylist()] , type=pa.string() , )
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"s-JoL/Open-Llama-V1": "https://huggingface.co/s-JoL/Open-Llama-V1/blob/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="open-llama"
def __init__( self :Union[str, Any] , _lowercase :List[Any]=100000 , _lowercase :Dict=4096 , _lowercase :List[Any]=11008 , _lowercase :Optional[int]=32 , _lowercase :Union[str, Any]=32 , _lowercase :List[str]="silu" , _lowercase :Union[str, Any]=2048 , _lowercase :Any=0.02 , _lowercase :Optional[Any]=1E-6 , _lowercase :str=True , _lowercase :str=0 , _lowercase :Any=1 , _lowercase :Optional[Any]=2 , _lowercase :str=False , _lowercase :Dict=True , _lowercase :Optional[Any]=0.1 , _lowercase :Tuple=0.1 , _lowercase :Dict=True , _lowercase :List[Any]=True , _lowercase :Dict=None , **_lowercase :Optional[int] , ) -> List[Any]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = rms_norm_eps
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = kwargs.pop(
'''use_memorry_efficient_attention''' , _lowercase)
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_dropout_prob
UpperCAmelCase_ = use_stable_embedding
UpperCAmelCase_ = shared_input_output_embedding
UpperCAmelCase_ = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , tie_word_embeddings=_lowercase , **_lowercase , )
def __a ( self :int) -> str:
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _lowercase) or len(self.rope_scaling) != 2:
raise ValueError(
'''`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, '''
f"got {self.rope_scaling}")
UpperCAmelCase_ = self.rope_scaling.get('''type''' , _lowercase)
UpperCAmelCase_ = self.rope_scaling.get('''factor''' , _lowercase)
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}")
if rope_scaling_factor is None or not isinstance(_lowercase , _lowercase) or rope_scaling_factor <= 1.0:
raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
| 344 | 1 |
def A ( __UpperCAmelCase = 100_0000 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = [i - 1 for i in range(limit + 1 )]
for i in range(2 , limit + 1 ):
if phi[i] == i - 1:
for j in range(2 * i , limit + 1 , __UpperCAmelCase ):
phi[j] -= phi[j] // i
return sum(phi[2 : limit + 1] )
if __name__ == "__main__":
print(solution())
| 344 |
import tempfile
import torch
from diffusers import (
DEISMultistepScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
UniPCMultistepScheduler,
)
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] =(DPMSolverSinglestepScheduler,)
UpperCamelCase__ : Tuple =(("num_inference_steps", 25),)
def __a ( self :List[Any] , **_lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
'''solver_order''': 2,
'''prediction_type''': '''epsilon''',
'''thresholding''': False,
'''sample_max_value''': 1.0,
'''algorithm_type''': '''dpmsolver++''',
'''solver_type''': '''midpoint''',
'''lambda_min_clipped''': -float('''inf'''),
'''variance_type''': None,
}
config.update(**_lowercase)
return config
def __a ( self :Union[str, Any] , _lowercase :List[Any]=0 , **_lowercase :Optional[int]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ , UpperCAmelCase_ = sample, sample
for t in range(_lowercase , time_step + scheduler.config.solver_order + 1):
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Union[str, Any]) -> List[Any]:
pass
def __a ( self :Optional[Any] , _lowercase :str=0 , **_lowercase :Union[str, Any]) -> Dict:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Dict , _lowercase :Union[str, Any]=None , **_lowercase :List[Any]) -> int:
if scheduler is None:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :int) -> Tuple:
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = 50
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
# make sure that the first t is uneven
for i, t in enumerate(scheduler.timesteps[3:]):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_574) < 1E-3
def __a ( self :List[Any]) -> List[Any]:
for timesteps in [25, 50, 100, 999, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :int) -> Optional[Any]:
# make sure that iterating over schedulers with same config names gives same results
# for defaults
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
UpperCAmelCase_ = DEISMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = UniPCMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverSinglestepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Tuple) -> int:
self.check_over_configs(thresholding=_lowercase)
for order in [1, 2, 3]:
for solver_type in ["midpoint", "heun"]:
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
thresholding=_lowercase , prediction_type=_lowercase , sample_max_value=_lowercase , algorithm_type='''dpmsolver++''' , solver_order=_lowercase , solver_type=_lowercase , )
def __a ( self :List[Any]) -> Any:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Optional[int]:
for algorithm_type in ["dpmsolver", "dpmsolver++"]:
for solver_type in ["midpoint", "heun"]:
for order in [1, 2, 3]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
UpperCAmelCase_ = self.full_loop(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
assert not torch.isnan(_lowercase).any(), "Samples have nan numbers"
def __a ( self :Tuple) -> int:
self.check_over_configs(lower_order_final=_lowercase)
self.check_over_configs(lower_order_final=_lowercase)
def __a ( self :Tuple) -> Optional[Any]:
self.check_over_configs(lambda_min_clipped=-float('''inf'''))
self.check_over_configs(lambda_min_clipped=-5.1)
def __a ( self :Any) -> List[str]:
self.check_over_configs(variance_type=_lowercase)
self.check_over_configs(variance_type='''learned_range''')
def __a ( self :Any) -> Dict:
for num_inference_steps in [1, 2, 3, 5, 10, 50, 100, 999, 1000]:
self.check_over_forward(num_inference_steps=_lowercase , time_step=0)
def __a ( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Any) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop(use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_248) < 1E-3
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.1_453) < 1E-3
def __a ( self :List[Any]) -> Dict:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''' , use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.0_649) < 1E-3
def __a ( self :Any) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(thresholding=_lowercase , dynamic_thresholding_ratio=0)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter.half()
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
assert sample.dtype == torch.floataa
| 344 | 1 |
from __future__ import annotations
from cmath import sqrt
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> tuple[complex, complex]:
'''simple docstring'''
if a == 0:
raise ValueError('''Coefficient \'a\' must not be zero.''' )
UpperCAmelCase_ = b * b - 4 * a * c
UpperCAmelCase_ = (-b + sqrt(__UpperCAmelCase )) / (2 * a)
UpperCAmelCase_ = (-b - sqrt(__UpperCAmelCase )) / (2 * a)
return (
root_a.real if not root_a.imag else root_a,
root_a.real if not root_a.imag else root_a,
)
def A ( ) -> Any:
'''simple docstring'''
UpperCAmelCase_ , UpperCAmelCase_ = quadratic_roots(a=5 , b=6 , c=1 )
print(f"The solutions are: {solutiona} and {solutiona}" )
if __name__ == "__main__":
main()
| 344 |
import inspect
import unittest
import torch
import torch.nn as nn
from accelerate.hooks import (
AlignDevicesHook,
ModelHook,
SequentialHook,
add_hook_to_module,
attach_align_device_hook,
remove_hook_from_module,
remove_hook_from_submodules,
)
from accelerate.test_utils import require_multi_gpu
class a_ ( nn.Module ):
def __init__( self :Optional[Any]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(3 , 4)
UpperCAmelCase_ = nn.BatchNormad(4)
UpperCAmelCase_ = nn.Linear(4 , 5)
def __a ( self :Dict , _lowercase :int) -> str:
return self.lineara(self.batchnorm(self.lineara(_lowercase)))
class a_ ( _snake_case ):
def __a ( self :Tuple , _lowercase :Optional[int] , *_lowercase :Union[str, Any] , **_lowercase :Any) -> Optional[Any]:
return (args[0] + 1,) + args[1:], kwargs
class a_ ( _snake_case ):
def __a ( self :Union[str, Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
return output + 1
class a_ ( unittest.TestCase ):
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
self.assertEqual(test_model._hf_hook , _lowercase)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[Any]) -> Any:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
add_hook_to_module(_lowercase , _lowercase , append=_lowercase)
self.assertEqual(isinstance(test_model._hf_hook , _lowercase) , _lowercase)
self.assertEqual(len(test_model._hf_hook.hooks) , 2)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[int]) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(x + 1)
UpperCAmelCase_ = test_model(x + 2)
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PreForwardHook() , PreForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , _lowercase , atol=1E-5)
def __a ( self :List[str]) -> int:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PostForwardHook() , PostForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , output + 2 , atol=1E-5)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1))
self.assertTrue(outputa.requires_grad)
UpperCAmelCase_ = True
UpperCAmelCase_ = test_model(_lowercase)
self.assertFalse(outputa.requires_grad)
@require_multi_gpu
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.batchnorm , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=1))
self.assertEqual(model.lineara.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.running_mean.device , torch.device(0))
self.assertEqual(model.lineara.weight.device , torch.device(1))
# We can still make a forward pass. The input does not need to be on any particular device
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(1))
# We can add a general hook to put back output on same device as input.
add_hook_to_module(_lowercase , AlignDevicesHook(io_same_device=_lowercase))
UpperCAmelCase_ = torch.randn(2 , 3).to(0)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(0))
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = {'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''', '''offload''': True}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(hook_kwargs['''execution_device'''])
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
UpperCAmelCase_ = {
'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''',
'''offload''': True,
'''offload_buffers''': True,
}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase)
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase , offload_buffers=_lowercase)
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict())
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict() , offload_buffers=_lowercase , )
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
| 344 | 1 |
import importlib
import json
import os
import sys
import tempfile
import unittest
from pathlib import Path
import transformers
import transformers.models.auto
from transformers.models.auto.configuration_auto import CONFIG_MAPPING, AutoConfig
from transformers.models.bert.configuration_bert import BertConfig
from transformers.models.roberta.configuration_roberta import RobertaConfig
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, get_tests_dir
sys.path.append(str(Path(__file__).parent.parent.parent.parent / "utils"))
from test_module.custom_configuration import CustomConfig # noqa E402
UpperCamelCase_ = get_tests_dir("fixtures/dummy-config.json")
class a_ ( unittest.TestCase ):
def __a ( self :Tuple) -> str:
UpperCAmelCase_ = 0
def __a ( self :Union[str, Any]) -> Optional[int]:
self.assertIsNotNone(transformers.models.auto.__spec__)
self.assertIsNotNone(importlib.util.find_spec('''transformers.models.auto'''))
def __a ( self :str) -> int:
UpperCAmelCase_ = AutoConfig.from_pretrained('''bert-base-uncased''')
self.assertIsInstance(_lowercase , _lowercase)
def __a ( self :List[str]) -> Optional[Any]:
UpperCAmelCase_ = AutoConfig.from_pretrained(_lowercase)
self.assertIsInstance(_lowercase , _lowercase)
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = AutoConfig.from_pretrained(_lowercase)
self.assertIsInstance(_lowercase , _lowercase)
def __a ( self :Optional[int]) -> Tuple:
UpperCAmelCase_ = AutoConfig.for_model('''roberta''')
self.assertIsInstance(_lowercase , _lowercase)
def __a ( self :Any) -> int:
with tempfile.TemporaryDirectory() as tmp_dir:
# This model name contains bert and roberta, but roberta ends up being picked.
UpperCAmelCase_ = os.path.join(_lowercase , '''fake-roberta''')
os.makedirs(_lowercase , exist_ok=_lowercase)
with open(os.path.join(_lowercase , '''config.json''') , '''w''') as f:
f.write(json.dumps({}))
UpperCAmelCase_ = AutoConfig.from_pretrained(_lowercase)
self.assertEqual(type(_lowercase) , _lowercase)
def __a ( self :Any) -> Any:
try:
AutoConfig.register('''custom''' , _lowercase)
# Wrong model type will raise an error
with self.assertRaises(_lowercase):
AutoConfig.register('''model''' , _lowercase)
# Trying to register something existing in the Transformers library will raise an error
with self.assertRaises(_lowercase):
AutoConfig.register('''bert''' , _lowercase)
# Now that the config is registered, it can be used as any other config with the auto-API
UpperCAmelCase_ = CustomConfig()
with tempfile.TemporaryDirectory() as tmp_dir:
config.save_pretrained(_lowercase)
UpperCAmelCase_ = AutoConfig.from_pretrained(_lowercase)
self.assertIsInstance(_lowercase , _lowercase)
finally:
if "custom" in CONFIG_MAPPING._extra_content:
del CONFIG_MAPPING._extra_content["custom"]
def __a ( self :Any) -> Optional[Any]:
with self.assertRaisesRegex(
_lowercase , '''bert-base is not a local folder and is not a valid model identifier'''):
UpperCAmelCase_ = AutoConfig.from_pretrained('''bert-base''')
def __a ( self :Optional[int]) -> int:
with self.assertRaisesRegex(
_lowercase , r'''aaaaaa is not a valid git identifier \(branch name, tag name or commit id\)'''):
UpperCAmelCase_ = AutoConfig.from_pretrained(_lowercase , revision='''aaaaaa''')
def __a ( self :str) -> Optional[int]:
with self.assertRaisesRegex(
_lowercase , '''hf-internal-testing/no-config-test-repo does not appear to have a file named config.json.''' , ):
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/no-config-test-repo''')
def __a ( self :Tuple) -> Dict:
# If remote code is not set, we will time out when asking whether to load the model.
with self.assertRaises(_lowercase):
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/test_dynamic_model''')
# If remote code is disabled, we can't load this config.
with self.assertRaises(_lowercase):
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/test_dynamic_model''' , trust_remote_code=_lowercase)
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/test_dynamic_model''' , trust_remote_code=_lowercase)
self.assertEqual(config.__class__.__name__ , '''NewModelConfig''')
# Test config can be reloaded.
with tempfile.TemporaryDirectory() as tmp_dir:
config.save_pretrained(_lowercase)
UpperCAmelCase_ = AutoConfig.from_pretrained(_lowercase , trust_remote_code=_lowercase)
self.assertEqual(reloaded_config.__class__.__name__ , '''NewModelConfig''')
def __a ( self :Any) -> Optional[Any]:
class a_ ( _snake_case ):
UpperCamelCase__ : Any ="new-model"
try:
AutoConfig.register('''new-model''' , _lowercase)
# If remote code is not set, the default is to use local
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/test_dynamic_model''')
self.assertEqual(config.__class__.__name__ , '''NewModelConfigLocal''')
# If remote code is disabled, we load the local one.
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/test_dynamic_model''' , trust_remote_code=_lowercase)
self.assertEqual(config.__class__.__name__ , '''NewModelConfigLocal''')
# If remote is enabled, we load from the Hub
UpperCAmelCase_ = AutoConfig.from_pretrained('''hf-internal-testing/test_dynamic_model''' , trust_remote_code=_lowercase)
self.assertEqual(config.__class__.__name__ , '''NewModelConfig''')
finally:
if "new-model" in CONFIG_MAPPING._extra_content:
del CONFIG_MAPPING._extra_content["new-model"]
| 344 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import (
AutoProcessor,
BertTokenizerFast,
BlipImageProcessor,
GPTaTokenizer,
InstructBlipProcessor,
PreTrainedTokenizerFast,
)
@require_vision
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = BlipImageProcessor()
UpperCAmelCase_ = GPTaTokenizer.from_pretrained('''hf-internal-testing/tiny-random-GPT2Model''')
UpperCAmelCase_ = BertTokenizerFast.from_pretrained('''hf-internal-testing/tiny-random-bert''')
UpperCAmelCase_ = InstructBlipProcessor(_lowercase , _lowercase , _lowercase)
processor.save_pretrained(self.tmpdirname)
def __a ( self :List[Any] , **_lowercase :Dict) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).tokenizer
def __a ( self :Optional[Any] , **_lowercase :Optional[Any]) -> Optional[int]:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).image_processor
def __a ( self :Dict , **_lowercase :Tuple) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).qformer_tokenizer
def __a ( self :Optional[int]) -> str:
shutil.rmtree(self.tmpdirname)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta)]
UpperCAmelCase_ = [Image.fromarray(np.moveaxis(_lowercase , 0 , -1)) for x in image_inputs]
return image_inputs
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() , qformer_tokenizer=self.get_qformer_tokenizer() , )
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = self.get_image_processor(do_normalize=_lowercase , padding_value=1.0)
UpperCAmelCase_ = InstructBlipProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowercase , padding_value=1.0)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
self.assertIsInstance(processor.tokenizer , _lowercase)
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string())
self.assertIsInstance(processor.image_processor , _lowercase)
self.assertIsInstance(processor.qformer_tokenizer , _lowercase)
def __a ( self :Dict) -> Any:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = image_processor(_lowercase , return_tensors='''np''')
UpperCAmelCase_ = processor(images=_lowercase , return_tensors='''np''')
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1E-2)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = processor(text=_lowercase)
UpperCAmelCase_ = tokenizer(_lowercase , return_token_type_ids=_lowercase)
UpperCAmelCase_ = qformer_tokenizer(_lowercase , return_token_type_ids=_lowercase)
for key in encoded_tokens.keys():
self.assertListEqual(encoded_tokens[key] , encoded_processor[key])
for key in encoded_tokens_qformer.keys():
self.assertListEqual(encoded_tokens_qformer[key] , encoded_processor['''qformer_''' + key])
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
# test if it raises when no input is passed
with pytest.raises(_lowercase):
processor()
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
UpperCAmelCase_ = processor.batch_decode(_lowercase)
UpperCAmelCase_ = tokenizer.batch_decode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :str) -> int:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
| 344 | 1 |
from __future__ import annotations
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Any: # noqa: E741
'''simple docstring'''
while r - l > 1:
UpperCAmelCase_ = (l + r) // 2
if v[m] >= key:
UpperCAmelCase_ = m
else:
UpperCAmelCase_ = m # noqa: E741
return r
def A ( __UpperCAmelCase ) -> int:
'''simple docstring'''
if len(__UpperCAmelCase ) == 0:
return 0
UpperCAmelCase_ = [0] * len(__UpperCAmelCase )
UpperCAmelCase_ = 1
UpperCAmelCase_ = v[0]
for i in range(1 , len(__UpperCAmelCase ) ):
if v[i] < tail[0]:
UpperCAmelCase_ = v[i]
elif v[i] > tail[length - 1]:
UpperCAmelCase_ = v[i]
length += 1
else:
UpperCAmelCase_ = v[i]
return length
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"facebook/levit-128S": "https://huggingface.co/facebook/levit-128S/resolve/main/config.json",
# See all LeViT models at https://huggingface.co/models?filter=levit
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[int] ="levit"
def __init__( self :List[str] , _lowercase :List[Any]=224 , _lowercase :str=3 , _lowercase :Optional[int]=3 , _lowercase :str=2 , _lowercase :List[Any]=1 , _lowercase :str=16 , _lowercase :Dict=[128, 256, 384] , _lowercase :Union[str, Any]=[4, 8, 12] , _lowercase :Tuple=[4, 4, 4] , _lowercase :Dict=[16, 16, 16] , _lowercase :Any=0 , _lowercase :Dict=[2, 2, 2] , _lowercase :Any=[2, 2, 2] , _lowercase :Tuple=0.02 , **_lowercase :Union[str, Any] , ) -> Optional[Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = kernel_size
UpperCAmelCase_ = stride
UpperCAmelCase_ = padding
UpperCAmelCase_ = hidden_sizes
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = depths
UpperCAmelCase_ = key_dim
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = attention_ratio
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = [
['''Subsample''', key_dim[0], hidden_sizes[0] // key_dim[0], 4, 2, 2],
['''Subsample''', key_dim[0], hidden_sizes[1] // key_dim[0], 4, 2, 2],
]
class a_ ( _snake_case ):
UpperCamelCase__ : Union[str, Any] =version.parse("1.11" )
@property
def __a ( self :Any) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
])
@property
def __a ( self :List[Any]) -> float:
return 1E-4
| 344 | 1 |
import unittest
from transformers import JukeboxTokenizer
from transformers.testing_utils import require_torch
class a_ ( unittest.TestCase ):
UpperCamelCase__ : Dict =JukeboxTokenizer
UpperCamelCase__ : Dict ={
"artist": "Zac Brown Band",
"genres": "Country",
"lyrics": "I met a traveller from an antique land,\n Who said \"Two vast and trunkless legs of stone\n Stand in the desert. . . . Near them, on the sand,\n Half sunk a shattered visage lies, whose frown,\n And wrinkled lip, and sneer of cold command,\n Tell that its sculptor well those passions read\n Which yet survive, stamped on these lifeless things,\n The hand that mocked them, and the heart that fed;\n And on the pedestal, these words appear:\n My name is Ozymandias, King of Kings;\n Look on my Works, ye Mighty, and despair!\n Nothing beside remains. Round the decay\n Of that colossal Wreck, boundless and bare\n The lone and level sands stretch far away\n ",
}
@require_torch
def __a ( self :Dict) -> Tuple:
import torch
UpperCAmelCase_ = JukeboxTokenizer.from_pretrained('''openai/jukebox-1b-lyrics''')
UpperCAmelCase_ = tokenizer(**self.metas)['''input_ids''']
# fmt: off
UpperCAmelCase_ = [
torch.tensor([[
0, 0, 0, 7169, 507, 9, 76, 39, 31, 46, 76, 27,
76, 46, 44, 27, 48, 31, 38, 38, 31, 44, 76, 32,
44, 41, 39, 76, 27, 40, 76, 27, 40, 46, 35, 43,
47, 31, 76, 38, 27, 40, 30, 64, 78, 76, 76, 76,
76, 76, 76, 76, 76, 23, 34, 41, 76, 45, 27, 35,
30, 76, 71, 20, 49, 41, 76, 48, 27, 45, 46, 76,
27, 40, 30, 76, 46, 44, 47, 40, 37, 38, 31, 45,
45, 76, 38, 31, 33, 45, 76, 41, 32, 76, 45, 46,
41, 40, 31, 78, 76, 76, 76, 76, 76, 76, 76, 76,
19, 46, 27, 40, 30, 76, 35, 40, 76, 46, 34, 31,
76, 30, 31, 45, 31, 44, 46, 63, 76, 63, 76, 63,
76, 63, 76, 14, 31, 27, 44, 76, 46, 34, 31, 39,
64, 76, 41, 40, 76, 46, 34, 31, 76, 45, 27, 40,
30, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76, 8,
27, 38, 32, 76, 45, 47, 40, 37, 76, 27, 76, 45,
34, 27, 46, 46, 31, 44, 31, 30, 76, 48, 35, 45,
27, 33, 31, 76, 38, 35, 31, 45, 64, 76, 49, 34,
41, 45, 31, 76, 32, 44, 41, 49, 40, 64, 78, 76,
76, 76, 76, 76, 76, 76, 76, 1, 40, 30, 76, 49,
44, 35, 40, 37, 38, 31, 30, 76, 38, 35, 42, 64,
76, 27, 40, 30, 76, 45, 40, 31, 31, 44, 76, 41,
32, 76, 29, 41, 38, 30, 76, 29, 41, 39, 39, 27,
40, 30, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76,
20, 31, 38, 38, 76, 46, 34, 27, 46, 76, 35, 46,
45, 76, 45, 29, 47, 38, 42, 46, 41, 44, 76, 49,
31, 38, 38, 76, 46, 34, 41, 45, 31, 76, 42, 27,
45, 45, 35, 41, 40, 45, 76, 44, 31, 27, 30, 78,
76, 76, 76, 76, 76, 76, 76, 76, 23, 34, 35, 29,
34, 76, 51, 31, 46, 76, 45, 47, 44, 48, 35, 48,
31, 64, 76, 45, 46, 27, 39, 42, 31, 30, 76, 41,
40, 76, 46, 34, 31, 45, 31, 76, 38, 35, 32, 31,
38, 31, 45, 45, 76, 46, 34, 35, 40, 33, 45, 64,
78, 76, 76, 76, 76, 76, 76, 76, 76, 20, 34, 31,
76, 34, 27, 40, 30, 76, 46, 34, 27, 46, 76, 39,
41, 29, 37, 31, 30, 76, 46, 34, 31, 39, 64, 76,
27, 40, 30, 76, 46, 34, 31, 76, 34, 31, 27, 44,
46, 76, 46, 34, 27, 46, 76, 32, 31, 30, 66, 78,
76, 76, 76, 76, 76, 76, 76, 76, 1, 40, 30, 76,
41, 40, 76, 46, 34, 31, 76, 42, 31, 30, 31, 45,
46, 27, 38, 64, 76, 46, 34, 31, 45, 31, 76, 49,
41, 44, 30, 45, 76, 27, 42, 42, 31, 27, 44, 65,
78, 76, 76, 76, 76, 76, 76, 76, 76, 13, 51, 76,
40, 27, 39, 31, 76, 35, 45, 76, 15, 52, 51, 39,
27, 40, 30, 35, 27, 45, 64, 76, 11, 35, 40, 33,
76, 41, 32, 76, 11, 35, 40, 33, 45, 66, 78, 76,
76, 76, 76, 76, 76, 76, 76, 12, 41, 41, 37, 76,
41, 40, 76, 39, 51, 76, 23, 41, 44, 37, 45, 64,
76, 51, 31, 76, 13, 35, 33, 34, 46, 51, 64, 76,
27, 40, 30, 76, 30, 31, 45, 42, 27, 35, 44, 67,
78, 76, 76, 76, 76, 76, 76, 76, 76, 14, 41, 46,
34, 35, 40, 33, 76, 28, 31, 45, 35, 30, 31, 76,
44, 31, 39, 27, 35, 40, 45, 63, 76, 18, 41, 47,
40, 30, 76, 46, 34, 31, 76, 30, 31, 29, 27, 51,
78, 76, 76, 76, 76, 76, 76, 76, 76, 15, 32, 76,
46, 34, 27, 46, 76, 29, 41, 38, 41, 45, 45, 27,
38, 76, 23, 44, 31, 29, 37, 64, 76, 28, 41, 47,
40, 30, 38, 31, 45, 45, 76, 27, 40, 30, 76, 28,
27, 44, 31, 78, 76, 76, 76, 76, 76, 76, 76, 76,
20, 34, 31, 76, 38, 41, 40, 31, 76, 27, 40, 30,
76, 38, 31, 48, 31, 38, 76, 45, 27, 40, 30, 45,
76, 45, 46, 44, 31, 46, 29, 34, 76, 32, 27, 44,
76, 27, 49, 27, 51, 78, 76, 76, 76, 76, 76, 76,
76, 76]]),
torch.tensor([[0, 0, 0, 1069, 11]]),
torch.tensor([[0, 0, 0, 1069, 11]]),
]
# fmt: on
self.assertTrue(torch.allclose(tokens[0] , EXPECTED_OUTPUT[0]))
self.assertTrue(torch.allclose(tokens[1] , EXPECTED_OUTPUT[1]))
self.assertTrue(torch.allclose(tokens[2] , EXPECTED_OUTPUT[2]))
@require_torch
def __a ( self :List[Any]) -> str:
import torch
UpperCAmelCase_ = JukeboxTokenizer.from_pretrained('''openai/jukebox-5b-lyrics''')
UpperCAmelCase_ = tokenizer(**self.metas)['''input_ids''']
# fmt: off
UpperCAmelCase_ = [
torch.tensor([[
0, 0, 0, 1069, 11, -1, -1, -1, -1, 9, 77, 39,
31, 46, 77, 27, 77, 46, 44, 27, 48, 31, 38, 38,
31, 44, 77, 32, 44, 41, 39, 77, 27, 40, 77, 27,
40, 46, 35, 43, 47, 31, 77, 38, 27, 40, 30, 64,
79, 77, 77, 77, 77, 77, 77, 77, 77, 23, 34, 41,
77, 45, 27, 35, 30, 77, 72, 20, 49, 41, 77, 48,
27, 45, 46, 77, 27, 40, 30, 77, 46, 44, 47, 40,
37, 38, 31, 45, 45, 77, 38, 31, 33, 45, 77, 41,
32, 77, 45, 46, 41, 40, 31, 79, 77, 77, 77, 77,
77, 77, 77, 77, 19, 46, 27, 40, 30, 77, 35, 40,
77, 46, 34, 31, 77, 30, 31, 45, 31, 44, 46, 63,
77, 63, 77, 63, 77, 63, 77, 14, 31, 27, 44, 77,
46, 34, 31, 39, 64, 77, 41, 40, 77, 46, 34, 31,
77, 45, 27, 40, 30, 64, 79, 77, 77, 77, 77, 77,
77, 77, 77, 8, 27, 38, 32, 77, 45, 47, 40, 37,
77, 27, 77, 45, 34, 27, 46, 46, 31, 44, 31, 30,
77, 48, 35, 45, 27, 33, 31, 77, 38, 35, 31, 45,
64, 77, 49, 34, 41, 45, 31, 77, 32, 44, 41, 49,
40, 64, 79, 77, 77, 77, 77, 77, 77, 77, 77, 1,
40, 30, 77, 49, 44, 35, 40, 37, 38, 31, 30, 77,
38, 35, 42, 64, 77, 27, 40, 30, 77, 45, 40, 31,
31, 44, 77, 41, 32, 77, 29, 41, 38, 30, 77, 29,
41, 39, 39, 27, 40, 30, 64, 79, 77, 77, 77, 77,
77, 77, 77, 77, 20, 31, 38, 38, 77, 46, 34, 27,
46, 77, 35, 46, 45, 77, 45, 29, 47, 38, 42, 46,
41, 44, 77, 49, 31, 38, 38, 77, 46, 34, 41, 45,
31, 77, 42, 27, 45, 45, 35, 41, 40, 45, 77, 44,
31, 27, 30, 79, 77, 77, 77, 77, 77, 77, 77, 77,
23, 34, 35, 29, 34, 77, 51, 31, 46, 77, 45, 47,
44, 48, 35, 48, 31, 64, 77, 45, 46, 27, 39, 42,
31, 30, 77, 41, 40, 77, 46, 34, 31, 45, 31, 77,
38, 35, 32, 31, 38, 31, 45, 45, 77, 46, 34, 35,
40, 33, 45, 64, 79, 77, 77, 77, 77, 77, 77, 77,
77, 20, 34, 31, 77, 34, 27, 40, 30, 77, 46, 34,
27, 46, 77, 39, 41, 29, 37, 31, 30, 77, 46, 34,
31, 39, 64, 77, 27, 40, 30, 77, 46, 34, 31, 77,
34, 31, 27, 44, 46, 77, 46, 34, 27, 46, 77, 32,
31, 30, 66, 79, 77, 77, 77, 77, 77, 77, 77, 77,
1, 40, 30, 77, 41, 40, 77, 46, 34, 31, 77, 42,
31, 30, 31, 45, 46, 27, 38, 64, 77, 46, 34, 31,
45, 31, 77, 49, 41, 44, 30, 45, 77, 27, 42, 42,
31, 27, 44, 65, 79, 77, 77, 77, 77, 77, 77, 77,
77, 13, 51, 77, 40, 27, 39, 31, 77, 35, 45, 77,
15, 52, 51, 39, 27, 40, 30, 35, 27, 45, 64, 77,
11, 35, 40, 33, 77, 41, 32, 77, 11, 35, 40, 33,
45, 66, 79, 77, 77, 77, 77, 77, 77, 77, 77, 12,
41, 41, 37, 77, 41, 40, 77, 39, 51, 77, 23, 41,
44, 37, 45, 64, 77, 51, 31, 77, 13, 35, 33, 34,
46, 51, 64, 77, 27, 40, 30, 77, 30, 31, 45, 42,
27, 35, 44, 67, 79, 77, 77, 77, 77, 77, 77, 77,
77, 14, 41, 46, 34, 35, 40, 33, 77, 28, 31, 45,
35, 30, 31, 77, 44, 31, 39, 27, 35, 40, 45, 63,
77, 18, 41, 47, 40, 30, 77, 46, 34, 31, 77, 30,
31, 29, 27, 51, 79, 77, 77, 77, 77, 77, 77, 77,
77, 15, 32, 77, 46, 34, 27, 46, 77, 29, 41, 38,
41, 45, 45, 27, 38, 77, 23, 44, 31, 29, 37, 64,
77, 28, 41, 47, 40, 30, 38, 31, 45, 45, 77, 27,
40, 30, 77, 28, 27, 44, 31, 79, 77, 77, 77, 77,
77, 77, 77, 77, 20, 34, 31, 77, 38, 41, 40, 31,
77, 27, 40, 30, 77, 38, 31, 48, 31, 38, 77, 45,
27, 40, 30, 45, 77, 45, 46, 44, 31, 46, 29, 34,
77, 32, 27, 44, 77, 27, 49, 27, 51, 79, 77, 77,
77, 77, 77, 77, 77, 77]]),
torch.tensor([[0, 0, 0, 1069, 11, -1, -1, -1, -1]]),
torch.tensor([[0, 0, 0, 1069, 11, -1, -1, -1, -1]]),
]
# fmt: on
self.assertTrue(torch.allclose(tokens[0] , EXPECTED_OUTPUT[0]))
self.assertTrue(torch.allclose(tokens[1] , EXPECTED_OUTPUT[1]))
self.assertTrue(torch.allclose(tokens[2] , EXPECTED_OUTPUT[2]))
| 344 |
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=None , __UpperCAmelCase="no" , __UpperCAmelCase="29500" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = False
UpperCAmelCase_ = False
if any(key.startswith('''KAGGLE''' ) for key in os.environ.keys() ):
UpperCAmelCase_ = True
elif "IPython" in sys.modules:
UpperCAmelCase_ = '''google.colab''' in str(sys.modules['''IPython'''].get_ipython() )
try:
UpperCAmelCase_ = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
f"Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}." )
if (in_colab or in_kaggle) and (os.environ.get('''TPU_NAME''' , __UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '''
'''your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if num_processes is None:
UpperCAmelCase_ = 8
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''TPU''' )
print(f"Launching a training on {num_processes} TPU cores." )
xmp.spawn(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on one CPU.''' )
function(*__UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'''You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.''' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '''
'''inside your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if torch.cuda.is_initialized():
raise ValueError(
'''To launch a multi-GPU training from your notebook, you need to avoid running any instruction '''
'''using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '''
'''function.''' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port=__UpperCAmelCase , mixed_precision=__UpperCAmelCase ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''MULTI_GPU''' )
print(f"Launching training on {num_processes} GPUs." )
try:
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'''CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '''
'''This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '''
'''Please review your imports and test them when running the `notebook_launcher()` to identify '''
'''which one is problematic.''' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
UpperCAmelCase_ = '''1'''
print('''Launching training on MPS.''' )
elif torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on CPU.''' )
function(*__UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=2 ) -> Optional[Any]:
'''simple docstring'''
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port='''29500''' , accelerate_mixed_precision='''no''' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='''yes''' , ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , debug=__UpperCAmelCase )
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
| 344 | 1 |
import inspect
import unittest
import torch
import torch.nn as nn
from accelerate.hooks import (
AlignDevicesHook,
ModelHook,
SequentialHook,
add_hook_to_module,
attach_align_device_hook,
remove_hook_from_module,
remove_hook_from_submodules,
)
from accelerate.test_utils import require_multi_gpu
class a_ ( nn.Module ):
def __init__( self :Optional[Any]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(3 , 4)
UpperCAmelCase_ = nn.BatchNormad(4)
UpperCAmelCase_ = nn.Linear(4 , 5)
def __a ( self :Dict , _lowercase :int) -> str:
return self.lineara(self.batchnorm(self.lineara(_lowercase)))
class a_ ( _snake_case ):
def __a ( self :Tuple , _lowercase :Optional[int] , *_lowercase :Union[str, Any] , **_lowercase :Any) -> Optional[Any]:
return (args[0] + 1,) + args[1:], kwargs
class a_ ( _snake_case ):
def __a ( self :Union[str, Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
return output + 1
class a_ ( unittest.TestCase ):
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
self.assertEqual(test_model._hf_hook , _lowercase)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[Any]) -> Any:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
add_hook_to_module(_lowercase , _lowercase , append=_lowercase)
self.assertEqual(isinstance(test_model._hf_hook , _lowercase) , _lowercase)
self.assertEqual(len(test_model._hf_hook.hooks) , 2)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[int]) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(x + 1)
UpperCAmelCase_ = test_model(x + 2)
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PreForwardHook() , PreForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , _lowercase , atol=1E-5)
def __a ( self :List[str]) -> int:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PostForwardHook() , PostForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , output + 2 , atol=1E-5)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1))
self.assertTrue(outputa.requires_grad)
UpperCAmelCase_ = True
UpperCAmelCase_ = test_model(_lowercase)
self.assertFalse(outputa.requires_grad)
@require_multi_gpu
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.batchnorm , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=1))
self.assertEqual(model.lineara.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.running_mean.device , torch.device(0))
self.assertEqual(model.lineara.weight.device , torch.device(1))
# We can still make a forward pass. The input does not need to be on any particular device
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(1))
# We can add a general hook to put back output on same device as input.
add_hook_to_module(_lowercase , AlignDevicesHook(io_same_device=_lowercase))
UpperCAmelCase_ = torch.randn(2 , 3).to(0)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(0))
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = {'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''', '''offload''': True}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(hook_kwargs['''execution_device'''])
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
UpperCAmelCase_ = {
'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''',
'''offload''': True,
'''offload_buffers''': True,
}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase)
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase , offload_buffers=_lowercase)
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict())
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict() , offload_buffers=_lowercase , )
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
| 344 |
import math
import os
import re
import sys
import unittest
from pathlib import Path
from typing import Tuple
from unittest.mock import patch
from parameterized import parameterized
from transformers.testing_utils import (
CaptureStderr,
ExtendSysPath,
TestCasePlus,
execute_subprocess_async,
get_gpu_count,
get_torch_dist_unique_port,
require_apex,
require_bitsandbytes,
require_fairscale,
require_torch,
require_torch_gpu,
require_torch_multi_gpu,
require_torch_non_multi_gpu,
slow,
)
from transformers.trainer_callback import TrainerState
from transformers.trainer_utils import set_seed
UpperCamelCase_ = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f"{bindir}/../../examples/pytorch/translation"):
from run_translation import main # noqa
set_seed(42)
UpperCamelCase_ = "sshleifer/student_marian_en_ro_6_1"
UpperCamelCase_ = "sshleifer/tiny-mbart"
@require_torch
class a_ ( _snake_case ):
def __a ( self :str , _lowercase :Any=False , _lowercase :Tuple=None , _lowercase :Dict=True , _lowercase :Tuple=True , _lowercase :List[Any]=True , _lowercase :List[str]=True , ) -> int:
UpperCAmelCase_ = self.run_trainer(
eval_steps=1 , max_len=12 , model_name=_lowercase , num_train_epochs=1 , distributed=_lowercase , extra_args_str=_lowercase , predict_with_generate=_lowercase , do_train=_lowercase , do_eval=_lowercase , do_predict=_lowercase , )
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
if not do_eval:
return
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
if predict_with_generate:
assert "eval_bleu" in first_step_stats
UpperCAmelCase_ = eval_metrics[-1]
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
assert not math.isnan(float(last_step_stats['''eval_loss'''])), "eval_loss must not be `nan`"
@require_torch_non_multi_gpu
def __a ( self :Dict) -> str:
self.run_seqaseq_quick()
@require_torch_multi_gpu
def __a ( self :Any) -> int:
self.run_seqaseq_quick(distributed=_lowercase)
@require_torch_multi_gpu
def __a ( self :int) -> Any:
self.run_seqaseq_quick(distributed=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> List[str]:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple --fp16''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Union[str, Any]) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :int) -> Any:
self.run_seqaseq_quick(
distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=_lowercase)
@require_apex
@require_torch_gpu
def __a ( self :Tuple) -> str:
# XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same
# program and it breaks other tests that run from the same pytest worker, therefore until this is
# sorted out it must be run only in an external program, that is distributed=True in this
# test and only under one or more gpus - if we want cpu will need to make a special test
#
# specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via
# 2nd main() call it botches the future eval.
#
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
# test 2nd time - was getting eval_loss': nan'
# to reproduce the problem set distributed=False
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
@parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''])
@require_torch_multi_gpu
def __a ( self :str , _lowercase :Any) -> List[str]:
# as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout
UpperCAmelCase_ = {
# test with the default log_level - should be info and thus log info once
'''base''': {'''extra_args_str''': '''''', '''n_matches''': 1},
# test with low log_level and log_level_replica - should be noisy on all processes
# now the info string should appear twice on 2 processes
'''low''': {'''extra_args_str''': '''--log_level debug --log_level_replica debug''', '''n_matches''': 2},
# test with high log_level and low log_level_replica
# now the info string should appear once only on the replica
'''high''': {'''extra_args_str''': '''--log_level error --log_level_replica debug''', '''n_matches''': 1},
# test with high log_level and log_level_replica - should be quiet on all processes
'''mixed''': {'''extra_args_str''': '''--log_level error --log_level_replica error''', '''n_matches''': 0},
}
UpperCAmelCase_ = experiments[experiment_id]
UpperCAmelCase_ = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False}
UpperCAmelCase_ = '''Running training'''
with CaptureStderr() as cl:
self.run_seqaseq_quick(**_lowercase , extra_args_str=data['''extra_args_str'''])
UpperCAmelCase_ = len(re.findall(_lowercase , cl.err))
self.assertEqual(_lowercase , data['''n_matches'''])
@slow
def __a ( self :Any) -> Dict:
UpperCAmelCase_ = self.run_trainer(
eval_steps=2 , max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=10 , distributed=_lowercase , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
UpperCAmelCase_ = eval_metrics[-1]
assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing"
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
# test if do_predict saves generations and metrics
UpperCAmelCase_ = os.listdir(_lowercase)
UpperCAmelCase_ = {os.path.basename(_lowercase) for p in contents}
assert "generated_predictions.txt" in contents
assert "predict_results.json" in contents
@slow
@require_bitsandbytes
def __a ( self :List[str]) -> str:
from transformers.training_args import OptimizerNames
def train_and_return_metrics(_lowercase :str) -> Tuple[int, float]:
UpperCAmelCase_ = '''--skip_memory_metrics 0'''
UpperCAmelCase_ = self.run_trainer(
max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=1 , optim=_lowercase , distributed=_lowercase , extra_args_str=_lowercase , do_eval=_lowercase , do_predict=_lowercase , n_gpus_to_use=1 , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(Path(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20)
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20)
UpperCAmelCase_ = logs[0]['''train_loss''']
return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value)
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value)
UpperCAmelCase_ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_peak_mem_orig + gpu_alloc_mem_orig
UpperCAmelCase_ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_total_mem_orig - gpu_total_mem_bnb
# sshleifer/student_marian_en_ro_6_1 has 54M parameter, 29M of which is `nn.Embedding` which
# doesn't get quantized and remains in fp32. Therefore we only have 25M parameters quantized
# in 2 bytes and the diff in optim memory usage is derived as so:
#
# - normal 25*8=~200MB (8 bytes per param)
# - bnb 25*2= ~50MB (2 bytes per param)
#
# Thus we should expect ~150MB total memory saved.
#
# Peak memory should be the same - the total should be different by about that same margin
#
# After leaving a small margin to accommodate for differences between gpus let's check
# that we have at least 120MB in savings
UpperCAmelCase_ = 120
# uncomment the following if this test starts failing - requires py38 for a new print feature
# gpu_peak_mem_diff = gpu_peak_mem_orig - gpu_peak_mem_bnb
# print(f"{gpu_alloc_mem_orig=}MB {gpu_peak_mem_orig=}MB {gpu_alloc_mem_orig+gpu_peak_mem_orig=}MB")
# print(f" {gpu_alloc_mem_bnb=}MB {gpu_peak_mem_bnb=}MB {gpu_alloc_mem_bnb+gpu_peak_mem_bnb=}MB")
# print(f"{gpu_alloc_mem_diff=}MB")
# print(f"{gpu_peak_mem_diff=}MB")
# print(f"{gpu_total_mem_orig=}MB, {gpu_total_mem_bnb=}MB")
# print(f"{gpu_total_mem_diff=}MB, {gpu_total_mem_diff=}MB")
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less alloc gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_alloc_mem_diff}MB, with gpu_alloc_mem_orig={gpu_alloc_mem_orig}MB and"
f" gpu_alloc_mem_bnb={gpu_alloc_mem_bnb}MB" , )
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less total gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_total_mem_diff}MB, with gpu_total_mem_orig={gpu_total_mem_orig}MB and"
f" gpu_total_mem_bnb={gpu_total_mem_bnb}MB" , )
self.assertEqual(
_lowercase , _lowercase , f"loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}")
def __a ( self :Any , _lowercase :int , _lowercase :str , _lowercase :int , _lowercase :float = 3E-3 , _lowercase :str = "adafactor" , _lowercase :bool = False , _lowercase :str = None , _lowercase :int = 0 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :int = None , ) -> List[Any]:
UpperCAmelCase_ = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro'''
UpperCAmelCase_ = self.get_auto_remove_tmp_dir()
UpperCAmelCase_ = f"\n --model_name_or_path {model_name}\n --train_file {data_dir}/train.json\n --validation_file {data_dir}/val.json\n --test_file {data_dir}/test.json\n --output_dir {output_dir}\n --overwrite_output_dir\n --max_train_samples 8\n --max_source_length {max_len}\n --max_target_length {max_len}\n --do_train\n --num_train_epochs {str(_lowercase)}\n --per_device_train_batch_size 4\n --learning_rate {learning_rate}\n --warmup_steps 8\n --logging_steps 0\n --logging_strategy no\n --save_steps {str(_lowercase)}\n --group_by_length\n --label_smoothing_factor 0.1\n --target_lang ro_RO\n --source_lang en_XX\n ".split()
UpperCAmelCase_ = f"\n --do_eval\n --per_device_eval_batch_size 4\n --max_eval_samples 8\n --val_max_target_length {max_len}\n --evaluation_strategy steps\n --eval_steps {str(_lowercase)}\n ".split()
UpperCAmelCase_ = '''
--do_predict
'''.split()
UpperCAmelCase_ = []
if do_train:
args += args_train
if do_eval:
args += args_eval
if do_predict:
args += args_predict
if predict_with_generate:
args += "--predict_with_generate".split()
if do_train:
if optim == "adafactor":
args += "--adafactor".split()
else:
args += f"--optim {optim}".split()
if extra_args_str is not None:
args += extra_args_str.split()
if distributed:
if n_gpus_to_use is None:
UpperCAmelCase_ = get_gpu_count()
UpperCAmelCase_ = get_torch_dist_unique_port()
UpperCAmelCase_ = f"\n -m torch.distributed.run\n --nproc_per_node={n_gpus_to_use}\n --master_port={master_port}\n {self.examples_dir_str}/pytorch/translation/run_translation.py\n ".split()
UpperCAmelCase_ = [sys.executable] + distributed_args + args
# keep for quick debug
# print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die
execute_subprocess_async(_lowercase , env=self.get_env())
else:
UpperCAmelCase_ = ['''run_translation.py'''] + args
with patch.object(_lowercase , '''argv''' , _lowercase):
main()
return output_dir
| 344 | 1 |
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
if index == number_of_items:
return 0
UpperCAmelCase_ = 0
UpperCAmelCase_ = 0
UpperCAmelCase_ = knapsack(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , index + 1 )
if weights[index] <= max_weight:
UpperCAmelCase_ = values[index] + knapsack(
__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , max_weight - weights[index] , index + 1 )
return max(__UpperCAmelCase , __UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
import functools
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = len(__UpperCAmelCase )
UpperCAmelCase_ = len(__UpperCAmelCase )
@functools.cache
def min_distance(__UpperCAmelCase , __UpperCAmelCase ) -> int:
# if first word index is overflow - delete all from the second word
if indexa >= len_worda:
return len_worda - indexa
# if second word index is overflow - delete all from the first word
if indexa >= len_worda:
return len_worda - indexa
UpperCAmelCase_ = int(worda[indexa] != worda[indexa] ) # current letters not identical
return min(
1 + min_distance(indexa + 1 , __UpperCAmelCase ) , 1 + min_distance(__UpperCAmelCase , indexa + 1 ) , diff + min_distance(indexa + 1 , indexa + 1 ) , )
return min_distance(0 , 0 )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 | 1 |
import unittest
import numpy as np
from transformers.testing_utils import require_pytesseract, require_torch
from transformers.utils import is_pytesseract_available, is_torch_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_pytesseract_available():
from PIL import Image
from transformers import LayoutLMvaImageProcessor
class a_ ( unittest.TestCase ):
def __init__( self :List[str] , _lowercase :List[str] , _lowercase :Optional[int]=7 , _lowercase :Optional[Any]=3 , _lowercase :str=18 , _lowercase :Optional[Any]=30 , _lowercase :Union[str, Any]=400 , _lowercase :Any=True , _lowercase :List[str]=None , _lowercase :Optional[Any]=True , ) -> Union[str, Any]:
UpperCAmelCase_ = size if size is not None else {'''height''': 18, '''width''': 18}
UpperCAmelCase_ = parent
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = image_size
UpperCAmelCase_ = min_resolution
UpperCAmelCase_ = max_resolution
UpperCAmelCase_ = do_resize
UpperCAmelCase_ = size
UpperCAmelCase_ = apply_ocr
def __a ( self :Union[str, Any]) -> Union[str, Any]:
return {"do_resize": self.do_resize, "size": self.size, "apply_ocr": self.apply_ocr}
@require_torch
@require_pytesseract
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : Any =LayoutLMvaImageProcessor if is_pytesseract_available() else None
def __a ( self :Optional[int]) -> Tuple:
UpperCAmelCase_ = LayoutLMvaImageProcessingTester(self)
@property
def __a ( self :str) -> List[Any]:
return self.image_processor_tester.prepare_image_processor_dict()
def __a ( self :Any) -> Dict:
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(_lowercase , '''do_resize'''))
self.assertTrue(hasattr(_lowercase , '''size'''))
self.assertTrue(hasattr(_lowercase , '''apply_ocr'''))
def __a ( self :str) -> Dict:
UpperCAmelCase_ = self.image_processing_class.from_dict(self.image_processor_dict)
self.assertEqual(image_processor.size , {'''height''': 18, '''width''': 18})
UpperCAmelCase_ = self.image_processing_class.from_dict(self.image_processor_dict , size=42)
self.assertEqual(image_processor.size , {'''height''': 42, '''width''': 42})
def __a ( self :Any) -> List[str]:
pass
def __a ( self :List[str]) -> List[str]:
# Initialize image_processing
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , Image.Image)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''')
self.assertEqual(
encoding.pixel_values.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.size['''height'''],
self.image_processor_tester.size['''width'''],
) , )
self.assertIsInstance(encoding.words , _lowercase)
self.assertIsInstance(encoding.boxes , _lowercase)
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size['''height'''],
self.image_processor_tester.size['''width'''],
) , )
def __a ( self :List[Any]) -> Union[str, Any]:
# Initialize image_processing
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , numpify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , np.ndarray)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.size['''height'''],
self.image_processor_tester.size['''width'''],
) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size['''height'''],
self.image_processor_tester.size['''width'''],
) , )
def __a ( self :Dict) -> Tuple:
# Initialize image_processing
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , torchify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , torch.Tensor)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.size['''height'''],
self.image_processor_tester.size['''width'''],
) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.size['''height'''],
self.image_processor_tester.size['''width'''],
) , )
def __a ( self :List[Any]) -> List[Any]:
# with apply_OCR = True
UpperCAmelCase_ = LayoutLMvaImageProcessor()
from datasets import load_dataset
UpperCAmelCase_ = load_dataset('''hf-internal-testing/fixtures_docvqa''' , split='''test''')
UpperCAmelCase_ = Image.open(ds[0]['''file''']).convert('''RGB''')
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''')
self.assertEqual(encoding.pixel_values.shape , (1, 3, 224, 224))
self.assertEqual(len(encoding.words) , len(encoding.boxes))
# fmt: off
# the words and boxes were obtained with Tesseract 4.1.1
UpperCAmelCase_ = [['''11:14''', '''to''', '''11:39''', '''a.m''', '''11:39''', '''to''', '''11:44''', '''a.m.''', '''11:44''', '''a.m.''', '''to''', '''12:25''', '''p.m.''', '''12:25''', '''to''', '''12:58''', '''p.m.''', '''12:58''', '''to''', '''4:00''', '''p.m.''', '''2:00''', '''to''', '''5:00''', '''p.m.''', '''Coffee''', '''Break''', '''Coffee''', '''will''', '''be''', '''served''', '''for''', '''men''', '''and''', '''women''', '''in''', '''the''', '''lobby''', '''adjacent''', '''to''', '''exhibit''', '''area.''', '''Please''', '''move''', '''into''', '''exhibit''', '''area.''', '''(Exhibits''', '''Open)''', '''TRRF''', '''GENERAL''', '''SESSION''', '''(PART''', '''|)''', '''Presiding:''', '''Lee''', '''A.''', '''Waller''', '''TRRF''', '''Vice''', '''President''', '''“Introductory''', '''Remarks”''', '''Lee''', '''A.''', '''Waller,''', '''TRRF''', '''Vice''', '''Presi-''', '''dent''', '''Individual''', '''Interviews''', '''with''', '''TRRF''', '''Public''', '''Board''', '''Members''', '''and''', '''Sci-''', '''entific''', '''Advisory''', '''Council''', '''Mem-''', '''bers''', '''Conducted''', '''by''', '''TRRF''', '''Treasurer''', '''Philip''', '''G.''', '''Kuehn''', '''to''', '''get''', '''answers''', '''which''', '''the''', '''public''', '''refrigerated''', '''warehousing''', '''industry''', '''is''', '''looking''', '''for.''', '''Plus''', '''questions''', '''from''', '''the''', '''floor.''', '''Dr.''', '''Emil''', '''M.''', '''Mrak,''', '''University''', '''of''', '''Cal-''', '''ifornia,''', '''Chairman,''', '''TRRF''', '''Board;''', '''Sam''', '''R.''', '''Cecil,''', '''University''', '''of''', '''Georgia''', '''College''', '''of''', '''Agriculture;''', '''Dr.''', '''Stanley''', '''Charm,''', '''Tufts''', '''University''', '''School''', '''of''', '''Medicine;''', '''Dr.''', '''Robert''', '''H.''', '''Cotton,''', '''ITT''', '''Continental''', '''Baking''', '''Company;''', '''Dr.''', '''Owen''', '''Fennema,''', '''University''', '''of''', '''Wis-''', '''consin;''', '''Dr.''', '''Robert''', '''E.''', '''Hardenburg,''', '''USDA.''', '''Questions''', '''and''', '''Answers''', '''Exhibits''', '''Open''', '''Capt.''', '''Jack''', '''Stoney''', '''Room''', '''TRRF''', '''Scientific''', '''Advisory''', '''Council''', '''Meeting''', '''Ballroom''', '''Foyer''']] # noqa: E231
UpperCAmelCase_ = [[[141, 57, 214, 69], [228, 58, 252, 69], [141, 75, 216, 88], [230, 79, 280, 88], [142, 260, 218, 273], [230, 261, 255, 273], [143, 279, 218, 290], [231, 282, 290, 291], [143, 342, 218, 354], [231, 345, 289, 355], [202, 362, 227, 373], [143, 379, 220, 392], [231, 382, 291, 394], [144, 714, 220, 726], [231, 715, 256, 726], [144, 732, 220, 745], [232, 736, 291, 747], [144, 769, 218, 782], [231, 770, 256, 782], [141, 788, 202, 801], [215, 791, 274, 804], [143, 826, 204, 838], [215, 826, 240, 838], [142, 844, 202, 857], [215, 847, 274, 859], [334, 57, 427, 69], [440, 57, 522, 69], [369, 75, 461, 88], [469, 75, 516, 88], [528, 76, 562, 88], [570, 76, 667, 88], [675, 75, 711, 87], [721, 79, 778, 88], [789, 75, 840, 88], [369, 97, 470, 107], [484, 94, 507, 106], [518, 94, 562, 107], [576, 94, 655, 110], [668, 94, 792, 109], [804, 95, 829, 107], [369, 113, 465, 125], [477, 116, 547, 125], [562, 113, 658, 125], [671, 116, 748, 125], [761, 113, 811, 125], [369, 131, 465, 143], [477, 133, 548, 143], [563, 130, 698, 145], [710, 130, 802, 146], [336, 171, 412, 183], [423, 171, 572, 183], [582, 170, 716, 184], [728, 171, 817, 187], [829, 171, 844, 186], [338, 197, 482, 212], [507, 196, 557, 209], [569, 196, 595, 208], [610, 196, 702, 209], [505, 214, 583, 226], [595, 214, 656, 227], [670, 215, 807, 227], [335, 259, 543, 274], [556, 259, 708, 272], [372, 279, 422, 291], [435, 279, 460, 291], [474, 279, 574, 292], [587, 278, 664, 291], [676, 278, 738, 291], [751, 279, 834, 291], [372, 298, 434, 310], [335, 341, 483, 354], [497, 341, 655, 354], [667, 341, 728, 354], [740, 341, 825, 354], [335, 360, 430, 372], [442, 360, 534, 372], [545, 359, 687, 372], [697, 360, 754, 372], [765, 360, 823, 373], [334, 378, 428, 391], [440, 378, 577, 394], [590, 378, 705, 391], [720, 378, 801, 391], [334, 397, 400, 409], [370, 416, 529, 429], [544, 416, 576, 432], [587, 416, 665, 428], [677, 416, 814, 429], [372, 435, 452, 450], [465, 434, 495, 447], [511, 434, 600, 447], [611, 436, 637, 447], [649, 436, 694, 451], [705, 438, 824, 447], [369, 453, 452, 466], [464, 454, 509, 466], [522, 453, 611, 469], [625, 453, 792, 469], [370, 472, 556, 488], [570, 472, 684, 487], [697, 472, 718, 485], [732, 472, 835, 488], [369, 490, 411, 503], [425, 490, 484, 503], [496, 490, 635, 506], [645, 490, 707, 503], [718, 491, 761, 503], [771, 490, 840, 503], [336, 510, 374, 521], [388, 510, 447, 522], [460, 510, 489, 521], [503, 510, 580, 522], [592, 509, 736, 525], [745, 509, 770, 522], [781, 509, 840, 522], [338, 528, 434, 541], [448, 528, 596, 541], [609, 527, 687, 540], [700, 528, 792, 541], [336, 546, 397, 559], [407, 546, 431, 559], [443, 546, 525, 560], [537, 546, 680, 562], [688, 546, 714, 559], [722, 546, 837, 562], [336, 565, 449, 581], [461, 565, 485, 577], [497, 565, 665, 581], [681, 565, 718, 577], [732, 565, 837, 580], [337, 584, 438, 597], [452, 583, 521, 596], [535, 584, 677, 599], [690, 583, 787, 596], [801, 583, 825, 596], [338, 602, 478, 615], [492, 602, 530, 614], [543, 602, 638, 615], [650, 602, 676, 614], [688, 602, 788, 615], [802, 602, 843, 614], [337, 621, 502, 633], [516, 621, 615, 637], [629, 621, 774, 636], [789, 621, 827, 633], [337, 639, 418, 652], [432, 640, 571, 653], [587, 639, 731, 655], [743, 639, 769, 652], [780, 639, 841, 652], [338, 658, 440, 673], [455, 658, 491, 670], [508, 658, 602, 671], [616, 658, 638, 670], [654, 658, 835, 674], [337, 677, 429, 689], [337, 714, 482, 726], [495, 714, 548, 726], [561, 714, 683, 726], [338, 770, 461, 782], [474, 769, 554, 785], [489, 788, 562, 803], [576, 788, 643, 801], [656, 787, 751, 804], [764, 788, 844, 801], [334, 825, 421, 838], [430, 824, 574, 838], [584, 824, 723, 841], [335, 844, 450, 857], [464, 843, 583, 860], [628, 862, 755, 875], [769, 861, 848, 878]]] # noqa: E231
# fmt: on
self.assertListEqual(encoding.words , _lowercase)
self.assertListEqual(encoding.boxes , _lowercase)
# with apply_OCR = False
UpperCAmelCase_ = LayoutLMvaImageProcessor(apply_ocr=_lowercase)
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''')
self.assertEqual(encoding.pixel_values.shape , (1, 3, 224, 224))
| 344 |
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import SPIECE_UNDERLINE, logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {"vocab_file": "spiece.model"}
UpperCamelCase_ = {
"vocab_file": {
"xlnet-base-cased": "https://huggingface.co/xlnet-base-cased/resolve/main/spiece.model",
"xlnet-large-cased": "https://huggingface.co/xlnet-large-cased/resolve/main/spiece.model",
}
}
UpperCamelCase_ = {
"xlnet-base-cased": None,
"xlnet-large-cased": None,
}
# Segments (not really needed)
UpperCamelCase_ = 0
UpperCamelCase_ = 1
UpperCamelCase_ = 2
UpperCamelCase_ = 3
UpperCamelCase_ = 4
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =VOCAB_FILES_NAMES
UpperCamelCase__ : Optional[Any] =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : Tuple =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Any ="left"
def __init__( self :Optional[int] , _lowercase :Union[str, Any] , _lowercase :Union[str, Any]=False , _lowercase :Optional[int]=True , _lowercase :Union[str, Any]=False , _lowercase :Tuple="<s>" , _lowercase :Any="</s>" , _lowercase :Dict="<unk>" , _lowercase :str="<sep>" , _lowercase :Tuple="<pad>" , _lowercase :Any="<cls>" , _lowercase :List[str]="<mask>" , _lowercase :Union[str, Any]=["<eop>", "<eod>"] , _lowercase :Optional[Dict[str, Any]] = None , **_lowercase :Union[str, Any] , ) -> None:
# Mask token behave like a normal word, i.e. include the space before it
UpperCAmelCase_ = AddedToken(_lowercase , lstrip=_lowercase , rstrip=_lowercase) if isinstance(_lowercase , _lowercase) else mask_token
UpperCAmelCase_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_lowercase , remove_space=_lowercase , keep_accents=_lowercase , bos_token=_lowercase , eos_token=_lowercase , unk_token=_lowercase , sep_token=_lowercase , pad_token=_lowercase , cls_token=_lowercase , mask_token=_lowercase , additional_special_tokens=_lowercase , sp_model_kwargs=self.sp_model_kwargs , **_lowercase , )
UpperCAmelCase_ = 3
UpperCAmelCase_ = do_lower_case
UpperCAmelCase_ = remove_space
UpperCAmelCase_ = keep_accents
UpperCAmelCase_ = vocab_file
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(_lowercase)
@property
def __a ( self :int) -> List[Any]:
return len(self.sp_model)
def __a ( self :Optional[int]) -> List[Any]:
UpperCAmelCase_ = {self.convert_ids_to_tokens(_lowercase): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def __getstate__( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.__dict__.copy()
UpperCAmelCase_ = None
return state
def __setstate__( self :Optional[Any] , _lowercase :Optional[Any]) -> List[Any]:
UpperCAmelCase_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs'''):
UpperCAmelCase_ = {}
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(self.vocab_file)
def __a ( self :List[str] , _lowercase :Tuple) -> Optional[int]:
if self.remove_space:
UpperCAmelCase_ = ''' '''.join(inputs.strip().split())
else:
UpperCAmelCase_ = inputs
UpperCAmelCase_ = outputs.replace('''``''' , '''"''').replace('''\'\'''' , '''"''')
if not self.keep_accents:
UpperCAmelCase_ = unicodedata.normalize('''NFKD''' , _lowercase)
UpperCAmelCase_ = ''''''.join([c for c in outputs if not unicodedata.combining(_lowercase)])
if self.do_lower_case:
UpperCAmelCase_ = outputs.lower()
return outputs
def __a ( self :str , _lowercase :str) -> List[str]:
UpperCAmelCase_ = self.preprocess_text(_lowercase)
UpperCAmelCase_ = self.sp_model.encode(_lowercase , out_type=_lowercase)
UpperCAmelCase_ = []
for piece in pieces:
if len(_lowercase) > 1 and piece[-1] == str(''',''') and piece[-2].isdigit():
UpperCAmelCase_ = self.sp_model.EncodeAsPieces(piece[:-1].replace(_lowercase , ''''''))
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0]) == 1:
UpperCAmelCase_ = cur_pieces[1:]
else:
UpperCAmelCase_ = cur_pieces[0][1:]
cur_pieces.append(piece[-1])
new_pieces.extend(_lowercase)
else:
new_pieces.append(_lowercase)
return new_pieces
def __a ( self :Optional[Any] , _lowercase :Union[str, Any]) -> Tuple:
return self.sp_model.PieceToId(_lowercase)
def __a ( self :Optional[int] , _lowercase :Optional[Any]) -> List[str]:
return self.sp_model.IdToPiece(_lowercase)
def __a ( self :List[Any] , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = ''''''.join(_lowercase).replace(_lowercase , ''' ''').strip()
return out_string
def __a ( self :Union[str, Any] , _lowercase :List[int] , _lowercase :bool = False , _lowercase :bool = None , _lowercase :bool = True , **_lowercase :Tuple , ) -> str:
UpperCAmelCase_ = kwargs.pop('''use_source_tokenizer''' , _lowercase)
UpperCAmelCase_ = self.convert_ids_to_tokens(_lowercase , skip_special_tokens=_lowercase)
# To avoid mixing byte-level and unicode for byte-level BPT
# we need to build string separately for added tokens and byte-level tokens
# cf. https://github.com/huggingface/transformers/issues/1133
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for token in filtered_tokens:
if skip_special_tokens and token in self.all_special_ids:
continue
if token in self.added_tokens_encoder:
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
UpperCAmelCase_ = []
sub_texts.append(_lowercase)
else:
current_sub_text.append(_lowercase)
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
# Mimic the behavior of the Rust tokenizer:
# By default, there are no spaces between special tokens
UpperCAmelCase_ = ''''''.join(_lowercase)
UpperCAmelCase_ = (
clean_up_tokenization_spaces
if clean_up_tokenization_spaces is not None
else self.clean_up_tokenization_spaces
)
if clean_up_tokenization_spaces:
UpperCAmelCase_ = self.clean_up_tokenization(_lowercase)
return clean_text
else:
return text
def __a ( self :str , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [self.cls_token_id]
if token_ids_a is None:
return token_ids_a + sep + cls
return token_ids_a + sep + token_ids_a + sep + cls
def __a ( self :Dict , _lowercase :List[int] , _lowercase :Optional[List[int]] = None , _lowercase :bool = False) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowercase , token_ids_a=_lowercase , already_has_special_tokens=_lowercase)
if token_ids_a is not None:
return ([0] * len(_lowercase)) + [1] + ([0] * len(_lowercase)) + [1, 1]
return ([0] * len(_lowercase)) + [1, 1]
def __a ( self :Optional[int] , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [2]
if token_ids_a is None:
return len(token_ids_a + sep) * [0] + cls_segment_id
return len(token_ids_a + sep) * [0] + len(token_ids_a + sep) * [1] + cls_segment_id
def __a ( self :str , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
if not os.path.isdir(_lowercase):
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
return
UpperCAmelCase_ = os.path.join(
_lowercase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''])
if os.path.abspath(self.vocab_file) != os.path.abspath(_lowercase) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , _lowercase)
elif not os.path.isfile(self.vocab_file):
with open(_lowercase , '''wb''') as fi:
UpperCAmelCase_ = self.sp_model.serialized_model_proto()
fi.write(_lowercase)
return (out_vocab_file,)
| 344 | 1 |
import unittest
from transformers import is_tf_available
from transformers.testing_utils import require_tf
if is_tf_available():
import tensorflow as tf
from tensorflow.python.eager import context
from tensorflow.python.framework import ops
from transformers import GradientAccumulator, create_optimizer
@require_tf
class a_ ( unittest.TestCase ):
def __a ( self :str , _lowercase :Union[str, Any] , _lowercase :Optional[int] , _lowercase :Dict) -> int:
self.assertEqual(len(_lowercase) , len(_lowercase))
for a, b in zip(_lowercase , _lowercase):
self.assertAlmostEqual(_lowercase , _lowercase , delta=_lowercase)
def __a ( self :str) -> Tuple:
UpperCAmelCase_ = GradientAccumulator()
accumulator([tf.constant([1.0, 2.0])])
accumulator([tf.constant([-2.0, 1.0])])
accumulator([tf.constant([-1.0, 2.0])])
with self.assertRaises(_lowercase):
accumulator([tf.constant([1.0, 1.0]), tf.constant([2.0, 2.0])])
self.assertEqual(accumulator.step , 3)
self.assertEqual(len(accumulator.gradients) , 1)
self.assertListAlmostEqual(accumulator.gradients[0].numpy().tolist() , [-2.0, 5.0] , tol=1E-2)
accumulator.reset()
self.assertEqual(accumulator.step , 0)
self.assertListAlmostEqual(accumulator.gradients[0].numpy().tolist() , [0.0, 0.0] , tol=1E-2)
def __a ( self :Any) -> Optional[Any]:
UpperCAmelCase_ = None
ops.enable_eager_execution_internal()
UpperCAmelCase_ = tf.config.list_physical_devices('''CPU''')
if len(_lowercase) == 1:
tf.config.set_logical_device_configuration(
physical_devices[0] , [tf.config.LogicalDeviceConfiguration(), tf.config.LogicalDeviceConfiguration()])
UpperCAmelCase_ = tf.config.list_logical_devices(device_type='''CPU''')
UpperCAmelCase_ = tf.distribute.MirroredStrategy(devices=devices[:2])
with strategy.scope():
UpperCAmelCase_ = GradientAccumulator()
UpperCAmelCase_ = tf.Variable([4.0, 3.0])
UpperCAmelCase_ , UpperCAmelCase_ = create_optimizer(5E-5 , 10 , 5)
UpperCAmelCase_ = tf.Variable([0.0, 0.0] , trainable=_lowercase)
def accumulate_on_replica(_lowercase :Optional[Any]):
accumulator([gradient])
def apply_on_replica():
optimizer.apply_gradients(list(zip(accumulator.gradients , [variable])))
@tf.function
def accumulate(_lowercase :str , _lowercase :Union[str, Any]):
with strategy.scope():
UpperCAmelCase_ = strategy.experimental_local_results(_lowercase)
local_variables[0].assign(_lowercase)
local_variables[1].assign(_lowercase)
strategy.run(_lowercase , args=(gradient_placeholder,))
@tf.function
def apply_grad():
with strategy.scope():
strategy.run(_lowercase)
def _check_local_values(_lowercase :List[Any] , _lowercase :Any):
UpperCAmelCase_ = strategy.experimental_local_results(accumulator._gradients[0])
self.assertListAlmostEqual(values[0].value() , _lowercase , tol=1E-2)
self.assertListAlmostEqual(values[1].value() , _lowercase , tol=1E-2)
accumulate([1.0, 2.0] , [-1.0, 1.0])
accumulate([3.0, -1.0] , [-1.0, -1.0])
accumulate([-2.0, 2.0] , [3.0, -2.0])
self.assertEqual(accumulator.step , 3)
_check_local_values([2.0, 3.0] , [1.0, -2.0])
apply_grad()
self.assertListAlmostEqual(variable.value() , [4.0, 3.0] , tol=1E-2)
accumulator.reset()
self.assertEqual(accumulator.step , 0)
_check_local_values([0.0, 0.0] , [0.0, 0.0])
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
UpperCamelCase_ = logging.get_logger(__name__)
class a_ ( _snake_case , _snake_case ):
UpperCamelCase__ : Union[str, Any] ="maskformer-swin"
UpperCamelCase__ : List[str] ={
"num_attention_heads": "num_heads",
"num_hidden_layers": "num_layers",
}
def __init__( self :Union[str, Any] , _lowercase :Optional[int]=224 , _lowercase :List[str]=4 , _lowercase :Tuple=3 , _lowercase :List[Any]=96 , _lowercase :Any=[2, 2, 6, 2] , _lowercase :int=[3, 6, 12, 24] , _lowercase :List[Any]=7 , _lowercase :Dict=4.0 , _lowercase :Any=True , _lowercase :int=0.0 , _lowercase :List[Any]=0.0 , _lowercase :Tuple=0.1 , _lowercase :str="gelu" , _lowercase :Union[str, Any]=False , _lowercase :Tuple=0.02 , _lowercase :List[str]=1E-5 , _lowercase :List[str]=None , _lowercase :Any=None , **_lowercase :Any , ) -> Union[str, Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = embed_dim
UpperCAmelCase_ = depths
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = num_heads
UpperCAmelCase_ = window_size
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = qkv_bias
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = use_absolute_embeddings
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = initializer_range
# we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
UpperCAmelCase_ = int(embed_dim * 2 ** (len(_lowercase) - 1))
UpperCAmelCase_ = ['''stem'''] + [f"stage{idx}" for idx in range(1 , len(_lowercase) + 1)]
UpperCAmelCase_ , UpperCAmelCase_ = get_aligned_output_features_output_indices(
out_features=_lowercase , out_indices=_lowercase , stage_names=self.stage_names)
| 344 | 1 |
import os
from tempfile import TemporaryDirectory
from unittest import TestCase
import pytest
from absl.testing import parameterized
from datasets import config
from datasets.arrow_reader import HF_GCP_BASE_URL
from datasets.builder import DatasetBuilder
from datasets.dataset_dict import IterableDatasetDict
from datasets.iterable_dataset import IterableDataset
from datasets.load import dataset_module_factory, import_main_class
from datasets.utils.file_utils import cached_path
UpperCamelCase_ = [
{"dataset": "wikipedia", "config_name": "20220301.de"},
{"dataset": "wikipedia", "config_name": "20220301.en"},
{"dataset": "wikipedia", "config_name": "20220301.fr"},
{"dataset": "wikipedia", "config_name": "20220301.frr"},
{"dataset": "wikipedia", "config_name": "20220301.it"},
{"dataset": "wikipedia", "config_name": "20220301.simple"},
{"dataset": "snli", "config_name": "plain_text"},
{"dataset": "eli5", "config_name": "LFQA_reddit"},
{"dataset": "wiki40b", "config_name": "en"},
{"dataset": "wiki_dpr", "config_name": "psgs_w100.nq.compressed"},
{"dataset": "wiki_dpr", "config_name": "psgs_w100.nq.no_index"},
{"dataset": "wiki_dpr", "config_name": "psgs_w100.multiset.no_index"},
{"dataset": "natural_questions", "config_name": "default"},
]
def A ( __UpperCAmelCase=True ) -> Optional[int]:
'''simple docstring'''
if with_config:
return [
{
"testcase_name": d["dataset"] + "/" + d["config_name"],
"dataset": d["dataset"],
"config_name": d["config_name"],
}
for d in DATASETS_ON_HF_GCP
]
else:
return [
{"testcase_name": dataset, "dataset": dataset} for dataset in {d["dataset"] for d in DATASETS_ON_HF_GCP}
]
@parameterized.named_parameters(list_datasets_on_hf_gcp_parameters(with_config=_snake_case ) )
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] =None
UpperCamelCase__ : List[Any] =None
def __a ( self :List[Any] , _lowercase :Dict , _lowercase :Any) -> List[Any]:
with TemporaryDirectory() as tmp_dir:
UpperCAmelCase_ = dataset_module_factory(_lowercase , cache_dir=_lowercase)
UpperCAmelCase_ = import_main_class(dataset_module.module_path , dataset=_lowercase)
UpperCAmelCase_ = builder_cls(
cache_dir=_lowercase , config_name=_lowercase , hash=dataset_module.hash , )
UpperCAmelCase_ = '''/'''.join(
[
HF_GCP_BASE_URL,
builder_instance._relative_data_dir(with_hash=_lowercase).replace(os.sep , '''/'''),
config.DATASET_INFO_FILENAME,
])
UpperCAmelCase_ = cached_path(_lowercase , cache_dir=_lowercase)
self.assertTrue(os.path.exists(_lowercase))
@pytest.mark.integration
def A ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
UpperCAmelCase_ = tmp_path_factory.mktemp('''test_hf_gcp''' ) / '''test_wikipedia_simple'''
UpperCAmelCase_ = dataset_module_factory('''wikipedia''' , cache_dir=__UpperCAmelCase )
UpperCAmelCase_ = import_main_class(dataset_module.module_path )
UpperCAmelCase_ = builder_cls(
cache_dir=__UpperCAmelCase , config_name='''20220301.frr''' , hash=dataset_module.hash , )
# use the HF cloud storage, not the original download_and_prepare that uses apache-beam
UpperCAmelCase_ = None
builder_instance.download_and_prepare()
UpperCAmelCase_ = builder_instance.as_dataset()
assert ds
@pytest.mark.integration
def A ( __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = dataset_module_factory('''wikipedia''' , cache_dir=__UpperCAmelCase )
UpperCAmelCase_ = import_main_class(dataset_module.module_path , dataset=__UpperCAmelCase )
UpperCAmelCase_ = builder_cls(
cache_dir=__UpperCAmelCase , config_name='''20220301.frr''' , hash=dataset_module.hash , )
UpperCAmelCase_ = builder_instance.as_streaming_dataset()
assert ds
assert isinstance(__UpperCAmelCase , __UpperCAmelCase )
assert "train" in ds
assert isinstance(ds['''train'''] , __UpperCAmelCase )
assert next(iter(ds['''train'''] ) )
| 344 |
import os
from argparse import ArgumentParser, Namespace
from ..data import SingleSentenceClassificationProcessor as Processor
from ..pipelines import TextClassificationPipeline
from ..utils import is_tf_available, is_torch_available, logging
from . import BaseTransformersCLICommand
if not is_tf_available() and not is_torch_available():
raise RuntimeError("At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training")
# TF training parameters
UpperCamelCase_ = False
UpperCamelCase_ = False
def A ( __UpperCAmelCase ) -> Any:
'''simple docstring'''
return TrainCommand(__UpperCAmelCase )
class a_ ( _snake_case ):
@staticmethod
def __a ( _lowercase :ArgumentParser) -> List[Any]:
UpperCAmelCase_ = parser.add_parser('''train''' , help='''CLI tool to train a model on a task.''')
