code
stringlengths 87
55.2k
| code_codestyle
int64 0
349
| style_context
stringlengths 135
49.1k
| style_context_codestyle
int64 0
349
| label
int64 0
1
|
|---|---|---|---|---|
'''simple docstring'''
from typing import Dict
from .base import GenericTensor, Pipeline
class a ( _lowerCamelCase ):
def A_ ( self : Any , lowercase_ : Optional[Any]=None , lowercase_ : Any=None , lowercase_ : List[Any]=None , **lowercase_ : Optional[Any] ):
if tokenize_kwargs is None:
snake_case_ = {}
if truncation is not None:
if "truncation" in tokenize_kwargs:
raise ValueError(
'''truncation parameter defined twice (given as keyword argument as well as in tokenize_kwargs)''' )
snake_case_ = truncation
snake_case_ = tokenize_kwargs
snake_case_ = {}
if return_tensors is not None:
snake_case_ = return_tensors
return preprocess_params, {}, postprocess_params
def A_ ( self : List[str] , lowercase_ : List[str] , **lowercase_ : Tuple ):
snake_case_ = self.framework
snake_case_ = self.tokenizer(lowercase_ , return_tensors=lowercase_ , **lowercase_ )
return model_inputs
def A_ ( self : Union[str, Any] , lowercase_ : Optional[Any] ):
snake_case_ = self.model(**lowercase_ )
return model_outputs
def A_ ( self : Any , lowercase_ : Tuple , lowercase_ : Optional[int]=False ):
# [0] is the first available tensor, logits or last_hidden_state.
if return_tensors:
return model_outputs[0]
if self.framework == "pt":
return model_outputs[0].tolist()
elif self.framework == "tf":
return model_outputs[0].numpy().tolist()
def __call__( self : Any , *lowercase_ : List[Any] , **lowercase_ : Union[str, Any] ):
return super().__call__(*lowercase_ , **lowercase_ )
| 56
|
'''simple docstring'''
import os
import re
import warnings
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_ta import TaTokenizer
else:
a : Dict = None
a : List[Any] = logging.get_logger(__name__)
a : List[Any] = {'vocab_file': 'spiece.model', 'tokenizer_file': 'tokenizer.json'}
a : str = {
'vocab_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/spiece.model',
't5-base': 'https://huggingface.co/t5-base/resolve/main/spiece.model',
't5-large': 'https://huggingface.co/t5-large/resolve/main/spiece.model',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/spiece.model',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/spiece.model',
},
'tokenizer_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/tokenizer.json',
't5-base': 'https://huggingface.co/t5-base/resolve/main/tokenizer.json',
't5-large': 'https://huggingface.co/t5-large/resolve/main/tokenizer.json',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/tokenizer.json',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/tokenizer.json',
},
}
# TODO(PVP) - this should be removed in Transformers v5
a : List[Any] = {
't5-small': 512,
't5-base': 512,
't5-large': 512,
't5-3b': 512,
't5-11b': 512,
}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
snake_case_ = TaTokenizer
snake_case_ = []
def __init__( self : List[Any] , lowercase_ : int=None , lowercase_ : Dict=None , lowercase_ : Dict="</s>" , lowercase_ : List[Any]="<unk>" , lowercase_ : int="<pad>" , lowercase_ : int=100 , lowercase_ : List[Any]=None , **lowercase_ : List[str] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
snake_case_ = [F"<extra_id_{i}>" for i in range(lowercase_ )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra special tokens
snake_case_ = len(set(filter(lambda lowercase_ : bool('''extra_id_''' in str(lowercase_ ) ) , lowercase_ ) ) )
if extra_tokens != extra_ids:
raise ValueError(
F"Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are"
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
super().__init__(
lowercase_ , tokenizer_file=lowercase_ , eos_token=lowercase_ , unk_token=lowercase_ , pad_token=lowercase_ , extra_ids=lowercase_ , additional_special_tokens=lowercase_ , **lowercase_ , )
snake_case_ = vocab_file
snake_case_ = False if not self.vocab_file else True
snake_case_ = extra_ids
@staticmethod
def A_ ( lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : int ):
if pretrained_model_name_or_path in TaTokenizerFast.max_model_input_sizes:
snake_case_ = TaTokenizerFast.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
F" {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this"
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
F" {pretrained_model_name_or_path} automatically truncating your input to"
F" {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences"
F" longer than {deprecated_max_model_length} you can either instantiate this tokenizer with"
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , lowercase_ , )
return max_model_length
def A_ ( self : Optional[Any] , lowercase_ : str , lowercase_ : Optional[str] = None ):
if not self.can_save_slow_tokenizer:
raise ValueError(
'''Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '''
'''tokenizer.''' )
if not os.path.isdir(lowercase_ ):
logger.error(F"Vocabulary path ({save_directory}) should be a directory" )
return
snake_case_ = 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_ ):
copyfile(self.vocab_file , lowercase_ )
logger.info(F"Copy vocab file to {out_vocab_file}" )
return (out_vocab_file,)
def A_ ( self : Optional[int] , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = token_ids_a + [self.eos_token_id]
if token_ids_a is None:
return self.prefix_tokens + token_ids_a
else:
snake_case_ = token_ids_a + [self.eos_token_id]
return self.prefix_tokens + token_ids_a + token_ids_a
def A_ ( self : int , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def A_ ( self : Dict ):
return list(
set(filter(lambda lowercase_ : bool(re.search(R'''<extra_id_\d+>''' , lowercase_ ) ) is not None , self.additional_special_tokens ) ) )
def A_ ( self : Any ):
return [self.convert_tokens_to_ids(lowercase_ ) for token in self.get_sentinel_tokens()]
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = abs(__UpperCAmelCase )
snake_case_ = 0
while n > 0:
res += n % 10
n //= 10
return res
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = abs(__UpperCAmelCase )
return n if n < 10 else n % 10 + sum_of_digits(n // 10 )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
return sum(int(__UpperCAmelCase ) for c in str(abs(__UpperCAmelCase ) ) )
def __magic_name__ ( ) -> None:
'''simple docstring'''
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(__UpperCAmelCase, __UpperCAmelCase ) -> None:
snake_case_ = F"{func.__name__}({value})"
snake_case_ = timeit(F"__main__.{call}", setup='''import __main__''' )
print(F"{call:56} = {func(__UpperCAmelCase )} -- {timing:.4f} seconds" )
for value in (26_2144, 1125_8999_0684_2624, 126_7650_6002_2822_9401_4967_0320_5376):
for func in (sum_of_digits, sum_of_digits_recursion, sum_of_digits_compact):
benchmark_a_function(__UpperCAmelCase, __UpperCAmelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 56
|
'''simple docstring'''
from __future__ import annotations
import math
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
if depth < 0:
raise ValueError('''Depth cannot be less than 0''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Scores cannot be empty''' )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
return min(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ = [90, 23, 6, 33, 21, 65, 123, 3_4423]
snake_case_ = math.log(len(__UpperCAmelCase ), 2 )
print('''Optimal value : ''', end='''''' )
print(minimax(0, 0, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 56
| 1
|
'''simple docstring'''
import unittest
from transformers import (
MODEL_FOR_OBJECT_DETECTION_MAPPING,
AutoFeatureExtractor,
AutoModelForObjectDetection,
ObjectDetectionPipeline,
is_vision_available,
pipeline,
)
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_pytesseract,
require_tf,
require_timm,
require_torch,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_vision_available():
from PIL import Image
else:
class a :
@staticmethod
def A_ ( *lowercase_ : int , **lowercase_ : str ):
pass
@is_pipeline_test
@require_vision
@require_timm
@require_torch
class a ( unittest.TestCase ):
snake_case_ = MODEL_FOR_OBJECT_DETECTION_MAPPING
def A_ ( self : Any , lowercase_ : List[Any] , lowercase_ : Optional[int] , lowercase_ : List[str] ):
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , image_processor=lowercase_ )
return object_detector, ["./tests/fixtures/tests_samples/COCO/000000039769.png"]
def A_ ( self : Dict , lowercase_ : List[Any] , lowercase_ : int ):
snake_case_ = object_detector('''./tests/fixtures/tests_samples/COCO/000000039769.png''' , threshold=0.0 )
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
import datasets
snake_case_ = datasets.load_dataset('''hf-internal-testing/fixtures_image_utils''' , '''image''' , split='''test''' )
snake_case_ = [
Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' ),
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
# RGBA
dataset[0]['''file'''],
# LA
dataset[1]['''file'''],
# L
dataset[2]['''file'''],
]
snake_case_ = object_detector(lowercase_ , threshold=0.0 )
self.assertEqual(len(lowercase_ ) , len(lowercase_ ) )
for outputs in batch_outputs:
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
@require_tf
@unittest.skip('''Object detection not implemented in TF''' )
def A_ ( self : int ):
pass
@require_torch
def A_ ( self : Tuple ):
snake_case_ = '''hf-internal-testing/tiny-detr-mobilenetsv3'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=0.0 )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] , threshold=0.0 , )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
] , )
@require_torch
@slow
def A_ ( self : Optional[int] ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : Tuple ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : str ):
snake_case_ = 0.9985
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=lowercase_ )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
@require_torch
@require_pytesseract
@slow
def A_ ( self : Dict ):
snake_case_ = '''Narsil/layoutlmv3-finetuned-funsd'''
snake_case_ = 0.9993
snake_case_ = pipeline('''object-detection''' , model=lowercase_ , threshold=lowercase_ )
snake_case_ = object_detector(
'''https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
] , )
| 56
|
'''simple docstring'''
import argparse
import torch
from torch import nn
from transformers import SpeechaTextConfig, SpeechaTextForConditionalGeneration
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = [
'''encoder.version''',
'''decoder.version''',
'''model.encoder.version''',
'''model.decoder.version''',
'''decoder.output_projection.weight''',
'''_float_tensor''',
'''encoder.embed_positions._float_tensor''',
'''decoder.embed_positions._float_tensor''',
]
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = list(s_dict.keys() )
for key in keys:
if "transformer_layers" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
elif "subsample" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ ,snake_case_ = emb.weight.shape
snake_case_ = nn.Linear(__UpperCAmelCase, __UpperCAmelCase, bias=__UpperCAmelCase )
snake_case_ = emb.weight.data
return lin_layer
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )
snake_case_ = mam_aaa['''args''']
snake_case_ = mam_aaa['''model''']
snake_case_ = state_dict['''decoder.output_projection.weight''']
remove_ignore_keys_(__UpperCAmelCase )
rename_keys(__UpperCAmelCase )
snake_case_ = state_dict['''decoder.embed_tokens.weight'''].shape[0]
snake_case_ = args.share_decoder_input_output_embed
snake_case_ = [int(__UpperCAmelCase ) for i in args.conv_kernel_sizes.split(''',''' )]
snake_case_ = SpeechaTextConfig(
vocab_size=__UpperCAmelCase, max_source_positions=args.max_source_positions, max_target_positions=args.max_target_positions, encoder_layers=args.encoder_layers, decoder_layers=args.decoder_layers, encoder_attention_heads=args.encoder_attention_heads, decoder_attention_heads=args.decoder_attention_heads, encoder_ffn_dim=args.encoder_ffn_embed_dim, decoder_ffn_dim=args.decoder_ffn_embed_dim, d_model=args.encoder_embed_dim, dropout=args.dropout, attention_dropout=args.attention_dropout, activation_dropout=args.activation_dropout, activation_function='''relu''', num_conv_layers=len(__UpperCAmelCase ), conv_channels=args.conv_channels, conv_kernel_sizes=__UpperCAmelCase, input_feat_per_channel=args.input_feat_per_channel, input_channels=args.input_channels, tie_word_embeddings=__UpperCAmelCase, num_beams=5, max_length=200, use_cache=__UpperCAmelCase, decoder_start_token_id=2, early_stopping=__UpperCAmelCase, )
snake_case_ = SpeechaTextForConditionalGeneration(__UpperCAmelCase )
snake_case_ ,snake_case_ = model.model.load_state_dict(__UpperCAmelCase, strict=__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0 and not set(__UpperCAmelCase ) <= {
"encoder.embed_positions.weights",
"decoder.embed_positions.weights",
}:
raise ValueError(
'''Only `encoder.embed_positions.weights` and `decoder.embed_positions.weights` are allowed to be missing,'''
F" but all the following weights are missing {missing}" )
if tie_embeds:
snake_case_ = make_linear_from_emb(model.model.decoder.embed_tokens )
else:
snake_case_ = lm_head_weights
model.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
a : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument('--fairseq_path', type=str, help='Path to the fairseq model (.pt) file.')
parser.add_argument('--pytorch_dump_folder_path', default=None, type=str, help='Path to the output PyTorch model.')
a : List[Any] = parser.parse_args()
convert_fairseq_sat_checkpoint_to_tfms(args.fairseq_path, args.pytorch_dump_folder_path)
| 56
| 1
|
'''simple docstring'''
import unittest
import numpy as np
from diffusers import LMSDiscreteScheduler, OnnxStableDiffusionInpaintPipeline
from diffusers.utils.testing_utils import (
is_onnx_available,
load_image,
nightly,
require_onnxruntime,
require_torch_gpu,
)
from ..test_pipelines_onnx_common import OnnxPipelineTesterMixin
if is_onnx_available():
import onnxruntime as ort
class a ( _lowerCamelCase , unittest.TestCase ):
# FIXME: add fast tests
pass
@nightly
@require_onnxruntime
@require_torch_gpu
class a ( unittest.TestCase ):
@property
def A_ ( self : List[Any] ):
return (
"CUDAExecutionProvider",
{
"gpu_mem_limit": "15000000000", # 15GB
"arena_extend_strategy": "kSameAsRequested",
},
)
@property
def A_ ( self : Union[str, Any] ):
snake_case_ = ort.SessionOptions()
snake_case_ = False
return options
def A_ ( self : Tuple ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo.png''' )
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo_mask.png''' )
snake_case_ = OnnxStableDiffusionInpaintPipeline.from_pretrained(
'''runwayml/stable-diffusion-inpainting''' , revision='''onnx''' , safety_checker=lowercase_ , feature_extractor=lowercase_ , provider=self.gpu_provider , sess_options=self.gpu_options , )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = '''A red cat sitting on a park bench'''
snake_case_ = np.random.RandomState(0 )
snake_case_ = pipe(
prompt=lowercase_ , image=lowercase_ , mask_image=lowercase_ , guidance_scale=7.5 , num_inference_steps=10 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
snake_case_ = images[0, 255:258, 255:258, -1]
assert images.shape == (1, 512, 512, 3)
snake_case_ = np.array([0.2514, 0.3007, 0.3517, 0.1790, 0.2382, 0.3167, 0.1944, 0.2273, 0.2464] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
def A_ ( self : Tuple ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo.png''' )
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/in_paint/overture-creations-5sI6fQgYIuo_mask.png''' )
snake_case_ = LMSDiscreteScheduler.from_pretrained(
'''runwayml/stable-diffusion-inpainting''' , subfolder='''scheduler''' , revision='''onnx''' )
snake_case_ = OnnxStableDiffusionInpaintPipeline.from_pretrained(
'''runwayml/stable-diffusion-inpainting''' , revision='''onnx''' , scheduler=lowercase_ , safety_checker=lowercase_ , feature_extractor=lowercase_ , provider=self.gpu_provider , sess_options=self.gpu_options , )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = '''A red cat sitting on a park bench'''
snake_case_ = np.random.RandomState(0 )
snake_case_ = pipe(
prompt=lowercase_ , image=lowercase_ , mask_image=lowercase_ , guidance_scale=7.5 , num_inference_steps=20 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
snake_case_ = images[0, 255:258, 255:258, -1]
assert images.shape == (1, 512, 512, 3)
snake_case_ = np.array([0.0086, 0.0077, 0.0083, 0.0093, 0.0107, 0.0139, 0.0094, 0.0097, 0.0125] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
| 56
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class a ( metaclass=_lowerCamelCase ):
snake_case_ = ["transformers", "torch", "note_seq"]
def __init__( self : Union[str, Any] , *lowercase_ : Optional[int] , **lowercase_ : int ):
requires_backends(self , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : List[Any] , *lowercase_ : Optional[Any] , **lowercase_ : str ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : Tuple , *lowercase_ : Union[str, Any] , **lowercase_ : List[Any] ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase = 200 ) -> int:
'''simple docstring'''
snake_case_ = [1, 2, 5, 10, 20, 50, 100, 200]
snake_case_ = [0] * (pence + 1)
snake_case_ = 1 # base case: 1 way to make 0 pence
for coin in coins:
for i in range(__UpperCAmelCase, pence + 1, 1 ):
number_of_ways[i] += number_of_ways[i - coin]
return number_of_ways[pence]
if __name__ == "__main__":
assert solution(200) == 7_3682
| 56
|
'''simple docstring'''
# tests directory-specific settings - this file is run automatically
# by pytest before any tests are run
import doctest
import sys
import warnings
from os.path import abspath, dirname, join
import _pytest
from transformers.testing_utils import HfDoctestModule, HfDocTestParser
# allow having multiple repository checkouts and not needing to remember to rerun
# 'pip install -e .[dev]' when switching between checkouts and running tests.
a : int = abspath(join(dirname(__file__), 'src'))
sys.path.insert(1, git_repo_path)
# silence FutureWarning warnings in tests since often we can't act on them until
# they become normal warnings - i.e. the tests still need to test the current functionality
warnings.simplefilter(action='ignore', category=FutureWarning)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
config.addinivalue_line(
'''markers''', '''is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested''' )
config.addinivalue_line(
'''markers''', '''is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested''' )
config.addinivalue_line('''markers''', '''is_pipeline_test: mark test to run only when pipelines are tested''' )
config.addinivalue_line('''markers''', '''is_staging_test: mark test to run only in the staging environment''' )
config.addinivalue_line('''markers''', '''accelerate_tests: mark test that require accelerate''' )
config.addinivalue_line('''markers''', '''tool_tests: mark the tool tests that are run on their specific schedule''' )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from transformers.testing_utils import pytest_addoption_shared
pytest_addoption_shared(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
from transformers.testing_utils import pytest_terminal_summary_main
snake_case_ = terminalreporter.config.getoption('''--make-reports''' )
if make_reports:
pytest_terminal_summary_main(__UpperCAmelCase, id=__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if exitstatus == 5:
snake_case_ = 0
# Doctest custom flag to ignore output.
a : Union[str, Any] = doctest.register_optionflag('IGNORE_RESULT')
a : Optional[int] = doctest.OutputChecker
class a ( _lowerCamelCase ):
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int] ):
if IGNORE_RESULT & optionflags:
return True
return OutputChecker.check_output(self , lowercase_ , lowercase_ , lowercase_ )
a : List[Any] = CustomOutputChecker
a : Optional[int] = HfDoctestModule
a : Tuple = HfDocTestParser
| 56
| 1
|
'''simple docstring'''
from .data_collator import (
DataCollatorForLanguageModeling,
DataCollatorForPermutationLanguageModeling,
DataCollatorForSeqaSeq,
DataCollatorForSOP,
DataCollatorForTokenClassification,
DataCollatorForWholeWordMask,
DataCollatorWithPadding,
DefaultDataCollator,
default_data_collator,
)
from .metrics import glue_compute_metrics, xnli_compute_metrics
from .processors import (
DataProcessor,
InputExample,
InputFeatures,
SingleSentenceClassificationProcessor,
SquadExample,
SquadFeatures,
SquadVaProcessor,
SquadVaProcessor,
glue_convert_examples_to_features,
glue_output_modes,
glue_processors,
glue_tasks_num_labels,
squad_convert_examples_to_features,
xnli_output_modes,
xnli_processors,
xnli_tasks_num_labels,
)
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import TensorType, is_torch_available, logging
a : Dict = logging.get_logger(__name__)
a : List[str] = {
'Helsinki-NLP/opus-mt-en-de': 'https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/config.json',
# See all Marian models at https://huggingface.co/models?filter=marian
}
class a ( _lowerCamelCase ):
snake_case_ = "marian"
snake_case_ = ["past_key_values"]
snake_case_ = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self : List[Any] , lowercase_ : Optional[Any]=5_8101 , lowercase_ : Dict=None , lowercase_ : List[str]=1024 , lowercase_ : Optional[Any]=12 , lowercase_ : int=4096 , lowercase_ : Any=16 , lowercase_ : Optional[int]=12 , lowercase_ : str=4096 , lowercase_ : Union[str, Any]=16 , lowercase_ : Dict=0.0 , lowercase_ : Union[str, Any]=0.0 , lowercase_ : Optional[Any]=True , lowercase_ : Union[str, Any]=True , lowercase_ : int="gelu" , lowercase_ : Dict=1024 , lowercase_ : int=0.1 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.02 , lowercase_ : int=5_8100 , lowercase_ : Optional[Any]=False , lowercase_ : Any=5_8100 , lowercase_ : Optional[int]=0 , lowercase_ : Dict=0 , lowercase_ : List[str]=True , **lowercase_ : Any , ):
snake_case_ = vocab_size
snake_case_ = decoder_vocab_size or vocab_size
snake_case_ = max_position_embeddings
snake_case_ = d_model
snake_case_ = encoder_ffn_dim
snake_case_ = encoder_layers
snake_case_ = encoder_attention_heads
snake_case_ = decoder_ffn_dim
snake_case_ = decoder_layers
snake_case_ = decoder_attention_heads
snake_case_ = dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = activation_function
snake_case_ = init_std
snake_case_ = encoder_layerdrop
snake_case_ = decoder_layerdrop
snake_case_ = use_cache
snake_case_ = encoder_layers
snake_case_ = scale_embedding # scale factor will be sqrt(d_model) if True
snake_case_ = share_encoder_decoder_embeddings
super().__init__(
pad_token_id=lowercase_ , eos_token_id=lowercase_ , is_encoder_decoder=lowercase_ , decoder_start_token_id=lowercase_ , forced_eos_token_id=lowercase_ , **lowercase_ , )
class a ( _lowerCamelCase ):
@property
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.inputs
def A_ ( self : Union[str, Any] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ = {0: '''batch'''}
snake_case_ = {0: '''batch''', 1: '''past_decoder_sequence + sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''decoder_sequence'''}
snake_case_ = {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.
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
else:
snake_case_ = 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
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.outputs
def A_ ( self : Dict ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super().outputs
else:
snake_case_ = super(lowercase_ , self ).outputs
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
return common_outputs
def A_ ( self : Dict , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
# Generate decoder inputs
snake_case_ = seq_length if not self.use_past else 1
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
snake_case_ = {F"decoder_{name}": tensor for name, tensor in decoder_inputs.items()}
snake_case_ = 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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
snake_case_ = common_inputs['''decoder_input_ids'''].shape[1]
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = decoder_seq_length + 3
snake_case_ = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
snake_case_ = torch.cat(
[common_inputs['''decoder_attention_mask'''], torch.ones(lowercase_ , lowercase_ )] , dim=1 )
snake_case_ = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
snake_case_ ,snake_case_ = self.num_layers
snake_case_ = min(lowercase_ , lowercase_ )
snake_case_ = max(lowercase_ , lowercase_ ) - min_num_layers
snake_case_ = '''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.