train_parser.add_argument(
'''--train_data''' , type=_lowercase , required=_lowercase , help='''path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.''' , )
train_parser.add_argument(
'''--column_label''' , type=_lowercase , default=0 , help='''Column of the dataset csv file with example labels.''')
train_parser.add_argument(
'''--column_text''' , type=_lowercase , default=1 , help='''Column of the dataset csv file with example texts.''')
train_parser.add_argument(
'''--column_id''' , type=_lowercase , default=2 , help='''Column of the dataset csv file with example ids.''')
train_parser.add_argument(
'''--skip_first_row''' , action='''store_true''' , help='''Skip the first row of the csv file (headers).''')
train_parser.add_argument('''--validation_data''' , type=_lowercase , default='''''' , help='''path to validation dataset.''')
train_parser.add_argument(
'''--validation_split''' , type=_lowercase , default=0.1 , help='''if validation dataset is not provided, fraction of train dataset to use as validation dataset.''' , )
train_parser.add_argument('''--output''' , type=_lowercase , default='''./''' , help='''path to saved the trained model.''')
train_parser.add_argument(
'''--task''' , type=_lowercase , default='''text_classification''' , help='''Task to train the model on.''')
train_parser.add_argument(
'''--model''' , type=_lowercase , default='''bert-base-uncased''' , help='''Model\'s name or path to stored model.''')
train_parser.add_argument('''--train_batch_size''' , type=_lowercase , default=32 , help='''Batch size for training.''')
train_parser.add_argument('''--valid_batch_size''' , type=_lowercase , default=64 , help='''Batch size for validation.''')
train_parser.add_argument('''--learning_rate''' , type=_lowercase , default=3E-5 , help='''Learning rate.''')
train_parser.add_argument('''--adam_epsilon''' , type=_lowercase , default=1E-0_8 , help='''Epsilon for Adam optimizer.''')
train_parser.set_defaults(func=_lowercase)
def __init__( self :Union[str, Any] , _lowercase :Namespace) -> Union[str, Any]:
UpperCAmelCase_ = logging.get_logger('''transformers-cli/training''')
UpperCAmelCase_ = '''tf''' if is_tf_available() else '''torch'''
os.makedirs(args.output , exist_ok=_lowercase)
UpperCAmelCase_ = args.output
UpperCAmelCase_ = args.column_label
UpperCAmelCase_ = args.column_text
UpperCAmelCase_ = args.column_id
self.logger.info(f"Loading {args.task} pipeline for {args.model}")
if args.task == "text_classification":
UpperCAmelCase_ = TextClassificationPipeline.from_pretrained(args.model)
elif args.task == "token_classification":
raise NotImplementedError
elif args.task == "question_answering":
raise NotImplementedError
self.logger.info(f"Loading dataset from {args.train_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.train_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = None
if args.validation_data:
self.logger.info(f"Loading validation dataset from {args.validation_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.validation_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = args.validation_split
UpperCAmelCase_ = args.train_batch_size
UpperCAmelCase_ = args.valid_batch_size
UpperCAmelCase_ = args.learning_rate
UpperCAmelCase_ = args.adam_epsilon
def __a ( self :int) -> Tuple:
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def __a ( self :Optional[Any]) -> Any:
raise NotImplementedError
def __a ( self :int) -> Optional[Any]:
self.pipeline.fit(
self.train_dataset , validation_data=self.valid_dataset , validation_split=self.validation_split , learning_rate=self.learning_rate , adam_epsilon=self.adam_epsilon , train_batch_size=self.train_batch_size , valid_batch_size=self.valid_batch_size , )
# Save trained pipeline
self.pipeline.save_pretrained(self.output)
| 344 | 1 |
from __future__ import annotations
import typing
from collections import Counter
def A ( __UpperCAmelCase ) -> typing.Counter[int]:
'''simple docstring'''
UpperCAmelCase_ = Counter()
for base in range(1 , max_perimeter + 1 ):
for perpendicular in range(__UpperCAmelCase , max_perimeter + 1 ):
UpperCAmelCase_ = (base * base + perpendicular * perpendicular) ** 0.5
if hypotenuse == int(__UpperCAmelCase ):
UpperCAmelCase_ = int(base + perpendicular + hypotenuse )
if perimeter > max_perimeter:
continue
triplets[perimeter] += 1
return triplets
def A ( __UpperCAmelCase = 1000 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = pythagorean_triple(__UpperCAmelCase )
return triplets.most_common(1 )[0][0]
if __name__ == "__main__":
print(f"Perimeter {solution()} has maximum solutions")
| 344 |
import unittest
from typing import Dict, List, Optional, Union
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import BridgeTowerImageProcessor
class a_ ( unittest.TestCase ):
def __init__( self :Tuple , _lowercase :List[Any] , _lowercase :bool = True , _lowercase :Dict[str, int] = None , _lowercase :int = 32 , _lowercase :bool = True , _lowercase :Union[int, float] = 1 / 255 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :Optional[Union[float, List[float]]] = [0.48_145_466, 0.4_578_275, 0.40_821_073] , _lowercase :Optional[Union[float, List[float]]] = [0.26_862_954, 0.26_130_258, 0.27_577_711] , _lowercase :bool = True , _lowercase :List[Any]=7 , _lowercase :Dict=30 , _lowercase :Optional[int]=400 , _lowercase :Any=3 , ) -> Any:
UpperCAmelCase_ = parent
UpperCAmelCase_ = do_resize
UpperCAmelCase_ = size if size is not None else {'''shortest_edge''': 288}
UpperCAmelCase_ = size_divisor
UpperCAmelCase_ = do_rescale
UpperCAmelCase_ = rescale_factor
UpperCAmelCase_ = do_normalize
UpperCAmelCase_ = do_center_crop
UpperCAmelCase_ = image_mean
UpperCAmelCase_ = image_std
UpperCAmelCase_ = do_pad
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = min_resolution
UpperCAmelCase_ = max_resolution
def __a ( self :str) -> Tuple:
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"size_divisor": self.size_divisor,
}
def __a ( self :List[Any] , _lowercase :Tuple , _lowercase :List[str]=False) -> int:
if not batched:
UpperCAmelCase_ = self.size['''shortest_edge''']
UpperCAmelCase_ = image_inputs[0]
if isinstance(_lowercase , Image.Image):
UpperCAmelCase_ , UpperCAmelCase_ = image.size
else:
UpperCAmelCase_ , UpperCAmelCase_ = image.shape[1], image.shape[2]
UpperCAmelCase_ = size / min(_lowercase , _lowercase)
if h < w:
UpperCAmelCase_ , UpperCAmelCase_ = size, scale * w
else:
UpperCAmelCase_ , UpperCAmelCase_ = scale * h, size
UpperCAmelCase_ = int((1333 / 800) * size)
if max(_lowercase , _lowercase) > max_size:
UpperCAmelCase_ = max_size / max(_lowercase , _lowercase)
UpperCAmelCase_ = newh * scale
UpperCAmelCase_ = neww * scale
UpperCAmelCase_ , UpperCAmelCase_ = int(newh + 0.5), int(neww + 0.5)
UpperCAmelCase_ , UpperCAmelCase_ = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
UpperCAmelCase_ = []
for image in image_inputs:
UpperCAmelCase_ , UpperCAmelCase_ = self.get_expected_values([image])
expected_values.append((expected_height, expected_width))
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[0])[0]
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[1])[1]
return expected_height, expected_width
@require_torch
@require_vision
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =BridgeTowerImageProcessor if is_vision_available() else None
def __a ( self :int) -> Dict:
UpperCAmelCase_ = BridgeTowerImageProcessingTester(self)
@property
def __a ( self :Dict) -> Any:
return self.image_processor_tester.prepare_image_processor_dict()
def __a ( self :Dict) -> Tuple:
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(_lowercase , '''image_mean'''))
self.assertTrue(hasattr(_lowercase , '''image_std'''))
self.assertTrue(hasattr(_lowercase , '''do_normalize'''))
self.assertTrue(hasattr(_lowercase , '''do_resize'''))
self.assertTrue(hasattr(_lowercase , '''size'''))
self.assertTrue(hasattr(_lowercase , '''size_divisor'''))
def __a ( self :Union[str, Any]) -> Tuple:
pass
def __a ( self :List[str]) -> Tuple:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , Image.Image)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :Union[str, Any]) -> Optional[int]:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , numpify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , np.ndarray)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :str) -> int:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , torchify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , torch.Tensor)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
| 344 | 1 |
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import AutoTokenizer, BarkProcessor
from transformers.testing_utils import require_torch, slow
@require_torch
class a_ ( unittest.TestCase ):
def __a ( self :Dict) -> str:
UpperCAmelCase_ = '''ylacombe/bark-small'''
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = '''en_speaker_1'''
UpperCAmelCase_ = '''This is a test string'''
UpperCAmelCase_ = '''speaker_embeddings_path.json'''
UpperCAmelCase_ = '''speaker_embeddings'''
def __a ( self :Tuple , **_lowercase :Tuple) -> str:
return AutoTokenizer.from_pretrained(self.checkpoint , **_lowercase)
def __a ( self :List[str]) -> Union[str, Any]:
shutil.rmtree(self.tmpdirname)
def __a ( self :str) -> str:
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = BarkProcessor(tokenizer=_lowercase)
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = BarkProcessor.from_pretrained(self.tmpdirname)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab())
@slow
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = BarkProcessor.from_pretrained(
pretrained_processor_name_or_path=self.checkpoint , speaker_embeddings_dict_path=self.speaker_embeddings_dict_path , )
processor.save_pretrained(
self.tmpdirname , speaker_embeddings_dict_path=self.speaker_embeddings_dict_path , speaker_embeddings_directory=self.speaker_embeddings_directory , )
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = BarkProcessor.from_pretrained(
self.tmpdirname , self.speaker_embeddings_dict_path , bos_token='''(BOS)''' , eos_token='''(EOS)''' , )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
def __a ( self :Any) -> Union[str, Any]:
UpperCAmelCase_ = BarkProcessor.from_pretrained(
pretrained_processor_name_or_path=self.checkpoint , speaker_embeddings_dict_path=self.speaker_embeddings_dict_path , )
UpperCAmelCase_ = 35
UpperCAmelCase_ = 2
UpperCAmelCase_ = 8
UpperCAmelCase_ = {
'''semantic_prompt''': np.ones(_lowercase),
'''coarse_prompt''': np.ones((nb_codebooks_coarse, seq_len)),
'''fine_prompt''': np.ones((nb_codebooks_total, seq_len)),
}
# test providing already loaded voice_preset
UpperCAmelCase_ = processor(text=self.input_string , voice_preset=_lowercase)
UpperCAmelCase_ = inputs['''history_prompt''']
for key in voice_preset:
self.assertListEqual(voice_preset[key].tolist() , processed_voice_preset.get(_lowercase , np.array([])).tolist())
# test loading voice preset from npz file
UpperCAmelCase_ = os.path.join(self.tmpdirname , '''file.npz''')
np.savez(_lowercase , **_lowercase)
UpperCAmelCase_ = processor(text=self.input_string , voice_preset=_lowercase)
UpperCAmelCase_ = inputs['''history_prompt''']
for key in voice_preset:
self.assertListEqual(voice_preset[key].tolist() , processed_voice_preset.get(_lowercase , np.array([])).tolist())
# test loading voice preset from the hub
UpperCAmelCase_ = processor(text=self.input_string , voice_preset=self.voice_preset)
def __a ( self :str) -> Tuple:
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = BarkProcessor(tokenizer=_lowercase)
UpperCAmelCase_ = processor(text=self.input_string)
UpperCAmelCase_ = tokenizer(
self.input_string , padding='''max_length''' , max_length=256 , add_special_tokens=_lowercase , return_attention_mask=_lowercase , return_token_type_ids=_lowercase , )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key].squeeze().tolist())
| 344 |
def A ( __UpperCAmelCase = 100_0000 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = [i - 1 for i in range(limit + 1 )]
for i in range(2 , limit + 1 ):
if phi[i] == i - 1:
for j in range(2 * i , limit + 1 , __UpperCAmelCase ):
phi[j] -= phi[j] // i
return sum(phi[2 : limit + 1] )
if __name__ == "__main__":
print(solution())
| 344 | 1 |
import os
from typing import Dict, List, Union
import tensorflow as tf
from keras_nlp.tokenizers import BytePairTokenizer
from tensorflow_text import pad_model_inputs
from .tokenization_gpta import GPTaTokenizer
class a_ ( tf.keras.layers.Layer ):
def __init__( self :str , _lowercase :Dict[str, int] , _lowercase :List[str] , _lowercase :int = None , _lowercase :int = None) -> Optional[Any]:
super().__init__()
UpperCAmelCase_ = pad_token_id
UpperCAmelCase_ = max_length
UpperCAmelCase_ = vocab
UpperCAmelCase_ = merges
UpperCAmelCase_ = BytePairTokenizer(_lowercase , _lowercase , sequence_length=_lowercase)
@classmethod
def __a ( cls :Tuple , _lowercase :GPTaTokenizer , *_lowercase :Union[str, Any] , **_lowercase :str) -> List[Any]:
UpperCAmelCase_ = [''' '''.join(_lowercase) for m in tokenizer.bpe_ranks.keys()]
UpperCAmelCase_ = tokenizer.get_vocab()
return cls(_lowercase , _lowercase , *_lowercase , **_lowercase)
@classmethod
def __a ( cls :Optional[int] , _lowercase :Union[str, os.PathLike] , *_lowercase :str , **_lowercase :int) -> Union[str, Any]:
UpperCAmelCase_ = GPTaTokenizer.from_pretrained(_lowercase , *_lowercase , **_lowercase)
return cls.from_tokenizer(_lowercase , *_lowercase , **_lowercase)
@classmethod
def __a ( cls :List[str] , _lowercase :Any) -> Optional[int]:
return cls(**_lowercase)
def __a ( self :Dict) -> Optional[int]:
return {
"vocab": self.vocab,
"merges": self.merges,
"max_length": self.max_length,
"pad_token_id": self.pad_token_id,
}
def __a ( self :Tuple , _lowercase :Union[str, Any] , _lowercase :int = None) -> Optional[int]:
UpperCAmelCase_ = self.tf_tokenizer(_lowercase)
UpperCAmelCase_ = tf.ones_like(_lowercase)
if self.pad_token_id is not None:
# pad the tokens up to max length
UpperCAmelCase_ = max_length if max_length is not None else self.max_length
if max_length is not None:
UpperCAmelCase_ , UpperCAmelCase_ = pad_model_inputs(
_lowercase , max_seq_length=_lowercase , pad_value=self.pad_token_id)
return {"attention_mask": attention_mask, "input_ids": input_ids}
| 344 |
import tempfile
import torch
from diffusers import PNDMScheduler
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =(PNDMScheduler,)
UpperCamelCase__ : Optional[Any] =(("num_inference_steps", 50),)
def __a ( self :Union[str, Any] , **_lowercase :Any) -> Union[str, Any]:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
}
config.update(**_lowercase)
return config
def __a ( self :str , _lowercase :List[Any]=0 , **_lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Any) -> Optional[Any]:
pass
def __a ( self :str , _lowercase :int=0 , **_lowercase :Union[str, Any]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :int , **_lowercase :str) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.prk_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
for i, t in enumerate(scheduler.plms_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :Union[str, Any]) -> int:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
if num_inference_steps is not None and hasattr(_lowercase , '''set_timesteps'''):
scheduler.set_timesteps(_lowercase)
elif num_inference_steps is not None and not hasattr(_lowercase , '''set_timesteps'''):
UpperCAmelCase_ = num_inference_steps
# copy over dummy past residuals (must be done after set_timesteps)
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
def __a ( self :Any) -> Dict:
for timesteps in [100, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :List[Any]) -> Any:
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(steps_offset=1)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(10)
assert torch.equal(
scheduler.timesteps , torch.LongTensor(
[901, 851, 851, 801, 801, 751, 751, 701, 701, 651, 651, 601, 601, 501, 401, 301, 201, 101, 1]) , )
def __a ( self :Optional[int]) -> str:
for beta_start, beta_end in zip([0.0_001, 0.001] , [0.002, 0.02]):
self.check_over_configs(beta_start=_lowercase , beta_end=_lowercase)
def __a ( self :Any) -> List[str]:
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=_lowercase)
def __a ( self :List[Any]) -> Dict:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Tuple:
for t in [1, 5, 10]:
self.check_over_forward(time_step=_lowercase)
def __a ( self :Tuple) -> Dict:
for t, num_inference_steps in zip([1, 5, 10] , [10, 50, 100]):
self.check_over_forward(num_inference_steps=_lowercase)
def __a ( self :str) -> List[Any]:
# earlier version of set_timesteps() caused an error indexing alpha's with inference steps as power of 3
UpperCAmelCase_ = 27
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# before power of 3 fix, would error on first step, so we only need to do two
for i, t in enumerate(scheduler.prk_timesteps[:2]):
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
def __a ( self :List[str]) -> int:
with self.assertRaises(_lowercase):
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.step_plms(self.dummy_sample , 1 , self.dummy_sample).prev_sample
def __a ( self :List[str]) -> Dict:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 198.1_318) < 1E-2
assert abs(result_mean.item() - 0.2_580) < 1E-3
def __a ( self :Any) -> Tuple:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 67.3_986) < 1E-2
assert abs(result_mean.item() - 0.0_878) < 1E-3
def __a ( self :int) -> Any:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 230.0_399) < 1E-2
assert abs(result_mean.item() - 0.2_995) < 1E-3
def __a ( self :Any) -> Dict:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 186.9_482) < 1E-2
assert abs(result_mean.item() - 0.2_434) < 1E-3
| 344 | 1 |
def A ( __UpperCAmelCase ) -> list[int]:
'''simple docstring'''
if num <= 0:
raise ValueError('''Input must be a positive integer''' )
UpperCAmelCase_ = [True] * (num + 1)
UpperCAmelCase_ = 2
while p * p <= num:
if primes[p]:
for i in range(p * p , num + 1 , __UpperCAmelCase ):
UpperCAmelCase_ = False
p += 1
return [prime for prime in range(2 , num + 1 ) if primes[prime]]
if __name__ == "__main__":
import doctest
doctest.testmod()
UpperCamelCase_ = int(input("Enter a positive integer: ").strip())
print(prime_sieve_eratosthenes(user_num))
| 344 |
import os
from dataclasses import dataclass, field
from io import BytesIO
from typing import TYPE_CHECKING, Any, ClassVar, Dict, Optional, Union
import numpy as np
import pyarrow as pa
from .. import config
from ..download.streaming_download_manager import xopen, xsplitext
from ..table import array_cast
from ..utils.py_utils import no_op_if_value_is_null, string_to_dict
if TYPE_CHECKING:
from .features import FeatureType
UpperCamelCase_ , UpperCamelCase_ , UpperCamelCase_ = False, False, False
@dataclass
class a_ :
UpperCamelCase__ : Optional[int] =None
UpperCamelCase__ : bool =True
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[str] =None
# Automatically constructed
UpperCamelCase__ : ClassVar[str] ="dict"
UpperCamelCase__ : ClassVar[Any] =pa.struct({"bytes": pa.binary(), "path": pa.string()} )
UpperCamelCase__ : str =field(default="Audio" , init=_snake_case , repr=_snake_case )
def __call__( self :List[Any]) -> List[Any]:
return self.pa_type
def __a ( self :Any , _lowercase :Union[str, bytes, dict]) -> dict:
try:
import soundfile as sf # soundfile is a dependency of librosa, needed to decode audio files.
except ImportError as err:
raise ImportError('''To support encoding audio data, please install \'soundfile\'.''') from err
if isinstance(_lowercase , _lowercase):
return {"bytes": None, "path": value}
elif isinstance(_lowercase , _lowercase):
return {"bytes": value, "path": None}
elif "array" in value:
# convert the audio array to wav bytes
UpperCAmelCase_ = BytesIO()
sf.write(_lowercase , value['''array'''] , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
elif value.get('''path''') is not None and os.path.isfile(value['''path''']):
# we set "bytes": None to not duplicate the data if they're already available locally
if value["path"].endswith('''pcm'''):
# "PCM" only has raw audio bytes
if value.get('''sampling_rate''') is None:
# At least, If you want to convert "PCM-byte" to "WAV-byte", you have to know sampling rate
raise KeyError('''To use PCM files, please specify a \'sampling_rate\' in Audio object''')
if value.get('''bytes'''):
# If we already had PCM-byte, we don`t have to make "read file, make bytes" (just use it!)
UpperCAmelCase_ = np.frombuffer(value['''bytes'''] , dtype=np.intaa).astype(np.floataa) / 32767
else:
UpperCAmelCase_ = np.memmap(value['''path'''] , dtype='''h''' , mode='''r''').astype(np.floataa) / 32767
UpperCAmelCase_ = BytesIO(bytes())
sf.write(_lowercase , _lowercase , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
else:
return {"bytes": None, "path": value.get('''path''')}
elif value.get('''bytes''') is not None or value.get('''path''') is not None:
# store the audio bytes, and path is used to infer the audio format using the file extension
return {"bytes": value.get('''bytes'''), "path": value.get('''path''')}
else:
raise ValueError(
f"An audio sample should have one of 'path' or 'bytes' but they are missing or None in {value}.")
def __a ( self :Dict , _lowercase :dict , _lowercase :Optional[Dict[str, Union[str, bool, None]]] = None) -> dict:
if not self.decode:
raise RuntimeError('''Decoding is disabled for this feature. Please use Audio(decode=True) instead.''')
UpperCAmelCase_ , UpperCAmelCase_ = (value['''path'''], BytesIO(value['''bytes'''])) if value['''bytes'''] is not None else (value['''path'''], None)
if path is None and file is None:
raise ValueError(f"An audio sample should have one of 'path' or 'bytes' but both are None in {value}.")
try:
import librosa
import soundfile as sf
except ImportError as err:
raise ImportError('''To support decoding audio files, please install \'librosa\' and \'soundfile\'.''') from err
UpperCAmelCase_ = xsplitext(_lowercase)[1][1:].lower() if path is not None else None
if not config.IS_OPUS_SUPPORTED and audio_format == "opus":
raise RuntimeError(
'''Decoding \'opus\' files requires system library \'libsndfile\'>=1.0.31, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
elif not config.IS_MP3_SUPPORTED and audio_format == "mp3":
raise RuntimeError(
'''Decoding \'mp3\' files requires system library \'libsndfile\'>=1.1.0, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
if file is None:
UpperCAmelCase_ = token_per_repo_id or {}
UpperCAmelCase_ = path.split('''::''')[-1]
try:
UpperCAmelCase_ = string_to_dict(_lowercase , config.HUB_DATASETS_URL)['''repo_id''']
UpperCAmelCase_ = token_per_repo_id[repo_id]
except (ValueError, KeyError):
UpperCAmelCase_ = None
with xopen(_lowercase , '''rb''' , use_auth_token=_lowercase) as f:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
else:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
UpperCAmelCase_ = array.T
if self.mono:
UpperCAmelCase_ = librosa.to_mono(_lowercase)
if self.sampling_rate and self.sampling_rate != sampling_rate:
UpperCAmelCase_ = librosa.resample(_lowercase , orig_sr=_lowercase , target_sr=self.sampling_rate)
UpperCAmelCase_ = self.sampling_rate
return {"path": path, "array": array, "sampling_rate": sampling_rate}
def __a ( self :Union[str, Any]) -> Union["FeatureType", Dict[str, "FeatureType"]]:
from .features import Value
if self.decode:
raise ValueError('''Cannot flatten a decoded Audio feature.''')
return {
"bytes": Value('''binary'''),
"path": Value('''string'''),
}
def __a ( self :int , _lowercase :Union[pa.StringArray, pa.StructArray]) -> pa.StructArray:
if pa.types.is_string(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, storage] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_binary(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([storage, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_struct(storage.type) and storage.type.get_all_field_indices('''array'''):
UpperCAmelCase_ = pa.array([Audio().encode_example(_lowercase) if x is not None else None for x in storage.to_pylist()])
elif pa.types.is_struct(storage.type):
if storage.type.get_field_index('''bytes''') >= 0:
UpperCAmelCase_ = storage.field('''bytes''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
if storage.type.get_field_index('''path''') >= 0:
UpperCAmelCase_ = storage.field('''path''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
return array_cast(_lowercase , self.pa_type)
def __a ( self :Any , _lowercase :pa.StructArray) -> pa.StructArray:
@no_op_if_value_is_null
def path_to_bytes(_lowercase :Tuple):
with xopen(_lowercase , '''rb''') as f:
UpperCAmelCase_ = f.read()
return bytes_
UpperCAmelCase_ = pa.array(
[
(path_to_bytes(x['''path''']) if x['''bytes'''] is None else x['''bytes''']) if x is not None else None
for x in storage.to_pylist()
] , type=pa.binary() , )
UpperCAmelCase_ = pa.array(
[os.path.basename(_lowercase) if path is not None else None for path in storage.field('''path''').to_pylist()] , type=pa.string() , )
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
| 344 | 1 |
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import SPIECE_UNDERLINE, logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {"vocab_file": "spiece.model"}
UpperCamelCase_ = {
"vocab_file": {
"xlnet-base-cased": "https://huggingface.co/xlnet-base-cased/resolve/main/spiece.model",
"xlnet-large-cased": "https://huggingface.co/xlnet-large-cased/resolve/main/spiece.model",
}
}
UpperCamelCase_ = {
"xlnet-base-cased": None,
"xlnet-large-cased": None,
}
# Segments (not really needed)
UpperCamelCase_ = 0
UpperCamelCase_ = 1
UpperCamelCase_ = 2
UpperCamelCase_ = 3
UpperCamelCase_ = 4
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =VOCAB_FILES_NAMES
UpperCamelCase__ : Optional[Any] =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : Tuple =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Any ="left"
def __init__( self :Optional[int] , _lowercase :Union[str, Any] , _lowercase :Union[str, Any]=False , _lowercase :Optional[int]=True , _lowercase :Union[str, Any]=False , _lowercase :Tuple="<s>" , _lowercase :Any="</s>" , _lowercase :Dict="<unk>" , _lowercase :str="<sep>" , _lowercase :Tuple="<pad>" , _lowercase :Any="<cls>" , _lowercase :List[str]="<mask>" , _lowercase :Union[str, Any]=["<eop>", "<eod>"] , _lowercase :Optional[Dict[str, Any]] = None , **_lowercase :Union[str, Any] , ) -> None:
# Mask token behave like a normal word, i.e. include the space before it
UpperCAmelCase_ = AddedToken(_lowercase , lstrip=_lowercase , rstrip=_lowercase) if isinstance(_lowercase , _lowercase) else mask_token
UpperCAmelCase_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_lowercase , remove_space=_lowercase , keep_accents=_lowercase , bos_token=_lowercase , eos_token=_lowercase , unk_token=_lowercase , sep_token=_lowercase , pad_token=_lowercase , cls_token=_lowercase , mask_token=_lowercase , additional_special_tokens=_lowercase , sp_model_kwargs=self.sp_model_kwargs , **_lowercase , )
UpperCAmelCase_ = 3
UpperCAmelCase_ = do_lower_case
UpperCAmelCase_ = remove_space
UpperCAmelCase_ = keep_accents
UpperCAmelCase_ = vocab_file
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(_lowercase)
@property
def __a ( self :int) -> List[Any]:
return len(self.sp_model)
def __a ( self :Optional[int]) -> List[Any]:
UpperCAmelCase_ = {self.convert_ids_to_tokens(_lowercase): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def __getstate__( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.__dict__.copy()
UpperCAmelCase_ = None
return state
def __setstate__( self :Optional[Any] , _lowercase :Optional[Any]) -> List[Any]:
UpperCAmelCase_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs'''):
UpperCAmelCase_ = {}
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(self.vocab_file)
def __a ( self :List[str] , _lowercase :Tuple) -> Optional[int]:
if self.remove_space:
UpperCAmelCase_ = ''' '''.join(inputs.strip().split())
else:
UpperCAmelCase_ = inputs
UpperCAmelCase_ = outputs.replace('''``''' , '''"''').replace('''\'\'''' , '''"''')
if not self.keep_accents:
UpperCAmelCase_ = unicodedata.normalize('''NFKD''' , _lowercase)
UpperCAmelCase_ = ''''''.join([c for c in outputs if not unicodedata.combining(_lowercase)])
if self.do_lower_case:
UpperCAmelCase_ = outputs.lower()
return outputs
def __a ( self :str , _lowercase :str) -> List[str]:
UpperCAmelCase_ = self.preprocess_text(_lowercase)
UpperCAmelCase_ = self.sp_model.encode(_lowercase , out_type=_lowercase)
UpperCAmelCase_ = []
for piece in pieces:
if len(_lowercase) > 1 and piece[-1] == str(''',''') and piece[-2].isdigit():
UpperCAmelCase_ = self.sp_model.EncodeAsPieces(piece[:-1].replace(_lowercase , ''''''))
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0]) == 1:
UpperCAmelCase_ = cur_pieces[1:]
else:
UpperCAmelCase_ = cur_pieces[0][1:]
cur_pieces.append(piece[-1])
new_pieces.extend(_lowercase)
else:
new_pieces.append(_lowercase)
return new_pieces
def __a ( self :Optional[Any] , _lowercase :Union[str, Any]) -> Tuple:
return self.sp_model.PieceToId(_lowercase)
def __a ( self :Optional[int] , _lowercase :Optional[Any]) -> List[str]:
return self.sp_model.IdToPiece(_lowercase)
def __a ( self :List[Any] , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = ''''''.join(_lowercase).replace(_lowercase , ''' ''').strip()
return out_string
def __a ( self :Union[str, Any] , _lowercase :List[int] , _lowercase :bool = False , _lowercase :bool = None , _lowercase :bool = True , **_lowercase :Tuple , ) -> str:
UpperCAmelCase_ = kwargs.pop('''use_source_tokenizer''' , _lowercase)
UpperCAmelCase_ = self.convert_ids_to_tokens(_lowercase , skip_special_tokens=_lowercase)
# To avoid mixing byte-level and unicode for byte-level BPT
# we need to build string separately for added tokens and byte-level tokens
# cf. https://github.com/huggingface/transformers/issues/1133
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for token in filtered_tokens:
if skip_special_tokens and token in self.all_special_ids:
continue
if token in self.added_tokens_encoder:
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
UpperCAmelCase_ = []
sub_texts.append(_lowercase)
else:
current_sub_text.append(_lowercase)
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
# Mimic the behavior of the Rust tokenizer:
# By default, there are no spaces between special tokens
UpperCAmelCase_ = ''''''.join(_lowercase)
UpperCAmelCase_ = (
clean_up_tokenization_spaces
if clean_up_tokenization_spaces is not None
else self.clean_up_tokenization_spaces
)
if clean_up_tokenization_spaces:
UpperCAmelCase_ = self.clean_up_tokenization(_lowercase)
return clean_text
else:
return text
def __a ( self :str , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [self.cls_token_id]
if token_ids_a is None:
return token_ids_a + sep + cls
return token_ids_a + sep + token_ids_a + sep + cls
def __a ( self :Dict , _lowercase :List[int] , _lowercase :Optional[List[int]] = None , _lowercase :bool = False) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowercase , token_ids_a=_lowercase , already_has_special_tokens=_lowercase)
if token_ids_a is not None:
return ([0] * len(_lowercase)) + [1] + ([0] * len(_lowercase)) + [1, 1]
return ([0] * len(_lowercase)) + [1, 1]
def __a ( self :Optional[int] , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [2]
if token_ids_a is None:
return len(token_ids_a + sep) * [0] + cls_segment_id
return len(token_ids_a + sep) * [0] + len(token_ids_a + sep) * [1] + cls_segment_id
def __a ( self :str , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
if not os.path.isdir(_lowercase):
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
return
UpperCAmelCase_ = os.path.join(
_lowercase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''])
if os.path.abspath(self.vocab_file) != os.path.abspath(_lowercase) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , _lowercase)
elif not os.path.isfile(self.vocab_file):
with open(_lowercase , '''wb''') as fi:
UpperCAmelCase_ = self.sp_model.serialized_model_proto()
fi.write(_lowercase)
return (out_vocab_file,)
| 344 |
from ..utils import DummyObject, requires_backends
class a_ ( metaclass=_snake_case ):
UpperCamelCase__ : Any =["torch", "scipy"]
def __init__( self :List[str] , *_lowercase :List[str] , **_lowercase :Union[str, Any]) -> List[Any]:
requires_backends(self , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Dict , *_lowercase :Any , **_lowercase :Dict) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Optional[Any] , *_lowercase :str , **_lowercase :Optional[Any]) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
| 344 | 1 |
import argparse
import json
import os
import fairseq
import torch
from fairseq.data import Dictionary
from transformers import (
HubertConfig,
HubertForCTC,
HubertModel,
WavaVecaCTCTokenizer,
WavaVecaFeatureExtractor,
WavaVecaProcessor,
logging,
)
logging.set_verbosity_info()
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"post_extract_proj": "feature_projection.projection",
"encoder.pos_conv.0": "encoder.pos_conv_embed.conv",
"self_attn.k_proj": "encoder.layers.*.attention.k_proj",
"self_attn.v_proj": "encoder.layers.*.attention.v_proj",
"self_attn.q_proj": "encoder.layers.*.attention.q_proj",
"self_attn.out_proj": "encoder.layers.*.attention.out_proj",
"self_attn_layer_norm": "encoder.layers.*.layer_norm",
"fc1": "encoder.layers.*.feed_forward.intermediate_dense",
"fc2": "encoder.layers.*.feed_forward.output_dense",
"final_layer_norm": "encoder.layers.*.final_layer_norm",
"encoder.layer_norm": "encoder.layer_norm",
"w2v_model.layer_norm": "feature_projection.layer_norm",
"w2v_encoder.proj": "lm_head",
"mask_emb": "masked_spec_embed",
}
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
for attribute in key.split('''.''' ):
UpperCAmelCase_ = getattr(__UpperCAmelCase , __UpperCAmelCase )
if weight_type is not None:
UpperCAmelCase_ = getattr(__UpperCAmelCase , __UpperCAmelCase ).shape
else:
UpperCAmelCase_ = hf_pointer.shape
assert hf_shape == value.shape, (
f"Shape of hf {key + '.' + weight_type if weight_type is not None else ''} is {hf_shape}, but should be"
f" {value.shape} for {full_name}"
)
if weight_type == "weight":
UpperCAmelCase_ = value
elif weight_type == "weight_g":
UpperCAmelCase_ = value
elif weight_type == "weight_v":
UpperCAmelCase_ = value
elif weight_type == "bias":
UpperCAmelCase_ = value
else:
UpperCAmelCase_ = value
logger.info(f"{key + '.' + weight_type if weight_type is not None else ''} was initialized from {full_name}." )
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = fairseq_model.state_dict()
UpperCAmelCase_ = hf_model.hubert.feature_extractor if is_finetuned else hf_model.feature_extractor
for name, value in fairseq_dict.items():
UpperCAmelCase_ = False
if "conv_layers" in name:
load_conv_layer(
__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , hf_model.config.feat_extract_norm == '''group''' , )
UpperCAmelCase_ = True
else:
for key, mapped_key in MAPPING.items():
UpperCAmelCase_ = '''hubert.''' + mapped_key if (is_finetuned and mapped_key != '''lm_head''') else mapped_key
if key in name or (key.split('''w2v_model.''' )[-1] == name.split('''.''' )[0] and not is_finetuned):
UpperCAmelCase_ = True
if "*" in mapped_key:
UpperCAmelCase_ = name.split(__UpperCAmelCase )[0].split('''.''' )[-2]
UpperCAmelCase_ = mapped_key.replace('''*''' , __UpperCAmelCase )
if "weight_g" in name:
UpperCAmelCase_ = '''weight_g'''
elif "weight_v" in name:
UpperCAmelCase_ = '''weight_v'''
elif "weight" in name:
UpperCAmelCase_ = '''weight'''
elif "bias" in name:
UpperCAmelCase_ = '''bias'''
else:
UpperCAmelCase_ = None
set_recursively(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
continue
if not is_used:
unused_weights.append(__UpperCAmelCase )
logger.warning(f"Unused weights: {unused_weights}" )
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
UpperCAmelCase_ = full_name.split('''conv_layers.''' )[-1]
UpperCAmelCase_ = name.split('''.''' )
UpperCAmelCase_ = int(items[0] )
UpperCAmelCase_ = int(items[1] )
if type_id == 0:
if "bias" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].conv.bias.data.shape, (
f"{full_name} has size {value.shape}, but"
f" {feature_extractor.conv_layers[layer_id].conv.bias.data.shape} was found."
)
UpperCAmelCase_ = value
logger.info(f"Feat extract conv layer {layer_id} was initialized from {full_name}." )
elif "weight" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].conv.weight.data.shape, (
f"{full_name} has size {value.shape}, but"
f" {feature_extractor.conv_layers[layer_id].conv.weight.data.shape} was found."
)
UpperCAmelCase_ = value
logger.info(f"Feat extract conv layer {layer_id} was initialized from {full_name}." )
elif (type_id == 2 and not use_group_norm) or (type_id == 2 and layer_id == 0 and use_group_norm):
if "bias" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape, (
f"{full_name} has size {value.shape}, but {feature_extractor[layer_id].layer_norm.bias.data.shape} was"
" found."
)
UpperCAmelCase_ = value
logger.info(f"Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}." )
elif "weight" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape, (
f"{full_name} has size {value.shape}, but"
f" {feature_extractor[layer_id].layer_norm.weight.data.shape} was found."
)
UpperCAmelCase_ = value
logger.info(f"Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}." )
else:
unused_weights.append(__UpperCAmelCase )
@torch.no_grad()
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase=None , __UpperCAmelCase=None , __UpperCAmelCase=True ) -> List[str]:
'''simple docstring'''
if config_path is not None:
UpperCAmelCase_ = HubertConfig.from_pretrained(__UpperCAmelCase )
else:
UpperCAmelCase_ = HubertConfig()
if is_finetuned:
if dict_path:
UpperCAmelCase_ = Dictionary.load(__UpperCAmelCase )
# important change bos & pad token id since CTC symbol is <pad> and
# not <s> as in fairseq
UpperCAmelCase_ = target_dict.pad_index
UpperCAmelCase_ = target_dict.bos_index
UpperCAmelCase_ = target_dict.eos_index
UpperCAmelCase_ = len(target_dict.symbols )
UpperCAmelCase_ = os.path.join(__UpperCAmelCase , '''vocab.json''' )
if not os.path.isdir(__UpperCAmelCase ):
logger.error('''--pytorch_dump_folder_path ({}) should be a directory'''.format(__UpperCAmelCase ) )
return
os.makedirs(__UpperCAmelCase , exist_ok=__UpperCAmelCase )
with open(__UpperCAmelCase , '''w''' , encoding='''utf-8''' ) as vocab_handle:
json.dump(target_dict.indices , __UpperCAmelCase )
UpperCAmelCase_ = WavaVecaCTCTokenizer(
__UpperCAmelCase , unk_token=target_dict.unk_word , pad_token=target_dict.pad_word , bos_token=target_dict.bos_word , eos_token=target_dict.eos_word , word_delimiter_token='''|''' , do_lower_case=__UpperCAmelCase , )
UpperCAmelCase_ = True if config.feat_extract_norm == '''layer''' else False
UpperCAmelCase_ = WavaVecaFeatureExtractor(
feature_size=1 , sampling_rate=1_6000 , padding_value=0 , do_normalize=__UpperCAmelCase , return_attention_mask=__UpperCAmelCase , )
UpperCAmelCase_ = WavaVecaProcessor(feature_extractor=__UpperCAmelCase , tokenizer=__UpperCAmelCase )
processor.save_pretrained(__UpperCAmelCase )
UpperCAmelCase_ = HubertForCTC(__UpperCAmelCase )
else:
UpperCAmelCase_ = HubertModel(__UpperCAmelCase )
if is_finetuned:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = fairseq.checkpoint_utils.load_model_ensemble_and_task(
[checkpoint_path] , arg_overrides={'''data''': '''/'''.join(dict_path.split('''/''' )[:-1] )} )
else:
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = fairseq.checkpoint_utils.load_model_ensemble_and_task([checkpoint_path] )
UpperCAmelCase_ = model[0].eval()
recursively_load_weights(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
hf_wavavec.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
UpperCamelCase_ = argparse.ArgumentParser()
parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.")
parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint")
parser.add_argument("--dict_path", default=None, type=str, help="Path to dict of fine-tuned model")
parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert")
parser.add_argument(
"--not_finetuned", action="store_true", help="Whether the model to convert is a fine-tuned model or not"
)
UpperCamelCase_ = parser.parse_args()
convert_hubert_checkpoint(
args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned
)
| 344 |
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def A ( __UpperCAmelCase ) -> Dict[str, torch.Tensor]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for rt in rc.restypes:
UpperCAmelCase_ = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
UpperCAmelCase_ = {name: i for i, name in enumerate(__UpperCAmelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.floataa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = protein['''aatype'''].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the corresponding mask
UpperCAmelCase_ = torch.zeros([21, 37] , dtype=torch.floataa , device=protein['''aatype'''].device )
for restype, restype_letter in enumerate(rc.restypes ):
UpperCAmelCase_ = rc.restype_atoa[restype_letter]
UpperCAmelCase_ = rc.residue_atoms[restype_name]
for atom_name in atom_names:
UpperCAmelCase_ = rc.atom_order[atom_name]
UpperCAmelCase_ = 1
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
return protein
def A ( __UpperCAmelCase ) -> Dict[str, np.ndarray]:
'''simple docstring'''
UpperCAmelCase_ = tree_map(lambda __UpperCAmelCase : torch.tensor(__UpperCAmelCase , device=batch['''aatype'''].device ) , __UpperCAmelCase , np.ndarray )
UpperCAmelCase_ = tensor_tree_map(lambda __UpperCAmelCase : np.array(__UpperCAmelCase ) , make_atomaa_masks(__UpperCAmelCase ) )
return out
| 344 | 1 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tf_available, is_torch_available
UpperCamelCase_ = {
"configuration_data2vec_audio": ["DATA2VEC_AUDIO_PRETRAINED_CONFIG_ARCHIVE_MAP", "Data2VecAudioConfig"],
"configuration_data2vec_text": [
"DATA2VEC_TEXT_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Data2VecTextConfig",
"Data2VecTextOnnxConfig",
],
"configuration_data2vec_vision": [
"DATA2VEC_VISION_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Data2VecVisionConfig",
"Data2VecVisionOnnxConfig",
],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"DATA2VEC_AUDIO_PRETRAINED_MODEL_ARCHIVE_LIST",
"Data2VecAudioForAudioFrameClassification",
"Data2VecAudioForCTC",
"Data2VecAudioForSequenceClassification",
"Data2VecAudioForXVector",
"Data2VecAudioModel",
"Data2VecAudioPreTrainedModel",
]
UpperCamelCase_ = [
"DATA2VEC_TEXT_PRETRAINED_MODEL_ARCHIVE_LIST",
"Data2VecTextForCausalLM",
"Data2VecTextForMaskedLM",
"Data2VecTextForMultipleChoice",
"Data2VecTextForQuestionAnswering",
"Data2VecTextForSequenceClassification",
"Data2VecTextForTokenClassification",
"Data2VecTextModel",
"Data2VecTextPreTrainedModel",
]
UpperCamelCase_ = [
"DATA2VEC_VISION_PRETRAINED_MODEL_ARCHIVE_LIST",
"Data2VecVisionForImageClassification",
"Data2VecVisionForMaskedImageModeling",
"Data2VecVisionForSemanticSegmentation",
"Data2VecVisionModel",
"Data2VecVisionPreTrainedModel",
]
if is_tf_available():
UpperCamelCase_ = [
"TFData2VecVisionForImageClassification",
"TFData2VecVisionForSemanticSegmentation",
"TFData2VecVisionModel",
"TFData2VecVisionPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_dataavec_audio import DATA2VEC_AUDIO_PRETRAINED_CONFIG_ARCHIVE_MAP, DataaVecAudioConfig
from .configuration_dataavec_text import (
DATA2VEC_TEXT_PRETRAINED_CONFIG_ARCHIVE_MAP,
DataaVecTextConfig,
DataaVecTextOnnxConfig,
)
from .configuration_dataavec_vision import (
DATA2VEC_VISION_PRETRAINED_CONFIG_ARCHIVE_MAP,
DataaVecVisionConfig,
DataaVecVisionOnnxConfig,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_dataavec_audio import (
DATA2VEC_AUDIO_PRETRAINED_MODEL_ARCHIVE_LIST,
DataaVecAudioForAudioFrameClassification,
DataaVecAudioForCTC,
DataaVecAudioForSequenceClassification,
DataaVecAudioForXVector,
DataaVecAudioModel,
DataaVecAudioPreTrainedModel,
)
from .modeling_dataavec_text import (
DATA2VEC_TEXT_PRETRAINED_MODEL_ARCHIVE_LIST,
DataaVecTextForCausalLM,
DataaVecTextForMaskedLM,
DataaVecTextForMultipleChoice,
DataaVecTextForQuestionAnswering,
DataaVecTextForSequenceClassification,
DataaVecTextForTokenClassification,
DataaVecTextModel,
DataaVecTextPreTrainedModel,
)
from .modeling_dataavec_vision import (
DATA2VEC_VISION_PRETRAINED_MODEL_ARCHIVE_LIST,
DataaVecVisionForImageClassification,
DataaVecVisionForMaskedImageModeling,
DataaVecVisionForSemanticSegmentation,
DataaVecVisionModel,
DataaVecVisionPreTrainedModel,
)
if is_tf_available():
from .modeling_tf_dataavec_vision import (
TFDataaVecVisionForImageClassification,
TFDataaVecVisionForSemanticSegmentation,
TFDataaVecVisionModel,
TFDataaVecVisionPreTrainedModel,
)
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from ..models.whisper import WhisperForConditionalGeneration, WhisperProcessor
from .base import PipelineTool
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="openai/whisper-base"
UpperCamelCase__ : int =(
"This is a tool that transcribes an audio into text. It takes an input named `audio` and returns the "
"transcribed text."