snake_case_ = 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 : Union[str, Any] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
# Not using the same length for past_key_values
snake_case_ = seqlen + 2
snake_case_ ,snake_case_ = self.num_layers
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = common_inputs['''attention_mask'''].dtype
snake_case_ = torch.cat(
[common_inputs['''attention_mask'''], torch.ones(lowercase_ , lowercase_ , dtype=lowercase_ )] , dim=1 )
snake_case_ = [
(torch.zeros(lowercase_ ), torch.zeros(lowercase_ )) for _ in range(lowercase_ )
]
return common_inputs
def A_ ( self : List[str] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
# 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
snake_case_ = 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
snake_case_ = tokenizer.num_special_tokens_to_add(lowercase_ )
snake_case_ = 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
snake_case_ = [''' '''.join([tokenizer.unk_token] ) * seq_length] * batch_size
snake_case_ = dict(tokenizer(lowercase_ , return_tensors=lowercase_ ) )
return common_inputs
def A_ ( self : Any , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
else:
snake_case_ = self._generate_dummy_inputs_for_causal_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
return common_inputs
def A_ ( self : Dict , lowercase_ : List[str] , lowercase_ : List[str] , lowercase_ : int , lowercase_ : List[str] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super()._flatten_past_key_values_(lowercase_ , lowercase_ , lowercase_ , lowercase_ )
else:
snake_case_ = super(lowercase_ , self )._flatten_past_key_values_(
lowercase_ , lowercase_ , lowercase_ , lowercase_ )
@property
def A_ ( self : List[str] ):
return 1e-4
| 56
| 1
|
'''simple docstring'''
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
a : Tuple = logging.get_logger(__name__)
a : List[str] = {
'facebook/data2vec-vision-base-ft': (
'https://huggingface.co/facebook/data2vec-vision-base-ft/resolve/main/config.json'
),
}
class a ( _lowerCamelCase ):
snake_case_ = "data2vec-vision"
def __init__( self : Optional[Any] , lowercase_ : str=768 , lowercase_ : Union[str, Any]=12 , lowercase_ : int=12 , lowercase_ : str=3072 , lowercase_ : Dict="gelu" , lowercase_ : Any=0.0 , lowercase_ : Optional[int]=0.0 , lowercase_ : int=0.02 , lowercase_ : Union[str, Any]=1e-12 , lowercase_ : Optional[Any]=224 , lowercase_ : List[Any]=16 , lowercase_ : List[Any]=3 , lowercase_ : Tuple=False , lowercase_ : Dict=False , lowercase_ : Union[str, Any]=False , lowercase_ : List[Any]=False , lowercase_ : Dict=0.1 , lowercase_ : Optional[int]=0.1 , lowercase_ : Optional[int]=True , lowercase_ : Optional[int]=[3, 5, 7, 11] , lowercase_ : Tuple=[1, 2, 3, 6] , lowercase_ : List[Any]=True , lowercase_ : Tuple=0.4 , lowercase_ : Tuple=256 , lowercase_ : Optional[int]=1 , lowercase_ : Optional[Any]=False , lowercase_ : Union[str, Any]=255 , **lowercase_ : str , ):
super().__init__(**lowercase_ )
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = image_size
snake_case_ = patch_size
snake_case_ = num_channels
snake_case_ = use_mask_token
snake_case_ = use_absolute_position_embeddings
snake_case_ = use_relative_position_bias
snake_case_ = use_shared_relative_position_bias
snake_case_ = layer_scale_init_value
snake_case_ = drop_path_rate
snake_case_ = use_mean_pooling
# decode head attributes (semantic segmentation)
snake_case_ = out_indices
snake_case_ = pool_scales
# auxiliary head attributes (semantic segmentation)
snake_case_ = use_auxiliary_head
snake_case_ = auxiliary_loss_weight
snake_case_ = auxiliary_channels
snake_case_ = auxiliary_num_convs
snake_case_ = auxiliary_concat_input
snake_case_ = semantic_loss_ignore_index
class a ( _lowerCamelCase ):
snake_case_ = version.parse("1.11" )
@property
def A_ ( self : Any ):
return OrderedDict(
[
('''pixel_values''', {0: '''batch''', 1: '''num_channels''', 2: '''height''', 3: '''width'''}),
] )
@property
def A_ ( self : str ):
return 1e-4
| 56
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import AutoencoderKL, CycleDiffusionPipeline, DDIMScheduler, UNetaDConditionModel
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, skip_mps
from ..pipeline_params import (
IMAGE_TO_IMAGE_IMAGE_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = CycleDiffusionPipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {
"negative_prompt",
"height",
"width",
"negative_prompt_embeds",
}
snake_case_ = PipelineTesterMixin.required_optional_params - {"latents"}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"source_prompt"} )
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
def A_ ( self : Tuple ):
torch.manual_seed(0 )
snake_case_ = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=32 , )
snake_case_ = DDIMScheduler(
beta_start=0.0_0085 , beta_end=0.012 , beta_schedule='''scaled_linear''' , num_train_timesteps=1000 , clip_sample=lowercase_ , set_alpha_to_one=lowercase_ , )
torch.manual_seed(0 )
snake_case_ = 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 )
snake_case_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , )
snake_case_ = CLIPTextModel(lowercase_ )
snake_case_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
snake_case_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
'''safety_checker''': None,
'''feature_extractor''': None,
}
return components
def A_ ( self : Any , lowercase_ : int , lowercase_ : Optional[Any]=0 ):
snake_case_ = floats_tensor((1, 3, 32, 32) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
snake_case_ = image / 2 + 0.5
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''prompt''': '''An astronaut riding an elephant''',
'''source_prompt''': '''An astronaut riding a horse''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''eta''': 0.1,
'''strength''': 0.8,
'''guidance_scale''': 3,
'''source_guidance_scale''': 1,
'''output_type''': '''numpy''',
}
return inputs
def A_ ( self : Union[str, Any] ):
snake_case_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
snake_case_ = self.get_dummy_components()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.4459, 0.4943, 0.4544, 0.6643, 0.5474, 0.4327, 0.5701, 0.5959, 0.5179] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.get_dummy_components()
for name, module in components.items():
if hasattr(lowercase_ , '''half''' ):
snake_case_ = module.half()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.3506, 0.4543, 0.446, 0.4575, 0.5195, 0.4155, 0.5273, 0.518, 0.4116] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@skip_mps
def A_ ( self : Optional[int] ):
return super().test_save_load_local()
@unittest.skip('''non-deterministic pipeline''' )
def A_ ( self : List[Any] ):
return super().test_inference_batch_single_identical()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_dict_tuple_outputs_equivalent()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_save_load_optional_components()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_attention_slicing_forward_pass()
@slow
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : List[Any] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Union[str, Any] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car_fp16.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(
lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , torch_dtype=torch.floataa , revision='''fp16''' )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
# the values aren't exactly equal, but the images look the same visually
assert np.abs(image - expected_image ).max() < 5e-1
def A_ ( self : List[str] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
assert np.abs(image - expected_image ).max() < 2e-2
| 56
| 1
|
'''simple docstring'''
import pickle
import unittest
import torch
from accelerate import Accelerator
from accelerate.state import AcceleratorState
from accelerate.test_utils import require_cpu
@require_cpu
class a ( unittest.TestCase ):
def A_ ( self : Any ):
snake_case_ = torch.nn.Linear(10 , 10 )
snake_case_ = torch.optim.SGD(model.parameters() , 0.1 )
snake_case_ = Accelerator()
snake_case_ = accelerator.prepare(lowercase_ )
try:
pickle.loads(pickle.dumps(lowercase_ ) )
except Exception as e:
self.fail(F"Accelerated optimizer pickling failed with {e}" )
AcceleratorState._reset_state()
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : str = logging.get_logger(__name__)
a : str = {
'google/bigbird-roberta-base': 'https://huggingface.co/google/bigbird-roberta-base/resolve/main/config.json',
'google/bigbird-roberta-large': 'https://huggingface.co/google/bigbird-roberta-large/resolve/main/config.json',
'google/bigbird-base-trivia-itc': 'https://huggingface.co/google/bigbird-base-trivia-itc/resolve/main/config.json',
# See all BigBird models at https://huggingface.co/models?filter=big_bird
}
class a ( _lowerCamelCase ):
snake_case_ = "big_bird"
def __init__( self : Union[str, Any] , lowercase_ : List[Any]=5_0358 , lowercase_ : Tuple=768 , lowercase_ : Dict=12 , lowercase_ : str=12 , lowercase_ : Tuple=3072 , lowercase_ : Any="gelu_new" , lowercase_ : Optional[Any]=0.1 , lowercase_ : List[Any]=0.1 , lowercase_ : List[Any]=4096 , lowercase_ : List[Any]=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[int]=1e-12 , lowercase_ : Tuple=True , lowercase_ : Tuple=0 , lowercase_ : str=1 , lowercase_ : Union[str, Any]=2 , lowercase_ : Optional[Any]=66 , lowercase_ : Optional[int]="block_sparse" , lowercase_ : Any=True , lowercase_ : List[str]=False , lowercase_ : Any=64 , lowercase_ : Tuple=3 , lowercase_ : Tuple=None , **lowercase_ : Tuple , ):
super().__init__(
pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , sep_token_id=lowercase_ , **lowercase_ , )
snake_case_ = vocab_size
snake_case_ = max_position_embeddings
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = type_vocab_size
snake_case_ = layer_norm_eps
snake_case_ = use_cache
snake_case_ = rescale_embeddings
snake_case_ = attention_type
snake_case_ = use_bias
snake_case_ = block_size
snake_case_ = num_random_blocks
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : str ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
import math
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
if depth < 0:
raise ValueError('''Depth cannot be less than 0''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Scores cannot be empty''' )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
return min(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ = [90, 23, 6, 33, 21, 65, 123, 3_4423]
snake_case_ = math.log(len(__UpperCAmelCase ), 2 )
print('''Optimal value : ''', end='''''' )
print(minimax(0, 0, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 56
|
'''simple docstring'''
import contextlib
import os
import sqlitea
import pytest
from datasets import Dataset, Features, Value
from datasets.io.sql import SqlDatasetReader, SqlDatasetWriter
from ..utils import assert_arrow_memory_doesnt_increase, assert_arrow_memory_increases, require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
assert isinstance(__UpperCAmelCase, __UpperCAmelCase )
assert dataset.num_rows == 4
assert dataset.num_columns == 3
assert dataset.column_names == ["col_1", "col_2", "col_3"]
for feature, expected_dtype in expected_features.items():
assert dataset.features[feature].dtype == expected_dtype
@require_sqlalchemy
@pytest.mark.parametrize('''keep_in_memory''', [False, True] )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase():
snake_case_ = SqlDatasetReader(
'''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase, keep_in_memory=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
@require_sqlalchemy
@pytest.mark.parametrize(
'''features''', [
None,
{'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''},
{'''col_1''': '''string''', '''col_2''': '''string''', '''col_3''': '''string'''},
{'''col_1''': '''int32''', '''col_2''': '''int32''', '''col_3''': '''int32'''},
{'''col_1''': '''float32''', '''col_2''': '''float32''', '''col_3''': '''float32'''},
], )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
snake_case_ = features.copy() if features else default_expected_features
snake_case_ = (
Features({feature: Value(__UpperCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None
)
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, features=__UpperCAmelCase, cache_dir=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
with contextlib.closing(sqlitea.connect(__UpperCAmelCase ) ) as con:
snake_case_ = con.cursor()
cur.execute('''SELECT * FROM dataset''' )
for row in cur:
yield row
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=1 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=2 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
with pytest.raises(__UpperCAmelCase ):
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=0 ).write()
| 56
| 1
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from diffusers import (
DDIMScheduler,
KandinskyVaaControlnetPipeline,
KandinskyVaaPriorPipeline,
UNetaDConditionModel,
VQModel,
)
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..test_pipelines_common import PipelineTesterMixin, assert_mean_pixel_difference
enable_full_determinism()
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = KandinskyVaaControlnetPipeline
snake_case_ = ["image_embeds", "negative_image_embeds", "hint"]
snake_case_ = ["image_embeds", "negative_image_embeds", "hint"]
snake_case_ = [
"generator",
"height",
"width",
"latents",
"guidance_scale",
"num_inference_steps",
"return_dict",
"guidance_scale",
"num_images_per_prompt",
"output_type",
"return_dict",
]
snake_case_ = False
@property
def A_ ( self : Union[str, Any] ):
return 32
@property
def A_ ( self : List[Any] ):
return 32
@property
def A_ ( self : List[Any] ):
return self.time_input_dim
@property
def A_ ( self : Any ):
return self.time_input_dim * 4
@property
def A_ ( self : Optional[int] ):
return 100
@property
def A_ ( self : Any ):
torch.manual_seed(0 )
snake_case_ = {
'''in_channels''': 8,
# Out channels is double in channels because predicts mean and variance
'''out_channels''': 8,
'''addition_embed_type''': '''image_hint''',
'''down_block_types''': ('''ResnetDownsampleBlock2D''', '''SimpleCrossAttnDownBlock2D'''),
'''up_block_types''': ('''SimpleCrossAttnUpBlock2D''', '''ResnetUpsampleBlock2D'''),
'''mid_block_type''': '''UNetMidBlock2DSimpleCrossAttn''',
'''block_out_channels''': (self.block_out_channels_a, self.block_out_channels_a * 2),
'''layers_per_block''': 1,
'''encoder_hid_dim''': self.text_embedder_hidden_size,
'''encoder_hid_dim_type''': '''image_proj''',
'''cross_attention_dim''': self.cross_attention_dim,
'''attention_head_dim''': 4,
'''resnet_time_scale_shift''': '''scale_shift''',
'''class_embed_type''': None,
}
snake_case_ = UNetaDConditionModel(**lowercase_ )
return model
@property
def A_ ( self : Any ):
return {
"block_out_channels": [32, 32, 64, 64],
"down_block_types": [
"DownEncoderBlock2D",
"DownEncoderBlock2D",
"DownEncoderBlock2D",
"AttnDownEncoderBlock2D",
],
"in_channels": 3,
"latent_channels": 4,
"layers_per_block": 1,
"norm_num_groups": 8,
"norm_type": "spatial",
"num_vq_embeddings": 12,
"out_channels": 3,
"up_block_types": ["AttnUpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D", "UpDecoderBlock2D"],
"vq_embed_dim": 4,
}
@property
def A_ ( self : int ):
torch.manual_seed(0 )
snake_case_ = VQModel(**self.dummy_movq_kwargs )
return model
def A_ ( self : Tuple ):
snake_case_ = self.dummy_unet
snake_case_ = self.dummy_movq
snake_case_ = DDIMScheduler(
num_train_timesteps=1000 , beta_schedule='''linear''' , beta_start=0.0_0085 , beta_end=0.012 , clip_sample=lowercase_ , set_alpha_to_one=lowercase_ , steps_offset=1 , prediction_type='''epsilon''' , thresholding=lowercase_ , )
snake_case_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''movq''': movq,
}
return components
def A_ ( self : Dict , lowercase_ : Dict , lowercase_ : Any=0 ):
snake_case_ = floats_tensor((1, self.text_embedder_hidden_size) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
snake_case_ = floats_tensor((1, self.text_embedder_hidden_size) , rng=random.Random(seed + 1 ) ).to(
lowercase_ )
# create hint
snake_case_ = floats_tensor((1, 3, 64, 64) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''image_embeds''': image_embeds,
'''negative_image_embeds''': negative_image_embeds,
'''hint''': hint,
'''generator''': generator,
'''height''': 64,
'''width''': 64,
'''guidance_scale''': 4.0,
'''num_inference_steps''': 2,
'''output_type''': '''np''',
}
return inputs
def A_ ( self : Optional[Any] ):
snake_case_ = '''cpu'''
snake_case_ = self.get_dummy_components()
snake_case_ = self.pipeline_class(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = pipe(**self.get_dummy_inputs(lowercase_ ) )
snake_case_ = output.images
snake_case_ = pipe(
**self.get_dummy_inputs(lowercase_ ) , return_dict=lowercase_ , )[0]
snake_case_ = image[0, -3:, -3:, -1]
snake_case_ = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 64, 64, 3)
snake_case_ = np.array(
[0.695_9826, 0.86_8279, 0.755_8092, 0.6876_9467, 0.8580_5804, 0.6597_7496, 0.4488_5302, 0.595_9111, 0.425_1595] )
assert (
np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
), F" expected_slice {expected_slice}, but got {image_slice.flatten()}"
assert (
np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1e-2
), F" expected_slice {expected_slice}, but got {image_from_tuple_slice.flatten()}"
@slow
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : str ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Tuple ):
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/kandinskyv22/kandinskyv22_controlnet_robotcat_fp16.npy''' )
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/kandinskyv22/hint_image_cat.png''' )
snake_case_ = torch.from_numpy(np.array(lowercase_ ) ).float() / 255.0
snake_case_ = hint.permute(2 , 0 , 1 ).unsqueeze(0 )
snake_case_ = KandinskyVaaPriorPipeline.from_pretrained(
'''kandinsky-community/kandinsky-2-2-prior''' , torch_dtype=torch.floataa )
pipe_prior.to(lowercase_ )
snake_case_ = KandinskyVaaControlnetPipeline.from_pretrained(
'''kandinsky-community/kandinsky-2-2-controlnet-depth''' , torch_dtype=torch.floataa )
snake_case_ = pipeline.to(lowercase_ )
pipeline.set_progress_bar_config(disable=lowercase_ )
snake_case_ = '''A robot, 4k photo'''
snake_case_ = torch.Generator(device='''cuda''' ).manual_seed(0 )
snake_case_ ,snake_case_ = pipe_prior(
lowercase_ , generator=lowercase_ , num_inference_steps=5 , negative_prompt='''''' , ).to_tuple()
snake_case_ = torch.Generator(device='''cuda''' ).manual_seed(0 )
snake_case_ = pipeline(
image_embeds=lowercase_ , negative_image_embeds=lowercase_ , hint=lowercase_ , generator=lowercase_ , num_inference_steps=100 , output_type='''np''' , )
snake_case_ = output.images[0]
assert image.shape == (512, 512, 3)
assert_mean_pixel_difference(lowercase_ , lowercase_ )
| 56
|
'''simple docstring'''
from collections import defaultdict
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 1
snake_case_ = True
for v in tree[start]:
if v not in visited:
ret += dfs(__UpperCAmelCase )
if ret % 2 == 0:
cuts.append(__UpperCAmelCase )
return ret
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
dfs(1 )
if __name__ == "__main__":
a ,a : Dict = 10, 9
a : Dict = defaultdict(list)
a : dict[int, bool] = {}
a : list[int] = []
a : Tuple = 0
a : str = [(2, 1), (3, 1), (4, 3), (5, 2), (6, 1), (7, 2), (8, 6), (9, 8), (10, 8)]
for u, v in edges:
tree[u].append(v)
tree[v].append(u)
even_tree()
print(len(cuts) - 1)
| 56
| 1
|
'''simple docstring'''
import timeit
import numpy as np
import datasets
from datasets.arrow_writer import ArrowWriter
from datasets.features.features import _ArrayXD
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
def wrapper(*__UpperCAmelCase, **__UpperCAmelCase ):
snake_case_ = timeit.default_timer()
snake_case_ = func(*__UpperCAmelCase, **__UpperCAmelCase )
snake_case_ = timeit.default_timer() - starttime
return delta
snake_case_ = func.__name__
return wrapper
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=100, __UpperCAmelCase=None ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = []
snake_case_ = seq_shapes or {}
for i in range(__UpperCAmelCase ):
snake_case_ = {}
for col_id, (k, v) in enumerate(features.items() ):
if isinstance(__UpperCAmelCase, _ArrayXD ):
snake_case_ = np.random.rand(*v.shape ).astype(v.dtype )
elif isinstance(__UpperCAmelCase, datasets.Value ):
if v.dtype == "string":
snake_case_ = '''The small grey turtle was surprisingly fast when challenged.'''
else:
snake_case_ = np.random.randint(10, size=1 ).astype(v.dtype ).item()
elif isinstance(__UpperCAmelCase, datasets.Sequence ):
while isinstance(__UpperCAmelCase, datasets.Sequence ):
snake_case_ = v.feature
snake_case_ = seq_shapes[k]
snake_case_ = np.random.rand(*__UpperCAmelCase ).astype(v.dtype )
snake_case_ = data
dummy_data.append((i, example) )
return dummy_data
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase=100, __UpperCAmelCase=None ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = generate_examples(__UpperCAmelCase, num_examples=__UpperCAmelCase, seq_shapes=__UpperCAmelCase )
with ArrowWriter(features=__UpperCAmelCase, path=__UpperCAmelCase ) as writer:
for key, record in dummy_data:
snake_case_ = features.encode_example(__UpperCAmelCase )
writer.write(__UpperCAmelCase )
snake_case_ ,snake_case_ = 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}." )
snake_case_ = datasets.Dataset.from_file(filename=__UpperCAmelCase, info=datasets.DatasetInfo(features=__UpperCAmelCase ) )
return dataset
| 56
|
'''simple docstring'''
import math
from collections.abc import Callable
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
snake_case_ = xa
snake_case_ = xa
while True:
if x_n == x_na or function(__UpperCAmelCase ) == function(__UpperCAmelCase ):
raise ZeroDivisionError('''float division by zero, could not find root''' )
snake_case_ = x_na - (
function(__UpperCAmelCase ) / ((function(__UpperCAmelCase ) - function(__UpperCAmelCase )) / (x_na - x_n))
)
if abs(x_na - x_na ) < 10**-5:
return x_na
snake_case_ = x_na
snake_case_ = x_na
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
return math.pow(__UpperCAmelCase, 3 ) - (2 * x) - 5
if __name__ == "__main__":
print(intersection(f, 3, 3.5))
| 56
| 1
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
a : Optional[Any] = logging.get_logger(__name__)
a : str = {
'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 ( _lowerCamelCase ):
snake_case_ = "funnel"
snake_case_ = {
"hidden_size": "d_model",
"num_attention_heads": "n_head",
}
def __init__( self : List[Any] , lowercase_ : Optional[int]=3_0522 , lowercase_ : Optional[int]=[4, 4, 4] , lowercase_ : str=None , lowercase_ : Union[str, Any]=2 , lowercase_ : Union[str, Any]=768 , lowercase_ : List[str]=12 , lowercase_ : List[str]=64 , lowercase_ : Optional[int]=3072 , lowercase_ : Optional[int]="gelu_new" , lowercase_ : List[Any]=0.1 , lowercase_ : Dict=0.1 , lowercase_ : str=0.0 , lowercase_ : Union[str, Any]=0.1 , lowercase_ : List[Any]=None , lowercase_ : int=1e-9 , lowercase_ : Union[str, Any]="mean" , lowercase_ : Dict="relative_shift" , lowercase_ : Optional[int]=True , lowercase_ : Optional[int]=True , lowercase_ : Optional[Any]=True , **lowercase_ : Dict , ):
snake_case_ = vocab_size
snake_case_ = block_sizes
snake_case_ = [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."
snake_case_ = num_decoder_layers
snake_case_ = d_model
snake_case_ = n_head
snake_case_ = d_head
snake_case_ = d_inner
snake_case_ = hidden_act
snake_case_ = hidden_dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = initializer_range
snake_case_ = initializer_std
snake_case_ = layer_norm_eps
assert pooling_type in [
"mean",
"max",
], F"Got {pooling_type} for `pooling_type` but only 'mean' and 'max' are supported."
snake_case_ = pooling_type
assert attention_type in [
"relative_shift",
"factorized",
], F"Got {attention_type} for `attention_type` but only 'relative_shift' and 'factorized' are supported."
snake_case_ = attention_type
snake_case_ = separate_cls
snake_case_ = truncate_seq
snake_case_ = pool_q_only
super().__init__(**lowercase_ )
@property
def A_ ( self : Optional[int] ):
return sum(self.block_sizes )
@num_hidden_layers.setter
def A_ ( self : Union[str, Any] , lowercase_ : Dict ):
raise NotImplementedError(
'''This model does not support the setting of `num_hidden_layers`. Please set `block_sizes`.''' )
@property
def A_ ( self : Optional[Any] ):
return len(self.block_sizes )
@num_blocks.setter
def A_ ( self : Tuple , lowercase_ : Union[str, Any] ):
raise NotImplementedError('''This model does not support the setting of `num_blocks`. Please set `block_sizes`.''' )
| 56
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_url
from PIL import Image
from transformers import DPTConfig, DPTForDepthEstimation, DPTForSemanticSegmentation, DPTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
a : Any = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = DPTConfig()
if "large" in checkpoint_url:
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 24
snake_case_ = 16
snake_case_ = [5, 11, 17, 23]
snake_case_ = [256, 512, 1024, 1024]
snake_case_ = (1, 384, 384)
if "ade" in checkpoint_url:
snake_case_ = True
snake_case_ = 150
snake_case_ = '''huggingface/label-files'''
snake_case_ = '''ade20k-id2label.json'''
snake_case_ = json.load(open(cached_download(hf_hub_url(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ) ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = idalabel
snake_case_ = {v: k for k, v in idalabel.items()}
snake_case_ = [1, 150, 480, 480]
return config, expected_shape
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = ['''pretrained.model.head.weight''', '''pretrained.model.head.bias''']
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if (
"pretrained.model" in name
and "cls_token" not in name
and "pos_embed" not in name
and "patch_embed" not in name
):
snake_case_ = name.replace('''pretrained.model''', '''dpt.encoder''' )
if "pretrained.model" in name:
snake_case_ = name.replace('''pretrained.model''', '''dpt.embeddings''' )
if "patch_embed" in name:
snake_case_ = name.replace('''patch_embed''', '''patch_embeddings''' )
if "pos_embed" in name:
snake_case_ = name.replace('''pos_embed''', '''position_embeddings''' )
if "attn.proj" in name:
snake_case_ = name.replace('''attn.proj''', '''attention.output.dense''' )
if "proj" in name and "project" not in name:
snake_case_ = name.replace('''proj''', '''projection''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''layer''' )
if "mlp.fc1" in name:
snake_case_ = name.replace('''mlp.fc1''', '''intermediate.dense''' )
if "mlp.fc2" in name:
snake_case_ = name.replace('''mlp.fc2''', '''output.dense''' )
if "norm1" in name:
snake_case_ = name.replace('''norm1''', '''layernorm_before''' )
if "norm2" in name:
snake_case_ = name.replace('''norm2''', '''layernorm_after''' )
if "scratch.output_conv" in name:
snake_case_ = name.replace('''scratch.output_conv''', '''head''' )
if "scratch" in name:
snake_case_ = name.replace('''scratch''', '''neck''' )
if "layer1_rn" in name:
snake_case_ = name.replace('''layer1_rn''', '''convs.0''' )
if "layer2_rn" in name:
snake_case_ = name.replace('''layer2_rn''', '''convs.1''' )
if "layer3_rn" in name:
snake_case_ = name.replace('''layer3_rn''', '''convs.2''' )
if "layer4_rn" in name:
snake_case_ = name.replace('''layer4_rn''', '''convs.3''' )
if "refinenet" in name:
snake_case_ = int(name[len('''neck.refinenet''' ) : len('''neck.refinenet''' ) + 1] )
# tricky here: we need to map 4 to 0, 3 to 1, 2 to 2 and 1 to 3
snake_case_ = name.replace(F"refinenet{layer_idx}", F"fusion_stage.layers.{abs(layer_idx-4 )}" )
if "out_conv" in name:
snake_case_ = name.replace('''out_conv''', '''projection''' )
if "resConfUnit1" in name:
snake_case_ = name.replace('''resConfUnit1''', '''residual_layer1''' )
if "resConfUnit2" in name:
snake_case_ = name.replace('''resConfUnit2''', '''residual_layer2''' )
if "conv1" in name:
snake_case_ = name.replace('''conv1''', '''convolution1''' )
if "conv2" in name:
snake_case_ = name.replace('''conv2''', '''convolution2''' )
# readout blocks
if "pretrained.act_postprocess1.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.0.project.0''', '''neck.reassemble_stage.readout_projects.0.0''' )
if "pretrained.act_postprocess2.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.0.project.0''', '''neck.reassemble_stage.readout_projects.1.0''' )