)
UpperCamelCase__ : Any ="transcriber"
UpperCamelCase__ : Optional[int] =WhisperProcessor
UpperCamelCase__ : List[str] =WhisperForConditionalGeneration
UpperCamelCase__ : List[Any] =["audio"]
UpperCamelCase__ : Union[str, Any] =["text"]
def __a ( self :int , _lowercase :Any) -> Tuple:
return self.pre_processor(_lowercase , return_tensors='''pt''').input_features
def __a ( self :Dict , _lowercase :Tuple) -> Any:
return self.model.generate(inputs=_lowercase)
def __a ( self :int , _lowercase :Union[str, Any]) -> Optional[Any]:
return self.pre_processor.batch_decode(_lowercase , skip_special_tokens=_lowercase)[0]
| 344 | 1 |
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> float:
'''simple docstring'''
UpperCAmelCase_ = (num_of_terms / 2) * (2 * first_term + (num_of_terms - 1) * common_diff)
# formula for sum of series
return total
def A ( ) -> Any:
'''simple docstring'''
print(sum_of_series(1 , 1 , 10 ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
UpperCamelCase_ = {"configuration_opt": ["OPT_PRETRAINED_CONFIG_ARCHIVE_MAP", "OPTConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"OPT_PRETRAINED_MODEL_ARCHIVE_LIST",
"OPTForCausalLM",
"OPTModel",
"OPTPreTrainedModel",
"OPTForSequenceClassification",
"OPTForQuestionAnswering",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["TFOPTForCausalLM", "TFOPTModel", "TFOPTPreTrainedModel"]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxOPTForCausalLM",
"FlaxOPTModel",
"FlaxOPTPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_opt import OPT_PRETRAINED_CONFIG_ARCHIVE_MAP, OPTConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_opt import (
OPT_PRETRAINED_MODEL_ARCHIVE_LIST,
OPTForCausalLM,
OPTForQuestionAnswering,
OPTForSequenceClassification,
OPTModel,
OPTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_opt import TFOPTForCausalLM, TFOPTModel, TFOPTPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_opt import FlaxOPTForCausalLM, FlaxOPTModel, FlaxOPTPreTrainedModel
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 | 1 |
import os
import tempfile
import unittest
from transformers import DistilBertConfig, is_torch_available
from transformers.testing_utils import require_torch, require_torch_gpu, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
DistilBertForMaskedLM,
DistilBertForMultipleChoice,
DistilBertForQuestionAnswering,
DistilBertForSequenceClassification,
DistilBertForTokenClassification,
DistilBertModel,
)
class a_ ( _snake_case ):
def __init__( self :int , _lowercase :int , _lowercase :Dict=13 , _lowercase :List[Any]=7 , _lowercase :Optional[int]=True , _lowercase :Tuple=True , _lowercase :List[Any]=False , _lowercase :Any=True , _lowercase :Dict=99 , _lowercase :List[str]=32 , _lowercase :Tuple=5 , _lowercase :Optional[Any]=4 , _lowercase :List[str]=37 , _lowercase :Optional[int]="gelu" , _lowercase :List[str]=0.1 , _lowercase :Dict=0.1 , _lowercase :Optional[int]=512 , _lowercase :Tuple=16 , _lowercase :Any=2 , _lowercase :List[str]=0.02 , _lowercase :int=3 , _lowercase :Union[str, Any]=4 , _lowercase :Tuple=None , ) -> Dict:
UpperCAmelCase_ = parent
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = seq_length
UpperCAmelCase_ = is_training
UpperCAmelCase_ = use_input_mask
UpperCAmelCase_ = use_token_type_ids
UpperCAmelCase_ = use_labels
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = type_vocab_size
UpperCAmelCase_ = type_sequence_label_size
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = num_labels
UpperCAmelCase_ = num_choices
UpperCAmelCase_ = scope
def __a ( self :str) -> Union[str, Any]:
UpperCAmelCase_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size)
UpperCAmelCase_ = None
if self.use_input_mask:
UpperCAmelCase_ = random_attention_mask([self.batch_size, self.seq_length])
UpperCAmelCase_ = None
UpperCAmelCase_ = None
UpperCAmelCase_ = None
if self.use_labels:
UpperCAmelCase_ = ids_tensor([self.batch_size] , self.type_sequence_label_size)
UpperCAmelCase_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels)
UpperCAmelCase_ = ids_tensor([self.batch_size] , self.num_choices)
UpperCAmelCase_ = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def __a ( self :Any) -> Dict:
return DistilBertConfig(
vocab_size=self.vocab_size , dim=self.hidden_size , n_layers=self.num_hidden_layers , n_heads=self.num_attention_heads , hidden_dim=self.intermediate_size , hidden_act=self.hidden_act , dropout=self.hidden_dropout_prob , attention_dropout=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , initializer_range=self.initializer_range , )
def __a ( self :Any , _lowercase :Union[str, Any] , _lowercase :Optional[int] , _lowercase :int , _lowercase :List[Any] , _lowercase :Tuple , _lowercase :Tuple) -> Optional[Any]:
UpperCAmelCase_ = DistilBertModel(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = model(_lowercase)
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size))
def __a ( self :List[str] , _lowercase :Dict , _lowercase :List[str] , _lowercase :Dict , _lowercase :Union[str, Any] , _lowercase :Any , _lowercase :Optional[Any]) -> Optional[Any]:
UpperCAmelCase_ = DistilBertForMaskedLM(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , labels=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size))
def __a ( self :Optional[int] , _lowercase :Optional[Any] , _lowercase :Any , _lowercase :List[Any] , _lowercase :Optional[int] , _lowercase :List[str] , _lowercase :Optional[Any]) -> Dict:
UpperCAmelCase_ = DistilBertForQuestionAnswering(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(
_lowercase , attention_mask=_lowercase , start_positions=_lowercase , end_positions=_lowercase)
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length))
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length))
def __a ( self :Any , _lowercase :Optional[int] , _lowercase :Optional[Any] , _lowercase :Union[str, Any] , _lowercase :Union[str, Any] , _lowercase :List[Any] , _lowercase :Union[str, Any]) -> str:
UpperCAmelCase_ = self.num_labels
UpperCAmelCase_ = DistilBertForSequenceClassification(_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , labels=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels))
def __a ( self :Dict , _lowercase :List[str] , _lowercase :List[Any] , _lowercase :Optional[Any] , _lowercase :int , _lowercase :List[str] , _lowercase :Union[str, Any]) -> Tuple:
UpperCAmelCase_ = self.num_labels
UpperCAmelCase_ = DistilBertForTokenClassification(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase , labels=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels))
def __a ( self :Optional[int] , _lowercase :Any , _lowercase :Dict , _lowercase :Dict , _lowercase :int , _lowercase :Dict , _lowercase :Optional[int]) -> Any:
UpperCAmelCase_ = self.num_choices
UpperCAmelCase_ = DistilBertForMultipleChoice(config=_lowercase)
model.to(_lowercase)
model.eval()
UpperCAmelCase_ = input_ids.unsqueeze(1).expand(-1 , self.num_choices , -1).contiguous()
UpperCAmelCase_ = input_mask.unsqueeze(1).expand(-1 , self.num_choices , -1).contiguous()
UpperCAmelCase_ = model(
_lowercase , attention_mask=_lowercase , labels=_lowercase , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices))
def __a ( self :List[str]) -> Optional[Any]:
UpperCAmelCase_ = self.prepare_config_and_inputs()
((UpperCAmelCase_) , (UpperCAmelCase_) , (UpperCAmelCase_) , (UpperCAmelCase_) , (UpperCAmelCase_) , (UpperCAmelCase_)) = config_and_inputs
UpperCAmelCase_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a_ ( _snake_case , _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =(
(
DistilBertModel,
DistilBertForMaskedLM,
DistilBertForMultipleChoice,
DistilBertForQuestionAnswering,
DistilBertForSequenceClassification,
DistilBertForTokenClassification,
)
if is_torch_available()
else None
)
UpperCamelCase__ : List[Any] =(
{
"feature-extraction": DistilBertModel,
"fill-mask": DistilBertForMaskedLM,
"question-answering": DistilBertForQuestionAnswering,
"text-classification": DistilBertForSequenceClassification,
"token-classification": DistilBertForTokenClassification,
"zero-shot": DistilBertForSequenceClassification,
}
if is_torch_available()
else {}
)
UpperCamelCase__ : Optional[int] =True
UpperCamelCase__ : Optional[int] =True
UpperCamelCase__ : Union[str, Any] =True
UpperCamelCase__ : Dict =True
def __a ( self :Any) -> Tuple:
UpperCAmelCase_ = DistilBertModelTester(self)
UpperCAmelCase_ = ConfigTester(self , config_class=_lowercase , dim=37)
def __a ( self :Optional[Any]) -> List[Any]:
self.config_tester.run_common_tests()
def __a ( self :Optional[Any]) -> Optional[Any]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_model(*_lowercase)
def __a ( self :Tuple) -> List[str]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_masked_lm(*_lowercase)
def __a ( self :Optional[int]) -> int:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_question_answering(*_lowercase)
def __a ( self :Any) -> int:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_sequence_classification(*_lowercase)
def __a ( self :Optional[int]) -> List[Any]:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_token_classification(*_lowercase)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_multiple_choice(*_lowercase)
@slow
def __a ( self :Tuple) -> Dict:
for model_name in DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
UpperCAmelCase_ = DistilBertModel.from_pretrained(_lowercase)
self.assertIsNotNone(_lowercase)
@slow
@require_torch_gpu
def __a ( self :List[Any]) -> Tuple:
UpperCAmelCase_ , UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
# BertForMultipleChoice behaves incorrectly in JIT environments.
if model_class == DistilBertForMultipleChoice:
return
UpperCAmelCase_ = True
UpperCAmelCase_ = model_class(config=_lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase)
UpperCAmelCase_ = torch.jit.trace(
_lowercase , (inputs_dict['''input_ids'''].to('''cpu'''), inputs_dict['''attention_mask'''].to('''cpu''')))
with tempfile.TemporaryDirectory() as tmp:
torch.jit.save(_lowercase , os.path.join(_lowercase , '''traced_model.pt'''))
UpperCAmelCase_ = torch.jit.load(os.path.join(_lowercase , '''traced_model.pt''') , map_location=_lowercase)
loaded(inputs_dict['''input_ids'''].to(_lowercase) , inputs_dict['''attention_mask'''].to(_lowercase))
@require_torch
class a_ ( unittest.TestCase ):
@slow
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = DistilBertModel.from_pretrained('''distilbert-base-uncased''')
UpperCAmelCase_ = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]])
UpperCAmelCase_ = torch.tensor([[0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])
with torch.no_grad():
UpperCAmelCase_ = model(_lowercase , attention_mask=_lowercase)[0]
UpperCAmelCase_ = torch.Size((1, 11, 768))
self.assertEqual(output.shape , _lowercase)
UpperCAmelCase_ = torch.tensor(
[[[-0.1_639, 0.3_299, 0.1_648], [-0.1_746, 0.3_289, 0.1_710], [-0.1_884, 0.3_357, 0.1_810]]])
self.assertTrue(torch.allclose(output[:, 1:4, 1:4] , _lowercase , atol=1E-4))
| 344 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
UpperCamelCase_ = "▁"
UpperCamelCase_ = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : str =BigBirdTokenizer
UpperCamelCase__ : Tuple =BigBirdTokenizerFast
UpperCamelCase__ : Union[str, Any] =True
UpperCamelCase__ : List[str] =True
def __a ( self :Any) -> List[str]:
super().setUp()
UpperCAmelCase_ = self.tokenizer_class(_lowercase , keep_accents=_lowercase)
tokenizer.save_pretrained(self.tmpdirname)
def __a ( self :Optional[int]) -> str:
UpperCAmelCase_ = '''<s>'''
UpperCAmelCase_ = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_lowercase) , _lowercase)
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_lowercase) , _lowercase)
def __a ( self :str) -> str:
UpperCAmelCase_ = list(self.get_tokenizer().get_vocab().keys())
self.assertEqual(vocab_keys[0] , '''<unk>''')
self.assertEqual(vocab_keys[1] , '''<s>''')
self.assertEqual(vocab_keys[-1] , '''[MASK]''')
self.assertEqual(len(_lowercase) , 1004)
def __a ( self :List[str]) -> int:
self.assertEqual(self.get_tokenizer().vocab_size , 1000)
def __a ( self :Tuple) -> int:
if not self.test_rust_tokenizer:
return
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = '''I was born in 92000, and this is falsé.'''
UpperCAmelCase_ = tokenizer.tokenize(_lowercase)
UpperCAmelCase_ = rust_tokenizer.tokenize(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = tokenizer.encode(_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :Optional[Any]) -> List[str]:
UpperCAmelCase_ = BigBirdTokenizer(_lowercase , keep_accents=_lowercase)
UpperCAmelCase_ = tokenizer.tokenize('''This is a test''')
self.assertListEqual(_lowercase , ['''▁This''', '''▁is''', '''▁a''', '''▁t''', '''est'''])
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_lowercase) , [285, 46, 10, 170, 382] , )
UpperCAmelCase_ = tokenizer.tokenize('''I was born in 92000, and this is falsé.''')
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''9''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''é''',
'''.''',
] , )
UpperCAmelCase_ = tokenizer.convert_tokens_to_ids(_lowercase)
self.assertListEqual(
_lowercase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
UpperCAmelCase_ = tokenizer.convert_ids_to_tokens(_lowercase)
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''<unk>''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''<unk>''',
'''.''',
] , )
@cached_property
def __a ( self :Any) -> List[Any]:
return BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
@slow
def __a ( self :int) -> List[Any]:
UpperCAmelCase_ = '''Hello World!'''
UpperCAmelCase_ = [65, 18536, 2260, 101, 66]
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@slow
def __a ( self :int) -> Any:
UpperCAmelCase_ = (
'''This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) " [ ] ! : - . Also we will'''
''' add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth'''
)
# fmt: off
UpperCAmelCase_ = [65, 871, 419, 358, 946, 991, 2521, 452, 358, 1357, 387, 7751, 3536, 112, 985, 456, 126, 865, 938, 5400, 5734, 458, 1368, 467, 786, 2462, 5246, 1159, 633, 865, 4519, 457, 582, 852, 2557, 427, 916, 508, 405, 34324, 497, 391, 408, 11342, 1244, 385, 100, 938, 985, 456, 574, 362, 12597, 3200, 3129, 1172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@require_torch
@slow
def __a ( self :Dict) -> Union[str, Any]:
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
UpperCAmelCase_ = list(self.big_tokenizer.get_vocab().keys())[:10]
UpperCAmelCase_ = ''' '''.join(_lowercase)
UpperCAmelCase_ = self.big_tokenizer.encode_plus(_lowercase , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = self.big_tokenizer.batch_encode_plus(
[sequence + ''' ''' + sequence] , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = BigBirdConfig(attention_type='''original_full''')
UpperCAmelCase_ = BigBirdModel(_lowercase)
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_lowercase)
model(**_lowercase)
@slow
def __a ( self :Optional[int]) -> Any:
UpperCAmelCase_ = BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
UpperCAmelCase_ = tokenizer.decode(tokenizer('''Paris is the [MASK].''').input_ids)
self.assertTrue(decoded_text == '''[CLS] Paris is the[MASK].[SEP]''')
@slow
def __a ( self :Dict) -> List[str]:
# fmt: off
UpperCAmelCase_ = {'''input_ids''': [[65, 39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114, 66], [65, 448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], '''attention_mask''': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_lowercase , model_name='''google/bigbird-roberta-base''' , revision='''215c99f1600e06f83acce68422f2035b2b5c3510''' , )
| 344 | 1 |
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
while a != 0:
UpperCAmelCase_ , UpperCAmelCase_ = b % a, a
return b
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
if gcd(__UpperCAmelCase , __UpperCAmelCase ) != 1:
UpperCAmelCase_ = f"mod inverse of {a!r} and {m!r} does not exist"
raise ValueError(__UpperCAmelCase )
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = 1, 0, a
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = 0, 1, m
while va != 0:
UpperCAmelCase_ = ua // va
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = (ua - q * va), (ua - q * va), (ua - q * va), va, va, va
return ua % m
| 344 |
from __future__ import annotations
import string
from itertools import cycle, product
from pathlib import Path
UpperCamelCase_ = (
string.ascii_letters + string.digits + string.punctuation + string.whitespace
)
UpperCamelCase_ = [ord(letter) for letter in string.ascii_lowercase]
UpperCamelCase_ = {ord(char) for char in VALID_CHARS}
UpperCamelCase_ = ["the", "be", "to", "of", "and", "in", "that", "have"]
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str | None:
'''simple docstring'''
UpperCAmelCase_ = ""
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
for keychar, cipherchar in zip(cycle(__UpperCAmelCase ) , __UpperCAmelCase ):
UpperCAmelCase_ = cipherchar ^ keychar
if decodedchar not in VALID_INTS:
return None
decoded += chr(__UpperCAmelCase )
return decoded
def A ( __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
for key in product(__UpperCAmelCase , repeat=3 ):
UpperCAmelCase_ = try_key(__UpperCAmelCase , __UpperCAmelCase )
if encoded is not None:
possibles.append(__UpperCAmelCase )
return possibles
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
return [possible for possible in possibles if common_word in possible.lower()]
def A ( __UpperCAmelCase = "p059_cipher.txt" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = Path(__UpperCAmelCase ).parent.joinpath(__UpperCAmelCase ).read_text(encoding='''utf-8''' )
UpperCAmelCase_ = [int(__UpperCAmelCase ) for number in data.strip().split(''',''' )]
UpperCAmelCase_ = filter_valid_chars(__UpperCAmelCase )
for common_word in COMMON_WORDS:
UpperCAmelCase_ = filter_common_word(__UpperCAmelCase , __UpperCAmelCase )
if len(__UpperCAmelCase ) == 1:
break
UpperCAmelCase_ = possibles[0]
return sum(ord(__UpperCAmelCase ) for char in decoded_text )
if __name__ == "__main__":
print(f"{solution() = }")
| 344 | 1 |
import os
import time
import pytest
from datasets.utils.filelock import FileLock, Timeout
def A ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
UpperCAmelCase_ = FileLock(str(tmpdir / '''foo.lock''' ) )
UpperCAmelCase_ = FileLock(str(tmpdir / '''foo.lock''' ) )
UpperCAmelCase_ = 0.01
with locka.acquire():
with pytest.raises(__UpperCAmelCase ):
UpperCAmelCase_ = time.time()
locka.acquire(__UpperCAmelCase )
assert time.time() - _start > timeout
def A ( __UpperCAmelCase ) -> Any:
'''simple docstring'''
UpperCAmelCase_ = '''a''' * 1000 + '''.lock'''
UpperCAmelCase_ = FileLock(str(tmpdir / filename ) )
assert locka._lock_file.endswith('''.lock''' )
assert not locka._lock_file.endswith(__UpperCAmelCase )
assert len(os.path.basename(locka._lock_file ) ) <= 255
UpperCAmelCase_ = FileLock(tmpdir / filename )
with locka.acquire():
with pytest.raises(__UpperCAmelCase ):
locka.acquire(0 )
| 344 |
import pytest
UpperCamelCase_ = "__dummy_dataset1__"
UpperCamelCase_ = "\nimport json\nimport os\n\nimport datasets\n\n\nREPO_URL = \"https://huggingface.co/datasets/albertvillanova/tests-raw-jsonl/resolve/main/\"\nURLS = {\"train\": REPO_URL + \"wikiann-bn-train.jsonl\", \"validation\": REPO_URL + \"wikiann-bn-validation.jsonl\"}\n\n\nclass __DummyDataset1__(datasets.GeneratorBasedBuilder):\n\n def _info(self):\n features = datasets.Features(\n {\n \"tokens\": datasets.Sequence(datasets.Value(\"string\")),\n \"ner_tags\": datasets.Sequence(\n datasets.features.ClassLabel(\n names=[\n \"O\",\n \"B-PER\",\n \"I-PER\",\n \"B-ORG\",\n \"I-ORG\",\n \"B-LOC\",\n \"I-LOC\",\n ]\n )\n ),\n \"langs\": datasets.Sequence(datasets.Value(\"string\")),\n \"spans\": datasets.Sequence(datasets.Value(\"string\")),\n }\n )\n return datasets.DatasetInfo(features=features)\n\n def _split_generators(self, dl_manager):\n dl_path = dl_manager.download(URLS)\n return [\n datasets.SplitGenerator(datasets.Split.TRAIN, gen_kwargs={\"filepath\": dl_path[\"train\"]}),\n datasets.SplitGenerator(datasets.Split.VALIDATION, gen_kwargs={\"filepath\": dl_path[\"validation\"]}),\n ]\n\n def _generate_examples(self, filepath):\n with open(filepath, \"r\", encoding=\"utf-8\") as f:\n for i, line in enumerate(f):\n yield i, json.loads(line)\n"
@pytest.fixture
def A ( ) -> str:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_NAME
@pytest.fixture
def A ( ) -> Any:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_CODE
@pytest.fixture
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = dataset_loading_script_name
UpperCAmelCase_ = tmp_path / '''datasets''' / script_name
script_dir.mkdir(parents=__UpperCAmelCase )
UpperCAmelCase_ = script_dir / f"{script_name}.py"
with open(__UpperCAmelCase , '''w''' ) as f:
f.write(__UpperCAmelCase )
return str(__UpperCAmelCase )
| 344 | 1 |
import operator as op
UpperCamelCase_ = "scaler.pt"
UpperCamelCase_ = "pytorch_model"
UpperCamelCase_ = "random_states"
UpperCamelCase_ = "optimizer"
UpperCamelCase_ = "scheduler"
UpperCamelCase_ = "pytorch_model.bin"
UpperCamelCase_ = "pytorch_model.bin.index.json"
UpperCamelCase_ = "model.safetensors"
UpperCamelCase_ = "model.safetensors.index.json"
UpperCamelCase_ = "1.10.2"
UpperCamelCase_ = "py38"
UpperCamelCase_ = "4.17.0"
UpperCamelCase_ = ["ml.p3.16xlarge", "ml.p3dn.24xlarge", "ml.p4dn.24xlarge"]
UpperCamelCase_ = ["FULL_SHARD", "SHARD_GRAD_OP", "NO_SHARD", "HYBRID_SHARD", "HYBRID_SHARD_ZERO2"]
UpperCamelCase_ = ["TRANSFORMER_BASED_WRAP", "SIZE_BASED_WRAP", "NO_WRAP"]
UpperCamelCase_ = ["BACKWARD_PRE", "BACKWARD_POST", "NO_PREFETCH"]
UpperCamelCase_ = ["FULL_STATE_DICT", "LOCAL_STATE_DICT", "SHARDED_STATE_DICT"]
UpperCamelCase_ = "2.0.1"
UpperCamelCase_ = ["pdsh", "standard", "openmpi", "mvapich"]
UpperCamelCase_ = ["default", "reduce-overhead", "max-autotune"]
UpperCamelCase_ = {">": op.gt, ">=": op.ge, "==": op.eq, "!=": op.ne, "<=": op.le, "<": op.lt}
# These are the args for `torch.distributed.launch` for pytorch < 1.9
UpperCamelCase_ = [
"nnodes",
"nproc_per_node",
"rdzv_backend",
"rdzv_endpoint",
"rdzv_id",
"rdzv_conf",
"standalone",
"max_restarts",
"monitor_interval",
"start_method",
"role",
"module",
"m",
"no_python",
"run_path",
"log_dir",
"r",
"redirects",
"t",
"tee",
"node_rank",
"master_addr",
"master_port",
]
UpperCamelCase_ = ["DEEPSPEED", "MULTI_GPU", "FSDP", "MEGATRON_LM"]
UpperCamelCase_ = ["DEEPSPEED", "MULTI_XPU", "FSDP"]
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"s-JoL/Open-Llama-V1": "https://huggingface.co/s-JoL/Open-Llama-V1/blob/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="open-llama"
def __init__( self :Union[str, Any] , _lowercase :List[Any]=100000 , _lowercase :Dict=4096 , _lowercase :List[Any]=11008 , _lowercase :Optional[int]=32 , _lowercase :Union[str, Any]=32 , _lowercase :List[str]="silu" , _lowercase :Union[str, Any]=2048 , _lowercase :Any=0.02 , _lowercase :Optional[Any]=1E-6 , _lowercase :str=True , _lowercase :str=0 , _lowercase :Any=1 , _lowercase :Optional[Any]=2 , _lowercase :str=False , _lowercase :Dict=True , _lowercase :Optional[Any]=0.1 , _lowercase :Tuple=0.1 , _lowercase :Dict=True , _lowercase :List[Any]=True , _lowercase :Dict=None , **_lowercase :Optional[int] , ) -> List[Any]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = rms_norm_eps
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = kwargs.pop(
'''use_memorry_efficient_attention''' , _lowercase)
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_dropout_prob
UpperCAmelCase_ = use_stable_embedding
UpperCAmelCase_ = shared_input_output_embedding
UpperCAmelCase_ = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , tie_word_embeddings=_lowercase , **_lowercase , )
def __a ( self :int) -> str:
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _lowercase) or len(self.rope_scaling) != 2:
raise ValueError(
'''`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, '''
f"got {self.rope_scaling}")
UpperCAmelCase_ = self.rope_scaling.get('''type''' , _lowercase)
UpperCAmelCase_ = self.rope_scaling.get('''factor''' , _lowercase)
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}")
if rope_scaling_factor is None or not isinstance(_lowercase , _lowercase) or rope_scaling_factor <= 1.0:
raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
| 344 | 1 |
class a_ :
def __init__( self :Optional[int] , _lowercase :List[Any] , _lowercase :Any , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = None
UpperCAmelCase_ = None
UpperCAmelCase_ = graph
self._normalize_graph(_lowercase , _lowercase)
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = None
def __a ( self :Any , _lowercase :List[Any] , _lowercase :Tuple) -> List[str]:
if sources is int:
UpperCAmelCase_ = [sources]
if sinks is int:
UpperCAmelCase_ = [sinks]
if len(_lowercase) == 0 or len(_lowercase) == 0:
return
UpperCAmelCase_ = sources[0]
UpperCAmelCase_ = sinks[0]
# make fake vertex if there are more
# than one source or sink
if len(_lowercase) > 1 or len(_lowercase) > 1:
UpperCAmelCase_ = 0
for i in sources:
max_input_flow += sum(self.graph[i])
UpperCAmelCase_ = len(self.graph) + 1
for room in self.graph:
room.insert(0 , 0)
self.graph.insert(0 , [0] * size)
for i in sources:
UpperCAmelCase_ = max_input_flow
UpperCAmelCase_ = 0
UpperCAmelCase_ = len(self.graph) + 1
for room in self.graph:
room.append(0)
self.graph.append([0] * size)
for i in sinks:
UpperCAmelCase_ = max_input_flow
UpperCAmelCase_ = size - 1
def __a ( self :int) -> Tuple:
if self.maximum_flow_algorithm is None:
raise Exception('''You need to set maximum flow algorithm before.''')
if self.source_index is None or self.sink_index is None:
return 0
self.maximum_flow_algorithm.execute()
return self.maximum_flow_algorithm.getMaximumFlow()
def __a ( self :Union[str, Any] , _lowercase :List[str]) -> List[Any]:
UpperCAmelCase_ = algorithm(self)
class a_ :
def __init__( self :Any , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = flow_network
UpperCAmelCase_ = flow_network.verticesCount
UpperCAmelCase_ = flow_network.sourceIndex
UpperCAmelCase_ = flow_network.sinkIndex
# it's just a reference, so you shouldn't change
# it in your algorithms, use deep copy before doing that
UpperCAmelCase_ = flow_network.graph
UpperCAmelCase_ = False
def __a ( self :str) -> List[Any]:
if not self.executed:
self._algorithm()
UpperCAmelCase_ = True
def __a ( self :Dict) -> Optional[int]:
pass
class a_ ( _snake_case ):
def __init__( self :Optional[Any] , _lowercase :List[str]) -> Tuple:
super().__init__(_lowercase)
# use this to save your result
UpperCAmelCase_ = -1
def __a ( self :Any) -> int:
if not self.executed:
raise Exception('''You should execute algorithm before using its result!''')
return self.maximum_flow
class a_ ( _snake_case ):
def __init__( self :str , _lowercase :List[Any]) -> Optional[int]:
super().__init__(_lowercase)
UpperCAmelCase_ = [[0] * self.verticies_count for i in range(self.verticies_count)]
UpperCAmelCase_ = [0] * self.verticies_count
UpperCAmelCase_ = [0] * self.verticies_count
def __a ( self :str) -> Tuple:
UpperCAmelCase_ = self.verticies_count
# push some substance to graph
for nextvertex_index, bandwidth in enumerate(self.graph[self.source_index]):
self.preflow[self.source_index][nextvertex_index] += bandwidth
self.preflow[nextvertex_index][self.source_index] -= bandwidth
self.excesses[nextvertex_index] += bandwidth
# Relabel-to-front selection rule
UpperCAmelCase_ = [
i
for i in range(self.verticies_count)
if i != self.source_index and i != self.sink_index
]
# move through list
UpperCAmelCase_ = 0
while i < len(_lowercase):
UpperCAmelCase_ = vertices_list[i]
UpperCAmelCase_ = self.heights[vertex_index]
self.process_vertex(_lowercase)
if self.heights[vertex_index] > previous_height:
# if it was relabeled, swap elements
# and start from 0 index
vertices_list.insert(0 , vertices_list.pop(_lowercase))
UpperCAmelCase_ = 0
else:
i += 1
UpperCAmelCase_ = sum(self.preflow[self.source_index])
def __a ( self :Union[str, Any] , _lowercase :Any) -> Tuple:
while self.excesses[vertex_index] > 0:
for neighbour_index in range(self.verticies_count):
# if it's neighbour and current vertex is higher
if (
self.graph[vertex_index][neighbour_index]
- self.preflow[vertex_index][neighbour_index]
> 0
and self.heights[vertex_index] > self.heights[neighbour_index]
):
self.push(_lowercase , _lowercase)
self.relabel(_lowercase)
def __a ( self :Any , _lowercase :int , _lowercase :str) -> str:
UpperCAmelCase_ = min(
self.excesses[from_index] , self.graph[from_index][to_index] - self.preflow[from_index][to_index] , )
self.preflow[from_index][to_index] += preflow_delta
self.preflow[to_index][from_index] -= preflow_delta
self.excesses[from_index] -= preflow_delta
self.excesses[to_index] += preflow_delta
def __a ( self :Optional[Any] , _lowercase :Any) -> int:
UpperCAmelCase_ = None
for to_index in range(self.verticies_count):
if (
self.graph[vertex_index][to_index]
- self.preflow[vertex_index][to_index]
> 0
) and (min_height is None or self.heights[to_index] < min_height):
UpperCAmelCase_ = self.heights[to_index]
if min_height is not None:
UpperCAmelCase_ = min_height + 1
if __name__ == "__main__":
UpperCamelCase_ = [0]
UpperCamelCase_ = [3]
# graph = [
# [0, 0, 4, 6, 0, 0],
# [0, 0, 5, 2, 0, 0],
# [0, 0, 0, 0, 4, 4],
# [0, 0, 0, 0, 6, 6],
# [0, 0, 0, 0, 0, 0],
# [0, 0, 0, 0, 0, 0],
# ]
UpperCamelCase_ = [[0, 7, 0, 0], [0, 0, 6, 0], [0, 0, 0, 8], [9, 0, 0, 0]]
# prepare our network
UpperCamelCase_ = FlowNetwork(graph, entrances, exits)
# set algorithm
flow_network.set_maximum_flow_algorithm(PushRelabelExecutor)
# and calculate
UpperCamelCase_ = flow_network.find_maximum_flow()
print(f"maximum flow is {maximum_flow}")
| 344 |
import tempfile
import torch
from diffusers import (
DEISMultistepScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
UniPCMultistepScheduler,
)
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] =(DPMSolverSinglestepScheduler,)
UpperCamelCase__ : Tuple =(("num_inference_steps", 25),)
def __a ( self :List[Any] , **_lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
'''solver_order''': 2,
'''prediction_type''': '''epsilon''',
'''thresholding''': False,
'''sample_max_value''': 1.0,
'''algorithm_type''': '''dpmsolver++''',
'''solver_type''': '''midpoint''',
'''lambda_min_clipped''': -float('''inf'''),
'''variance_type''': None,
}
config.update(**_lowercase)
return config
def __a ( self :Union[str, Any] , _lowercase :List[Any]=0 , **_lowercase :Optional[int]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ , UpperCAmelCase_ = sample, sample
for t in range(_lowercase , time_step + scheduler.config.solver_order + 1):
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Union[str, Any]) -> List[Any]:
pass
def __a ( self :Optional[Any] , _lowercase :str=0 , **_lowercase :Union[str, Any]) -> Dict:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Dict , _lowercase :Union[str, Any]=None , **_lowercase :List[Any]) -> int:
if scheduler is None:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :int) -> Tuple:
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = 50
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
# make sure that the first t is uneven
for i, t in enumerate(scheduler.timesteps[3:]):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_574) < 1E-3
def __a ( self :List[Any]) -> List[Any]:
for timesteps in [25, 50, 100, 999, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :int) -> Optional[Any]:
# make sure that iterating over schedulers with same config names gives same results
# for defaults
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
UpperCAmelCase_ = DEISMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = UniPCMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverSinglestepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Tuple) -> int:
self.check_over_configs(thresholding=_lowercase)
for order in [1, 2, 3]:
for solver_type in ["midpoint", "heun"]:
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
thresholding=_lowercase , prediction_type=_lowercase , sample_max_value=_lowercase , algorithm_type='''dpmsolver++''' , solver_order=_lowercase , solver_type=_lowercase , )
def __a ( self :List[Any]) -> Any:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Optional[int]:
for algorithm_type in ["dpmsolver", "dpmsolver++"]:
for solver_type in ["midpoint", "heun"]:
for order in [1, 2, 3]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
UpperCAmelCase_ = self.full_loop(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
assert not torch.isnan(_lowercase).any(), "Samples have nan numbers"
def __a ( self :Tuple) -> int:
self.check_over_configs(lower_order_final=_lowercase)
self.check_over_configs(lower_order_final=_lowercase)
def __a ( self :Tuple) -> Optional[Any]:
self.check_over_configs(lambda_min_clipped=-float('''inf'''))
self.check_over_configs(lambda_min_clipped=-5.1)
def __a ( self :Any) -> List[str]:
self.check_over_configs(variance_type=_lowercase)
self.check_over_configs(variance_type='''learned_range''')
def __a ( self :Any) -> Dict:
for num_inference_steps in [1, 2, 3, 5, 10, 50, 100, 999, 1000]:
self.check_over_forward(num_inference_steps=_lowercase , time_step=0)
def __a ( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Any) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop(use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_248) < 1E-3
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.1_453) < 1E-3
def __a ( self :List[Any]) -> Dict:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''' , use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.0_649) < 1E-3
def __a ( self :Any) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(thresholding=_lowercase , dynamic_thresholding_ratio=0)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter.half()
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
assert sample.dtype == torch.floataa
| 344 | 1 |
import math
class a_ :
def __init__( self :Dict , _lowercase :Union[str, Any]=0) -> Union[str, Any]: # a graph with Node 0,1,...,N-1
UpperCAmelCase_ = n
UpperCAmelCase_ = [
[math.inf for j in range(0 , _lowercase)] for i in range(0 , _lowercase)
] # adjacency matrix for weight
UpperCAmelCase_ = [
[math.inf for j in range(0 , _lowercase)] for i in range(0 , _lowercase)
] # dp[i][j] stores minimum distance from i to j
def __a ( self :Any , _lowercase :List[str] , _lowercase :List[str] , _lowercase :Optional[int]) -> Any:
UpperCAmelCase_ = w
def __a ( self :str) -> Optional[Any]:
for k in range(0 , self.n):
for i in range(0 , self.n):
for j in range(0 , self.n):
UpperCAmelCase_ = min(self.dp[i][j] , self.dp[i][k] + self.dp[k][j])
def __a ( self :int , _lowercase :Tuple , _lowercase :Optional[int]) -> List[str]:
return self.dp[u][v]
if __name__ == "__main__":
UpperCamelCase_ = Graph(5)
graph.add_edge(0, 2, 9)
graph.add_edge(0, 4, 10)
graph.add_edge(1, 3, 5)
graph.add_edge(2, 3, 7)
graph.add_edge(3, 0, 10)
graph.add_edge(3, 1, 2)
graph.add_edge(3, 2, 1)
graph.add_edge(3, 4, 6)
graph.add_edge(4, 1, 3)
graph.add_edge(4, 2, 4)
graph.add_edge(4, 3, 9)
graph.floyd_warshall()
graph.show_min(1, 4)
graph.show_min(0, 3)
| 344 |
import inspect
import unittest
import torch
import torch.nn as nn
from accelerate.hooks import (
AlignDevicesHook,
ModelHook,
SequentialHook,
add_hook_to_module,
attach_align_device_hook,
remove_hook_from_module,
remove_hook_from_submodules,
)
from accelerate.test_utils import require_multi_gpu
class a_ ( nn.Module ):
def __init__( self :Optional[Any]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(3 , 4)
UpperCAmelCase_ = nn.BatchNormad(4)
UpperCAmelCase_ = nn.Linear(4 , 5)
def __a ( self :Dict , _lowercase :int) -> str:
return self.lineara(self.batchnorm(self.lineara(_lowercase)))
class a_ ( _snake_case ):
def __a ( self :Tuple , _lowercase :Optional[int] , *_lowercase :Union[str, Any] , **_lowercase :Any) -> Optional[Any]:
return (args[0] + 1,) + args[1:], kwargs
class a_ ( _snake_case ):
def __a ( self :Union[str, Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
return output + 1
class a_ ( unittest.TestCase ):
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
self.assertEqual(test_model._hf_hook , _lowercase)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[Any]) -> Any:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
add_hook_to_module(_lowercase , _lowercase , append=_lowercase)
self.assertEqual(isinstance(test_model._hf_hook , _lowercase) , _lowercase)
self.assertEqual(len(test_model._hf_hook.hooks) , 2)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[int]) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(x + 1)
UpperCAmelCase_ = test_model(x + 2)
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PreForwardHook() , PreForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , _lowercase , atol=1E-5)
def __a ( self :List[str]) -> int:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PostForwardHook() , PostForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , output + 2 , atol=1E-5)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1))
self.assertTrue(outputa.requires_grad)
UpperCAmelCase_ = True
UpperCAmelCase_ = test_model(_lowercase)
self.assertFalse(outputa.requires_grad)
@require_multi_gpu
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.batchnorm , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=1))
self.assertEqual(model.lineara.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.running_mean.device , torch.device(0))
self.assertEqual(model.lineara.weight.device , torch.device(1))
# We can still make a forward pass. The input does not need to be on any particular device
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(1))
# We can add a general hook to put back output on same device as input.
add_hook_to_module(_lowercase , AlignDevicesHook(io_same_device=_lowercase))
UpperCAmelCase_ = torch.randn(2 , 3).to(0)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(0))
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = {'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''', '''offload''': True}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(hook_kwargs['''execution_device'''])
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
UpperCAmelCase_ = {
'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''',
'''offload''': True,
'''offload_buffers''': True,
}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase)
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase , offload_buffers=_lowercase)
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict())
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict() , offload_buffers=_lowercase , )
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
| 344 | 1 |
import importlib
import torch
import yaml
from omegaconf import OmegaConf
from taming.models.vqgan import VQModel
def A ( __UpperCAmelCase , __UpperCAmelCase=False ) -> Optional[Any]:
'''simple docstring'''
UpperCAmelCase_ = OmegaConf.load(__UpperCAmelCase )
if display:
print(yaml.dump(OmegaConf.to_container(__UpperCAmelCase ) ) )
return config
def A ( __UpperCAmelCase , __UpperCAmelCase=None , __UpperCAmelCase=None ) -> List[Any]:
'''simple docstring'''
if conf_path is None:
UpperCAmelCase_ = '''./model_checkpoints/vqgan_only.yaml'''
UpperCAmelCase_ = load_config(__UpperCAmelCase , display=__UpperCAmelCase )
UpperCAmelCase_ = VQModel(**config.model.params )
if ckpt_path is None:
UpperCAmelCase_ = '''./model_checkpoints/vqgan_only.pt'''
UpperCAmelCase_ = torch.load(__UpperCAmelCase , map_location=__UpperCAmelCase )
if ".ckpt" in ckpt_path:
UpperCAmelCase_ = sd['''state_dict''']
model.load_state_dict(__UpperCAmelCase , strict=__UpperCAmelCase )
model.to(__UpperCAmelCase )
del sd
return model
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = model.encode(__UpperCAmelCase )
print(f"VQGAN --- {model.__class__.__name__}: latent shape: {z.shape[2:]}" )
UpperCAmelCase_ = model.decode(__UpperCAmelCase )
return xrec
def A ( __UpperCAmelCase , __UpperCAmelCase=False ) -> Optional[int]:
'''simple docstring'''
UpperCAmelCase_ , UpperCAmelCase_ = string.rsplit('''.''' , 1 )
if reload:
UpperCAmelCase_ = importlib.import_module(__UpperCAmelCase )
importlib.reload(__UpperCAmelCase )
return getattr(importlib.import_module(__UpperCAmelCase , package=__UpperCAmelCase ) , cls )
def A ( __UpperCAmelCase ) -> str:
'''simple docstring'''
if "target" not in config:
raise KeyError('''Expected key `target` to instantiate.''' )
return get_obj_from_str(config['''target'''] )(**config.get('''params''' , {} ) )
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase=True , __UpperCAmelCase=True ) -> Optional[int]:
'''simple docstring'''
UpperCAmelCase_ = instantiate_from_config(__UpperCAmelCase )
if sd is not None:
model.load_state_dict(__UpperCAmelCase )
if gpu:
model.cuda()
if eval_mode:
model.eval()
return {"model": model}
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> str:
'''simple docstring'''
if ckpt:
UpperCAmelCase_ = torch.load(__UpperCAmelCase , map_location='''cpu''' )
UpperCAmelCase_ = pl_sd['''global_step''']
print(f"loaded model from global step {global_step}." )
else:
UpperCAmelCase_ = {'''state_dict''': None}
UpperCAmelCase_ = None
UpperCAmelCase_ = load_model_from_config(config.model , pl_sd['''state_dict'''] , gpu=__UpperCAmelCase , eval_mode=__UpperCAmelCase )['''model''']
return model, global_step
| 344 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import (
AutoProcessor,
BertTokenizerFast,
BlipImageProcessor,
GPTaTokenizer,
InstructBlipProcessor,
PreTrainedTokenizerFast,
)
@require_vision
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = BlipImageProcessor()
UpperCAmelCase_ = GPTaTokenizer.from_pretrained('''hf-internal-testing/tiny-random-GPT2Model''')
UpperCAmelCase_ = BertTokenizerFast.from_pretrained('''hf-internal-testing/tiny-random-bert''')
UpperCAmelCase_ = InstructBlipProcessor(_lowercase , _lowercase , _lowercase)
processor.save_pretrained(self.tmpdirname)
def __a ( self :List[Any] , **_lowercase :Dict) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).tokenizer
def __a ( self :Optional[Any] , **_lowercase :Optional[Any]) -> Optional[int]:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).image_processor
def __a ( self :Dict , **_lowercase :Tuple) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).qformer_tokenizer
def __a ( self :Optional[int]) -> str:
shutil.rmtree(self.tmpdirname)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta)]
UpperCAmelCase_ = [Image.fromarray(np.moveaxis(_lowercase , 0 , -1)) for x in image_inputs]
return image_inputs
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() , qformer_tokenizer=self.get_qformer_tokenizer() , )
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = self.get_image_processor(do_normalize=_lowercase , padding_value=1.0)
UpperCAmelCase_ = InstructBlipProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowercase , padding_value=1.0)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
self.assertIsInstance(processor.tokenizer , _lowercase)
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string())
self.assertIsInstance(processor.image_processor , _lowercase)
self.assertIsInstance(processor.qformer_tokenizer , _lowercase)
def __a ( self :Dict) -> Any:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = image_processor(_lowercase , return_tensors='''np''')
UpperCAmelCase_ = processor(images=_lowercase , return_tensors='''np''')
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1E-2)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = processor(text=_lowercase)
UpperCAmelCase_ = tokenizer(_lowercase , return_token_type_ids=_lowercase)
UpperCAmelCase_ = qformer_tokenizer(_lowercase , return_token_type_ids=_lowercase)
for key in encoded_tokens.keys():
self.assertListEqual(encoded_tokens[key] , encoded_processor[key])
for key in encoded_tokens_qformer.keys():
self.assertListEqual(encoded_tokens_qformer[key] , encoded_processor['''qformer_''' + key])
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
# test if it raises when no input is passed
with pytest.raises(_lowercase):
processor()
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
UpperCAmelCase_ = processor.batch_decode(_lowercase)
UpperCAmelCase_ = tokenizer.batch_decode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :str) -> int:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
| 344 | 1 |
def A ( __UpperCAmelCase , __UpperCAmelCase = 0 ) -> list:
'''simple docstring'''
UpperCAmelCase_ = length or len(__UpperCAmelCase )
UpperCAmelCase_ = False
for i in range(length - 1 ):
if list_data[i] > list_data[i + 1]:
UpperCAmelCase_ , UpperCAmelCase_ = list_data[i + 1], list_data[i]
UpperCAmelCase_ = True
return list_data if not swapped else bubble_sort(__UpperCAmelCase , length - 1 )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"facebook/levit-128S": "https://huggingface.co/facebook/levit-128S/resolve/main/config.json",
# See all LeViT models at https://huggingface.co/models?filter=levit
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[int] ="levit"
def __init__( self :List[str] , _lowercase :List[Any]=224 , _lowercase :str=3 , _lowercase :Optional[int]=3 , _lowercase :str=2 , _lowercase :List[Any]=1 , _lowercase :str=16 , _lowercase :Dict=[128, 256, 384] , _lowercase :Union[str, Any]=[4, 8, 12] , _lowercase :Tuple=[4, 4, 4] , _lowercase :Dict=[16, 16, 16] , _lowercase :Any=0 , _lowercase :Dict=[2, 2, 2] , _lowercase :Any=[2, 2, 2] , _lowercase :Tuple=0.02 , **_lowercase :Union[str, Any] , ) -> Optional[Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = kernel_size
UpperCAmelCase_ = stride
UpperCAmelCase_ = padding
UpperCAmelCase_ = hidden_sizes
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = depths
UpperCAmelCase_ = key_dim
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = attention_ratio
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = [
['''Subsample''', key_dim[0], hidden_sizes[0] // key_dim[0], 4, 2, 2],
['''Subsample''', key_dim[0], hidden_sizes[1] // key_dim[0], 4, 2, 2],
]
class a_ ( _snake_case ):
UpperCamelCase__ : Union[str, Any] =version.parse("1.11" )
@property
def __a ( self :Any) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
])
@property
def __a ( self :List[Any]) -> float:
return 1E-4
| 344 | 1 |
from collections import deque
from .hash_table import HashTable
class a_ ( _snake_case ):
def __init__( self :Union[str, Any] , *_lowercase :Optional[int] , **_lowercase :List[str]) -> int:
super().__init__(*_lowercase , **_lowercase)
def __a ( self :Optional[int] , _lowercase :str , _lowercase :Optional[Any]) -> List[str]:
UpperCAmelCase_ = deque([]) if self.values[key] is None else self.values[key]
self.values[key].appendleft(_lowercase)
UpperCAmelCase_ = self.values[key]
def __a ( self :Optional[Any]) -> List[Any]:
return (
sum(self.charge_factor - len(_lowercase) for slot in self.values)
/ self.size_table
* self.charge_factor
)
def __a ( self :Any , _lowercase :Union[str, Any] , _lowercase :int=None) -> Tuple:
if not (
len(self.values[key]) == self.charge_factor and self.values.count(_lowercase) == 0
):
return key
return super()._collision_resolution(_lowercase , _lowercase)
| 344 |
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=None , __UpperCAmelCase="no" , __UpperCAmelCase="29500" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = False
UpperCAmelCase_ = False
if any(key.startswith('''KAGGLE''' ) for key in os.environ.keys() ):
UpperCAmelCase_ = True
elif "IPython" in sys.modules:
UpperCAmelCase_ = '''google.colab''' in str(sys.modules['''IPython'''].get_ipython() )
try:
UpperCAmelCase_ = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
f"Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}." )
if (in_colab or in_kaggle) and (os.environ.get('''TPU_NAME''' , __UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '''
'''your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if num_processes is None:
UpperCAmelCase_ = 8
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''TPU''' )
print(f"Launching a training on {num_processes} TPU cores." )
xmp.spawn(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on one CPU.''' )
function(*__UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'''You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.''' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '''
'''inside your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if torch.cuda.is_initialized():
raise ValueError(
'''To launch a multi-GPU training from your notebook, you need to avoid running any instruction '''
'''using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '''
'''function.''' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port=__UpperCAmelCase , mixed_precision=__UpperCAmelCase ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''MULTI_GPU''' )
print(f"Launching training on {num_processes} GPUs." )
try:
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'''CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '''
'''This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '''
'''Please review your imports and test them when running the `notebook_launcher()` to identify '''
'''which one is problematic.''' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
UpperCAmelCase_ = '''1'''
print('''Launching training on MPS.''' )
elif torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on CPU.''' )
function(*__UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=2 ) -> Optional[Any]:
'''simple docstring'''
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port='''29500''' , accelerate_mixed_precision='''no''' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='''yes''' , ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , debug=__UpperCAmelCase )
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
| 344 | 1 |
import sys
from .dependency_versions_table import deps
from .utils.versions import require_version, require_version_core
# define which module versions we always want to check at run time
# (usually the ones defined in `install_requires` in setup.py)
#
# order specific notes:
# - tqdm must be checked before tokenizers
UpperCamelCase_ = "python tqdm regex requests packaging filelock numpy tokenizers".split()
if sys.version_info < (3, 7):
pkgs_to_check_at_runtime.append("dataclasses")
if sys.version_info < (3, 8):
pkgs_to_check_at_runtime.append("importlib_metadata")
for pkg in pkgs_to_check_at_runtime:
if pkg in deps:
if pkg == "tokenizers":
# must be loaded here, or else tqdm check may fail
from .utils import is_tokenizers_available
if not is_tokenizers_available():
continue # not required, check version only if installed
require_version_core(deps[pkg])
else:
raise ValueError(f"can't find {pkg} in {deps.keys()}, check dependency_versions_table.py")
def A ( __UpperCAmelCase , __UpperCAmelCase=None ) -> Dict:
'''simple docstring'''
require_version(deps[pkg] , __UpperCAmelCase )
| 344 |
import math
import os
import re
import sys
import unittest
from pathlib import Path
from typing import Tuple
from unittest.mock import patch
from parameterized import parameterized
from transformers.testing_utils import (
CaptureStderr,
ExtendSysPath,
TestCasePlus,
execute_subprocess_async,
get_gpu_count,
get_torch_dist_unique_port,
require_apex,
require_bitsandbytes,
require_fairscale,
require_torch,
require_torch_gpu,
require_torch_multi_gpu,
require_torch_non_multi_gpu,
slow,
)
from transformers.trainer_callback import TrainerState
from transformers.trainer_utils import set_seed
UpperCamelCase_ = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f"{bindir}/../../examples/pytorch/translation"):
from run_translation import main # noqa
set_seed(42)
UpperCamelCase_ = "sshleifer/student_marian_en_ro_6_1"
UpperCamelCase_ = "sshleifer/tiny-mbart"
@require_torch
class a_ ( _snake_case ):
def __a ( self :str , _lowercase :Any=False , _lowercase :Tuple=None , _lowercase :Dict=True , _lowercase :Tuple=True , _lowercase :List[Any]=True , _lowercase :List[str]=True , ) -> int:
UpperCAmelCase_ = self.run_trainer(
eval_steps=1 , max_len=12 , model_name=_lowercase , num_train_epochs=1 , distributed=_lowercase , extra_args_str=_lowercase , predict_with_generate=_lowercase , do_train=_lowercase , do_eval=_lowercase , do_predict=_lowercase , )
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
if not do_eval:
return
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
if predict_with_generate:
assert "eval_bleu" in first_step_stats
UpperCAmelCase_ = eval_metrics[-1]
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
assert not math.isnan(float(last_step_stats['''eval_loss'''])), "eval_loss must not be `nan`"
@require_torch_non_multi_gpu
def __a ( self :Dict) -> str:
self.run_seqaseq_quick()
@require_torch_multi_gpu
def __a ( self :Any) -> int:
self.run_seqaseq_quick(distributed=_lowercase)
@require_torch_multi_gpu
def __a ( self :int) -> Any:
self.run_seqaseq_quick(distributed=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> List[str]:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple --fp16''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Union[str, Any]) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :int) -> Any:
self.run_seqaseq_quick(
distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=_lowercase)
@require_apex
@require_torch_gpu
def __a ( self :Tuple) -> str:
# XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same
# program and it breaks other tests that run from the same pytest worker, therefore until this is
# sorted out it must be run only in an external program, that is distributed=True in this
# test and only under one or more gpus - if we want cpu will need to make a special test
#
# specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via
# 2nd main() call it botches the future eval.
#
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
# test 2nd time - was getting eval_loss': nan'
# to reproduce the problem set distributed=False
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
@parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''])
@require_torch_multi_gpu
def __a ( self :str , _lowercase :Any) -> List[str]:
# as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout
UpperCAmelCase_ = {
# test with the default log_level - should be info and thus log info once
'''base''': {'''extra_args_str''': '''''', '''n_matches''': 1},
# test with low log_level and log_level_replica - should be noisy on all processes
# now the info string should appear twice on 2 processes
'''low''': {'''extra_args_str''': '''--log_level debug --log_level_replica debug''', '''n_matches''': 2},
# test with high log_level and low log_level_replica
# now the info string should appear once only on the replica
'''high''': {'''extra_args_str''': '''--log_level error --log_level_replica debug''', '''n_matches''': 1},
# test with high log_level and log_level_replica - should be quiet on all processes
'''mixed''': {'''extra_args_str''': '''--log_level error --log_level_replica error''', '''n_matches''': 0},
}
UpperCAmelCase_ = experiments[experiment_id]
UpperCAmelCase_ = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False}
UpperCAmelCase_ = '''Running training'''
with CaptureStderr() as cl:
self.run_seqaseq_quick(**_lowercase , extra_args_str=data['''extra_args_str'''])
UpperCAmelCase_ = len(re.findall(_lowercase , cl.err))
self.assertEqual(_lowercase , data['''n_matches'''])
@slow
def __a ( self :Any) -> Dict:
UpperCAmelCase_ = self.run_trainer(
eval_steps=2 , max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=10 , distributed=_lowercase , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
UpperCAmelCase_ = eval_metrics[-1]
assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing"
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
# test if do_predict saves generations and metrics
UpperCAmelCase_ = os.listdir(_lowercase)
UpperCAmelCase_ = {os.path.basename(_lowercase) for p in contents}
assert "generated_predictions.txt" in contents
assert "predict_results.json" in contents
@slow
@require_bitsandbytes
def __a ( self :List[str]) -> str:
from transformers.training_args import OptimizerNames
def train_and_return_metrics(_lowercase :str) -> Tuple[int, float]:
UpperCAmelCase_ = '''--skip_memory_metrics 0'''
UpperCAmelCase_ = self.run_trainer(
max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=1 , optim=_lowercase , distributed=_lowercase , extra_args_str=_lowercase , do_eval=_lowercase , do_predict=_lowercase , n_gpus_to_use=1 , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(Path(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20)
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20)
UpperCAmelCase_ = logs[0]['''train_loss''']
return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value)
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value)
UpperCAmelCase_ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_peak_mem_orig + gpu_alloc_mem_orig
UpperCAmelCase_ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_total_mem_orig - gpu_total_mem_bnb
# sshleifer/student_marian_en_ro_6_1 has 54M parameter, 29M of which is `nn.Embedding` which
# doesn't get quantized and remains in fp32. Therefore we only have 25M parameters quantized
# in 2 bytes and the diff in optim memory usage is derived as so:
#
# - normal 25*8=~200MB (8 bytes per param)
# - bnb 25*2= ~50MB (2 bytes per param)
#
# Thus we should expect ~150MB total memory saved.
#
# Peak memory should be the same - the total should be different by about that same margin
#
# After leaving a small margin to accommodate for differences between gpus let's check
# that we have at least 120MB in savings
UpperCAmelCase_ = 120
# uncomment the following if this test starts failing - requires py38 for a new print feature
# gpu_peak_mem_diff = gpu_peak_mem_orig - gpu_peak_mem_bnb
# print(f"{gpu_alloc_mem_orig=}MB {gpu_peak_mem_orig=}MB {gpu_alloc_mem_orig+gpu_peak_mem_orig=}MB")
# print(f" {gpu_alloc_mem_bnb=}MB {gpu_peak_mem_bnb=}MB {gpu_alloc_mem_bnb+gpu_peak_mem_bnb=}MB")
# print(f"{gpu_alloc_mem_diff=}MB")
# print(f"{gpu_peak_mem_diff=}MB")
# print(f"{gpu_total_mem_orig=}MB, {gpu_total_mem_bnb=}MB")
# print(f"{gpu_total_mem_diff=}MB, {gpu_total_mem_diff=}MB")
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less alloc gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_alloc_mem_diff}MB, with gpu_alloc_mem_orig={gpu_alloc_mem_orig}MB and"
f" gpu_alloc_mem_bnb={gpu_alloc_mem_bnb}MB" , )
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less total gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_total_mem_diff}MB, with gpu_total_mem_orig={gpu_total_mem_orig}MB and"
f" gpu_total_mem_bnb={gpu_total_mem_bnb}MB" , )
self.assertEqual(
_lowercase , _lowercase , f"loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}")
def __a ( self :Any , _lowercase :int , _lowercase :str , _lowercase :int , _lowercase :float = 3E-3 , _lowercase :str = "adafactor" , _lowercase :bool = False , _lowercase :str = None , _lowercase :int = 0 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :int = None , ) -> List[Any]:
UpperCAmelCase_ = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro'''
UpperCAmelCase_ = self.get_auto_remove_tmp_dir()
UpperCAmelCase_ = f"\n --model_name_or_path {model_name}\n --train_file {data_dir}/train.json\n --validation_file {data_dir}/val.json\n --test_file {data_dir}/test.json\n --output_dir {output_dir}\n --overwrite_output_dir\n --max_train_samples 8\n --max_source_length {max_len}\n --max_target_length {max_len}\n --do_train\n --num_train_epochs {str(_lowercase)}\n --per_device_train_batch_size 4\n --learning_rate {learning_rate}\n --warmup_steps 8\n --logging_steps 0\n --logging_strategy no\n --save_steps {str(_lowercase)}\n --group_by_length\n --label_smoothing_factor 0.1\n --target_lang ro_RO\n --source_lang en_XX\n ".split()
UpperCAmelCase_ = f"\n --do_eval\n --per_device_eval_batch_size 4\n --max_eval_samples 8\n --val_max_target_length {max_len}\n --evaluation_strategy steps\n --eval_steps {str(_lowercase)}\n ".split()
UpperCAmelCase_ = '''
--do_predict
'''.split()
UpperCAmelCase_ = []
if do_train:
args += args_train
if do_eval:
args += args_eval
if do_predict:
args += args_predict
if predict_with_generate:
args += "--predict_with_generate".split()
if do_train:
if optim == "adafactor":
args += "--adafactor".split()
else:
args += f"--optim {optim}".split()
if extra_args_str is not None:
args += extra_args_str.split()
if distributed:
if n_gpus_to_use is None:
UpperCAmelCase_ = get_gpu_count()
UpperCAmelCase_ = get_torch_dist_unique_port()
UpperCAmelCase_ = f"\n -m torch.distributed.run\n --nproc_per_node={n_gpus_to_use}\n --master_port={master_port}\n {self.examples_dir_str}/pytorch/translation/run_translation.py\n ".split()
UpperCAmelCase_ = [sys.executable] + distributed_args + args
# keep for quick debug
# print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die
execute_subprocess_async(_lowercase , env=self.get_env())
else:
UpperCAmelCase_ = ['''run_translation.py'''] + args
with patch.object(_lowercase , '''argv''' , _lowercase):
main()
return output_dir
| 344 | 1 |
import torch
from diffusers import DDPMParallelScheduler
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : Tuple =(DDPMParallelScheduler,)
def __a ( self :Tuple , **_lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
'''variance_type''': '''fixed_small''',
'''clip_sample''': True,
}
config.update(**_lowercase)
return config
def __a ( self :Dict) -> Tuple:
for timesteps in [1, 5, 100, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :Tuple) -> str:
for beta_start, beta_end in zip([0.0_001, 0.001, 0.01, 0.1] , [0.002, 0.02, 0.2, 2]):
self.check_over_configs(beta_start=_lowercase , beta_end=_lowercase)
def __a ( self :Dict) -> Any:
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=_lowercase)
def __a ( self :Any) -> Union[str, Any]:
for variance in ["fixed_small", "fixed_large", "other"]:
self.check_over_configs(variance_type=_lowercase)
def __a ( self :Tuple) -> str:
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=_lowercase)
def __a ( self :Union[str, Any]) -> Optional[int]:
self.check_over_configs(thresholding=_lowercase)
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample", "v_prediction"]:
self.check_over_configs(
thresholding=_lowercase , prediction_type=_lowercase , sample_max_value=_lowercase , )
def __a ( self :Dict) -> int:
for prediction_type in ["epsilon", "sample", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Tuple) -> List[Any]:
for t in [0, 500, 999]:
self.check_over_forward(time_step=_lowercase)
def __a ( self :Union[str, Any]) -> Optional[int]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
assert torch.sum(torch.abs(scheduler._get_variance(0) - 0.0)) < 1E-5
assert torch.sum(torch.abs(scheduler._get_variance(487) - 0.00_979)) < 1E-5
assert torch.sum(torch.abs(scheduler._get_variance(999) - 0.02)) < 1E-5
def __a ( self :Optional[Any]) -> Dict:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
UpperCAmelCase_ = self.dummy_sample_deter + 0.1
UpperCAmelCase_ = self.dummy_sample_deter - 0.1
UpperCAmelCase_ = samplea.shape[0]
UpperCAmelCase_ = torch.stack([samplea, samplea, samplea] , dim=0)
UpperCAmelCase_ = torch.arange(_lowercase)[0:3, None].repeat(1 , _lowercase)
UpperCAmelCase_ = model(samples.flatten(0 , 1) , timesteps.flatten(0 , 1))
UpperCAmelCase_ = scheduler.batch_step_no_noise(_lowercase , timesteps.flatten(0 , 1) , samples.flatten(0 , 1))
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 1_153.1_833) < 1E-2
assert abs(result_mean.item() - 0.5_005) < 1E-3
def __a ( self :Optional[int]) -> Dict:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
UpperCAmelCase_ = torch.manual_seed(0)
for t in reversed(range(_lowercase)):
# 1. predict noise residual
UpperCAmelCase_ = model(_lowercase , _lowercase)
# 2. predict previous mean of sample x_t-1
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , generator=_lowercase).prev_sample
UpperCAmelCase_ = pred_prev_sample
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 258.9_606) < 1E-2
assert abs(result_mean.item() - 0.3_372) < 1E-3
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(prediction_type='''v_prediction''')
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
UpperCAmelCase_ = torch.manual_seed(0)
for t in reversed(range(_lowercase)):
# 1. predict noise residual
UpperCAmelCase_ = model(_lowercase , _lowercase)
# 2. predict previous mean of sample x_t-1
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , generator=_lowercase).prev_sample
UpperCAmelCase_ = pred_prev_sample
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 202.0_296) < 1E-2
assert abs(result_mean.item() - 0.2_631) < 1E-3
def __a ( self :Optional[int]) -> List[str]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = [100, 87, 50, 1, 0]
scheduler.set_timesteps(timesteps=_lowercase)
UpperCAmelCase_ = scheduler.timesteps
for i, timestep in enumerate(_lowercase):
if i == len(_lowercase) - 1:
UpperCAmelCase_ = -1
else:
UpperCAmelCase_ = timesteps[i + 1]
UpperCAmelCase_ = scheduler.previous_timestep(_lowercase)
UpperCAmelCase_ = prev_t.item()
self.assertEqual(_lowercase , _lowercase)
def __a ( self :Dict) -> Dict:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = [100, 87, 50, 51, 0]
with self.assertRaises(_lowercase , msg='''`custom_timesteps` must be in descending order.'''):
scheduler.set_timesteps(timesteps=_lowercase)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = [100, 87, 50, 1, 0]
UpperCAmelCase_ = len(_lowercase)
with self.assertRaises(_lowercase , msg='''Can only pass one of `num_inference_steps` or `custom_timesteps`.'''):
scheduler.set_timesteps(num_inference_steps=_lowercase , timesteps=_lowercase)
def __a ( self :Union[str, Any]) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = [scheduler.config.num_train_timesteps]
with self.assertRaises(
_lowercase , msg='''`timesteps` must start before `self.config.train_timesteps`: {scheduler.config.num_train_timesteps}}''' , ):
scheduler.set_timesteps(timesteps=_lowercase)
| 344 |
import functools
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = len(__UpperCAmelCase )
UpperCAmelCase_ = len(__UpperCAmelCase )
@functools.cache
def min_distance(__UpperCAmelCase , __UpperCAmelCase ) -> int:
# if first word index is overflow - delete all from the second word
if indexa >= len_worda:
return len_worda - indexa
# if second word index is overflow - delete all from the first word
if indexa >= len_worda:
return len_worda - indexa
UpperCAmelCase_ = int(worda[indexa] != worda[indexa] ) # current letters not identical
return min(
1 + min_distance(indexa + 1 , __UpperCAmelCase ) , 1 + min_distance(__UpperCAmelCase , indexa + 1 ) , diff + min_distance(indexa + 1 , indexa + 1 ) , )
return min_distance(0 , 0 )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 | 1 |
from queue import PriorityQueue
from typing import Any
import numpy as np
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , ) -> float | int:
'''simple docstring'''
for nxt, d in graph[v]:
if nxt in visited_forward:
continue
UpperCAmelCase_ = cst_fwd.get(__UpperCAmelCase , np.inf )
UpperCAmelCase_ = cst_fwd[v] + d
if new_cost_f < old_cost_f:
queue.put((new_cost_f, nxt) )
UpperCAmelCase_ = new_cost_f
UpperCAmelCase_ = v
if nxt in visited_backward:
if cst_fwd[v] + d + cst_bwd[nxt] < shortest_distance:
UpperCAmelCase_ = cst_fwd[v] + d + cst_bwd[nxt]
return shortest_distance
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = -1
UpperCAmelCase_ = set()
UpperCAmelCase_ = set()
UpperCAmelCase_ = {source: 0}
UpperCAmelCase_ = {destination: 0}
UpperCAmelCase_ = {source: None}
UpperCAmelCase_ = {destination: None}
UpperCAmelCase_ = PriorityQueue()
UpperCAmelCase_ = PriorityQueue()
UpperCAmelCase_ = np.inf
queue_forward.put((0, source) )
queue_backward.put((0, destination) )
if source == destination:
return 0
while not queue_forward.empty() and not queue_backward.empty():
UpperCAmelCase_ , UpperCAmelCase_ = queue_forward.get()
visited_forward.add(__UpperCAmelCase )
UpperCAmelCase_ , UpperCAmelCase_ = queue_backward.get()
visited_backward.add(__UpperCAmelCase )
UpperCAmelCase_ = pass_and_relaxation(
__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , )
UpperCAmelCase_ = pass_and_relaxation(
__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , )
if cst_fwd[v_fwd] + cst_bwd[v_bwd] >= shortest_distance:
break
if shortest_distance != np.inf:
UpperCAmelCase_ = shortest_distance
return shortest_path_distance
UpperCamelCase_ = {
"B": [["C", 1]],
"C": [["D", 1]],
"D": [["F", 1]],
"E": [["B", 1], ["G", 2]],
"F": [],
"G": [["F", 1]],
}
UpperCamelCase_ = {
"B": [["E", 1]],
"C": [["B", 1]],
"D": [["C", 1]],
"F": [["D", 1], ["G", 1]],
"E": [[None, np.inf]],
"G": [["E", 2]],
}
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import SPIECE_UNDERLINE, logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {"vocab_file": "spiece.model"}
UpperCamelCase_ = {
"vocab_file": {
"xlnet-base-cased": "https://huggingface.co/xlnet-base-cased/resolve/main/spiece.model",
"xlnet-large-cased": "https://huggingface.co/xlnet-large-cased/resolve/main/spiece.model",
}
}
UpperCamelCase_ = {
"xlnet-base-cased": None,
"xlnet-large-cased": None,
}
# Segments (not really needed)
UpperCamelCase_ = 0
UpperCamelCase_ = 1
UpperCamelCase_ = 2
UpperCamelCase_ = 3
UpperCamelCase_ = 4
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =VOCAB_FILES_NAMES
UpperCamelCase__ : Optional[Any] =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : Tuple =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Any ="left"
def __init__( self :Optional[int] , _lowercase :Union[str, Any] , _lowercase :Union[str, Any]=False , _lowercase :Optional[int]=True , _lowercase :Union[str, Any]=False , _lowercase :Tuple="<s>" , _lowercase :Any="</s>" , _lowercase :Dict="<unk>" , _lowercase :str="<sep>" , _lowercase :Tuple="<pad>" , _lowercase :Any="<cls>" , _lowercase :List[str]="<mask>" , _lowercase :Union[str, Any]=["<eop>", "<eod>"] , _lowercase :Optional[Dict[str, Any]] = None , **_lowercase :Union[str, Any] , ) -> None:
# Mask token behave like a normal word, i.e. include the space before it
UpperCAmelCase_ = AddedToken(_lowercase , lstrip=_lowercase , rstrip=_lowercase) if isinstance(_lowercase , _lowercase) else mask_token
UpperCAmelCase_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_lowercase , remove_space=_lowercase , keep_accents=_lowercase , bos_token=_lowercase , eos_token=_lowercase , unk_token=_lowercase , sep_token=_lowercase , pad_token=_lowercase , cls_token=_lowercase , mask_token=_lowercase , additional_special_tokens=_lowercase , sp_model_kwargs=self.sp_model_kwargs , **_lowercase , )
UpperCAmelCase_ = 3
UpperCAmelCase_ = do_lower_case
UpperCAmelCase_ = remove_space
UpperCAmelCase_ = keep_accents
UpperCAmelCase_ = vocab_file
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(_lowercase)
@property
def __a ( self :int) -> List[Any]:
return len(self.sp_model)
def __a ( self :Optional[int]) -> List[Any]:
UpperCAmelCase_ = {self.convert_ids_to_tokens(_lowercase): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def __getstate__( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.__dict__.copy()
UpperCAmelCase_ = None
return state
def __setstate__( self :Optional[Any] , _lowercase :Optional[Any]) -> List[Any]:
UpperCAmelCase_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs'''):
UpperCAmelCase_ = {}
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(self.vocab_file)
def __a ( self :List[str] , _lowercase :Tuple) -> Optional[int]:
if self.remove_space:
UpperCAmelCase_ = ''' '''.join(inputs.strip().split())
else:
UpperCAmelCase_ = inputs
UpperCAmelCase_ = outputs.replace('''``''' , '''"''').replace('''\'\'''' , '''"''')
if not self.keep_accents:
UpperCAmelCase_ = unicodedata.normalize('''NFKD''' , _lowercase)
UpperCAmelCase_ = ''''''.join([c for c in outputs if not unicodedata.combining(_lowercase)])
if self.do_lower_case:
UpperCAmelCase_ = outputs.lower()
return outputs
def __a ( self :str , _lowercase :str) -> List[str]:
UpperCAmelCase_ = self.preprocess_text(_lowercase)
UpperCAmelCase_ = self.sp_model.encode(_lowercase , out_type=_lowercase)
UpperCAmelCase_ = []
for piece in pieces:
if len(_lowercase) > 1 and piece[-1] == str(''',''') and piece[-2].isdigit():
UpperCAmelCase_ = self.sp_model.EncodeAsPieces(piece[:-1].replace(_lowercase , ''''''))
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0]) == 1:
UpperCAmelCase_ = cur_pieces[1:]
else:
UpperCAmelCase_ = cur_pieces[0][1:]
cur_pieces.append(piece[-1])
new_pieces.extend(_lowercase)
else:
new_pieces.append(_lowercase)
return new_pieces
def __a ( self :Optional[Any] , _lowercase :Union[str, Any]) -> Tuple:
return self.sp_model.PieceToId(_lowercase)
def __a ( self :Optional[int] , _lowercase :Optional[Any]) -> List[str]:
return self.sp_model.IdToPiece(_lowercase)
def __a ( self :List[Any] , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = ''''''.join(_lowercase).replace(_lowercase , ''' ''').strip()
return out_string
def __a ( self :Union[str, Any] , _lowercase :List[int] , _lowercase :bool = False , _lowercase :bool = None , _lowercase :bool = True , **_lowercase :Tuple , ) -> str:
UpperCAmelCase_ = kwargs.pop('''use_source_tokenizer''' , _lowercase)
UpperCAmelCase_ = self.convert_ids_to_tokens(_lowercase , skip_special_tokens=_lowercase)
# To avoid mixing byte-level and unicode for byte-level BPT
# we need to build string separately for added tokens and byte-level tokens
# cf. https://github.com/huggingface/transformers/issues/1133
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for token in filtered_tokens:
if skip_special_tokens and token in self.all_special_ids:
continue
if token in self.added_tokens_encoder:
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
UpperCAmelCase_ = []
sub_texts.append(_lowercase)
else:
current_sub_text.append(_lowercase)
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
# Mimic the behavior of the Rust tokenizer:
# By default, there are no spaces between special tokens
UpperCAmelCase_ = ''''''.join(_lowercase)
UpperCAmelCase_ = (
clean_up_tokenization_spaces
if clean_up_tokenization_spaces is not None
else self.clean_up_tokenization_spaces
)
if clean_up_tokenization_spaces:
UpperCAmelCase_ = self.clean_up_tokenization(_lowercase)
return clean_text
else:
return text
def __a ( self :str , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [self.cls_token_id]
if token_ids_a is None:
return token_ids_a + sep + cls
return token_ids_a + sep + token_ids_a + sep + cls
def __a ( self :Dict , _lowercase :List[int] , _lowercase :Optional[List[int]] = None , _lowercase :bool = False) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowercase , token_ids_a=_lowercase , already_has_special_tokens=_lowercase)
if token_ids_a is not None:
return ([0] * len(_lowercase)) + [1] + ([0] * len(_lowercase)) + [1, 1]
return ([0] * len(_lowercase)) + [1, 1]
def __a ( self :Optional[int] , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [2]
if token_ids_a is None:
return len(token_ids_a + sep) * [0] + cls_segment_id
return len(token_ids_a + sep) * [0] + len(token_ids_a + sep) * [1] + cls_segment_id
def __a ( self :str , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
if not os.path.isdir(_lowercase):
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
return
UpperCAmelCase_ = os.path.join(
_lowercase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''])
if os.path.abspath(self.vocab_file) != os.path.abspath(_lowercase) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , _lowercase)
elif not os.path.isfile(self.vocab_file):
with open(_lowercase , '''wb''') as fi:
UpperCAmelCase_ = self.sp_model.serialized_model_proto()
fi.write(_lowercase)
return (out_vocab_file,)
| 344 | 1 |
import pytest
UpperCamelCase_ = "__dummy_dataset1__"
UpperCamelCase_ = "\nimport json\nimport os\n\nimport datasets\n\n\nREPO_URL = \"https://huggingface.co/datasets/albertvillanova/tests-raw-jsonl/resolve/main/\"\nURLS = {\"train\": REPO_URL + \"wikiann-bn-train.jsonl\", \"validation\": REPO_URL + \"wikiann-bn-validation.jsonl\"}\n\n\nclass __DummyDataset1__(datasets.GeneratorBasedBuilder):\n\n def _info(self):\n features = datasets.Features(\n {\n \"tokens\": datasets.Sequence(datasets.Value(\"string\")),\n \"ner_tags\": datasets.Sequence(\n datasets.features.ClassLabel(\n names=[\n \"O\",\n \"B-PER\",\n \"I-PER\",\n \"B-ORG\",\n \"I-ORG\",\n \"B-LOC\",\n \"I-LOC\",\n ]\n )\n ),\n \"langs\": datasets.Sequence(datasets.Value(\"string\")),\n \"spans\": datasets.Sequence(datasets.Value(\"string\")),\n }\n )\n return datasets.DatasetInfo(features=features)\n\n def _split_generators(self, dl_manager):\n dl_path = dl_manager.download(URLS)\n return [\n datasets.SplitGenerator(datasets.Split.TRAIN, gen_kwargs={\"filepath\": dl_path[\"train\"]}),\n datasets.SplitGenerator(datasets.Split.VALIDATION, gen_kwargs={\"filepath\": dl_path[\"validation\"]}),\n ]\n\n def _generate_examples(self, filepath):\n with open(filepath, \"r\", encoding=\"utf-8\") as f:\n for i, line in enumerate(f):\n yield i, json.loads(line)\n"
@pytest.fixture
def A ( ) -> str:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_NAME
@pytest.fixture
def A ( ) -> Any:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_CODE
@pytest.fixture
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = dataset_loading_script_name
UpperCAmelCase_ = tmp_path / '''datasets''' / script_name
script_dir.mkdir(parents=__UpperCAmelCase )
UpperCAmelCase_ = script_dir / f"{script_name}.py"
with open(__UpperCAmelCase , '''w''' ) as f:
f.write(__UpperCAmelCase )
return str(__UpperCAmelCase )
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
UpperCamelCase_ = logging.get_logger(__name__)
class a_ ( _snake_case , _snake_case ):
UpperCamelCase__ : Union[str, Any] ="maskformer-swin"
UpperCamelCase__ : List[str] ={
"num_attention_heads": "num_heads",
"num_hidden_layers": "num_layers",
}
def __init__( self :Union[str, Any] , _lowercase :Optional[int]=224 , _lowercase :List[str]=4 , _lowercase :Tuple=3 , _lowercase :List[Any]=96 , _lowercase :Any=[2, 2, 6, 2] , _lowercase :int=[3, 6, 12, 24] , _lowercase :List[Any]=7 , _lowercase :Dict=4.0 , _lowercase :Any=True , _lowercase :int=0.0 , _lowercase :List[Any]=0.0 , _lowercase :Tuple=0.1 , _lowercase :str="gelu" , _lowercase :Union[str, Any]=False , _lowercase :Tuple=0.02 , _lowercase :List[str]=1E-5 , _lowercase :List[str]=None , _lowercase :Any=None , **_lowercase :Any , ) -> Union[str, Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = embed_dim
UpperCAmelCase_ = depths
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = num_heads
UpperCAmelCase_ = window_size
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = qkv_bias
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = use_absolute_embeddings
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = initializer_range
# we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
UpperCAmelCase_ = int(embed_dim * 2 ** (len(_lowercase) - 1))
UpperCAmelCase_ = ['''stem'''] + [f"stage{idx}" for idx in range(1 , len(_lowercase) + 1)]
UpperCAmelCase_ , UpperCAmelCase_ = get_aligned_output_features_output_indices(
out_features=_lowercase , out_indices=_lowercase , stage_names=self.stage_names)
| 344 | 1 |
import unittest
from typing import Tuple
import torch
from diffusers.utils import floats_tensor, randn_tensor, torch_all_close, torch_device
from diffusers.utils.testing_utils import require_torch
@require_torch
class a_ :
@property
def __a ( self :Union[str, Any]) -> Any:
return self.get_dummy_input()
@property
def __a ( self :int) -> Any:
if self.block_type == "down":
return (4, 32, 16, 16)
elif self.block_type == "mid":
return (4, 32, 32, 32)
elif self.block_type == "up":
return (4, 32, 64, 64)
raise ValueError(f"'{self.block_type}' is not a supported block_type. Set it to 'up', 'mid', or 'down'.")
def __a ( self :List[Any] , _lowercase :List[str]=True , _lowercase :str=False , _lowercase :Any=False , _lowercase :str=False , ) -> Optional[int]:
UpperCAmelCase_ = 4
UpperCAmelCase_ = 32
UpperCAmelCase_ = (32, 32)
UpperCAmelCase_ = torch.manual_seed(0)
UpperCAmelCase_ = torch.device(_lowercase)
UpperCAmelCase_ = (batch_size, num_channels) + sizes
UpperCAmelCase_ = randn_tensor(_lowercase , generator=_lowercase , device=_lowercase)
UpperCAmelCase_ = {'''hidden_states''': hidden_states}
if include_temb:
UpperCAmelCase_ = 128
UpperCAmelCase_ = randn_tensor((batch_size, temb_channels) , generator=_lowercase , device=_lowercase)
if include_res_hidden_states_tuple:
UpperCAmelCase_ = torch.manual_seed(1)
UpperCAmelCase_ = (randn_tensor(_lowercase , generator=_lowercase , device=_lowercase),)
if include_encoder_hidden_states:
UpperCAmelCase_ = floats_tensor((batch_size, 32, 32)).to(_lowercase)
if include_skip_sample:
UpperCAmelCase_ = randn_tensor(((batch_size, 3) + sizes) , generator=_lowercase , device=_lowercase)
return dummy_input
def __a ( self :List[Any]) -> Any:
UpperCAmelCase_ = {
'''in_channels''': 32,
'''out_channels''': 32,
'''temb_channels''': 128,
}
if self.block_type == "up":
UpperCAmelCase_ = 32
if self.block_type == "mid":
init_dict.pop('''out_channels''')
UpperCAmelCase_ = self.dummy_input
return init_dict, inputs_dict
def __a ( self :Optional[Any] , _lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ , UpperCAmelCase_ = self.prepare_init_args_and_inputs_for_common()
UpperCAmelCase_ = self.block_class(**_lowercase)
unet_block.to(_lowercase)
unet_block.eval()
with torch.no_grad():
UpperCAmelCase_ = unet_block(**_lowercase)
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = output[0]
self.assertEqual(output.shape , self.output_shape)
UpperCAmelCase_ = output[0, -1, -3:, -3:]
UpperCAmelCase_ = torch.tensor(_lowercase).to(_lowercase)
assert torch_all_close(output_slice.flatten() , _lowercase , atol=5E-3)
@unittest.skipIf(torch_device == '''mps''' , '''Training is not supported in mps''')
def __a ( self :List[Any]) -> Any:
UpperCAmelCase_ , UpperCAmelCase_ = self.prepare_init_args_and_inputs_for_common()
UpperCAmelCase_ = self.block_class(**_lowercase)
model.to(_lowercase)
model.train()
UpperCAmelCase_ = model(**_lowercase)
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = output[0]
UpperCAmelCase_ = torch.device(_lowercase)
UpperCAmelCase_ = randn_tensor(output.shape , device=_lowercase)
UpperCAmelCase_ = torch.nn.functional.mse_loss(_lowercase , _lowercase)
loss.backward()
| 344 |
import os
from argparse import ArgumentParser, Namespace
from ..data import SingleSentenceClassificationProcessor as Processor
from ..pipelines import TextClassificationPipeline
from ..utils import is_tf_available, is_torch_available, logging
from . import BaseTransformersCLICommand
if not is_tf_available() and not is_torch_available():
raise RuntimeError("At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training")
# TF training parameters
UpperCamelCase_ = False
UpperCamelCase_ = False
def A ( __UpperCAmelCase ) -> Any:
'''simple docstring'''
return TrainCommand(__UpperCAmelCase )
class a_ ( _snake_case ):
@staticmethod
def __a ( _lowercase :ArgumentParser) -> List[Any]:
UpperCAmelCase_ = parser.add_parser('''train''' , help='''CLI tool to train a model on a task.''')
train_parser.add_argument(
'''--train_data''' , type=_lowercase , required=_lowercase , help='''path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.''' , )
train_parser.add_argument(
'''--column_label''' , type=_lowercase , default=0 , help='''Column of the dataset csv file with example labels.''')
train_parser.add_argument(
'''--column_text''' , type=_lowercase , default=1 , help='''Column of the dataset csv file with example texts.''')
train_parser.add_argument(
'''--column_id''' , type=_lowercase , default=2 , help='''Column of the dataset csv file with example ids.''')
train_parser.add_argument(
'''--skip_first_row''' , action='''store_true''' , help='''Skip the first row of the csv file (headers).''')
train_parser.add_argument('''--validation_data''' , type=_lowercase , default='''''' , help='''path to validation dataset.''')
train_parser.add_argument(
'''--validation_split''' , type=_lowercase , default=0.1 , help='''if validation dataset is not provided, fraction of train dataset to use as validation dataset.''' , )
train_parser.add_argument('''--output''' , type=_lowercase , default='''./''' , help='''path to saved the trained model.''')
train_parser.add_argument(
'''--task''' , type=_lowercase , default='''text_classification''' , help='''Task to train the model on.''')
train_parser.add_argument(
'''--model''' , type=_lowercase , default='''bert-base-uncased''' , help='''Model\'s name or path to stored model.''')
train_parser.add_argument('''--train_batch_size''' , type=_lowercase , default=32 , help='''Batch size for training.''')
train_parser.add_argument('''--valid_batch_size''' , type=_lowercase , default=64 , help='''Batch size for validation.''')
train_parser.add_argument('''--learning_rate''' , type=_lowercase , default=3E-5 , help='''Learning rate.''')
train_parser.add_argument('''--adam_epsilon''' , type=_lowercase , default=1E-0_8 , help='''Epsilon for Adam optimizer.''')
train_parser.set_defaults(func=_lowercase)
def __init__( self :Union[str, Any] , _lowercase :Namespace) -> Union[str, Any]:
UpperCAmelCase_ = logging.get_logger('''transformers-cli/training''')
UpperCAmelCase_ = '''tf''' if is_tf_available() else '''torch'''
os.makedirs(args.output , exist_ok=_lowercase)
UpperCAmelCase_ = args.output
UpperCAmelCase_ = args.column_label
UpperCAmelCase_ = args.column_text
UpperCAmelCase_ = args.column_id
self.logger.info(f"Loading {args.task} pipeline for {args.model}")
if args.task == "text_classification":
UpperCAmelCase_ = TextClassificationPipeline.from_pretrained(args.model)
elif args.task == "token_classification":
raise NotImplementedError
elif args.task == "question_answering":
raise NotImplementedError
self.logger.info(f"Loading dataset from {args.train_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.train_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = None
if args.validation_data:
self.logger.info(f"Loading validation dataset from {args.validation_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.validation_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = args.validation_split
UpperCAmelCase_ = args.train_batch_size
UpperCAmelCase_ = args.valid_batch_size
UpperCAmelCase_ = args.learning_rate
UpperCAmelCase_ = args.adam_epsilon
def __a ( self :int) -> Tuple:
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def __a ( self :Optional[Any]) -> Any:
raise NotImplementedError
def __a ( self :int) -> Optional[Any]:
self.pipeline.fit(
self.train_dataset , validation_data=self.valid_dataset , validation_split=self.validation_split , learning_rate=self.learning_rate , adam_epsilon=self.adam_epsilon , train_batch_size=self.train_batch_size , valid_batch_size=self.valid_batch_size , )
# Save trained pipeline
self.pipeline.save_pretrained(self.output)
| 344 | 1 |
import copy
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ..auto import CONFIG_MAPPING
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"SenseTime/deformable-detr": "https://huggingface.co/sensetime/deformable-detr/resolve/main/config.json",
# See all Deformable DETR models at https://huggingface.co/models?filter=deformable-detr
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] ="deformable_detr"
UpperCamelCase__ : Any ={
"hidden_size": "d_model",
"num_attention_heads": "encoder_attention_heads",
}
def __init__( self :Optional[Any] , _lowercase :Optional[Any]=True , _lowercase :Any=None , _lowercase :Optional[Any]=3 , _lowercase :int=300 , _lowercase :List[Any]=1024 , _lowercase :Any=6 , _lowercase :Dict=1024 , _lowercase :Union[str, Any]=8 , _lowercase :int=6 , _lowercase :str=1024 , _lowercase :List[str]=8 , _lowercase :Tuple=0.0 , _lowercase :List[str]=True , _lowercase :Optional[int]="relu" , _lowercase :Optional[int]=256 , _lowercase :str=0.1 , _lowercase :Union[str, Any]=0.0 , _lowercase :List[Any]=0.0 , _lowercase :Any=0.02 , _lowercase :Tuple=1.0 , _lowercase :Union[str, Any]=True , _lowercase :int=False , _lowercase :List[Any]="sine" , _lowercase :Optional[int]="resnet50" , _lowercase :Optional[int]=True , _lowercase :Any=False , _lowercase :Optional[Any]=4 , _lowercase :Optional[Any]=4 , _lowercase :Dict=4 , _lowercase :List[str]=False , _lowercase :Union[str, Any]=300 , _lowercase :List[Any]=False , _lowercase :int=1 , _lowercase :str=5 , _lowercase :List[Any]=2 , _lowercase :str=1 , _lowercase :str=1 , _lowercase :Optional[int]=5 , _lowercase :Optional[int]=2 , _lowercase :Union[str, Any]=0.1 , _lowercase :List[Any]=0.25 , _lowercase :Dict=False , **_lowercase :List[Any] , ) -> Optional[int]:
if backbone_config is not None and use_timm_backbone:
raise ValueError('''You can\'t specify both `backbone_config` and `use_timm_backbone`.''')
if not use_timm_backbone:
if backbone_config is None:
logger.info('''`backbone_config` is `None`. Initializing the config with the default `ResNet` backbone.''')
UpperCAmelCase_ = CONFIG_MAPPING['''resnet'''](out_features=['''stage4'''])
elif isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = backbone_config.get('''model_type''')
UpperCAmelCase_ = CONFIG_MAPPING[backbone_model_type]
UpperCAmelCase_ = config_class.from_dict(_lowercase)
UpperCAmelCase_ = use_timm_backbone
UpperCAmelCase_ = backbone_config
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = num_queries
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = d_model
UpperCAmelCase_ = encoder_ffn_dim
UpperCAmelCase_ = encoder_layers
UpperCAmelCase_ = encoder_attention_heads
UpperCAmelCase_ = decoder_ffn_dim
UpperCAmelCase_ = decoder_layers
UpperCAmelCase_ = decoder_attention_heads
UpperCAmelCase_ = dropout
UpperCAmelCase_ = attention_dropout
UpperCAmelCase_ = activation_dropout
UpperCAmelCase_ = activation_function
UpperCAmelCase_ = init_std
UpperCAmelCase_ = init_xavier_std
UpperCAmelCase_ = encoder_layerdrop
UpperCAmelCase_ = auxiliary_loss
UpperCAmelCase_ = position_embedding_type
UpperCAmelCase_ = backbone
UpperCAmelCase_ = use_pretrained_backbone
UpperCAmelCase_ = dilation
# deformable attributes
UpperCAmelCase_ = num_feature_levels
UpperCAmelCase_ = encoder_n_points
UpperCAmelCase_ = decoder_n_points
UpperCAmelCase_ = two_stage
UpperCAmelCase_ = two_stage_num_proposals
UpperCAmelCase_ = with_box_refine
if two_stage is True and with_box_refine is False:
raise ValueError('''If two_stage is True, with_box_refine must be True.''')
# Hungarian matcher
UpperCAmelCase_ = class_cost
UpperCAmelCase_ = bbox_cost
UpperCAmelCase_ = giou_cost
# Loss coefficients
UpperCAmelCase_ = mask_loss_coefficient
UpperCAmelCase_ = dice_loss_coefficient
UpperCAmelCase_ = bbox_loss_coefficient
UpperCAmelCase_ = giou_loss_coefficient
UpperCAmelCase_ = eos_coefficient
UpperCAmelCase_ = focal_alpha
UpperCAmelCase_ = disable_custom_kernels
super().__init__(is_encoder_decoder=_lowercase , **_lowercase)
@property
def __a ( self :Optional[Any]) -> int:
return self.encoder_attention_heads
@property
def __a ( self :int) -> int:
return self.d_model
def __a ( self :int) -> Optional[int]:
UpperCAmelCase_ = copy.deepcopy(self.__dict__)
if self.backbone_config is not None:
UpperCAmelCase_ = self.backbone_config.to_dict()
UpperCAmelCase_ = self.__class__.model_type
return output
| 344 |
import unittest
from typing import Dict, List, Optional, Union
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import BridgeTowerImageProcessor
class a_ ( unittest.TestCase ):
def __init__( self :Tuple , _lowercase :List[Any] , _lowercase :bool = True , _lowercase :Dict[str, int] = None , _lowercase :int = 32 , _lowercase :bool = True , _lowercase :Union[int, float] = 1 / 255 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :Optional[Union[float, List[float]]] = [0.48_145_466, 0.4_578_275, 0.40_821_073] , _lowercase :Optional[Union[float, List[float]]] = [0.26_862_954, 0.26_130_258, 0.27_577_711] , _lowercase :bool = True , _lowercase :List[Any]=7 , _lowercase :Dict=30 , _lowercase :Optional[int]=400 , _lowercase :Any=3 , ) -> Any:
UpperCAmelCase_ = parent
UpperCAmelCase_ = do_resize
UpperCAmelCase_ = size if size is not None else {'''shortest_edge''': 288}
UpperCAmelCase_ = size_divisor
UpperCAmelCase_ = do_rescale
UpperCAmelCase_ = rescale_factor
UpperCAmelCase_ = do_normalize
UpperCAmelCase_ = do_center_crop
UpperCAmelCase_ = image_mean
UpperCAmelCase_ = image_std
UpperCAmelCase_ = do_pad
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = min_resolution
UpperCAmelCase_ = max_resolution
def __a ( self :str) -> Tuple:
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"size_divisor": self.size_divisor,
}
def __a ( self :List[Any] , _lowercase :Tuple , _lowercase :List[str]=False) -> int:
if not batched:
UpperCAmelCase_ = self.size['''shortest_edge''']
UpperCAmelCase_ = image_inputs[0]
if isinstance(_lowercase , Image.Image):
UpperCAmelCase_ , UpperCAmelCase_ = image.size
else:
UpperCAmelCase_ , UpperCAmelCase_ = image.shape[1], image.shape[2]
UpperCAmelCase_ = size / min(_lowercase , _lowercase)
if h < w:
UpperCAmelCase_ , UpperCAmelCase_ = size, scale * w
else:
UpperCAmelCase_ , UpperCAmelCase_ = scale * h, size
UpperCAmelCase_ = int((1333 / 800) * size)
if max(_lowercase , _lowercase) > max_size:
UpperCAmelCase_ = max_size / max(_lowercase , _lowercase)
UpperCAmelCase_ = newh * scale
UpperCAmelCase_ = neww * scale
UpperCAmelCase_ , UpperCAmelCase_ = int(newh + 0.5), int(neww + 0.5)
UpperCAmelCase_ , UpperCAmelCase_ = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
UpperCAmelCase_ = []
for image in image_inputs:
UpperCAmelCase_ , UpperCAmelCase_ = self.get_expected_values([image])
expected_values.append((expected_height, expected_width))
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[0])[0]
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[1])[1]
return expected_height, expected_width
@require_torch
@require_vision
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =BridgeTowerImageProcessor if is_vision_available() else None
def __a ( self :int) -> Dict:
UpperCAmelCase_ = BridgeTowerImageProcessingTester(self)
@property
def __a ( self :Dict) -> Any:
return self.image_processor_tester.prepare_image_processor_dict()
def __a ( self :Dict) -> Tuple:
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(_lowercase , '''image_mean'''))
self.assertTrue(hasattr(_lowercase , '''image_std'''))
self.assertTrue(hasattr(_lowercase , '''do_normalize'''))
self.assertTrue(hasattr(_lowercase , '''do_resize'''))
self.assertTrue(hasattr(_lowercase , '''size'''))
self.assertTrue(hasattr(_lowercase , '''size_divisor'''))
def __a ( self :Union[str, Any]) -> Tuple:
pass
def __a ( self :List[str]) -> Tuple:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , Image.Image)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :Union[str, Any]) -> Optional[int]:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , numpify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , np.ndarray)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :str) -> int:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , torchify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , torch.Tensor)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
| 344 | 1 |
import faiss # noqa: F401 # Here to have a nice missing dependency error message early on
import numpy # noqa: F401 # Here to have a nice missing dependency error message early on
import requests # noqa: F401 # Here to have a nice missing dependency error message early on
import sklearn # noqa: F401 # Here to have a nice missing dependency error message early on
import tqdm # noqa: F401 # Here to have a nice missing dependency error message early on
from mauve import compute_mauve # From: mauve-text
import datasets
UpperCamelCase_ = "\\n@inproceedings{pillutla-etal:mauve:neurips2021,\n title={MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontiers},\n author={Pillutla, Krishna and Swayamdipta, Swabha and Zellers, Rowan and Thickstun, John and Welleck, Sean and Choi, Yejin and Harchaoui, Zaid},\n booktitle = {NeurIPS},\n year = {2021}\n}\n\n"
UpperCamelCase_ = "\\nMAUVE is a library built on PyTorch and HuggingFace Transformers to measure the gap between neural text and human text with the eponymous MAUVE measure.\n\nMAUVE summarizes both Type I and Type II errors measured softly using Kullback–Leibler (KL) divergences.\n\nFor details, see the MAUVE paper: https://arxiv.org/abs/2102.01454 (Neurips, 2021).\n\nThis metrics is a wrapper around the official implementation of MAUVE:\nhttps://github.com/krishnap25/mauve\n"
UpperCamelCase_ = "\nCalculates MAUVE scores between two lists of generated text and reference text.\nArgs:\n predictions: list of generated text to score. Each predictions\n should be a string with tokens separated by spaces.\n references: list of reference for each prediction. Each\n reference should be a string with tokens separated by spaces.\nOptional Args:\n num_buckets: the size of the histogram to quantize P and Q. Options: 'auto' (default) or an integer\n pca_max_data: the number data points to use for PCA dimensionality reduction prior to clustering. If -1, use all the data. Default -1\n kmeans_explained_var: amount of variance of the data to keep in dimensionality reduction by PCA. Default 0.9\n kmeans_num_redo: number of times to redo k-means clustering (the best objective is kept). Default 5\n kmeans_max_iter: maximum number of k-means iterations. Default 500\n featurize_model_name: name of the model from which features are obtained. Default 'gpt2-large' Use one of ['gpt2', 'gpt2-medium', 'gpt2-large', 'gpt2-xl'].\n device_id: Device for featurization. Supply a GPU id (e.g. 0 or 3) to use GPU. If no GPU with this id is found, use CPU\n max_text_length: maximum number of tokens to consider. Default 1024\n divergence_curve_discretization_size: Number of points to consider on the divergence curve. Default 25\n mauve_scaling_factor: \"c\" from the paper. Default 5.\n verbose: If True (default), print running time updates\n seed: random seed to initialize k-means cluster assignments.\nReturns:\n mauve: MAUVE score, a number between 0 and 1. Larger values indicate that P and Q are closer,\n frontier_integral: Frontier Integral, a number between 0 and 1. Smaller values indicate that P and Q are closer,\n divergence_curve: a numpy.ndarray of shape (m, 2); plot it with matplotlib to view the divergence curve,\n p_hist: a discrete distribution, which is a quantized version of the text distribution p_text,\n q_hist: same as above, but with q_text.\nExamples:\n\n >>> # faiss segfaults in doctest for some reason, so the .compute call is not tested with doctest\n >>> import datasets\n >>> mauve = datasets.load_metric('mauve')\n >>> predictions = [\"hello there\", \"general kenobi\"]\n >>> references = [\"hello there\", \"general kenobi\"]\n >>> out = mauve.compute(predictions=predictions, references=references) # doctest: +SKIP\n >>> print(out.mauve) # doctest: +SKIP\n 1.0\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class a_ ( datasets.Metric ):
def __a ( self :List[str]) -> List[Any]:
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , homepage='''https://github.com/krishnap25/mauve''' , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
'''predictions''': datasets.Value('''string''' , id='''sequence'''),
'''references''': datasets.Value('''string''' , id='''sequence'''),
}) , codebase_urls=['''https://github.com/krishnap25/mauve'''] , reference_urls=[
'''https://arxiv.org/abs/2102.01454''',
'''https://github.com/krishnap25/mauve''',
] , )
def __a ( self :List[str] , _lowercase :Tuple , _lowercase :Any , _lowercase :Optional[Any]=None , _lowercase :Optional[Any]=None , _lowercase :Optional[Any]=None , _lowercase :List[Any]=None , _lowercase :List[Any]="auto" , _lowercase :List[str]=-1 , _lowercase :List[str]=0.9 , _lowercase :Dict=5 , _lowercase :Dict=500 , _lowercase :Any="gpt2-large" , _lowercase :str=-1 , _lowercase :List[str]=1024 , _lowercase :str=25 , _lowercase :Optional[Any]=5 , _lowercase :List[Any]=True , _lowercase :int=25 , ) -> Optional[Any]:
UpperCAmelCase_ = compute_mauve(
p_text=_lowercase , q_text=_lowercase , p_features=_lowercase , q_features=_lowercase , p_tokens=_lowercase , q_tokens=_lowercase , num_buckets=_lowercase , pca_max_data=_lowercase , kmeans_explained_var=_lowercase , kmeans_num_redo=_lowercase , kmeans_max_iter=_lowercase , featurize_model_name=_lowercase , device_id=_lowercase , max_text_length=_lowercase , divergence_curve_discretization_size=_lowercase , mauve_scaling_factor=_lowercase , verbose=_lowercase , seed=_lowercase , )
return out
| 344 |
def A ( __UpperCAmelCase = 100_0000 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = [i - 1 for i in range(limit + 1 )]
for i in range(2 , limit + 1 ):
if phi[i] == i - 1:
for j in range(2 * i , limit + 1 , __UpperCAmelCase ):
phi[j] -= phi[j] // i
return sum(phi[2 : limit + 1] )
if __name__ == "__main__":
print(solution())
| 344 | 1 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
UpperCamelCase_ = {
"configuration_roberta": ["ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP", "RobertaConfig", "RobertaOnnxConfig"],
"tokenization_roberta": ["RobertaTokenizer"],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["RobertaTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST",
"RobertaForCausalLM",
"RobertaForMaskedLM",
"RobertaForMultipleChoice",
"RobertaForQuestionAnswering",
"RobertaForSequenceClassification",
"RobertaForTokenClassification",
"RobertaModel",
"RobertaPreTrainedModel",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"TF_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST",
"TFRobertaForCausalLM",
"TFRobertaForMaskedLM",
"TFRobertaForMultipleChoice",
"TFRobertaForQuestionAnswering",
"TFRobertaForSequenceClassification",
"TFRobertaForTokenClassification",
"TFRobertaMainLayer",
"TFRobertaModel",
"TFRobertaPreTrainedModel",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxRobertaForCausalLM",
"FlaxRobertaForMaskedLM",
"FlaxRobertaForMultipleChoice",
"FlaxRobertaForQuestionAnswering",
"FlaxRobertaForSequenceClassification",
"FlaxRobertaForTokenClassification",
"FlaxRobertaModel",
"FlaxRobertaPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_roberta import ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP, RobertaConfig, RobertaOnnxConfig
from .tokenization_roberta import RobertaTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_roberta_fast import RobertaTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_roberta import (
ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST,
RobertaForCausalLM,
RobertaForMaskedLM,
RobertaForMultipleChoice,
RobertaForQuestionAnswering,
RobertaForSequenceClassification,
RobertaForTokenClassification,
RobertaModel,
RobertaPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_roberta import (
TF_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST,
TFRobertaForCausalLM,
TFRobertaForMaskedLM,
TFRobertaForMultipleChoice,
TFRobertaForQuestionAnswering,
TFRobertaForSequenceClassification,
TFRobertaForTokenClassification,
TFRobertaMainLayer,
TFRobertaModel,
TFRobertaPreTrainedModel,
)
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_roberta import (
FlaxRobertaForCausalLM,
FlaxRobertaForMaskedLM,
FlaxRobertaForMultipleChoice,
FlaxRobertaForQuestionAnswering,
FlaxRobertaForSequenceClassification,
FlaxRobertaForTokenClassification,
FlaxRobertaModel,
FlaxRobertaPreTrainedModel,
)
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
import tempfile
import torch
from diffusers import PNDMScheduler
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =(PNDMScheduler,)
UpperCamelCase__ : Optional[Any] =(("num_inference_steps", 50),)
def __a ( self :Union[str, Any] , **_lowercase :Any) -> Union[str, Any]:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
}
config.update(**_lowercase)
return config
def __a ( self :str , _lowercase :List[Any]=0 , **_lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Any) -> Optional[Any]:
pass
def __a ( self :str , _lowercase :int=0 , **_lowercase :Union[str, Any]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :int , **_lowercase :str) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.prk_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
for i, t in enumerate(scheduler.plms_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :Union[str, Any]) -> int:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
if num_inference_steps is not None and hasattr(_lowercase , '''set_timesteps'''):
scheduler.set_timesteps(_lowercase)
elif num_inference_steps is not None and not hasattr(_lowercase , '''set_timesteps'''):
UpperCAmelCase_ = num_inference_steps
# copy over dummy past residuals (must be done after set_timesteps)
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
def __a ( self :Any) -> Dict:
for timesteps in [100, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :List[Any]) -> Any:
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(steps_offset=1)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(10)
assert torch.equal(
scheduler.timesteps , torch.LongTensor(
[901, 851, 851, 801, 801, 751, 751, 701, 701, 651, 651, 601, 601, 501, 401, 301, 201, 101, 1]) , )
def __a ( self :Optional[int]) -> str:
for beta_start, beta_end in zip([0.0_001, 0.001] , [0.002, 0.02]):
self.check_over_configs(beta_start=_lowercase , beta_end=_lowercase)
def __a ( self :Any) -> List[str]:
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=_lowercase)
def __a ( self :List[Any]) -> Dict:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Tuple:
for t in [1, 5, 10]:
self.check_over_forward(time_step=_lowercase)
def __a ( self :Tuple) -> Dict:
for t, num_inference_steps in zip([1, 5, 10] , [10, 50, 100]):
self.check_over_forward(num_inference_steps=_lowercase)
def __a ( self :str) -> List[Any]:
# earlier version of set_timesteps() caused an error indexing alpha's with inference steps as power of 3
UpperCAmelCase_ = 27
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# before power of 3 fix, would error on first step, so we only need to do two
for i, t in enumerate(scheduler.prk_timesteps[:2]):
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
def __a ( self :List[str]) -> int:
with self.assertRaises(_lowercase):
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.step_plms(self.dummy_sample , 1 , self.dummy_sample).prev_sample
def __a ( self :List[str]) -> Dict:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 198.1_318) < 1E-2
assert abs(result_mean.item() - 0.2_580) < 1E-3
def __a ( self :Any) -> Tuple:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 67.3_986) < 1E-2
assert abs(result_mean.item() - 0.0_878) < 1E-3
def __a ( self :int) -> Any:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 230.0_399) < 1E-2
assert abs(result_mean.item() - 0.2_995) < 1E-3
def __a ( self :Any) -> Dict:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 186.9_482) < 1E-2
assert abs(result_mean.item() - 0.2_434) < 1E-3
| 344 | 1 |
import copy
import os
from typing import Union
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"microsoft/git-base": "https://huggingface.co/microsoft/git-base/resolve/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : List[str] ="git_vision_model"
def __init__( self :Optional[Any] , _lowercase :Optional[int]=768 , _lowercase :Any=3072 , _lowercase :Dict=12 , _lowercase :Optional[Any]=12 , _lowercase :int=3 , _lowercase :List[str]=224 , _lowercase :Dict=16 , _lowercase :List[str]="quick_gelu" , _lowercase :List[str]=1E-5 , _lowercase :Any=0.0 , _lowercase :Optional[int]=0.02 , **_lowercase :List[Any] , ) -> Optional[Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = image_size
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = attention_dropout
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = hidden_act
@classmethod
def __a ( cls :Any , _lowercase :Union[str, os.PathLike] , **_lowercase :Optional[int]) -> "PretrainedConfig":
cls._set_token_in_kwargs(_lowercase)
UpperCAmelCase_ , UpperCAmelCase_ = cls.get_config_dict(_lowercase , **_lowercase)
# get the vision config dict if we are loading from GITConfig
if config_dict.get('''model_type''') == "git":
UpperCAmelCase_ = config_dict['''vision_config''']
if "model_type" in config_dict and hasattr(cls , '''model_type''') and config_dict["model_type"] != cls.model_type:
logger.warning(
f"You are using a model of type {config_dict['model_type']} to instantiate a model of type "
f"{cls.model_type}. This is not supported for all configurations of models and can yield errors.")
return cls.from_dict(_lowercase , **_lowercase)
class a_ ( _snake_case ):
UpperCamelCase__ : Any ="git"
def __init__( self :int , _lowercase :Optional[Any]=None , _lowercase :Optional[int]=30522 , _lowercase :Union[str, Any]=768 , _lowercase :Union[str, Any]=6 , _lowercase :Dict=12 , _lowercase :Dict=3072 , _lowercase :Any="gelu" , _lowercase :Union[str, Any]=0.1 , _lowercase :Any=0.1 , _lowercase :Any=1024 , _lowercase :Union[str, Any]=0.02 , _lowercase :Any=1E-1_2 , _lowercase :Tuple=0 , _lowercase :Tuple="absolute" , _lowercase :Optional[int]=True , _lowercase :Dict=False , _lowercase :Tuple=101 , _lowercase :Optional[int]=102 , _lowercase :Optional[Any]=None , **_lowercase :Dict , ) -> Optional[Any]:
super().__init__(bos_token_id=_lowercase , eos_token_id=_lowercase , pad_token_id=_lowercase , **_lowercase)
if vision_config is None:
UpperCAmelCase_ = {}
logger.info('''vision_config is None. initializing the GitVisionConfig with default values.''')
UpperCAmelCase_ = GitVisionConfig(**_lowercase)
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = position_embedding_type
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = tie_word_embeddings
UpperCAmelCase_ = num_image_with_embedding
UpperCAmelCase_ = bos_token_id
UpperCAmelCase_ = eos_token_id
def __a ( self :Tuple) -> Tuple:
UpperCAmelCase_ = copy.deepcopy(self.__dict__)
UpperCAmelCase_ = self.vision_config.to_dict()
UpperCAmelCase_ = self.__class__.model_type
return output
| 344 |
import os
from dataclasses import dataclass, field
from io import BytesIO
from typing import TYPE_CHECKING, Any, ClassVar, Dict, Optional, Union
import numpy as np
import pyarrow as pa
from .. import config
from ..download.streaming_download_manager import xopen, xsplitext
from ..table import array_cast
from ..utils.py_utils import no_op_if_value_is_null, string_to_dict
if TYPE_CHECKING:
from .features import FeatureType
UpperCamelCase_ , UpperCamelCase_ , UpperCamelCase_ = False, False, False
@dataclass
class a_ :
UpperCamelCase__ : Optional[int] =None
UpperCamelCase__ : bool =True
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[str] =None
# Automatically constructed
UpperCamelCase__ : ClassVar[str] ="dict"
UpperCamelCase__ : ClassVar[Any] =pa.struct({"bytes": pa.binary(), "path": pa.string()} )
UpperCamelCase__ : str =field(default="Audio" , init=_snake_case , repr=_snake_case )
def __call__( self :List[Any]) -> List[Any]:
return self.pa_type
def __a ( self :Any , _lowercase :Union[str, bytes, dict]) -> dict:
try:
import soundfile as sf # soundfile is a dependency of librosa, needed to decode audio files.
except ImportError as err:
raise ImportError('''To support encoding audio data, please install \'soundfile\'.''') from err
if isinstance(_lowercase , _lowercase):
return {"bytes": None, "path": value}
elif isinstance(_lowercase , _lowercase):
return {"bytes": value, "path": None}
elif "array" in value:
# convert the audio array to wav bytes
UpperCAmelCase_ = BytesIO()
sf.write(_lowercase , value['''array'''] , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
elif value.get('''path''') is not None and os.path.isfile(value['''path''']):
# we set "bytes": None to not duplicate the data if they're already available locally
if value["path"].endswith('''pcm'''):
# "PCM" only has raw audio bytes
if value.get('''sampling_rate''') is None:
# At least, If you want to convert "PCM-byte" to "WAV-byte", you have to know sampling rate
raise KeyError('''To use PCM files, please specify a \'sampling_rate\' in Audio object''')
if value.get('''bytes'''):
# If we already had PCM-byte, we don`t have to make "read file, make bytes" (just use it!)
UpperCAmelCase_ = np.frombuffer(value['''bytes'''] , dtype=np.intaa).astype(np.floataa) / 32767
else:
UpperCAmelCase_ = np.memmap(value['''path'''] , dtype='''h''' , mode='''r''').astype(np.floataa) / 32767
UpperCAmelCase_ = BytesIO(bytes())
sf.write(_lowercase , _lowercase , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
else:
return {"bytes": None, "path": value.get('''path''')}
elif value.get('''bytes''') is not None or value.get('''path''') is not None:
# store the audio bytes, and path is used to infer the audio format using the file extension
return {"bytes": value.get('''bytes'''), "path": value.get('''path''')}
else:
raise ValueError(
f"An audio sample should have one of 'path' or 'bytes' but they are missing or None in {value}.")
def __a ( self :Dict , _lowercase :dict , _lowercase :Optional[Dict[str, Union[str, bool, None]]] = None) -> dict:
if not self.decode:
raise RuntimeError('''Decoding is disabled for this feature. Please use Audio(decode=True) instead.''')
UpperCAmelCase_ , UpperCAmelCase_ = (value['''path'''], BytesIO(value['''bytes'''])) if value['''bytes'''] is not None else (value['''path'''], None)
if path is None and file is None:
raise ValueError(f"An audio sample should have one of 'path' or 'bytes' but both are None in {value}.")
try:
import librosa
import soundfile as sf
except ImportError as err:
raise ImportError('''To support decoding audio files, please install \'librosa\' and \'soundfile\'.''') from err
UpperCAmelCase_ = xsplitext(_lowercase)[1][1:].lower() if path is not None else None
if not config.IS_OPUS_SUPPORTED and audio_format == "opus":
raise RuntimeError(
'''Decoding \'opus\' files requires system library \'libsndfile\'>=1.0.31, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
elif not config.IS_MP3_SUPPORTED and audio_format == "mp3":
raise RuntimeError(
'''Decoding \'mp3\' files requires system library \'libsndfile\'>=1.1.0, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
if file is None:
UpperCAmelCase_ = token_per_repo_id or {}
UpperCAmelCase_ = path.split('''::''')[-1]
try:
UpperCAmelCase_ = string_to_dict(_lowercase , config.HUB_DATASETS_URL)['''repo_id''']
UpperCAmelCase_ = token_per_repo_id[repo_id]
except (ValueError, KeyError):
UpperCAmelCase_ = None
with xopen(_lowercase , '''rb''' , use_auth_token=_lowercase) as f:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
else:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
UpperCAmelCase_ = array.T
if self.mono:
UpperCAmelCase_ = librosa.to_mono(_lowercase)
if self.sampling_rate and self.sampling_rate != sampling_rate:
UpperCAmelCase_ = librosa.resample(_lowercase , orig_sr=_lowercase , target_sr=self.sampling_rate)
UpperCAmelCase_ = self.sampling_rate
return {"path": path, "array": array, "sampling_rate": sampling_rate}
def __a ( self :Union[str, Any]) -> Union["FeatureType", Dict[str, "FeatureType"]]:
from .features import Value
if self.decode:
raise ValueError('''Cannot flatten a decoded Audio feature.''')
return {
"bytes": Value('''binary'''),
"path": Value('''string'''),
}
def __a ( self :int , _lowercase :Union[pa.StringArray, pa.StructArray]) -> pa.StructArray:
if pa.types.is_string(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, storage] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_binary(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([storage, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_struct(storage.type) and storage.type.get_all_field_indices('''array'''):
UpperCAmelCase_ = pa.array([Audio().encode_example(_lowercase) if x is not None else None for x in storage.to_pylist()])
elif pa.types.is_struct(storage.type):
if storage.type.get_field_index('''bytes''') >= 0:
UpperCAmelCase_ = storage.field('''bytes''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
if storage.type.get_field_index('''path''') >= 0:
UpperCAmelCase_ = storage.field('''path''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
return array_cast(_lowercase , self.pa_type)
def __a ( self :Any , _lowercase :pa.StructArray) -> pa.StructArray:
@no_op_if_value_is_null
def path_to_bytes(_lowercase :Tuple):
with xopen(_lowercase , '''rb''') as f:
UpperCAmelCase_ = f.read()
return bytes_
UpperCAmelCase_ = pa.array(
[
(path_to_bytes(x['''path''']) if x['''bytes'''] is None else x['''bytes''']) if x is not None else None
for x in storage.to_pylist()
] , type=pa.binary() , )
UpperCAmelCase_ = pa.array(
[os.path.basename(_lowercase) if path is not None else None for path in storage.field('''path''').to_pylist()] , type=pa.string() , )
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
| 344 | 1 |
import math
def A ( __UpperCAmelCase , __UpperCAmelCase = 0 , __UpperCAmelCase = 0 ) -> list:
'''simple docstring'''
UpperCAmelCase_ = end or len(__UpperCAmelCase )
for i in range(__UpperCAmelCase , __UpperCAmelCase ):
UpperCAmelCase_ = i
UpperCAmelCase_ = array[i]
while temp_index != start and temp_index_value < array[temp_index - 1]:
UpperCAmelCase_ = array[temp_index - 1]
temp_index -= 1
UpperCAmelCase_ = temp_index_value
return array
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> None: # Max Heap
'''simple docstring'''
UpperCAmelCase_ = index
UpperCAmelCase_ = 2 * index + 1 # Left Node
UpperCAmelCase_ = 2 * index + 2 # Right Node
if left_index < heap_size and array[largest] < array[left_index]:
UpperCAmelCase_ = left_index
if right_index < heap_size and array[largest] < array[right_index]:
UpperCAmelCase_ = right_index
if largest != index:
UpperCAmelCase_ , UpperCAmelCase_ = array[largest], array[index]
heapify(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
def A ( __UpperCAmelCase ) -> list:
'''simple docstring'''
UpperCAmelCase_ = len(__UpperCAmelCase )
for i in range(n // 2 , -1 , -1 ):
heapify(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
for i in range(n - 1 , 0 , -1 ):
UpperCAmelCase_ , UpperCAmelCase_ = array[0], array[i]
heapify(__UpperCAmelCase , 0 , __UpperCAmelCase )
return array
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
if (array[first_index] > array[middle_index]) != (
array[first_index] > array[last_index]
):
return array[first_index]
elif (array[middle_index] > array[first_index]) != (
array[middle_index] > array[last_index]
):
return array[middle_index]
else:
return array[last_index]
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = low
UpperCAmelCase_ = high
while True:
while array[i] < pivot:
i += 1
j -= 1
while pivot < array[j]:
j -= 1
if i >= j:
return i
UpperCAmelCase_ , UpperCAmelCase_ = array[j], array[i]
i += 1
def A ( __UpperCAmelCase ) -> list:
'''simple docstring'''
if len(__UpperCAmelCase ) == 0:
return array
UpperCAmelCase_ = 2 * math.ceil(math.loga(len(__UpperCAmelCase ) ) )
UpperCAmelCase_ = 16
return intro_sort(__UpperCAmelCase , 0 , len(__UpperCAmelCase ) , __UpperCAmelCase , __UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> list:
'''simple docstring'''
while end - start > size_threshold:
if max_depth == 0:
return heap_sort(__UpperCAmelCase )
max_depth -= 1
UpperCAmelCase_ = median_of_a(__UpperCAmelCase , __UpperCAmelCase , start + ((end - start) // 2) + 1 , end - 1 )
UpperCAmelCase_ = partition(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
intro_sort(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
UpperCAmelCase_ = p
return insertion_sort(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
UpperCamelCase_ = input("Enter numbers separated by a comma : ").strip()
UpperCamelCase_ = [float(item) for item in user_input.split(",")]
print(sort(unsorted))
| 344 |
from ..utils import DummyObject, requires_backends
class a_ ( metaclass=_snake_case ):
UpperCamelCase__ : Any =["torch", "scipy"]
def __init__( self :List[str] , *_lowercase :List[str] , **_lowercase :Union[str, Any]) -> List[Any]:
requires_backends(self , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Dict , *_lowercase :Any , **_lowercase :Dict) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Optional[Any] , *_lowercase :str , **_lowercase :Optional[Any]) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
| 344 | 1 |
from sklearn.metrics import recall_score
import datasets
UpperCamelCase_ = "\nRecall is the fraction of the positive examples that were correctly labeled by the model as positive. It can be computed with the equation:\nRecall = TP / (TP + FN)\nWhere TP is the true positives and FN is the false negatives.\n"
UpperCamelCase_ = "\nArgs:\n- **predictions** (`list` of `int`): The predicted labels.\n- **references** (`list` of `int`): The ground truth labels.\n- **labels** (`list` of `int`): The set of labels to include when `average` is not set to `binary`, and their order when average is `None`. Labels present in the data can be excluded in this input, for example to calculate a multiclass average ignoring a majority negative class, while labels not present in the data will result in 0 components in a macro average. For multilabel targets, labels are column indices. By default, all labels in y_true and y_pred are used in sorted order. Defaults to None.\n- **pos_label** (`int`): The class label to use as the 'positive class' when calculating the recall. Defaults to `1`.\n- **average** (`string`): This parameter is required for multiclass/multilabel targets. If None, the scores for each class are returned. Otherwise, this determines the type of averaging performed on the data. Defaults to `'binary'`.\n - `'binary'`: Only report results for the class specified by `pos_label`. This is applicable only if the target labels and predictions are binary.\n - `'micro'`: Calculate metrics globally by counting the total true positives, false negatives, and false positives.\n - `'macro'`: Calculate metrics for each label, and find their unweighted mean. This does not take label imbalance into account.\n - `'weighted'`: Calculate metrics for each label, and find their average weighted by support (the number of true instances for each label). This alters `'macro'` to account for label imbalance. Note that it can result in an F-score that is not between precision and recall.\n - `'samples'`: Calculate metrics for each instance, and find their average (only meaningful for multilabel classification).\n- **sample_weight** (`list` of `float`): Sample weights Defaults to `None`.\n- **zero_division** (): Sets the value to return when there is a zero division. Defaults to .\n - `'warn'`: If there is a zero division, the return value is `0`, but warnings are also raised.\n - `0`: If there is a zero division, the return value is `0`.\n - `1`: If there is a zero division, the return value is `1`.\n\nReturns:\n- **recall** (`float`, or `array` of `float`): Either the general recall score, or the recall scores for individual classes, depending on the values input to `labels` and `average`. Minimum possible value is 0. Maximum possible value is 1. A higher recall means that more of the positive examples have been labeled correctly. Therefore, a higher recall is generally considered better.\n\nExamples:\n\n Example 1-A simple example with some errors\n >>> recall_metric = datasets.load_metric('recall')\n >>> results = recall_metric.compute(references=[0, 0, 1, 1, 1], predictions=[0, 1, 0, 1, 1])\n >>> print(results)\n {'recall': 0.6666666666666666}\n\n Example 2-The same example as Example 1, but with `pos_label=0` instead of the default `pos_label=1`.\n >>> recall_metric = datasets.load_metric('recall')\n >>> results = recall_metric.compute(references=[0, 0, 1, 1, 1], predictions=[0, 1, 0, 1, 1], pos_label=0)\n >>> print(results)\n {'recall': 0.5}\n\n Example 3-The same example as Example 1, but with `sample_weight` included.\n >>> recall_metric = datasets.load_metric('recall')\n >>> sample_weight = [0.9, 0.2, 0.9, 0.3, 0.8]\n >>> results = recall_metric.compute(references=[0, 0, 1, 1, 1], predictions=[0, 1, 0, 1, 1], sample_weight=sample_weight)\n >>> print(results)\n {'recall': 0.55}\n\n Example 4-A multiclass example, using different averages.\n >>> recall_metric = datasets.load_metric('recall')\n >>> predictions = [0, 2, 1, 0, 0, 1]\n >>> references = [0, 1, 2, 0, 1, 2]\n >>> results = recall_metric.compute(predictions=predictions, references=references, average='macro')\n >>> print(results)\n {'recall': 0.3333333333333333}\n >>> results = recall_metric.compute(predictions=predictions, references=references, average='micro')\n >>> print(results)\n {'recall': 0.3333333333333333}\n >>> results = recall_metric.compute(predictions=predictions, references=references, average='weighted')\n >>> print(results)\n {'recall': 0.3333333333333333}\n >>> results = recall_metric.compute(predictions=predictions, references=references, average=None)\n >>> print(results)\n {'recall': array([1., 0., 0.])}\n"
UpperCamelCase_ = "\n@article{scikit-learn, title={Scikit-learn: Machine Learning in {P}ython}, author={Pedregosa, F. and Varoquaux, G. and Gramfort, A. and Michel, V. and Thirion, B. and Grisel, O. and Blondel, M. and Prettenhofer, P. and Weiss, R. and Dubourg, V. and Vanderplas, J. and Passos, A. and Cournapeau, D. and Brucher, M. and Perrot, M. and Duchesnay, E.}, journal={Journal of Machine Learning Research}, volume={12}, pages={2825--2830}, year={2011}\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class a_ ( datasets.Metric ):
def __a ( self :int) -> Dict:
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
'''predictions''': datasets.Sequence(datasets.Value('''int32''')),
'''references''': datasets.Sequence(datasets.Value('''int32''')),
}
if self.config_name == '''multilabel'''
else {
'''predictions''': datasets.Value('''int32'''),
'''references''': datasets.Value('''int32'''),
}) , reference_urls=['''https://scikit-learn.org/stable/modules/generated/sklearn.metrics.recall_score.html'''] , )
def __a ( self :Tuple , _lowercase :Optional[Any] , _lowercase :str , _lowercase :Optional[int]=None , _lowercase :Dict=1 , _lowercase :List[str]="binary" , _lowercase :Tuple=None , _lowercase :str="warn" , ) -> Dict:
UpperCAmelCase_ = recall_score(
_lowercase , _lowercase , labels=_lowercase , pos_label=_lowercase , average=_lowercase , sample_weight=_lowercase , zero_division=_lowercase , )
return {"recall": float(_lowercase) if score.size == 1 else score}
| 344 |
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def A ( __UpperCAmelCase ) -> Dict[str, torch.Tensor]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for rt in rc.restypes:
UpperCAmelCase_ = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
UpperCAmelCase_ = {name: i for i, name in enumerate(__UpperCAmelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.floataa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = protein['''aatype'''].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the corresponding mask
UpperCAmelCase_ = torch.zeros([21, 37] , dtype=torch.floataa , device=protein['''aatype'''].device )
for restype, restype_letter in enumerate(rc.restypes ):
UpperCAmelCase_ = rc.restype_atoa[restype_letter]
UpperCAmelCase_ = rc.residue_atoms[restype_name]
for atom_name in atom_names:
UpperCAmelCase_ = rc.atom_order[atom_name]
UpperCAmelCase_ = 1
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
return protein
def A ( __UpperCAmelCase ) -> Dict[str, np.ndarray]:
'''simple docstring'''
UpperCAmelCase_ = tree_map(lambda __UpperCAmelCase : torch.tensor(__UpperCAmelCase , device=batch['''aatype'''].device ) , __UpperCAmelCase , np.ndarray )
UpperCAmelCase_ = tensor_tree_map(lambda __UpperCAmelCase : np.array(__UpperCAmelCase ) , make_atomaa_masks(__UpperCAmelCase ) )
return out
| 344 | 1 |
from typing import Dict, List, Optional, Tuple, Union
import numpy as np
from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
from ...image_transforms import (
center_crop,
get_resize_output_image_size,
normalize,
rescale,
resize,
to_channel_dimension_format,
)
from ...image_utils import (
IMAGENET_STANDARD_MEAN,
IMAGENET_STANDARD_STD,
ChannelDimension,
ImageInput,
PILImageResampling,
make_list_of_images,
to_numpy_array,
valid_images,
)
from ...utils import TensorType, is_torch_available, is_torch_tensor, logging
if is_torch_available():
import torch
UpperCamelCase_ = logging.get_logger(__name__)
class a_ ( _snake_case ):
UpperCamelCase__ : Tuple =["pixel_values"]
def __init__( self :List[str] , _lowercase :bool = True , _lowercase :Optional[Dict[str, int]] = None , _lowercase :PILImageResampling = PILImageResampling.BILINEAR , _lowercase :bool = True , _lowercase :Dict[str, int] = None , _lowercase :bool = True , _lowercase :Union[int, float] = 1 / 255 , _lowercase :bool = True , _lowercase :Optional[Union[float, List[float]]] = None , _lowercase :Optional[Union[float, List[float]]] = None , **_lowercase :Tuple , ) -> None:
super().__init__(**_lowercase)
UpperCAmelCase_ = size if size is not None else {'''shortest_edge''': 256}
UpperCAmelCase_ = get_size_dict(_lowercase , default_to_square=_lowercase)
UpperCAmelCase_ = crop_size if crop_size is not None else {'''height''': 224, '''width''': 224}
UpperCAmelCase_ = get_size_dict(_lowercase , param_name='''crop_size''')
UpperCAmelCase_ = do_resize
UpperCAmelCase_ = size
UpperCAmelCase_ = resample
UpperCAmelCase_ = do_center_crop
UpperCAmelCase_ = crop_size
UpperCAmelCase_ = do_rescale
UpperCAmelCase_ = rescale_factor
UpperCAmelCase_ = do_normalize
UpperCAmelCase_ = image_mean if image_mean is not None else IMAGENET_STANDARD_MEAN
UpperCAmelCase_ = image_std if image_std is not None else IMAGENET_STANDARD_STD
def __a ( self :Any , _lowercase :np.ndarray , _lowercase :Dict[str, int] , _lowercase :PILImageResampling = PILImageResampling.BICUBIC , _lowercase :Optional[Union[str, ChannelDimension]] = None , **_lowercase :Union[str, Any] , ) -> np.ndarray:
UpperCAmelCase_ = get_size_dict(_lowercase , default_to_square=_lowercase)
if "shortest_edge" not in size:
raise ValueError(f"The `size` parameter must contain the key `shortest_edge`. Got {size.keys()}")
UpperCAmelCase_ = get_resize_output_image_size(_lowercase , size=size['''shortest_edge'''] , default_to_square=_lowercase)
return resize(_lowercase , size=_lowercase , resample=_lowercase , data_format=_lowercase , **_lowercase)
def __a ( self :Optional[int] , _lowercase :np.ndarray , _lowercase :Dict[str, int] , _lowercase :Optional[Union[str, ChannelDimension]] = None , **_lowercase :Union[str, Any] , ) -> np.ndarray:
UpperCAmelCase_ = get_size_dict(_lowercase)
if "height" not in size or "width" not in size:
raise ValueError(f"The `size` parameter must contain the keys `height` and `width`. Got {size.keys()}")
return center_crop(_lowercase , size=(size['''height'''], size['''width''']) , data_format=_lowercase , **_lowercase)
def __a ( self :List[str] , _lowercase :np.ndarray , _lowercase :float , _lowercase :Optional[Union[str, ChannelDimension]] = None , **_lowercase :Optional[Any]) -> np.ndarray:
return rescale(_lowercase , scale=_lowercase , data_format=_lowercase , **_lowercase)
def __a ( self :Union[str, Any] , _lowercase :np.ndarray , _lowercase :Union[float, List[float]] , _lowercase :Union[float, List[float]] , _lowercase :Optional[Union[str, ChannelDimension]] = None , **_lowercase :int , ) -> np.ndarray:
return normalize(_lowercase , mean=_lowercase , std=_lowercase , data_format=_lowercase , **_lowercase)
def __a ( self :Optional[int] , _lowercase :ImageInput , _lowercase :Optional[bool] = None , _lowercase :Dict[str, int] = None , _lowercase :PILImageResampling = None , _lowercase :bool = None , _lowercase :Dict[str, int] = None , _lowercase :Optional[bool] = None , _lowercase :Optional[float] = None , _lowercase :Optional[bool] = None , _lowercase :Optional[Union[float, List[float]]] = None , _lowercase :Optional[Union[float, List[float]]] = None , _lowercase :Optional[Union[str, TensorType]] = None , _lowercase :Union[str, ChannelDimension] = ChannelDimension.FIRST , **_lowercase :Dict , ) -> int:
UpperCAmelCase_ = do_resize if do_resize is not None else self.do_resize
UpperCAmelCase_ = size if size is not None else self.size
UpperCAmelCase_ = get_size_dict(_lowercase , default_to_square=_lowercase)
UpperCAmelCase_ = resample if resample is not None else self.resample
UpperCAmelCase_ = do_center_crop if do_center_crop is not None else self.do_center_crop
UpperCAmelCase_ = crop_size if crop_size is not None else self.crop_size
UpperCAmelCase_ = get_size_dict(_lowercase , param_name='''crop_size''')
UpperCAmelCase_ = do_rescale if do_rescale is not None else self.do_rescale
UpperCAmelCase_ = rescale_factor if rescale_factor is not None else self.rescale_factor
UpperCAmelCase_ = do_normalize if do_normalize is not None else self.do_normalize
UpperCAmelCase_ = image_mean if image_mean is not None else self.image_mean
UpperCAmelCase_ = image_std if image_std is not None else self.image_std
UpperCAmelCase_ = make_list_of_images(_lowercase)
if not valid_images(_lowercase):
raise ValueError(
'''Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, '''
'''torch.Tensor, tf.Tensor or jax.ndarray.''')
if do_resize and size is None:
raise ValueError('''Size must be specified if do_resize is True.''')
if do_center_crop and crop_size is None:
raise ValueError('''Crop size must be specified if do_center_crop is True.''')
if do_rescale and rescale_factor is None:
raise ValueError('''Rescale factor must be specified if do_rescale is True.''')
if do_normalize and (image_mean is None or image_std is None):
raise ValueError('''Image mean and std must be specified if do_normalize is True.''')