if "pretrained.act_postprocess3.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.0.project.0''', '''neck.reassemble_stage.readout_projects.2.0''' )
if "pretrained.act_postprocess4.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.0.project.0''', '''neck.reassemble_stage.readout_projects.3.0''' )
# resize blocks
if "pretrained.act_postprocess1.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.3''', '''neck.reassemble_stage.layers.0.projection''' )
if "pretrained.act_postprocess1.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.4''', '''neck.reassemble_stage.layers.0.resize''' )
if "pretrained.act_postprocess2.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.3''', '''neck.reassemble_stage.layers.1.projection''' )
if "pretrained.act_postprocess2.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.4''', '''neck.reassemble_stage.layers.1.resize''' )
if "pretrained.act_postprocess3.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.3''', '''neck.reassemble_stage.layers.2.projection''' )
if "pretrained.act_postprocess4.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.3''', '''neck.reassemble_stage.layers.3.projection''' )
if "pretrained.act_postprocess4.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.4''', '''neck.reassemble_stage.layers.3.resize''' )
if "pretrained" in name:
snake_case_ = name.replace('''pretrained''', '''dpt''' )
if "bn" in name:
snake_case_ = name.replace('''bn''', '''batch_norm''' )
if "head" in name:
snake_case_ = name.replace('''head''', '''head.head''' )
if "encoder.norm" in name:
snake_case_ = name.replace('''encoder.norm''', '''layernorm''' )
if "auxlayer" in name:
snake_case_ = name.replace('''auxlayer''', '''auxiliary_head.head''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.weight" )
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.bias" )
# next, add query, keys and values (in that order) to the state dict
snake_case_ = in_proj_weight[: config.hidden_size, :]
snake_case_ = in_proj_bias[: config.hidden_size]
snake_case_ = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
snake_case_ = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
snake_case_ = in_proj_weight[
-config.hidden_size :, :
]
snake_case_ = in_proj_bias[-config.hidden_size :]
def __magic_name__ ( ) -> Any:
'''simple docstring'''
snake_case_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
snake_case_ = Image.open(requests.get(__UpperCAmelCase, stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dpt_config(__UpperCAmelCase )
# load original state_dict from URL
snake_case_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase, map_location='''cpu''' )
# remove certain keys
remove_ignore_keys_(__UpperCAmelCase )
# rename keys
for key in state_dict.copy().keys():
snake_case_ = state_dict.pop(__UpperCAmelCase )
snake_case_ = val
# read in qkv matrices
read_in_q_k_v(__UpperCAmelCase, __UpperCAmelCase )
# load HuggingFace model
snake_case_ = DPTForSemanticSegmentation(__UpperCAmelCase ) if '''ade''' in checkpoint_url else DPTForDepthEstimation(__UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
# Check outputs on an image
snake_case_ = 480 if '''ade''' in checkpoint_url else 384
snake_case_ = DPTImageProcessor(size=__UpperCAmelCase )
snake_case_ = prepare_img()
snake_case_ = image_processor(__UpperCAmelCase, return_tensors='''pt''' )
# forward pass
snake_case_ = model(**__UpperCAmelCase ).logits if '''ade''' in checkpoint_url else model(**__UpperCAmelCase ).predicted_depth
# Assert logits
snake_case_ = torch.tensor([[6.3_1_9_9, 6.3_6_2_9, 6.4_1_4_8], [6.3_8_5_0, 6.3_6_1_5, 6.4_1_6_6], [6.3_5_1_9, 6.3_1_7_6, 6.3_5_7_5]] )
if "ade" in checkpoint_url:
snake_case_ = torch.tensor([[4.0_4_8_0, 4.2_4_2_0, 4.4_3_6_0], [4.3_1_2_4, 4.5_6_9_3, 4.8_2_6_1], [4.5_7_6_8, 4.8_9_6_5, 5.2_1_6_3]] )
assert outputs.shape == torch.Size(__UpperCAmelCase )
assert (
torch.allclose(outputs[0, 0, :3, :3], __UpperCAmelCase, atol=1e-4 )
if "ade" in checkpoint_url
else torch.allclose(outputs[0, :3, :3], __UpperCAmelCase )
)
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
print(F"Saving model to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
print(F"Saving image processor to {pytorch_dump_folder_path}" )
image_processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print('''Pushing model to hub...''' )
model.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add model''', use_temp_dir=__UpperCAmelCase, )
image_processor.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add image processor''', use_temp_dir=__UpperCAmelCase, )
if __name__ == "__main__":
a : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--checkpoint_url',
default='https://github.com/intel-isl/DPT/releases/download/1_0/dpt_large-midas-2f21e586.pt',
type=str,
help='URL of the original DPT checkpoint you\'d like to convert.',
)
parser.add_argument(
'--pytorch_dump_folder_path',
default=None,
type=str,
required=True,
help='Path to the output PyTorch model directory.',
)
parser.add_argument(
'--push_to_hub',
action='store_true',
)
parser.add_argument(
'--model_name',
default='dpt-large',
type=str,
help='Name of the model, in case you\'re pushing to the hub.',
)
a : List[Any] = parser.parse_args()
convert_dpt_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 56
| 1
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
a : List[str] = logging.get_logger(__name__)
a : Any = {
'MIT/ast-finetuned-audioset-10-10-0.4593': (
'https://huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593/resolve/main/config.json'
),
}
class a ( _lowerCamelCase ):
snake_case_ = "audio-spectrogram-transformer"
def __init__( self : Optional[Any] , lowercase_ : Union[str, Any]=768 , lowercase_ : str=12 , lowercase_ : Any=12 , lowercase_ : Optional[Any]=3072 , lowercase_ : Optional[Any]="gelu" , lowercase_ : str=0.0 , lowercase_ : Union[str, Any]=0.0 , lowercase_ : Tuple=0.02 , lowercase_ : str=1e-12 , lowercase_ : str=16 , lowercase_ : Optional[Any]=True , lowercase_ : List[Any]=10 , lowercase_ : int=10 , lowercase_ : Any=1024 , lowercase_ : Union[str, Any]=128 , **lowercase_ : Tuple , ):
super().__init__(**lowercase_ )
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = patch_size
snake_case_ = qkv_bias
snake_case_ = frequency_stride
snake_case_ = time_stride
snake_case_ = max_length
snake_case_ = num_mel_bins
| 56
|
'''simple docstring'''
import re
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
snake_case_ = re.compile(
r'''^(?:0|94|\+94|0{2}94)''' r'''7(0|1|2|4|5|6|7|8)''' r'''(-| |)''' r'''\d{7}$''' )
return bool(re.search(__UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
a : Any = '0094702343221'
print(is_sri_lankan_phone_number(phone))
| 56
| 1
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
a : Tuple = {
'configuration_mobilebert': [
'MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP',
'MobileBertConfig',
'MobileBertOnnxConfig',
],
'tokenization_mobilebert': ['MobileBertTokenizer'],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Tuple = ['MobileBertTokenizerFast']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Union[str, Any] = [
'MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST',
'MobileBertForMaskedLM',
'MobileBertForMultipleChoice',
'MobileBertForNextSentencePrediction',
'MobileBertForPreTraining',
'MobileBertForQuestionAnswering',
'MobileBertForSequenceClassification',
'MobileBertForTokenClassification',
'MobileBertLayer',
'MobileBertModel',
'MobileBertPreTrainedModel',
'load_tf_weights_in_mobilebert',
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : List[Any] = [
'TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST',
'TFMobileBertForMaskedLM',
'TFMobileBertForMultipleChoice',
'TFMobileBertForNextSentencePrediction',
'TFMobileBertForPreTraining',
'TFMobileBertForQuestionAnswering',
'TFMobileBertForSequenceClassification',
'TFMobileBertForTokenClassification',
'TFMobileBertMainLayer',
'TFMobileBertModel',
'TFMobileBertPreTrainedModel',
]
if TYPE_CHECKING:
from .configuration_mobilebert import (
MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
MobileBertConfig,
MobileBertOnnxConfig,
)
from .tokenization_mobilebert import MobileBertTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_mobilebert_fast import MobileBertTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mobilebert import (
MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
MobileBertForMaskedLM,
MobileBertForMultipleChoice,
MobileBertForNextSentencePrediction,
MobileBertForPreTraining,
MobileBertForQuestionAnswering,
MobileBertForSequenceClassification,
MobileBertForTokenClassification,
MobileBertLayer,
MobileBertModel,
MobileBertPreTrainedModel,
load_tf_weights_in_mobilebert,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mobilebert import (
TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFMobileBertForMaskedLM,
TFMobileBertForMultipleChoice,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertMainLayer,
TFMobileBertModel,
TFMobileBertPreTrainedModel,
)
else:
import sys
a : List[Any] = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
|
'''simple docstring'''
import re
from filelock import FileLock
try:
import nltk
a : Union[str, Any] = True
except (ImportError, ModuleNotFoundError):
a : Any = False
if NLTK_AVAILABLE:
with FileLock('.lock') as lock:
nltk.download('punkt', quiet=True)
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
re.sub('''<n>''', '''''', __UpperCAmelCase ) # remove pegasus newline char
assert NLTK_AVAILABLE, "nltk must be installed to separate newlines between sentences. (pip install nltk)"
return "\n".join(nltk.sent_tokenize(__UpperCAmelCase ) )
| 56
| 1
|
'''simple docstring'''
import math
import os
import sys
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = ''''''
try:
with open(__UpperCAmelCase, '''rb''' ) as binary_file:
snake_case_ = binary_file.read()
for dat in data:
snake_case_ = F"{dat:08b}"
result += curr_byte
return result
except OSError:
print('''File not accessible''' )
sys.exit()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> None:
'''simple docstring'''
lexicon.pop(__UpperCAmelCase )
snake_case_ = last_match_id
if math.loga(__UpperCAmelCase ).is_integer():
for curr_key in lexicon:
snake_case_ = '''0''' + lexicon[curr_key]
snake_case_ = bin(__UpperCAmelCase )[2:]
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = {'''0''': '''0''', '''1''': '''1'''}
snake_case_ ,snake_case_ = '''''', ''''''
snake_case_ = len(__UpperCAmelCase )
for i in range(len(__UpperCAmelCase ) ):
curr_string += data_bits[i]
if curr_string not in lexicon:
continue
snake_case_ = lexicon[curr_string]
result += last_match_id
add_key_to_lexicon(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
index += 1
snake_case_ = ''''''
while curr_string != "" and curr_string not in lexicon:
curr_string += "0"
if curr_string != "":
snake_case_ = lexicon[curr_string]
result += last_match_id
return result
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = os.path.getsize(__UpperCAmelCase )
snake_case_ = bin(__UpperCAmelCase )[2:]
snake_case_ = len(__UpperCAmelCase )
return "0" * (length_length - 1) + file_length_binary + compressed
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> None:
'''simple docstring'''
snake_case_ = 8
try:
with open(__UpperCAmelCase, '''wb''' ) as opened_file:
snake_case_ = [
to_write[i : i + byte_length]
for i in range(0, len(__UpperCAmelCase ), __UpperCAmelCase )
]
if len(result_byte_array[-1] ) % byte_length == 0:
result_byte_array.append('''10000000''' )
else:
result_byte_array[-1] += "1" + "0" * (
byte_length - len(result_byte_array[-1] ) - 1
)
for elem in result_byte_array:
opened_file.write(int(__UpperCAmelCase, 2 ).to_bytes(1, byteorder='''big''' ) )
except OSError:
print('''File not accessible''' )
sys.exit()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> None:
'''simple docstring'''
snake_case_ = read_file_binary(__UpperCAmelCase )
snake_case_ = compress_data(__UpperCAmelCase )
snake_case_ = add_file_length(__UpperCAmelCase, __UpperCAmelCase )
write_file_binary(__UpperCAmelCase, __UpperCAmelCase )
if __name__ == "__main__":
compress(sys.argv[1], sys.argv[2])
| 56
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
a : Tuple = {
'configuration_llama': ['LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP', 'LlamaConfig'],
}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = ['LlamaTokenizer']
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : str = ['LlamaTokenizerFast']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = [
'LlamaForCausalLM',
'LlamaModel',
'LlamaPreTrainedModel',
'LlamaForSequenceClassification',
]
if TYPE_CHECKING:
from .configuration_llama import LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP, LlamaConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama import LlamaTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama_fast import LlamaTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_llama import LlamaForCausalLM, LlamaForSequenceClassification, LlamaModel, LlamaPreTrainedModel
else:
import sys
a : str = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
| 1
|
'''simple docstring'''
import argparse
import gc
import json
import os
import shutil
import warnings
import torch
from transformers import LlamaConfig, LlamaForCausalLM, LlamaTokenizer
try:
from transformers import LlamaTokenizerFast
except ImportError as e:
warnings.warn(e)
warnings.warn(
'The converted tokenizer will be the `slow` tokenizer. To use the fast, update your `tokenizers` library and re-run the tokenizer conversion'
)
a : Dict = None
a : Optional[int] = {
'7B': 1_1008,
'13B': 1_3824,
'30B': 1_7920,
'65B': 2_2016,
'70B': 2_8672,
}
a : Union[str, Any] = {
'7B': 1,
'7Bf': 1,
'13B': 2,
'13Bf': 2,
'30B': 4,
'65B': 8,
'70B': 8,
'70Bf': 8,
}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=1, __UpperCAmelCase=256 ) -> Dict:
'''simple docstring'''
return multiple_of * ((int(ffn_dim_multiplier * int(8 * n / 3 ) ) + multiple_of - 1) // multiple_of)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
with open(__UpperCAmelCase, '''r''' ) as f:
return json.load(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
with open(__UpperCAmelCase, '''w''' ) as f:
json.dump(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase=True ) -> Optional[int]:
'''simple docstring'''
os.makedirs(__UpperCAmelCase, exist_ok=__UpperCAmelCase )
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp''' )
os.makedirs(__UpperCAmelCase, exist_ok=__UpperCAmelCase )
snake_case_ = read_json(os.path.join(__UpperCAmelCase, '''params.json''' ) )
snake_case_ = NUM_SHARDS[model_size]
snake_case_ = params['''n_layers''']
snake_case_ = params['''n_heads''']
snake_case_ = n_heads // num_shards
snake_case_ = params['''dim''']
snake_case_ = dim // n_heads
snake_case_ = 1_0_0_0_0.0
snake_case_ = 1.0 / (base ** (torch.arange(0, __UpperCAmelCase, 2 ).float() / dims_per_head))
if "n_kv_heads" in params:
snake_case_ = params['''n_kv_heads'''] # for GQA / MQA
snake_case_ = n_heads_per_shard // num_key_value_heads
snake_case_ = dim // num_key_value_heads
else: # compatibility with other checkpoints
snake_case_ = n_heads
snake_case_ = n_heads_per_shard
snake_case_ = dim
# permute for sliced rotary
def permute(__UpperCAmelCase, __UpperCAmelCase=n_heads, __UpperCAmelCase=dim, __UpperCAmelCase=dim ):
return w.view(__UpperCAmelCase, dima // n_heads // 2, 2, __UpperCAmelCase ).transpose(1, 2 ).reshape(__UpperCAmelCase, __UpperCAmelCase )
print(F"Fetching all parameters from the checkpoint at {input_base_path}." )
# Load weights
if model_size == "7B":
# Not sharded
# (The sharded implementation would also work, but this is simpler.)
snake_case_ = torch.load(os.path.join(__UpperCAmelCase, '''consolidated.00.pth''' ), map_location='''cpu''' )
else:
# Sharded
snake_case_ = [
torch.load(os.path.join(__UpperCAmelCase, F"consolidated.{i:02d}.pth" ), map_location='''cpu''' )
for i in range(__UpperCAmelCase )
]
snake_case_ = 0
snake_case_ = {'''weight_map''': {}}
for layer_i in range(__UpperCAmelCase ):
snake_case_ = F"pytorch_model-{layer_i + 1}-of-{n_layers + 1}.bin"
if model_size == "7B":
# Unsharded
snake_case_ = {
F"model.layers.{layer_i}.self_attn.q_proj.weight": permute(
loaded[F"layers.{layer_i}.attention.wq.weight"] ),
F"model.layers.{layer_i}.self_attn.k_proj.weight": permute(
loaded[F"layers.{layer_i}.attention.wk.weight"] ),
F"model.layers.{layer_i}.self_attn.v_proj.weight": loaded[F"layers.{layer_i}.attention.wv.weight"],
F"model.layers.{layer_i}.self_attn.o_proj.weight": loaded[F"layers.{layer_i}.attention.wo.weight"],
F"model.layers.{layer_i}.mlp.gate_proj.weight": loaded[F"layers.{layer_i}.feed_forward.w1.weight"],
F"model.layers.{layer_i}.mlp.down_proj.weight": loaded[F"layers.{layer_i}.feed_forward.w2.weight"],
F"model.layers.{layer_i}.mlp.up_proj.weight": loaded[F"layers.{layer_i}.feed_forward.w3.weight"],
F"model.layers.{layer_i}.input_layernorm.weight": loaded[F"layers.{layer_i}.attention_norm.weight"],
F"model.layers.{layer_i}.post_attention_layernorm.weight": loaded[F"layers.{layer_i}.ffn_norm.weight"],
}
else:
# Sharded
# Note that attention.w{q,k,v,o}, feed_fordward.w[1,2,3], attention_norm.weight and ffn_norm.weight share
# the same storage object, saving attention_norm and ffn_norm will save other weights too, which is
# redundant as other weights will be stitched from multiple shards. To avoid that, they are cloned.
snake_case_ = {
F"model.layers.{layer_i}.input_layernorm.weight": loaded[0][
F"layers.{layer_i}.attention_norm.weight"
].clone(),
F"model.layers.{layer_i}.post_attention_layernorm.weight": loaded[0][
F"layers.{layer_i}.ffn_norm.weight"
].clone(),
}
snake_case_ = permute(
torch.cat(
[
loaded[i][F"layers.{layer_i}.attention.wq.weight"].view(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
for i in range(__UpperCAmelCase )
], dim=0, ).reshape(__UpperCAmelCase, __UpperCAmelCase ) )
snake_case_ = permute(
torch.cat(
[
loaded[i][F"layers.{layer_i}.attention.wk.weight"].view(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
for i in range(__UpperCAmelCase )
], dim=0, ).reshape(__UpperCAmelCase, __UpperCAmelCase ), __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, )
snake_case_ = torch.cat(
[
loaded[i][F"layers.{layer_i}.attention.wv.weight"].view(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
for i in range(__UpperCAmelCase )
], dim=0, ).reshape(__UpperCAmelCase, __UpperCAmelCase )
snake_case_ = torch.cat(
[loaded[i][F"layers.{layer_i}.attention.wo.weight"] for i in range(__UpperCAmelCase )], dim=1 )
snake_case_ = torch.cat(
[loaded[i][F"layers.{layer_i}.feed_forward.w1.weight"] for i in range(__UpperCAmelCase )], dim=0 )
snake_case_ = torch.cat(
[loaded[i][F"layers.{layer_i}.feed_forward.w2.weight"] for i in range(__UpperCAmelCase )], dim=1 )
snake_case_ = torch.cat(
[loaded[i][F"layers.{layer_i}.feed_forward.w3.weight"] for i in range(__UpperCAmelCase )], dim=0 )
snake_case_ = inv_freq
for k, v in state_dict.items():
snake_case_ = filename
param_count += v.numel()
torch.save(__UpperCAmelCase, os.path.join(__UpperCAmelCase, __UpperCAmelCase ) )
snake_case_ = F"pytorch_model-{n_layers + 1}-of-{n_layers + 1}.bin"
if model_size == "7B":
# Unsharded
snake_case_ = {
'''model.embed_tokens.weight''': loaded['''tok_embeddings.weight'''],
'''model.norm.weight''': loaded['''norm.weight'''],
'''lm_head.weight''': loaded['''output.weight'''],
}
else:
snake_case_ = {
'''model.norm.weight''': loaded[0]['''norm.weight'''],
'''model.embed_tokens.weight''': torch.cat(
[loaded[i]['''tok_embeddings.weight'''] for i in range(__UpperCAmelCase )], dim=1 ),
'''lm_head.weight''': torch.cat([loaded[i]['''output.weight'''] for i in range(__UpperCAmelCase )], dim=0 ),
}
for k, v in state_dict.items():
snake_case_ = filename
param_count += v.numel()
torch.save(__UpperCAmelCase, os.path.join(__UpperCAmelCase, __UpperCAmelCase ) )
# Write configs
snake_case_ = {'''total_size''': param_count * 2}
write_json(__UpperCAmelCase, os.path.join(__UpperCAmelCase, '''pytorch_model.bin.index.json''' ) )
snake_case_ = params['''ffn_dim_multiplier'''] if '''ffn_dim_multiplier''' in params else 1
snake_case_ = params['''multiple_of'''] if '''multiple_of''' in params else 256
snake_case_ = LlamaConfig(
hidden_size=__UpperCAmelCase, intermediate_size=compute_intermediate_size(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), num_attention_heads=params['''n_heads'''], num_hidden_layers=params['''n_layers'''], rms_norm_eps=params['''norm_eps'''], num_key_value_heads=__UpperCAmelCase, )
config.save_pretrained(__UpperCAmelCase )
# Make space so we can load the model properly now.
del state_dict
del loaded
gc.collect()
print('''Loading the checkpoint in a Llama model.''' )
snake_case_ = LlamaForCausalLM.from_pretrained(__UpperCAmelCase, torch_dtype=torch.floataa, low_cpu_mem_usage=__UpperCAmelCase )
# Avoid saving this as part of the config.
del model.config._name_or_path
print('''Saving in the Transformers format.''' )
model.save_pretrained(__UpperCAmelCase, safe_serialization=__UpperCAmelCase )
shutil.rmtree(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = LlamaTokenizer if LlamaTokenizerFast is None else LlamaTokenizerFast
print(F"Saving a {tokenizer_class.__name__} to {tokenizer_path}." )
snake_case_ = tokenizer_class(__UpperCAmelCase )
tokenizer.save_pretrained(__UpperCAmelCase )
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = argparse.ArgumentParser()
parser.add_argument(
'''--input_dir''', help='''Location of LLaMA weights, which contains tokenizer.model and model folders''', )
parser.add_argument(
'''--model_size''', choices=['''7B''', '''7Bf''', '''13B''', '''13Bf''', '''30B''', '''65B''', '''70B''', '''70Bf''', '''tokenizer_only'''], )
parser.add_argument(
'''--output_dir''', help='''Location to write HF model and tokenizer''', )
parser.add_argument('''--safe_serialization''', type=__UpperCAmelCase, help='''Whether or not to save using `safetensors`.''' )
snake_case_ = parser.parse_args()
if args.model_size != "tokenizer_only":
write_model(
model_path=args.output_dir, input_base_path=os.path.join(args.input_dir, args.model_size ), model_size=args.model_size, safe_serialization=args.safe_serialization, )
snake_case_ = os.path.join(args.input_dir, '''tokenizer.model''' )
write_tokenizer(args.output_dir, __UpperCAmelCase )
if __name__ == "__main__":
main()
| 56
|
'''simple docstring'''
import re
from typing import Callable, List, Optional, Union
import tensorflow as tf
try:
from tensorflow.keras.optimizers.legacy import Adam
except ImportError:
from tensorflow.keras.optimizers import Adam
class a ( tf.keras.optimizers.schedules.LearningRateSchedule ):
def __init__( self : Optional[Any] , lowercase_ : float , lowercase_ : Callable , lowercase_ : int , lowercase_ : float = 1.0 , lowercase_ : str = None , ):
super().__init__()
snake_case_ = initial_learning_rate
snake_case_ = warmup_steps
snake_case_ = power
snake_case_ = decay_schedule_fn
snake_case_ = name
def __call__( self : Tuple , lowercase_ : str ):
with tf.name_scope(self.name or '''WarmUp''' ) as name:
# Implements polynomial warmup. i.e., if global_step < warmup_steps, the
# learning rate will be `global_step/num_warmup_steps * init_lr`.
snake_case_ = tf.cast(lowercase_ , tf.floataa )
snake_case_ = tf.cast(self.warmup_steps , tf.floataa )
snake_case_ = global_step_float / warmup_steps_float
snake_case_ = self.initial_learning_rate * tf.math.pow(lowercase_ , self.power )
return tf.cond(
global_step_float < warmup_steps_float , lambda: warmup_learning_rate , lambda: self.decay_schedule_fn(step - self.warmup_steps ) , name=lowercase_ , )
def A_ ( self : Any ):
return {
"initial_learning_rate": self.initial_learning_rate,
"decay_schedule_fn": self.decay_schedule_fn,
"warmup_steps": self.warmup_steps,
"power": self.power,
"name": self.name,
}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, __UpperCAmelCase = 0.9, __UpperCAmelCase = 0.9_9_9, __UpperCAmelCase = 1e-8, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = 0.0, __UpperCAmelCase = 1.0, __UpperCAmelCase = None, ) -> List[str]:
'''simple docstring'''
snake_case_ = tf.keras.optimizers.schedules.PolynomialDecay(
initial_learning_rate=__UpperCAmelCase, decay_steps=num_train_steps - num_warmup_steps, end_learning_rate=init_lr * min_lr_ratio, power=__UpperCAmelCase, )
if num_warmup_steps:
snake_case_ = WarmUp(
initial_learning_rate=__UpperCAmelCase, decay_schedule_fn=__UpperCAmelCase, warmup_steps=__UpperCAmelCase, )
if weight_decay_rate > 0.0:
snake_case_ = AdamWeightDecay(
learning_rate=__UpperCAmelCase, weight_decay_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, exclude_from_weight_decay=['''LayerNorm''', '''layer_norm''', '''bias'''], include_in_weight_decay=__UpperCAmelCase, )
else:
snake_case_ = tf.keras.optimizers.Adam(
learning_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, )
# We return the optimizer and the LR scheduler in order to better track the
# evolution of the LR independently of the optimizer.
return optimizer, lr_schedule
class a ( _lowerCamelCase ):
def __init__( self : Dict , lowercase_ : Union[float, tf.keras.optimizers.schedules.LearningRateSchedule] = 0.001 , lowercase_ : float = 0.9 , lowercase_ : float = 0.999 , lowercase_ : float = 1e-7 , lowercase_ : bool = False , lowercase_ : float = 0.0 , lowercase_ : Optional[List[str]] = None , lowercase_ : Optional[List[str]] = None , lowercase_ : str = "AdamWeightDecay" , **lowercase_ : Optional[int] , ):
super().__init__(lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
snake_case_ = weight_decay_rate
snake_case_ = include_in_weight_decay
snake_case_ = exclude_from_weight_decay
@classmethod
def A_ ( cls : Dict , lowercase_ : Union[str, Any] ):
snake_case_ = {'''WarmUp''': WarmUp}
return super(lowercase_ , cls ).from_config(lowercase_ , custom_objects=lowercase_ )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : Optional[int] ):
super(lowercase_ , self )._prepare_local(lowercase_ , lowercase_ , lowercase_ )
snake_case_ = tf.constant(
self.weight_decay_rate , name='''adam_weight_decay_rate''' )
def A_ ( self : str , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Any ):
snake_case_ = self._do_use_weight_decay(var.name )
if do_decay:
return var.assign_sub(
learning_rate * var * apply_state[(var.device, var.dtype.base_dtype)]['''weight_decay_rate'''] , use_locking=self._use_locking , )
return tf.no_op()
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : str=None , **lowercase_ : List[str] ):
snake_case_ ,snake_case_ = list(zip(*lowercase_ ) )
return super(lowercase_ , self ).apply_gradients(zip(lowercase_ , lowercase_ ) , name=lowercase_ , **lowercase_ )
def A_ ( self : List[Any] , lowercase_ : str , lowercase_ : str , lowercase_ : Any ):
if apply_state is None:
return self._decayed_lr_t[var_dtype], {}
snake_case_ = apply_state or {}
snake_case_ = apply_state.get((var_device, var_dtype) )
if coefficients is None:
snake_case_ = self._fallback_apply_state(lowercase_ , lowercase_ )
snake_case_ = coefficients
return coefficients["lr_t"], {"apply_state": apply_state}
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : Union[str, Any] , lowercase_ : Optional[int]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_dense(lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : int , lowercase_ : str , lowercase_ : List[Any]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_sparse(lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = super().get_config()
config.update({'''weight_decay_rate''': self.weight_decay_rate} )
return config
def A_ ( self : Optional[int] , lowercase_ : int ):
if self.weight_decay_rate == 0:
return False
if self._include_in_weight_decay:
for r in self._include_in_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return True
if self._exclude_from_weight_decay:
for r in self._exclude_from_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return False
return True
class a ( _lowerCamelCase ):
def __init__( self : List[Any] ):
snake_case_ = []
snake_case_ = None
@property
def A_ ( self : Union[str, Any] ):
if self._accum_steps is None:
snake_case_ = tf.Variable(
tf.constant(0 , dtype=tf.intaa ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
return self._accum_steps.value()
@property
def A_ ( self : Dict ):
if not self._gradients:
raise ValueError('''The accumulator should be called first to initialize the gradients''' )
return [gradient.value() if gradient is not None else gradient for gradient in self._gradients]
def __call__( self : Any , lowercase_ : int ):
if not self._gradients:
snake_case_ = self.step # Create the step variable.