# All transformations expect numpy arrays.
UpperCAmelCase_ = [to_numpy_array(_lowercase) for image in images]
if do_resize:
UpperCAmelCase_ = [self.resize(image=_lowercase , size=_lowercase , resample=_lowercase) for image in images]
if do_center_crop:
UpperCAmelCase_ = [self.center_crop(image=_lowercase , size=_lowercase) for image in images]
if do_rescale:
UpperCAmelCase_ = [self.rescale(image=_lowercase , scale=_lowercase) for image in images]
if do_normalize:
UpperCAmelCase_ = [self.normalize(image=_lowercase , mean=_lowercase , std=_lowercase) for image in images]
UpperCAmelCase_ = [to_channel_dimension_format(_lowercase , _lowercase) for image in images]
UpperCAmelCase_ = {'''pixel_values''': images}
return BatchFeature(data=_lowercase , tensor_type=_lowercase)
def __a ( self :Tuple , _lowercase :Optional[Any] , _lowercase :List[Tuple] = None) -> List[Any]:
UpperCAmelCase_ = outputs.logits
# Resize logits and compute semantic segmentation maps
if target_sizes is not None:
if len(_lowercase) != len(_lowercase):
raise ValueError(
'''Make sure that you pass in as many target sizes as the batch dimension of the logits''')
if is_torch_tensor(_lowercase):
UpperCAmelCase_ = target_sizes.numpy()
UpperCAmelCase_ = []
for idx in range(len(_lowercase)):
UpperCAmelCase_ = torch.nn.functional.interpolate(
logits[idx].unsqueeze(dim=0) , size=target_sizes[idx] , mode='''bilinear''' , align_corners=_lowercase)
UpperCAmelCase_ = resized_logits[0].argmax(dim=0)
semantic_segmentation.append(_lowercase)
else:
UpperCAmelCase_ = logits.argmax(dim=1)
UpperCAmelCase_ = [semantic_segmentation[i] for i in range(semantic_segmentation.shape[0])]
return semantic_segmentation
| 344 |
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from ..models.whisper import WhisperForConditionalGeneration, WhisperProcessor
from .base import PipelineTool
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="openai/whisper-base"
UpperCamelCase__ : int =(
"This is a tool that transcribes an audio into text. It takes an input named `audio` and returns the "
"transcribed text."
)
UpperCamelCase__ : Any ="transcriber"
UpperCamelCase__ : Optional[int] =WhisperProcessor
UpperCamelCase__ : List[str] =WhisperForConditionalGeneration
UpperCamelCase__ : List[Any] =["audio"]
UpperCamelCase__ : Union[str, Any] =["text"]
def __a ( self :int , _lowercase :Any) -> Tuple:
return self.pre_processor(_lowercase , return_tensors='''pt''').input_features
def __a ( self :Dict , _lowercase :Tuple) -> Any:
return self.model.generate(inputs=_lowercase)
def __a ( self :int , _lowercase :Union[str, Any]) -> Optional[Any]:
return self.pre_processor.batch_decode(_lowercase , skip_special_tokens=_lowercase)[0]
| 344 | 1 |
import math
def A ( __UpperCAmelCase ) -> str:
'''simple docstring'''
UpperCAmelCase_ = 0
UpperCAmelCase_ = 0
while num > 0:
UpperCAmelCase_ = num % 8
UpperCAmelCase_ = octal + (remainder * math.floor(math.pow(10 , __UpperCAmelCase ) ))
counter += 1
UpperCAmelCase_ = math.floor(num / 8 ) # basically /= 8 without remainder if any
# This formatting removes trailing '.0' from `octal`.
return f"0o{int(__UpperCAmelCase )}"
def A ( ) -> None:
'''simple docstring'''
print('''\n2 in octal is:''' )
print(decimal_to_octal(2 ) ) # = 2
print('''\n8 in octal is:''' )
print(decimal_to_octal(8 ) ) # = 10
print('''\n65 in octal is:''' )
print(decimal_to_octal(65 ) ) # = 101
print('''\n216 in octal is:''' )
print(decimal_to_octal(216 ) ) # = 330
print('''\n512 in octal is:''' )
print(decimal_to_octal(512 ) ) # = 1000
print('''\n''' )
if __name__ == "__main__":
main()
| 344 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
UpperCamelCase_ = {"configuration_opt": ["OPT_PRETRAINED_CONFIG_ARCHIVE_MAP", "OPTConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"OPT_PRETRAINED_MODEL_ARCHIVE_LIST",
"OPTForCausalLM",
"OPTModel",
"OPTPreTrainedModel",
"OPTForSequenceClassification",
"OPTForQuestionAnswering",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["TFOPTForCausalLM", "TFOPTModel", "TFOPTPreTrainedModel"]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxOPTForCausalLM",
"FlaxOPTModel",
"FlaxOPTPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_opt import OPT_PRETRAINED_CONFIG_ARCHIVE_MAP, OPTConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_opt import (
OPT_PRETRAINED_MODEL_ARCHIVE_LIST,
OPTForCausalLM,
OPTForQuestionAnswering,
OPTForSequenceClassification,
OPTModel,
OPTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_opt import TFOPTForCausalLM, TFOPTModel, TFOPTPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_opt import FlaxOPTForCausalLM, FlaxOPTModel, FlaxOPTPreTrainedModel
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 | 1 |
import copy
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
class a_ ( _snake_case ):
UpperCamelCase__ : int ="encoder-decoder"
UpperCamelCase__ : List[Any] =True
def __init__( self :str , **_lowercase :Any) -> List[Any]:
super().__init__(**_lowercase)
assert (
"encoder" in kwargs and "decoder" in kwargs
), "Config has to be initialized with encoder and decoder config"
UpperCAmelCase_ = kwargs.pop('''encoder''')
UpperCAmelCase_ = encoder_config.pop('''model_type''')
UpperCAmelCase_ = kwargs.pop('''decoder''')
UpperCAmelCase_ = decoder_config.pop('''model_type''')
from ..auto.configuration_auto import AutoConfig
UpperCAmelCase_ = AutoConfig.for_model(_lowercase , **_lowercase)
UpperCAmelCase_ = AutoConfig.for_model(_lowercase , **_lowercase)
UpperCAmelCase_ = True
@classmethod
def __a ( cls :int , _lowercase :PretrainedConfig , _lowercase :PretrainedConfig , **_lowercase :str) -> PretrainedConfig:
logger.info('''Set `config.is_decoder=True` and `config.add_cross_attention=True` for decoder_config''')
UpperCAmelCase_ = True
UpperCAmelCase_ = True
return cls(encoder=encoder_config.to_dict() , decoder=decoder_config.to_dict() , **_lowercase)
def __a ( self :List[Any]) -> int:
UpperCAmelCase_ = copy.deepcopy(self.__dict__)
UpperCAmelCase_ = self.encoder.to_dict()
UpperCAmelCase_ = self.decoder.to_dict()
UpperCAmelCase_ = self.__class__.model_type
return output
| 344 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
UpperCamelCase_ = "▁"
UpperCamelCase_ = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : str =BigBirdTokenizer
UpperCamelCase__ : Tuple =BigBirdTokenizerFast
UpperCamelCase__ : Union[str, Any] =True
UpperCamelCase__ : List[str] =True
def __a ( self :Any) -> List[str]:
super().setUp()
UpperCAmelCase_ = self.tokenizer_class(_lowercase , keep_accents=_lowercase)
tokenizer.save_pretrained(self.tmpdirname)
def __a ( self :Optional[int]) -> str:
UpperCAmelCase_ = '''<s>'''
UpperCAmelCase_ = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_lowercase) , _lowercase)
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_lowercase) , _lowercase)
def __a ( self :str) -> str:
UpperCAmelCase_ = list(self.get_tokenizer().get_vocab().keys())
self.assertEqual(vocab_keys[0] , '''<unk>''')
self.assertEqual(vocab_keys[1] , '''<s>''')
self.assertEqual(vocab_keys[-1] , '''[MASK]''')
self.assertEqual(len(_lowercase) , 1004)
def __a ( self :List[str]) -> int:
self.assertEqual(self.get_tokenizer().vocab_size , 1000)
def __a ( self :Tuple) -> int:
if not self.test_rust_tokenizer:
return
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = '''I was born in 92000, and this is falsé.'''
UpperCAmelCase_ = tokenizer.tokenize(_lowercase)
UpperCAmelCase_ = rust_tokenizer.tokenize(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = tokenizer.encode(_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :Optional[Any]) -> List[str]:
UpperCAmelCase_ = BigBirdTokenizer(_lowercase , keep_accents=_lowercase)
UpperCAmelCase_ = tokenizer.tokenize('''This is a test''')
self.assertListEqual(_lowercase , ['''▁This''', '''▁is''', '''▁a''', '''▁t''', '''est'''])
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_lowercase) , [285, 46, 10, 170, 382] , )
UpperCAmelCase_ = tokenizer.tokenize('''I was born in 92000, and this is falsé.''')
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''9''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''é''',
'''.''',
] , )
UpperCAmelCase_ = tokenizer.convert_tokens_to_ids(_lowercase)
self.assertListEqual(
_lowercase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
UpperCAmelCase_ = tokenizer.convert_ids_to_tokens(_lowercase)
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''<unk>''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''<unk>''',
'''.''',
] , )
@cached_property
def __a ( self :Any) -> List[Any]:
return BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
@slow
def __a ( self :int) -> List[Any]:
UpperCAmelCase_ = '''Hello World!'''
UpperCAmelCase_ = [65, 18536, 2260, 101, 66]
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@slow
def __a ( self :int) -> Any:
UpperCAmelCase_ = (
'''This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) " [ ] ! : - . Also we will'''
''' add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth'''
)
# fmt: off
UpperCAmelCase_ = [65, 871, 419, 358, 946, 991, 2521, 452, 358, 1357, 387, 7751, 3536, 112, 985, 456, 126, 865, 938, 5400, 5734, 458, 1368, 467, 786, 2462, 5246, 1159, 633, 865, 4519, 457, 582, 852, 2557, 427, 916, 508, 405, 34324, 497, 391, 408, 11342, 1244, 385, 100, 938, 985, 456, 574, 362, 12597, 3200, 3129, 1172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@require_torch
@slow
def __a ( self :Dict) -> Union[str, Any]:
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
UpperCAmelCase_ = list(self.big_tokenizer.get_vocab().keys())[:10]
UpperCAmelCase_ = ''' '''.join(_lowercase)
UpperCAmelCase_ = self.big_tokenizer.encode_plus(_lowercase , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = self.big_tokenizer.batch_encode_plus(
[sequence + ''' ''' + sequence] , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = BigBirdConfig(attention_type='''original_full''')
UpperCAmelCase_ = BigBirdModel(_lowercase)
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_lowercase)
model(**_lowercase)
@slow
def __a ( self :Optional[int]) -> Any:
UpperCAmelCase_ = BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
UpperCAmelCase_ = tokenizer.decode(tokenizer('''Paris is the [MASK].''').input_ids)
self.assertTrue(decoded_text == '''[CLS] Paris is the[MASK].[SEP]''')
@slow
def __a ( self :Dict) -> List[str]:
# fmt: off
UpperCAmelCase_ = {'''input_ids''': [[65, 39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114, 66], [65, 448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], '''attention_mask''': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_lowercase , model_name='''google/bigbird-roberta-base''' , revision='''215c99f1600e06f83acce68422f2035b2b5c3510''' , )
| 344 | 1 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_torch_available,
is_vision_available,
)
UpperCamelCase_ = {"configuration_vit": ["VIT_PRETRAINED_CONFIG_ARCHIVE_MAP", "ViTConfig", "ViTOnnxConfig"]}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["ViTFeatureExtractor"]
UpperCamelCase_ = ["ViTImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"VIT_PRETRAINED_MODEL_ARCHIVE_LIST",
"ViTForImageClassification",
"ViTForMaskedImageModeling",
"ViTModel",
"ViTPreTrainedModel",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"TFViTForImageClassification",
"TFViTModel",
"TFViTPreTrainedModel",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxViTForImageClassification",
"FlaxViTModel",
"FlaxViTPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_vit import VIT_PRETRAINED_CONFIG_ARCHIVE_MAP, ViTConfig, ViTOnnxConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_vit import ViTFeatureExtractor
from .image_processing_vit import ViTImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_vit import (
VIT_PRETRAINED_MODEL_ARCHIVE_LIST,
ViTForImageClassification,
ViTForMaskedImageModeling,
ViTModel,
ViTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_vit import TFViTForImageClassification, TFViTModel, TFViTPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_vit import FlaxViTForImageClassification, FlaxViTModel, FlaxViTPreTrainedModel
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
from __future__ import annotations
import string
from itertools import cycle, product
from pathlib import Path
UpperCamelCase_ = (
string.ascii_letters + string.digits + string.punctuation + string.whitespace
)
UpperCamelCase_ = [ord(letter) for letter in string.ascii_lowercase]
UpperCamelCase_ = {ord(char) for char in VALID_CHARS}
UpperCamelCase_ = ["the", "be", "to", "of", "and", "in", "that", "have"]
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str | None:
'''simple docstring'''
UpperCAmelCase_ = ""
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
for keychar, cipherchar in zip(cycle(__UpperCAmelCase ) , __UpperCAmelCase ):
UpperCAmelCase_ = cipherchar ^ keychar
if decodedchar not in VALID_INTS:
return None
decoded += chr(__UpperCAmelCase )
return decoded
def A ( __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
for key in product(__UpperCAmelCase , repeat=3 ):
UpperCAmelCase_ = try_key(__UpperCAmelCase , __UpperCAmelCase )
if encoded is not None:
possibles.append(__UpperCAmelCase )
return possibles
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
return [possible for possible in possibles if common_word in possible.lower()]
def A ( __UpperCAmelCase = "p059_cipher.txt" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = Path(__UpperCAmelCase ).parent.joinpath(__UpperCAmelCase ).read_text(encoding='''utf-8''' )
UpperCAmelCase_ = [int(__UpperCAmelCase ) for number in data.strip().split(''',''' )]
UpperCAmelCase_ = filter_valid_chars(__UpperCAmelCase )
for common_word in COMMON_WORDS:
UpperCAmelCase_ = filter_common_word(__UpperCAmelCase , __UpperCAmelCase )
if len(__UpperCAmelCase ) == 1:
break
UpperCAmelCase_ = possibles[0]
return sum(ord(__UpperCAmelCase ) for char in decoded_text )
if __name__ == "__main__":
print(f"{solution() = }")
| 344 | 1 |
import json
import os
import unittest
from transformers.models.xlm.tokenization_xlm import VOCAB_FILES_NAMES, XLMTokenizer
from transformers.testing_utils import slow
from ...test_tokenization_common import TokenizerTesterMixin
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : List[Any] =XLMTokenizer
UpperCamelCase__ : Tuple =False
def __a ( self :Optional[Any]) -> Tuple:
super().setUp()
# Adapted from Sennrich et al. 2015 and https://github.com/rsennrich/subword-nmt
UpperCAmelCase_ = [
'''l''',
'''o''',
'''w''',
'''e''',
'''r''',
'''s''',
'''t''',
'''i''',
'''d''',
'''n''',
'''w</w>''',
'''r</w>''',
'''t</w>''',
'''lo''',
'''low''',
'''er</w>''',
'''low</w>''',
'''lowest</w>''',
'''newer</w>''',
'''wider</w>''',
'''<unk>''',
]
UpperCAmelCase_ = dict(zip(_lowercase , range(len(_lowercase))))
UpperCAmelCase_ = ['''l o 123''', '''lo w 1456''', '''e r</w> 1789''', '''''']
UpperCAmelCase_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''])
UpperCAmelCase_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''merges_file'''])
with open(self.vocab_file , '''w''') as fp:
fp.write(json.dumps(_lowercase))
with open(self.merges_file , '''w''') as fp:
fp.write('''\n'''.join(_lowercase))
def __a ( self :Union[str, Any] , _lowercase :List[Any]) -> Tuple:
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = '''lower newer'''
return input_text, output_text
def __a ( self :str) -> Any:
UpperCAmelCase_ = XLMTokenizer(self.vocab_file , self.merges_file)
UpperCAmelCase_ = '''lower'''
UpperCAmelCase_ = ['''low''', '''er</w>''']
UpperCAmelCase_ = tokenizer.tokenize(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = tokens + ['''<unk>''']
UpperCAmelCase_ = [14, 15, 20]
self.assertListEqual(tokenizer.convert_tokens_to_ids(_lowercase) , _lowercase)
@slow
def __a ( self :List[Any]) -> List[Any]:
UpperCAmelCase_ = XLMTokenizer.from_pretrained('''xlm-mlm-en-2048''')
UpperCAmelCase_ = tokenizer.encode('''sequence builders''' , add_special_tokens=_lowercase)
UpperCAmelCase_ = tokenizer.encode('''multi-sequence build''' , add_special_tokens=_lowercase)
UpperCAmelCase_ = tokenizer.build_inputs_with_special_tokens(_lowercase)
UpperCAmelCase_ = tokenizer.build_inputs_with_special_tokens(_lowercase , _lowercase)
assert encoded_sentence == [0] + text + [1]
assert encoded_pair == [0] + text + [1] + text_a + [1]
| 344 |
import pytest
UpperCamelCase_ = "__dummy_dataset1__"
UpperCamelCase_ = "\nimport json\nimport os\n\nimport datasets\n\n\nREPO_URL = \"https://huggingface.co/datasets/albertvillanova/tests-raw-jsonl/resolve/main/\"\nURLS = {\"train\": REPO_URL + \"wikiann-bn-train.jsonl\", \"validation\": REPO_URL + \"wikiann-bn-validation.jsonl\"}\n\n\nclass __DummyDataset1__(datasets.GeneratorBasedBuilder):\n\n def _info(self):\n features = datasets.Features(\n {\n \"tokens\": datasets.Sequence(datasets.Value(\"string\")),\n \"ner_tags\": datasets.Sequence(\n datasets.features.ClassLabel(\n names=[\n \"O\",\n \"B-PER\",\n \"I-PER\",\n \"B-ORG\",\n \"I-ORG\",\n \"B-LOC\",\n \"I-LOC\",\n ]\n )\n ),\n \"langs\": datasets.Sequence(datasets.Value(\"string\")),\n \"spans\": datasets.Sequence(datasets.Value(\"string\")),\n }\n )\n return datasets.DatasetInfo(features=features)\n\n def _split_generators(self, dl_manager):\n dl_path = dl_manager.download(URLS)\n return [\n datasets.SplitGenerator(datasets.Split.TRAIN, gen_kwargs={\"filepath\": dl_path[\"train\"]}),\n datasets.SplitGenerator(datasets.Split.VALIDATION, gen_kwargs={\"filepath\": dl_path[\"validation\"]}),\n ]\n\n def _generate_examples(self, filepath):\n with open(filepath, \"r\", encoding=\"utf-8\") as f:\n for i, line in enumerate(f):\n yield i, json.loads(line)\n"
@pytest.fixture
def A ( ) -> str:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_NAME
@pytest.fixture
def A ( ) -> Any:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_CODE
@pytest.fixture
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = dataset_loading_script_name
UpperCAmelCase_ = tmp_path / '''datasets''' / script_name
script_dir.mkdir(parents=__UpperCAmelCase )
UpperCAmelCase_ = script_dir / f"{script_name}.py"
with open(__UpperCAmelCase , '''w''' ) as f:
f.write(__UpperCAmelCase )
return str(__UpperCAmelCase )
| 344 | 1 |
import functools
import operator
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"microsoft/unispeech-large-1500h-cv": (
"https://huggingface.co/microsoft/unispeech-large-1500h-cv/resolve/main/config.json"
),
# See all UniSpeech models at https://huggingface.co/models?filter=unispeech
}
class a_ ( _snake_case ):
UpperCamelCase__ : Union[str, Any] ="unispeech"
def __init__( self :Dict , _lowercase :Optional[int]=32 , _lowercase :List[Any]=768 , _lowercase :List[str]=12 , _lowercase :Tuple=12 , _lowercase :List[str]=3072 , _lowercase :Dict="gelu" , _lowercase :Any=0.1 , _lowercase :Optional[int]=0.1 , _lowercase :Dict=0.1 , _lowercase :List[str]=0.0 , _lowercase :Dict=0.0 , _lowercase :List[Any]=0.1 , _lowercase :str=0.1 , _lowercase :Optional[Any]=0.02 , _lowercase :List[str]=1E-5 , _lowercase :List[Any]="group" , _lowercase :str="gelu" , _lowercase :Any=(512, 512, 512, 512, 512, 512, 512) , _lowercase :Optional[Any]=(5, 2, 2, 2, 2, 2, 2) , _lowercase :Optional[Any]=(10, 3, 3, 3, 3, 2, 2) , _lowercase :Dict=False , _lowercase :str=128 , _lowercase :List[str]=16 , _lowercase :Union[str, Any]=False , _lowercase :Union[str, Any]=True , _lowercase :Any=0.05 , _lowercase :List[Any]=10 , _lowercase :Optional[Any]=2 , _lowercase :Tuple=0.0 , _lowercase :Union[str, Any]=10 , _lowercase :Union[str, Any]=0 , _lowercase :List[Any]=320 , _lowercase :Union[str, Any]=2 , _lowercase :int=0.1 , _lowercase :int=100 , _lowercase :List[str]=256 , _lowercase :Tuple=256 , _lowercase :List[str]=0.1 , _lowercase :List[Any]="mean" , _lowercase :Any=False , _lowercase :Union[str, Any]=False , _lowercase :str=256 , _lowercase :int=80 , _lowercase :List[str]=0 , _lowercase :Union[str, Any]=1 , _lowercase :str=2 , _lowercase :Dict=0.5 , **_lowercase :List[Any] , ) -> Dict:
super().__init__(**_lowercase , pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase)
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = feat_extract_norm
UpperCAmelCase_ = feat_extract_activation
UpperCAmelCase_ = list(_lowercase)
UpperCAmelCase_ = list(_lowercase)
UpperCAmelCase_ = list(_lowercase)
UpperCAmelCase_ = conv_bias
UpperCAmelCase_ = num_conv_pos_embeddings
UpperCAmelCase_ = num_conv_pos_embedding_groups
UpperCAmelCase_ = len(self.conv_dim)
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_dropout
UpperCAmelCase_ = attention_dropout
UpperCAmelCase_ = activation_dropout
UpperCAmelCase_ = feat_proj_dropout
UpperCAmelCase_ = final_dropout
UpperCAmelCase_ = layerdrop
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = num_ctc_classes
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = do_stable_layer_norm
UpperCAmelCase_ = use_weighted_layer_sum
UpperCAmelCase_ = classifier_proj_size
if (
(len(self.conv_stride) != self.num_feat_extract_layers)
or (len(self.conv_kernel) != self.num_feat_extract_layers)
or (len(self.conv_dim) != self.num_feat_extract_layers)
):
raise ValueError(
'''Configuration for convolutional layers is incorrect. It is required that `len(config.conv_dim)` =='''
''' `len(config.conv_stride)` == `len(config.conv_kernel)`, but is `len(config.conv_dim) ='''
f" {len(self.conv_dim)}`, `len(config.conv_stride) = {len(self.conv_stride)}`,"
f" `len(config.conv_kernel) = {len(self.conv_kernel)}`.")
# fine-tuning config parameters for SpecAugment: https://arxiv.org/abs/1904.08779
UpperCAmelCase_ = apply_spec_augment
UpperCAmelCase_ = mask_time_prob
UpperCAmelCase_ = mask_time_length
UpperCAmelCase_ = mask_time_min_masks
UpperCAmelCase_ = mask_feature_prob
UpperCAmelCase_ = mask_feature_length
UpperCAmelCase_ = mask_feature_min_masks
# parameters for pretraining with codevector quantized representations
UpperCAmelCase_ = num_codevectors_per_group
UpperCAmelCase_ = num_codevector_groups
UpperCAmelCase_ = contrastive_logits_temperature
UpperCAmelCase_ = feat_quantizer_dropout
UpperCAmelCase_ = num_negatives
UpperCAmelCase_ = codevector_dim
UpperCAmelCase_ = proj_codevector_dim
UpperCAmelCase_ = diversity_loss_weight
# ctc loss
UpperCAmelCase_ = ctc_loss_reduction
UpperCAmelCase_ = ctc_zero_infinity
# pretraining loss
UpperCAmelCase_ = replace_prob
@property
def __a ( self :Dict) -> Dict:
return functools.reduce(operator.mul , self.conv_stride , 1)
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"s-JoL/Open-Llama-V1": "https://huggingface.co/s-JoL/Open-Llama-V1/blob/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="open-llama"
def __init__( self :Union[str, Any] , _lowercase :List[Any]=100000 , _lowercase :Dict=4096 , _lowercase :List[Any]=11008 , _lowercase :Optional[int]=32 , _lowercase :Union[str, Any]=32 , _lowercase :List[str]="silu" , _lowercase :Union[str, Any]=2048 , _lowercase :Any=0.02 , _lowercase :Optional[Any]=1E-6 , _lowercase :str=True , _lowercase :str=0 , _lowercase :Any=1 , _lowercase :Optional[Any]=2 , _lowercase :str=False , _lowercase :Dict=True , _lowercase :Optional[Any]=0.1 , _lowercase :Tuple=0.1 , _lowercase :Dict=True , _lowercase :List[Any]=True , _lowercase :Dict=None , **_lowercase :Optional[int] , ) -> List[Any]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = rms_norm_eps
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = kwargs.pop(
'''use_memorry_efficient_attention''' , _lowercase)
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_dropout_prob
UpperCAmelCase_ = use_stable_embedding
UpperCAmelCase_ = shared_input_output_embedding
UpperCAmelCase_ = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , tie_word_embeddings=_lowercase , **_lowercase , )
def __a ( self :int) -> str:
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _lowercase) or len(self.rope_scaling) != 2:
raise ValueError(
'''`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, '''
f"got {self.rope_scaling}")
UpperCAmelCase_ = self.rope_scaling.get('''type''' , _lowercase)
UpperCAmelCase_ = self.rope_scaling.get('''factor''' , _lowercase)
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}")
if rope_scaling_factor is None or not isinstance(_lowercase , _lowercase) or rope_scaling_factor <= 1.0:
raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
| 344 | 1 |
import argparse
import torch
from transformers import (
UniSpeechSatConfig,
UniSpeechSatForAudioFrameClassification,
UniSpeechSatForSequenceClassification,
UniSpeechSatForXVector,
WavaVecaFeatureExtractor,
logging,
)
logging.set_verbosity_info()
UpperCamelCase_ = logging.get_logger(__name__)
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
UpperCAmelCase_ = UniSpeechSatForSequenceClassification.from_pretrained(__UpperCAmelCase , config=__UpperCAmelCase )
UpperCAmelCase_ = downstream_dict['''projector.weight''']
UpperCAmelCase_ = downstream_dict['''projector.bias''']
UpperCAmelCase_ = downstream_dict['''model.post_net.linear.weight''']
UpperCAmelCase_ = downstream_dict['''model.post_net.linear.bias''']
return model
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Any:
'''simple docstring'''
UpperCAmelCase_ = UniSpeechSatForAudioFrameClassification.from_pretrained(__UpperCAmelCase , config=__UpperCAmelCase )
UpperCAmelCase_ = downstream_dict['''model.linear.weight''']
UpperCAmelCase_ = downstream_dict['''model.linear.bias''']
return model
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Any:
'''simple docstring'''
UpperCAmelCase_ = UniSpeechSatForXVector.from_pretrained(__UpperCAmelCase , config=__UpperCAmelCase )
UpperCAmelCase_ = downstream_dict['''connector.weight''']
UpperCAmelCase_ = downstream_dict['''connector.bias''']
for i, kernel_size in enumerate(hf_config.tdnn_kernel ):
UpperCAmelCase_ = downstream_dict[
f"model.framelevel_feature_extractor.module.{i}.kernel.weight"
]
UpperCAmelCase_ = downstream_dict[f"model.framelevel_feature_extractor.module.{i}.kernel.bias"]
UpperCAmelCase_ = downstream_dict['''model.utterancelevel_feature_extractor.linear1.weight''']
UpperCAmelCase_ = downstream_dict['''model.utterancelevel_feature_extractor.linear1.bias''']
UpperCAmelCase_ = downstream_dict['''model.utterancelevel_feature_extractor.linear2.weight''']
UpperCAmelCase_ = downstream_dict['''model.utterancelevel_feature_extractor.linear2.bias''']
UpperCAmelCase_ = downstream_dict['''objective.W''']
return model
@torch.no_grad()
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
UpperCAmelCase_ = torch.load(__UpperCAmelCase , map_location='''cpu''' )
UpperCAmelCase_ = checkpoint['''Downstream''']
UpperCAmelCase_ = UniSpeechSatConfig.from_pretrained(__UpperCAmelCase )
UpperCAmelCase_ = WavaVecaFeatureExtractor.from_pretrained(
__UpperCAmelCase , return_attention_mask=__UpperCAmelCase , do_normalize=__UpperCAmelCase )
UpperCAmelCase_ = hf_config.architectures[0]
if arch.endswith('''ForSequenceClassification''' ):
UpperCAmelCase_ = convert_classification(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
elif arch.endswith('''ForAudioFrameClassification''' ):
UpperCAmelCase_ = convert_diarization(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
elif arch.endswith('''ForXVector''' ):
UpperCAmelCase_ = convert_xvector(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
else:
raise NotImplementedError(f"S3PRL weights conversion is not supported for {arch}" )
if hf_config.use_weighted_layer_sum:
UpperCAmelCase_ = checkpoint['''Featurizer''']['''weights''']
hf_feature_extractor.save_pretrained(__UpperCAmelCase )
hf_model.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
UpperCamelCase_ = argparse.ArgumentParser()
parser.add_argument(
"--base_model_name", default=None, type=str, help="Name of the huggingface pretrained base model."
)
parser.add_argument("--config_path", default=None, type=str, help="Path to the huggingface classifier config.")
parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to the s3prl checkpoint.")
parser.add_argument("--model_dump_path", default=None, type=str, help="Path to the final converted model.")
UpperCamelCase_ = parser.parse_args()
convert_saprl_checkpoint(args.base_model_name, args.config_path, args.checkpoint_path, args.model_dump_path)
| 344 |
import tempfile
import torch
from diffusers import (
DEISMultistepScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
UniPCMultistepScheduler,
)
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] =(DPMSolverSinglestepScheduler,)
UpperCamelCase__ : Tuple =(("num_inference_steps", 25),)
def __a ( self :List[Any] , **_lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
'''solver_order''': 2,
'''prediction_type''': '''epsilon''',
'''thresholding''': False,
'''sample_max_value''': 1.0,
'''algorithm_type''': '''dpmsolver++''',
'''solver_type''': '''midpoint''',
'''lambda_min_clipped''': -float('''inf'''),
'''variance_type''': None,
}
config.update(**_lowercase)
return config
def __a ( self :Union[str, Any] , _lowercase :List[Any]=0 , **_lowercase :Optional[int]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ , UpperCAmelCase_ = sample, sample
for t in range(_lowercase , time_step + scheduler.config.solver_order + 1):
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Union[str, Any]) -> List[Any]:
pass
def __a ( self :Optional[Any] , _lowercase :str=0 , **_lowercase :Union[str, Any]) -> Dict:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Dict , _lowercase :Union[str, Any]=None , **_lowercase :List[Any]) -> int:
if scheduler is None:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :int) -> Tuple:
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = 50
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
# make sure that the first t is uneven
for i, t in enumerate(scheduler.timesteps[3:]):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_574) < 1E-3
def __a ( self :List[Any]) -> List[Any]:
for timesteps in [25, 50, 100, 999, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :int) -> Optional[Any]:
# make sure that iterating over schedulers with same config names gives same results
# for defaults
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
UpperCAmelCase_ = DEISMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = UniPCMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverSinglestepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Tuple) -> int:
self.check_over_configs(thresholding=_lowercase)
for order in [1, 2, 3]:
for solver_type in ["midpoint", "heun"]:
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
thresholding=_lowercase , prediction_type=_lowercase , sample_max_value=_lowercase , algorithm_type='''dpmsolver++''' , solver_order=_lowercase , solver_type=_lowercase , )
def __a ( self :List[Any]) -> Any:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Optional[int]:
for algorithm_type in ["dpmsolver", "dpmsolver++"]:
for solver_type in ["midpoint", "heun"]:
for order in [1, 2, 3]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
UpperCAmelCase_ = self.full_loop(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
assert not torch.isnan(_lowercase).any(), "Samples have nan numbers"
def __a ( self :Tuple) -> int:
self.check_over_configs(lower_order_final=_lowercase)
self.check_over_configs(lower_order_final=_lowercase)
def __a ( self :Tuple) -> Optional[Any]:
self.check_over_configs(lambda_min_clipped=-float('''inf'''))
self.check_over_configs(lambda_min_clipped=-5.1)
def __a ( self :Any) -> List[str]:
self.check_over_configs(variance_type=_lowercase)
self.check_over_configs(variance_type='''learned_range''')
def __a ( self :Any) -> Dict:
for num_inference_steps in [1, 2, 3, 5, 10, 50, 100, 999, 1000]:
self.check_over_forward(num_inference_steps=_lowercase , time_step=0)
def __a ( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Any) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop(use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_248) < 1E-3
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.1_453) < 1E-3
def __a ( self :List[Any]) -> Dict:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''' , use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.0_649) < 1E-3
def __a ( self :Any) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(thresholding=_lowercase , dynamic_thresholding_ratio=0)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter.half()
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
assert sample.dtype == torch.floataa
| 344 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"s-JoL/Open-Llama-V1": "https://huggingface.co/s-JoL/Open-Llama-V1/blob/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="open-llama"
def __init__( self :Union[str, Any] , _lowercase :List[Any]=100000 , _lowercase :Dict=4096 , _lowercase :List[Any]=11008 , _lowercase :Optional[int]=32 , _lowercase :Union[str, Any]=32 , _lowercase :List[str]="silu" , _lowercase :Union[str, Any]=2048 , _lowercase :Any=0.02 , _lowercase :Optional[Any]=1E-6 , _lowercase :str=True , _lowercase :str=0 , _lowercase :Any=1 , _lowercase :Optional[Any]=2 , _lowercase :str=False , _lowercase :Dict=True , _lowercase :Optional[Any]=0.1 , _lowercase :Tuple=0.1 , _lowercase :Dict=True , _lowercase :List[Any]=True , _lowercase :Dict=None , **_lowercase :Optional[int] , ) -> List[Any]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = rms_norm_eps
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = kwargs.pop(
'''use_memorry_efficient_attention''' , _lowercase)
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_dropout_prob
UpperCAmelCase_ = use_stable_embedding
UpperCAmelCase_ = shared_input_output_embedding
UpperCAmelCase_ = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , tie_word_embeddings=_lowercase , **_lowercase , )
def __a ( self :int) -> str:
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _lowercase) or len(self.rope_scaling) != 2:
raise ValueError(
'''`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, '''
f"got {self.rope_scaling}")
UpperCAmelCase_ = self.rope_scaling.get('''type''' , _lowercase)
UpperCAmelCase_ = self.rope_scaling.get('''factor''' , _lowercase)
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}")
if rope_scaling_factor is None or not isinstance(_lowercase , _lowercase) or rope_scaling_factor <= 1.0:
raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
| 344 |
import inspect
import unittest
import torch
import torch.nn as nn
from accelerate.hooks import (
AlignDevicesHook,
ModelHook,
SequentialHook,
add_hook_to_module,
attach_align_device_hook,
remove_hook_from_module,
remove_hook_from_submodules,
)
from accelerate.test_utils import require_multi_gpu
class a_ ( nn.Module ):
def __init__( self :Optional[Any]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(3 , 4)
UpperCAmelCase_ = nn.BatchNormad(4)
UpperCAmelCase_ = nn.Linear(4 , 5)
def __a ( self :Dict , _lowercase :int) -> str:
return self.lineara(self.batchnorm(self.lineara(_lowercase)))
class a_ ( _snake_case ):
def __a ( self :Tuple , _lowercase :Optional[int] , *_lowercase :Union[str, Any] , **_lowercase :Any) -> Optional[Any]:
return (args[0] + 1,) + args[1:], kwargs
class a_ ( _snake_case ):
def __a ( self :Union[str, Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
return output + 1
class a_ ( unittest.TestCase ):
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
self.assertEqual(test_model._hf_hook , _lowercase)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[Any]) -> Any:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
add_hook_to_module(_lowercase , _lowercase , append=_lowercase)
self.assertEqual(isinstance(test_model._hf_hook , _lowercase) , _lowercase)
self.assertEqual(len(test_model._hf_hook.hooks) , 2)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[int]) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(x + 1)
UpperCAmelCase_ = test_model(x + 2)
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PreForwardHook() , PreForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , _lowercase , atol=1E-5)
def __a ( self :List[str]) -> int:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PostForwardHook() , PostForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , output + 2 , atol=1E-5)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1))
self.assertTrue(outputa.requires_grad)
UpperCAmelCase_ = True
UpperCAmelCase_ = test_model(_lowercase)
self.assertFalse(outputa.requires_grad)
@require_multi_gpu
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.batchnorm , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=1))
self.assertEqual(model.lineara.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.running_mean.device , torch.device(0))
self.assertEqual(model.lineara.weight.device , torch.device(1))
# We can still make a forward pass. The input does not need to be on any particular device
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(1))
# We can add a general hook to put back output on same device as input.
add_hook_to_module(_lowercase , AlignDevicesHook(io_same_device=_lowercase))
UpperCAmelCase_ = torch.randn(2 , 3).to(0)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(0))
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = {'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''', '''offload''': True}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(hook_kwargs['''execution_device'''])
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
UpperCAmelCase_ = {
'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''',
'''offload''': True,
'''offload_buffers''': True,
}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase)
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase , offload_buffers=_lowercase)
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict())
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict() , offload_buffers=_lowercase , )
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
| 344 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"funnel-transformer/small": "https://huggingface.co/funnel-transformer/small/resolve/main/config.json",
"funnel-transformer/small-base": "https://huggingface.co/funnel-transformer/small-base/resolve/main/config.json",
"funnel-transformer/medium": "https://huggingface.co/funnel-transformer/medium/resolve/main/config.json",
"funnel-transformer/medium-base": "https://huggingface.co/funnel-transformer/medium-base/resolve/main/config.json",
"funnel-transformer/intermediate": (
"https://huggingface.co/funnel-transformer/intermediate/resolve/main/config.json"
),
"funnel-transformer/intermediate-base": (
"https://huggingface.co/funnel-transformer/intermediate-base/resolve/main/config.json"
),
"funnel-transformer/large": "https://huggingface.co/funnel-transformer/large/resolve/main/config.json",
"funnel-transformer/large-base": "https://huggingface.co/funnel-transformer/large-base/resolve/main/config.json",
"funnel-transformer/xlarge": "https://huggingface.co/funnel-transformer/xlarge/resolve/main/config.json",
"funnel-transformer/xlarge-base": "https://huggingface.co/funnel-transformer/xlarge-base/resolve/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="funnel"
UpperCamelCase__ : Optional[Any] ={
"hidden_size": "d_model",
"num_attention_heads": "n_head",
}
def __init__( self :Union[str, Any] , _lowercase :int=30522 , _lowercase :str=[4, 4, 4] , _lowercase :str=None , _lowercase :Optional[int]=2 , _lowercase :Union[str, Any]=768 , _lowercase :List[str]=12 , _lowercase :Optional[int]=64 , _lowercase :int=3072 , _lowercase :Dict="gelu_new" , _lowercase :Optional[Any]=0.1 , _lowercase :Union[str, Any]=0.1 , _lowercase :str=0.0 , _lowercase :Dict=0.1 , _lowercase :str=None , _lowercase :Tuple=1E-9 , _lowercase :List[Any]="mean" , _lowercase :Optional[int]="relative_shift" , _lowercase :List[str]=True , _lowercase :Optional[Any]=True , _lowercase :Optional[int]=True , **_lowercase :str , ) -> List[str]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = block_sizes
UpperCAmelCase_ = [1] * len(_lowercase) if block_repeats is None else block_repeats
assert len(_lowercase) == len(
self.block_repeats), "`block_sizes` and `block_repeats` should have the same length."
UpperCAmelCase_ = num_decoder_layers
UpperCAmelCase_ = d_model
UpperCAmelCase_ = n_head
UpperCAmelCase_ = d_head
UpperCAmelCase_ = d_inner
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = hidden_dropout
UpperCAmelCase_ = attention_dropout
UpperCAmelCase_ = activation_dropout
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = initializer_std
UpperCAmelCase_ = layer_norm_eps
assert pooling_type in [
"mean",
"max",
], f"Got {pooling_type} for `pooling_type` but only 'mean' and 'max' are supported."
UpperCAmelCase_ = pooling_type
assert attention_type in [
"relative_shift",
"factorized",
], f"Got {attention_type} for `attention_type` but only 'relative_shift' and 'factorized' are supported."
UpperCAmelCase_ = attention_type
UpperCAmelCase_ = separate_cls
UpperCAmelCase_ = truncate_seq
UpperCAmelCase_ = pool_q_only
super().__init__(**_lowercase)
@property
def __a ( self :int) -> Any:
return sum(self.block_sizes)
@num_hidden_layers.setter
def __a ( self :int , _lowercase :str) -> str:
raise NotImplementedError(
'''This model does not support the setting of `num_hidden_layers`. Please set `block_sizes`.''')
@property
def __a ( self :Optional[int]) -> Dict:
return len(self.block_sizes)
@num_blocks.setter
def __a ( self :Union[str, Any] , _lowercase :int) -> Optional[Any]:
raise NotImplementedError('''This model does not support the setting of `num_blocks`. Please set `block_sizes`.''')
| 344 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import (
AutoProcessor,
BertTokenizerFast,
BlipImageProcessor,
GPTaTokenizer,
InstructBlipProcessor,
PreTrainedTokenizerFast,
)
@require_vision
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = BlipImageProcessor()
UpperCAmelCase_ = GPTaTokenizer.from_pretrained('''hf-internal-testing/tiny-random-GPT2Model''')
UpperCAmelCase_ = BertTokenizerFast.from_pretrained('''hf-internal-testing/tiny-random-bert''')
UpperCAmelCase_ = InstructBlipProcessor(_lowercase , _lowercase , _lowercase)
processor.save_pretrained(self.tmpdirname)
def __a ( self :List[Any] , **_lowercase :Dict) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).tokenizer
def __a ( self :Optional[Any] , **_lowercase :Optional[Any]) -> Optional[int]:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).image_processor
def __a ( self :Dict , **_lowercase :Tuple) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).qformer_tokenizer
def __a ( self :Optional[int]) -> str:
shutil.rmtree(self.tmpdirname)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta)]
UpperCAmelCase_ = [Image.fromarray(np.moveaxis(_lowercase , 0 , -1)) for x in image_inputs]
return image_inputs
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() , qformer_tokenizer=self.get_qformer_tokenizer() , )
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = self.get_image_processor(do_normalize=_lowercase , padding_value=1.0)
UpperCAmelCase_ = InstructBlipProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowercase , padding_value=1.0)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
self.assertIsInstance(processor.tokenizer , _lowercase)
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string())
self.assertIsInstance(processor.image_processor , _lowercase)
self.assertIsInstance(processor.qformer_tokenizer , _lowercase)
def __a ( self :Dict) -> Any:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = image_processor(_lowercase , return_tensors='''np''')
UpperCAmelCase_ = processor(images=_lowercase , return_tensors='''np''')
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1E-2)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = processor(text=_lowercase)
UpperCAmelCase_ = tokenizer(_lowercase , return_token_type_ids=_lowercase)
UpperCAmelCase_ = qformer_tokenizer(_lowercase , return_token_type_ids=_lowercase)
for key in encoded_tokens.keys():
self.assertListEqual(encoded_tokens[key] , encoded_processor[key])
for key in encoded_tokens_qformer.keys():
self.assertListEqual(encoded_tokens_qformer[key] , encoded_processor['''qformer_''' + key])
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
# test if it raises when no input is passed
with pytest.raises(_lowercase):
processor()
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
UpperCAmelCase_ = processor.batch_decode(_lowercase)
UpperCAmelCase_ = tokenizer.batch_decode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :str) -> int:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
| 344 | 1 |
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class a_ ( unittest.TestCase ):
@slow
def __a ( self :Any) -> Any:
UpperCAmelCase_ = XLMRobertaModel.from_pretrained('''xlm-roberta-base''')
UpperCAmelCase_ = torch.tensor([[0, 581, 10269, 83, 99942, 136, 60742, 23, 70, 80583, 18276, 2]])
# The dog is cute and lives in the garden house
UpperCAmelCase_ = torch.Size((1, 12, 768)) # batch_size, sequence_length, embedding_vector_dim
UpperCAmelCase_ = torch.tensor(
[[-0.0_101, 0.1_218, -0.0_803, 0.0_801, 0.1_327, 0.0_776, -0.1_215, 0.2_383, 0.3_338, 0.3_106, 0.0_300, 0.0_252]])
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
UpperCAmelCase_ = model(_lowercase)['''last_hidden_state'''].detach()
self.assertEqual(output.shape , _lowercase)
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] , _lowercase , atol=1E-3))
@slow
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = XLMRobertaModel.from_pretrained('''xlm-roberta-large''')
UpperCAmelCase_ = torch.tensor([[0, 581, 10269, 83, 99942, 136, 60742, 23, 70, 80583, 18276, 2]])
# The dog is cute and lives in the garden house
UpperCAmelCase_ = torch.Size((1, 12, 1024)) # batch_size, sequence_length, embedding_vector_dim
UpperCAmelCase_ = torch.tensor(
[[-0.0_699, -0.0_318, 0.0_705, -0.1_241, 0.0_999, -0.0_520, 0.1_004, -0.1_838, -0.4_704, 0.1_437, 0.0_821, 0.0_126]])
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.large')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
UpperCAmelCase_ = model(_lowercase)['''last_hidden_state'''].detach()
self.assertEqual(output.shape , _lowercase)
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] , _lowercase , atol=1E-3))
| 344 |
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"facebook/levit-128S": "https://huggingface.co/facebook/levit-128S/resolve/main/config.json",
# See all LeViT models at https://huggingface.co/models?filter=levit
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[int] ="levit"
def __init__( self :List[str] , _lowercase :List[Any]=224 , _lowercase :str=3 , _lowercase :Optional[int]=3 , _lowercase :str=2 , _lowercase :List[Any]=1 , _lowercase :str=16 , _lowercase :Dict=[128, 256, 384] , _lowercase :Union[str, Any]=[4, 8, 12] , _lowercase :Tuple=[4, 4, 4] , _lowercase :Dict=[16, 16, 16] , _lowercase :Any=0 , _lowercase :Dict=[2, 2, 2] , _lowercase :Any=[2, 2, 2] , _lowercase :Tuple=0.02 , **_lowercase :Union[str, Any] , ) -> Optional[Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = kernel_size
UpperCAmelCase_ = stride
UpperCAmelCase_ = padding
UpperCAmelCase_ = hidden_sizes
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = depths
UpperCAmelCase_ = key_dim
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = attention_ratio
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = [
['''Subsample''', key_dim[0], hidden_sizes[0] // key_dim[0], 4, 2, 2],
['''Subsample''', key_dim[0], hidden_sizes[1] // key_dim[0], 4, 2, 2],
]
class a_ ( _snake_case ):
UpperCamelCase__ : Union[str, Any] =version.parse("1.11" )
@property
def __a ( self :Any) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
])
@property
def __a ( self :List[Any]) -> float:
return 1E-4
| 344 | 1 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_sentencepiece_available
UpperCamelCase_ = {}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["BartphoTokenizer"]
if TYPE_CHECKING:
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_bartpho import BartphoTokenizer
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=None , __UpperCAmelCase="no" , __UpperCAmelCase="29500" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = False
UpperCAmelCase_ = False
if any(key.startswith('''KAGGLE''' ) for key in os.environ.keys() ):
UpperCAmelCase_ = True
elif "IPython" in sys.modules:
UpperCAmelCase_ = '''google.colab''' in str(sys.modules['''IPython'''].get_ipython() )
try:
UpperCAmelCase_ = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
f"Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}." )
if (in_colab or in_kaggle) and (os.environ.get('''TPU_NAME''' , __UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '''
'''your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if num_processes is None:
UpperCAmelCase_ = 8
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''TPU''' )
print(f"Launching a training on {num_processes} TPU cores." )
xmp.spawn(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on one CPU.''' )
function(*__UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'''You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.''' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '''
'''inside your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if torch.cuda.is_initialized():
raise ValueError(
'''To launch a multi-GPU training from your notebook, you need to avoid running any instruction '''
'''using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '''
'''function.''' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port=__UpperCAmelCase , mixed_precision=__UpperCAmelCase ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''MULTI_GPU''' )
print(f"Launching training on {num_processes} GPUs." )
try:
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'''CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '''
'''This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '''
'''Please review your imports and test them when running the `notebook_launcher()` to identify '''
'''which one is problematic.''' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
UpperCAmelCase_ = '''1'''
print('''Launching training on MPS.''' )
elif torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on CPU.''' )
function(*__UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=2 ) -> Optional[Any]:
'''simple docstring'''
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port='''29500''' , accelerate_mixed_precision='''no''' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='''yes''' , ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , debug=__UpperCAmelCase )
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
| 344 | 1 |
import argparse
import json
from collections import OrderedDict
from pathlib import Path
import requests
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import (
ConditionalDetrConfig,
ConditionalDetrForObjectDetection,
ConditionalDetrForSegmentation,
ConditionalDetrImageProcessor,
)
from transformers.utils import logging
logging.set_verbosity_info()
UpperCamelCase_ = logging.get_logger(__name__)
# here we list all keys to be renamed (original name on the left, our name on the right)
UpperCamelCase_ = []
for i in range(6):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append(
(f"transformer.encoder.layers.{i}.self_attn.out_proj.weight", f"encoder.layers.{i}.self_attn.out_proj.weight")
)
rename_keys.append(
(f"transformer.encoder.layers.{i}.self_attn.out_proj.bias", f"encoder.layers.{i}.self_attn.out_proj.bias")
)
rename_keys.append((f"transformer.encoder.layers.{i}.linear1.weight", f"encoder.layers.{i}.fc1.weight"))
rename_keys.append((f"transformer.encoder.layers.{i}.linear1.bias", f"encoder.layers.{i}.fc1.bias"))
rename_keys.append((f"transformer.encoder.layers.{i}.linear2.weight", f"encoder.layers.{i}.fc2.weight"))
rename_keys.append((f"transformer.encoder.layers.{i}.linear2.bias", f"encoder.layers.{i}.fc2.bias"))
rename_keys.append(
(f"transformer.encoder.layers.{i}.norm1.weight", f"encoder.layers.{i}.self_attn_layer_norm.weight")
)
rename_keys.append((f"transformer.encoder.layers.{i}.norm1.bias", f"encoder.layers.{i}.self_attn_layer_norm.bias"))
rename_keys.append((f"transformer.encoder.layers.{i}.norm2.weight", f"encoder.layers.{i}.final_layer_norm.weight"))
rename_keys.append((f"transformer.encoder.layers.{i}.norm2.bias", f"encoder.layers.{i}.final_layer_norm.bias"))
# decoder layers: 2 times output projection, 2 feedforward neural networks and 3 layernorms
rename_keys.append(
(f"transformer.decoder.layers.{i}.self_attn.out_proj.weight", f"decoder.layers.{i}.self_attn.out_proj.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.self_attn.out_proj.bias", f"decoder.layers.{i}.self_attn.out_proj.bias")
)
rename_keys.append(
(
f"transformer.decoder.layers.{i}.cross_attn.out_proj.weight",
f"decoder.layers.{i}.encoder_attn.out_proj.weight",
)
)
rename_keys.append(
(
f"transformer.decoder.layers.{i}.cross_attn.out_proj.bias",
f"decoder.layers.{i}.encoder_attn.out_proj.bias",
)
)
rename_keys.append((f"transformer.decoder.layers.{i}.linear1.weight", f"decoder.layers.{i}.fc1.weight"))
rename_keys.append((f"transformer.decoder.layers.{i}.linear1.bias", f"decoder.layers.{i}.fc1.bias"))
rename_keys.append((f"transformer.decoder.layers.{i}.linear2.weight", f"decoder.layers.{i}.fc2.weight"))
rename_keys.append((f"transformer.decoder.layers.{i}.linear2.bias", f"decoder.layers.{i}.fc2.bias"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.norm1.weight", f"decoder.layers.{i}.self_attn_layer_norm.weight")
)
rename_keys.append((f"transformer.decoder.layers.{i}.norm1.bias", f"decoder.layers.{i}.self_attn_layer_norm.bias"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.norm2.weight", f"decoder.layers.{i}.encoder_attn_layer_norm.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.norm2.bias", f"decoder.layers.{i}.encoder_attn_layer_norm.bias")
)
rename_keys.append((f"transformer.decoder.layers.{i}.norm3.weight", f"decoder.layers.{i}.final_layer_norm.weight"))
rename_keys.append((f"transformer.decoder.layers.{i}.norm3.bias", f"decoder.layers.{i}.final_layer_norm.bias"))
# q, k, v projections in self/cross-attention in decoder for conditional DETR
rename_keys.append(
(f"transformer.decoder.layers.{i}.sa_qcontent_proj.weight", f"decoder.layers.{i}.sa_qcontent_proj.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.sa_kcontent_proj.weight", f"decoder.layers.{i}.sa_kcontent_proj.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.sa_qpos_proj.weight", f"decoder.layers.{i}.sa_qpos_proj.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.sa_kpos_proj.weight", f"decoder.layers.{i}.sa_kpos_proj.weight")
)
rename_keys.append((f"transformer.decoder.layers.{i}.sa_v_proj.weight", f"decoder.layers.{i}.sa_v_proj.weight"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_qcontent_proj.weight", f"decoder.layers.{i}.ca_qcontent_proj.weight")
)
# rename_keys.append((f"transformer.decoder.layers.{i}.ca_qpos_proj.weight", f"decoder.layers.{i}.ca_qpos_proj.weight"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_kcontent_proj.weight", f"decoder.layers.{i}.ca_kcontent_proj.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_kpos_proj.weight", f"decoder.layers.{i}.ca_kpos_proj.weight")
)
rename_keys.append((f"transformer.decoder.layers.{i}.ca_v_proj.weight", f"decoder.layers.{i}.ca_v_proj.weight"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_qpos_sine_proj.weight", f"decoder.layers.{i}.ca_qpos_sine_proj.weight")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.sa_qcontent_proj.bias", f"decoder.layers.{i}.sa_qcontent_proj.bias")
)
rename_keys.append(
(f"transformer.decoder.layers.{i}.sa_kcontent_proj.bias", f"decoder.layers.{i}.sa_kcontent_proj.bias")
)
rename_keys.append((f"transformer.decoder.layers.{i}.sa_qpos_proj.bias", f"decoder.layers.{i}.sa_qpos_proj.bias"))
rename_keys.append((f"transformer.decoder.layers.{i}.sa_kpos_proj.bias", f"decoder.layers.{i}.sa_kpos_proj.bias"))
rename_keys.append((f"transformer.decoder.layers.{i}.sa_v_proj.bias", f"decoder.layers.{i}.sa_v_proj.bias"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_qcontent_proj.bias", f"decoder.layers.{i}.ca_qcontent_proj.bias")
)
# rename_keys.append((f"transformer.decoder.layers.{i}.ca_qpos_proj.bias", f"decoder.layers.{i}.ca_qpos_proj.bias"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_kcontent_proj.bias", f"decoder.layers.{i}.ca_kcontent_proj.bias")
)
rename_keys.append((f"transformer.decoder.layers.{i}.ca_kpos_proj.bias", f"decoder.layers.{i}.ca_kpos_proj.bias"))
rename_keys.append((f"transformer.decoder.layers.{i}.ca_v_proj.bias", f"decoder.layers.{i}.ca_v_proj.bias"))
rename_keys.append(
(f"transformer.decoder.layers.{i}.ca_qpos_sine_proj.bias", f"decoder.layers.{i}.ca_qpos_sine_proj.bias")
)
# convolutional projection + query embeddings + layernorm of decoder + class and bounding box heads
# for conditional DETR, also convert reference point head and query scale MLP
rename_keys.extend(
[
("input_proj.weight", "input_projection.weight"),
("input_proj.bias", "input_projection.bias"),
("query_embed.weight", "query_position_embeddings.weight"),
("transformer.decoder.norm.weight", "decoder.layernorm.weight"),
("transformer.decoder.norm.bias", "decoder.layernorm.bias"),
("class_embed.weight", "class_labels_classifier.weight"),
("class_embed.bias", "class_labels_classifier.bias"),
("bbox_embed.layers.0.weight", "bbox_predictor.layers.0.weight"),
("bbox_embed.layers.0.bias", "bbox_predictor.layers.0.bias"),
("bbox_embed.layers.1.weight", "bbox_predictor.layers.1.weight"),
("bbox_embed.layers.1.bias", "bbox_predictor.layers.1.bias"),
("bbox_embed.layers.2.weight", "bbox_predictor.layers.2.weight"),
("bbox_embed.layers.2.bias", "bbox_predictor.layers.2.bias"),
("transformer.decoder.ref_point_head.layers.0.weight", "decoder.ref_point_head.layers.0.weight"),
("transformer.decoder.ref_point_head.layers.0.bias", "decoder.ref_point_head.layers.0.bias"),
("transformer.decoder.ref_point_head.layers.1.weight", "decoder.ref_point_head.layers.1.weight"),
("transformer.decoder.ref_point_head.layers.1.bias", "decoder.ref_point_head.layers.1.bias"),
("transformer.decoder.query_scale.layers.0.weight", "decoder.query_scale.layers.0.weight"),
("transformer.decoder.query_scale.layers.0.bias", "decoder.query_scale.layers.0.bias"),
("transformer.decoder.query_scale.layers.1.weight", "decoder.query_scale.layers.1.weight"),
("transformer.decoder.query_scale.layers.1.bias", "decoder.query_scale.layers.1.bias"),
("transformer.decoder.layers.0.ca_qpos_proj.weight", "decoder.layers.0.ca_qpos_proj.weight"),
("transformer.decoder.layers.0.ca_qpos_proj.bias", "decoder.layers.0.ca_qpos_proj.bias"),
]
)
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
UpperCAmelCase_ = state_dict.pop(__UpperCAmelCase )
UpperCAmelCase_ = val
def A ( __UpperCAmelCase ) -> Tuple:
'''simple docstring'''
UpperCAmelCase_ = OrderedDict()
for key, value in state_dict.items():
if "backbone.0.body" in key:
UpperCAmelCase_ = key.replace('''backbone.0.body''' , '''backbone.conv_encoder.model''' )
UpperCAmelCase_ = value
else:
UpperCAmelCase_ = value
return new_state_dict
def A ( __UpperCAmelCase , __UpperCAmelCase=False ) -> Tuple:
'''simple docstring'''
UpperCAmelCase_ = ''''''
if is_panoptic:
UpperCAmelCase_ = '''conditional_detr.'''
# first: transformer encoder
for i in range(6 ):
# read in weights + bias of input projection layer (in PyTorch's MultiHeadAttention, this is a single matrix + bias)
UpperCAmelCase_ = state_dict.pop(f"{prefix}transformer.encoder.layers.{i}.self_attn.in_proj_weight" )
UpperCAmelCase_ = state_dict.pop(f"{prefix}transformer.encoder.layers.{i}.self_attn.in_proj_bias" )
# next, add query, keys and values (in that order) to the state dict
UpperCAmelCase_ = in_proj_weight[:256, :]
UpperCAmelCase_ = in_proj_bias[:256]
UpperCAmelCase_ = in_proj_weight[256:512, :]
UpperCAmelCase_ = in_proj_bias[256:512]
UpperCAmelCase_ = in_proj_weight[-256:, :]
UpperCAmelCase_ = in_proj_bias[-256:]
def A ( ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
UpperCAmelCase_ = Image.open(requests.get(__UpperCAmelCase , stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str:
'''simple docstring'''
UpperCAmelCase_ = ConditionalDetrConfig()
# set backbone and dilation attributes
if "resnet101" in model_name:
UpperCAmelCase_ = '''resnet101'''
if "dc5" in model_name:
UpperCAmelCase_ = True
UpperCAmelCase_ = '''panoptic''' in model_name
if is_panoptic:
UpperCAmelCase_ = 250
else:
UpperCAmelCase_ = 91
UpperCAmelCase_ = '''huggingface/label-files'''
UpperCAmelCase_ = '''coco-detection-id2label.json'''
UpperCAmelCase_ = json.load(open(hf_hub_download(__UpperCAmelCase , __UpperCAmelCase , repo_type='''dataset''' ) , '''r''' ) )
UpperCAmelCase_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
UpperCAmelCase_ = idalabel
UpperCAmelCase_ = {v: k for k, v in idalabel.items()}
# load image processor
UpperCAmelCase_ = '''coco_panoptic''' if is_panoptic else '''coco_detection'''
UpperCAmelCase_ = ConditionalDetrImageProcessor(format=__UpperCAmelCase )
# prepare image
UpperCAmelCase_ = prepare_img()
UpperCAmelCase_ = image_processor(images=__UpperCAmelCase , return_tensors='''pt''' )
UpperCAmelCase_ = encoding['''pixel_values''']
logger.info(f"Converting model {model_name}..." )
# load original model from torch hub
UpperCAmelCase_ = torch.hub.load('''DeppMeng/ConditionalDETR''' , __UpperCAmelCase , pretrained=__UpperCAmelCase ).eval()
UpperCAmelCase_ = conditional_detr.state_dict()
# rename keys
for src, dest in rename_keys:
if is_panoptic:
UpperCAmelCase_ = '''conditional_detr.''' + src
rename_key(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase )
UpperCAmelCase_ = rename_backbone_keys(__UpperCAmelCase )
# query, key and value matrices need special treatment
read_in_q_k_v(__UpperCAmelCase , is_panoptic=__UpperCAmelCase )
# important: we need to prepend a prefix to each of the base model keys as the head models use different attributes for them
UpperCAmelCase_ = '''conditional_detr.model.''' if is_panoptic else '''model.'''
for key in state_dict.copy().keys():
if is_panoptic:
if (
key.startswith('''conditional_detr''' )
and not key.startswith('''class_labels_classifier''' )
and not key.startswith('''bbox_predictor''' )
):
UpperCAmelCase_ = state_dict.pop(__UpperCAmelCase )
UpperCAmelCase_ = val
elif "class_labels_classifier" in key or "bbox_predictor" in key:
UpperCAmelCase_ = state_dict.pop(__UpperCAmelCase )
UpperCAmelCase_ = val
elif key.startswith('''bbox_attention''' ) or key.startswith('''mask_head''' ):
continue
else:
UpperCAmelCase_ = state_dict.pop(__UpperCAmelCase )
UpperCAmelCase_ = val
else:
if not key.startswith('''class_labels_classifier''' ) and not key.startswith('''bbox_predictor''' ):
UpperCAmelCase_ = state_dict.pop(__UpperCAmelCase )
UpperCAmelCase_ = val
# finally, create HuggingFace model and load state dict
UpperCAmelCase_ = ConditionalDetrForSegmentation(__UpperCAmelCase ) if is_panoptic else ConditionalDetrForObjectDetection(__UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
model.push_to_hub(repo_id=__UpperCAmelCase , organization='''DepuMeng''' , commit_message='''Add model''' )
# verify our conversion
UpperCAmelCase_ = conditional_detr(__UpperCAmelCase )
UpperCAmelCase_ = model(__UpperCAmelCase )
assert torch.allclose(outputs.logits , original_outputs['''pred_logits'''] , atol=1e-4 )
assert torch.allclose(outputs.pred_boxes , original_outputs['''pred_boxes'''] , atol=1e-4 )
if is_panoptic:
assert torch.allclose(outputs.pred_masks , original_outputs['''pred_masks'''] , atol=1e-4 )
# Save model and image processor
logger.info(f"Saving PyTorch model and image processor to {pytorch_dump_folder_path}..." )
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
model.save_pretrained(__UpperCAmelCase )
image_processor.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
UpperCamelCase_ = argparse.ArgumentParser()
parser.add_argument(
"--model_name",
default="conditional_detr_resnet50",
type=str,
help="Name of the CONDITIONAL_DETR model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the folder to output PyTorch model."
)
UpperCamelCase_ = parser.parse_args()
convert_conditional_detr_checkpoint(args.model_name, args.pytorch_dump_folder_path)
| 344 |
import math
import os
import re
import sys
import unittest
from pathlib import Path
from typing import Tuple
from unittest.mock import patch
from parameterized import parameterized
from transformers.testing_utils import (
CaptureStderr,
ExtendSysPath,
TestCasePlus,
execute_subprocess_async,
get_gpu_count,
get_torch_dist_unique_port,
require_apex,
require_bitsandbytes,
require_fairscale,
require_torch,
require_torch_gpu,
require_torch_multi_gpu,
require_torch_non_multi_gpu,
slow,
)
from transformers.trainer_callback import TrainerState
from transformers.trainer_utils import set_seed
UpperCamelCase_ = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f"{bindir}/../../examples/pytorch/translation"):
from run_translation import main # noqa
set_seed(42)
UpperCamelCase_ = "sshleifer/student_marian_en_ro_6_1"
UpperCamelCase_ = "sshleifer/tiny-mbart"
@require_torch
class a_ ( _snake_case ):
def __a ( self :str , _lowercase :Any=False , _lowercase :Tuple=None , _lowercase :Dict=True , _lowercase :Tuple=True , _lowercase :List[Any]=True , _lowercase :List[str]=True , ) -> int:
UpperCAmelCase_ = self.run_trainer(
eval_steps=1 , max_len=12 , model_name=_lowercase , num_train_epochs=1 , distributed=_lowercase , extra_args_str=_lowercase , predict_with_generate=_lowercase , do_train=_lowercase , do_eval=_lowercase , do_predict=_lowercase , )
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
if not do_eval:
return
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
if predict_with_generate:
assert "eval_bleu" in first_step_stats
UpperCAmelCase_ = eval_metrics[-1]
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
assert not math.isnan(float(last_step_stats['''eval_loss'''])), "eval_loss must not be `nan`"
@require_torch_non_multi_gpu
def __a ( self :Dict) -> str:
self.run_seqaseq_quick()
@require_torch_multi_gpu
def __a ( self :Any) -> int:
self.run_seqaseq_quick(distributed=_lowercase)
@require_torch_multi_gpu
def __a ( self :int) -> Any:
self.run_seqaseq_quick(distributed=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> List[str]:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple --fp16''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Union[str, Any]) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :int) -> Any:
self.run_seqaseq_quick(
distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=_lowercase)
@require_apex
@require_torch_gpu
def __a ( self :Tuple) -> str:
# XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same
# program and it breaks other tests that run from the same pytest worker, therefore until this is
# sorted out it must be run only in an external program, that is distributed=True in this
# test and only under one or more gpus - if we want cpu will need to make a special test
#
# specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via
# 2nd main() call it botches the future eval.
#
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
# test 2nd time - was getting eval_loss': nan'
# to reproduce the problem set distributed=False
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
@parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''])
@require_torch_multi_gpu
def __a ( self :str , _lowercase :Any) -> List[str]:
# as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout
UpperCAmelCase_ = {
# test with the default log_level - should be info and thus log info once
'''base''': {'''extra_args_str''': '''''', '''n_matches''': 1},
# test with low log_level and log_level_replica - should be noisy on all processes
# now the info string should appear twice on 2 processes
'''low''': {'''extra_args_str''': '''--log_level debug --log_level_replica debug''', '''n_matches''': 2},
# test with high log_level and low log_level_replica
# now the info string should appear once only on the replica
'''high''': {'''extra_args_str''': '''--log_level error --log_level_replica debug''', '''n_matches''': 1},
# test with high log_level and log_level_replica - should be quiet on all processes
'''mixed''': {'''extra_args_str''': '''--log_level error --log_level_replica error''', '''n_matches''': 0},
}
UpperCAmelCase_ = experiments[experiment_id]
UpperCAmelCase_ = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False}
UpperCAmelCase_ = '''Running training'''
with CaptureStderr() as cl:
self.run_seqaseq_quick(**_lowercase , extra_args_str=data['''extra_args_str'''])
UpperCAmelCase_ = len(re.findall(_lowercase , cl.err))
self.assertEqual(_lowercase , data['''n_matches'''])
@slow
def __a ( self :Any) -> Dict:
UpperCAmelCase_ = self.run_trainer(
eval_steps=2 , max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=10 , distributed=_lowercase , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
UpperCAmelCase_ = eval_metrics[-1]
assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing"
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
# test if do_predict saves generations and metrics
UpperCAmelCase_ = os.listdir(_lowercase)
UpperCAmelCase_ = {os.path.basename(_lowercase) for p in contents}
assert "generated_predictions.txt" in contents
assert "predict_results.json" in contents
@slow
@require_bitsandbytes
def __a ( self :List[str]) -> str:
from transformers.training_args import OptimizerNames
def train_and_return_metrics(_lowercase :str) -> Tuple[int, float]:
UpperCAmelCase_ = '''--skip_memory_metrics 0'''
UpperCAmelCase_ = self.run_trainer(
max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=1 , optim=_lowercase , distributed=_lowercase , extra_args_str=_lowercase , do_eval=_lowercase , do_predict=_lowercase , n_gpus_to_use=1 , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(Path(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20)
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20)
UpperCAmelCase_ = logs[0]['''train_loss''']
return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value)
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value)
UpperCAmelCase_ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_peak_mem_orig + gpu_alloc_mem_orig
UpperCAmelCase_ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_total_mem_orig - gpu_total_mem_bnb
# sshleifer/student_marian_en_ro_6_1 has 54M parameter, 29M of which is `nn.Embedding` which
# doesn't get quantized and remains in fp32. Therefore we only have 25M parameters quantized
# in 2 bytes and the diff in optim memory usage is derived as so:
#
# - normal 25*8=~200MB (8 bytes per param)
# - bnb 25*2= ~50MB (2 bytes per param)
#
# Thus we should expect ~150MB total memory saved.
#
# Peak memory should be the same - the total should be different by about that same margin
#
# After leaving a small margin to accommodate for differences between gpus let's check
# that we have at least 120MB in savings
UpperCAmelCase_ = 120
# uncomment the following if this test starts failing - requires py38 for a new print feature
# gpu_peak_mem_diff = gpu_peak_mem_orig - gpu_peak_mem_bnb
# print(f"{gpu_alloc_mem_orig=}MB {gpu_peak_mem_orig=}MB {gpu_alloc_mem_orig+gpu_peak_mem_orig=}MB")
# print(f" {gpu_alloc_mem_bnb=}MB {gpu_peak_mem_bnb=}MB {gpu_alloc_mem_bnb+gpu_peak_mem_bnb=}MB")
# print(f"{gpu_alloc_mem_diff=}MB")
# print(f"{gpu_peak_mem_diff=}MB")
# print(f"{gpu_total_mem_orig=}MB, {gpu_total_mem_bnb=}MB")
# print(f"{gpu_total_mem_diff=}MB, {gpu_total_mem_diff=}MB")
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less alloc gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_alloc_mem_diff}MB, with gpu_alloc_mem_orig={gpu_alloc_mem_orig}MB and"
f" gpu_alloc_mem_bnb={gpu_alloc_mem_bnb}MB" , )
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less total gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_total_mem_diff}MB, with gpu_total_mem_orig={gpu_total_mem_orig}MB and"
f" gpu_total_mem_bnb={gpu_total_mem_bnb}MB" , )
self.assertEqual(
_lowercase , _lowercase , f"loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}")
def __a ( self :Any , _lowercase :int , _lowercase :str , _lowercase :int , _lowercase :float = 3E-3 , _lowercase :str = "adafactor" , _lowercase :bool = False , _lowercase :str = None , _lowercase :int = 0 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :int = None , ) -> List[Any]:
UpperCAmelCase_ = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro'''
UpperCAmelCase_ = self.get_auto_remove_tmp_dir()
UpperCAmelCase_ = f"\n --model_name_or_path {model_name}\n --train_file {data_dir}/train.json\n --validation_file {data_dir}/val.json\n --test_file {data_dir}/test.json\n --output_dir {output_dir}\n --overwrite_output_dir\n --max_train_samples 8\n --max_source_length {max_len}\n --max_target_length {max_len}\n --do_train\n --num_train_epochs {str(_lowercase)}\n --per_device_train_batch_size 4\n --learning_rate {learning_rate}\n --warmup_steps 8\n --logging_steps 0\n --logging_strategy no\n --save_steps {str(_lowercase)}\n --group_by_length\n --label_smoothing_factor 0.1\n --target_lang ro_RO\n --source_lang en_XX\n ".split()
UpperCAmelCase_ = f"\n --do_eval\n --per_device_eval_batch_size 4\n --max_eval_samples 8\n --val_max_target_length {max_len}\n --evaluation_strategy steps\n --eval_steps {str(_lowercase)}\n ".split()
UpperCAmelCase_ = '''
--do_predict
'''.split()
UpperCAmelCase_ = []
if do_train:
args += args_train
if do_eval:
args += args_eval
if do_predict:
args += args_predict
if predict_with_generate:
args += "--predict_with_generate".split()
if do_train:
if optim == "adafactor":
args += "--adafactor".split()
else:
args += f"--optim {optim}".split()
if extra_args_str is not None:
args += extra_args_str.split()
if distributed:
if n_gpus_to_use is None:
UpperCAmelCase_ = get_gpu_count()
UpperCAmelCase_ = get_torch_dist_unique_port()
UpperCAmelCase_ = f"\n -m torch.distributed.run\n --nproc_per_node={n_gpus_to_use}\n --master_port={master_port}\n {self.examples_dir_str}/pytorch/translation/run_translation.py\n ".split()
UpperCAmelCase_ = [sys.executable] + distributed_args + args
# keep for quick debug
# print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die
execute_subprocess_async(_lowercase , env=self.get_env())
else:
UpperCAmelCase_ = ['''run_translation.py'''] + args
with patch.object(_lowercase , '''argv''' , _lowercase):
main()
return output_dir
| 344 | 1 |
def A ( __UpperCAmelCase ) -> list:
'''simple docstring'''
if len(__UpperCAmelCase ) <= 1:
return [tuple(__UpperCAmelCase )]
UpperCAmelCase_ = []
def generate(__UpperCAmelCase , __UpperCAmelCase ):
if k == 1:
res.append(tuple(arr[:] ) )
return
generate(k - 1 , __UpperCAmelCase )
for i in range(k - 1 ):
if k % 2 == 0: # k is even
UpperCAmelCase_ , UpperCAmelCase_ = arr[k - 1], arr[i]
else: # k is odd
UpperCAmelCase_ , UpperCAmelCase_ = arr[k - 1], arr[0]
generate(k - 1 , __UpperCAmelCase )
generate(len(__UpperCAmelCase ) , __UpperCAmelCase )
return res
if __name__ == "__main__":
UpperCamelCase_ = input("Enter numbers separated by a comma:\n").strip()
UpperCamelCase_ = [int(item) for item in user_input.split(",")]
print(heaps(arr))
| 344 |
import functools
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = len(__UpperCAmelCase )
UpperCAmelCase_ = len(__UpperCAmelCase )
@functools.cache
def min_distance(__UpperCAmelCase , __UpperCAmelCase ) -> int:
# if first word index is overflow - delete all from the second word
if indexa >= len_worda:
return len_worda - indexa
# if second word index is overflow - delete all from the first word
if indexa >= len_worda:
return len_worda - indexa
UpperCAmelCase_ = int(worda[indexa] != worda[indexa] ) # current letters not identical
return min(
1 + min_distance(indexa + 1 , __UpperCAmelCase ) , 1 + min_distance(__UpperCAmelCase , indexa + 1 ) , diff + min_distance(indexa + 1 , indexa + 1 ) , )
return min_distance(0 , 0 )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 | 1 |
from __future__ import annotations
import requests
UpperCamelCase_ = set(
"approved_at_utc approved_by author_flair_background_color\nauthor_flair_css_class author_flair_richtext author_flair_template_id author_fullname\nauthor_premium can_mod_post category clicked content_categories created_utc downs\nedited gilded gildings hidden hide_score is_created_from_ads_ui is_meta\nis_original_content is_reddit_media_domain is_video link_flair_css_class\nlink_flair_richtext link_flair_text link_flair_text_color media_embed mod_reason_title\nname permalink pwls quarantine saved score secure_media secure_media_embed selftext\nsubreddit subreddit_name_prefixed subreddit_type thumbnail title top_awarded_type\ntotal_awards_received ups upvote_ratio url user_reports".split()
)
def A ( __UpperCAmelCase , __UpperCAmelCase = 1 , __UpperCAmelCase = "new" , __UpperCAmelCase = None ) -> dict:
'''simple docstring'''
UpperCAmelCase_ = wanted_data or []
if invalid_search_terms := ", ".join(sorted(set(__UpperCAmelCase ) - valid_terms ) ):
UpperCAmelCase_ = f"Invalid search term: {invalid_search_terms}"
raise ValueError(__UpperCAmelCase )
UpperCAmelCase_ = requests.get(
f"https://reddit.com/r/{subreddit}/{age}.json?limit={limit}" , headers={'''User-agent''': '''A random string'''} , )
if response.status_code == 429:
raise requests.HTTPError
UpperCAmelCase_ = response.json()
if not wanted_data:
return {id_: data["data"]["children"][id_] for id_ in range(__UpperCAmelCase )}
UpperCAmelCase_ = {}
for id_ in range(__UpperCAmelCase ):
UpperCAmelCase_ = {
item: data['''data''']['''children'''][id_]['''data'''][item] for item in wanted_data
}
return data_dict
if __name__ == "__main__":
# If you get Error 429, that means you are rate limited.Try after some time
print(get_subreddit_data("learnpython", wanted_data=["title", "url", "selftext"]))
| 344 |
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import SPIECE_UNDERLINE, logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {"vocab_file": "spiece.model"}
UpperCamelCase_ = {
"vocab_file": {
"xlnet-base-cased": "https://huggingface.co/xlnet-base-cased/resolve/main/spiece.model",
"xlnet-large-cased": "https://huggingface.co/xlnet-large-cased/resolve/main/spiece.model",
}
}
UpperCamelCase_ = {
"xlnet-base-cased": None,
"xlnet-large-cased": None,
}
# Segments (not really needed)
UpperCamelCase_ = 0
UpperCamelCase_ = 1
UpperCamelCase_ = 2
UpperCamelCase_ = 3
UpperCamelCase_ = 4
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =VOCAB_FILES_NAMES
UpperCamelCase__ : Optional[Any] =PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ : Tuple =PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ : Any ="left"
def __init__( self :Optional[int] , _lowercase :Union[str, Any] , _lowercase :Union[str, Any]=False , _lowercase :Optional[int]=True , _lowercase :Union[str, Any]=False , _lowercase :Tuple="<s>" , _lowercase :Any="</s>" , _lowercase :Dict="<unk>" , _lowercase :str="<sep>" , _lowercase :Tuple="<pad>" , _lowercase :Any="<cls>" , _lowercase :List[str]="<mask>" , _lowercase :Union[str, Any]=["<eop>", "<eod>"] , _lowercase :Optional[Dict[str, Any]] = None , **_lowercase :Union[str, Any] , ) -> None:
# Mask token behave like a normal word, i.e. include the space before it
UpperCAmelCase_ = AddedToken(_lowercase , lstrip=_lowercase , rstrip=_lowercase) if isinstance(_lowercase , _lowercase) else mask_token
UpperCAmelCase_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_lowercase , remove_space=_lowercase , keep_accents=_lowercase , bos_token=_lowercase , eos_token=_lowercase , unk_token=_lowercase , sep_token=_lowercase , pad_token=_lowercase , cls_token=_lowercase , mask_token=_lowercase , additional_special_tokens=_lowercase , sp_model_kwargs=self.sp_model_kwargs , **_lowercase , )
UpperCAmelCase_ = 3
UpperCAmelCase_ = do_lower_case
UpperCAmelCase_ = remove_space
UpperCAmelCase_ = keep_accents
UpperCAmelCase_ = vocab_file
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(_lowercase)
@property
def __a ( self :int) -> List[Any]:
return len(self.sp_model)
def __a ( self :Optional[int]) -> List[Any]:
UpperCAmelCase_ = {self.convert_ids_to_tokens(_lowercase): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def __getstate__( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.__dict__.copy()
UpperCAmelCase_ = None
return state
def __setstate__( self :Optional[Any] , _lowercase :Optional[Any]) -> List[Any]:
UpperCAmelCase_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs'''):
UpperCAmelCase_ = {}
UpperCAmelCase_ = spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(self.vocab_file)
def __a ( self :List[str] , _lowercase :Tuple) -> Optional[int]:
if self.remove_space:
UpperCAmelCase_ = ''' '''.join(inputs.strip().split())
else:
UpperCAmelCase_ = inputs
UpperCAmelCase_ = outputs.replace('''``''' , '''"''').replace('''\'\'''' , '''"''')
if not self.keep_accents:
UpperCAmelCase_ = unicodedata.normalize('''NFKD''' , _lowercase)
UpperCAmelCase_ = ''''''.join([c for c in outputs if not unicodedata.combining(_lowercase)])
if self.do_lower_case:
UpperCAmelCase_ = outputs.lower()
return outputs
def __a ( self :str , _lowercase :str) -> List[str]:
UpperCAmelCase_ = self.preprocess_text(_lowercase)
UpperCAmelCase_ = self.sp_model.encode(_lowercase , out_type=_lowercase)
UpperCAmelCase_ = []
for piece in pieces:
if len(_lowercase) > 1 and piece[-1] == str(''',''') and piece[-2].isdigit():
UpperCAmelCase_ = self.sp_model.EncodeAsPieces(piece[:-1].replace(_lowercase , ''''''))
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0]) == 1:
UpperCAmelCase_ = cur_pieces[1:]
else:
UpperCAmelCase_ = cur_pieces[0][1:]
cur_pieces.append(piece[-1])
new_pieces.extend(_lowercase)
else:
new_pieces.append(_lowercase)
return new_pieces
def __a ( self :Optional[Any] , _lowercase :Union[str, Any]) -> Tuple:
return self.sp_model.PieceToId(_lowercase)
def __a ( self :Optional[int] , _lowercase :Optional[Any]) -> List[str]:
return self.sp_model.IdToPiece(_lowercase)
def __a ( self :List[Any] , _lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = ''''''.join(_lowercase).replace(_lowercase , ''' ''').strip()
return out_string
def __a ( self :Union[str, Any] , _lowercase :List[int] , _lowercase :bool = False , _lowercase :bool = None , _lowercase :bool = True , **_lowercase :Tuple , ) -> str:
UpperCAmelCase_ = kwargs.pop('''use_source_tokenizer''' , _lowercase)
UpperCAmelCase_ = self.convert_ids_to_tokens(_lowercase , skip_special_tokens=_lowercase)
# To avoid mixing byte-level and unicode for byte-level BPT
# we need to build string separately for added tokens and byte-level tokens
# cf. https://github.com/huggingface/transformers/issues/1133
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for token in filtered_tokens:
if skip_special_tokens and token in self.all_special_ids:
continue
if token in self.added_tokens_encoder:
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
UpperCAmelCase_ = []
sub_texts.append(_lowercase)
else:
current_sub_text.append(_lowercase)
if current_sub_text:
sub_texts.append(self.convert_tokens_to_string(_lowercase))
# Mimic the behavior of the Rust tokenizer:
# By default, there are no spaces between special tokens
UpperCAmelCase_ = ''''''.join(_lowercase)
UpperCAmelCase_ = (
clean_up_tokenization_spaces
if clean_up_tokenization_spaces is not None
else self.clean_up_tokenization_spaces
)
if clean_up_tokenization_spaces:
UpperCAmelCase_ = self.clean_up_tokenization(_lowercase)
return clean_text
else:
return text
def __a ( self :str , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [self.cls_token_id]
if token_ids_a is None:
return token_ids_a + sep + cls
return token_ids_a + sep + token_ids_a + sep + cls
def __a ( self :Dict , _lowercase :List[int] , _lowercase :Optional[List[int]] = None , _lowercase :bool = False) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_lowercase , token_ids_a=_lowercase , already_has_special_tokens=_lowercase)
if token_ids_a is not None:
return ([0] * len(_lowercase)) + [1] + ([0] * len(_lowercase)) + [1, 1]
return ([0] * len(_lowercase)) + [1, 1]
def __a ( self :Optional[int] , _lowercase :List[int] , _lowercase :Optional[List[int]] = None) -> List[int]:
UpperCAmelCase_ = [self.sep_token_id]
UpperCAmelCase_ = [2]
if token_ids_a is None:
return len(token_ids_a + sep) * [0] + cls_segment_id
return len(token_ids_a + sep) * [0] + len(token_ids_a + sep) * [1] + cls_segment_id
def __a ( self :str , _lowercase :str , _lowercase :Optional[str] = None) -> Tuple[str]:
if not os.path.isdir(_lowercase):
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
return
UpperCAmelCase_ = os.path.join(
_lowercase , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''])
if os.path.abspath(self.vocab_file) != os.path.abspath(_lowercase) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , _lowercase)
elif not os.path.isfile(self.vocab_file):
with open(_lowercase , '''wb''') as fi:
UpperCAmelCase_ = self.sp_model.serialized_model_proto()
fi.write(_lowercase)
return (out_vocab_file,)
| 344 | 1 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import (
AutoProcessor,
BertTokenizerFast,
BlipImageProcessor,
GPTaTokenizer,
InstructBlipProcessor,
PreTrainedTokenizerFast,
)
@require_vision
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = BlipImageProcessor()
UpperCAmelCase_ = GPTaTokenizer.from_pretrained('''hf-internal-testing/tiny-random-GPT2Model''')
UpperCAmelCase_ = BertTokenizerFast.from_pretrained('''hf-internal-testing/tiny-random-bert''')
UpperCAmelCase_ = InstructBlipProcessor(_lowercase , _lowercase , _lowercase)
processor.save_pretrained(self.tmpdirname)
def __a ( self :List[Any] , **_lowercase :Dict) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).tokenizer
def __a ( self :Optional[Any] , **_lowercase :Optional[Any]) -> Optional[int]:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).image_processor
def __a ( self :Dict , **_lowercase :Tuple) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).qformer_tokenizer
def __a ( self :Optional[int]) -> str:
shutil.rmtree(self.tmpdirname)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta)]
UpperCAmelCase_ = [Image.fromarray(np.moveaxis(_lowercase , 0 , -1)) for x in image_inputs]
return image_inputs
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() , qformer_tokenizer=self.get_qformer_tokenizer() , )
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = self.get_image_processor(do_normalize=_lowercase , padding_value=1.0)
UpperCAmelCase_ = InstructBlipProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowercase , padding_value=1.0)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
self.assertIsInstance(processor.tokenizer , _lowercase)
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string())
self.assertIsInstance(processor.image_processor , _lowercase)
self.assertIsInstance(processor.qformer_tokenizer , _lowercase)
def __a ( self :Dict) -> Any:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = image_processor(_lowercase , return_tensors='''np''')
UpperCAmelCase_ = processor(images=_lowercase , return_tensors='''np''')
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1E-2)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = processor(text=_lowercase)
UpperCAmelCase_ = tokenizer(_lowercase , return_token_type_ids=_lowercase)
UpperCAmelCase_ = qformer_tokenizer(_lowercase , return_token_type_ids=_lowercase)
for key in encoded_tokens.keys():
self.assertListEqual(encoded_tokens[key] , encoded_processor[key])
for key in encoded_tokens_qformer.keys():
self.assertListEqual(encoded_tokens_qformer[key] , encoded_processor['''qformer_''' + key])
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
# test if it raises when no input is passed
with pytest.raises(_lowercase):
processor()
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
UpperCAmelCase_ = processor.batch_decode(_lowercase)
UpperCAmelCase_ = tokenizer.batch_decode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :str) -> int:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
UpperCamelCase_ = logging.get_logger(__name__)
class a_ ( _snake_case , _snake_case ):
UpperCamelCase__ : Union[str, Any] ="maskformer-swin"
UpperCamelCase__ : List[str] ={
"num_attention_heads": "num_heads",
"num_hidden_layers": "num_layers",
}
def __init__( self :Union[str, Any] , _lowercase :Optional[int]=224 , _lowercase :List[str]=4 , _lowercase :Tuple=3 , _lowercase :List[Any]=96 , _lowercase :Any=[2, 2, 6, 2] , _lowercase :int=[3, 6, 12, 24] , _lowercase :List[Any]=7 , _lowercase :Dict=4.0 , _lowercase :Any=True , _lowercase :int=0.0 , _lowercase :List[Any]=0.0 , _lowercase :Tuple=0.1 , _lowercase :str="gelu" , _lowercase :Union[str, Any]=False , _lowercase :Tuple=0.02 , _lowercase :List[str]=1E-5 , _lowercase :List[str]=None , _lowercase :Any=None , **_lowercase :Any , ) -> Union[str, Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = embed_dim
UpperCAmelCase_ = depths
UpperCAmelCase_ = len(_lowercase)
UpperCAmelCase_ = num_heads
UpperCAmelCase_ = window_size
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = qkv_bias
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = use_absolute_embeddings
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = initializer_range
# we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
UpperCAmelCase_ = int(embed_dim * 2 ** (len(_lowercase) - 1))
UpperCAmelCase_ = ['''stem'''] + [f"stage{idx}" for idx in range(1 , len(_lowercase) + 1)]
UpperCAmelCase_ , UpperCAmelCase_ = get_aligned_output_features_output_indices(
out_features=_lowercase , out_indices=_lowercase , stage_names=self.stage_names)
| 344 | 1 |
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...file_utils import TensorType, is_torch_available
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"facebook/blenderbot_small-90M": "https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/config.json",
# See all BlenderbotSmall models at https://huggingface.co/models?filter=blenderbot_small
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] ="blenderbot-small"
UpperCamelCase__ : Any =["past_key_values"]
UpperCamelCase__ : Optional[Any] ={"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self :Union[str, Any] , _lowercase :List[str]=50265 , _lowercase :Any=512 , _lowercase :str=8 , _lowercase :Union[str, Any]=2048 , _lowercase :Union[str, Any]=16 , _lowercase :Tuple=8 , _lowercase :Tuple=2048 , _lowercase :str=16 , _lowercase :Any=0.0 , _lowercase :Optional[int]=0.0 , _lowercase :Union[str, Any]=True , _lowercase :int=True , _lowercase :int="gelu" , _lowercase :List[Any]=512 , _lowercase :Tuple=0.1 , _lowercase :Dict=0.0 , _lowercase :Dict=0.0 , _lowercase :Optional[int]=0.02 , _lowercase :str=1 , _lowercase :Optional[int]=False , _lowercase :Optional[Any]=0 , _lowercase :Any=1 , _lowercase :str=2 , _lowercase :Any=2 , **_lowercase :Optional[int] , ) -> List[str]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = d_model
UpperCAmelCase_ = encoder_ffn_dim
UpperCAmelCase_ = encoder_layers
UpperCAmelCase_ = encoder_attention_heads
UpperCAmelCase_ = decoder_ffn_dim
UpperCAmelCase_ = decoder_layers
UpperCAmelCase_ = decoder_attention_heads
UpperCAmelCase_ = dropout
UpperCAmelCase_ = attention_dropout
UpperCAmelCase_ = activation_dropout
UpperCAmelCase_ = activation_function
UpperCAmelCase_ = init_std
UpperCAmelCase_ = encoder_layerdrop
UpperCAmelCase_ = decoder_layerdrop
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = encoder_layers
UpperCAmelCase_ = scale_embedding # scale factor will be sqrt(d_model) if True
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , is_encoder_decoder=_lowercase , decoder_start_token_id=_lowercase , forced_eos_token_id=_lowercase , **_lowercase , )
class a_ ( _snake_case ):
@property
def __a ( self :Optional[int]) -> Mapping[str, Mapping[int, str]]:
if self.task in ["default", "seq2seq-lm"]:
UpperCAmelCase_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
])
if self.use_past:
UpperCAmelCase_ = {0: '''batch'''}
UpperCAmelCase_ = {0: '''batch''', 1: '''past_decoder_sequence + sequence'''}
else:
UpperCAmelCase_ = {0: '''batch''', 1: '''decoder_sequence'''}
UpperCAmelCase_ = {0: '''batch''', 1: '''decoder_sequence'''}
if self.use_past:
self.fill_with_past_key_values_(_lowercase , direction='''inputs''')
elif self.task == "causal-lm":
# TODO: figure this case out.
UpperCAmelCase_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
])
if self.use_past:
UpperCAmelCase_ , UpperCAmelCase_ = self.num_layers
for i in range(_lowercase):
UpperCAmelCase_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
UpperCAmelCase_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
else:
UpperCAmelCase_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''decoder_input_ids''', {0: '''batch''', 1: '''decoder_sequence'''}),
('''decoder_attention_mask''', {0: '''batch''', 1: '''decoder_sequence'''}),
])
return common_inputs
@property
def __a ( self :Any) -> Mapping[str, Mapping[int, str]]:
if self.task in ["default", "seq2seq-lm"]:
UpperCAmelCase_ = super().outputs
else:
UpperCAmelCase_ = super(_lowercase , self).outputs
if self.use_past:
UpperCAmelCase_ , UpperCAmelCase_ = self.num_layers
for i in range(_lowercase):
UpperCAmelCase_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
UpperCAmelCase_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
return common_outputs
def __a ( self :List[str] , _lowercase :PreTrainedTokenizer , _lowercase :int = -1 , _lowercase :int = -1 , _lowercase :bool = False , _lowercase :Optional[TensorType] = None , ) -> Mapping[str, Any]:
UpperCAmelCase_ = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
_lowercase , _lowercase , _lowercase , _lowercase , _lowercase)
# Generate decoder inputs
UpperCAmelCase_ = seq_length if not self.use_past else 1
UpperCAmelCase_ = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
_lowercase , _lowercase , _lowercase , _lowercase , _lowercase)
UpperCAmelCase_ = {f"decoder_{name}": tensor for name, tensor in decoder_inputs.items()}
UpperCAmelCase_ = dict(**_lowercase , **_lowercase)
if self.use_past:
if not is_torch_available():
raise ValueError('''Cannot generate dummy past_keys inputs without PyTorch installed.''')
else:
import torch
UpperCAmelCase_ , UpperCAmelCase_ = common_inputs['''input_ids'''].shape
UpperCAmelCase_ = common_inputs['''decoder_input_ids'''].shape[1]
UpperCAmelCase_ , UpperCAmelCase_ = self.num_attention_heads
UpperCAmelCase_ = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
UpperCAmelCase_ = decoder_seq_length + 3
UpperCAmelCase_ = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
UpperCAmelCase_ = torch.cat(
[common_inputs['''decoder_attention_mask'''], torch.ones(_lowercase , _lowercase)] , dim=1)
UpperCAmelCase_ = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
UpperCAmelCase_ , UpperCAmelCase_ = self.num_layers
UpperCAmelCase_ = min(_lowercase , _lowercase)
UpperCAmelCase_ = max(_lowercase , _lowercase) - min_num_layers
UpperCAmelCase_ = '''encoder''' if num_encoder_layers > num_decoder_layers else '''decoder'''
for _ in range(_lowercase):
common_inputs["past_key_values"].append(
(
torch.zeros(_lowercase),
torch.zeros(_lowercase),
torch.zeros(_lowercase),
torch.zeros(_lowercase),
))
# TODO: test this.
UpperCAmelCase_ = encoder_shape if remaining_side_name == '''encoder''' else decoder_shape
for _ in range(_lowercase , _lowercase):
common_inputs["past_key_values"].append((torch.zeros(_lowercase), torch.zeros(_lowercase)))
return common_inputs
def __a ( self :List[Any] , _lowercase :PreTrainedTokenizer , _lowercase :int = -1 , _lowercase :int = -1 , _lowercase :bool = False , _lowercase :Optional[TensorType] = None , ) -> Mapping[str, Any]:
UpperCAmelCase_ = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
_lowercase , _lowercase , _lowercase , _lowercase , _lowercase)
if self.use_past:
if not is_torch_available():
raise ValueError('''Cannot generate dummy past_keys inputs without PyTorch installed.''')
else:
import torch
UpperCAmelCase_ , UpperCAmelCase_ = common_inputs['''input_ids'''].shape
# Not using the same length for past_key_values
UpperCAmelCase_ = seqlen + 2
UpperCAmelCase_ , UpperCAmelCase_ = self.num_layers
UpperCAmelCase_ , UpperCAmelCase_ = self.num_attention_heads
UpperCAmelCase_ = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
UpperCAmelCase_ = common_inputs['''attention_mask'''].dtype
UpperCAmelCase_ = torch.cat(
[common_inputs['''attention_mask'''], torch.ones(_lowercase , _lowercase , dtype=_lowercase)] , dim=1)
UpperCAmelCase_ = [
(torch.zeros(_lowercase), torch.zeros(_lowercase)) for _ in range(_lowercase)
]
return common_inputs
def __a ( self :Optional[int] , _lowercase :PreTrainedTokenizer , _lowercase :int = -1 , _lowercase :int = -1 , _lowercase :bool = False , _lowercase :Optional[TensorType] = None , ) -> Mapping[str, Any]:
# Copied from OnnxConfig.generate_dummy_inputs
# Did not use super(OnnxConfigWithPast, self).generate_dummy_inputs for code clarity.
# If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX
UpperCAmelCase_ = compute_effective_axis_dimension(
_lowercase , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0)
# If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX
UpperCAmelCase_ = tokenizer.num_special_tokens_to_add(_lowercase)
UpperCAmelCase_ = compute_effective_axis_dimension(
_lowercase , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=_lowercase)
# Generate dummy inputs according to compute batch and sequence
UpperCAmelCase_ = [''' '''.join([tokenizer.unk_token]) * seq_length] * batch_size
UpperCAmelCase_ = dict(tokenizer(_lowercase , return_tensors=_lowercase))
return common_inputs
def __a ( self :Optional[Any] , _lowercase :PreTrainedTokenizer , _lowercase :int = -1 , _lowercase :int = -1 , _lowercase :bool = False , _lowercase :Optional[TensorType] = None , ) -> Mapping[str, Any]:
if self.task in ["default", "seq2seq-lm"]:
UpperCAmelCase_ = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
_lowercase , batch_size=_lowercase , seq_length=_lowercase , is_pair=_lowercase , framework=_lowercase)
elif self.task == "causal-lm":
UpperCAmelCase_ = self._generate_dummy_inputs_for_causal_lm(
_lowercase , batch_size=_lowercase , seq_length=_lowercase , is_pair=_lowercase , framework=_lowercase)
else:
UpperCAmelCase_ = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
_lowercase , batch_size=_lowercase , seq_length=_lowercase , is_pair=_lowercase , framework=_lowercase)
return common_inputs
def __a ( self :Tuple , _lowercase :int , _lowercase :Optional[int] , _lowercase :Union[str, Any] , _lowercase :Optional[Any]) -> List[str]:
if self.task in ["default", "seq2seq-lm"]:
UpperCAmelCase_ = super()._flatten_past_key_values_(_lowercase , _lowercase , _lowercase , _lowercase)
else:
UpperCAmelCase_ = super(_lowercase , self)._flatten_past_key_values_(
_lowercase , _lowercase , _lowercase , _lowercase)
| 344 |
import os
from argparse import ArgumentParser, Namespace
from ..data import SingleSentenceClassificationProcessor as Processor
from ..pipelines import TextClassificationPipeline
from ..utils import is_tf_available, is_torch_available, logging
from . import BaseTransformersCLICommand
if not is_tf_available() and not is_torch_available():
raise RuntimeError("At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training")
# TF training parameters
UpperCamelCase_ = False
UpperCamelCase_ = False
def A ( __UpperCAmelCase ) -> Any:
'''simple docstring'''
return TrainCommand(__UpperCAmelCase )
class a_ ( _snake_case ):
@staticmethod
def __a ( _lowercase :ArgumentParser) -> List[Any]:
UpperCAmelCase_ = parser.add_parser('''train''' , help='''CLI tool to train a model on a task.''')
train_parser.add_argument(
'''--train_data''' , type=_lowercase , required=_lowercase , help='''path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.''' , )
train_parser.add_argument(
'''--column_label''' , type=_lowercase , default=0 , help='''Column of the dataset csv file with example labels.''')
train_parser.add_argument(
'''--column_text''' , type=_lowercase , default=1 , help='''Column of the dataset csv file with example texts.''')
train_parser.add_argument(
'''--column_id''' , type=_lowercase , default=2 , help='''Column of the dataset csv file with example ids.''')
train_parser.add_argument(
'''--skip_first_row''' , action='''store_true''' , help='''Skip the first row of the csv file (headers).''')
train_parser.add_argument('''--validation_data''' , type=_lowercase , default='''''' , help='''path to validation dataset.''')
train_parser.add_argument(
'''--validation_split''' , type=_lowercase , default=0.1 , help='''if validation dataset is not provided, fraction of train dataset to use as validation dataset.''' , )
train_parser.add_argument('''--output''' , type=_lowercase , default='''./''' , help='''path to saved the trained model.''')
train_parser.add_argument(
'''--task''' , type=_lowercase , default='''text_classification''' , help='''Task to train the model on.''')
train_parser.add_argument(
'''--model''' , type=_lowercase , default='''bert-base-uncased''' , help='''Model\'s name or path to stored model.''')
train_parser.add_argument('''--train_batch_size''' , type=_lowercase , default=32 , help='''Batch size for training.''')
train_parser.add_argument('''--valid_batch_size''' , type=_lowercase , default=64 , help='''Batch size for validation.''')
train_parser.add_argument('''--learning_rate''' , type=_lowercase , default=3E-5 , help='''Learning rate.''')
train_parser.add_argument('''--adam_epsilon''' , type=_lowercase , default=1E-0_8 , help='''Epsilon for Adam optimizer.''')
train_parser.set_defaults(func=_lowercase)
def __init__( self :Union[str, Any] , _lowercase :Namespace) -> Union[str, Any]:
UpperCAmelCase_ = logging.get_logger('''transformers-cli/training''')
UpperCAmelCase_ = '''tf''' if is_tf_available() else '''torch'''
os.makedirs(args.output , exist_ok=_lowercase)
UpperCAmelCase_ = args.output
UpperCAmelCase_ = args.column_label
UpperCAmelCase_ = args.column_text
UpperCAmelCase_ = args.column_id
self.logger.info(f"Loading {args.task} pipeline for {args.model}")
if args.task == "text_classification":
UpperCAmelCase_ = TextClassificationPipeline.from_pretrained(args.model)
elif args.task == "token_classification":
raise NotImplementedError
elif args.task == "question_answering":
raise NotImplementedError
self.logger.info(f"Loading dataset from {args.train_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.train_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = None
if args.validation_data:
self.logger.info(f"Loading validation dataset from {args.validation_data}")
UpperCAmelCase_ = Processor.create_from_csv(
args.validation_data , column_label=args.column_label , column_text=args.column_text , column_id=args.column_id , skip_first_row=args.skip_first_row , )
UpperCAmelCase_ = args.validation_split
UpperCAmelCase_ = args.train_batch_size
UpperCAmelCase_ = args.valid_batch_size
UpperCAmelCase_ = args.learning_rate
UpperCAmelCase_ = args.adam_epsilon
def __a ( self :int) -> Tuple:
if self.framework == "tf":
return self.run_tf()
return self.run_torch()
def __a ( self :Optional[Any]) -> Any:
raise NotImplementedError
def __a ( self :int) -> Optional[Any]:
self.pipeline.fit(
self.train_dataset , validation_data=self.valid_dataset , validation_split=self.validation_split , learning_rate=self.learning_rate , adam_epsilon=self.adam_epsilon , train_batch_size=self.train_batch_size , valid_batch_size=self.valid_batch_size , )
# Save trained pipeline
self.pipeline.save_pretrained(self.output)
| 344 | 1 |
import numpy as np
import torch
from torch.utils.data import DataLoader
from accelerate.utils.dataclasses import DistributedType
class a_ :
def __init__( self :Dict , _lowercase :Tuple=2 , _lowercase :str=3 , _lowercase :Optional[Any]=64 , _lowercase :Optional[int]=None) -> List[str]:
UpperCAmelCase_ = np.random.default_rng(_lowercase)
UpperCAmelCase_ = length
UpperCAmelCase_ = rng.normal(size=(length,)).astype(np.floataa)
UpperCAmelCase_ = a * self.x + b + rng.normal(scale=0.1 , size=(length,)).astype(np.floataa)
def __len__( self :Union[str, Any]) -> List[Any]:
return self.length
def __getitem__( self :int , _lowercase :Optional[int]) -> Union[str, Any]:
return {"x": self.x[i], "y": self.y[i]}
class a_ ( torch.nn.Module ):
def __init__( self :Dict , _lowercase :str=0 , _lowercase :Optional[Any]=0 , _lowercase :Optional[int]=False) -> Optional[int]:
super().__init__()
UpperCAmelCase_ = torch.nn.Parameter(torch.tensor([2, 3]).float())
UpperCAmelCase_ = torch.nn.Parameter(torch.tensor([2, 3]).float())
UpperCAmelCase_ = True
def __a ( self :Tuple , _lowercase :Tuple=None) -> str:
if self.first_batch:
print(f"Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}")
UpperCAmelCase_ = False
return x * self.a[0] + self.b[0]
class a_ ( torch.nn.Module ):
def __init__( self :Union[str, Any] , _lowercase :int=0 , _lowercase :List[Any]=0 , _lowercase :Optional[Any]=False) -> Optional[int]:
super().__init__()
UpperCAmelCase_ = torch.nn.Parameter(torch.tensor(_lowercase).float())
UpperCAmelCase_ = torch.nn.Parameter(torch.tensor(_lowercase).float())
UpperCAmelCase_ = True
def __a ( self :List[Any] , _lowercase :List[Any]=None) -> Dict:
if self.first_batch:
print(f"Model dtype: {self.a.dtype}, {self.b.dtype}. Input dtype: {x.dtype}")
UpperCAmelCase_ = False
return x * self.a + self.b
def A ( __UpperCAmelCase , __UpperCAmelCase = 16 ) -> List[str]:
'''simple docstring'''
from datasets import load_dataset
from transformers import AutoTokenizer
UpperCAmelCase_ = AutoTokenizer.from_pretrained('''bert-base-cased''' )
UpperCAmelCase_ = {'''train''': '''tests/test_samples/MRPC/train.csv''', '''validation''': '''tests/test_samples/MRPC/dev.csv'''}
UpperCAmelCase_ = load_dataset('''csv''' , data_files=__UpperCAmelCase )
UpperCAmelCase_ = datasets['''train'''].unique('''label''' )
UpperCAmelCase_ = {v: i for i, v in enumerate(__UpperCAmelCase )}
def tokenize_function(__UpperCAmelCase ):
# max_length=None => use the model max length (it's actually the default)
UpperCAmelCase_ = tokenizer(
examples['''sentence1'''] , examples['''sentence2'''] , truncation=__UpperCAmelCase , max_length=__UpperCAmelCase , padding='''max_length''' )
if "label" in examples:
UpperCAmelCase_ = [label_to_id[l] for l in examples['''label''']]
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
UpperCAmelCase_ = datasets.map(
__UpperCAmelCase , batched=__UpperCAmelCase , remove_columns=['''sentence1''', '''sentence2''', '''label'''] , )
def collate_fn(__UpperCAmelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
if accelerator.distributed_type == DistributedType.TPU:
return tokenizer.pad(__UpperCAmelCase , padding='''max_length''' , max_length=128 , return_tensors='''pt''' )
return tokenizer.pad(__UpperCAmelCase , padding='''longest''' , return_tensors='''pt''' )
# Instantiate dataloaders.
UpperCAmelCase_ = DataLoader(tokenized_datasets['''train'''] , shuffle=__UpperCAmelCase , collate_fn=__UpperCAmelCase , batch_size=2 )
UpperCAmelCase_ = DataLoader(tokenized_datasets['''validation'''] , shuffle=__UpperCAmelCase , collate_fn=__UpperCAmelCase , batch_size=1 )
return train_dataloader, eval_dataloader
| 344 |
import unittest
from typing import Dict, List, Optional, Union
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import BridgeTowerImageProcessor
class a_ ( unittest.TestCase ):
def __init__( self :Tuple , _lowercase :List[Any] , _lowercase :bool = True , _lowercase :Dict[str, int] = None , _lowercase :int = 32 , _lowercase :bool = True , _lowercase :Union[int, float] = 1 / 255 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :Optional[Union[float, List[float]]] = [0.48_145_466, 0.4_578_275, 0.40_821_073] , _lowercase :Optional[Union[float, List[float]]] = [0.26_862_954, 0.26_130_258, 0.27_577_711] , _lowercase :bool = True , _lowercase :List[Any]=7 , _lowercase :Dict=30 , _lowercase :Optional[int]=400 , _lowercase :Any=3 , ) -> Any:
UpperCAmelCase_ = parent
UpperCAmelCase_ = do_resize
UpperCAmelCase_ = size if size is not None else {'''shortest_edge''': 288}
UpperCAmelCase_ = size_divisor
UpperCAmelCase_ = do_rescale
UpperCAmelCase_ = rescale_factor
UpperCAmelCase_ = do_normalize
UpperCAmelCase_ = do_center_crop
UpperCAmelCase_ = image_mean
UpperCAmelCase_ = image_std
UpperCAmelCase_ = do_pad
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = min_resolution
UpperCAmelCase_ = max_resolution
def __a ( self :str) -> Tuple:
return {
"image_mean": self.image_mean,
"image_std": self.image_std,
"do_normalize": self.do_normalize,
"do_resize": self.do_resize,
"size": self.size,
"size_divisor": self.size_divisor,
}
def __a ( self :List[Any] , _lowercase :Tuple , _lowercase :List[str]=False) -> int:
if not batched:
UpperCAmelCase_ = self.size['''shortest_edge''']
UpperCAmelCase_ = image_inputs[0]
if isinstance(_lowercase , Image.Image):
UpperCAmelCase_ , UpperCAmelCase_ = image.size
else:
UpperCAmelCase_ , UpperCAmelCase_ = image.shape[1], image.shape[2]
UpperCAmelCase_ = size / min(_lowercase , _lowercase)
if h < w:
UpperCAmelCase_ , UpperCAmelCase_ = size, scale * w
else:
UpperCAmelCase_ , UpperCAmelCase_ = scale * h, size
UpperCAmelCase_ = int((1333 / 800) * size)
if max(_lowercase , _lowercase) > max_size:
UpperCAmelCase_ = max_size / max(_lowercase , _lowercase)
UpperCAmelCase_ = newh * scale
UpperCAmelCase_ = neww * scale
UpperCAmelCase_ , UpperCAmelCase_ = int(newh + 0.5), int(neww + 0.5)
UpperCAmelCase_ , UpperCAmelCase_ = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
UpperCAmelCase_ = []
for image in image_inputs:
UpperCAmelCase_ , UpperCAmelCase_ = self.get_expected_values([image])
expected_values.append((expected_height, expected_width))
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[0])[0]
UpperCAmelCase_ = max(_lowercase , key=lambda _lowercase: item[1])[1]
return expected_height, expected_width
@require_torch
@require_vision
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =BridgeTowerImageProcessor if is_vision_available() else None
def __a ( self :int) -> Dict:
UpperCAmelCase_ = BridgeTowerImageProcessingTester(self)
@property
def __a ( self :Dict) -> Any:
return self.image_processor_tester.prepare_image_processor_dict()
def __a ( self :Dict) -> Tuple:
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
self.assertTrue(hasattr(_lowercase , '''image_mean'''))
self.assertTrue(hasattr(_lowercase , '''image_std'''))
self.assertTrue(hasattr(_lowercase , '''do_normalize'''))
self.assertTrue(hasattr(_lowercase , '''do_resize'''))
self.assertTrue(hasattr(_lowercase , '''size'''))
self.assertTrue(hasattr(_lowercase , '''size_divisor'''))
def __a ( self :Union[str, Any]) -> Tuple:
pass
def __a ( self :List[str]) -> Tuple:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PIL images
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , Image.Image)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :Union[str, Any]) -> Optional[int]:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random numpy tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , numpify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , np.ndarray)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
def __a ( self :str) -> int:
# Initialize image processor
UpperCAmelCase_ = self.image_processing_class(**self.image_processor_dict)
# create random PyTorch tensors
UpperCAmelCase_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=_lowercase , torchify=_lowercase)
for image in image_inputs:
self.assertIsInstance(_lowercase , torch.Tensor)
# Test not batched input
UpperCAmelCase_ = image_processing(image_inputs[0] , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase)
self.assertEqual(
encoded_images.shape , (1, self.image_processor_tester.num_channels, expected_height, expected_width) , )
# Test batched
UpperCAmelCase_ = image_processing(_lowercase , return_tensors='''pt''').pixel_values
UpperCAmelCase_ , UpperCAmelCase_ = self.image_processor_tester.get_expected_values(_lowercase , batched=_lowercase)
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
expected_height,
expected_width,
) , )
| 344 | 1 |
from __future__ import annotations
def A ( __UpperCAmelCase ) -> int:
'''simple docstring'''
for i in range(1 , len(matrix[0] ) ):
matrix[0][i] += matrix[0][i - 1]
# preprocessing the first column
for i in range(1 , len(__UpperCAmelCase ) ):
matrix[i][0] += matrix[i - 1][0]
# updating the path cost for current position
for i in range(1 , len(__UpperCAmelCase ) ):
for j in range(1 , len(matrix[0] ) ):
matrix[i][j] += min(matrix[i - 1][j] , matrix[i][j - 1] )
return matrix[-1][-1]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
def A ( __UpperCAmelCase = 100_0000 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = [i - 1 for i in range(limit + 1 )]
for i in range(2 , limit + 1 ):
if phi[i] == i - 1:
for j in range(2 * i , limit + 1 , __UpperCAmelCase ):
phi[j] -= phi[j] // i
return sum(phi[2 : limit + 1] )
if __name__ == "__main__":
print(solution())
| 344 | 1 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
is_vision_available,
)
UpperCamelCase_ = {"processing_layoutxlm": ["LayoutXLMProcessor"]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["LayoutXLMTokenizer"]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["LayoutXLMTokenizerFast"]
if TYPE_CHECKING:
from .processing_layoutxlm import LayoutXLMProcessor
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_layoutxlm import LayoutXLMTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_layoutxlm_fast import LayoutXLMTokenizerFast
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 |
import tempfile
import torch
from diffusers import PNDMScheduler
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] =(PNDMScheduler,)
UpperCamelCase__ : Optional[Any] =(("num_inference_steps", 50),)
def __a ( self :Union[str, Any] , **_lowercase :Any) -> Union[str, Any]:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
}
config.update(**_lowercase)
return config
def __a ( self :str , _lowercase :List[Any]=0 , **_lowercase :str) -> Union[str, Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Any) -> Optional[Any]:
pass
def __a ( self :str , _lowercase :int=0 , **_lowercase :Union[str, Any]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_prk(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step_plms(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :int , **_lowercase :str) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.prk_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
for i, t in enumerate(scheduler.plms_timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :Union[str, Any]) -> int:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
if num_inference_steps is not None and hasattr(_lowercase , '''set_timesteps'''):
scheduler.set_timesteps(_lowercase)
elif num_inference_steps is not None and not hasattr(_lowercase , '''set_timesteps'''):
UpperCAmelCase_ = num_inference_steps
# copy over dummy past residuals (must be done after set_timesteps)
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.1, residual + 0.05]
UpperCAmelCase_ = dummy_past_residuals[:]
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_prk(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 0 , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = scheduler.step_plms(_lowercase , 1 , _lowercase , **_lowercase).prev_sample
self.assertEqual(output_a.shape , sample.shape)
self.assertEqual(output_a.shape , output_a.shape)
def __a ( self :Any) -> Dict:
for timesteps in [100, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :List[Any]) -> Any:
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(steps_offset=1)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(10)
assert torch.equal(
scheduler.timesteps , torch.LongTensor(
[901, 851, 851, 801, 801, 751, 751, 701, 701, 651, 651, 601, 601, 501, 401, 301, 201, 101, 1]) , )
def __a ( self :Optional[int]) -> str:
for beta_start, beta_end in zip([0.0_001, 0.001] , [0.002, 0.02]):
self.check_over_configs(beta_start=_lowercase , beta_end=_lowercase)
def __a ( self :Any) -> List[str]:
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=_lowercase)
def __a ( self :List[Any]) -> Dict:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Tuple:
for t in [1, 5, 10]:
self.check_over_forward(time_step=_lowercase)
def __a ( self :Tuple) -> Dict:
for t, num_inference_steps in zip([1, 5, 10] , [10, 50, 100]):
self.check_over_forward(num_inference_steps=_lowercase)
def __a ( self :str) -> List[Any]:
# earlier version of set_timesteps() caused an error indexing alpha's with inference steps as power of 3
UpperCAmelCase_ = 27
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# before power of 3 fix, would error on first step, so we only need to do two
for i, t in enumerate(scheduler.prk_timesteps[:2]):
UpperCAmelCase_ = scheduler.step_prk(_lowercase , _lowercase , _lowercase).prev_sample
def __a ( self :List[str]) -> int:
with self.assertRaises(_lowercase):
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.step_plms(self.dummy_sample , 1 , self.dummy_sample).prev_sample
def __a ( self :List[str]) -> Dict:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 198.1_318) < 1E-2
assert abs(result_mean.item() - 0.2_580) < 1E-3
def __a ( self :Any) -> Tuple:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 67.3_986) < 1E-2
assert abs(result_mean.item() - 0.0_878) < 1E-3
def __a ( self :int) -> Any:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 230.0_399) < 1E-2
assert abs(result_mean.item() - 0.2_995) < 1E-3
def __a ( self :Any) -> Dict:
# We specify different beta, so that the first alpha is 0.99
UpperCAmelCase_ = self.full_loop(set_alpha_to_one=_lowercase , beta_start=0.01)
UpperCAmelCase_ = torch.sum(torch.abs(_lowercase))
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_sum.item() - 186.9_482) < 1E-2
assert abs(result_mean.item() - 0.2_434) < 1E-3
| 344 | 1 |
import pytest
from datasets import inspect_metric, list_metrics, load_metric
@pytest.fixture
def A ( __UpperCAmelCase ) -> Any:
'''simple docstring'''
monkeypatch.setattr('''datasets.utils.deprecation_utils._emitted_deprecation_warnings''' , set() )
@pytest.fixture
def A ( __UpperCAmelCase ) -> str:
'''simple docstring'''
class a_ :
def __init__( self :Union[str, Any] , _lowercase :Dict) -> List[Any]:
UpperCAmelCase_ = metric_id
class a_ :
UpperCamelCase__ : Dict =[MetricMock(_snake_case ) for metric_id in ["accuracy", "mse", "precision", "codeparrot/apps_metric"]]
def __a ( self :Optional[Any]) -> List[str]:
return self._metrics
monkeypatch.setattr('''datasets.inspect.huggingface_hub''' , HfhMock() )
@pytest.mark.parametrize(
'''func, args''' , [(load_metric, ('''metrics/mse''',)), (list_metrics, ()), (inspect_metric, ('''metrics/mse''', '''tmp_path'''))] )
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
if "tmp_path" in args:
UpperCAmelCase_ = tuple(arg if arg != '''tmp_path''' else tmp_path for arg in args )
with pytest.warns(__UpperCAmelCase , match='''https://huggingface.co/docs/evaluate''' ):
func(*__UpperCAmelCase )
| 344 |
import os
from dataclasses import dataclass, field
from io import BytesIO
from typing import TYPE_CHECKING, Any, ClassVar, Dict, Optional, Union
import numpy as np
import pyarrow as pa
from .. import config
from ..download.streaming_download_manager import xopen, xsplitext
from ..table import array_cast
from ..utils.py_utils import no_op_if_value_is_null, string_to_dict
if TYPE_CHECKING:
from .features import FeatureType
UpperCamelCase_ , UpperCamelCase_ , UpperCamelCase_ = False, False, False
@dataclass
class a_ :
UpperCamelCase__ : Optional[int] =None
UpperCamelCase__ : bool =True
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[str] =None
# Automatically constructed
UpperCamelCase__ : ClassVar[str] ="dict"
UpperCamelCase__ : ClassVar[Any] =pa.struct({"bytes": pa.binary(), "path": pa.string()} )
UpperCamelCase__ : str =field(default="Audio" , init=_snake_case , repr=_snake_case )
def __call__( self :List[Any]) -> List[Any]:
return self.pa_type
def __a ( self :Any , _lowercase :Union[str, bytes, dict]) -> dict:
try:
import soundfile as sf # soundfile is a dependency of librosa, needed to decode audio files.
except ImportError as err:
raise ImportError('''To support encoding audio data, please install \'soundfile\'.''') from err
if isinstance(_lowercase , _lowercase):
return {"bytes": None, "path": value}
elif isinstance(_lowercase , _lowercase):
return {"bytes": value, "path": None}
elif "array" in value:
# convert the audio array to wav bytes
UpperCAmelCase_ = BytesIO()
sf.write(_lowercase , value['''array'''] , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
elif value.get('''path''') is not None and os.path.isfile(value['''path''']):
# we set "bytes": None to not duplicate the data if they're already available locally
if value["path"].endswith('''pcm'''):
# "PCM" only has raw audio bytes
if value.get('''sampling_rate''') is None:
# At least, If you want to convert "PCM-byte" to "WAV-byte", you have to know sampling rate
raise KeyError('''To use PCM files, please specify a \'sampling_rate\' in Audio object''')
if value.get('''bytes'''):
# If we already had PCM-byte, we don`t have to make "read file, make bytes" (just use it!)
UpperCAmelCase_ = np.frombuffer(value['''bytes'''] , dtype=np.intaa).astype(np.floataa) / 32767
else:
UpperCAmelCase_ = np.memmap(value['''path'''] , dtype='''h''' , mode='''r''').astype(np.floataa) / 32767
UpperCAmelCase_ = BytesIO(bytes())
sf.write(_lowercase , _lowercase , value['''sampling_rate'''] , format='''wav''')
return {"bytes": buffer.getvalue(), "path": None}
else:
return {"bytes": None, "path": value.get('''path''')}
elif value.get('''bytes''') is not None or value.get('''path''') is not None:
# store the audio bytes, and path is used to infer the audio format using the file extension
return {"bytes": value.get('''bytes'''), "path": value.get('''path''')}
else:
raise ValueError(
f"An audio sample should have one of 'path' or 'bytes' but they are missing or None in {value}.")
def __a ( self :Dict , _lowercase :dict , _lowercase :Optional[Dict[str, Union[str, bool, None]]] = None) -> dict:
if not self.decode:
raise RuntimeError('''Decoding is disabled for this feature. Please use Audio(decode=True) instead.''')