self._gradients.extend(
[
tf.Variable(
tf.zeros_like(lowercase_ ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
if gradient is not None
else gradient
for gradient in gradients
] )
if len(lowercase_ ) != len(self._gradients ):
raise ValueError(F"Expected {len(self._gradients )} gradients, but got {len(lowercase_ )}" )
for accum_gradient, gradient in zip(self._gradients , lowercase_ ):
if accum_gradient is not None and gradient is not None:
accum_gradient.assign_add(lowercase_ )
self._accum_steps.assign_add(1 )
def A_ ( self : Optional[int] ):
if not self._gradients:
return
self._accum_steps.assign(0 )
for gradient in self._gradients:
if gradient is not None:
gradient.assign(tf.zeros_like(lowercase_ ) )
| 56
| 1
|
'''simple docstring'''
import tempfile
import unittest
import numpy as np
from huggingface_hub import HfFolder, delete_repo
from requests.exceptions import HTTPError
from transformers import BertConfig, is_flax_available
from transformers.testing_utils import TOKEN, USER, is_staging_test, require_flax
if is_flax_available():
import os
from flax.core.frozen_dict import unfreeze
from flax.traverse_util import flatten_dict
from transformers import FlaxBertModel
a : List[str] = '0.12' # assumed parallelism: 8
@require_flax
@is_staging_test
class a ( unittest.TestCase ):
@classmethod
def A_ ( cls : str ):
snake_case_ = TOKEN
HfFolder.save_token(lowercase_ )
@classmethod
def A_ ( cls : Optional[Any] ):
try:
delete_repo(token=cls._token , repo_id='''test-model-flax''' )
except HTTPError:
pass
try:
delete_repo(token=cls._token , repo_id='''valid_org/test-model-flax-org''' )
except HTTPError:
pass
def A_ ( self : Tuple ):
snake_case_ = BertConfig(
vocab_size=99 , hidden_size=32 , num_hidden_layers=5 , num_attention_heads=4 , intermediate_size=37 )
snake_case_ = FlaxBertModel(lowercase_ )
model.push_to_hub('''test-model-flax''' , use_auth_token=self._token )
snake_case_ = FlaxBertModel.from_pretrained(F"{USER}/test-model-flax" )
snake_case_ = flatten_dict(unfreeze(model.params ) )
snake_case_ = flatten_dict(unfreeze(new_model.params ) )
for key in base_params.keys():
snake_case_ = (base_params[key] - new_params[key]).sum().item()
self.assertLessEqual(lowercase_ , 1e-3 , msg=F"{key} not identical" )
# Reset repo
delete_repo(token=self._token , repo_id='''test-model-flax''' )
# Push to hub via save_pretrained
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(lowercase_ , repo_id='''test-model-flax''' , push_to_hub=lowercase_ , use_auth_token=self._token )
snake_case_ = FlaxBertModel.from_pretrained(F"{USER}/test-model-flax" )
snake_case_ = flatten_dict(unfreeze(model.params ) )
snake_case_ = flatten_dict(unfreeze(new_model.params ) )
for key in base_params.keys():
snake_case_ = (base_params[key] - new_params[key]).sum().item()
self.assertLessEqual(lowercase_ , 1e-3 , msg=F"{key} not identical" )
def A_ ( self : Optional[Any] ):
snake_case_ = BertConfig(
vocab_size=99 , hidden_size=32 , num_hidden_layers=5 , num_attention_heads=4 , intermediate_size=37 )
snake_case_ = FlaxBertModel(lowercase_ )
model.push_to_hub('''valid_org/test-model-flax-org''' , use_auth_token=self._token )
snake_case_ = FlaxBertModel.from_pretrained('''valid_org/test-model-flax-org''' )
snake_case_ = flatten_dict(unfreeze(model.params ) )
snake_case_ = flatten_dict(unfreeze(new_model.params ) )
for key in base_params.keys():
snake_case_ = (base_params[key] - new_params[key]).sum().item()
self.assertLessEqual(lowercase_ , 1e-3 , msg=F"{key} not identical" )
# Reset repo
delete_repo(token=self._token , repo_id='''valid_org/test-model-flax-org''' )
# Push to hub via save_pretrained
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(
lowercase_ , repo_id='''valid_org/test-model-flax-org''' , push_to_hub=lowercase_ , use_auth_token=self._token )
snake_case_ = FlaxBertModel.from_pretrained('''valid_org/test-model-flax-org''' )
snake_case_ = flatten_dict(unfreeze(model.params ) )
snake_case_ = flatten_dict(unfreeze(new_model.params ) )
for key in base_params.keys():
snake_case_ = (base_params[key] - new_params[key]).sum().item()
self.assertLessEqual(lowercase_ , 1e-3 , msg=F"{key} not identical" )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = True
snake_case_ = flatten_dict(modela.params )
snake_case_ = flatten_dict(modela.params )
for key in flat_params_a.keys():
if np.sum(np.abs(flat_params_a[key] - flat_params_a[key] ) ) > 1e-4:
snake_case_ = False
return models_are_equal
@require_flax
class a ( unittest.TestCase ):
def A_ ( self : Union[str, Any] ):
snake_case_ = BertConfig.from_pretrained('''hf-internal-testing/tiny-bert-flax-only''' )
snake_case_ = FlaxBertModel(lowercase_ )
snake_case_ = '''bert'''
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(os.path.join(lowercase_ , lowercase_ ) )
with self.assertRaises(lowercase_ ):
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ )
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ , subfolder=lowercase_ )
self.assertTrue(check_models_equal(lowercase_ , lowercase_ ) )
def A_ ( self : Union[str, Any] ):
snake_case_ = BertConfig.from_pretrained('''hf-internal-testing/tiny-bert-flax-only''' )
snake_case_ = FlaxBertModel(lowercase_ )
snake_case_ = '''bert'''
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(os.path.join(lowercase_ , lowercase_ ) , max_shard_size='''10KB''' )
with self.assertRaises(lowercase_ ):
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ )
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ , subfolder=lowercase_ )
self.assertTrue(check_models_equal(lowercase_ , lowercase_ ) )
def A_ ( self : str ):
snake_case_ = '''bert'''
snake_case_ = '''hf-internal-testing/tiny-random-bert-subfolder'''
with self.assertRaises(lowercase_ ):
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ )
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ , subfolder=lowercase_ )
self.assertIsNotNone(lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = '''bert'''
snake_case_ = '''hf-internal-testing/tiny-random-bert-sharded-subfolder'''
with self.assertRaises(lowercase_ ):
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ )
snake_case_ = FlaxBertModel.from_pretrained(lowercase_ , subfolder=lowercase_ )
self.assertIsNotNone(lowercase_ )
| 56
|
'''simple docstring'''
import gc
import unittest
import torch
from parameterized import parameterized
from diffusers import AutoencoderKL
from diffusers.utils import floats_tensor, load_hf_numpy, require_torch_gpu, slow, torch_all_close, torch_device
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = AutoencoderKL
snake_case_ = "sample"
snake_case_ = 1e-2
@property
def A_ ( self : Dict ):
snake_case_ = 4
snake_case_ = 3
snake_case_ = (32, 32)
snake_case_ = floats_tensor((batch_size, num_channels) + sizes ).to(lowercase_ )
return {"sample": image}
@property
def A_ ( self : List[Any] ):
return (3, 32, 32)
@property
def A_ ( self : Dict ):
return (3, 32, 32)
def A_ ( self : Union[str, Any] ):
snake_case_ = {
'''block_out_channels''': [32, 64],
'''in_channels''': 3,
'''out_channels''': 3,
'''down_block_types''': ['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''],
'''up_block_types''': ['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''],
'''latent_channels''': 4,
}
snake_case_ = self.dummy_input
return init_dict, inputs_dict
def A_ ( self : Any ):
pass
def A_ ( self : str ):
pass
@unittest.skipIf(torch_device == '''mps''' , '''Gradient checkpointing skipped on MPS''' )
def A_ ( self : Dict ):
# enable deterministic behavior for gradient checkpointing
snake_case_ ,snake_case_ = self.prepare_init_args_and_inputs_for_common()
snake_case_ = self.model_class(**lowercase_ )
model.to(lowercase_ )
assert not model.is_gradient_checkpointing and model.training
snake_case_ = model(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model.zero_grad()
snake_case_ = torch.randn_like(lowercase_ )
snake_case_ = (out - labels).mean()
loss.backward()
# re-instantiate the model now enabling gradient checkpointing
snake_case_ = self.model_class(**lowercase_ )
# clone model
model_a.load_state_dict(model.state_dict() )
model_a.to(lowercase_ )
model_a.enable_gradient_checkpointing()
assert model_a.is_gradient_checkpointing and model_a.training
snake_case_ = model_a(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model_a.zero_grad()
snake_case_ = (out_a - labels).mean()
loss_a.backward()
# compare the output and parameters gradients
self.assertTrue((loss - loss_a).abs() < 1e-5 )
snake_case_ = dict(model.named_parameters() )
snake_case_ = dict(model_a.named_parameters() )
for name, param in named_params.items():
self.assertTrue(torch_all_close(param.grad.data , named_params_a[name].grad.data , atol=5e-5 ) )
def A_ ( self : Tuple ):
snake_case_ ,snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' , output_loading_info=lowercase_ )
self.assertIsNotNone(lowercase_ )
self.assertEqual(len(loading_info['''missing_keys'''] ) , 0 )
model.to(lowercase_ )
snake_case_ = model(**self.dummy_input )
assert image is not None, "Make sure output is not None"
def A_ ( self : Tuple ):
snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' )
snake_case_ = model.to(lowercase_ )
model.eval()
if torch_device == "mps":
snake_case_ = torch.manual_seed(0 )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(0 )
snake_case_ = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ = image.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , sample_posterior=lowercase_ , generator=lowercase_ ).sample
snake_case_ = output[0, -1, -3:, -3:].flatten().cpu()
# Since the VAE Gaussian prior's generator is seeded on the appropriate device,
# the expected output slices are not the same for CPU and GPU.
if torch_device == "mps":
snake_case_ = torch.tensor(
[
-4.0_078e-01,
-3.8_323e-04,
-1.2_681e-01,
-1.1_462e-01,
2.0_095e-01,
1.0_893e-01,
-8.8_247e-02,
-3.0_361e-01,
-9.8_644e-03,
] )
elif torch_device == "cpu":
snake_case_ = torch.tensor(
[-0.1352, 0.0878, 0.0419, -0.0818, -0.1069, 0.0688, -0.1458, -0.4446, -0.0026] )
else:
snake_case_ = torch.tensor(
[-0.2421, 0.4642, 0.2507, -0.0438, 0.0682, 0.3160, -0.2018, -0.0727, 0.2485] )
self.assertTrue(torch_all_close(lowercase_ , lowercase_ , rtol=1e-2 ) )
@slow
class a ( unittest.TestCase ):
def A_ ( self : Union[str, Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] ):
return F"gaussian_noise_s={seed}_shape={'_'.join([str(lowercase_ ) for s in shape] )}.npy"
def A_ ( self : Any ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Dict , lowercase_ : List[Any]=0 , lowercase_ : Union[str, Any]=(4, 3, 512, 512) , lowercase_ : Optional[Any]=False ):
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = torch.from_numpy(load_hf_numpy(self.get_file_format(lowercase_ , lowercase_ ) ) ).to(lowercase_ ).to(lowercase_ )
return image
def A_ ( self : Any , lowercase_ : Dict="CompVis/stable-diffusion-v1-4" , lowercase_ : List[str]=False ):
snake_case_ = '''fp16''' if fpaa else None
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = AutoencoderKL.from_pretrained(
lowercase_ , subfolder='''vae''' , torch_dtype=lowercase_ , revision=lowercase_ , )
model.to(lowercase_ ).eval()
return model
def A_ ( self : Any , lowercase_ : int=0 ):
if torch_device == "mps":
return torch.manual_seed(lowercase_ )
return torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
@parameterized.expand(
[
# fmt: off
[33, [-0.1603, 0.9878, -0.0495, -0.0790, -0.2709, 0.8375, -0.2060, -0.0824], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2376, 0.1168, 0.1332, -0.4840, -0.2508, -0.0791, -0.0493, -0.4089], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Union[str, Any] , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[33, [-0.0513, 0.0289, 1.3799, 0.2166, -0.2573, -0.0871, 0.5103, -0.0999]],
[47, [-0.4128, -0.1320, -0.3704, 0.1965, -0.4116, -0.2332, -0.3340, 0.2247]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Optional[int] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , fpaa=lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.1609, 0.9866, -0.0487, -0.0777, -0.2716, 0.8368, -0.2055, -0.0814], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2377, 0.1147, 0.1333, -0.4841, -0.2506, -0.0805, -0.0491, -0.4085], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Tuple , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[13, [-0.2051, -0.1803, -0.2311, -0.2114, -0.3292, -0.3574, -0.2953, -0.3323]],
[37, [-0.2632, -0.2625, -0.2199, -0.2741, -0.4539, -0.4990, -0.3720, -0.4925]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Dict , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-3 )
@parameterized.expand(
[
# fmt: off
[27, [-0.0369, 0.0207, -0.0776, -0.0682, -0.1747, -0.1930, -0.1465, -0.2039]],
[16, [-0.1628, -0.2134, -0.2747, -0.2642, -0.3774, -0.4404, -0.3687, -0.4277]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Union[str, Any] , lowercase_ : Any , lowercase_ : Optional[Any] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=5e-3 )
@parameterized.expand([(13,), (16,), (27,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : List[str] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-1 )
@parameterized.expand([(13,), (16,), (37,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.3001, 0.0918, -2.6984, -3.9720, -3.2099, -5.0353, 1.7338, -0.2065, 3.4267]],
[47, [-1.5030, -4.3871, -6.0355, -9.1157, -1.6661, -2.7853, 2.1607, -5.0823, 2.5633]],
# fmt: on
] )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model.encode(lowercase_ ).latent_dist
snake_case_ = dist.sample(generator=lowercase_ )
assert list(sample.shape ) == [image.shape[0], 4] + [i // 8 for i in image.shape[2:]]
snake_case_ = sample[0, -1, -3:, -3:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
snake_case_ = 3e-3 if torch_device != '''mps''' else 1e-2
assert torch_all_close(lowercase_ , lowercase_ , atol=lowercase_ )
| 56
| 1
|
'''simple docstring'''
import importlib.util
import json
import os
import warnings
from dataclasses import dataclass, field
import torch
from ..training_args import TrainingArguments
from ..utils import cached_property, is_sagemaker_dp_enabled, logging
a : List[Any] = logging.get_logger(__name__)
def __magic_name__ ( ) -> str:
'''simple docstring'''
snake_case_ = os.getenv('''SM_HP_MP_PARAMETERS''', '''{}''' )
try:
# Parse it and check the field "partitions" is included, it is required for model parallel.
snake_case_ = json.loads(__UpperCAmelCase )
if "partitions" not in smp_options:
return False
except json.JSONDecodeError:
return False
# Get the sagemaker specific framework parameters from mpi_options variable.
snake_case_ = os.getenv('''SM_FRAMEWORK_PARAMS''', '''{}''' )
try:
# Parse it and check the field "sagemaker_distributed_dataparallel_enabled".
snake_case_ = json.loads(__UpperCAmelCase )
if not mpi_options.get('''sagemaker_mpi_enabled''', __UpperCAmelCase ):
return False
except json.JSONDecodeError:
return False
# Lastly, check if the `smdistributed` module is present.
return importlib.util.find_spec('''smdistributed''' ) is not None
if is_sagemaker_model_parallel_available():
import smdistributed.modelparallel.torch as smp
smp.init()
@dataclass
class a ( _lowerCamelCase ):
snake_case_ = field(
default="" , metadata={"help": "Used by the SageMaker launcher to send mp-specific args. Ignored in SageMakerTrainer"} , )
def A_ ( self : Optional[Any] ):
super().__post_init__()
warnings.warn(
'''`SageMakerTrainingArguments` is deprecated and will be removed in v5 of Transformers. You can use '''
'''`TrainingArguments` instead.''' , lowercase_ , )
@cached_property
def A_ ( self : List[Any] ):
logger.info('''PyTorch: setting up devices''' )
if torch.distributed.is_available() and torch.distributed.is_initialized() and self.local_rank == -1:
logger.warning(
'''torch.distributed process group is initialized, but local_rank == -1. '''
'''In order to use Torch DDP, launch your script with `python -m torch.distributed.launch''' )
if self.no_cuda:
snake_case_ = torch.device('''cpu''' )
snake_case_ = 0
elif is_sagemaker_model_parallel_available():
snake_case_ = smp.local_rank()
snake_case_ = torch.device('''cuda''' , lowercase_ )
snake_case_ = 1
elif is_sagemaker_dp_enabled():
import smdistributed.dataparallel.torch.torch_smddp # noqa: F401
torch.distributed.init_process_group(backend='''smddp''' , timeout=self.ddp_timeout_delta )
snake_case_ = int(os.getenv('''SMDATAPARALLEL_LOCAL_RANK''' ) )
snake_case_ = torch.device('''cuda''' , self.local_rank )
snake_case_ = 1
elif self.local_rank == -1:
# if n_gpu is > 1 we'll use nn.DataParallel.
# If you only want to use a specific subset of GPUs use `CUDA_VISIBLE_DEVICES=0`
# Explicitly set CUDA to the first (index 0) CUDA device, otherwise `set_device` will
# trigger an error that a device index is missing. Index 0 takes into account the
# GPUs available in the environment, so `CUDA_VISIBLE_DEVICES=1,2` with `cuda:0`
# will use the first GPU in that env, i.e. GPU#1
snake_case_ = torch.device('''cuda:0''' if torch.cuda.is_available() else '''cpu''' )
# Sometimes the line in the postinit has not been run before we end up here, so just checking we're not at
# the default value.
snake_case_ = torch.cuda.device_count()
else:
# Here, we'll use torch.distributed.
# Initializes the distributed backend which will take care of synchronizing nodes/GPUs
if not torch.distributed.is_initialized():
torch.distributed.init_process_group(backend='''nccl''' , timeout=self.ddp_timeout_delta )
snake_case_ = torch.device('''cuda''' , self.local_rank )
snake_case_ = 1
if device.type == "cuda":
torch.cuda.set_device(lowercase_ )
return device
@property
def A_ ( self : Optional[int] ):
if is_sagemaker_model_parallel_available():
return smp.dp_size()
return super().world_size
@property
def A_ ( self : Dict ):
return not is_sagemaker_model_parallel_available()
@property
def A_ ( self : Optional[Any] ):
return False
| 56
|
'''simple docstring'''
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 ( _lowerCamelCase ):
snake_case_ = 42
@flax_register_to_config
class a ( nn.Module , _lowerCamelCase , _lowerCamelCase ):
snake_case_ = 32
snake_case_ = 4
snake_case_ = 4
snake_case_ = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
snake_case_ = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
snake_case_ = False
snake_case_ = (320, 640, 1_280, 1_280)
snake_case_ = 2
snake_case_ = 8
snake_case_ = None
snake_case_ = 1_280
snake_case_ = 0.0
snake_case_ = False
snake_case_ = jnp.floataa
snake_case_ = True
snake_case_ = 0
snake_case_ = False
def A_ ( self : Optional[int] , lowercase_ : jax.random.KeyArray ):
# init input tensors
snake_case_ = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case_ = jnp.zeros(lowercase_ , dtype=jnp.floataa )
snake_case_ = jnp.ones((1,) , dtype=jnp.intaa )
snake_case_ = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case_ ,snake_case_ = jax.random.split(lowercase_ )
snake_case_ = {'''params''': params_rng, '''dropout''': dropout_rng}
return self.init(lowercase_ , lowercase_ , lowercase_ , lowercase_ )["params"]
def A_ ( self : List[str] ):
snake_case_ = self.block_out_channels
snake_case_ = 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.
snake_case_ = self.num_attention_heads or self.attention_head_dim
# input
snake_case_ = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case_ = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case_ = FlaxTimestepEmbedding(lowercase_ , dtype=self.dtype )
snake_case_ = self.only_cross_attention
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (only_cross_attention,) * len(self.down_block_types )
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case_ = []
snake_case_ = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case_ = output_channel
snake_case_ = block_out_channels[i]
snake_case_ = i == len(lowercase_ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = down_blocks
# mid
snake_case_ = 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
snake_case_ = []
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case_ = output_channel
snake_case_ = reversed_block_out_channels[i]
snake_case_ = reversed_block_out_channels[min(i + 1 , len(lowercase_ ) - 1 )]
snake_case_ = i == len(lowercase_ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = output_channel
snake_case_ = up_blocks
# out
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : int , lowercase_ : Any , lowercase_ : int=None , lowercase_ : Any=None , lowercase_ : bool = True , lowercase_ : bool = False , ):
# 1. time
if not isinstance(lowercase_ , jnp.ndarray ):
snake_case_ = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(lowercase_ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case_ = timesteps.astype(dtype=jnp.floataa )
snake_case_ = jnp.expand_dims(lowercase_ , 0 )
snake_case_ = self.time_proj(lowercase_ )
snake_case_ = self.time_embedding(lowercase_ )
# 2. pre-process
snake_case_ = jnp.transpose(lowercase_ , (0, 2, 3, 1) )
snake_case_ = self.conv_in(lowercase_ )
# 3. down
snake_case_ = (sample,)
for down_block in self.down_blocks:
if isinstance(lowercase_ , lowercase_ ):
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , lowercase_ , deterministic=not train )
else:
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case_ = ()
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,)
snake_case_ = new_down_block_res_samples
# 4. mid
snake_case_ = 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:
snake_case_ = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case_ = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = up_block(
lowercase_ , temb=lowercase_ , encoder_hidden_states=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train , )
else:
snake_case_ = up_block(lowercase_ , temb=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train )
# 6. post-process
snake_case_ = self.conv_norm_out(lowercase_ )
snake_case_ = nn.silu(lowercase_ )
snake_case_ = self.conv_out(lowercase_ )
snake_case_ = jnp.transpose(lowercase_ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=lowercase_ )
| 56
| 1
|
'''simple docstring'''
import importlib
import os
import sys
# This is required to make the module import works (when the python process is running from the root of the repo)
sys.path.append('.')
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = test_file.split(os.path.sep )
if components[0:2] != ["tests", "models"]:
raise ValueError(
'''`test_file` should start with `tests/models/` (with `/` being the OS specific path separator). Got '''
F"{test_file} instead." )
snake_case_ = components[-1]
if not test_fn.endswith('''py''' ):
raise ValueError(F"`test_file` should be a python file. Got {test_fn} instead." )
if not test_fn.startswith('''test_modeling_''' ):
raise ValueError(
F"`test_file` should point to a file name of the form `test_modeling_*.py`. Got {test_fn} instead." )
snake_case_ = components[:-1] + [test_fn.replace('''.py''', '''''' )]
snake_case_ = '''.'''.join(__UpperCAmelCase )
return test_module_path
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = get_module_path(__UpperCAmelCase )
snake_case_ = importlib.import_module(__UpperCAmelCase )
return test_module
def __magic_name__ ( __UpperCAmelCase ) -> Tuple:
'''simple docstring'''
snake_case_ = []
snake_case_ = get_test_module(__UpperCAmelCase )
for attr in dir(__UpperCAmelCase ):
if attr.endswith('''ModelTester''' ):
tester_classes.append(getattr(__UpperCAmelCase, __UpperCAmelCase ) )
# sort with class names
return sorted(__UpperCAmelCase, key=lambda __UpperCAmelCase : x.__name__ )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = []
snake_case_ = get_test_module(__UpperCAmelCase )
for attr in dir(__UpperCAmelCase ):
snake_case_ = getattr(__UpperCAmelCase, __UpperCAmelCase )
# (TF/Flax)ModelTesterMixin is also an attribute in specific model test module. Let's exclude them by checking
# `all_model_classes` is not empty (which also excludes other special classes).
snake_case_ = getattr(__UpperCAmelCase, '''all_model_classes''', [] )
if len(__UpperCAmelCase ) > 0:
test_classes.append(__UpperCAmelCase )
# sort with class names
return sorted(__UpperCAmelCase, key=lambda __UpperCAmelCase : x.__name__ )
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = get_test_classes(__UpperCAmelCase )
snake_case_ = set()
for test_class in test_classes:
model_classes.update(test_class.all_model_classes )
# sort with class names
return sorted(__UpperCAmelCase, key=lambda __UpperCAmelCase : x.__name__ )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = test_class()
if hasattr(__UpperCAmelCase, '''setUp''' ):
test.setUp()
snake_case_ = None
if hasattr(__UpperCAmelCase, '''model_tester''' ):
# `(TF/Flax)ModelTesterMixin` has this attribute default to `None`. Let's skip this case.
if test.model_tester is not None:
snake_case_ = test.model_tester.__class__
return model_tester
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = get_test_classes(__UpperCAmelCase )
snake_case_ = []
for test_class in test_classes:
if model_class in test_class.all_model_classes:
target_test_classes.append(__UpperCAmelCase )
# sort with class names
return sorted(__UpperCAmelCase, key=lambda __UpperCAmelCase : x.__name__ )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = get_test_classes_for_model(__UpperCAmelCase, __UpperCAmelCase )
snake_case_ = []
for test_class in test_classes:
snake_case_ = get_model_tester_from_test_class(__UpperCAmelCase )
if tester_class is not None:
tester_classes.append(__UpperCAmelCase )
# sort with class names
return sorted(__UpperCAmelCase, key=lambda __UpperCAmelCase : x.__name__ )
def __magic_name__ ( __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = get_test_classes(__UpperCAmelCase )
snake_case_ = {test_class: get_model_tester_from_test_class(__UpperCAmelCase ) for test_class in test_classes}
return test_tester_mapping
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = get_model_classes(__UpperCAmelCase )
snake_case_ = {
model_class: get_test_classes_for_model(__UpperCAmelCase, __UpperCAmelCase ) for model_class in model_classes
}
return model_test_mapping
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = get_model_classes(__UpperCAmelCase )
snake_case_ = {
model_class: get_tester_classes_for_model(__UpperCAmelCase, __UpperCAmelCase ) for model_class in model_classes
}
return model_to_tester_mapping
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if isinstance(__UpperCAmelCase, __UpperCAmelCase ):
return o
elif isinstance(__UpperCAmelCase, __UpperCAmelCase ):
return o.__name__
elif isinstance(__UpperCAmelCase, (list, tuple) ):
return [to_json(__UpperCAmelCase ) for x in o]
elif isinstance(__UpperCAmelCase, __UpperCAmelCase ):
return {to_json(__UpperCAmelCase ): to_json(__UpperCAmelCase ) for k, v in o.items()}
else:
return o
| 56
|
'''simple docstring'''
import glob
import os
import random
from string import ascii_lowercase, digits
import cva
import numpy as np
# Parrameters
a : Dict = (720, 1280) # Height, Width
a : Tuple = (0.4, 0.6) # if height or width lower than this scale, drop it.