UpperCAmelCase_ , UpperCAmelCase_ = (value['''path'''], BytesIO(value['''bytes'''])) if value['''bytes'''] is not None else (value['''path'''], None)
if path is None and file is None:
raise ValueError(f"An audio sample should have one of 'path' or 'bytes' but both are None in {value}.")
try:
import librosa
import soundfile as sf
except ImportError as err:
raise ImportError('''To support decoding audio files, please install \'librosa\' and \'soundfile\'.''') from err
UpperCAmelCase_ = xsplitext(_lowercase)[1][1:].lower() if path is not None else None
if not config.IS_OPUS_SUPPORTED and audio_format == "opus":
raise RuntimeError(
'''Decoding \'opus\' files requires system library \'libsndfile\'>=1.0.31, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
elif not config.IS_MP3_SUPPORTED and audio_format == "mp3":
raise RuntimeError(
'''Decoding \'mp3\' files requires system library \'libsndfile\'>=1.1.0, '''
'''You can try to update `soundfile` python library: `pip install "soundfile>=0.12.1"`. ''')
if file is None:
UpperCAmelCase_ = token_per_repo_id or {}
UpperCAmelCase_ = path.split('''::''')[-1]
try:
UpperCAmelCase_ = string_to_dict(_lowercase , config.HUB_DATASETS_URL)['''repo_id''']
UpperCAmelCase_ = token_per_repo_id[repo_id]
except (ValueError, KeyError):
UpperCAmelCase_ = None
with xopen(_lowercase , '''rb''' , use_auth_token=_lowercase) as f:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
else:
UpperCAmelCase_ , UpperCAmelCase_ = sf.read(_lowercase)
UpperCAmelCase_ = array.T
if self.mono:
UpperCAmelCase_ = librosa.to_mono(_lowercase)
if self.sampling_rate and self.sampling_rate != sampling_rate:
UpperCAmelCase_ = librosa.resample(_lowercase , orig_sr=_lowercase , target_sr=self.sampling_rate)
UpperCAmelCase_ = self.sampling_rate
return {"path": path, "array": array, "sampling_rate": sampling_rate}
def __a ( self :Union[str, Any]) -> Union["FeatureType", Dict[str, "FeatureType"]]:
from .features import Value
if self.decode:
raise ValueError('''Cannot flatten a decoded Audio feature.''')
return {
"bytes": Value('''binary'''),
"path": Value('''string'''),
}
def __a ( self :int , _lowercase :Union[pa.StringArray, pa.StructArray]) -> pa.StructArray:
if pa.types.is_string(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, storage] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_binary(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([storage, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_struct(storage.type) and storage.type.get_all_field_indices('''array'''):
UpperCAmelCase_ = pa.array([Audio().encode_example(_lowercase) if x is not None else None for x in storage.to_pylist()])
elif pa.types.is_struct(storage.type):
if storage.type.get_field_index('''bytes''') >= 0:
UpperCAmelCase_ = storage.field('''bytes''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
if storage.type.get_field_index('''path''') >= 0:
UpperCAmelCase_ = storage.field('''path''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
return array_cast(_lowercase , self.pa_type)
def __a ( self :Any , _lowercase :pa.StructArray) -> pa.StructArray:
@no_op_if_value_is_null
def path_to_bytes(_lowercase :Tuple):
with xopen(_lowercase , '''rb''') as f:
UpperCAmelCase_ = f.read()
return bytes_
UpperCAmelCase_ = pa.array(
[
(path_to_bytes(x['''path''']) if x['''bytes'''] is None else x['''bytes''']) if x is not None else None
for x in storage.to_pylist()
] , type=pa.binary() , )
UpperCAmelCase_ = pa.array(
[os.path.basename(_lowercase) if path is not None else None for path in storage.field('''path''').to_pylist()] , type=pa.string() , )
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
| 344 | 1 |
import argparse
import json
import os
import sys
import tempfile
import unittest
from argparse import Namespace
from dataclasses import dataclass, field
from enum import Enum
from pathlib import Path
from typing import List, Literal, Optional
import yaml
from transformers import HfArgumentParser, TrainingArguments
from transformers.hf_argparser import make_choice_type_function, string_to_bool
# Since Python 3.10, we can use the builtin `|` operator for Union types
# See PEP 604: https://peps.python.org/pep-0604
UpperCamelCase_ = sys.version_info >= (3, 10)
def A ( __UpperCAmelCase=None , __UpperCAmelCase=None ) -> Tuple:
'''simple docstring'''
return field(default_factory=lambda: default , metadata=__UpperCAmelCase )
@dataclass
class a_ :
UpperCamelCase__ : int
UpperCamelCase__ : float
UpperCamelCase__ : str
UpperCamelCase__ : bool
@dataclass
class a_ :
UpperCamelCase__ : int =42
UpperCamelCase__ : str =field(default="toto" , metadata={"help": "help message"} )
@dataclass
class a_ :
UpperCamelCase__ : bool =False
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[bool] =None
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] ="titi"
UpperCamelCase__ : int ="toto"
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="titi"
UpperCamelCase__ : Tuple ="toto"
UpperCamelCase__ : Dict =42
@dataclass
class a_ :
UpperCamelCase__ : BasicEnum ="toto"
def __a ( self :List[Any]) -> int:
UpperCAmelCase_ = BasicEnum(self.foo)
@dataclass
class a_ :
UpperCamelCase__ : MixedTypeEnum ="toto"
def __a ( self :Any) -> Optional[int]:
UpperCAmelCase_ = MixedTypeEnum(self.foo)
@dataclass
class a_ :
UpperCamelCase__ : Optional[int] =None
UpperCamelCase__ : Optional[float] =field(default=_snake_case , metadata={"help": "help message"} )
UpperCamelCase__ : Optional[str] =None
UpperCamelCase__ : Optional[List[str]] =list_field(default=[] )
UpperCamelCase__ : Optional[List[int]] =list_field(default=[] )
@dataclass
class a_ :
UpperCamelCase__ : List[int] =list_field(default=[] )
UpperCamelCase__ : List[int] =list_field(default=[1, 2, 3] )
UpperCamelCase__ : List[str] =list_field(default=["Hallo", "Bonjour", "Hello"] )
UpperCamelCase__ : List[float] =list_field(default=[0.1, 0.2, 0.3] )
@dataclass
class a_ :
UpperCamelCase__ : List[int] =field()
UpperCamelCase__ : str =field()
UpperCamelCase__ : BasicEnum =field()
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = BasicEnum(self.required_enum)
@dataclass
class a_ :
UpperCamelCase__ : int
UpperCamelCase__ : "BasicEnum" =field()
UpperCamelCase__ : "Optional[bool]" =None
UpperCamelCase__ : "str" =field(default="toto" , metadata={"help": "help message"} )
UpperCamelCase__ : "List[str]" =list_field(default=["Hallo", "Bonjour", "Hello"] )
if is_python_no_less_than_3_10:
@dataclass
class a_ :
UpperCamelCase__ : bool =False
UpperCamelCase__ : bool =True
UpperCamelCase__ : bool | None =None
@dataclass
class a_ :
UpperCamelCase__ : int | None =None
UpperCamelCase__ : float | None =field(default=_snake_case , metadata={"help": "help message"} )
UpperCamelCase__ : str | None =None
UpperCamelCase__ : list[str] | None =list_field(default=[] )
UpperCamelCase__ : list[int] | None =list_field(default=[] )
class a_ ( unittest.TestCase ):
def __a ( self :List[str] , _lowercase :argparse.ArgumentParser , _lowercase :argparse.ArgumentParser) -> Optional[int]:
self.assertEqual(len(a._actions) , len(b._actions))
for x, y in zip(a._actions , b._actions):
UpperCAmelCase_ = {k: v for k, v in vars(_lowercase).items() if k != '''container'''}
UpperCAmelCase_ = {k: v for k, v in vars(_lowercase).items() if k != '''container'''}
# Choices with mixed type have custom function as "type"
# So we need to compare results directly for equality
if xx.get('''choices''' , _lowercase) and yy.get('''choices''' , _lowercase):
for expected_choice in yy["choices"] + xx["choices"]:
self.assertEqual(xx['''type'''](_lowercase) , yy['''type'''](_lowercase))
del xx["type"], yy["type"]
self.assertEqual(_lowercase , _lowercase)
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--foo''' , type=_lowercase , required=_lowercase)
expected.add_argument('''--bar''' , type=_lowercase , required=_lowercase)
expected.add_argument('''--baz''' , type=_lowercase , required=_lowercase)
expected.add_argument('''--flag''' , type=_lowercase , default=_lowercase , const=_lowercase , nargs='''?''')
self.argparsersEqual(_lowercase , _lowercase)
UpperCAmelCase_ = ['''--foo''', '''1''', '''--baz''', '''quux''', '''--bar''', '''0.5''']
((UpperCAmelCase_) , ) = parser.parse_args_into_dataclasses(_lowercase , look_for_args_file=_lowercase)
self.assertFalse(example.flag)
def __a ( self :int) -> int:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--foo''' , default=42 , type=_lowercase)
expected.add_argument('''--baz''' , default='''toto''' , type=_lowercase , help='''help message''')
self.argparsersEqual(_lowercase , _lowercase)
def __a ( self :List[Any]) -> Union[str, Any]:
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--foo''' , type=_lowercase , default=_lowercase , const=_lowercase , nargs='''?''')
expected.add_argument('''--baz''' , type=_lowercase , default=_lowercase , const=_lowercase , nargs='''?''')
# A boolean no_* argument always has to come after its "default: True" regular counter-part
# and its default must be set to False
expected.add_argument('''--no_baz''' , action='''store_false''' , default=_lowercase , dest='''baz''')
expected.add_argument('''--opt''' , type=_lowercase , default=_lowercase)
UpperCAmelCase_ = [WithDefaultBoolExample]
if is_python_no_less_than_3_10:
dataclass_types.append(_lowercase)
for dataclass_type in dataclass_types:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
self.argparsersEqual(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_args([])
self.assertEqual(_lowercase , Namespace(foo=_lowercase , baz=_lowercase , opt=_lowercase))
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''--no_baz'''])
self.assertEqual(_lowercase , Namespace(foo=_lowercase , baz=_lowercase , opt=_lowercase))
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''--baz'''])
self.assertEqual(_lowercase , Namespace(foo=_lowercase , baz=_lowercase , opt=_lowercase))
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''True''', '''--baz''', '''True''', '''--opt''', '''True'''])
self.assertEqual(_lowercase , Namespace(foo=_lowercase , baz=_lowercase , opt=_lowercase))
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''False''', '''--baz''', '''False''', '''--opt''', '''False'''])
self.assertEqual(_lowercase , Namespace(foo=_lowercase , baz=_lowercase , opt=_lowercase))
def __a ( self :List[Any]) -> Union[str, Any]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument(
'''--foo''' , default='''toto''' , choices=['''titi''', '''toto''', 42] , type=make_choice_type_function(['''titi''', '''toto''', 42]) , )
self.argparsersEqual(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_args([])
self.assertEqual(args.foo , '''toto''')
UpperCAmelCase_ = parser.parse_args_into_dataclasses([])[0]
self.assertEqual(enum_ex.foo , MixedTypeEnum.toto)
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''titi'''])
self.assertEqual(args.foo , '''titi''')
UpperCAmelCase_ = parser.parse_args_into_dataclasses(['''--foo''', '''titi'''])[0]
self.assertEqual(enum_ex.foo , MixedTypeEnum.titi)
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''42'''])
self.assertEqual(args.foo , 42)
UpperCAmelCase_ = parser.parse_args_into_dataclasses(['''--foo''', '''42'''])[0]
self.assertEqual(enum_ex.foo , MixedTypeEnum.fourtytwo)
def __a ( self :Optional[int]) -> Tuple:
@dataclass
class a_ :
UpperCamelCase__ : Literal["titi", "toto", 42] ="toto"
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument(
'''--foo''' , default='''toto''' , choices=('''titi''', '''toto''', 42) , type=make_choice_type_function(['''titi''', '''toto''', 42]) , )
self.argparsersEqual(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_args([])
self.assertEqual(args.foo , '''toto''')
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''titi'''])
self.assertEqual(args.foo , '''titi''')
UpperCAmelCase_ = parser.parse_args(['''--foo''', '''42'''])
self.assertEqual(args.foo , 42)
def __a ( self :Union[str, Any]) -> str:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--foo_int''' , nargs='''+''' , default=[] , type=_lowercase)
expected.add_argument('''--bar_int''' , nargs='''+''' , default=[1, 2, 3] , type=_lowercase)
expected.add_argument('''--foo_str''' , nargs='''+''' , default=['''Hallo''', '''Bonjour''', '''Hello'''] , type=_lowercase)
expected.add_argument('''--foo_float''' , nargs='''+''' , default=[0.1, 0.2, 0.3] , type=_lowercase)
self.argparsersEqual(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_args([])
self.assertEqual(
_lowercase , Namespace(foo_int=[] , bar_int=[1, 2, 3] , foo_str=['''Hallo''', '''Bonjour''', '''Hello'''] , foo_float=[0.1, 0.2, 0.3]) , )
UpperCAmelCase_ = parser.parse_args('''--foo_int 1 --bar_int 2 3 --foo_str a b c --foo_float 0.1 0.7'''.split())
self.assertEqual(_lowercase , Namespace(foo_int=[1] , bar_int=[2, 3] , foo_str=['''a''', '''b''', '''c'''] , foo_float=[0.1, 0.7]))
def __a ( self :Optional[Any]) -> List[Any]:
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--foo''' , default=_lowercase , type=_lowercase)
expected.add_argument('''--bar''' , default=_lowercase , type=_lowercase , help='''help message''')
expected.add_argument('''--baz''' , default=_lowercase , type=_lowercase)
expected.add_argument('''--ces''' , nargs='''+''' , default=[] , type=_lowercase)
expected.add_argument('''--des''' , nargs='''+''' , default=[] , type=_lowercase)
UpperCAmelCase_ = [OptionalExample]
if is_python_no_less_than_3_10:
dataclass_types.append(_lowercase)
for dataclass_type in dataclass_types:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
self.argparsersEqual(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_args([])
self.assertEqual(_lowercase , Namespace(foo=_lowercase , bar=_lowercase , baz=_lowercase , ces=[] , des=[]))
UpperCAmelCase_ = parser.parse_args('''--foo 12 --bar 3.14 --baz 42 --ces a b c --des 1 2 3'''.split())
self.assertEqual(_lowercase , Namespace(foo=12 , bar=3.14 , baz='''42''' , ces=['''a''', '''b''', '''c'''] , des=[1, 2, 3]))
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--required_list''' , nargs='''+''' , type=_lowercase , required=_lowercase)
expected.add_argument('''--required_str''' , type=_lowercase , required=_lowercase)
expected.add_argument(
'''--required_enum''' , type=make_choice_type_function(['''titi''', '''toto''']) , choices=['''titi''', '''toto'''] , required=_lowercase , )
self.argparsersEqual(_lowercase , _lowercase)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = argparse.ArgumentParser()
expected.add_argument('''--foo''' , type=_lowercase , required=_lowercase)
expected.add_argument(
'''--required_enum''' , type=make_choice_type_function(['''titi''', '''toto''']) , choices=['''titi''', '''toto'''] , required=_lowercase , )
expected.add_argument('''--opt''' , type=_lowercase , default=_lowercase)
expected.add_argument('''--baz''' , default='''toto''' , type=_lowercase , help='''help message''')
expected.add_argument('''--foo_str''' , nargs='''+''' , default=['''Hallo''', '''Bonjour''', '''Hello'''] , type=_lowercase)
self.argparsersEqual(_lowercase , _lowercase)
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = {
'''foo''': 12,
'''bar''': 3.14,
'''baz''': '''42''',
'''flag''': True,
}
UpperCAmelCase_ = parser.parse_dict(_lowercase)[0]
UpperCAmelCase_ = BasicExample(**_lowercase)
self.assertEqual(_lowercase , _lowercase)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = {
'''foo''': 12,
'''bar''': 3.14,
'''baz''': '''42''',
'''flag''': True,
'''extra''': 42,
}
self.assertRaises(_lowercase , parser.parse_dict , _lowercase , allow_extra_keys=_lowercase)
def __a ( self :Optional[Any]) -> Optional[int]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = {
'''foo''': 12,
'''bar''': 3.14,
'''baz''': '''42''',
'''flag''': True,
}
with tempfile.TemporaryDirectory() as tmp_dir:
UpperCAmelCase_ = os.path.join(_lowercase , '''temp_json''')
os.mkdir(_lowercase)
with open(temp_local_path + '''.json''' , '''w+''') as f:
json.dump(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_yaml_file(Path(temp_local_path + '''.json'''))[0]
UpperCAmelCase_ = BasicExample(**_lowercase)
self.assertEqual(_lowercase , _lowercase)
def __a ( self :Optional[Any]) -> Optional[Any]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
UpperCAmelCase_ = {
'''foo''': 12,
'''bar''': 3.14,
'''baz''': '''42''',
'''flag''': True,
}
with tempfile.TemporaryDirectory() as tmp_dir:
UpperCAmelCase_ = os.path.join(_lowercase , '''temp_yaml''')
os.mkdir(_lowercase)
with open(temp_local_path + '''.yaml''' , '''w+''') as f:
yaml.dump(_lowercase , _lowercase)
UpperCAmelCase_ = parser.parse_yaml_file(Path(temp_local_path + '''.yaml'''))[0]
UpperCAmelCase_ = BasicExample(**_lowercase)
self.assertEqual(_lowercase , _lowercase)
def __a ( self :Optional[int]) -> List[str]:
UpperCAmelCase_ = HfArgumentParser(_lowercase)
self.assertIsNotNone(_lowercase)
| 344 |
from ..utils import DummyObject, requires_backends
class a_ ( metaclass=_snake_case ):
UpperCamelCase__ : Any =["torch", "scipy"]
def __init__( self :List[str] , *_lowercase :List[str] , **_lowercase :Union[str, Any]) -> List[Any]:
requires_backends(self , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Dict , *_lowercase :Any , **_lowercase :Dict) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
@classmethod
def __a ( cls :Optional[Any] , *_lowercase :str , **_lowercase :Optional[Any]) -> Union[str, Any]:
requires_backends(cls , ['''torch''', '''scipy'''])
| 344 | 1 |
# We ignore warnings about stepping the scheduler since we step it ourselves during gradient accumulation
import warnings
from .state import AcceleratorState, GradientState
warnings.filterwarnings("ignore", category=UserWarning, module="torch.optim.lr_scheduler")
class a_ :
def __init__( self :int , _lowercase :List[str] , _lowercase :Union[str, Any] , _lowercase :bool = True , _lowercase :bool = False) -> Any:
UpperCAmelCase_ = scheduler
UpperCAmelCase_ = optimizers if isinstance(_lowercase , (list, tuple)) else [optimizers]
UpperCAmelCase_ = split_batches
UpperCAmelCase_ = step_with_optimizer
UpperCAmelCase_ = GradientState()
def __a ( self :str , *_lowercase :Optional[int] , **_lowercase :Union[str, Any]) -> List[str]:
if not self.step_with_optimizer:
# No link between scheduler and optimizer -> just step
self.scheduler.step(*_lowercase , **_lowercase)
return
# Otherwise, first make sure the optimizer was stepped.
if not self.gradient_state.sync_gradients:
if self.gradient_state.adjust_scheduler:
self.scheduler._step_count += 1
return
for opt in self.optimizers:
if opt.step_was_skipped:
return
if self.split_batches:
# Split batches -> the training dataloader batch size is not changed so one step per training step
self.scheduler.step(*_lowercase , **_lowercase)
else:
# Otherwise the training dataloader batch size was multiplied by `num_processes`, so we need to do
# num_processes steps per training step
UpperCAmelCase_ = AcceleratorState().num_processes
for _ in range(_lowercase):
# Special case when using OneCycle and `drop_last` was not used
if hasattr(self.scheduler , '''total_steps'''):
if self.scheduler._step_count <= self.scheduler.total_steps:
self.scheduler.step(*_lowercase , **_lowercase)
else:
self.scheduler.step(*_lowercase , **_lowercase)
def __a ( self :Dict) -> Dict:
return self.scheduler.get_last_lr()
def __a ( self :str) -> List[str]:
return self.scheduler.state_dict()
def __a ( self :Any , _lowercase :int) -> int:
self.scheduler.load_state_dict(_lowercase)
def __a ( self :str) -> Tuple:
return self.scheduler.get_lr()
def __a ( self :Optional[int] , *_lowercase :Optional[int] , **_lowercase :Union[str, Any]) -> List[str]:
return self.scheduler.print_lr(*_lowercase , **_lowercase)
| 344 |
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def A ( __UpperCAmelCase ) -> Dict[str, torch.Tensor]:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for rt in rc.restypes:
UpperCAmelCase_ = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
UpperCAmelCase_ = {name: i for i, name in enumerate(__UpperCAmelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.intaa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = torch.tensor(
__UpperCAmelCase , dtype=torch.floataa , device=protein['''aatype'''].device , )
UpperCAmelCase_ = protein['''aatype'''].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
UpperCAmelCase_ = restype_atomaa_to_atomaa[protein_aatype]
UpperCAmelCase_ = residx_atomaa_to_atomaa.long()
# create the corresponding mask
UpperCAmelCase_ = torch.zeros([21, 37] , dtype=torch.floataa , device=protein['''aatype'''].device )
for restype, restype_letter in enumerate(rc.restypes ):
UpperCAmelCase_ = rc.restype_atoa[restype_letter]
UpperCAmelCase_ = rc.residue_atoms[restype_name]
for atom_name in atom_names:
UpperCAmelCase_ = rc.atom_order[atom_name]
UpperCAmelCase_ = 1
UpperCAmelCase_ = restype_atomaa_mask[protein_aatype]
UpperCAmelCase_ = residx_atomaa_mask
return protein
def A ( __UpperCAmelCase ) -> Dict[str, np.ndarray]:
'''simple docstring'''
UpperCAmelCase_ = tree_map(lambda __UpperCAmelCase : torch.tensor(__UpperCAmelCase , device=batch['''aatype'''].device ) , __UpperCAmelCase , np.ndarray )
UpperCAmelCase_ = tensor_tree_map(lambda __UpperCAmelCase : np.array(__UpperCAmelCase ) , make_atomaa_masks(__UpperCAmelCase ) )
return out
| 344 | 1 |
import copy
import os
from typing import Union
from ...configuration_utils import PretrainedConfig
from ...models.auto.modeling_auto import MODEL_FOR_CAUSAL_LM_MAPPING_NAMES
from ...utils import logging
from ..auto import CONFIG_MAPPING
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"salesforce/blip2-opt-2.7b": "https://huggingface.co/salesforce/blip2-opt-2.7b/resolve/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : int ="blip_2_vision_model"
def __init__( self :str , _lowercase :int=1408 , _lowercase :Union[str, Any]=6144 , _lowercase :str=39 , _lowercase :str=16 , _lowercase :int=224 , _lowercase :List[str]=14 , _lowercase :List[Any]="gelu" , _lowercase :str=0.00_001 , _lowercase :Optional[Any]=0.0 , _lowercase :Optional[int]=1E-1_0 , _lowercase :str=True , **_lowercase :Tuple , ) -> List[str]:
super().__init__(**_lowercase)
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = image_size
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = attention_dropout
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = qkv_bias
@classmethod
def __a ( cls :List[str] , _lowercase :Union[str, os.PathLike] , **_lowercase :Dict) -> "PretrainedConfig":
cls._set_token_in_kwargs(_lowercase)
UpperCAmelCase_ , UpperCAmelCase_ = cls.get_config_dict(_lowercase , **_lowercase)
# get the vision config dict if we are loading from Blip2Config
if config_dict.get('''model_type''') == "blip-2":
UpperCAmelCase_ = config_dict['''vision_config''']
if "model_type" in config_dict and hasattr(cls , '''model_type''') and config_dict["model_type"] != cls.model_type:
logger.warning(
f"You are using a model of type {config_dict['model_type']} to instantiate a model of type "
f"{cls.model_type}. This is not supported for all configurations of models and can yield errors.")
return cls.from_dict(_lowercase , **_lowercase)
class a_ ( _snake_case ):
UpperCamelCase__ : int ="blip_2_qformer"
def __init__( self :Optional[int] , _lowercase :str=30522 , _lowercase :Tuple=768 , _lowercase :Optional[int]=12 , _lowercase :List[Any]=12 , _lowercase :List[str]=3072 , _lowercase :Any="gelu" , _lowercase :Optional[int]=0.1 , _lowercase :Any=0.1 , _lowercase :Any=512 , _lowercase :List[str]=0.02 , _lowercase :List[Any]=1E-1_2 , _lowercase :List[Any]=0 , _lowercase :List[Any]="absolute" , _lowercase :List[str]=2 , _lowercase :Optional[Any]=1408 , **_lowercase :Optional[Any] , ) -> Union[str, Any]:
super().__init__(pad_token_id=_lowercase , **_lowercase)
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_probs_dropout_prob
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = position_embedding_type
UpperCAmelCase_ = cross_attention_frequency
UpperCAmelCase_ = encoder_hidden_size
@classmethod
def __a ( cls :List[str] , _lowercase :Union[str, os.PathLike] , **_lowercase :Any) -> "PretrainedConfig":
cls._set_token_in_kwargs(_lowercase)
UpperCAmelCase_ , UpperCAmelCase_ = cls.get_config_dict(_lowercase , **_lowercase)
# get the qformer config dict if we are loading from Blip2Config
if config_dict.get('''model_type''') == "blip-2":
UpperCAmelCase_ = config_dict['''qformer_config''']
if "model_type" in config_dict and hasattr(cls , '''model_type''') and config_dict["model_type"] != cls.model_type:
logger.warning(
f"You are using a model of type {config_dict['model_type']} to instantiate a model of type "
f"{cls.model_type}. This is not supported for all configurations of models and can yield errors.")
return cls.from_dict(_lowercase , **_lowercase)
class a_ ( _snake_case ):
UpperCamelCase__ : List[Any] ="blip-2"
UpperCamelCase__ : List[str] =True
def __init__( self :Optional[Any] , _lowercase :Optional[int]=None , _lowercase :List[str]=None , _lowercase :Tuple=None , _lowercase :Any=32 , **_lowercase :Dict) -> Any:
super().__init__(**_lowercase)
if vision_config is None:
UpperCAmelCase_ = {}
logger.info('''vision_config is None. initializing the Blip2VisionConfig with default values.''')
if qformer_config is None:
UpperCAmelCase_ = {}
logger.info('''qformer_config is None. Initializing the Blip2QFormerConfig with default values.''')
if text_config is None:
UpperCAmelCase_ = {}
logger.info('''text_config is None. Initializing the text config with default values (`OPTConfig`).''')
UpperCAmelCase_ = BlipaVisionConfig(**_lowercase)
UpperCAmelCase_ = BlipaQFormerConfig(**_lowercase)
UpperCAmelCase_ = text_config['''model_type'''] if '''model_type''' in text_config else '''opt'''
UpperCAmelCase_ = CONFIG_MAPPING[text_model_type](**_lowercase)
UpperCAmelCase_ = self.text_config.tie_word_embeddings
UpperCAmelCase_ = self.text_config.is_encoder_decoder
UpperCAmelCase_ = num_query_tokens
UpperCAmelCase_ = self.vision_config.hidden_size
UpperCAmelCase_ = self.text_config.model_type in MODEL_FOR_CAUSAL_LM_MAPPING_NAMES
UpperCAmelCase_ = 1.0
UpperCAmelCase_ = 0.02
@classmethod
def __a ( cls :str , _lowercase :BlipaVisionConfig , _lowercase :BlipaQFormerConfig , _lowercase :PretrainedConfig , **_lowercase :Union[str, Any] , ) -> Dict:
return cls(
vision_config=vision_config.to_dict() , qformer_config=qformer_config.to_dict() , text_config=text_config.to_dict() , **_lowercase , )
def __a ( self :Dict) -> List[Any]:
UpperCAmelCase_ = copy.deepcopy(self.__dict__)
UpperCAmelCase_ = self.vision_config.to_dict()
UpperCAmelCase_ = self.qformer_config.to_dict()
UpperCAmelCase_ = self.text_config.to_dict()
UpperCAmelCase_ = self.__class__.model_type
return output
| 344 |
# Copyright 2023 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from ..models.whisper import WhisperForConditionalGeneration, WhisperProcessor
from .base import PipelineTool
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="openai/whisper-base"
UpperCamelCase__ : int =(
"This is a tool that transcribes an audio into text. It takes an input named `audio` and returns the "
"transcribed text."
)
UpperCamelCase__ : Any ="transcriber"
UpperCamelCase__ : Optional[int] =WhisperProcessor
UpperCamelCase__ : List[str] =WhisperForConditionalGeneration
UpperCamelCase__ : List[Any] =["audio"]
UpperCamelCase__ : Union[str, Any] =["text"]
def __a ( self :int , _lowercase :Any) -> Tuple:
return self.pre_processor(_lowercase , return_tensors='''pt''').input_features
def __a ( self :Dict , _lowercase :Tuple) -> Any:
return self.model.generate(inputs=_lowercase)
def __a ( self :int , _lowercase :Union[str, Any]) -> Optional[Any]:
return self.pre_processor.batch_decode(_lowercase , skip_special_tokens=_lowercase)[0]
| 344 | 1 |
from sympy import diff, lambdify, symbols
from sympy.functions import * # noqa: F403
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase = "x" , __UpperCAmelCase = 10**-10 , __UpperCAmelCase = 1 , ) -> complex:
'''simple docstring'''
UpperCAmelCase_ = symbols(__UpperCAmelCase )
UpperCAmelCase_ = lambdify(__UpperCAmelCase , __UpperCAmelCase )
UpperCAmelCase_ = lambdify(__UpperCAmelCase , diff(__UpperCAmelCase , __UpperCAmelCase ) )
UpperCAmelCase_ = starting_point
while True:
if diff_function(__UpperCAmelCase ) != 0:
UpperCAmelCase_ = prev_guess - multiplicity * func(__UpperCAmelCase ) / diff_function(
__UpperCAmelCase )
else:
raise ZeroDivisionError('''Could not find root''' ) from None
# Precision is checked by comparing the difference of consecutive guesses
if abs(next_guess - prev_guess ) < precision:
return next_guess
UpperCAmelCase_ = next_guess
# Let's Execute
if __name__ == "__main__":
# Find root of trigonometric function
# Find value of pi
print(f"The root of sin(x) = 0 is {newton_raphson('sin(x)', 2)}")
# Find root of polynomial
# Find fourth Root of 5
print(f"The root of x**4 - 5 = 0 is {newton_raphson('x**4 -5', 0.4 +5J)}")
# Find value of e
print(
"The root of log(y) - 1 = 0 is ",
f"{newton_raphson('log(y) - 1', 2, variable='y')}",
)
# Exponential Roots
print(
"The root of exp(x) - 1 = 0 is",
f"{newton_raphson('exp(x) - 1', 10, precision=0.0_05)}",
)
# Find root of cos(x)
print(f"The root of cos(x) = 0 is {newton_raphson('cos(x)', 0)}")
| 344 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
UpperCamelCase_ = {"configuration_opt": ["OPT_PRETRAINED_CONFIG_ARCHIVE_MAP", "OPTConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"OPT_PRETRAINED_MODEL_ARCHIVE_LIST",
"OPTForCausalLM",
"OPTModel",
"OPTPreTrainedModel",
"OPTForSequenceClassification",
"OPTForQuestionAnswering",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = ["TFOPTForCausalLM", "TFOPTModel", "TFOPTPreTrainedModel"]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
UpperCamelCase_ = [
"FlaxOPTForCausalLM",
"FlaxOPTModel",
"FlaxOPTPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_opt import OPT_PRETRAINED_CONFIG_ARCHIVE_MAP, OPTConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_opt import (
OPT_PRETRAINED_MODEL_ARCHIVE_LIST,
OPTForCausalLM,
OPTForQuestionAnswering,
OPTForSequenceClassification,
OPTModel,
OPTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_opt import TFOPTForCausalLM, TFOPTModel, TFOPTPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_opt import FlaxOPTForCausalLM, FlaxOPTModel, FlaxOPTPreTrainedModel
else:
import sys
UpperCamelCase_ = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 344 | 1 |
def A ( __UpperCAmelCase = 100 ) -> int:
'''simple docstring'''
UpperCAmelCase_ = n * (n + 1) * (2 * n + 1) / 6
UpperCAmelCase_ = (n * (n + 1) / 2) ** 2
return int(square_of_sum - sum_of_squares )
if __name__ == "__main__":
print(f"{solution() = }")
| 344 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
UpperCamelCase_ = "▁"
UpperCamelCase_ = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class a_ ( _snake_case , unittest.TestCase ):
UpperCamelCase__ : str =BigBirdTokenizer
UpperCamelCase__ : Tuple =BigBirdTokenizerFast
UpperCamelCase__ : Union[str, Any] =True
UpperCamelCase__ : List[str] =True
def __a ( self :Any) -> List[str]:
super().setUp()
UpperCAmelCase_ = self.tokenizer_class(_lowercase , keep_accents=_lowercase)
tokenizer.save_pretrained(self.tmpdirname)
def __a ( self :Optional[int]) -> str:
UpperCAmelCase_ = '''<s>'''
UpperCAmelCase_ = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_lowercase) , _lowercase)
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_lowercase) , _lowercase)
def __a ( self :str) -> str:
UpperCAmelCase_ = list(self.get_tokenizer().get_vocab().keys())
self.assertEqual(vocab_keys[0] , '''<unk>''')
self.assertEqual(vocab_keys[1] , '''<s>''')
self.assertEqual(vocab_keys[-1] , '''[MASK]''')
self.assertEqual(len(_lowercase) , 1004)
def __a ( self :List[str]) -> int:
self.assertEqual(self.get_tokenizer().vocab_size , 1000)
def __a ( self :Tuple) -> int:
if not self.test_rust_tokenizer:
return
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = '''I was born in 92000, and this is falsé.'''
UpperCAmelCase_ = tokenizer.tokenize(_lowercase)
UpperCAmelCase_ = rust_tokenizer.tokenize(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase , add_special_tokens=_lowercase)
self.assertListEqual(_lowercase , _lowercase)
UpperCAmelCase_ = self.get_rust_tokenizer()
UpperCAmelCase_ = tokenizer.encode(_lowercase)
UpperCAmelCase_ = rust_tokenizer.encode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :Optional[Any]) -> List[str]:
UpperCAmelCase_ = BigBirdTokenizer(_lowercase , keep_accents=_lowercase)
UpperCAmelCase_ = tokenizer.tokenize('''This is a test''')
self.assertListEqual(_lowercase , ['''▁This''', '''▁is''', '''▁a''', '''▁t''', '''est'''])
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_lowercase) , [285, 46, 10, 170, 382] , )
UpperCAmelCase_ = tokenizer.tokenize('''I was born in 92000, and this is falsé.''')
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''9''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''é''',
'''.''',
] , )
UpperCAmelCase_ = tokenizer.convert_tokens_to_ids(_lowercase)
self.assertListEqual(
_lowercase , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
UpperCAmelCase_ = tokenizer.convert_ids_to_tokens(_lowercase)
self.assertListEqual(
_lowercase , [
SPIECE_UNDERLINE + '''I''',
SPIECE_UNDERLINE + '''was''',
SPIECE_UNDERLINE + '''b''',
'''or''',
'''n''',
SPIECE_UNDERLINE + '''in''',
SPIECE_UNDERLINE + '''''',
'''<unk>''',
'''2''',
'''0''',
'''0''',
'''0''',
''',''',
SPIECE_UNDERLINE + '''and''',
SPIECE_UNDERLINE + '''this''',
SPIECE_UNDERLINE + '''is''',
SPIECE_UNDERLINE + '''f''',
'''al''',
'''s''',
'''<unk>''',
'''.''',
] , )
@cached_property
def __a ( self :Any) -> List[Any]:
return BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
@slow
def __a ( self :int) -> List[Any]:
UpperCAmelCase_ = '''Hello World!'''
UpperCAmelCase_ = [65, 18536, 2260, 101, 66]
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@slow
def __a ( self :int) -> Any:
UpperCAmelCase_ = (
'''This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) " [ ] ! : - . Also we will'''
''' add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth'''
)
# fmt: off
UpperCAmelCase_ = [65, 871, 419, 358, 946, 991, 2521, 452, 358, 1357, 387, 7751, 3536, 112, 985, 456, 126, 865, 938, 5400, 5734, 458, 1368, 467, 786, 2462, 5246, 1159, 633, 865, 4519, 457, 582, 852, 2557, 427, 916, 508, 405, 34324, 497, 391, 408, 11342, 1244, 385, 100, 938, 985, 456, 574, 362, 12597, 3200, 3129, 1172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_lowercase , self.big_tokenizer.encode(_lowercase))
@require_torch
@slow
def __a ( self :Dict) -> Union[str, Any]:
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
UpperCAmelCase_ = list(self.big_tokenizer.get_vocab().keys())[:10]
UpperCAmelCase_ = ''' '''.join(_lowercase)
UpperCAmelCase_ = self.big_tokenizer.encode_plus(_lowercase , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = self.big_tokenizer.batch_encode_plus(
[sequence + ''' ''' + sequence] , return_tensors='''pt''' , return_token_type_ids=_lowercase)
UpperCAmelCase_ = BigBirdConfig(attention_type='''original_full''')
UpperCAmelCase_ = BigBirdModel(_lowercase)
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_lowercase)
model(**_lowercase)
@slow
def __a ( self :Optional[int]) -> Any:
UpperCAmelCase_ = BigBirdTokenizer.from_pretrained('''google/bigbird-roberta-base''')
UpperCAmelCase_ = tokenizer.decode(tokenizer('''Paris is the [MASK].''').input_ids)
self.assertTrue(decoded_text == '''[CLS] Paris is the[MASK].[SEP]''')
@slow
def __a ( self :Dict) -> List[str]:
# fmt: off
UpperCAmelCase_ = {'''input_ids''': [[65, 39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114, 66], [65, 448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], '''attention_mask''': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_lowercase , model_name='''google/bigbird-roberta-base''' , revision='''215c99f1600e06f83acce68422f2035b2b5c3510''' , )
| 344 | 1 |
from manim import *
class a_ ( _snake_case ):
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = Rectangle(height=0.5 , width=0.5)
UpperCAmelCase_ = Rectangle(height=0.46 , width=0.46).set_stroke(width=0)
UpperCAmelCase_ = [mem.copy() for i in range(6)]
UpperCAmelCase_ = [mem.copy() for i in range(6)]
UpperCAmelCase_ = VGroup(*_lowercase).arrange(_lowercase , buff=0)
UpperCAmelCase_ = VGroup(*_lowercase).arrange(_lowercase , buff=0)
UpperCAmelCase_ = VGroup(_lowercase , _lowercase).arrange(_lowercase , buff=0)
UpperCAmelCase_ = Text('''CPU''' , font_size=24)
UpperCAmelCase_ = Group(_lowercase , _lowercase).arrange(_lowercase , buff=0.5 , aligned_edge=_lowercase)
cpu.move_to([-2.5, -0.5, 0])
self.add(_lowercase)
UpperCAmelCase_ = [mem.copy() for i in range(4)]
UpperCAmelCase_ = VGroup(*_lowercase).arrange(_lowercase , buff=0)
UpperCAmelCase_ = Text('''GPU''' , font_size=24)
UpperCAmelCase_ = Group(_lowercase , _lowercase).arrange(_lowercase , buff=0.5 , aligned_edge=_lowercase)
gpu.move_to([-1, -1, 0])
self.add(_lowercase)
UpperCAmelCase_ = [mem.copy() for i in range(6)]
UpperCAmelCase_ = VGroup(*_lowercase).arrange(_lowercase , buff=0)
UpperCAmelCase_ = Text('''Model''' , font_size=24)
UpperCAmelCase_ = Group(_lowercase , _lowercase).arrange(_lowercase , buff=0.5 , aligned_edge=_lowercase)
model.move_to([3, -1.0, 0])
self.add(_lowercase)
UpperCAmelCase_ = []
for i, rect in enumerate(_lowercase):
rect.set_stroke(_lowercase)
# target = fill.copy().set_fill(YELLOW, opacity=0.7)
# target.move_to(rect)
# self.add(target)
UpperCAmelCase_ = Rectangle(height=0.46 / 4 , width=0.46 / 3).set_stroke(width=0.0).set_fill(_lowercase , opacity=0.7)
if i == 0:
cpu_target.next_to(cpu_left_col_base[0].get_corner(DOWN + LEFT) , buff=0.02 , direction=_lowercase)
cpu_target.set_x(cpu_target.get_x() + 0.1)
elif i == 3:
cpu_target.next_to(cpu_targs[0] , direction=_lowercase , buff=0.0)
else:
cpu_target.next_to(cpu_targs[i - 1] , direction=_lowercase , buff=0.0)
self.add(_lowercase)
cpu_targs.append(_lowercase)
UpperCAmelCase_ = [mem.copy() for i in range(6)]
UpperCAmelCase_ = VGroup(*_lowercase).arrange(_lowercase , buff=0)
UpperCAmelCase_ = Text('''Loaded Checkpoint''' , font_size=24)
UpperCAmelCase_ = Group(_lowercase , _lowercase).arrange(_lowercase , aligned_edge=_lowercase , buff=0.4)
checkpoint.move_to([3, 0.5, 0])
UpperCAmelCase_ = Square(side_length=2.2)
key.move_to([-5, 2, 0])
UpperCAmelCase_ = MarkupText(
f"<b>Key:</b>\n\n<span fgcolor='{YELLOW}'>●</span> Empty Model" , font_size=18 , )
key_text.move_to([-5, 2.4, 0])
self.add(_lowercase , _lowercase)
UpperCAmelCase_ = MarkupText(
f"<span fgcolor='{BLUE}'>●</span> Checkpoint" , font_size=18 , )
blue_text.next_to(_lowercase , DOWN * 2.4 , aligned_edge=key_text.get_left())
UpperCAmelCase_ = MarkupText(
f"Next, a <i><span fgcolor=\"{BLUE}\">second</span></i> model is loaded into memory,\nwith the weights of a <span fgcolor=\"{BLUE}\">single shard</span>." , font_size=24 , )
step_a.move_to([2, 2, 0])
self.play(Write(_lowercase) , Write(_lowercase))
self.play(Write(_lowercase , run_time=1) , Create(_lowercase , run_time=1))
UpperCAmelCase_ = []
UpperCAmelCase_ = []
for i, rect in enumerate(_lowercase):
UpperCAmelCase_ = fill.copy().set_fill(_lowercase , opacity=0.7)
target.move_to(_lowercase)
first_animations.append(GrowFromCenter(_lowercase , run_time=1))
UpperCAmelCase_ = target.copy()
cpu_target.generate_target()
if i < 5:
cpu_target.target.move_to(cpu_left_col_base[i + 1])
else:
cpu_target.target.move_to(cpu_right_col_base[i - 5])
second_animations.append(MoveToTarget(_lowercase , run_time=1.5))
self.play(*_lowercase)
self.play(*_lowercase)
self.wait()
| 344 |
from __future__ import annotations
import string
from itertools import cycle, product
from pathlib import Path
UpperCamelCase_ = (
string.ascii_letters + string.digits + string.punctuation + string.whitespace
)
UpperCamelCase_ = [ord(letter) for letter in string.ascii_lowercase]
UpperCamelCase_ = {ord(char) for char in VALID_CHARS}
UpperCamelCase_ = ["the", "be", "to", "of", "and", "in", "that", "have"]
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> str | None:
'''simple docstring'''
UpperCAmelCase_ = ""
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
for keychar, cipherchar in zip(cycle(__UpperCAmelCase ) , __UpperCAmelCase ):
UpperCAmelCase_ = cipherchar ^ keychar
if decodedchar not in VALID_INTS:
return None
decoded += chr(__UpperCAmelCase )
return decoded
def A ( __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
UpperCAmelCase_ = []
for key in product(__UpperCAmelCase , repeat=3 ):
UpperCAmelCase_ = try_key(__UpperCAmelCase , __UpperCAmelCase )
if encoded is not None:
possibles.append(__UpperCAmelCase )
return possibles
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> list[str]:
'''simple docstring'''
return [possible for possible in possibles if common_word in possible.lower()]
def A ( __UpperCAmelCase = "p059_cipher.txt" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = 42
UpperCAmelCase_ = Path(__UpperCAmelCase ).parent.joinpath(__UpperCAmelCase ).read_text(encoding='''utf-8''' )
UpperCAmelCase_ = [int(__UpperCAmelCase ) for number in data.strip().split(''',''' )]
UpperCAmelCase_ = filter_valid_chars(__UpperCAmelCase )
for common_word in COMMON_WORDS:
UpperCAmelCase_ = filter_common_word(__UpperCAmelCase , __UpperCAmelCase )
if len(__UpperCAmelCase ) == 1:
break
UpperCAmelCase_ = possibles[0]
return sum(ord(__UpperCAmelCase ) for char in decoded_text )
if __name__ == "__main__":
print(f"{solution() = }")
| 344 | 1 |
import timeit
import numpy as np
import datasets
from datasets.arrow_writer import ArrowWriter
from datasets.features.features import _ArrayXD
def A ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
def wrapper(*__UpperCAmelCase , **__UpperCAmelCase ):
UpperCAmelCase_ = timeit.default_timer()
UpperCAmelCase_ = func(*__UpperCAmelCase , **__UpperCAmelCase )
UpperCAmelCase_ = timeit.default_timer() - starttime
return delta
UpperCAmelCase_ = func.__name__
return wrapper
def A ( __UpperCAmelCase , __UpperCAmelCase=100 , __UpperCAmelCase=None ) -> Tuple:
'''simple docstring'''
UpperCAmelCase_ = []
UpperCAmelCase_ = seq_shapes or {}
for i in range(__UpperCAmelCase ):
UpperCAmelCase_ = {}
for col_id, (k, v) in enumerate(features.items() ):
if isinstance(__UpperCAmelCase , _ArrayXD ):
UpperCAmelCase_ = np.random.rand(*v.shape ).astype(v.dtype )
elif isinstance(__UpperCAmelCase , datasets.Value ):
if v.dtype == "string":
UpperCAmelCase_ = '''The small grey turtle was surprisingly fast when challenged.'''
else:
UpperCAmelCase_ = np.random.randint(10 , size=1 ).astype(v.dtype ).item()
elif isinstance(__UpperCAmelCase , datasets.Sequence ):
while isinstance(__UpperCAmelCase , datasets.Sequence ):
UpperCAmelCase_ = v.feature
UpperCAmelCase_ = seq_shapes[k]
UpperCAmelCase_ = np.random.rand(*__UpperCAmelCase ).astype(v.dtype )
UpperCAmelCase_ = data
dummy_data.append((i, example) )
return dummy_data
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase=100 , __UpperCAmelCase=None ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = generate_examples(__UpperCAmelCase , num_examples=__UpperCAmelCase , seq_shapes=__UpperCAmelCase )
with ArrowWriter(features=__UpperCAmelCase , path=__UpperCAmelCase ) as writer:
for key, record in dummy_data:
UpperCAmelCase_ = features.encode_example(__UpperCAmelCase )
writer.write(__UpperCAmelCase )
UpperCAmelCase_ , UpperCAmelCase_ = writer.finalize()
if not num_final_examples == num_examples:
raise ValueError(
f"Error writing the dataset, wrote {num_final_examples} examples but should have written {num_examples}." )
UpperCAmelCase_ = datasets.Dataset.from_file(filename=__UpperCAmelCase , info=datasets.DatasetInfo(features=__UpperCAmelCase ) )
return dataset
| 344 |
import pytest
UpperCamelCase_ = "__dummy_dataset1__"
UpperCamelCase_ = "\nimport json\nimport os\n\nimport datasets\n\n\nREPO_URL = \"https://huggingface.co/datasets/albertvillanova/tests-raw-jsonl/resolve/main/\"\nURLS = {\"train\": REPO_URL + \"wikiann-bn-train.jsonl\", \"validation\": REPO_URL + \"wikiann-bn-validation.jsonl\"}\n\n\nclass __DummyDataset1__(datasets.GeneratorBasedBuilder):\n\n def _info(self):\n features = datasets.Features(\n {\n \"tokens\": datasets.Sequence(datasets.Value(\"string\")),\n \"ner_tags\": datasets.Sequence(\n datasets.features.ClassLabel(\n names=[\n \"O\",\n \"B-PER\",\n \"I-PER\",\n \"B-ORG\",\n \"I-ORG\",\n \"B-LOC\",\n \"I-LOC\",\n ]\n )\n ),\n \"langs\": datasets.Sequence(datasets.Value(\"string\")),\n \"spans\": datasets.Sequence(datasets.Value(\"string\")),\n }\n )\n return datasets.DatasetInfo(features=features)\n\n def _split_generators(self, dl_manager):\n dl_path = dl_manager.download(URLS)\n return [\n datasets.SplitGenerator(datasets.Split.TRAIN, gen_kwargs={\"filepath\": dl_path[\"train\"]}),\n datasets.SplitGenerator(datasets.Split.VALIDATION, gen_kwargs={\"filepath\": dl_path[\"validation\"]}),\n ]\n\n def _generate_examples(self, filepath):\n with open(filepath, \"r\", encoding=\"utf-8\") as f:\n for i, line in enumerate(f):\n yield i, json.loads(line)\n"
@pytest.fixture
def A ( ) -> str:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_NAME
@pytest.fixture
def A ( ) -> Any:
'''simple docstring'''
return DATASET_LOADING_SCRIPT_CODE
@pytest.fixture
def A ( __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = dataset_loading_script_name
UpperCAmelCase_ = tmp_path / '''datasets''' / script_name
script_dir.mkdir(parents=__UpperCAmelCase )
UpperCAmelCase_ = script_dir / f"{script_name}.py"
with open(__UpperCAmelCase , '''w''' ) as f:
f.write(__UpperCAmelCase )
return str(__UpperCAmelCase )
| 344 | 1 |
def A ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
return number & 1 == 0
if __name__ == "__main__":
import doctest
doctest.testmod()
| 344 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"s-JoL/Open-Llama-V1": "https://huggingface.co/s-JoL/Open-Llama-V1/blob/main/config.json",
}
class a_ ( _snake_case ):
UpperCamelCase__ : Dict ="open-llama"
def __init__( self :Union[str, Any] , _lowercase :List[Any]=100000 , _lowercase :Dict=4096 , _lowercase :List[Any]=11008 , _lowercase :Optional[int]=32 , _lowercase :Union[str, Any]=32 , _lowercase :List[str]="silu" , _lowercase :Union[str, Any]=2048 , _lowercase :Any=0.02 , _lowercase :Optional[Any]=1E-6 , _lowercase :str=True , _lowercase :str=0 , _lowercase :Any=1 , _lowercase :Optional[Any]=2 , _lowercase :str=False , _lowercase :Dict=True , _lowercase :Optional[Any]=0.1 , _lowercase :Tuple=0.1 , _lowercase :Dict=True , _lowercase :List[Any]=True , _lowercase :Dict=None , **_lowercase :Optional[int] , ) -> List[Any]:
UpperCAmelCase_ = vocab_size
UpperCAmelCase_ = max_position_embeddings
UpperCAmelCase_ = hidden_size
UpperCAmelCase_ = intermediate_size
UpperCAmelCase_ = num_hidden_layers
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = rms_norm_eps
UpperCAmelCase_ = use_cache
UpperCAmelCase_ = kwargs.pop(
'''use_memorry_efficient_attention''' , _lowercase)
UpperCAmelCase_ = hidden_dropout_prob
UpperCAmelCase_ = attention_dropout_prob
UpperCAmelCase_ = use_stable_embedding
UpperCAmelCase_ = shared_input_output_embedding
UpperCAmelCase_ = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_lowercase , bos_token_id=_lowercase , eos_token_id=_lowercase , tie_word_embeddings=_lowercase , **_lowercase , )
def __a ( self :int) -> str:
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _lowercase) or len(self.rope_scaling) != 2:
raise ValueError(
'''`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, '''
f"got {self.rope_scaling}")
UpperCAmelCase_ = self.rope_scaling.get('''type''' , _lowercase)
UpperCAmelCase_ = self.rope_scaling.get('''factor''' , _lowercase)
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}")
if rope_scaling_factor is None or not isinstance(_lowercase , _lowercase) or rope_scaling_factor <= 1.0:
raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
| 344 | 1 |
import os
import sys
import warnings
from dataclasses import dataclass, field
from io import BytesIO
from typing import TYPE_CHECKING, Any, ClassVar, Dict, List, Optional, Union
import numpy as np
import pyarrow as pa
from .. import config
from ..download.streaming_download_manager import xopen
from ..table import array_cast
from ..utils.file_utils import is_local_path
from ..utils.py_utils import first_non_null_value, no_op_if_value_is_null, string_to_dict
if TYPE_CHECKING:
import PIL.Image
from .features import FeatureType
UpperCamelCase_ = None
UpperCamelCase_ = "<" if sys.byteorder == "little" else ">"
# Origin: https://github.com/python-pillow/Pillow/blob/698951e19e19972aeed56df686868f1329981c12/src/PIL/Image.py#L3126 minus "|i1" which values are not preserved correctly when saving and loading an image
UpperCamelCase_ = [
np.dtype("|b1"),
np.dtype("|u1"),
np.dtype("<u2"),
np.dtype(">u2"),
np.dtype("<i2"),
np.dtype(">i2"),
np.dtype("<u4"),
np.dtype(">u4"),
np.dtype("<i4"),
np.dtype(">i4"),
np.dtype("<f4"),
np.dtype(">f4"),
np.dtype("<f8"),
np.dtype(">f8"),
]
@dataclass
class a_ :
UpperCamelCase__ : bool =True
UpperCamelCase__ : Optional[str] =None
# Automatically constructed
UpperCamelCase__ : ClassVar[str] ="PIL.Image.Image"
UpperCamelCase__ : ClassVar[Any] =pa.struct({"bytes": pa.binary(), "path": pa.string()} )
UpperCamelCase__ : str =field(default="Image" , init=_snake_case , repr=_snake_case )
def __call__( self :Optional[int]) -> Dict:
return self.pa_type
def __a ( self :str , _lowercase :Union[str, bytes, dict, np.ndarray, "PIL.Image.Image"]) -> dict:
if config.PIL_AVAILABLE:
import PIL.Image
else:
raise ImportError('''To support encoding images, please install \'Pillow\'.''')
if isinstance(_lowercase , _lowercase):
UpperCAmelCase_ = np.array(_lowercase)
if isinstance(_lowercase , _lowercase):
return {"path": value, "bytes": None}
elif isinstance(_lowercase , _lowercase):
return {"path": None, "bytes": value}
elif isinstance(_lowercase , np.ndarray):
# convert the image array to PNG/TIFF bytes
return encode_np_array(_lowercase)
elif isinstance(_lowercase , PIL.Image.Image):
# convert the PIL image to bytes (default format is PNG/TIFF)
return encode_pil_image(_lowercase)
elif value.get('''path''') is not None and os.path.isfile(value['''path''']):
# we set "bytes": None to not duplicate the data if they're already available locally
return {"bytes": None, "path": value.get('''path''')}
elif value.get('''bytes''') is not None or value.get('''path''') is not None:
# store the image bytes, and path is used to infer the image format using the file extension
return {"bytes": value.get('''bytes'''), "path": value.get('''path''')}
else:
raise ValueError(
f"An image sample should have one of 'path' or 'bytes' but they are missing or None in {value}.")
def __a ( self :List[Any] , _lowercase :dict , _lowercase :Any=None) -> "PIL.Image.Image":
if not self.decode:
raise RuntimeError('''Decoding is disabled for this feature. Please use Image(decode=True) instead.''')
if config.PIL_AVAILABLE:
import PIL.Image
else:
raise ImportError('''To support decoding images, please install \'Pillow\'.''')
if token_per_repo_id is None:
UpperCAmelCase_ = {}
UpperCAmelCase_ , UpperCAmelCase_ = value['''path'''], value['''bytes''']
if bytes_ is None:
if path is None:
raise ValueError(f"An image should have one of 'path' or 'bytes' but both are None in {value}.")