a : Dict = 1 / 100
a : str = ''
a : Any = ''
a : Optional[int] = ''
a : List[str] = 250
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dataset(__UpperCAmelCase, __UpperCAmelCase )
for index in range(__UpperCAmelCase ):
snake_case_ = random.sample(range(len(__UpperCAmelCase ) ), 4 )
snake_case_ ,snake_case_ ,snake_case_ = update_image_and_anno(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, filter_scale=__UpperCAmelCase, )
# Get random string code: '7b7ad245cdff75241935e4dd860f3bad'
snake_case_ = random_chars(32 )
snake_case_ = path.split(os.sep )[-1].rsplit('''.''', 1 )[0]
snake_case_ = F"{OUTPUT_DIR}/{file_name}_MOSAIC_{letter_code}"
cva.imwrite(F"{file_root}.jpg", __UpperCAmelCase, [cva.IMWRITE_JPEG_QUALITY, 85] )
print(F"Succeeded {index+1}/{NUMBER_IMAGES} with {file_name}" )
snake_case_ = []
for anno in new_annos:
snake_case_ = anno[3] - anno[1]
snake_case_ = anno[4] - anno[2]
snake_case_ = anno[1] + width / 2
snake_case_ = anno[2] + height / 2
snake_case_ = F"{anno[0]} {x_center} {y_center} {width} {height}"
annos_list.append(__UpperCAmelCase )
with open(F"{file_root}.txt", '''w''' ) as outfile:
outfile.write('''\n'''.join(line for line in annos_list ) )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> tuple[list, list]:
'''simple docstring'''
snake_case_ = []
snake_case_ = []
for label_file in glob.glob(os.path.join(__UpperCAmelCase, '''*.txt''' ) ):
snake_case_ = label_file.split(os.sep )[-1].rsplit('''.''', 1 )[0]
with open(__UpperCAmelCase ) as in_file:
snake_case_ = in_file.readlines()
snake_case_ = os.path.join(__UpperCAmelCase, F"{label_name}.jpg" )
snake_case_ = []
for obj_list in obj_lists:
snake_case_ = obj_list.rstrip('''\n''' ).split(''' ''' )
snake_case_ = float(obj[1] ) - float(obj[3] ) / 2
snake_case_ = float(obj[2] ) - float(obj[4] ) / 2
snake_case_ = float(obj[1] ) + float(obj[3] ) / 2
snake_case_ = float(obj[2] ) + float(obj[4] ) / 2
boxes.append([int(obj[0] ), xmin, ymin, xmax, ymax] )
if not boxes:
continue
img_paths.append(__UpperCAmelCase )
labels.append(__UpperCAmelCase )
return img_paths, labels
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, ) -> tuple[list, list, str]:
'''simple docstring'''
snake_case_ = np.zeros([output_size[0], output_size[1], 3], dtype=np.uinta )
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = int(scale_x * output_size[1] )
snake_case_ = int(scale_y * output_size[0] )
snake_case_ = []
snake_case_ = []
for i, index in enumerate(__UpperCAmelCase ):
snake_case_ = all_img_list[index]
path_list.append(__UpperCAmelCase )
snake_case_ = all_annos[index]
snake_case_ = cva.imread(__UpperCAmelCase )
if i == 0: # top-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = bbox[2] * scale_y
snake_case_ = bbox[3] * scale_x
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 1: # top-right
snake_case_ = cva.resize(__UpperCAmelCase, (output_size[1] - divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = bbox[2] * scale_y
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 2: # bottom-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = bbox[3] * scale_x
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
else: # bottom-right
snake_case_ = cva.resize(
__UpperCAmelCase, (output_size[1] - divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
# Remove bounding box small than scale of filter
if filter_scale > 0:
snake_case_ = [
anno
for anno in new_anno
if filter_scale < (anno[3] - anno[1]) and filter_scale < (anno[4] - anno[2])
]
return output_img, new_anno, path_list[0]
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
assert number_char > 1, "The number of character should greater than 1"
snake_case_ = ascii_lowercase + digits
return "".join(random.choice(__UpperCAmelCase ) for _ in range(__UpperCAmelCase ) )
if __name__ == "__main__":
main()
print('DONE ✅')
| 56
| 1
|
'''simple docstring'''
import math
import os
from copy import deepcopy
import datasets
import evaluate
import torch
import transformers
from datasets import load_dataset
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer
from accelerate import Accelerator
from accelerate.test_utils import RegressionDataset, RegressionModel
from accelerate.utils import is_tpu_available, set_seed
a : Optional[Any] = 'true'
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=82, __UpperCAmelCase=16 ) -> str:
'''simple docstring'''
set_seed(42 )
snake_case_ = RegressionModel()
snake_case_ = deepcopy(__UpperCAmelCase )
snake_case_ = RegressionDataset(length=__UpperCAmelCase )
snake_case_ = DataLoader(__UpperCAmelCase, batch_size=__UpperCAmelCase )
model.to(accelerator.device )
snake_case_ ,snake_case_ = accelerator.prepare(__UpperCAmelCase, __UpperCAmelCase )
return model, ddp_model, dataloader
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=False ) -> int:
'''simple docstring'''
snake_case_ = AutoTokenizer.from_pretrained('''hf-internal-testing/mrpc-bert-base-cased''' )
snake_case_ = load_dataset('''glue''', '''mrpc''', split='''validation''' )
def tokenize_function(__UpperCAmelCase ):
snake_case_ = tokenizer(examples['''sentence1'''], examples['''sentence2'''], truncation=__UpperCAmelCase, max_length=__UpperCAmelCase )
return outputs
with accelerator.main_process_first():
snake_case_ = dataset.map(
__UpperCAmelCase, batched=__UpperCAmelCase, remove_columns=['''idx''', '''sentence1''', '''sentence2'''], )
snake_case_ = tokenized_datasets.rename_column('''label''', '''labels''' )
def collate_fn(__UpperCAmelCase ):
if use_longest:
return tokenizer.pad(__UpperCAmelCase, padding='''longest''', return_tensors='''pt''' )
return tokenizer.pad(__UpperCAmelCase, padding='''max_length''', max_length=128, return_tensors='''pt''' )
return DataLoader(__UpperCAmelCase, shuffle=__UpperCAmelCase, collate_fn=__UpperCAmelCase, batch_size=16 )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = Accelerator(dispatch_batches=__UpperCAmelCase, split_batches=__UpperCAmelCase )
snake_case_ = get_dataloader(__UpperCAmelCase, not dispatch_batches )
snake_case_ = AutoModelForSequenceClassification.from_pretrained(
'''hf-internal-testing/mrpc-bert-base-cased''', return_dict=__UpperCAmelCase )
snake_case_ ,snake_case_ = accelerator.prepare(__UpperCAmelCase, __UpperCAmelCase )
return {"ddp": [ddp_model, ddp_dataloader, "cuda:0"], "no": [model, dataloader, accelerator.device]}, accelerator
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = []
for batch in dataloader:
snake_case_ ,snake_case_ = batch.values()
with torch.no_grad():
snake_case_ = model(__UpperCAmelCase )
snake_case_ ,snake_case_ = accelerator.gather_for_metrics((logit, target) )
logits_and_targets.append((logit, target) )
snake_case_ ,snake_case_ = [], []
for logit, targ in logits_and_targets:
logits.append(__UpperCAmelCase )
targs.append(__UpperCAmelCase )
snake_case_ ,snake_case_ = torch.cat(__UpperCAmelCase ), torch.cat(__UpperCAmelCase )
return logits, targs
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=82, __UpperCAmelCase=False, __UpperCAmelCase=False, __UpperCAmelCase=16 ) -> Dict:
'''simple docstring'''
snake_case_ ,snake_case_ ,snake_case_ = get_basic_setup(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
snake_case_ ,snake_case_ = generate_predictions(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
assert (
len(__UpperCAmelCase ) == num_samples
), F"Unexpected number of inputs:\n Expected: {num_samples}\n Actual: {len(__UpperCAmelCase )}"
def __magic_name__ ( __UpperCAmelCase = False, __UpperCAmelCase = False ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = evaluate.load('''glue''', '''mrpc''' )
snake_case_ ,snake_case_ = get_mrpc_setup(__UpperCAmelCase, __UpperCAmelCase )
# First do baseline
snake_case_ ,snake_case_ ,snake_case_ = setup['''no''']
model.to(__UpperCAmelCase )
model.eval()
for batch in dataloader:
batch.to(__UpperCAmelCase )
with torch.inference_mode():
snake_case_ = model(**__UpperCAmelCase )
snake_case_ = outputs.logits.argmax(dim=-1 )
metric.add_batch(predictions=__UpperCAmelCase, references=batch['''labels'''] )
snake_case_ = metric.compute()
# Then do distributed
snake_case_ ,snake_case_ ,snake_case_ = setup['''ddp''']
model.eval()
for batch in dataloader:
with torch.inference_mode():
snake_case_ = model(**__UpperCAmelCase )
snake_case_ = outputs.logits.argmax(dim=-1 )
snake_case_ = batch['''labels''']
snake_case_ ,snake_case_ = accelerator.gather_for_metrics((preds, references) )
metric.add_batch(predictions=__UpperCAmelCase, references=__UpperCAmelCase )
snake_case_ = metric.compute()
for key in "accuracy f1".split():
assert math.isclose(
baseline[key], distributed[key] ), F"Baseline and Distributed are not the same for key {key}:\n\tBaseline: {baseline[key]}\n\tDistributed: {distributed[key]}\n"
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = Accelerator(split_batches=__UpperCAmelCase, dispatch_batches=__UpperCAmelCase )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_warning()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
# These are a bit slower so they should only be ran on the GPU or TPU
if torch.cuda.is_available() or is_tpu_available():
if accelerator.is_local_main_process:
print('''**Testing gather_for_metrics**''' )
for split_batches in [True, False]:
for dispatch_batches in [True, False]:
if accelerator.is_local_main_process:
print(F"With: `split_batches={split_batches}`, `dispatch_batches={dispatch_batches}`" )
test_mrpc(__UpperCAmelCase, __UpperCAmelCase )
accelerator.state._reset_state()
if accelerator.is_local_main_process:
print('''**Test torch metrics**''' )
for split_batches in [True, False]:
for dispatch_batches in [True, False]:
snake_case_ = Accelerator(split_batches=__UpperCAmelCase, dispatch_batches=__UpperCAmelCase )
if accelerator.is_local_main_process:
print(F"With: `split_batches={split_batches}`, `dispatch_batches={dispatch_batches}`, length=99" )
test_torch_metrics(__UpperCAmelCase, 99 )
accelerator.state._reset_state()
if accelerator.is_local_main_process:
print('''**Test last batch is not dropped when perfectly divisible**''' )
snake_case_ = Accelerator()
test_torch_metrics(__UpperCAmelCase, 512 )
accelerator.state._reset_state()
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
main()
if __name__ == "__main__":
main()
| 56
|
'''simple docstring'''
import unittest
from transformers import (
MODEL_FOR_OBJECT_DETECTION_MAPPING,
AutoFeatureExtractor,
AutoModelForObjectDetection,
ObjectDetectionPipeline,
is_vision_available,
pipeline,
)
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_pytesseract,
require_tf,
require_timm,
require_torch,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_vision_available():
from PIL import Image
else:
class a :
@staticmethod
def A_ ( *lowercase_ : int , **lowercase_ : str ):
pass
@is_pipeline_test
@require_vision
@require_timm
@require_torch
class a ( unittest.TestCase ):
snake_case_ = MODEL_FOR_OBJECT_DETECTION_MAPPING
def A_ ( self : Any , lowercase_ : List[Any] , lowercase_ : Optional[int] , lowercase_ : List[str] ):
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , image_processor=lowercase_ )
return object_detector, ["./tests/fixtures/tests_samples/COCO/000000039769.png"]
def A_ ( self : Dict , lowercase_ : List[Any] , lowercase_ : int ):
snake_case_ = object_detector('''./tests/fixtures/tests_samples/COCO/000000039769.png''' , threshold=0.0 )
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
import datasets
snake_case_ = datasets.load_dataset('''hf-internal-testing/fixtures_image_utils''' , '''image''' , split='''test''' )
snake_case_ = [
Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' ),
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
# RGBA
dataset[0]['''file'''],
# LA
dataset[1]['''file'''],
# L
dataset[2]['''file'''],
]
snake_case_ = object_detector(lowercase_ , threshold=0.0 )
self.assertEqual(len(lowercase_ ) , len(lowercase_ ) )
for outputs in batch_outputs:
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
@require_tf
@unittest.skip('''Object detection not implemented in TF''' )
def A_ ( self : int ):
pass
@require_torch
def A_ ( self : Tuple ):
snake_case_ = '''hf-internal-testing/tiny-detr-mobilenetsv3'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=0.0 )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] , threshold=0.0 , )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
] , )
@require_torch
@slow
def A_ ( self : Optional[int] ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : Tuple ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : str ):
snake_case_ = 0.9985
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=lowercase_ )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
@require_torch
@require_pytesseract
@slow
def A_ ( self : Dict ):
snake_case_ = '''Narsil/layoutlmv3-finetuned-funsd'''
snake_case_ = 0.9993
snake_case_ = pipeline('''object-detection''' , model=lowercase_ , threshold=lowercase_ )
snake_case_ = object_detector(
'''https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
] , )
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
from PIL import Image
# Define glider example
a : Optional[int] = [
[0, 1, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 0, 0],
[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],
]
# Define blinker example
a : List[Any] = [[0, 1, 0], [0, 1, 0], [0, 1, 0]]
def __magic_name__ ( __UpperCAmelCase ) -> list[list[int]]:
'''simple docstring'''
snake_case_ = []
for i in range(len(__UpperCAmelCase ) ):
snake_case_ = []
for j in range(len(cells[i] ) ):
# Get the number of live neighbours
snake_case_ = 0
if i > 0 and j > 0:
neighbour_count += cells[i - 1][j - 1]
if i > 0:
neighbour_count += cells[i - 1][j]
if i > 0 and j < len(cells[i] ) - 1:
neighbour_count += cells[i - 1][j + 1]
if j > 0:
neighbour_count += cells[i][j - 1]
if j < len(cells[i] ) - 1:
neighbour_count += cells[i][j + 1]
if i < len(__UpperCAmelCase ) - 1 and j > 0:
neighbour_count += cells[i + 1][j - 1]
if i < len(__UpperCAmelCase ) - 1:
neighbour_count += cells[i + 1][j]
if i < len(__UpperCAmelCase ) - 1 and j < len(cells[i] ) - 1:
neighbour_count += cells[i + 1][j + 1]
# Rules of the game of life (excerpt from Wikipedia):
# 1. Any live cell with two or three live neighbours survives.
# 2. Any dead cell with three live neighbours becomes a live cell.
# 3. All other live cells die in the next generation.
# Similarly, all other dead cells stay dead.
snake_case_ = cells[i][j] == 1
if (
(alive and 2 <= neighbour_count <= 3)
or not alive
and neighbour_count == 3
):
next_generation_row.append(1 )
else:
next_generation_row.append(0 )
next_generation.append(__UpperCAmelCase )
return next_generation
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> list[Image.Image]:
'''simple docstring'''
snake_case_ = []
for _ in range(__UpperCAmelCase ):
# Create output image
snake_case_ = Image.new('''RGB''', (len(cells[0] ), len(__UpperCAmelCase )) )
snake_case_ = img.load()
# Save cells to image
for x in range(len(__UpperCAmelCase ) ):
for y in range(len(cells[0] ) ):
snake_case_ = 255 - cells[y][x] * 255
snake_case_ = (colour, colour, colour)
# Save image
images.append(__UpperCAmelCase )
snake_case_ = new_generation(__UpperCAmelCase )
return images
if __name__ == "__main__":
a : Dict = generate_images(GLIDER, 16)
images[0].save('out.gif', save_all=True, append_images=images[1:])
| 56
|
'''simple docstring'''
import unittest
from transformers import MPNetConfig, is_torch_available
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 (
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
class a :
def __init__( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Any=13 , lowercase_ : Optional[Any]=7 , lowercase_ : Optional[Any]=True , lowercase_ : Dict=True , lowercase_ : Tuple=False , lowercase_ : Optional[Any]=True , lowercase_ : Any=99 , lowercase_ : Union[str, Any]=64 , lowercase_ : str=5 , lowercase_ : int=4 , lowercase_ : List[Any]=64 , lowercase_ : Dict="gelu" , lowercase_ : Optional[int]=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Tuple=512 , lowercase_ : List[Any]=16 , lowercase_ : str=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[Any]=3 , lowercase_ : Optional[Any]=4 , lowercase_ : List[Any]=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_input_mask
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = scope
def A_ ( self : List[str] ):
return MPNetConfig.from_pretrained('''microsoft/mpnet-base''' )
def A_ ( self : str ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_input_mask:
snake_case_ = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = ids_tensor([self.batch_size] , self.num_choices )
snake_case_ = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def A_ ( self : Tuple ):
return MPNetConfig(
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 , initializer_range=self.initializer_range , )
def A_ ( self : Any , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Optional[int] ):
snake_case_ = MPNetModel(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , lowercase_ )
snake_case_ = 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 : str , lowercase_ : Optional[Any] , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = MPNetForQuestionAnswering(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Tuple , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : str , lowercase_ : Optional[int] , lowercase_ : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = MPNetForSequenceClassification(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , attention_mask=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : Any , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.num_choices
snake_case_ = MPNetForMultipleChoice(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = model(
lowercase_ , attention_mask=lowercase_ , labels=lowercase_ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def A_ ( self : Union[str, Any] , lowercase_ : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] , lowercase_ : str , lowercase_ : int , lowercase_ : List[str] ):
snake_case_ = self.num_labels
snake_case_ = MPNetForTokenClassification(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Union[str, Any] ):
snake_case_ = self.prepare_config_and_inputs()
((snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_)) = config_and_inputs
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
if is_torch_available()
else ()
)
snake_case_ = (
{
"feature-extraction": MPNetModel,
"fill-mask": MPNetForMaskedLM,
"question-answering": MPNetForQuestionAnswering,
"text-classification": MPNetForSequenceClassification,
"token-classification": MPNetForTokenClassification,
"zero-shot": MPNetForSequenceClassification,
}
if is_torch_available()
else {}
)
snake_case_ = False
snake_case_ = True
def A_ ( self : Tuple ):
snake_case_ = MPNetModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , hidden_size=37 )
def A_ ( self : Union[str, Any] ):
self.config_tester.run_common_tests()
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_model(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_sequence_classification(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_multiple_choice(*lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_token_classification(*lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_question_answering(*lowercase_ )
@require_torch
class a ( unittest.TestCase ):
@slow
def A_ ( self : List[Any] ):
snake_case_ = MPNetModel.from_pretrained('''microsoft/mpnet-base''' )
snake_case_ = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]] )
snake_case_ = model(lowercase_ )[0]
snake_case_ = torch.Size((1, 11, 768) )
self.assertEqual(output.shape , lowercase_ )
snake_case_ = torch.tensor(
[[[-0.0550, 0.1943, -0.0740], [-0.0562, 0.2211, -0.0579], [-0.0437, 0.3337, -0.0641]]] )
# compare the actual values for a slice.
self.assertTrue(torch.allclose(output[:, :3, :3] , lowercase_ , atol=1e-4 ) )
| 56
| 1
|
'''simple docstring'''
a : Tuple = '\n# Transformers installation\n! pip install transformers datasets\n# To install from source instead of the last release, comment the command above and uncomment the following one.\n# ! pip install git+https://github.com/huggingface/transformers.git\n'
a : Any = [{'type': 'code', 'content': INSTALL_CONTENT}]
a : Optional[int] = {
'{processor_class}': 'FakeProcessorClass',
'{model_class}': 'FakeModelClass',
'{object_class}': 'FakeObjectClass',
}
| 56
|
'''simple docstring'''
import json
import os
import pickle
import shutil
import tempfile
from unittest import TestCase
from unittest.mock import patch
import numpy as np
from datasets import Dataset
from transformers import is_faiss_available
from transformers.models.bart.configuration_bart import BartConfig
from transformers.models.bart.tokenization_bart import BartTokenizer
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES as DPR_VOCAB_FILES_NAMES
from transformers.models.dpr.configuration_dpr import DPRConfig
from transformers.models.dpr.tokenization_dpr import DPRContextEncoderTokenizer, DPRQuestionEncoderTokenizer
from transformers.models.rag.configuration_rag import RagConfig
from transformers.models.rag.retrieval_rag import CustomHFIndex, RagRetriever
from transformers.models.roberta.tokenization_roberta import VOCAB_FILES_NAMES as BART_VOCAB_FILES_NAMES
from transformers.testing_utils import require_faiss, require_sentencepiece, require_tokenizers, require_torch
if is_faiss_available():
import faiss
@require_faiss
class a ( _lowerCamelCase ):
def A_ ( self : str ):
snake_case_ = tempfile.mkdtemp()
snake_case_ = 8
# DPR tok
snake_case_ = [
'''[UNK]''',
'''[CLS]''',
'''[SEP]''',
'''[PAD]''',
'''[MASK]''',
'''want''',
'''##want''',
'''##ed''',
'''wa''',
'''un''',
'''runn''',
'''##ing''',
''',''',
'''low''',
'''lowest''',
]
snake_case_ = os.path.join(self.tmpdirname , '''dpr_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , DPR_VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
# BART tok
snake_case_ = [
'''l''',
'''o''',
'''w''',
'''e''',
'''r''',
'''s''',
'''t''',
'''i''',
'''d''',
'''n''',
'''\u0120''',
'''\u0120l''',
'''\u0120n''',
'''\u0120lo''',
'''\u0120low''',
'''er''',
'''\u0120lowest''',
'''\u0120newer''',
'''\u0120wider''',
'''<unk>''',
]
snake_case_ = dict(zip(lowercase_ , range(len(lowercase_ ) ) ) )
snake_case_ = ['''#version: 0.2''', '''\u0120 l''', '''\u0120l o''', '''\u0120lo w''', '''e r''', '''''']
snake_case_ = {'''unk_token''': '''<unk>'''}
snake_case_ = os.path.join(self.tmpdirname , '''bart_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''merges_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write(json.dumps(lowercase_ ) + '''\n''' )
with open(self.merges_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write('''\n'''.join(lowercase_ ) )
def A_ ( self : Union[str, Any] ):
return DPRQuestionEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : Union[str, Any] ):
return DPRContextEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : int ):
return BartTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''bart_tokenizer''' ) )
def A_ ( self : str ):
shutil.rmtree(self.tmpdirname )
def A_ ( self : str ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size ), 2 * np.ones(self.retrieval_vector_size )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
return dataset
def A_ ( self : str ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , )
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
return retriever
def A_ ( self : str , lowercase_ : bool ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''custom''' , )
if from_disk:
snake_case_ = os.path.join(self.tmpdirname , '''dataset''' )
snake_case_ = os.path.join(self.tmpdirname , '''index.faiss''' )
dataset.get_index('''embeddings''' ).save(os.path.join(self.tmpdirname , '''index.faiss''' ) )
dataset.drop_index('''embeddings''' )
dataset.save_to_disk(os.path.join(self.tmpdirname , '''dataset''' ) )
del dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
else:
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , index=CustomHFIndex(config.retrieval_vector_size , lowercase_ ) , )
return retriever
def A_ ( self : Tuple ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size + 1 ), 2 * np.ones(self.retrieval_vector_size + 1 )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
snake_case_ = os.path.join(self.tmpdirname , '''hf_bert_base.hnswSQ8_correct_phi_128.c_index''' )
dataset.save_faiss_index('''embeddings''' , index_file_name + '''.index.dpr''' )
pickle.dump(dataset['''id'''] , open(index_file_name + '''.index_meta.dpr''' , '''wb''' ) )
snake_case_ = os.path.join(self.tmpdirname , '''psgs_w100.tsv.pkl''' )
snake_case_ = {sample['''id''']: [sample['''text'''], sample['''title''']] for sample in dataset}
pickle.dump(lowercase_ , open(lowercase_ , '''wb''' ) )
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''legacy''' , index_path=self.tmpdirname , )
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() )
return retriever
def A_ ( self : Optional[Any] ):
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : str ):
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = self.get_dummy_dataset()
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : int ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : str ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : Any ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : Any ):
snake_case_ = 1
snake_case_ = self.get_dummy_legacy_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''text'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''text'''][0] , '''bar''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''text'''][0] , '''foo''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_legacy_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : List[str] ):
import torch
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
snake_case_ ,snake_case_ ,snake_case_ = (
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , np.ndarray )
snake_case_ = retriever(
lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ , return_tensors='''pt''' , )
snake_case_ ,snake_case_ ,snake_case_ ,snake_case_ = ( # noqa: F841
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
out['''doc_ids'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : Tuple ):
snake_case_ = self.get_dpr_ctx_encoder_tokenizer()
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
retriever.set_ctx_encoder_tokenizer(lowercase_ )
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
self.assertEqual(
len(lowercase_ ) , 6 ) # check whether the retriever output consist of 6 attributes including tokenized docs
self.assertEqual(
all(k in out for k in ('''tokenized_doc_ids''', '''tokenized_doc_attention_mask''') ) , lowercase_ ) # check for doc token related keys in dictionary.
| 56
| 1
|
'''simple docstring'''
import argparse
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
CLIPTokenizer,
CLIPTokenizerFast,
VideoMAEImageProcessor,
XCLIPConfig,
XCLIPModel,
XCLIPProcessor,
XCLIPTextConfig,
XCLIPVisionConfig,
)
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = XCLIPTextConfig()
# derive patch size from model name
snake_case_ = model_name.find('''patch''' )
snake_case_ = int(model_name[start_idx + len('''patch''' ) : start_idx + len('''patch''' ) + 2] )
snake_case_ = XCLIPVisionConfig(patch_size=__UpperCAmelCase, num_frames=__UpperCAmelCase )
if "large" in model_name:
snake_case_ = 768
snake_case_ = 3072
snake_case_ = 12
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 16
snake_case_ = 24
snake_case_ = 768
snake_case_ = 3072
if model_name == "xclip-large-patch14-16-frames":
snake_case_ = 336
snake_case_ = XCLIPConfig.from_text_vision_configs(__UpperCAmelCase, __UpperCAmelCase )
if "large" in model_name:
snake_case_ = 768
return config
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if name == "token_embedding.weight":
snake_case_ = name.replace('''token_embedding.weight''', '''text_model.embeddings.token_embedding.weight''' )
if name == "positional_embedding":
snake_case_ = name.replace('''positional_embedding''', '''text_model.embeddings.position_embedding.weight''' )
if "ln_1" in name:
snake_case_ = name.replace('''ln_1''', '''layer_norm1''' )
if "ln_2" in name:
snake_case_ = name.replace('''ln_2''', '''layer_norm2''' )
if "c_fc" in name:
snake_case_ = name.replace('''c_fc''', '''fc1''' )
if "c_proj" in name:
snake_case_ = name.replace('''c_proj''', '''fc2''' )
if name.startswith('''transformer.resblocks''' ):
snake_case_ = name.replace('''transformer.resblocks''', '''text_model.encoder.layers''' )
if "attn.out_proj" in name and "message" not in name:
snake_case_ = name.replace('''attn.out_proj''', '''self_attn.out_proj''' )
if "ln_final" in name:
snake_case_ = name.replace('''ln_final''', '''text_model.final_layer_norm''' )
# visual encoder
if name == "visual.class_embedding":
snake_case_ = name.replace('''visual.class_embedding''', '''vision_model.embeddings.class_embedding''' )
if name == "visual.positional_embedding":
snake_case_ = name.replace('''visual.positional_embedding''', '''vision_model.embeddings.position_embedding.weight''' )
if name.startswith('''visual.transformer.resblocks''' ):
snake_case_ = name.replace('''visual.transformer.resblocks''', '''vision_model.encoder.layers''' )
if "visual.conv1" in name:
snake_case_ = name.replace('''visual.conv1''', '''vision_model.embeddings.patch_embedding''' )
if "visual.ln_pre" in name:
snake_case_ = name.replace('''visual.ln_pre''', '''vision_model.pre_layernorm''' )
if "visual.ln_post" in name:
snake_case_ = name.replace('''visual.ln_post''', '''vision_model.post_layernorm''' )
if "visual.proj" in name:
snake_case_ = name.replace('''visual.proj''', '''visual_projection.weight''' )
if "text_projection" in name:
snake_case_ = name.replace('''text_projection''', '''text_projection.weight''' )
# things on top
if "prompts_visual_proj" in name:
snake_case_ = name.replace('''prompts_visual_proj''', '''prompts_visual_projection''' )
if "prompts_visual_ln" in name:
snake_case_ = name.replace('''prompts_visual_ln''', '''prompts_visual_layernorm''' )
# mit
if name == "mit.positional_embedding":
snake_case_ = name.replace('''positional''', '''position''' )
if name.startswith('''mit.resblocks''' ):
snake_case_ = name.replace('''mit.resblocks''', '''mit.encoder.layers''' )
# prompts generator
if name.startswith('''prompts_generator.norm''' ):
snake_case_ = name.replace('''prompts_generator.norm''', '''prompts_generator.layernorm''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ = orig_state_dict.pop(__UpperCAmelCase )
if "attn.in_proj" in key:
snake_case_ = key.split('''.''' )
if key.startswith('''visual''' ):
snake_case_ = key_split[3]
snake_case_ = config.vision_config.hidden_size
if "message_attn" in key:
if "weight" in key:
snake_case_ = val[
:dim, :
]
snake_case_ = val[
dim : dim * 2, :
]
snake_case_ = val[
-dim:, :
]
else:
snake_case_ = val[
:dim
]
snake_case_ = val[
dim : dim * 2
]
snake_case_ = val[
-dim:
]
else:
if "weight" in key:
snake_case_ = val[
:dim, :
]
snake_case_ = val[
dim : dim * 2, :
]
snake_case_ = val[
-dim:, :
]
else:
snake_case_ = val[:dim]
snake_case_ = val[
dim : dim * 2
]
snake_case_ = val[-dim:]
elif key.startswith('''mit''' ):
snake_case_ = key_split[2]
snake_case_ = config.vision_config.mit_hidden_size
if "weight" in key:
snake_case_ = val[:dim, :]
snake_case_ = val[dim : dim * 2, :]
snake_case_ = val[-dim:, :]
else:
snake_case_ = val[:dim]
snake_case_ = val[dim : dim * 2]
snake_case_ = val[-dim:]
else:
snake_case_ = key_split[2]
snake_case_ = config.text_config.hidden_size
if "weight" in key:
snake_case_ = val[:dim, :]
snake_case_ = val[
dim : dim * 2, :
]
snake_case_ = val[-dim:, :]
else:
snake_case_ = val[:dim]
snake_case_ = val[
dim : dim * 2
]
snake_case_ = val[-dim:]
else:
snake_case_ = rename_key(__UpperCAmelCase )
if new_key_name in ["visual_projection.weight", "text_projection.weight"]:
snake_case_ = val.T
snake_case_ = val
return orig_state_dict
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if num_frames == 8:
snake_case_ = '''eating_spaghetti_8_frames.npy'''
elif num_frames == 16:
snake_case_ = '''eating_spaghetti.npy'''
elif num_frames == 32:
snake_case_ = '''eating_spaghetti_32_frames.npy'''
snake_case_ = hf_hub_download(
repo_id='''hf-internal-testing/spaghetti-video''', filename=__UpperCAmelCase, repo_type='''dataset''', )
snake_case_ = np.load(__UpperCAmelCase )
return list(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=None, __UpperCAmelCase=False ) -> List[str]:
'''simple docstring'''
snake_case_ = {
# fully supervised kinetics-400 checkpoints
'''xclip-base-patch32''': '''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth''',
'''xclip-base-patch32-16-frames''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth'''
),
'''xclip-base-patch16''': '''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth''',
'''xclip-base-patch16-16-frames''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth'''
),
'''xclip-large-patch14''': '''https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&export=download&confirm=t&uuid=b26caedc-88e2-473e-830a-9d158b653cdb''',
'''xclip-large-patch14-16-frames''': '''https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&export=download&confirm=t&uuid=538fa810-e671-4050-b385-9a623f89804f''',
# fully supervised kinetics-600 checkpoints
'''xclip-base-patch16-kinetics-600''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth'''
),
'''xclip-base-patch16-kinetics-600-16-frames''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth'''
),
'''xclip-large-patch14-kinetics-600''': '''https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&export=download&confirm=t&uuid=141d4977-4a65-44ae-864f-4b0c19f838be''',
# few shot
'''xclip-base-patch16-hmdb-2-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth'''
),
'''xclip-base-patch16-hmdb-4-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth'''
),
'''xclip-base-patch16-hmdb-8-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth'''
),
'''xclip-base-patch16-hmdb-16-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth'''
),
'''xclip-base-patch16-ucf-2-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth'''
),
'''xclip-base-patch16-ucf-4-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth'''
),
'''xclip-base-patch16-ucf-8-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth'''
),
'''xclip-base-patch16-ucf-16-shot''': (
'''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth'''
),
# zero shot
'''xclip-base-patch16-zero-shot''': '''https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth''',
}
snake_case_ = model_to_url[model_name]
snake_case_ = 8
if "16-frames" in model_name:
snake_case_ = 16
elif "shot" in model_name:
snake_case_ = 32
snake_case_ = get_xclip_config(__UpperCAmelCase, __UpperCAmelCase )
snake_case_ = XCLIPModel(__UpperCAmelCase )
model.eval()
if "drive" in checkpoint_url:
snake_case_ = '''pytorch_model.bin'''
gdown.cached_download(__UpperCAmelCase, __UpperCAmelCase, quiet=__UpperCAmelCase )
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )['''model''']
else:
snake_case_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase )['''model''']
snake_case_ = convert_state_dict(__UpperCAmelCase, __UpperCAmelCase )
snake_case_ = XCLIPModel(__UpperCAmelCase )
snake_case_ ,snake_case_ = model.load_state_dict(__UpperCAmelCase, strict=__UpperCAmelCase )
assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"]
model.eval()
snake_case_ = 336 if model_name == '''xclip-large-patch14-16-frames''' else 224
snake_case_ = VideoMAEImageProcessor(size=__UpperCAmelCase )
snake_case_ = CLIPTokenizer.from_pretrained('''openai/clip-vit-base-patch32''' )
snake_case_ = CLIPTokenizerFast.from_pretrained('''openai/clip-vit-base-patch32''' )
snake_case_ = XCLIPProcessor(image_processor=__UpperCAmelCase, tokenizer=__UpperCAmelCase )
snake_case_ = prepare_video(__UpperCAmelCase )
snake_case_ = processor(
text=['''playing sports''', '''eating spaghetti''', '''go shopping'''], videos=__UpperCAmelCase, return_tensors='''pt''', padding=__UpperCAmelCase )
print('''Shape of pixel values:''', inputs.pixel_values.shape )
with torch.no_grad():
snake_case_ = model(**__UpperCAmelCase )
# Verify outputs
snake_case_ = outputs.logits_per_video
snake_case_ = logits_per_video.softmax(dim=1 )
print('''Probs:''', __UpperCAmelCase )
# kinetics-400
if model_name == "xclip-base-patch32":
snake_case_ = torch.tensor([[0.0_0_1_9, 0.9_9_5_1, 0.0_0_3_0]] )
elif model_name == "xclip-base-patch32-16-frames":
snake_case_ = torch.tensor([[7.0999e-04, 9.9883e-01, 4.5580e-04]] )
elif model_name == "xclip-base-patch16":
snake_case_ = torch.tensor([[0.0_0_8_3, 0.9_6_8_1, 0.0_2_3_6]] )
elif model_name == "xclip-base-patch16-16-frames":
snake_case_ = torch.tensor([[7.6937e-04, 9.9728e-01, 1.9473e-03]] )
elif model_name == "xclip-large-patch14":
snake_case_ = torch.tensor([[0.0_0_6_2, 0.9_8_6_4, 0.0_0_7_5]] )
elif model_name == "xclip-large-patch14-16-frames":
snake_case_ = torch.tensor([[3.3877e-04, 9.9937e-01, 2.8888e-04]] )
# kinetics-600
elif model_name == "xclip-base-patch16-kinetics-600":
snake_case_ = torch.tensor([[0.0_5_5_5, 0.8_9_1_4, 0.0_5_3_1]] )
elif model_name == "xclip-base-patch16-kinetics-600-16-frames":
snake_case_ = torch.tensor([[3.8554e-04, 9.9929e-01, 3.2754e-04]] )
elif model_name == "xclip-large-patch14-kinetics-600":
snake_case_ = torch.tensor([[0.0_0_3_6, 0.9_9_2_0, 0.0_0_4_5]] )
# few shot
elif model_name == "xclip-base-patch16-hmdb-2-shot":
snake_case_ = torch.tensor([[7.1890e-06, 9.9994e-01, 5.6559e-05]] )
elif model_name == "xclip-base-patch16-hmdb-4-shot":
snake_case_ = torch.tensor([[1.0320e-05, 9.9993e-01, 6.2435e-05]] )
elif model_name == "xclip-base-patch16-hmdb-8-shot":
snake_case_ = torch.tensor([[4.1377e-06, 9.9990e-01, 9.8386e-05]] )
elif model_name == "xclip-base-patch16-hmdb-16-shot":
snake_case_ = torch.tensor([[4.1347e-05, 9.9962e-01, 3.3411e-04]] )
elif model_name == "xclip-base-patch16-ucf-2-shot":
snake_case_ = torch.tensor([[8.5857e-05, 9.9928e-01, 6.3291e-04]] )
elif model_name == "xclip-base-patch16-ucf-4-shot":
snake_case_ = torch.tensor([[8.5857e-05, 9.9928e-01, 6.3291e-04]] )
elif model_name == "xclip-base-patch16-ucf-8-shot":
snake_case_ = torch.tensor([[0.0_0_2_7, 0.9_9_0_4, 0.0_0_7_0]] )
elif model_name == "xclip-base-patch16-ucf-16-shot":
snake_case_ = torch.tensor([[9.8219e-04, 9.9593e-01, 3.0863e-03]] )
# zero shot
elif model_name == "xclip-base-patch16-zero-shot":
snake_case_ = torch.tensor([[3.5082e-04, 9.9785e-01, 1.7966e-03]] )
else:
raise ValueError(F"Model name {model_name} not supported" )
assert torch.allclose(__UpperCAmelCase, __UpperCAmelCase, atol=1e-3 )
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 )
if push_to_hub:
print('''Pushing model, processor and slow tokenizer files to the hub...''' )
model.push_to_hub(__UpperCAmelCase, organization='''nielsr''' )
processor.push_to_hub(__UpperCAmelCase, organization='''nielsr''' )
slow_tokenizer.push_to_hub(__UpperCAmelCase, organization='''nielsr''' )
if __name__ == "__main__":
a : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--model_name',
default='xclip-base-patch32',
type=str,
help='Name of the model.',
)
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.'
)
a : Optional[int] = parser.parse_args()
convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 56
|
'''simple docstring'''
import os
import re
import warnings
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_ta import TaTokenizer
else:
a : Dict = None
a : List[Any] = logging.get_logger(__name__)
a : List[Any] = {'vocab_file': 'spiece.model', 'tokenizer_file': 'tokenizer.json'}
a : str = {
'vocab_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/spiece.model',
't5-base': 'https://huggingface.co/t5-base/resolve/main/spiece.model',
't5-large': 'https://huggingface.co/t5-large/resolve/main/spiece.model',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/spiece.model',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/spiece.model',
},
'tokenizer_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/tokenizer.json',
't5-base': 'https://huggingface.co/t5-base/resolve/main/tokenizer.json',
't5-large': 'https://huggingface.co/t5-large/resolve/main/tokenizer.json',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/tokenizer.json',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/tokenizer.json',
},
}
# TODO(PVP) - this should be removed in Transformers v5
a : List[Any] = {
't5-small': 512,
't5-base': 512,
't5-large': 512,
't5-3b': 512,
't5-11b': 512,
}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
snake_case_ = TaTokenizer
snake_case_ = []
def __init__( self : List[Any] , lowercase_ : int=None , lowercase_ : Dict=None , lowercase_ : Dict="</s>" , lowercase_ : List[Any]="<unk>" , lowercase_ : int="<pad>" , lowercase_ : int=100 , lowercase_ : List[Any]=None , **lowercase_ : List[str] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
snake_case_ = [F"<extra_id_{i}>" for i in range(lowercase_ )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra special tokens
snake_case_ = len(set(filter(lambda lowercase_ : bool('''extra_id_''' in str(lowercase_ ) ) , lowercase_ ) ) )
if extra_tokens != extra_ids:
raise ValueError(
F"Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are"
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
super().__init__(
lowercase_ , tokenizer_file=lowercase_ , eos_token=lowercase_ , unk_token=lowercase_ , pad_token=lowercase_ , extra_ids=lowercase_ , additional_special_tokens=lowercase_ , **lowercase_ , )
snake_case_ = vocab_file
snake_case_ = False if not self.vocab_file else True
snake_case_ = extra_ids
@staticmethod
def A_ ( lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : int ):
if pretrained_model_name_or_path in TaTokenizerFast.max_model_input_sizes:
snake_case_ = TaTokenizerFast.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
F" {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this"
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
F" {pretrained_model_name_or_path} automatically truncating your input to"
F" {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences"
F" longer than {deprecated_max_model_length} you can either instantiate this tokenizer with"
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , lowercase_ , )
return max_model_length
def A_ ( self : Optional[Any] , lowercase_ : str , lowercase_ : Optional[str] = None ):
if not self.can_save_slow_tokenizer:
raise ValueError(
'''Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '''
'''tokenizer.''' )
if not os.path.isdir(lowercase_ ):
logger.error(F"Vocabulary path ({save_directory}) should be a directory" )
return
snake_case_ = 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_ ):
copyfile(self.vocab_file , lowercase_ )
logger.info(F"Copy vocab file to {out_vocab_file}" )
return (out_vocab_file,)
def A_ ( self : Optional[int] , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = token_ids_a + [self.eos_token_id]
if token_ids_a is None:
return self.prefix_tokens + token_ids_a
else:
snake_case_ = token_ids_a + [self.eos_token_id]
return self.prefix_tokens + token_ids_a + token_ids_a
def A_ ( self : int , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def A_ ( self : Dict ):
return list(
set(filter(lambda lowercase_ : bool(re.search(R'''<extra_id_\d+>''' , lowercase_ ) ) is not None , self.additional_special_tokens ) ) )
def A_ ( self : Any ):
return [self.convert_tokens_to_ids(lowercase_ ) for token in self.get_sentinel_tokens()]
| 56
| 1
|
'''simple docstring'''
import flax.linen as nn
import jax
import jax.numpy as jnp
class a ( nn.Module ):
snake_case_ = 42
snake_case_ = jnp.floataa
def A_ ( self : int ):
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Optional[int] , lowercase_ : Tuple ):
snake_case_ ,snake_case_ ,snake_case_ ,snake_case_ = hidden_states.shape
snake_case_ = jax.image.resize(
lowercase_ , shape=(batch, height * 2, width * 2, channels) , method='''nearest''' , )
snake_case_ = self.conv(lowercase_ )
return hidden_states
class a ( nn.Module ):
snake_case_ = 42
snake_case_ = jnp.floataa
def A_ ( self : str ):
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : List[Any] , lowercase_ : List[Any] ):
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
snake_case_ = self.conv(lowercase_ )
return hidden_states
class a ( nn.Module ):
snake_case_ = 42
snake_case_ = None
snake_case_ = 0.0
snake_case_ = None
snake_case_ = jnp.floataa
def A_ ( self : Optional[Any] ):
snake_case_ = self.in_channels if self.out_channels is None else self.out_channels
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Conv(
lowercase_ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
snake_case_ = nn.Dense(lowercase_ , dtype=self.dtype )
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Dropout(self.dropout_prob )
snake_case_ = nn.Conv(
lowercase_ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
snake_case_ = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
snake_case_ = None
if use_nin_shortcut:
snake_case_ = nn.Conv(
lowercase_ , kernel_size=(1, 1) , strides=(1, 1) , padding='''VALID''' , dtype=self.dtype , )
def __call__( self : Any , lowercase_ : Optional[int] , lowercase_ : Tuple , lowercase_ : Any=True ):
snake_case_ = hidden_states
snake_case_ = self.norma(lowercase_ )
snake_case_ = nn.swish(lowercase_ )
snake_case_ = self.conva(lowercase_ )
snake_case_ = self.time_emb_proj(nn.swish(lowercase_ ) )
snake_case_ = jnp.expand_dims(jnp.expand_dims(lowercase_ , 1 ) , 1 )
snake_case_ = hidden_states + temb
snake_case_ = self.norma(lowercase_ )
snake_case_ = nn.swish(lowercase_ )
snake_case_ = self.dropout(lowercase_ , lowercase_ )
snake_case_ = self.conva(lowercase_ )
if self.conv_shortcut is not None:
snake_case_ = self.conv_shortcut(lowercase_ )
return hidden_states + residual
| 56
|
'''simple docstring'''
from __future__ import annotations
import math
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
if depth < 0:
raise ValueError('''Depth cannot be less than 0''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Scores cannot be empty''' )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
return min(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ = [90, 23, 6, 33, 21, 65, 123, 3_4423]
snake_case_ = math.log(len(__UpperCAmelCase ), 2 )
print('''Optimal value : ''', end='''''' )
print(minimax(0, 0, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 56
| 1
|
'''simple docstring'''
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,
flip_channel_order,
get_resize_output_image_size,
rescale,
resize,
to_channel_dimension_format,
)
from ...image_utils import (
ChannelDimension,
ImageInput,
PILImageResampling,
make_list_of_images,
to_numpy_array,
valid_images,
)
from ...utils import TensorType, is_torch_available, is_torch_tensor, is_vision_available, logging
if is_vision_available():
import PIL
if is_torch_available():
import torch
a : List[str] = logging.get_logger(__name__)
class a ( _lowerCamelCase ):
snake_case_ = ["pixel_values"]
def __init__( self : Optional[Any] , lowercase_ : bool = True , lowercase_ : Dict[str, int] = None , lowercase_ : PILImageResampling = PILImageResampling.BILINEAR , lowercase_ : bool = True , lowercase_ : Union[int, float] = 1 / 255 , lowercase_ : bool = True , lowercase_ : Dict[str, int] = None , lowercase_ : bool = True , **lowercase_ : Union[str, Any] , ):
super().__init__(**lowercase_ )
snake_case_ = size if size is not None else {'''shortest_edge''': 224}
snake_case_ = get_size_dict(lowercase_ , default_to_square=lowercase_ )
snake_case_ = crop_size if crop_size is not None else {'''height''': 256, '''width''': 256}
snake_case_ = get_size_dict(lowercase_ , param_name='''crop_size''' )
snake_case_ = do_resize
snake_case_ = size
snake_case_ = resample
snake_case_ = do_rescale
snake_case_ = rescale_factor
snake_case_ = do_center_crop
snake_case_ = crop_size
snake_case_ = do_flip_channel_order
def A_ ( self : int , lowercase_ : np.ndarray , lowercase_ : Dict[str, int] , lowercase_ : PILImageResampling = PIL.Image.BILINEAR , lowercase_ : Optional[Union[str, ChannelDimension]] = None , **lowercase_ : Optional[int] , ):
snake_case_ = get_size_dict(lowercase_ , default_to_square=lowercase_ )
if "shortest_edge" not in size:
raise ValueError(F"The `size` dictionary must contain the key `shortest_edge`. Got {size.keys()}" )
snake_case_ = 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 : List[str] , lowercase_ : np.ndarray , lowercase_ : Dict[str, int] , lowercase_ : Optional[Union[str, ChannelDimension]] = None , **lowercase_ : Tuple , ):
snake_case_ = get_size_dict(lowercase_ )
if "height" not in size or "width" not in size:
raise ValueError(F"The `size` dictionary 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 : Optional[Any] , lowercase_ : np.ndarray , lowercase_ : Union[int, float] , lowercase_ : Optional[Union[str, ChannelDimension]] = None , **lowercase_ : int , ):
return rescale(lowercase_ , scale=lowercase_ , data_format=lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : np.ndarray , lowercase_ : Optional[Union[str, ChannelDimension]] = None ):
return flip_channel_order(lowercase_ , data_format=lowercase_ )
def A_ ( self : Dict , lowercase_ : ImageInput , lowercase_ : bool = None , lowercase_ : Dict[str, int] = None , lowercase_ : PILImageResampling = None , lowercase_ : bool = None , lowercase_ : float = None , lowercase_ : bool = None , lowercase_ : Dict[str, int] = None , lowercase_ : bool = None , lowercase_ : Optional[Union[str, TensorType]] = None , lowercase_ : ChannelDimension = ChannelDimension.FIRST , **lowercase_ : Dict , ):
snake_case_ = do_resize if do_resize is not None else self.do_resize
snake_case_ = resample if resample is not None else self.resample
snake_case_ = do_rescale if do_rescale is not None else self.do_rescale
snake_case_ = rescale_factor if rescale_factor is not None else self.rescale_factor
snake_case_ = do_center_crop if do_center_crop is not None else self.do_center_crop
snake_case_ = (
do_flip_channel_order if do_flip_channel_order is not None else self.do_flip_channel_order
)
snake_case_ = size if size is not None else self.size
snake_case_ = get_size_dict(lowercase_ , default_to_square=lowercase_ )
snake_case_ = crop_size if crop_size is not None else self.crop_size
snake_case_ = get_size_dict(lowercase_ , param_name='''crop_size''' )
snake_case_ = 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_rescale and rescale_factor is None:
raise ValueError('''Rescale factor must be specified if do_rescale is True.''' )
if do_center_crop and crop_size is None:
raise ValueError('''Crop size must be specified if do_center_crop is True.''' )
# All transformations expect numpy arrays.
snake_case_ = [to_numpy_array(lowercase_ ) for image in images]
if do_resize:
snake_case_ = [self.resize(image=lowercase_ , size=lowercase_ , resample=lowercase_ ) for image in images]
if do_center_crop:
snake_case_ = [self.center_crop(image=lowercase_ , size=lowercase_ ) for image in images]
if do_rescale:
snake_case_ = [self.rescale(image=lowercase_ , scale=lowercase_ ) for image in images]
# the pretrained checkpoints assume images are BGR, not RGB
if do_flip_channel_order:
snake_case_ = [self.flip_channel_order(image=lowercase_ ) for image in images]
snake_case_ = [to_channel_dimension_format(lowercase_ , lowercase_ ) for image in images]
snake_case_ = {'''pixel_values''': images}
return BatchFeature(data=lowercase_ , tensor_type=lowercase_ )
def A_ ( self : Optional[int] , lowercase_ : Optional[int] , lowercase_ : List[Tuple] = None ):
snake_case_ = 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_ ):
snake_case_ = target_sizes.numpy()
snake_case_ = []
for idx in range(len(lowercase_ ) ):
snake_case_ = torch.nn.functional.interpolate(
logits[idx].unsqueeze(dim=0 ) , size=target_sizes[idx] , mode='''bilinear''' , align_corners=lowercase_ )
snake_case_ = resized_logits[0].argmax(dim=0 )
semantic_segmentation.append(lowercase_ )
else:
snake_case_ = logits.argmax(dim=1 )
snake_case_ = [semantic_segmentation[i] for i in range(semantic_segmentation.shape[0] )]
return semantic_segmentation
| 56
|
'''simple docstring'''
import argparse
import torch
from torch import nn
from transformers import SpeechaTextConfig, SpeechaTextForConditionalGeneration
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = [
'''encoder.version''',
'''decoder.version''',
'''model.encoder.version''',
'''model.decoder.version''',
'''decoder.output_projection.weight''',
'''_float_tensor''',
'''encoder.embed_positions._float_tensor''',
'''decoder.embed_positions._float_tensor''',
]
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = list(s_dict.keys() )
for key in keys:
if "transformer_layers" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
elif "subsample" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ ,snake_case_ = emb.weight.shape
snake_case_ = nn.Linear(__UpperCAmelCase, __UpperCAmelCase, bias=__UpperCAmelCase )
snake_case_ = emb.weight.data
return lin_layer
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )
snake_case_ = mam_aaa['''args''']
snake_case_ = mam_aaa['''model''']
snake_case_ = state_dict['''decoder.output_projection.weight''']
remove_ignore_keys_(__UpperCAmelCase )
rename_keys(__UpperCAmelCase )
snake_case_ = state_dict['''decoder.embed_tokens.weight'''].shape[0]
snake_case_ = args.share_decoder_input_output_embed
snake_case_ = [int(__UpperCAmelCase ) for i in args.conv_kernel_sizes.split(''',''' )]
snake_case_ = SpeechaTextConfig(
vocab_size=__UpperCAmelCase, max_source_positions=args.max_source_positions, max_target_positions=args.max_target_positions, encoder_layers=args.encoder_layers, decoder_layers=args.decoder_layers, encoder_attention_heads=args.encoder_attention_heads, decoder_attention_heads=args.decoder_attention_heads, encoder_ffn_dim=args.encoder_ffn_embed_dim, decoder_ffn_dim=args.decoder_ffn_embed_dim, d_model=args.encoder_embed_dim, dropout=args.dropout, attention_dropout=args.attention_dropout, activation_dropout=args.activation_dropout, activation_function='''relu''', num_conv_layers=len(__UpperCAmelCase ), conv_channels=args.conv_channels, conv_kernel_sizes=__UpperCAmelCase, input_feat_per_channel=args.input_feat_per_channel, input_channels=args.input_channels, tie_word_embeddings=__UpperCAmelCase, num_beams=5, max_length=200, use_cache=__UpperCAmelCase, decoder_start_token_id=2, early_stopping=__UpperCAmelCase, )
snake_case_ = SpeechaTextForConditionalGeneration(__UpperCAmelCase )
snake_case_ ,snake_case_ = model.model.load_state_dict(__UpperCAmelCase, strict=__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0 and not set(__UpperCAmelCase ) <= {
"encoder.embed_positions.weights",
"decoder.embed_positions.weights",
}:
raise ValueError(
'''Only `encoder.embed_positions.weights` and `decoder.embed_positions.weights` are allowed to be missing,'''
F" but all the following weights are missing {missing}" )
if tie_embeds:
snake_case_ = make_linear_from_emb(model.model.decoder.embed_tokens )
else:
snake_case_ = lm_head_weights
model.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
a : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument('--fairseq_path', type=str, help='Path to the fairseq model (.pt) file.')
parser.add_argument('--pytorch_dump_folder_path', default=None, type=str, help='Path to the output PyTorch model.')
a : List[Any] = parser.parse_args()
convert_fairseq_sat_checkpoint_to_tfms(args.fairseq_path, args.pytorch_dump_folder_path)
| 56
| 1
|
'''simple docstring'''
import functools
import operator
from ...configuration_utils import PretrainedConfig
from ...utils import logging
a : Union[str, Any] = logging.get_logger(__name__)
a : Optional[int] = {
'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 ( _lowerCamelCase ):
snake_case_ = "unispeech"
def __init__( self : int , lowercase_ : Any=32 , lowercase_ : Optional[int]=768 , lowercase_ : Union[str, Any]=12 , lowercase_ : Optional[int]=12 , lowercase_ : List[Any]=3072 , lowercase_ : Tuple="gelu" , lowercase_ : Tuple=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Tuple=0.1 , lowercase_ : List[str]=0.0 , lowercase_ : List[str]=0.0 , lowercase_ : Tuple=0.1 , lowercase_ : Tuple=0.1 , lowercase_ : List[str]=0.02 , lowercase_ : str=1e-5 , lowercase_ : Optional[int]="group" , lowercase_ : Union[str, Any]="gelu" , lowercase_ : List[Any]=(512, 512, 512, 512, 512, 512, 512) , lowercase_ : List[str]=(5, 2, 2, 2, 2, 2, 2) , lowercase_ : Dict=(10, 3, 3, 3, 3, 2, 2) , lowercase_ : Optional[Any]=False , lowercase_ : Optional[Any]=128 , lowercase_ : Any=16 , lowercase_ : Optional[int]=False , lowercase_ : List[str]=True , lowercase_ : Optional[int]=0.05 , lowercase_ : Dict=10 , lowercase_ : int=2 , lowercase_ : Optional[int]=0.0 , lowercase_ : Optional[Any]=10 , lowercase_ : Dict=0 , lowercase_ : str=320 , lowercase_ : Dict=2 , lowercase_ : Tuple=0.1 , lowercase_ : Tuple=100 , lowercase_ : Tuple=256 , lowercase_ : str=256 , lowercase_ : Union[str, Any]=0.1 , lowercase_ : Any="mean" , lowercase_ : Union[str, Any]=False , lowercase_ : Tuple=False , lowercase_ : Tuple=256 , lowercase_ : Tuple=80 , lowercase_ : int=0 , lowercase_ : List[str]=1 , lowercase_ : Optional[int]=2 , lowercase_ : Optional[Any]=0.5 , **lowercase_ : Dict , ):
super().__init__(**lowercase_ , pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ )
snake_case_ = hidden_size
snake_case_ = feat_extract_norm
snake_case_ = feat_extract_activation
snake_case_ = list(lowercase_ )
snake_case_ = list(lowercase_ )
snake_case_ = list(lowercase_ )
snake_case_ = conv_bias
snake_case_ = num_conv_pos_embeddings
snake_case_ = num_conv_pos_embedding_groups
snake_case_ = len(self.conv_dim )
snake_case_ = num_hidden_layers
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = num_attention_heads
snake_case_ = hidden_dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = feat_proj_dropout
snake_case_ = final_dropout
snake_case_ = layerdrop
snake_case_ = layer_norm_eps
snake_case_ = initializer_range
snake_case_ = num_ctc_classes
snake_case_ = vocab_size
snake_case_ = do_stable_layer_norm
snake_case_ = use_weighted_layer_sum
snake_case_ = 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
snake_case_ = apply_spec_augment
snake_case_ = mask_time_prob
snake_case_ = mask_time_length
snake_case_ = mask_time_min_masks
snake_case_ = mask_feature_prob
snake_case_ = mask_feature_length
snake_case_ = mask_feature_min_masks
# parameters for pretraining with codevector quantized representations
snake_case_ = num_codevectors_per_group
snake_case_ = num_codevector_groups
snake_case_ = contrastive_logits_temperature
snake_case_ = feat_quantizer_dropout
snake_case_ = num_negatives
snake_case_ = codevector_dim
snake_case_ = proj_codevector_dim
snake_case_ = diversity_loss_weight
# ctc loss
snake_case_ = ctc_loss_reduction
snake_case_ = ctc_zero_infinity
# pretraining loss
snake_case_ = replace_prob
@property
def A_ ( self : str ):
return functools.reduce(operator.mul , self.conv_stride , 1 )
| 56
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class a ( metaclass=_lowerCamelCase ):
snake_case_ = ["transformers", "torch", "note_seq"]
def __init__( self : Union[str, Any] , *lowercase_ : Optional[int] , **lowercase_ : int ):
requires_backends(self , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : List[Any] , *lowercase_ : Optional[Any] , **lowercase_ : str ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : Tuple , *lowercase_ : Union[str, Any] , **lowercase_ : List[Any] ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
| 56
| 1
|
'''simple docstring'''
from maths.prime_factors import prime_factors
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
if not isinstance(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ = F"Input value of [number={number}] must be an integer"
raise TypeError(__UpperCAmelCase )
if number < 1:
raise ValueError('''Input must be a positive integer''' )
return -1 if len(prime_factors(__UpperCAmelCase ) ) % 2 else 1
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
# tests directory-specific settings - this file is run automatically
# by pytest before any tests are run
import doctest
import sys
import warnings
from os.path import abspath, dirname, join
import _pytest
from transformers.testing_utils import HfDoctestModule, HfDocTestParser
# allow having multiple repository checkouts and not needing to remember to rerun
# 'pip install -e .[dev]' when switching between checkouts and running tests.
a : int = abspath(join(dirname(__file__), 'src'))
sys.path.insert(1, git_repo_path)
# silence FutureWarning warnings in tests since often we can't act on them until
# they become normal warnings - i.e. the tests still need to test the current functionality
warnings.simplefilter(action='ignore', category=FutureWarning)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
config.addinivalue_line(
'''markers''', '''is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested''' )
config.addinivalue_line(
'''markers''', '''is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested''' )
config.addinivalue_line('''markers''', '''is_pipeline_test: mark test to run only when pipelines are tested''' )
config.addinivalue_line('''markers''', '''is_staging_test: mark test to run only in the staging environment''' )
config.addinivalue_line('''markers''', '''accelerate_tests: mark test that require accelerate''' )
config.addinivalue_line('''markers''', '''tool_tests: mark the tool tests that are run on their specific schedule''' )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from transformers.testing_utils import pytest_addoption_shared
pytest_addoption_shared(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
from transformers.testing_utils import pytest_terminal_summary_main
snake_case_ = terminalreporter.config.getoption('''--make-reports''' )
if make_reports:
pytest_terminal_summary_main(__UpperCAmelCase, id=__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if exitstatus == 5:
snake_case_ = 0
# Doctest custom flag to ignore output.
a : Union[str, Any] = doctest.register_optionflag('IGNORE_RESULT')
a : Optional[int] = doctest.OutputChecker
class a ( _lowerCamelCase ):
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int] ):
if IGNORE_RESULT & optionflags:
return True
return OutputChecker.check_output(self , lowercase_ , lowercase_ , lowercase_ )
a : List[Any] = CustomOutputChecker
a : Optional[int] = HfDoctestModule
a : Tuple = HfDocTestParser
| 56
| 1
|
'''simple docstring'''
import argparse
import dataclasses
import json
import logging
import os
import shutil
from typing import List, Optional
import datasets
from accelerate import Accelerator
from datasets import load_dataset
from finetuning import finetune
from tqdm.auto import tqdm
import transformers
from transformers import AutoConfig, set_seed
from transformers.trainer_utils import IntervalStrategy
a : List[str] = logging.getLogger(__name__)
a : int = 'pytorch_model.bin'
@dataclasses.dataclass
class a :
snake_case_ = dataclasses.field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models."} )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co."} , )
@dataclasses.dataclass
class a :
snake_case_ = dataclasses.field(metadata={"help": "A csv or a json file containing the training data."} )
snake_case_ = dataclasses.field(metadata={"help": "A csv or a json file containing the data to predict on."} )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "A csv or a json file containing the validation data."} )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "The name of the task to train on."} , )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "The list of labels for the task."} )
@dataclasses.dataclass
class a :
snake_case_ = dataclasses.field(
metadata={"help": "The output directory where the model predictions and checkpoints will be written."} )
snake_case_ = dataclasses.field(
default="accuracy" , metadata={"help": "The evaluation metric used for the task."} )
snake_case_ = dataclasses.field(
default="no" , metadata={
"help": "The evaluation strategy to adopt during training. Possible values are: [\"no\", \"step\", \"epoch]"
} , )
snake_case_ = dataclasses.field(
default=10 , metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} , )
snake_case_ = dataclasses.field(
default=0.0 , metadata={
"help": "How much the specified evaluation metric must improve to satisfy early stopping conditions."