else:
if is_local_path(_lowercase):
UpperCAmelCase_ = PIL.Image.open(_lowercase)
else:
UpperCAmelCase_ = path.split('''::''')[-1]
try:
UpperCAmelCase_ = string_to_dict(_lowercase , config.HUB_DATASETS_URL)['''repo_id''']
UpperCAmelCase_ = token_per_repo_id.get(_lowercase)
except ValueError:
UpperCAmelCase_ = None
with xopen(_lowercase , '''rb''' , use_auth_token=_lowercase) as f:
UpperCAmelCase_ = BytesIO(f.read())
UpperCAmelCase_ = PIL.Image.open(bytes_)
else:
UpperCAmelCase_ = PIL.Image.open(BytesIO(bytes_))
image.load() # to avoid "Too many open files" errors
return image
def __a ( self :str) -> Union["FeatureType", Dict[str, "FeatureType"]]:
from .features import Value
return (
self
if self.decode
else {
"bytes": Value('''binary'''),
"path": Value('''string'''),
}
)
def __a ( self :int , _lowercase :Union[pa.StringArray, pa.StructArray, pa.ListArray]) -> pa.StructArray:
if pa.types.is_string(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, storage] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_binary(storage.type):
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([storage, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_struct(storage.type):
if storage.type.get_field_index('''bytes''') >= 0:
UpperCAmelCase_ = storage.field('''bytes''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.binary())
if storage.type.get_field_index('''path''') >= 0:
UpperCAmelCase_ = storage.field('''path''')
else:
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=storage.is_null())
elif pa.types.is_list(storage.type):
UpperCAmelCase_ = pa.array(
[encode_np_array(np.array(_lowercase))['''bytes'''] if arr is not None else None for arr in storage.to_pylist()] , type=pa.binary() , )
UpperCAmelCase_ = pa.array([None] * len(_lowercase) , type=pa.string())
UpperCAmelCase_ = pa.StructArray.from_arrays(
[bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
def __a ( self :List[Any] , _lowercase :pa.StructArray) -> pa.StructArray:
@no_op_if_value_is_null
def path_to_bytes(_lowercase :List[Any]):
with xopen(_lowercase , '''rb''') as f:
UpperCAmelCase_ = f.read()
return bytes_
UpperCAmelCase_ = pa.array(
[
(path_to_bytes(x['''path''']) if x['''bytes'''] is None else x['''bytes''']) if x is not None else None
for x in storage.to_pylist()
] , type=pa.binary() , )
UpperCAmelCase_ = pa.array(
[os.path.basename(_lowercase) if path is not None else None for path in storage.field('''path''').to_pylist()] , type=pa.string() , )
UpperCAmelCase_ = pa.StructArray.from_arrays([bytes_array, path_array] , ['''bytes''', '''path'''] , mask=bytes_array.is_null())
return array_cast(_lowercase , self.pa_type)
def A ( ) -> List[str]:
'''simple docstring'''
if config.PIL_AVAILABLE:
import PIL.Image
else:
raise ImportError('''To support encoding images, please install \'Pillow\'.''' )
global _IMAGE_COMPRESSION_FORMATS
if _IMAGE_COMPRESSION_FORMATS is None:
PIL.Image.init()
UpperCAmelCase_ = list(set(PIL.Image.OPEN.keys() ) & set(PIL.Image.SAVE.keys() ) )
return _IMAGE_COMPRESSION_FORMATS
def A ( __UpperCAmelCase ) -> bytes:
'''simple docstring'''
UpperCAmelCase_ = BytesIO()
if image.format in list_image_compression_formats():
UpperCAmelCase_ = image.format
else:
UpperCAmelCase_ = '''PNG''' if image.mode in ['''1''', '''L''', '''LA''', '''RGB''', '''RGBA'''] else '''TIFF'''
image.save(__UpperCAmelCase , format=__UpperCAmelCase )
return buffer.getvalue()
def A ( __UpperCAmelCase ) -> dict:
'''simple docstring'''
if hasattr(__UpperCAmelCase , '''filename''' ) and image.filename != "":
return {"path": image.filename, "bytes": None}
else:
return {"path": None, "bytes": image_to_bytes(__UpperCAmelCase )}
def A ( __UpperCAmelCase ) -> dict:
'''simple docstring'''
if config.PIL_AVAILABLE:
import PIL.Image
else:
raise ImportError('''To support encoding images, please install \'Pillow\'.''' )
UpperCAmelCase_ = array.dtype
UpperCAmelCase_ = dtype.byteorder if dtype.byteorder != '''=''' else _NATIVE_BYTEORDER
UpperCAmelCase_ = dtype.kind
UpperCAmelCase_ = dtype.itemsize
UpperCAmelCase_ = None
# Multi-channel array case (only np.dtype("|u1") is allowed)
if array.shape[2:]:
UpperCAmelCase_ = np.dtype('''|u1''' )
if dtype_kind not in ["u", "i"]:
raise TypeError(
f"Unsupported array dtype {dtype} for image encoding. Only {dest_dtype} is supported for multi-channel arrays." )
if dtype is not dest_dtype:
warnings.warn(f"Downcasting array dtype {dtype} to {dest_dtype} to be compatible with 'Pillow'" )
# Exact match
elif dtype in _VALID_IMAGE_ARRAY_DTPYES:
UpperCAmelCase_ = dtype
else: # Downcast the type within the kind (np.can_cast(from_type, to_type, casting="same_kind") doesn't behave as expected, so do it manually)
while dtype_itemsize >= 1:
UpperCAmelCase_ = dtype_byteorder + dtype_kind + str(__UpperCAmelCase )
UpperCAmelCase_ = np.dtype(__UpperCAmelCase )
if dest_dtype in _VALID_IMAGE_ARRAY_DTPYES:
warnings.warn(f"Downcasting array dtype {dtype} to {dest_dtype} to be compatible with 'Pillow'" )
break
else:
dtype_itemsize //= 2
if dest_dtype is None:
raise TypeError(
f"Cannot convert dtype {dtype} to a valid image dtype. Valid image dtypes: {_VALID_IMAGE_ARRAY_DTPYES}" )
UpperCAmelCase_ = PIL.Image.fromarray(array.astype(__UpperCAmelCase ) )
return {"path": None, "bytes": image_to_bytes(__UpperCAmelCase )}
def A ( __UpperCAmelCase ) -> List[dict]:
'''simple docstring'''
if config.PIL_AVAILABLE:
import PIL.Image
else:
raise ImportError('''To support encoding images, please install \'Pillow\'.''' )
if objs:
UpperCAmelCase_ , UpperCAmelCase_ = first_non_null_value(__UpperCAmelCase )
if isinstance(__UpperCAmelCase , __UpperCAmelCase ):
return [{"path": obj, "bytes": None} if obj is not None else None for obj in objs]
if isinstance(__UpperCAmelCase , np.ndarray ):
UpperCAmelCase_ = no_op_if_value_is_null(__UpperCAmelCase )
return [obj_to_image_dict_func(__UpperCAmelCase ) for obj in objs]
elif isinstance(__UpperCAmelCase , PIL.Image.Image ):
UpperCAmelCase_ = no_op_if_value_is_null(__UpperCAmelCase )
return [obj_to_image_dict_func(__UpperCAmelCase ) for obj in objs]
else:
return objs
else:
return objs
| 344 |
import tempfile
import torch
from diffusers import (
DEISMultistepScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
UniPCMultistepScheduler,
)
from .test_schedulers import SchedulerCommonTest
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[Any] =(DPMSolverSinglestepScheduler,)
UpperCamelCase__ : Tuple =(("num_inference_steps", 25),)
def __a ( self :List[Any] , **_lowercase :Optional[Any]) -> int:
UpperCAmelCase_ = {
'''num_train_timesteps''': 1000,
'''beta_start''': 0.0_001,
'''beta_end''': 0.02,
'''beta_schedule''': '''linear''',
'''solver_order''': 2,
'''prediction_type''': '''epsilon''',
'''thresholding''': False,
'''sample_max_value''': 1.0,
'''algorithm_type''': '''dpmsolver++''',
'''solver_type''': '''midpoint''',
'''lambda_min_clipped''': -float('''inf'''),
'''variance_type''': None,
}
config.update(**_lowercase)
return config
def __a ( self :Union[str, Any] , _lowercase :List[Any]=0 , **_lowercase :Optional[int]) -> List[Any]:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ , UpperCAmelCase_ = sample, sample
for t in range(_lowercase , time_step + scheduler.config.solver_order + 1):
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Union[str, Any]) -> List[Any]:
pass
def __a ( self :Optional[Any] , _lowercase :str=0 , **_lowercase :Union[str, Any]) -> Dict:
UpperCAmelCase_ = dict(self.forward_default_kwargs)
UpperCAmelCase_ = kwargs.pop('''num_inference_steps''' , _lowercase)
UpperCAmelCase_ = self.dummy_sample
UpperCAmelCase_ = 0.1 * sample
UpperCAmelCase_ = [residual + 0.2, residual + 0.15, residual + 0.10]
for scheduler_class in self.scheduler_classes:
UpperCAmelCase_ = self.get_scheduler_config()
UpperCAmelCase_ = scheduler_class(**_lowercase)
scheduler.set_timesteps(_lowercase)
# copy over dummy past residuals (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(_lowercase)
UpperCAmelCase_ = scheduler_class.from_pretrained(_lowercase)
# copy over dummy past residuals
new_scheduler.set_timesteps(_lowercase)
# copy over dummy past residual (must be after setting timesteps)
UpperCAmelCase_ = dummy_past_residuals[: new_scheduler.config.solver_order]
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
UpperCAmelCase_ = new_scheduler.step(_lowercase , _lowercase , _lowercase , **_lowercase).prev_sample
assert torch.sum(torch.abs(output - new_output)) < 1E-5, "Scheduler outputs are not identical"
def __a ( self :Dict , _lowercase :Union[str, Any]=None , **_lowercase :List[Any]) -> int:
if scheduler is None:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(**_lowercase)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
return sample
def __a ( self :int) -> Tuple:
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = 50
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter
scheduler.set_timesteps(_lowercase)
# make sure that the first t is uneven
for i, t in enumerate(scheduler.timesteps[3:]):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_574) < 1E-3
def __a ( self :List[Any]) -> List[Any]:
for timesteps in [25, 50, 100, 999, 1000]:
self.check_over_configs(num_train_timesteps=_lowercase)
def __a ( self :int) -> Optional[Any]:
# make sure that iterating over schedulers with same config names gives same results
# for defaults
UpperCAmelCase_ = DPMSolverSinglestepScheduler(**self.get_scheduler_config())
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
UpperCAmelCase_ = DEISMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = UniPCMultistepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = DPMSolverSinglestepScheduler.from_config(scheduler.config)
UpperCAmelCase_ = self.full_loop(scheduler=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Tuple) -> int:
self.check_over_configs(thresholding=_lowercase)
for order in [1, 2, 3]:
for solver_type in ["midpoint", "heun"]:
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
thresholding=_lowercase , prediction_type=_lowercase , sample_max_value=_lowercase , algorithm_type='''dpmsolver++''' , solver_order=_lowercase , solver_type=_lowercase , )
def __a ( self :List[Any]) -> Any:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=_lowercase)
def __a ( self :Any) -> Optional[int]:
for algorithm_type in ["dpmsolver", "dpmsolver++"]:
for solver_type in ["midpoint", "heun"]:
for order in [1, 2, 3]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
UpperCAmelCase_ = self.full_loop(
solver_order=_lowercase , solver_type=_lowercase , prediction_type=_lowercase , algorithm_type=_lowercase , )
assert not torch.isnan(_lowercase).any(), "Samples have nan numbers"
def __a ( self :Tuple) -> int:
self.check_over_configs(lower_order_final=_lowercase)
self.check_over_configs(lower_order_final=_lowercase)
def __a ( self :Tuple) -> Optional[Any]:
self.check_over_configs(lambda_min_clipped=-float('''inf'''))
self.check_over_configs(lambda_min_clipped=-5.1)
def __a ( self :Any) -> List[str]:
self.check_over_configs(variance_type=_lowercase)
self.check_over_configs(variance_type='''learned_range''')
def __a ( self :Any) -> Dict:
for num_inference_steps in [1, 2, 3, 5, 10, 50, 100, 999, 1000]:
self.check_over_forward(num_inference_steps=_lowercase , time_step=0)
def __a ( self :Dict) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop()
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_791) < 1E-3
def __a ( self :Any) -> Union[str, Any]:
UpperCAmelCase_ = self.full_loop(use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.2_248) < 1E-3
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''')
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.1_453) < 1E-3
def __a ( self :List[Any]) -> Dict:
UpperCAmelCase_ = self.full_loop(prediction_type='''v_prediction''' , use_karras_sigmas=_lowercase)
UpperCAmelCase_ = torch.mean(torch.abs(_lowercase))
assert abs(result_mean.item() - 0.0_649) < 1E-3
def __a ( self :Any) -> Optional[Any]:
UpperCAmelCase_ = self.scheduler_classes[0]
UpperCAmelCase_ = self.get_scheduler_config(thresholding=_lowercase , dynamic_thresholding_ratio=0)
UpperCAmelCase_ = scheduler_class(**_lowercase)
UpperCAmelCase_ = 10
UpperCAmelCase_ = self.dummy_model()
UpperCAmelCase_ = self.dummy_sample_deter.half()
scheduler.set_timesteps(_lowercase)
for i, t in enumerate(scheduler.timesteps):
UpperCAmelCase_ = model(_lowercase , _lowercase)
UpperCAmelCase_ = scheduler.step(_lowercase , _lowercase , _lowercase).prev_sample
assert sample.dtype == torch.floataa
| 344 | 1 |
from __future__ import annotations
import inspect
import unittest
from typing import List, Tuple
from transformers import RegNetConfig
from transformers.testing_utils import require_tf, require_vision, slow
from transformers.utils import cached_property, is_tf_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TF_REGNET_PRETRAINED_MODEL_ARCHIVE_LIST, TFRegNetForImageClassification, TFRegNetModel
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class a_ :
def __init__( self :List[str] , _lowercase :List[Any] , _lowercase :List[Any]=3 , _lowercase :Any=32 , _lowercase :int=3 , _lowercase :List[Any]=10 , _lowercase :Any=[10, 20, 30, 40] , _lowercase :Tuple=[1, 1, 2, 1] , _lowercase :Tuple=True , _lowercase :Tuple=True , _lowercase :Optional[Any]="relu" , _lowercase :str=3 , _lowercase :Optional[Any]=None , ) -> Any:
UpperCAmelCase_ = parent
UpperCAmelCase_ = batch_size
UpperCAmelCase_ = image_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = embeddings_size
UpperCAmelCase_ = hidden_sizes
UpperCAmelCase_ = depths
UpperCAmelCase_ = is_training
UpperCAmelCase_ = use_labels
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = num_labels
UpperCAmelCase_ = scope
UpperCAmelCase_ = len(_lowercase)
def __a ( self :Tuple) -> Optional[Any]:
UpperCAmelCase_ = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size])
UpperCAmelCase_ = None
if self.use_labels:
UpperCAmelCase_ = ids_tensor([self.batch_size] , self.num_labels)
UpperCAmelCase_ = self.get_config()
return config, pixel_values, labels
def __a ( self :Optional[Any]) -> Optional[int]:
return RegNetConfig(
num_channels=self.num_channels , embeddings_size=self.embeddings_size , hidden_sizes=self.hidden_sizes , depths=self.depths , hidden_act=self.hidden_act , num_labels=self.num_labels , )
def __a ( self :Dict , _lowercase :List[Any] , _lowercase :List[str] , _lowercase :Dict) -> Dict:
UpperCAmelCase_ = TFRegNetModel(config=_lowercase)
UpperCAmelCase_ = model(_lowercase , training=_lowercase)
# expected last hidden states: B, C, H // 32, W // 32
self.parent.assertEqual(
result.last_hidden_state.shape , (self.batch_size, self.hidden_sizes[-1], self.image_size // 32, self.image_size // 32) , )
def __a ( self :Union[str, Any] , _lowercase :int , _lowercase :List[str] , _lowercase :int) -> Union[str, Any]:
UpperCAmelCase_ = self.num_labels
UpperCAmelCase_ = TFRegNetForImageClassification(_lowercase)
UpperCAmelCase_ = model(_lowercase , labels=_lowercase , training=_lowercase)
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels))
def __a ( self :Tuple) -> Optional[Any]:
UpperCAmelCase_ = self.prepare_config_and_inputs()
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = config_and_inputs
UpperCAmelCase_ = {'''pixel_values''': pixel_values}
return config, inputs_dict
@require_tf
class a_ ( _snake_case , _snake_case , unittest.TestCase ):
UpperCamelCase__ : Any =(TFRegNetModel, TFRegNetForImageClassification) if is_tf_available() else ()
UpperCamelCase__ : str =(
{"feature-extraction": TFRegNetModel, "image-classification": TFRegNetForImageClassification}
if is_tf_available()
else {}
)
UpperCamelCase__ : Optional[int] =False
UpperCamelCase__ : List[Any] =False
UpperCamelCase__ : str =False
UpperCamelCase__ : List[Any] =False
UpperCamelCase__ : str =False
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = TFRegNetModelTester(self)
UpperCAmelCase_ = ConfigTester(self , config_class=_lowercase , has_text_modality=_lowercase)
def __a ( self :Optional[int]) -> int:
return
@unittest.skip(reason='''RegNet does not use inputs_embeds''')
def __a ( self :Union[str, Any]) -> List[str]:
pass
@unittest.skipIf(
not is_tf_available() or len(tf.config.list_physical_devices('''GPU''')) == 0 , reason='''TF does not support backprop for grouped convolutions on CPU.''' , )
@slow
def __a ( self :List[str]) -> int:
super().test_keras_fit()
@unittest.skip(reason='''RegNet does not support input and output embeddings''')
def __a ( self :Union[str, Any]) -> List[Any]:
pass
def __a ( self :Union[str, Any]) -> Optional[int]:
UpperCAmelCase_ , UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
UpperCAmelCase_ = model_class(_lowercase)
UpperCAmelCase_ = inspect.signature(model.call)
# signature.parameters is an OrderedDict => so arg_names order is deterministic
UpperCAmelCase_ = [*signature.parameters.keys()]
UpperCAmelCase_ = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , _lowercase)
def __a ( self :Optional[int]) -> Tuple:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*_lowercase)
def __a ( self :int) -> Optional[int]:
def check_hidden_states_output(_lowercase :List[str] , _lowercase :Tuple , _lowercase :Tuple):
UpperCAmelCase_ = model_class(_lowercase)
UpperCAmelCase_ = model(**self._prepare_for_class(_lowercase , _lowercase) , training=_lowercase)
UpperCAmelCase_ = outputs.encoder_hidden_states if config.is_encoder_decoder else outputs.hidden_states
UpperCAmelCase_ = self.model_tester.num_stages
self.assertEqual(len(_lowercase) , expected_num_stages + 1)
# RegNet's feature maps are of shape (batch_size, num_channels, height, width)
self.assertListEqual(
list(hidden_states[0].shape[-2:]) , [self.model_tester.image_size // 2, self.model_tester.image_size // 2] , )
UpperCAmelCase_ , UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs_for_common()
UpperCAmelCase_ = ['''basic''', '''bottleneck''']
for model_class in self.all_model_classes:
for layer_type in layers_type:
UpperCAmelCase_ = layer_type
UpperCAmelCase_ = True
check_hidden_states_output(_lowercase , _lowercase , _lowercase)
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
UpperCAmelCase_ = True
check_hidden_states_output(_lowercase , _lowercase , _lowercase)
def __a ( self :Union[str, Any]) -> Union[str, Any]:
UpperCAmelCase_ , UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs_for_common()
def check_equivalence(_lowercase :Union[str, Any] , _lowercase :str , _lowercase :int , _lowercase :Any={}):
UpperCAmelCase_ = model(_lowercase , return_dict=_lowercase , **_lowercase)
UpperCAmelCase_ = model(_lowercase , return_dict=_lowercase , **_lowercase).to_tuple()
def recursive_check(_lowercase :Dict , _lowercase :Union[str, Any]):
if isinstance(_lowercase , (List, Tuple)):
for tuple_iterable_value, dict_iterable_value in zip(_lowercase , _lowercase):
recursive_check(_lowercase , _lowercase)
elif tuple_object is None:
return
else:
self.assertTrue(
all(tf.equal(_lowercase , _lowercase)) , msg=(
'''Tuple and dict output are not equal. Difference:'''
f" {tf.math.reduce_max(tf.abs(tuple_object - dict_object))}"
) , )
recursive_check(_lowercase , _lowercase)
for model_class in self.all_model_classes:
UpperCAmelCase_ = model_class(_lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase)
check_equivalence(_lowercase , _lowercase , _lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase , return_labels=_lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase , return_labels=_lowercase)
check_equivalence(_lowercase , _lowercase , _lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase)
check_equivalence(_lowercase , _lowercase , _lowercase , {'''output_hidden_states''': True})
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase , return_labels=_lowercase)
UpperCAmelCase_ = self._prepare_for_class(_lowercase , _lowercase , return_labels=_lowercase)
check_equivalence(_lowercase , _lowercase , _lowercase , {'''output_hidden_states''': True})
def __a ( self :Any) -> int:
UpperCAmelCase_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*_lowercase)
@slow
def __a ( self :Optional[int]) -> List[str]:
for model_name in TF_REGNET_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
UpperCAmelCase_ = TFRegNetModel.from_pretrained(_lowercase)
self.assertIsNotNone(_lowercase)
def A ( ) -> Tuple:
'''simple docstring'''
UpperCAmelCase_ = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_tf
@require_vision
class a_ ( unittest.TestCase ):
@cached_property
def __a ( self :Optional[int]) -> Any:
return (
AutoImageProcessor.from_pretrained(TF_REGNET_PRETRAINED_MODEL_ARCHIVE_LIST[0])
if is_vision_available()
else None
)
@slow
def __a ( self :int) -> Union[str, Any]:
UpperCAmelCase_ = TFRegNetForImageClassification.from_pretrained(TF_REGNET_PRETRAINED_MODEL_ARCHIVE_LIST[0])
UpperCAmelCase_ = self.default_image_processor
UpperCAmelCase_ = prepare_img()
UpperCAmelCase_ = image_processor(images=_lowercase , return_tensors='''tf''')
# forward pass
UpperCAmelCase_ = model(**_lowercase , training=_lowercase)
# verify the logits
UpperCAmelCase_ = tf.TensorShape((1, 1000))
self.assertEqual(outputs.logits.shape , _lowercase)
UpperCAmelCase_ = tf.constant([-0.4_180, -1.5_051, -3.4_836])
tf.debugging.assert_near(outputs.logits[0, :3] , _lowercase , atol=1E-4)
| 344 |
import inspect
import unittest
import torch
import torch.nn as nn
from accelerate.hooks import (
AlignDevicesHook,
ModelHook,
SequentialHook,
add_hook_to_module,
attach_align_device_hook,
remove_hook_from_module,
remove_hook_from_submodules,
)
from accelerate.test_utils import require_multi_gpu
class a_ ( nn.Module ):
def __init__( self :Optional[Any]) -> Union[str, Any]:
super().__init__()
UpperCAmelCase_ = nn.Linear(3 , 4)
UpperCAmelCase_ = nn.BatchNormad(4)
UpperCAmelCase_ = nn.Linear(4 , 5)
def __a ( self :Dict , _lowercase :int) -> str:
return self.lineara(self.batchnorm(self.lineara(_lowercase)))
class a_ ( _snake_case ):
def __a ( self :Tuple , _lowercase :Optional[int] , *_lowercase :Union[str, Any] , **_lowercase :Any) -> Optional[Any]:
return (args[0] + 1,) + args[1:], kwargs
class a_ ( _snake_case ):
def __a ( self :Union[str, Any] , _lowercase :Dict , _lowercase :Tuple) -> int:
return output + 1
class a_ ( unittest.TestCase ):
def __a ( self :str) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
self.assertEqual(test_model._hf_hook , _lowercase)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[Any]) -> Any:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = ModelHook()
add_hook_to_module(_lowercase , _lowercase)
add_hook_to_module(_lowercase , _lowercase , append=_lowercase)
self.assertEqual(isinstance(test_model._hf_hook , _lowercase) , _lowercase)
self.assertEqual(len(test_model._hf_hook.hooks) , 2)
self.assertTrue(hasattr(_lowercase , '''_old_forward'''))
# Check adding the hook did not change the name or the signature
self.assertEqual(test_model.forward.__name__ , '''forward''')
self.assertListEqual(list(inspect.signature(test_model.forward).parameters) , ['''x'''])
remove_hook_from_module(_lowercase)
self.assertFalse(hasattr(_lowercase , '''_hf_hook'''))
self.assertFalse(hasattr(_lowercase , '''_old_forward'''))
def __a ( self :Optional[int]) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(x + 1)
UpperCAmelCase_ = test_model(x + 2)
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PreForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , _lowercase , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PreForwardHook() , PreForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , _lowercase , atol=1E-5)
def __a ( self :List[str]) -> int:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# Attaching a hook to a model when it already has one replaces, does not chain
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1 , atol=1E-5))
# You need to use the sequential hook to chain two or more hooks
UpperCAmelCase_ = SequentialHook(PostForwardHook() , PostForwardHook())
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
assert torch.allclose(_lowercase , output + 2 , atol=1E-5)
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = test_model(_lowercase)
UpperCAmelCase_ = PostForwardHook()
add_hook_to_module(_lowercase , _lowercase)
UpperCAmelCase_ = test_model(_lowercase)
self.assertTrue(torch.allclose(_lowercase , output + 1))
self.assertTrue(outputa.requires_grad)
UpperCAmelCase_ = True
UpperCAmelCase_ = test_model(_lowercase)
self.assertFalse(outputa.requires_grad)
@require_multi_gpu
def __a ( self :Tuple) -> Optional[int]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.batchnorm , AlignDevicesHook(execution_device=0))
add_hook_to_module(model.lineara , AlignDevicesHook(execution_device=1))
self.assertEqual(model.lineara.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.weight.device , torch.device(0))
self.assertEqual(model.batchnorm.running_mean.device , torch.device(0))
self.assertEqual(model.lineara.weight.device , torch.device(1))
# We can still make a forward pass. The input does not need to be on any particular device
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(1))
# We can add a general hook to put back output on same device as input.
add_hook_to_module(_lowercase , AlignDevicesHook(io_same_device=_lowercase))
UpperCAmelCase_ = torch.randn(2 , 3).to(0)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , torch.device(0))
def __a ( self :str) -> List[Any]:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = {'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''', '''offload''': True}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(hook_kwargs['''execution_device'''])
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
UpperCAmelCase_ = {
'''execution_device''': 0 if torch.cuda.is_available() else '''cpu''',
'''offload''': True,
'''offload_buffers''': True,
}
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.batchnorm , AlignDevicesHook(**_lowercase))
add_hook_to_module(model.lineara , AlignDevicesHook(**_lowercase))
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_module(model.lineara)
remove_hook_from_module(model.batchnorm)
remove_hook_from_module(model.lineara)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :List[Any]) -> str:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase)
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(_lowercase , execution_device=_lowercase , offload=_lowercase , offload_buffers=_lowercase)
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = ModelForTest()
# Everything is on CPU
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# This will move each submodule on different devices
UpperCAmelCase_ = 0 if torch.cuda.is_available() else '''cpu'''
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict())
# Parameters have been offloaded, so on the meta device
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
# Buffers are not included in the offload by default, so are on the execution device
UpperCAmelCase_ = torch.device(_lowercase)
self.assertEqual(model.batchnorm.running_mean.device , _lowercase)
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
# Now test with buffers included in the offload
attach_align_device_hook(
_lowercase , execution_device=_lowercase , offload=_lowercase , weights_map=model.state_dict() , offload_buffers=_lowercase , )
# Parameters have been offloaded, so on the meta device, buffers included
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''meta'''))
self.assertEqual(model.lineara.weight.device , torch.device('''meta'''))
self.assertEqual(model.batchnorm.running_mean.device , torch.device('''meta'''))
UpperCAmelCase_ = torch.randn(2 , 3)
UpperCAmelCase_ = model(_lowercase)
self.assertEqual(output.device , _lowercase)
# Removing hooks loads back the weights in the model.
remove_hook_from_submodules(_lowercase)
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
self.assertEqual(model.batchnorm.weight.device , torch.device('''cpu'''))
self.assertEqual(model.lineara.weight.device , torch.device('''cpu'''))
| 344 | 1 |
import gc
import random
import unittest
import numpy as np
import torch
from PIL import Image
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
EulerAncestralDiscreteScheduler,
LMSDiscreteScheduler,
PNDMScheduler,
StableDiffusionInstructPixaPixPipeline,
UNetaDConditionModel,
)
from diffusers.image_processor import VaeImageProcessor
from diffusers.utils import floats_tensor, load_image, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import (
IMAGE_TO_IMAGE_IMAGE_PARAMS,
TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineKarrasSchedulerTesterMixin, PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class a_ ( _snake_case , _snake_case , _snake_case , unittest.TestCase ):
UpperCamelCase__ : Tuple =StableDiffusionInstructPixaPixPipeline
UpperCamelCase__ : List[str] =TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {"height", "width", "cross_attention_kwargs"}
UpperCamelCase__ : Any =TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS
UpperCamelCase__ : int =IMAGE_TO_IMAGE_IMAGE_PARAMS
UpperCamelCase__ : Optional[int] =IMAGE_TO_IMAGE_IMAGE_PARAMS
def __a ( self :List[Any]) -> Any:
torch.manual_seed(0)
UpperCAmelCase_ = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=8 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=32 , )
UpperCAmelCase_ = PNDMScheduler(skip_prk_steps=_lowercase)
torch.manual_seed(0)
UpperCAmelCase_ = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''] , up_block_types=['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''] , latent_channels=4 , )
torch.manual_seed(0)
UpperCAmelCase_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1E-0_5 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , )
UpperCAmelCase_ = CLIPTextModel(_lowercase)
UpperCAmelCase_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''')
UpperCAmelCase_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
'''safety_checker''': None,
'''feature_extractor''': None,
}
return components
def __a ( self :Tuple , _lowercase :Optional[int] , _lowercase :int=0) -> Union[str, Any]:
UpperCAmelCase_ = floats_tensor((1, 3, 32, 32) , rng=random.Random(_lowercase)).to(_lowercase)
UpperCAmelCase_ = image.cpu().permute(0 , 2 , 3 , 1)[0]
UpperCAmelCase_ = Image.fromarray(np.uinta(_lowercase)).convert('''RGB''')
if str(_lowercase).startswith('''mps'''):
UpperCAmelCase_ = torch.manual_seed(_lowercase)
else:
UpperCAmelCase_ = torch.Generator(device=_lowercase).manual_seed(_lowercase)
UpperCAmelCase_ = {
'''prompt''': '''A painting of a squirrel eating a burger''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''guidance_scale''': 6.0,
'''image_guidance_scale''': 1,
'''output_type''': '''numpy''',
}
return inputs
def __a ( self :List[Any]) -> Any:
UpperCAmelCase_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase_ = self.get_dummy_components()
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline(**_lowercase)
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
UpperCAmelCase_ = self.get_dummy_inputs(_lowercase)
UpperCAmelCase_ = sd_pipe(**_lowercase).images
UpperCAmelCase_ = image[0, -3:, -3:, -1]
assert image.shape == (1, 32, 32, 3)
UpperCAmelCase_ = np.array([0.7_526, 0.3_750, 0.4_547, 0.6_117, 0.5_866, 0.5_016, 0.4_327, 0.5_642, 0.4_815])
assert np.abs(image_slice.flatten() - expected_slice).max() < 1E-3
def __a ( self :Optional[int]) -> int:
UpperCAmelCase_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase_ = self.get_dummy_components()
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline(**_lowercase)
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
UpperCAmelCase_ = self.get_dummy_inputs(_lowercase)
UpperCAmelCase_ = '''french fries'''
UpperCAmelCase_ = sd_pipe(**_lowercase , negative_prompt=_lowercase)
UpperCAmelCase_ = output.images
UpperCAmelCase_ = image[0, -3:, -3:, -1]
assert image.shape == (1, 32, 32, 3)
UpperCAmelCase_ = np.array([0.7_511, 0.3_642, 0.4_553, 0.6_236, 0.5_797, 0.5_013, 0.4_343, 0.5_611, 0.4_831])
assert np.abs(image_slice.flatten() - expected_slice).max() < 1E-3
def __a ( self :List[str]) -> Union[str, Any]:
UpperCAmelCase_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase_ = self.get_dummy_components()
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline(**_lowercase)
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
UpperCAmelCase_ = self.get_dummy_inputs(_lowercase)
UpperCAmelCase_ = [inputs['''prompt''']] * 2
UpperCAmelCase_ = np.array(inputs['''image''']).astype(np.floataa) / 255.0
UpperCAmelCase_ = torch.from_numpy(_lowercase).unsqueeze(0).to(_lowercase)
UpperCAmelCase_ = image / 2 + 0.5
UpperCAmelCase_ = image.permute(0 , 3 , 1 , 2)
UpperCAmelCase_ = image.repeat(2 , 1 , 1 , 1)
UpperCAmelCase_ = sd_pipe(**_lowercase).images
UpperCAmelCase_ = image[-1, -3:, -3:, -1]
assert image.shape == (2, 32, 32, 3)
UpperCAmelCase_ = np.array([0.5_812, 0.5_748, 0.5_222, 0.5_908, 0.5_695, 0.7_174, 0.6_804, 0.5_523, 0.5_579])
assert np.abs(image_slice.flatten() - expected_slice).max() < 1E-3
def __a ( self :int) -> List[Any]:
UpperCAmelCase_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
UpperCAmelCase_ = self.get_dummy_components()
UpperCAmelCase_ = EulerAncestralDiscreteScheduler(
beta_start=0.00_085 , beta_end=0.012 , beta_schedule='''scaled_linear''')
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline(**_lowercase)
UpperCAmelCase_ = sd_pipe.to(_lowercase)
sd_pipe.set_progress_bar_config(disable=_lowercase)
UpperCAmelCase_ = self.get_dummy_inputs(_lowercase)
UpperCAmelCase_ = sd_pipe(**_lowercase).images
UpperCAmelCase_ = image[0, -3:, -3:, -1]
UpperCAmelCase_ = [round(_lowercase , 4) for x in image_slice.flatten().tolist()]
print(''','''.join([str(_lowercase) for x in slice]))
assert image.shape == (1, 32, 32, 3)
UpperCAmelCase_ = np.array([0.7_417, 0.3_842, 0.4_732, 0.5_776, 0.5_891, 0.5_139, 0.4_052, 0.5_673, 0.4_986])
assert np.abs(image_slice.flatten() - expected_slice).max() < 1E-3
def __a ( self :Dict) -> List[str]:
super().test_inference_batch_single_identical(expected_max_diff=3E-3)
def __a ( self :Dict) -> int:
UpperCAmelCase_ = self.get_dummy_components()
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline(**_lowercase)
UpperCAmelCase_ = VaeImageProcessor(do_resize=_lowercase , do_normalize=_lowercase)
UpperCAmelCase_ = pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
UpperCAmelCase_ = pipe(**self.get_dummy_inputs_by_type(_lowercase , input_image_type='''pt'''))[0]
UpperCAmelCase_ = components['''vae''']
UpperCAmelCase_ = self.get_dummy_inputs_by_type(_lowercase , input_image_type='''pt''')
for image_param in self.image_latents_params:
if image_param in inputs.keys():
UpperCAmelCase_ = vae.encode(inputs[image_param]).latent_dist.mode()
UpperCAmelCase_ = pipe(**_lowercase)[0]
UpperCAmelCase_ = np.abs(out - out_latents_inputs).max()
self.assertLess(_lowercase , 1E-4 , '''passing latents as image input generate different result from passing image''')
@slow
@require_torch_gpu
class a_ ( unittest.TestCase ):
def __a ( self :Optional[int]) -> List[Any]:
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def __a ( self :List[str] , _lowercase :List[str]=0) -> Optional[Any]:
UpperCAmelCase_ = torch.manual_seed(_lowercase)
UpperCAmelCase_ = load_image(
'''https://huggingface.co/datasets/diffusers/test-arrays/resolve/main/stable_diffusion_pix2pix/example.jpg''')
UpperCAmelCase_ = {
'''prompt''': '''turn him into a cyborg''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 3,
'''guidance_scale''': 7.5,
'''image_guidance_scale''': 1.0,
'''output_type''': '''numpy''',
}
return inputs
def __a ( self :List[str]) -> Union[str, Any]:
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline.from_pretrained(
'''timbrooks/instruct-pix2pix''' , safety_checker=_lowercase)
pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
pipe.enable_attention_slicing()
UpperCAmelCase_ = self.get_inputs()
UpperCAmelCase_ = pipe(**_lowercase).images
UpperCAmelCase_ = image[0, -3:, -3:, -1].flatten()
assert image.shape == (1, 512, 512, 3)
UpperCAmelCase_ = np.array([0.5_902, 0.6_015, 0.6_027, 0.5_983, 0.6_092, 0.6_061, 0.5_765, 0.5_785, 0.5_555])
assert np.abs(expected_slice - image_slice).max() < 1E-3
def __a ( self :Dict) -> str:
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline.from_pretrained(
'''timbrooks/instruct-pix2pix''' , safety_checker=_lowercase)
UpperCAmelCase_ = LMSDiscreteScheduler.from_config(pipe.scheduler.config)
pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
pipe.enable_attention_slicing()
UpperCAmelCase_ = self.get_inputs()
UpperCAmelCase_ = pipe(**_lowercase).images
UpperCAmelCase_ = image[0, -3:, -3:, -1].flatten()
assert image.shape == (1, 512, 512, 3)
UpperCAmelCase_ = np.array([0.6_578, 0.6_817, 0.6_972, 0.6_761, 0.6_856, 0.6_916, 0.6_428, 0.6_516, 0.6_301])
assert np.abs(expected_slice - image_slice).max() < 1E-3
def __a ( self :Optional[Any]) -> Optional[Any]:
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline.from_pretrained(
'''timbrooks/instruct-pix2pix''' , safety_checker=_lowercase)
UpperCAmelCase_ = DDIMScheduler.from_config(pipe.scheduler.config)
pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
pipe.enable_attention_slicing()
UpperCAmelCase_ = self.get_inputs()
UpperCAmelCase_ = pipe(**_lowercase).images
UpperCAmelCase_ = image[0, -3:, -3:, -1].flatten()
assert image.shape == (1, 512, 512, 3)
UpperCAmelCase_ = np.array([0.3_828, 0.3_834, 0.3_818, 0.3_792, 0.3_865, 0.3_752, 0.3_792, 0.3_847, 0.3_753])
assert np.abs(expected_slice - image_slice).max() < 1E-3
def __a ( self :Tuple) -> str:
UpperCAmelCase_ = 0
def callback_fn(_lowercase :int , _lowercase :int , _lowercase :torch.FloatTensor) -> None:
UpperCAmelCase_ = True
nonlocal number_of_steps
number_of_steps += 1
if step == 1:
UpperCAmelCase_ = latents.detach().cpu().numpy()
assert latents.shape == (1, 4, 64, 64)
UpperCAmelCase_ = latents[0, -3:, -3:, -1]
UpperCAmelCase_ = np.array([-0.2_463, -0.4_644, -0.9_756, 1.5_176, 1.4_414, 0.7_866, 0.9_897, 0.8_521, 0.7_983])
assert np.abs(latents_slice.flatten() - expected_slice).max() < 5E-2
elif step == 2:
UpperCAmelCase_ = latents.detach().cpu().numpy()
assert latents.shape == (1, 4, 64, 64)
UpperCAmelCase_ = latents[0, -3:, -3:, -1]
UpperCAmelCase_ = np.array([-0.2_644, -0.4_626, -0.9_653, 1.5_176, 1.4_551, 0.7_686, 0.9_805, 0.8_452, 0.8_115])
assert np.abs(latents_slice.flatten() - expected_slice).max() < 5E-2
UpperCAmelCase_ = False
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline.from_pretrained(
'''timbrooks/instruct-pix2pix''' , safety_checker=_lowercase , torch_dtype=torch.floataa)
UpperCAmelCase_ = pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
pipe.enable_attention_slicing()
UpperCAmelCase_ = self.get_inputs()
pipe(**_lowercase , callback=_lowercase , callback_steps=1)
assert callback_fn.has_been_called
assert number_of_steps == 3
def __a ( self :int) -> Dict:
torch.cuda.empty_cache()
torch.cuda.reset_max_memory_allocated()
torch.cuda.reset_peak_memory_stats()
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline.from_pretrained(
'''timbrooks/instruct-pix2pix''' , safety_checker=_lowercase , torch_dtype=torch.floataa)
UpperCAmelCase_ = pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
pipe.enable_attention_slicing(1)
pipe.enable_sequential_cpu_offload()
UpperCAmelCase_ = self.get_inputs()
UpperCAmelCase_ = pipe(**_lowercase)
UpperCAmelCase_ = torch.cuda.max_memory_allocated()
# make sure that less than 2.2 GB is allocated
assert mem_bytes < 2.2 * 10**9
def __a ( self :Any) -> Optional[int]:
UpperCAmelCase_ = self.get_inputs()
# resize to resolution that is divisible by 8 but not 16 or 32
UpperCAmelCase_ = inputs['''image'''].resize((504, 504))
UpperCAmelCase_ = '''timbrooks/instruct-pix2pix'''
UpperCAmelCase_ = StableDiffusionInstructPixaPixPipeline.from_pretrained(
_lowercase , safety_checker=_lowercase , )
pipe.to(_lowercase)
pipe.set_progress_bar_config(disable=_lowercase)
pipe.enable_attention_slicing()
UpperCAmelCase_ = pipe(**_lowercase)
UpperCAmelCase_ = output.images[0]
UpperCAmelCase_ = image[255:258, 383:386, -1]
assert image.shape == (504, 504, 3)
UpperCAmelCase_ = np.array([0.2_726, 0.2_529, 0.2_664, 0.2_655, 0.2_641, 0.2_642, 0.2_591, 0.2_649, 0.2_590])
assert np.abs(image_slice.flatten() - expected_slice).max() < 5E-3
| 344 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers.testing_utils import require_vision
from transformers.utils import is_vision_available
if is_vision_available():
from PIL import Image
from transformers import (
AutoProcessor,
BertTokenizerFast,
BlipImageProcessor,
GPTaTokenizer,
InstructBlipProcessor,
PreTrainedTokenizerFast,
)
@require_vision
class a_ ( unittest.TestCase ):
def __a ( self :Optional[Any]) -> int:
UpperCAmelCase_ = tempfile.mkdtemp()
UpperCAmelCase_ = BlipImageProcessor()
UpperCAmelCase_ = GPTaTokenizer.from_pretrained('''hf-internal-testing/tiny-random-GPT2Model''')
UpperCAmelCase_ = BertTokenizerFast.from_pretrained('''hf-internal-testing/tiny-random-bert''')
UpperCAmelCase_ = InstructBlipProcessor(_lowercase , _lowercase , _lowercase)
processor.save_pretrained(self.tmpdirname)
def __a ( self :List[Any] , **_lowercase :Dict) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).tokenizer
def __a ( self :Optional[Any] , **_lowercase :Optional[Any]) -> Optional[int]:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).image_processor
def __a ( self :Dict , **_lowercase :Tuple) -> str:
return AutoProcessor.from_pretrained(self.tmpdirname , **_lowercase).qformer_tokenizer
def __a ( self :Optional[int]) -> str:
shutil.rmtree(self.tmpdirname)
def __a ( self :Any) -> List[str]:
UpperCAmelCase_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta)]
UpperCAmelCase_ = [Image.fromarray(np.moveaxis(_lowercase , 0 , -1)) for x in image_inputs]
return image_inputs
def __a ( self :Tuple) -> int:
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() , qformer_tokenizer=self.get_qformer_tokenizer() , )
processor.save_pretrained(self.tmpdirname)
UpperCAmelCase_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''')
UpperCAmelCase_ = self.get_image_processor(do_normalize=_lowercase , padding_value=1.0)
UpperCAmelCase_ = InstructBlipProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=_lowercase , padding_value=1.0)
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab())
self.assertIsInstance(processor.tokenizer , _lowercase)
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string())
self.assertIsInstance(processor.image_processor , _lowercase)
self.assertIsInstance(processor.qformer_tokenizer , _lowercase)
def __a ( self :Dict) -> Any:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = image_processor(_lowercase , return_tensors='''np''')
UpperCAmelCase_ = processor(images=_lowercase , return_tensors='''np''')
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1E-2)
def __a ( self :Union[str, Any]) -> Dict:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = processor(text=_lowercase)
UpperCAmelCase_ = tokenizer(_lowercase , return_token_type_ids=_lowercase)
UpperCAmelCase_ = qformer_tokenizer(_lowercase , return_token_type_ids=_lowercase)
for key in encoded_tokens.keys():
self.assertListEqual(encoded_tokens[key] , encoded_processor[key])
for key in encoded_tokens_qformer.keys():
self.assertListEqual(encoded_tokens_qformer[key] , encoded_processor['''qformer_''' + key])
def __a ( self :Dict) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
# test if it raises when no input is passed
with pytest.raises(_lowercase):
processor()
def __a ( self :Optional[int]) -> Optional[Any]:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
UpperCAmelCase_ = processor.batch_decode(_lowercase)
UpperCAmelCase_ = tokenizer.batch_decode(_lowercase)
self.assertListEqual(_lowercase , _lowercase)
def __a ( self :str) -> int:
UpperCAmelCase_ = self.get_image_processor()
UpperCAmelCase_ = self.get_tokenizer()
UpperCAmelCase_ = self.get_qformer_tokenizer()
UpperCAmelCase_ = InstructBlipProcessor(
tokenizer=_lowercase , image_processor=_lowercase , qformer_tokenizer=_lowercase)
UpperCAmelCase_ = '''lower newer'''
UpperCAmelCase_ = self.prepare_image_inputs()
UpperCAmelCase_ = processor(text=_lowercase , images=_lowercase)
self.assertListEqual(
list(inputs.keys()) , ['''input_ids''', '''attention_mask''', '''qformer_input_ids''', '''qformer_attention_mask''', '''pixel_values'''] , )
| 344 | 1 |
import argparse
import re
from typing import Dict
import torch
from datasets import Audio, Dataset, load_dataset, load_metric
from transformers import AutoFeatureExtractor, pipeline
def A ( __UpperCAmelCase , __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
UpperCAmelCase_ = args.log_outputs
UpperCAmelCase_ = '''_'''.join(args.dataset.split('''/''' ) + [args.config, args.split] )
# load metric
UpperCAmelCase_ = load_metric('''wer''' )
UpperCAmelCase_ = load_metric('''cer''' )
# compute metrics
UpperCAmelCase_ = wer.compute(references=result['''target'''] , predictions=result['''prediction'''] )
UpperCAmelCase_ = cer.compute(references=result['''target'''] , predictions=result['''prediction'''] )
# print & log results
UpperCAmelCase_ = f"WER: {wer_result}\nCER: {cer_result}"
print(__UpperCAmelCase )
with open(f"{dataset_id}_eval_results.txt" , '''w''' ) as f:
f.write(__UpperCAmelCase )
# log all results in text file. Possibly interesting for analysis
if log_outputs is not None:
UpperCAmelCase_ = f"log_{dataset_id}_predictions.txt"
UpperCAmelCase_ = f"log_{dataset_id}_targets.txt"
with open(__UpperCAmelCase , '''w''' ) as p, open(__UpperCAmelCase , '''w''' ) as t:
# mapping function to write output
def write_to_file(__UpperCAmelCase , __UpperCAmelCase ):
p.write(f"{i}" + '''\n''' )
p.write(batch['''prediction'''] + '''\n''' )
t.write(f"{i}" + '''\n''' )
t.write(batch['''target'''] + '''\n''' )
result.map(__UpperCAmelCase , with_indices=__UpperCAmelCase )
def A ( __UpperCAmelCase ) -> str:
'''simple docstring'''
UpperCAmelCase_ = '''[,?.!\-\;\:"“%‘”�—’…–]''' # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
UpperCAmelCase_ = re.sub(__UpperCAmelCase , '''''' , text.lower() )
# In addition, we can normalize the target text, e.g. removing new lines characters etc...
# note that order is important here!
UpperCAmelCase_ = ['''\n\n''', '''\n''', ''' ''', ''' ''']
for t in token_sequences_to_ignore:
UpperCAmelCase_ = ''' '''.join(text.split(__UpperCAmelCase ) )
return text
def A ( __UpperCAmelCase ) -> int:
'''simple docstring'''
UpperCAmelCase_ = load_dataset(args.dataset , args.config , split=args.split , use_auth_token=__UpperCAmelCase )
# for testing: only process the first two examples as a test
# dataset = dataset.select(range(10))
# load processor
UpperCAmelCase_ = AutoFeatureExtractor.from_pretrained(args.model_id )
UpperCAmelCase_ = feature_extractor.sampling_rate
# resample audio
UpperCAmelCase_ = dataset.cast_column('''audio''' , Audio(sampling_rate=__UpperCAmelCase ) )
# load eval pipeline
if args.device is None:
UpperCAmelCase_ = 0 if torch.cuda.is_available() else -1
UpperCAmelCase_ = pipeline('''automatic-speech-recognition''' , model=args.model_id , device=args.device )
# map function to decode audio
def map_to_pred(__UpperCAmelCase ):
UpperCAmelCase_ = asr(
batch['''audio''']['''array'''] , chunk_length_s=args.chunk_length_s , stride_length_s=args.stride_length_s )
UpperCAmelCase_ = prediction['''text''']
UpperCAmelCase_ = normalize_text(batch['''sentence'''] )
return batch
# run inference on all examples
UpperCAmelCase_ = dataset.map(__UpperCAmelCase , remove_columns=dataset.column_names )
# compute and log_results
# do not change function below
log_results(__UpperCAmelCase , __UpperCAmelCase )
if __name__ == "__main__":
UpperCamelCase_ = argparse.ArgumentParser()
parser.add_argument(
"--model_id", type=str, required=True, help="Model identifier. Should be loadable with 🤗 Transformers"
)
parser.add_argument(
"--dataset",
type=str,
required=True,
help="Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets",
)
parser.add_argument(
"--config", type=str, required=True, help="Config of the dataset. *E.g.* `'en'` for Common Voice"
)
parser.add_argument("--split", type=str, required=True, help="Split of the dataset. *E.g.* `'test'`")
parser.add_argument(
"--chunk_length_s", type=float, default=None, help="Chunk length in seconds. Defaults to 5 seconds."
)
parser.add_argument(
"--stride_length_s", type=float, default=None, help="Stride of the audio chunks. Defaults to 1 second."
)
parser.add_argument(
"--log_outputs", action="store_true", help="If defined, write outputs to log file for analysis."
)
parser.add_argument(
"--device",
type=int,
default=None,
help="The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.",
)
UpperCamelCase_ = parser.parse_args()
main(args)
| 344 |
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
UpperCamelCase_ = logging.get_logger(__name__)
UpperCamelCase_ = {
"facebook/levit-128S": "https://huggingface.co/facebook/levit-128S/resolve/main/config.json",
# See all LeViT models at https://huggingface.co/models?filter=levit
}
class a_ ( _snake_case ):
UpperCamelCase__ : Optional[int] ="levit"
def __init__( self :List[str] , _lowercase :List[Any]=224 , _lowercase :str=3 , _lowercase :Optional[int]=3 , _lowercase :str=2 , _lowercase :List[Any]=1 , _lowercase :str=16 , _lowercase :Dict=[128, 256, 384] , _lowercase :Union[str, Any]=[4, 8, 12] , _lowercase :Tuple=[4, 4, 4] , _lowercase :Dict=[16, 16, 16] , _lowercase :Any=0 , _lowercase :Dict=[2, 2, 2] , _lowercase :Any=[2, 2, 2] , _lowercase :Tuple=0.02 , **_lowercase :Union[str, Any] , ) -> Optional[Any]:
super().__init__(**_lowercase)
UpperCAmelCase_ = image_size
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = kernel_size
UpperCAmelCase_ = stride
UpperCAmelCase_ = padding
UpperCAmelCase_ = hidden_sizes
UpperCAmelCase_ = num_attention_heads
UpperCAmelCase_ = depths
UpperCAmelCase_ = key_dim
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = attention_ratio
UpperCAmelCase_ = mlp_ratio
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = [
['''Subsample''', key_dim[0], hidden_sizes[0] // key_dim[0], 4, 2, 2],
['''Subsample''', key_dim[0], hidden_sizes[1] // key_dim[0], 4, 2, 2],
]
class a_ ( _snake_case ):
UpperCamelCase__ : Union[str, Any] =version.parse("1.11" )
@property
def __a ( self :Any) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
])
@property
def __a ( self :List[Any]) -> float:
return 1E-4
| 344 | 1 |
import math
import numpy as np
import qiskit
from qiskit import Aer, ClassicalRegister, QuantumCircuit, QuantumRegister, execute
def A ( __UpperCAmelCase = 3 ) -> qiskit.result.counts.Counts:
'''simple docstring'''
if isinstance(__UpperCAmelCase , __UpperCAmelCase ):
raise TypeError('''number of qubits must be a integer.''' )
if number_of_qubits <= 0:
raise ValueError('''number of qubits must be > 0.''' )
if math.floor(__UpperCAmelCase ) != number_of_qubits:
raise ValueError('''number of qubits must be exact integer.''' )
if number_of_qubits > 10:
raise ValueError('''number of qubits too large to simulate(>10).''' )
UpperCAmelCase_ = QuantumRegister(__UpperCAmelCase , '''qr''' )
UpperCAmelCase_ = ClassicalRegister(__UpperCAmelCase , '''cr''' )
UpperCAmelCase_ = QuantumCircuit(__UpperCAmelCase , __UpperCAmelCase )
UpperCAmelCase_ = number_of_qubits
for i in range(__UpperCAmelCase ):
quantum_circuit.h(number_of_qubits - i - 1 )
counter -= 1
for j in range(__UpperCAmelCase ):
quantum_circuit.cp(np.pi / 2 ** (counter - j) , __UpperCAmelCase , __UpperCAmelCase )
for k in range(number_of_qubits // 2 ):
quantum_circuit.swap(__UpperCAmelCase , number_of_qubits - k - 1 )
# measure all the qubits
quantum_circuit.measure(__UpperCAmelCase , __UpperCAmelCase )
# simulate with 10000 shots
UpperCAmelCase_ = Aer.get_backend('''qasm_simulator''' )
UpperCAmelCase_ = execute(__UpperCAmelCase , __UpperCAmelCase , shots=1_0000 )
return job.result().get_counts(__UpperCAmelCase )
if __name__ == "__main__":
print(
f"Total count for quantum fourier transform state is: \
{quantum_fourier_transform(3)}"
)
| 344 |
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=None , __UpperCAmelCase="no" , __UpperCAmelCase="29500" ) -> int:
'''simple docstring'''
UpperCAmelCase_ = False
UpperCAmelCase_ = False
if any(key.startswith('''KAGGLE''' ) for key in os.environ.keys() ):
UpperCAmelCase_ = True
elif "IPython" in sys.modules:
UpperCAmelCase_ = '''google.colab''' in str(sys.modules['''IPython'''].get_ipython() )
try:
UpperCAmelCase_ = PrecisionType(mixed_precision.lower() )
except ValueError:
raise ValueError(
f"Unknown mixed_precision mode: {args.mixed_precision.lower()}. Choose between {PrecisionType.list()}." )
if (in_colab or in_kaggle) and (os.environ.get('''TPU_NAME''' , __UpperCAmelCase ) is not None):
# TPU launch
import torch_xla.distributed.xla_multiprocessing as xmp
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To train on TPU in Colab or Kaggle Kernel, the `Accelerator` should only be initialized inside '''
'''your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if num_processes is None:
UpperCAmelCase_ = 8
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''TPU''' )
print(f"Launching a training on {num_processes} TPU cores." )
xmp.spawn(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
elif in_colab:
# No need for a distributed launch otherwise as it's either CPU or one GPU.
if torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on one CPU.''' )
function(*__UpperCAmelCase )
else:
if num_processes is None:
raise ValueError(
'''You have to specify the number of GPUs you would like to use, add `num_processes=...` to your call.''' )
if num_processes > 1:
# Multi-GPU launch
from torch.multiprocessing import start_processes
from torch.multiprocessing.spawn import ProcessRaisedException
if len(AcceleratorState._shared_state ) > 0:
raise ValueError(
'''To launch a multi-GPU training from your notebook, the `Accelerator` should only be initialized '''
'''inside your training function. Restart your notebook and make sure no cells initializes an '''
'''`Accelerator`.''' )
if torch.cuda.is_initialized():
raise ValueError(
'''To launch a multi-GPU training from your notebook, you need to avoid running any instruction '''
'''using `torch.cuda` in any cell. Restart your notebook and make sure no cells use any CUDA '''
'''function.''' )
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port=__UpperCAmelCase , mixed_precision=__UpperCAmelCase ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , distributed_type='''MULTI_GPU''' )
print(f"Launching training on {num_processes} GPUs." )
try:
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
except ProcessRaisedException as e:
if "Cannot re-initialize CUDA in forked subprocess" in e.args[0]:
raise RuntimeError(
'''CUDA has been initialized before the `notebook_launcher` could create a forked subprocess. '''
'''This likely stems from an outside import causing issues once the `notebook_launcher()` is called. '''
'''Please review your imports and test them when running the `notebook_launcher()` to identify '''
'''which one is problematic.''' ) from e
else:
# No need for a distributed launch otherwise as it's either CPU, GPU or MPS.
if is_mps_available():
UpperCAmelCase_ = '''1'''
print('''Launching training on MPS.''' )
elif torch.cuda.is_available():
print('''Launching training on one GPU.''' )
else:
print('''Launching training on CPU.''' )
function(*__UpperCAmelCase )
def A ( __UpperCAmelCase , __UpperCAmelCase=() , __UpperCAmelCase=2 ) -> Optional[Any]:
'''simple docstring'''
from torch.multiprocessing import start_processes
with tempfile.NamedTemporaryFile() as tmp_file:
# torch.distributed will expect a few environment variable to be here. We set the ones common to each
# process here (the other ones will be set be the launcher).
with patch_environment(
world_size=__UpperCAmelCase , master_addr='''127.0.01''' , master_port='''29500''' , accelerate_mixed_precision='''no''' , accelerate_debug_rdv_file=tmp_file.name , accelerate_use_cpu='''yes''' , ):
UpperCAmelCase_ = PrepareForLaunch(__UpperCAmelCase , debug=__UpperCAmelCase )
start_processes(__UpperCAmelCase , args=__UpperCAmelCase , nprocs=__UpperCAmelCase , start_method='''fork''' )
| 344 | 1 |
import sacrebleu as scb
from packaging import version
from sacrebleu import CHRF
import datasets
UpperCamelCase_ = "\\n@inproceedings{popovic-2015-chrf,\n title = \"chr{F}: character n-gram {F}-score for automatic {MT} evaluation\",\n author = \"Popovi{\'c}, Maja\",\n booktitle = \"Proceedings of the Tenth Workshop on Statistical Machine Translation\",\n month = sep,\n year = \"2015\",\n address = \"Lisbon, Portugal\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/W15-3049\",\n doi = \"10.18653/v1/W15-3049\",\n pages = \"392--395\",\n}\n@inproceedings{popovic-2017-chrf,\n title = \"chr{F}++: words helping character n-grams\",\n author = \"Popovi{\'c}, Maja\",\n booktitle = \"Proceedings of the Second Conference on Machine Translation\",\n month = sep,\n year = \"2017\",\n address = \"Copenhagen, Denmark\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/W17-4770\",\n doi = \"10.18653/v1/W17-4770\",\n pages = \"612--618\",\n}\n@inproceedings{post-2018-call,\n title = \"A Call for Clarity in Reporting {BLEU} Scores\",\n author = \"Post, Matt\",\n booktitle = \"Proceedings of the Third Conference on Machine Translation: Research Papers\",\n month = oct,\n year = \"2018\",\n address = \"Belgium, Brussels\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W18-6319\",\n pages = \"186--191\",\n}\n"
UpperCamelCase_ = "\\nChrF and ChrF++ are two MT evaluation metrics. They both use the F-score statistic for character n-gram matches,\nand ChrF++ adds word n-grams as well which correlates more strongly with direct assessment. We use the implementation\nthat is already present in sacrebleu.\n\nThe implementation here is slightly different from sacrebleu in terms of the required input format. The length of\nthe references and hypotheses lists need to be the same, so you may need to transpose your references compared to\nsacrebleu's required input format. See https://github.com/huggingface/datasets/issues/3154#issuecomment-950746534\n\nSee the README.md file at https://github.com/mjpost/sacreBLEU#chrf--chrf for more information.\n"
UpperCamelCase_ = "\nProduces ChrF(++) scores for hypotheses given reference translations.\n\nArgs:\n predictions (list of str): The predicted sentences.\n references (list of list of str): The references. There should be one reference sub-list for each prediction sentence.\n char_order (int): Character n-gram order. Defaults to `6`.\n word_order (int): Word n-gram order. If equals to `2`, the metric is referred to as chrF++. Defaults to `0`.\n beta (int): Determine the importance of recall w.r.t precision. Defaults to `2`.\n lowercase (bool): if `True`, enables case-insensitivity. Defaults to `False`.\n whitespace (bool): If `True`, include whitespaces when extracting character n-grams.\n eps_smoothing (bool): If `True`, applies epsilon smoothing similar\n to reference chrF++.py, NLTK and Moses implementations. If `False`,\n it takes into account effective match order similar to sacreBLEU < 2.0.0. Defaults to `False`.\n\nReturns:\n 'score' (float): The chrF (chrF++) score,\n 'char_order' (int): The character n-gram order,\n 'word_order' (int): The word n-gram order. If equals to 2, the metric is referred to as chrF++,\n 'beta' (int): Determine the importance of recall w.r.t precision\n\nExamples:\n Example 1--a simple example of calculating chrF:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction, references=reference)\n >>> print(results)\n {'score': 84.64214891738334, 'char_order': 6, 'word_order': 0, 'beta': 2}\n\n Example 2--the same example, but with the argument word_order=2, to calculate chrF++ instead of chrF:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction,\n ... references=reference,\n ... word_order=2)\n >>> print(results)\n {'score': 82.87263732906315, 'char_order': 6, 'word_order': 2, 'beta': 2}\n\n Example 3--the same chrF++ example as above, but with `lowercase=True` to normalize all case:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction,\n ... references=reference,\n ... word_order=2,\n ... lowercase=True)\n >>> print(results)\n {'score': 92.12853119829202, 'char_order': 6, 'word_order': 2, 'beta': 2}\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class a_ ( datasets.Metric ):
def __a ( self :Any) -> List[str]:
if version.parse(scb.__version__) < version.parse('''1.4.12'''):
raise ImportWarning(
'''To use `sacrebleu`, the module `sacrebleu>=1.4.12` is required, and the current version of `sacrebleu` doesn\'t match this condition.\n'''
'''You can install it with `pip install "sacrebleu>=1.4.12"`.''')
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , homepage='''https://github.com/mjpost/sacreBLEU#chrf--chrf''' , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
'''predictions''': datasets.Value('''string''' , id='''sequence'''),
'''references''': datasets.Sequence(datasets.Value('''string''' , id='''sequence''') , id='''references'''),
}) , codebase_urls=['''https://github.com/mjpost/sacreBLEU#chrf--chrf'''] , reference_urls=[
'''https://github.com/m-popovic/chrF''',
] , )
def __a ( self :Dict , _lowercase :Tuple , _lowercase :str , _lowercase :int = CHRF.CHAR_ORDER , _lowercase :int = CHRF.WORD_ORDER , _lowercase :int = CHRF.BETA , _lowercase :bool = False , _lowercase :bool = False , _lowercase :bool = False , ) -> Any:
UpperCAmelCase_ = len(references[0])
if any(len(_lowercase) != references_per_prediction for refs in references):
raise ValueError('''Sacrebleu requires the same number of references for each prediction''')
UpperCAmelCase_ = [[refs[i] for refs in references] for i in range(_lowercase)]
UpperCAmelCase_ = CHRF(_lowercase , _lowercase , _lowercase , _lowercase , _lowercase , _lowercase)
UpperCAmelCase_ = sb_chrf.corpus_score(_lowercase , _lowercase)
return {
"score": output.score,
"char_order": output.char_order,
"word_order": output.word_order,
"beta": output.beta,
}
| 344 |
import math
import os
import re
import sys
import unittest
from pathlib import Path
from typing import Tuple
from unittest.mock import patch
from parameterized import parameterized
from transformers.testing_utils import (
CaptureStderr,
ExtendSysPath,
TestCasePlus,
execute_subprocess_async,
get_gpu_count,
get_torch_dist_unique_port,
require_apex,
require_bitsandbytes,
require_fairscale,
require_torch,
require_torch_gpu,
require_torch_multi_gpu,
require_torch_non_multi_gpu,
slow,
)
from transformers.trainer_callback import TrainerState
from transformers.trainer_utils import set_seed
UpperCamelCase_ = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f"{bindir}/../../examples/pytorch/translation"):
from run_translation import main # noqa
set_seed(42)
UpperCamelCase_ = "sshleifer/student_marian_en_ro_6_1"
UpperCamelCase_ = "sshleifer/tiny-mbart"
@require_torch
class a_ ( _snake_case ):
def __a ( self :str , _lowercase :Any=False , _lowercase :Tuple=None , _lowercase :Dict=True , _lowercase :Tuple=True , _lowercase :List[Any]=True , _lowercase :List[str]=True , ) -> int:
UpperCAmelCase_ = self.run_trainer(
eval_steps=1 , max_len=12 , model_name=_lowercase , num_train_epochs=1 , distributed=_lowercase , extra_args_str=_lowercase , predict_with_generate=_lowercase , do_train=_lowercase , do_eval=_lowercase , do_predict=_lowercase , )
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
if not do_eval:
return
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
if predict_with_generate:
assert "eval_bleu" in first_step_stats
UpperCAmelCase_ = eval_metrics[-1]
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
assert not math.isnan(float(last_step_stats['''eval_loss'''])), "eval_loss must not be `nan`"
@require_torch_non_multi_gpu
def __a ( self :Dict) -> str:
self.run_seqaseq_quick()
@require_torch_multi_gpu
def __a ( self :Any) -> int:
self.run_seqaseq_quick(distributed=_lowercase)
@require_torch_multi_gpu
def __a ( self :int) -> Any:
self.run_seqaseq_quick(distributed=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Tuple) -> List[str]:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp simple --fp16''')
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :Union[str, Any]) -> Any:
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=_lowercase)
@unittest.skip('''Requires an update of the env running those tests''')
@require_torch_multi_gpu
@require_fairscale
def __a ( self :int) -> Any:
self.run_seqaseq_quick(
distributed=_lowercase , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=_lowercase)
@require_apex
@require_torch_gpu
def __a ( self :Tuple) -> str:
# XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same
# program and it breaks other tests that run from the same pytest worker, therefore until this is
# sorted out it must be run only in an external program, that is distributed=True in this
# test and only under one or more gpus - if we want cpu will need to make a special test
#
# specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via
# 2nd main() call it botches the future eval.
#
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
# test 2nd time - was getting eval_loss': nan'
# to reproduce the problem set distributed=False
self.run_seqaseq_quick(distributed=_lowercase , extra_args_str='''--fp16 --fp16_backend=apex''')
@parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''])
@require_torch_multi_gpu
def __a ( self :str , _lowercase :Any) -> List[str]:
# as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout
UpperCAmelCase_ = {
# test with the default log_level - should be info and thus log info once
'''base''': {'''extra_args_str''': '''''', '''n_matches''': 1},
# test with low log_level and log_level_replica - should be noisy on all processes
# now the info string should appear twice on 2 processes
'''low''': {'''extra_args_str''': '''--log_level debug --log_level_replica debug''', '''n_matches''': 2},
# test with high log_level and low log_level_replica
# now the info string should appear once only on the replica
'''high''': {'''extra_args_str''': '''--log_level error --log_level_replica debug''', '''n_matches''': 1},
# test with high log_level and log_level_replica - should be quiet on all processes
'''mixed''': {'''extra_args_str''': '''--log_level error --log_level_replica error''', '''n_matches''': 0},
}
UpperCAmelCase_ = experiments[experiment_id]
UpperCAmelCase_ = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False}
UpperCAmelCase_ = '''Running training'''
with CaptureStderr() as cl:
self.run_seqaseq_quick(**_lowercase , extra_args_str=data['''extra_args_str'''])
UpperCAmelCase_ = len(re.findall(_lowercase , cl.err))
self.assertEqual(_lowercase , data['''n_matches'''])
@slow
def __a ( self :Any) -> Dict:
UpperCAmelCase_ = self.run_trainer(
eval_steps=2 , max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=10 , distributed=_lowercase , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(os.path.join(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = [log for log in logs if '''eval_loss''' in log.keys()]
UpperCAmelCase_ = eval_metrics[0]
UpperCAmelCase_ = eval_metrics[-1]
assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing"
assert isinstance(last_step_stats['''eval_bleu'''] , _lowercase)
# test if do_predict saves generations and metrics
UpperCAmelCase_ = os.listdir(_lowercase)
UpperCAmelCase_ = {os.path.basename(_lowercase) for p in contents}
assert "generated_predictions.txt" in contents
assert "predict_results.json" in contents
@slow
@require_bitsandbytes
def __a ( self :List[str]) -> str:
from transformers.training_args import OptimizerNames
def train_and_return_metrics(_lowercase :str) -> Tuple[int, float]:
UpperCAmelCase_ = '''--skip_memory_metrics 0'''
UpperCAmelCase_ = self.run_trainer(
max_len=128 , model_name=_lowercase , learning_rate=3E-4 , num_train_epochs=1 , optim=_lowercase , distributed=_lowercase , extra_args_str=_lowercase , do_eval=_lowercase , do_predict=_lowercase , n_gpus_to_use=1 , )
# Check metrics
UpperCAmelCase_ = TrainerState.load_from_json(Path(_lowercase , '''trainer_state.json''')).log_history
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20)
UpperCAmelCase_ = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20)
UpperCAmelCase_ = logs[0]['''train_loss''']
return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value)
UpperCAmelCase_ , UpperCAmelCase_ , UpperCAmelCase_ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value)
UpperCAmelCase_ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_peak_mem_orig + gpu_alloc_mem_orig
UpperCAmelCase_ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb
UpperCAmelCase_ = gpu_total_mem_orig - gpu_total_mem_bnb
# sshleifer/student_marian_en_ro_6_1 has 54M parameter, 29M of which is `nn.Embedding` which
# doesn't get quantized and remains in fp32. Therefore we only have 25M parameters quantized
# in 2 bytes and the diff in optim memory usage is derived as so:
#
# - normal 25*8=~200MB (8 bytes per param)
# - bnb 25*2= ~50MB (2 bytes per param)
#
# Thus we should expect ~150MB total memory saved.
#
# Peak memory should be the same - the total should be different by about that same margin
#
# After leaving a small margin to accommodate for differences between gpus let's check
# that we have at least 120MB in savings
UpperCAmelCase_ = 120
# uncomment the following if this test starts failing - requires py38 for a new print feature
# gpu_peak_mem_diff = gpu_peak_mem_orig - gpu_peak_mem_bnb
# print(f"{gpu_alloc_mem_orig=}MB {gpu_peak_mem_orig=}MB {gpu_alloc_mem_orig+gpu_peak_mem_orig=}MB")
# print(f" {gpu_alloc_mem_bnb=}MB {gpu_peak_mem_bnb=}MB {gpu_alloc_mem_bnb+gpu_peak_mem_bnb=}MB")
# print(f"{gpu_alloc_mem_diff=}MB")
# print(f"{gpu_peak_mem_diff=}MB")
# print(f"{gpu_total_mem_orig=}MB, {gpu_total_mem_bnb=}MB")
# print(f"{gpu_total_mem_diff=}MB, {gpu_total_mem_diff=}MB")
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less alloc gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_alloc_mem_diff}MB, with gpu_alloc_mem_orig={gpu_alloc_mem_orig}MB and"
f" gpu_alloc_mem_bnb={gpu_alloc_mem_bnb}MB" , )
self.assertGreater(
_lowercase , _lowercase , '''should use ~150MB less total gpu memory with BNB, compared to without it for this model but got'''
f" a difference of {gpu_total_mem_diff}MB, with gpu_total_mem_orig={gpu_total_mem_orig}MB and"
f" gpu_total_mem_bnb={gpu_total_mem_bnb}MB" , )
self.assertEqual(
_lowercase , _lowercase , f"loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}")
def __a ( self :Any , _lowercase :int , _lowercase :str , _lowercase :int , _lowercase :float = 3E-3 , _lowercase :str = "adafactor" , _lowercase :bool = False , _lowercase :str = None , _lowercase :int = 0 , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :bool = True , _lowercase :int = None , ) -> List[Any]:
UpperCAmelCase_ = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro'''
UpperCAmelCase_ = self.get_auto_remove_tmp_dir()
UpperCAmelCase_ = f"\n --model_name_or_path {model_name}\n --train_file {data_dir}/train.json\n --validation_file {data_dir}/val.json\n --test_file {data_dir}/test.json\n --output_dir {output_dir}\n --overwrite_output_dir\n --max_train_samples 8\n --max_source_length {max_len}\n --max_target_length {max_len}\n --do_train\n --num_train_epochs {str(_lowercase)}\n --per_device_train_batch_size 4\n --learning_rate {learning_rate}\n --warmup_steps 8\n --logging_steps 0\n --logging_strategy no\n --save_steps {str(_lowercase)}\n --group_by_length\n --label_smoothing_factor 0.1\n --target_lang ro_RO\n --source_lang en_XX\n ".split()
UpperCAmelCase_ = f"\n --do_eval\n --per_device_eval_batch_size 4\n --max_eval_samples 8\n --val_max_target_length {max_len}\n --evaluation_strategy steps\n --eval_steps {str(_lowercase)}\n ".split()
UpperCAmelCase_ = '''
--do_predict
'''.split()
UpperCAmelCase_ = []
if do_train:
args += args_train
if do_eval:
args += args_eval
if do_predict:
args += args_predict
if predict_with_generate:
args += "--predict_with_generate".split()
if do_train:
if optim == "adafactor":
args += "--adafactor".split()
else:
args += f"--optim {optim}".split()
if extra_args_str is not None:
args += extra_args_str.split()
if distributed:
if n_gpus_to_use is None:
UpperCAmelCase_ = get_gpu_count()
UpperCAmelCase_ = get_torch_dist_unique_port()
UpperCAmelCase_ = f"\n -m torch.distributed.run\n --nproc_per_node={n_gpus_to_use}\n --master_port={master_port}\n {self.examples_dir_str}/pytorch/translation/run_translation.py\n ".split()
UpperCAmelCase_ = [sys.executable] + distributed_args + args
# keep for quick debug
# print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die
execute_subprocess_async(_lowercase , env=self.get_env())
else:
UpperCAmelCase_ = ['''run_translation.py'''] + args
with patch.object(_lowercase , '''argv''' , _lowercase):
main()
return output_dir
| 344 | 1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.