} , )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "Whether to filter the pseudo-labeled data based on the confidence score."} , )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "Whether to filter the pseudo-labeled data based on the validation performance."} , )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "Whether to fine-tune on labeled data after pseudo training."} , )
snake_case_ = dataclasses.field(
default=0.0 , metadata={"help": "Confidence threshold for pseudo-labeled data filtering."} , )
snake_case_ = dataclasses.field(
default=100 , metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} , )
snake_case_ = dataclasses.field(
default=_lowerCamelCase , metadata={"help": "Random seed for initialization."} , )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = datasets.concatenate_datasets([infer_input, infer_output], axis=1 )
if args.do_filter_by_confidence:
snake_case_ = dataset.filter(lambda __UpperCAmelCase : example["probability"] > args.confidence_threshold )
if args.do_filter_by_val_performance:
assert eval_result >= 0.0 and eval_result <= 1.0
snake_case_ = int(eval_result * len(__UpperCAmelCase ) )
print(__UpperCAmelCase )
snake_case_ = dataset.sort('''probability''', reverse=__UpperCAmelCase )
snake_case_ = dataset.select(range(__UpperCAmelCase ) )
snake_case_ = dataset.remove_columns(['''label''', '''probability'''] )
snake_case_ = dataset.rename_column('''prediction''', '''label''' )
snake_case_ = dataset.map(lambda __UpperCAmelCase : {"label": idalabel[example["label"]]} )
snake_case_ = dataset.shuffle(seed=args.seed )
snake_case_ = os.path.join(__UpperCAmelCase, F"train_pseudo.{args.data_file_extension}" )
if args.data_file_extension == "csv":
dataset.to_csv(__UpperCAmelCase, index=__UpperCAmelCase )
else:
dataset.to_json(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, **__UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format='''%(asctime)s - %(levelname)s - %(name)s - %(message)s''', datefmt='''%m/%d/%Y %H:%M:%S''', level=logging.INFO, )
logger.info(accelerator.state )
# Setup logging, we only want one process per machine to log things on the
# screen. accelerator.is_local_main_process is only True for one process per
# machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
snake_case_ = STModelArguments(model_name_or_path=__UpperCAmelCase )
snake_case_ = STDataArguments(train_file=__UpperCAmelCase, infer_file=__UpperCAmelCase )
snake_case_ = STTrainingArguments(output_dir=__UpperCAmelCase )
snake_case_ = argparse.Namespace()
for arg_class in (model_args, data_args, training_args):
for key, value in vars(__UpperCAmelCase ).items():
setattr(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
for key, value in kwargs.items():
if hasattr(__UpperCAmelCase, __UpperCAmelCase ):
setattr(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
# Sanity checks
snake_case_ = {}
snake_case_ = None
# You need to provide the training data and the data to predict on
assert args.train_file is not None
assert args.infer_file is not None
snake_case_ = args.train_file
snake_case_ = args.infer_file
if args.evaluation_strategy != IntervalStrategy.NO.value:
assert args.eval_file is not None
snake_case_ = args.eval_file
for key in data_files:
snake_case_ = data_files[key].split('''.''' )[-1]
assert extension in ["csv", "json"], F"`{key}_file` should be a csv or a json file."
if args.data_file_extension is None:
snake_case_ = extension
else:
assert extension == args.data_file_extension, F"`{key}_file` should be a {args.data_file_extension} file`."
assert (
args.eval_metric in datasets.list_metrics()
), F"{args.eval_metric} not in the list of supported metrics {datasets.list_metrics()}."
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed )
logger.info('''Creating the initial data directory for self-training...''' )
snake_case_ = F"{args.output_dir}/self-train_iter-{{}}".format
snake_case_ = data_dir_format(0 )
if accelerator.is_main_process:
if args.output_dir is not None:
os.makedirs(args.output_dir, exist_ok=__UpperCAmelCase )
os.makedirs(__UpperCAmelCase, exist_ok=__UpperCAmelCase )
accelerator.wait_for_everyone()
snake_case_ = None
snake_case_ = None
snake_case_ = 0
snake_case_ = False
# Show the progress bar
snake_case_ = tqdm(range(args.max_selftrain_iterations ), disable=not accelerator.is_local_main_process )
# Self-train
for iteration in range(0, int(args.max_selftrain_iterations ) ):
snake_case_ = data_dir_format(__UpperCAmelCase )
assert os.path.exists(__UpperCAmelCase )
# Stage 1: initial fine-tuning for iteration = 0 or pseudo-training for
# iteration > 0
snake_case_ = os.path.join(__UpperCAmelCase, '''stage-1''' )
snake_case_ = {
'''accelerator''': accelerator,
'''model_name_or_path''': args.model_name_or_path,
'''cache_dir''': args.cache_dir,
'''do_train''': True,
'''train_file''': data_files['''train'''] if iteration == 0 else data_files['''train_pseudo'''],
'''do_eval''': True if args.eval_file is not None else False,
'''eval_file''': data_files['''eval'''],
'''do_predict''': True,
'''infer_file''': data_files['''infer'''],
'''task_name''': args.task_name,
'''label_list''': args.label_list,
'''output_dir''': current_output_dir,
'''eval_metric''': args.eval_metric,
'''evaluation_strategy''': args.evaluation_strategy,
'''early_stopping_patience''': args.early_stopping_patience,
'''early_stopping_threshold''': args.early_stopping_threshold,
'''seed''': args.seed,
}
# Add additional training arguments
for key, value in kwargs.items():
if key not in arguments_dict and not hasattr(__UpperCAmelCase, __UpperCAmelCase ):
arguments_dict.update({key: value} )
snake_case_ = os.path.join(__UpperCAmelCase, '''best-checkpoint''', __UpperCAmelCase )
if os.path.exists(__UpperCAmelCase ):
logger.info(
'''Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 1.''', __UpperCAmelCase, __UpperCAmelCase, )
else:
logger.info('''***** Running self-training: iteration: %d, stage: 1 *****''', __UpperCAmelCase )
finetune(**__UpperCAmelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__UpperCAmelCase )
logger.info('''Self-training job completed: iteration: %d, stage: 1.''', __UpperCAmelCase )
if iteration > 0 and args.finetune_on_labeled_data:
# Stage 2 (optional): fine-tuning on the original labeled data
snake_case_ = os.path.join(__UpperCAmelCase, '''best-checkpoint''' )
snake_case_ = os.path.join(__UpperCAmelCase, '''stage-2''' )
# Update arguments_dict
snake_case_ = model_path
snake_case_ = data_files['''train''']
snake_case_ = current_output_dir
snake_case_ = os.path.join(__UpperCAmelCase, '''best-checkpoint''', __UpperCAmelCase )
if os.path.exists(__UpperCAmelCase ):
logger.info(
'''Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 2.''', __UpperCAmelCase, __UpperCAmelCase, )
else:
logger.info('''***** Running self-training: iteration: %d, stage: 2 *****''', __UpperCAmelCase )
finetune(**__UpperCAmelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__UpperCAmelCase )
logger.info('''Self-training job completed: iteration: %d, stage: 2.''', __UpperCAmelCase )
snake_case_ = iteration
snake_case_ = data_dir_format(iteration + 1 )
snake_case_ = AutoConfig.from_pretrained(os.path.join(__UpperCAmelCase, '''best-checkpoint''' ) )
snake_case_ = config.idalabel
snake_case_ = os.path.join(__UpperCAmelCase, '''eval_results_best-checkpoint.json''' )
snake_case_ = os.path.join(__UpperCAmelCase, '''test_results_best-checkpoint.json''' )
assert os.path.exists(__UpperCAmelCase )
with open(__UpperCAmelCase, '''r''' ) as f:
snake_case_ = float(json.load(__UpperCAmelCase )[args.eval_metric] )
snake_case_ = os.path.join(__UpperCAmelCase, '''infer_output_best-checkpoint.csv''' )
assert os.path.exists(__UpperCAmelCase )
# Loading the dataset from local csv or json files.
snake_case_ = load_dataset(args.data_file_extension, data_files={'''data''': data_files['''infer''']} )['''data''']
snake_case_ = load_dataset('''csv''', data_files={'''data''': infer_output_file} )['''data''']
if accelerator.is_main_process:
os.makedirs(__UpperCAmelCase, exist_ok=__UpperCAmelCase )
shutil.copy(__UpperCAmelCase, os.path.join(__UpperCAmelCase, F"eval_results_iter-{iteration}.json" ) )
if os.path.exists(__UpperCAmelCase ):
shutil.copy(__UpperCAmelCase, os.path.join(__UpperCAmelCase, F"test_results_iter-{iteration}.json" ) )
create_pseudo_labeled_data(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
accelerator.wait_for_everyone()
snake_case_ = os.path.join(__UpperCAmelCase, F"train_pseudo.{args.data_file_extension}" )
if args.evaluation_strategy != IntervalStrategy.NO.value:
snake_case_ = eval_result
if best_iteration is None:
snake_case_ = new_iteration
snake_case_ = new_eval_result
else:
if new_eval_result - best_eval_result > args.early_stopping_threshold:
snake_case_ = new_iteration
snake_case_ = new_eval_result
snake_case_ = 0
else:
if new_eval_result == best_eval_result:
snake_case_ = new_iteration
snake_case_ = new_eval_result
early_stopping_patience_counter += 1
if early_stopping_patience_counter >= args.early_stopping_patience:
snake_case_ = True
progress_bar.update(1 )
if should_training_stop:
break
if best_iteration is not None:
# Save the best iteration
logger.info('''Best iteration: %d''', __UpperCAmelCase )
logger.info('''Best evaluation result: %s = %f''', args.eval_metric, __UpperCAmelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__UpperCAmelCase, F"eval_results_iter-{iteration}.json" ), os.path.join(__UpperCAmelCase, '''eval_results_best-iteration.json''' ), )
else:
# Assume that the last iteration is the best
logger.info('''Best iteration: %d''', args.max_selftrain_iterations - 1 )
logger.info('''Best evaluation result: %s = %f''', args.eval_metric, __UpperCAmelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__UpperCAmelCase, F"eval_results_iter-{args.max_selftrain_iterations - 1}.json" ), os.path.join(__UpperCAmelCase, '''eval_results_best-iteration.json''' ), )
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import TensorType, is_torch_available, logging
a : Dict = logging.get_logger(__name__)
a : List[str] = {
'Helsinki-NLP/opus-mt-en-de': 'https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/config.json',
# See all Marian models at https://huggingface.co/models?filter=marian
}
class a ( _lowerCamelCase ):
snake_case_ = "marian"
snake_case_ = ["past_key_values"]
snake_case_ = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self : List[Any] , lowercase_ : Optional[Any]=5_8101 , lowercase_ : Dict=None , lowercase_ : List[str]=1024 , lowercase_ : Optional[Any]=12 , lowercase_ : int=4096 , lowercase_ : Any=16 , lowercase_ : Optional[int]=12 , lowercase_ : str=4096 , lowercase_ : Union[str, Any]=16 , lowercase_ : Dict=0.0 , lowercase_ : Union[str, Any]=0.0 , lowercase_ : Optional[Any]=True , lowercase_ : Union[str, Any]=True , lowercase_ : int="gelu" , lowercase_ : Dict=1024 , lowercase_ : int=0.1 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.02 , lowercase_ : int=5_8100 , lowercase_ : Optional[Any]=False , lowercase_ : Any=5_8100 , lowercase_ : Optional[int]=0 , lowercase_ : Dict=0 , lowercase_ : List[str]=True , **lowercase_ : Any , ):
snake_case_ = vocab_size
snake_case_ = decoder_vocab_size or vocab_size
snake_case_ = max_position_embeddings
snake_case_ = d_model
snake_case_ = encoder_ffn_dim
snake_case_ = encoder_layers
snake_case_ = encoder_attention_heads
snake_case_ = decoder_ffn_dim
snake_case_ = decoder_layers
snake_case_ = decoder_attention_heads
snake_case_ = dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = activation_function
snake_case_ = init_std
snake_case_ = encoder_layerdrop
snake_case_ = decoder_layerdrop
snake_case_ = use_cache
snake_case_ = encoder_layers
snake_case_ = scale_embedding # scale factor will be sqrt(d_model) if True
snake_case_ = share_encoder_decoder_embeddings
super().__init__(
pad_token_id=lowercase_ , eos_token_id=lowercase_ , is_encoder_decoder=lowercase_ , decoder_start_token_id=lowercase_ , forced_eos_token_id=lowercase_ , **lowercase_ , )
class a ( _lowerCamelCase ):
@property
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.inputs
def A_ ( self : Union[str, Any] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ = {0: '''batch'''}
snake_case_ = {0: '''batch''', 1: '''past_decoder_sequence + sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''decoder_sequence'''}
snake_case_ = {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.
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
else:
snake_case_ = 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
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.outputs
def A_ ( self : Dict ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super().outputs
else:
snake_case_ = super(lowercase_ , self ).outputs
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
return common_outputs
def A_ ( self : Dict , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
# Generate decoder inputs
snake_case_ = seq_length if not self.use_past else 1
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
snake_case_ = {F"decoder_{name}": tensor for name, tensor in decoder_inputs.items()}
snake_case_ = 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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
snake_case_ = common_inputs['''decoder_input_ids'''].shape[1]
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = decoder_seq_length + 3
snake_case_ = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
snake_case_ = torch.cat(
[common_inputs['''decoder_attention_mask'''], torch.ones(lowercase_ , lowercase_ )] , dim=1 )
snake_case_ = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
snake_case_ ,snake_case_ = self.num_layers
snake_case_ = min(lowercase_ , lowercase_ )
snake_case_ = max(lowercase_ , lowercase_ ) - min_num_layers
snake_case_ = '''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.
snake_case_ = 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 : Union[str, Any] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
# Not using the same length for past_key_values
snake_case_ = seqlen + 2
snake_case_ ,snake_case_ = self.num_layers
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = common_inputs['''attention_mask'''].dtype
snake_case_ = torch.cat(
[common_inputs['''attention_mask'''], torch.ones(lowercase_ , lowercase_ , dtype=lowercase_ )] , dim=1 )
snake_case_ = [
(torch.zeros(lowercase_ ), torch.zeros(lowercase_ )) for _ in range(lowercase_ )
]
return common_inputs
def A_ ( self : List[str] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
# 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
snake_case_ = 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
snake_case_ = tokenizer.num_special_tokens_to_add(lowercase_ )
snake_case_ = 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
snake_case_ = [''' '''.join([tokenizer.unk_token] ) * seq_length] * batch_size
snake_case_ = dict(tokenizer(lowercase_ , return_tensors=lowercase_ ) )
return common_inputs
def A_ ( self : Any , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
else:
snake_case_ = self._generate_dummy_inputs_for_causal_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
return common_inputs
def A_ ( self : Dict , lowercase_ : List[str] , lowercase_ : List[str] , lowercase_ : int , lowercase_ : List[str] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super()._flatten_past_key_values_(lowercase_ , lowercase_ , lowercase_ , lowercase_ )
else:
snake_case_ = super(lowercase_ , self )._flatten_past_key_values_(
lowercase_ , lowercase_ , lowercase_ , lowercase_ )
@property
def A_ ( self : List[str] ):
return 1e-4
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
if not isinstance(__UpperCAmelCase, __UpperCAmelCase ):
raise ValueError('''iterations must be defined as integers''' )
if not isinstance(__UpperCAmelCase, __UpperCAmelCase ) or not number >= 1:
raise ValueError(
'''starting number must be
and integer and be more than 0''' )
if not iterations >= 1:
raise ValueError('''Iterations must be done more than 0 times to play FizzBuzz''' )
snake_case_ = ''''''
while number <= iterations:
if number % 3 == 0:
out += "Fizz"
if number % 5 == 0:
out += "Buzz"
if 0 not in (number % 3, number % 5):
out += str(__UpperCAmelCase )
# print(out)
number += 1
out += " "
return out
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import AutoencoderKL, CycleDiffusionPipeline, DDIMScheduler, UNetaDConditionModel
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, skip_mps
from ..pipeline_params import (
IMAGE_TO_IMAGE_IMAGE_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = CycleDiffusionPipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {
"negative_prompt",
"height",
"width",
"negative_prompt_embeds",
}
snake_case_ = PipelineTesterMixin.required_optional_params - {"latents"}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"source_prompt"} )
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
def A_ ( self : Tuple ):
torch.manual_seed(0 )
snake_case_ = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=32 , )
snake_case_ = DDIMScheduler(
beta_start=0.0_0085 , beta_end=0.012 , beta_schedule='''scaled_linear''' , num_train_timesteps=1000 , clip_sample=lowercase_ , set_alpha_to_one=lowercase_ , )
torch.manual_seed(0 )
snake_case_ = 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 )
snake_case_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , )
snake_case_ = CLIPTextModel(lowercase_ )
snake_case_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
snake_case_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
'''safety_checker''': None,
'''feature_extractor''': None,
}
return components
def A_ ( self : Any , lowercase_ : int , lowercase_ : Optional[Any]=0 ):
snake_case_ = floats_tensor((1, 3, 32, 32) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
snake_case_ = image / 2 + 0.5
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''prompt''': '''An astronaut riding an elephant''',
'''source_prompt''': '''An astronaut riding a horse''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''eta''': 0.1,
'''strength''': 0.8,
'''guidance_scale''': 3,
'''source_guidance_scale''': 1,
'''output_type''': '''numpy''',
}
return inputs
def A_ ( self : Union[str, Any] ):
snake_case_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
snake_case_ = self.get_dummy_components()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.4459, 0.4943, 0.4544, 0.6643, 0.5474, 0.4327, 0.5701, 0.5959, 0.5179] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.get_dummy_components()
for name, module in components.items():
if hasattr(lowercase_ , '''half''' ):
snake_case_ = module.half()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.3506, 0.4543, 0.446, 0.4575, 0.5195, 0.4155, 0.5273, 0.518, 0.4116] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@skip_mps
def A_ ( self : Optional[int] ):
return super().test_save_load_local()
@unittest.skip('''non-deterministic pipeline''' )
def A_ ( self : List[Any] ):
return super().test_inference_batch_single_identical()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_dict_tuple_outputs_equivalent()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_save_load_optional_components()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_attention_slicing_forward_pass()
@slow
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : List[Any] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Union[str, Any] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car_fp16.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(
lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , torch_dtype=torch.floataa , revision='''fp16''' )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
# the values aren't exactly equal, but the images look the same visually
assert np.abs(image - expected_image ).max() < 5e-1
def A_ ( self : List[str] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
assert np.abs(image - expected_image ).max() < 2e-2
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
return len(set(__UpperCAmelCase ) ) == len(__UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : str = logging.get_logger(__name__)
a : str = {
'google/bigbird-roberta-base': 'https://huggingface.co/google/bigbird-roberta-base/resolve/main/config.json',
'google/bigbird-roberta-large': 'https://huggingface.co/google/bigbird-roberta-large/resolve/main/config.json',
'google/bigbird-base-trivia-itc': 'https://huggingface.co/google/bigbird-base-trivia-itc/resolve/main/config.json',
# See all BigBird models at https://huggingface.co/models?filter=big_bird
}
class a ( _lowerCamelCase ):
snake_case_ = "big_bird"
def __init__( self : Union[str, Any] , lowercase_ : List[Any]=5_0358 , lowercase_ : Tuple=768 , lowercase_ : Dict=12 , lowercase_ : str=12 , lowercase_ : Tuple=3072 , lowercase_ : Any="gelu_new" , lowercase_ : Optional[Any]=0.1 , lowercase_ : List[Any]=0.1 , lowercase_ : List[Any]=4096 , lowercase_ : List[Any]=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[int]=1e-12 , lowercase_ : Tuple=True , lowercase_ : Tuple=0 , lowercase_ : str=1 , lowercase_ : Union[str, Any]=2 , lowercase_ : Optional[Any]=66 , lowercase_ : Optional[int]="block_sparse" , lowercase_ : Any=True , lowercase_ : List[str]=False , lowercase_ : Any=64 , lowercase_ : Tuple=3 , lowercase_ : Tuple=None , **lowercase_ : Tuple , ):
super().__init__(
pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , sep_token_id=lowercase_ , **lowercase_ , )
snake_case_ = vocab_size
snake_case_ = max_position_embeddings
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = type_vocab_size
snake_case_ = layer_norm_eps
snake_case_ = use_cache
snake_case_ = rescale_embeddings
snake_case_ = attention_type
snake_case_ = use_bias
snake_case_ = block_size
snake_case_ = num_random_blocks
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : str ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
| 1
|
'''simple docstring'''
import inspect
from typing import Callable, List, Optional, Union
import torch
from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
from diffusers import DiffusionPipeline
from diffusers.models import AutoencoderKL, UNetaDConditionModel
from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput
from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
from diffusers.schedulers import DDIMScheduler, LMSDiscreteScheduler, PNDMScheduler
from diffusers.utils import logging
a : str = logging.get_logger(__name__) # pylint: disable=invalid-name
class a ( _lowerCamelCase ):
def __init__( self : Dict , lowercase_ : AutoencoderKL , lowercase_ : CLIPTextModel , lowercase_ : CLIPTokenizer , lowercase_ : UNetaDConditionModel , lowercase_ : Union[DDIMScheduler, PNDMScheduler, LMSDiscreteScheduler] , lowercase_ : StableDiffusionSafetyChecker , lowercase_ : CLIPImageProcessor , ):
super().__init__()
self.register_modules(
vae=lowercase_ , text_encoder=lowercase_ , tokenizer=lowercase_ , unet=lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , feature_extractor=lowercase_ , )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Union[str, int]] = "auto" ):
if slice_size == "auto":
# half the attention head size is usually a good trade-off between
# speed and memory
snake_case_ = self.unet.config.attention_head_dim // 2
self.unet.set_attention_slice(lowercase_ )
def A_ ( self : List[str] ):
self.enable_attention_slicing(lowercase_ )
@torch.no_grad()
def __call__( self : int , lowercase_ : Union[str, List[str]] , lowercase_ : int = 512 , lowercase_ : int = 512 , lowercase_ : int = 50 , lowercase_ : float = 7.5 , lowercase_ : Optional[Union[str, List[str]]] = None , lowercase_ : Optional[int] = 1 , lowercase_ : float = 0.0 , lowercase_ : Optional[torch.Generator] = None , lowercase_ : Optional[torch.FloatTensor] = None , lowercase_ : Optional[str] = "pil" , lowercase_ : bool = True , lowercase_ : Optional[Callable[[int, int, torch.FloatTensor], None]] = None , lowercase_ : int = 1 , lowercase_ : Optional[torch.FloatTensor] = None , **lowercase_ : Optional[int] , ):
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = 1
elif isinstance(lowercase_ , lowercase_ ):
snake_case_ = len(lowercase_ )
else:
raise ValueError(F"`prompt` has to be of type `str` or `list` but is {type(lowercase_ )}" )
if height % 8 != 0 or width % 8 != 0:
raise ValueError(F"`height` and `width` have to be divisible by 8 but are {height} and {width}." )
if (callback_steps is None) or (
callback_steps is not None and (not isinstance(lowercase_ , lowercase_ ) or callback_steps <= 0)
):
raise ValueError(
F"`callback_steps` has to be a positive integer but is {callback_steps} of type"
F" {type(lowercase_ )}." )
# get prompt text embeddings
snake_case_ = self.tokenizer(
lowercase_ , padding='''max_length''' , max_length=self.tokenizer.model_max_length , return_tensors='''pt''' , )
snake_case_ = text_inputs.input_ids
if text_input_ids.shape[-1] > self.tokenizer.model_max_length:
snake_case_ = self.tokenizer.batch_decode(text_input_ids[:, self.tokenizer.model_max_length :] )
logger.warning(
'''The following part of your input was truncated because CLIP can only handle sequences up to'''
F" {self.tokenizer.model_max_length} tokens: {removed_text}" )
snake_case_ = text_input_ids[:, : self.tokenizer.model_max_length]
if text_embeddings is None:
snake_case_ = self.text_encoder(text_input_ids.to(self.device ) )[0]
# duplicate text embeddings for each generation per prompt, using mps friendly method
snake_case_ ,snake_case_ ,snake_case_ = text_embeddings.shape
snake_case_ = text_embeddings.repeat(1 , lowercase_ , 1 )
snake_case_ = text_embeddings.view(bs_embed * num_images_per_prompt , lowercase_ , -1 )
# here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
# of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
# corresponds to doing no classifier free guidance.
snake_case_ = guidance_scale > 1.0
# get unconditional embeddings for classifier free guidance
if do_classifier_free_guidance:
snake_case_ = 42
if negative_prompt is None:
snake_case_ = ['''''']
elif type(lowercase_ ) is not type(lowercase_ ):
raise TypeError(
F"`negative_prompt` should be the same type to `prompt`, but got {type(lowercase_ )} !="
F" {type(lowercase_ )}." )
elif isinstance(lowercase_ , lowercase_ ):
snake_case_ = [negative_prompt]
elif batch_size != len(lowercase_ ):
raise ValueError(
F"`negative_prompt`: {negative_prompt} has batch size {len(lowercase_ )}, but `prompt`:"
F" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
''' the batch size of `prompt`.''' )
else:
snake_case_ = negative_prompt
snake_case_ = text_input_ids.shape[-1]
snake_case_ = self.tokenizer(
lowercase_ , padding='''max_length''' , max_length=lowercase_ , truncation=lowercase_ , return_tensors='''pt''' , )
snake_case_ = self.text_encoder(uncond_input.input_ids.to(self.device ) )[0]
# duplicate unconditional embeddings for each generation per prompt, using mps friendly method
snake_case_ = uncond_embeddings.shape[1]
snake_case_ = uncond_embeddings.repeat(lowercase_ , lowercase_ , 1 )
snake_case_ = uncond_embeddings.view(batch_size * num_images_per_prompt , lowercase_ , -1 )
# For classifier free guidance, we need to do two forward passes.
# Here we concatenate the unconditional and text embeddings into a single batch
# to avoid doing two forward passes
snake_case_ = torch.cat([uncond_embeddings, text_embeddings] )
# get the initial random noise unless the user supplied it
# Unlike in other pipelines, latents need to be generated in the target device
# for 1-to-1 results reproducibility with the CompVis implementation.
# However this currently doesn't work in `mps`.
snake_case_ = (batch_size * num_images_per_prompt, self.unet.config.in_channels, height // 8, width // 8)
snake_case_ = (batch_size * num_images_per_prompt, self.unet.config.in_channels, 64, 64)
snake_case_ = text_embeddings.dtype
if latents is None:
if self.device.type == "mps":
# randn does not exist on mps
snake_case_ = torch.randn(
lowercase_ , generator=lowercase_ , device='''cpu''' , dtype=lowercase_ ).to(self.device )
snake_case_ = torch.randn(lowercase_ , generator=lowercase_ , device='''cpu''' , dtype=lowercase_ ).to(
self.device )
else:
snake_case_ = torch.randn(
lowercase_ , generator=lowercase_ , device=self.device , dtype=lowercase_ )
snake_case_ = torch.randn(lowercase_ , generator=lowercase_ , device=self.device , dtype=lowercase_ )
else:
if latents_reference.shape != latents_shape:
raise ValueError(F"Unexpected latents shape, got {latents.shape}, expected {latents_shape}" )
snake_case_ = latents_reference.to(self.device )
snake_case_ = latents.to(self.device )
# This is the key part of the pipeline where we
# try to ensure that the generated images w/ the same seed
# but different sizes actually result in similar images
snake_case_ = (latents_shape[3] - latents_shape_reference[3]) // 2
snake_case_ = (latents_shape[2] - latents_shape_reference[2]) // 2
snake_case_ = latents_shape_reference[3] if dx >= 0 else latents_shape_reference[3] + 2 * dx
snake_case_ = latents_shape_reference[2] if dy >= 0 else latents_shape_reference[2] + 2 * dy
snake_case_ = 0 if dx < 0 else dx
snake_case_ = 0 if dy < 0 else dy
snake_case_ = max(-dx , 0 )
snake_case_ = max(-dy , 0 )
# import pdb
# pdb.set_trace()
snake_case_ = latents_reference[:, :, dy : dy + h, dx : dx + w]
# set timesteps
self.scheduler.set_timesteps(lowercase_ )
# Some schedulers like PNDM have timesteps as arrays
# It's more optimized to move all timesteps to correct device beforehand
snake_case_ = self.scheduler.timesteps.to(self.device )
# scale the initial noise by the standard deviation required by the scheduler
snake_case_ = latents * self.scheduler.init_noise_sigma
# prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
# eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
# eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
# and should be between [0, 1]
snake_case_ = '''eta''' in set(inspect.signature(self.scheduler.step ).parameters.keys() )
snake_case_ = {}
if accepts_eta:
snake_case_ = eta
for i, t in enumerate(self.progress_bar(lowercase_ ) ):
# expand the latents if we are doing classifier free guidance
snake_case_ = torch.cat([latents] * 2 ) if do_classifier_free_guidance else latents
snake_case_ = self.scheduler.scale_model_input(lowercase_ , lowercase_ )
# predict the noise residual
snake_case_ = self.unet(lowercase_ , lowercase_ , encoder_hidden_states=lowercase_ ).sample
# perform guidance
if do_classifier_free_guidance:
snake_case_ ,snake_case_ = noise_pred.chunk(2 )
snake_case_ = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
# compute the previous noisy sample x_t -> x_t-1
snake_case_ = self.scheduler.step(lowercase_ , lowercase_ , lowercase_ , **lowercase_ ).prev_sample
# call the callback, if provided
if callback is not None and i % callback_steps == 0:
callback(lowercase_ , lowercase_ , lowercase_ )
snake_case_ = 1 / 0.1_8215 * latents
snake_case_ = self.vae.decode(lowercase_ ).sample
snake_case_ = (image / 2 + 0.5).clamp(0 , 1 )
# we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
snake_case_ = image.cpu().permute(0 , 2 , 3 , 1 ).float().numpy()
if self.safety_checker is not None:
snake_case_ = self.feature_extractor(self.numpy_to_pil(lowercase_ ) , return_tensors='''pt''' ).to(
self.device )
snake_case_ ,snake_case_ = self.safety_checker(
images=lowercase_ , clip_input=safety_checker_input.pixel_values.to(text_embeddings.dtype ) )
else:
snake_case_ = None
if output_type == "pil":
snake_case_ = self.numpy_to_pil(lowercase_ )
if not return_dict:
return (image, has_nsfw_concept)
return StableDiffusionPipelineOutput(images=lowercase_ , nsfw_content_detected=lowercase_ )
| 56
|
'''simple docstring'''
import contextlib
import os
import sqlitea
import pytest
from datasets import Dataset, Features, Value
from datasets.io.sql import SqlDatasetReader, SqlDatasetWriter
from ..utils import assert_arrow_memory_doesnt_increase, assert_arrow_memory_increases, require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
assert isinstance(__UpperCAmelCase, __UpperCAmelCase )
assert dataset.num_rows == 4
assert dataset.num_columns == 3
assert dataset.column_names == ["col_1", "col_2", "col_3"]
for feature, expected_dtype in expected_features.items():
assert dataset.features[feature].dtype == expected_dtype
@require_sqlalchemy
@pytest.mark.parametrize('''keep_in_memory''', [False, True] )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase():
snake_case_ = SqlDatasetReader(
'''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase, keep_in_memory=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
@require_sqlalchemy
@pytest.mark.parametrize(
'''features''', [
None,
{'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''},
{'''col_1''': '''string''', '''col_2''': '''string''', '''col_3''': '''string'''},
{'''col_1''': '''int32''', '''col_2''': '''int32''', '''col_3''': '''int32'''},
{'''col_1''': '''float32''', '''col_2''': '''float32''', '''col_3''': '''float32'''},
], )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
snake_case_ = features.copy() if features else default_expected_features
snake_case_ = (
Features({feature: Value(__UpperCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None
)
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, features=__UpperCAmelCase, cache_dir=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
with contextlib.closing(sqlitea.connect(__UpperCAmelCase ) ) as con:
snake_case_ = con.cursor()
cur.execute('''SELECT * FROM dataset''' )
for row in cur:
yield row
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=1 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=2 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
with pytest.raises(__UpperCAmelCase ):
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=0 ).write()
| 56
| 1
|
'''simple docstring'''
import collections
import os
import re
from pathlib import Path
a : str = 'src/transformers'
# Matches is_xxx_available()
a : Tuple = re.compile(r'is\_([a-z_]*)_available()')
# Catches a one-line _import_struct = {xxx}
a : List[str] = re.compile(r'^_import_structure\s+=\s+\{([^\}]+)\}')
# Catches a line with a key-values pattern: "bla": ["foo", "bar"]
a : List[str] = re.compile(r'\s+"\S*":\s+\[([^\]]*)\]')
# Catches a line if not is_foo_available
a : Optional[int] = re.compile(r'^\s*if\s+not\s+is\_[a-z_]*\_available\(\)')
# Catches a line _import_struct["bla"].append("foo")
a : str = re.compile(r'^\s*_import_structure\["\S*"\]\.append\("(\S*)"\)')
# Catches a line _import_struct["bla"].extend(["foo", "bar"]) or _import_struct["bla"] = ["foo", "bar"]
a : Tuple = re.compile(r'^\s*_import_structure\[\S*\](?:\.extend\(|\s*=\s+)\[([^\]]*)\]')
# Catches a line with an object between quotes and a comma: "MyModel",
a : int = re.compile(r'^\s+"([^"]+)",')
# Catches a line with objects between brackets only: ["foo", "bar"],
a : Optional[int] = re.compile(r'^\s+\[([^\]]+)\]')
# Catches a line with from foo import bar, bla, boo
a : Optional[int] = re.compile(r'\s+from\s+\S*\s+import\s+([^\(\s].*)\n')
# Catches a line with try:
a : List[str] = re.compile(r'^\s*try:')
# Catches a line with else:
a : str = re.compile(r'^\s*else:')
def __magic_name__ ( __UpperCAmelCase ) -> Dict:
'''simple docstring'''
if _re_test_backend.search(__UpperCAmelCase ) is None:
return None
snake_case_ = [b[0] for b in _re_backend.findall(__UpperCAmelCase )]
backends.sort()
return "_and_".join(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
with open(__UpperCAmelCase, '''r''', encoding='''utf-8''', newline='''\n''' ) as f:
snake_case_ = f.readlines()
snake_case_ = 0
while line_index < len(__UpperCAmelCase ) and not lines[line_index].startswith('''_import_structure = {''' ):
line_index += 1
# If this is a traditional init, just return.
if line_index >= len(__UpperCAmelCase ):
return None
# First grab the objects without a specific backend in _import_structure
snake_case_ = []
while not lines[line_index].startswith('''if TYPE_CHECKING''' ) and find_backend(lines[line_index] ) is None:
snake_case_ = lines[line_index]
# If we have everything on a single line, let's deal with it.
if _re_one_line_import_struct.search(__UpperCAmelCase ):
snake_case_ = _re_one_line_import_struct.search(__UpperCAmelCase ).groups()[0]
snake_case_ = re.findall(r'''\[([^\]]+)\]''', __UpperCAmelCase )
for imp in imports:
objects.extend([obj[1:-1] for obj in imp.split(''', ''' )] )
line_index += 1
continue
snake_case_ = _re_import_struct_key_value.search(__UpperCAmelCase )
if single_line_import_search is not None:
snake_case_ = [obj[1:-1] for obj in single_line_import_search.groups()[0].split(''', ''' ) if len(__UpperCAmelCase ) > 0]
objects.extend(__UpperCAmelCase )
elif line.startswith(''' ''' * 8 + '''"''' ):
objects.append(line[9:-3] )
line_index += 1
snake_case_ = {'''none''': objects}
# Let's continue with backend-specific objects in _import_structure
while not lines[line_index].startswith('''if TYPE_CHECKING''' ):
# If the line is an if not is_backend_available, we grab all objects associated.
snake_case_ = find_backend(lines[line_index] )
# Check if the backend declaration is inside a try block:
if _re_try.search(lines[line_index - 1] ) is None:
snake_case_ = None
if backend is not None:
line_index += 1
# Scroll until we hit the else block of try-except-else
while _re_else.search(lines[line_index] ) is None:
line_index += 1
line_index += 1
snake_case_ = []
# Until we unindent, add backend objects to the list
while len(lines[line_index] ) <= 1 or lines[line_index].startswith(''' ''' * 4 ):
snake_case_ = lines[line_index]
if _re_import_struct_add_one.search(__UpperCAmelCase ) is not None:
objects.append(_re_import_struct_add_one.search(__UpperCAmelCase ).groups()[0] )
elif _re_import_struct_add_many.search(__UpperCAmelCase ) is not None:
snake_case_ = _re_import_struct_add_many.search(__UpperCAmelCase ).groups()[0].split(''', ''' )
snake_case_ = [obj[1:-1] for obj in imports if len(__UpperCAmelCase ) > 0]
objects.extend(__UpperCAmelCase )
elif _re_between_brackets.search(__UpperCAmelCase ) is not None:
snake_case_ = _re_between_brackets.search(__UpperCAmelCase ).groups()[0].split(''', ''' )
snake_case_ = [obj[1:-1] for obj in imports if len(__UpperCAmelCase ) > 0]
objects.extend(__UpperCAmelCase )
elif _re_quote_object.search(__UpperCAmelCase ) is not None:
objects.append(_re_quote_object.search(__UpperCAmelCase ).groups()[0] )
elif line.startswith(''' ''' * 8 + '''"''' ):
objects.append(line[9:-3] )
elif line.startswith(''' ''' * 12 + '''"''' ):
objects.append(line[13:-3] )
line_index += 1
snake_case_ = objects
else:
line_index += 1
# At this stage we are in the TYPE_CHECKING part, first grab the objects without a specific backend
snake_case_ = []
while (
line_index < len(__UpperCAmelCase )
and find_backend(lines[line_index] ) is None
and not lines[line_index].startswith('''else''' )
):
snake_case_ = lines[line_index]
snake_case_ = _re_import.search(__UpperCAmelCase )
if single_line_import_search is not None:
objects.extend(single_line_import_search.groups()[0].split(''', ''' ) )
elif line.startswith(''' ''' * 8 ):
objects.append(line[8:-2] )
line_index += 1
snake_case_ = {'''none''': objects}
# Let's continue with backend-specific objects
while line_index < len(__UpperCAmelCase ):
# If the line is an if is_backend_available, we grab all objects associated.
snake_case_ = find_backend(lines[line_index] )
# Check if the backend declaration is inside a try block:
if _re_try.search(lines[line_index - 1] ) is None:
snake_case_ = None
if backend is not None:
line_index += 1
# Scroll until we hit the else block of try-except-else
while _re_else.search(lines[line_index] ) is None:
line_index += 1
line_index += 1
snake_case_ = []
# Until we unindent, add backend objects to the list
while len(lines[line_index] ) <= 1 or lines[line_index].startswith(''' ''' * 8 ):
snake_case_ = lines[line_index]
snake_case_ = _re_import.search(__UpperCAmelCase )
if single_line_import_search is not None:
objects.extend(single_line_import_search.groups()[0].split(''', ''' ) )
elif line.startswith(''' ''' * 12 ):
objects.append(line[12:-2] )
line_index += 1
snake_case_ = objects
else:
line_index += 1
return import_dict_objects, type_hint_objects
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
def find_duplicates(__UpperCAmelCase ):
return [k for k, v in collections.Counter(__UpperCAmelCase ).items() if v > 1]
if list(import_dict_objects.keys() ) != list(type_hint_objects.keys() ):
return ["Both sides of the init do not have the same backends!"]
snake_case_ = []
for key in import_dict_objects.keys():
snake_case_ = find_duplicates(import_dict_objects[key] )
if duplicate_imports:
errors.append(F"Duplicate _import_structure definitions for: {duplicate_imports}" )
snake_case_ = find_duplicates(type_hint_objects[key] )
if duplicate_type_hints:
errors.append(F"Duplicate TYPE_CHECKING objects for: {duplicate_type_hints}" )
if sorted(set(import_dict_objects[key] ) ) != sorted(set(type_hint_objects[key] ) ):
snake_case_ = '''base imports''' if key == '''none''' else F"{key} backend"
errors.append(F"Differences for {name}:" )
for a in type_hint_objects[key]:
if a not in import_dict_objects[key]:
errors.append(F" {a} in TYPE_HINT but not in _import_structure." )
for a in import_dict_objects[key]:
if a not in type_hint_objects[key]:
errors.append(F" {a} in _import_structure but not in TYPE_HINT." )
return errors
def __magic_name__ ( ) -> Tuple:
'''simple docstring'''
snake_case_ = []
for root, _, files in os.walk(__UpperCAmelCase ):
if "__init__.py" in files:
snake_case_ = os.path.join(__UpperCAmelCase, '''__init__.py''' )
snake_case_ = parse_init(__UpperCAmelCase )
if objects is not None:
snake_case_ = analyze_results(*__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0:
snake_case_ = F"Problem in {fname}, both halves do not define the same objects.\n{errors[0]}"
failures.append('''\n'''.join(__UpperCAmelCase ) )
if len(__UpperCAmelCase ) > 0:
raise ValueError('''\n\n'''.join(__UpperCAmelCase ) )
def __magic_name__ ( ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = []
for path, directories, files in os.walk(__UpperCAmelCase ):
for folder in directories:
# Ignore private modules
if folder.startswith('''_''' ):
directories.remove(__UpperCAmelCase )
continue
# Ignore leftovers from branches (empty folders apart from pycache)
if len(list((Path(__UpperCAmelCase ) / folder).glob('''*.py''' ) ) ) == 0:
continue
snake_case_ = str((Path(__UpperCAmelCase ) / folder).relative_to(__UpperCAmelCase ) )
snake_case_ = short_path.replace(os.path.sep, '''.''' )
submodules.append(__UpperCAmelCase )
for fname in files:
if fname == "__init__.py":
continue
snake_case_ = str((Path(__UpperCAmelCase ) / fname).relative_to(__UpperCAmelCase ) )
snake_case_ = short_path.replace('''.py''', '''''' ).replace(os.path.sep, '''.''' )
if len(submodule.split('''.''' ) ) == 1:
submodules.append(__UpperCAmelCase )
return submodules
a : Dict = [
'convert_pytorch_checkpoint_to_tf2',
'modeling_flax_pytorch_utils',
'models.esm.openfold_utils',
]
def __magic_name__ ( ) -> Optional[int]:
'''simple docstring'''
from transformers.utils import direct_transformers_import
snake_case_ = direct_transformers_import(__UpperCAmelCase )
snake_case_ = set(transformers._import_structure.keys() )
# This contains all the base keys of the _import_structure object defined in the init, but if the user is missing
# some optional dependencies, they may not have all of them. Thus we read the init to read all additions and
# (potentiall re-) add them.
with open(os.path.join(__UpperCAmelCase, '''__init__.py''' ), '''r''' ) as f:
snake_case_ = f.read()
import_structure_keys.update(set(re.findall(r'''import_structure\[\"([^\"]*)\"\]''', __UpperCAmelCase ) ) )
snake_case_ = [
module
for module in get_transformers_submodules()
if module not in IGNORE_SUBMODULES and module not in import_structure_keys
]
if len(__UpperCAmelCase ) > 0:
snake_case_ = '''\n'''.join(F"- {module}" for module in module_not_registered )
raise ValueError(
'''The following submodules are not properly registed in the main init of Transformers:\n'''
F"{list_of_modules}\n"
'''Make sure they appear somewhere in the keys of `_import_structure` with an empty list as value.''' )
if __name__ == "__main__":
check_all_inits()
check_submodules()
| 56
|
'''simple docstring'''
from collections import defaultdict
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 1
snake_case_ = True
for v in tree[start]:
if v not in visited:
ret += dfs(__UpperCAmelCase )
if ret % 2 == 0:
cuts.append(__UpperCAmelCase )
return ret
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
dfs(1 )
if __name__ == "__main__":
a ,a : Dict = 10, 9
a : Dict = defaultdict(list)
a : dict[int, bool] = {}
a : list[int] = []
a : Tuple = 0
a : str = [(2, 1), (3, 1), (4, 3), (5, 2), (6, 1), (7, 2), (8, 6), (9, 8), (10, 8)]
for u, v in edges:
tree[u].append(v)
tree[v].append(u)
even_tree()
print(len(cuts) - 1)
| 56
| 1
|
'''simple docstring'''
from typing import Any, Callable, Dict, List, Optional, Union
import torch
from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
DiffusionPipeline,
LMSDiscreteScheduler,
PNDMScheduler,
StableDiffusionPipeline,
UNetaDConditionModel,
)
from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput
from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
a : List[Any] = 'CompVis/stable-diffusion-v1-1'
a : Union[str, Any] = 'CompVis/stable-diffusion-v1-2'
a : Tuple = 'CompVis/stable-diffusion-v1-3'
a : Optional[Any] = 'CompVis/stable-diffusion-v1-4'
class a ( _lowerCamelCase ):
def __init__( self : Any , lowercase_ : AutoencoderKL , lowercase_ : CLIPTextModel , lowercase_ : CLIPTokenizer , lowercase_ : UNetaDConditionModel , lowercase_ : Union[DDIMScheduler, PNDMScheduler, LMSDiscreteScheduler] , lowercase_ : StableDiffusionSafetyChecker , lowercase_ : CLIPImageProcessor , lowercase_ : bool = True , ):
super()._init_()
snake_case_ = StableDiffusionPipeline.from_pretrained(lowercase_ )
snake_case_ = StableDiffusionPipeline.from_pretrained(lowercase_ )
snake_case_ = StableDiffusionPipeline.from_pretrained(lowercase_ )
snake_case_ = StableDiffusionPipeline(
vae=lowercase_ , text_encoder=lowercase_ , tokenizer=lowercase_ , unet=lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , feature_extractor=lowercase_ , requires_safety_checker=lowercase_ , )
self.register_modules(pipelinea=self.pipea , pipelinea=self.pipea , pipelinea=self.pipea , pipelinea=self.pipea )
@property
def A_ ( self : Optional[int] ):
return {k: getattr(self , lowercase_ ) for k in self.config.keys() if not k.startswith('''_''' )}
def A_ ( self : List[Any] , lowercase_ : Optional[Union[str, int]] = "auto" ):
if slice_size == "auto":
# half the attention head size is usually a good trade-off between
# speed and memory
snake_case_ = self.unet.config.attention_head_dim // 2
self.unet.set_attention_slice(lowercase_ )
def A_ ( self : Optional[int] ):
self.enable_attention_slicing(lowercase_ )
@torch.no_grad()
def A_ ( self : Any , lowercase_ : Union[str, List[str]] , lowercase_ : int = 512 , lowercase_ : int = 512 , lowercase_ : int = 50 , lowercase_ : float = 7.5 , lowercase_ : Optional[Union[str, List[str]]] = None , lowercase_ : Optional[int] = 1 , lowercase_ : float = 0.0 , lowercase_ : Optional[torch.Generator] = None , lowercase_ : Optional[torch.FloatTensor] = None , lowercase_ : Optional[str] = "pil" , lowercase_ : bool = True , lowercase_ : Optional[Callable[[int, int, torch.FloatTensor], None]] = None , lowercase_ : int = 1 , **lowercase_ : Union[str, Any] , ):
return self.pipea(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
@torch.no_grad()
def A_ ( self : str , lowercase_ : Union[str, List[str]] , lowercase_ : int = 512 , lowercase_ : int = 512 , lowercase_ : int = 50 , lowercase_ : float = 7.5 , lowercase_ : Optional[Union[str, List[str]]] = None , lowercase_ : Optional[int] = 1 , lowercase_ : float = 0.0 , lowercase_ : Optional[torch.Generator] = None , lowercase_ : Optional[torch.FloatTensor] = None , lowercase_ : Optional[str] = "pil" , lowercase_ : bool = True , lowercase_ : Optional[Callable[[int, int, torch.FloatTensor], None]] = None , lowercase_ : int = 1 , **lowercase_ : Optional[int] , ):
return self.pipea(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
@torch.no_grad()
def A_ ( self : Tuple , lowercase_ : Union[str, List[str]] , lowercase_ : int = 512 , lowercase_ : int = 512 , lowercase_ : int = 50 , lowercase_ : float = 7.5 , lowercase_ : Optional[Union[str, List[str]]] = None , lowercase_ : Optional[int] = 1 , lowercase_ : float = 0.0 , lowercase_ : Optional[torch.Generator] = None , lowercase_ : Optional[torch.FloatTensor] = None , lowercase_ : Optional[str] = "pil" , lowercase_ : bool = True , lowercase_ : Optional[Callable[[int, int, torch.FloatTensor], None]] = None , lowercase_ : int = 1 , **lowercase_ : Tuple , ):
return self.pipea(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
@torch.no_grad()
def A_ ( self : Optional[Any] , lowercase_ : Union[str, List[str]] , lowercase_ : int = 512 , lowercase_ : int = 512 , lowercase_ : int = 50 , lowercase_ : float = 7.5 , lowercase_ : Optional[Union[str, List[str]]] = None , lowercase_ : Optional[int] = 1 , lowercase_ : float = 0.0 , lowercase_ : Optional[torch.Generator] = None , lowercase_ : Optional[torch.FloatTensor] = None , lowercase_ : Optional[str] = "pil" , lowercase_ : bool = True , lowercase_ : Optional[Callable[[int, int, torch.FloatTensor], None]] = None , lowercase_ : int = 1 , **lowercase_ : int , ):
return self.pipea(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
@torch.no_grad()
def A_ ( self : Optional[int] , lowercase_ : Union[str, List[str]] , lowercase_ : int = 512 , lowercase_ : int = 512 , lowercase_ : int = 50 , lowercase_ : float = 7.5 , lowercase_ : Optional[Union[str, List[str]]] = None , lowercase_ : Optional[int] = 1 , lowercase_ : float = 0.0 , lowercase_ : Optional[torch.Generator] = None , lowercase_ : Optional[torch.FloatTensor] = None , lowercase_ : Optional[str] = "pil" , lowercase_ : bool = True , lowercase_ : Optional[Callable[[int, int, torch.FloatTensor], None]] = None , lowercase_ : int = 1 , **lowercase_ : int , ):
snake_case_ = '''cuda''' if torch.cuda.is_available() else '''cpu'''
self.to(lowercase_ )
# Checks if the height and width are divisible by 8 or not
if height % 8 != 0 or width % 8 != 0:
raise ValueError(F"`height` and `width` must be divisible by 8 but are {height} and {width}." )
# Get first result from Stable Diffusion Checkpoint v1.1
snake_case_ = self.textaimg_sda_a(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
# Get first result from Stable Diffusion Checkpoint v1.2
snake_case_ = self.textaimg_sda_a(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
# Get first result from Stable Diffusion Checkpoint v1.3
snake_case_ = self.textaimg_sda_a(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
# Get first result from Stable Diffusion Checkpoint v1.4
snake_case_ = self.textaimg_sda_a(
prompt=lowercase_ , height=lowercase_ , width=lowercase_ , num_inference_steps=lowercase_ , guidance_scale=lowercase_ , negative_prompt=lowercase_ , num_images_per_prompt=lowercase_ , eta=lowercase_ , generator=lowercase_ , latents=lowercase_ , output_type=lowercase_ , return_dict=lowercase_ , callback=lowercase_ , callback_steps=lowercase_ , **lowercase_ , )
# Get all result images into a single list and pass it via StableDiffusionPipelineOutput for final result
return StableDiffusionPipelineOutput([resa[0], resa[0], resa[0], resa[0]] )
| 56
|
'''simple docstring'''
import math
from collections.abc import Callable
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
snake_case_ = xa
snake_case_ = xa
while True:
if x_n == x_na or function(__UpperCAmelCase ) == function(__UpperCAmelCase ):
raise ZeroDivisionError('''float division by zero, could not find root''' )
snake_case_ = x_na - (
function(__UpperCAmelCase ) / ((function(__UpperCAmelCase ) - function(__UpperCAmelCase )) / (x_na - x_n))
)
if abs(x_na - x_na ) < 10**-5:
return x_na
snake_case_ = x_na
snake_case_ = x_na
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
return math.pow(__UpperCAmelCase, 3 ) - (2 * x) - 5
if __name__ == "__main__":
print(intersection(f, 3, 3.5))
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
from numpy import array, cos, cross, floataa, radians, sin
from numpy.typing import NDArray
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = False ) -> list[float]:
'''simple docstring'''
if radian_mode:
return [magnitude * cos(__UpperCAmelCase ), magnitude * sin(__UpperCAmelCase )]
return [magnitude * cos(radians(__UpperCAmelCase ) ), magnitude * sin(radians(__UpperCAmelCase ) )]
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 10**-1 ) -> bool:
'''simple docstring'''
snake_case_ = cross(__UpperCAmelCase, __UpperCAmelCase )
snake_case_ = sum(__UpperCAmelCase )
return abs(__UpperCAmelCase ) < eps
if __name__ == "__main__":
# Test to check if it works
a : Tuple = array(
[
polar_force(718.4, 180 - 30),
polar_force(879.54, 45),
polar_force(100, -90),
]
)
a : NDArray[floataa] = array([[0, 0], [0, 0], [0, 0]])
assert in_static_equilibrium(forces, location)
# Problem 1 in image_data/2D_problems.jpg
a : Union[str, Any] = array(
[
polar_force(30 * 9.81, 15),
polar_force(215, 180 - 45),
polar_force(264, 90 - 30),
]
)
a : List[Any] = array([[0, 0], [0, 0], [0, 0]])
assert in_static_equilibrium(forces, location)
# Problem in image_data/2D_problems_1.jpg
a : List[Any] = array([[0, -2000], [0, -1200], [0, 1_5600], [0, -1_2400]])
a : str = array([[0, 0], [6, 0], [10, 0], [12, 0]])
assert in_static_equilibrium(forces, location)
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_url
from PIL import Image
from transformers import DPTConfig, DPTForDepthEstimation, DPTForSemanticSegmentation, DPTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
a : Any = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = DPTConfig()
if "large" in checkpoint_url:
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 24
snake_case_ = 16
snake_case_ = [5, 11, 17, 23]
snake_case_ = [256, 512, 1024, 1024]
snake_case_ = (1, 384, 384)
if "ade" in checkpoint_url:
snake_case_ = True
snake_case_ = 150
snake_case_ = '''huggingface/label-files'''
snake_case_ = '''ade20k-id2label.json'''
snake_case_ = json.load(open(cached_download(hf_hub_url(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ) ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = idalabel
snake_case_ = {v: k for k, v in idalabel.items()}
snake_case_ = [1, 150, 480, 480]
return config, expected_shape
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = ['''pretrained.model.head.weight''', '''pretrained.model.head.bias''']
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if (
"pretrained.model" in name
and "cls_token" not in name
and "pos_embed" not in name
and "patch_embed" not in name
):
snake_case_ = name.replace('''pretrained.model''', '''dpt.encoder''' )
if "pretrained.model" in name:
snake_case_ = name.replace('''pretrained.model''', '''dpt.embeddings''' )
if "patch_embed" in name:
snake_case_ = name.replace('''patch_embed''', '''patch_embeddings''' )
if "pos_embed" in name:
snake_case_ = name.replace('''pos_embed''', '''position_embeddings''' )
if "attn.proj" in name:
snake_case_ = name.replace('''attn.proj''', '''attention.output.dense''' )
if "proj" in name and "project" not in name:
snake_case_ = name.replace('''proj''', '''projection''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''layer''' )
if "mlp.fc1" in name:
snake_case_ = name.replace('''mlp.fc1''', '''intermediate.dense''' )
if "mlp.fc2" in name:
snake_case_ = name.replace('''mlp.fc2''', '''output.dense''' )
if "norm1" in name:
snake_case_ = name.replace('''norm1''', '''layernorm_before''' )
if "norm2" in name:
snake_case_ = name.replace('''norm2''', '''layernorm_after''' )
if "scratch.output_conv" in name:
snake_case_ = name.replace('''scratch.output_conv''', '''head''' )
if "scratch" in name:
snake_case_ = name.replace('''scratch''', '''neck''' )
if "layer1_rn" in name:
snake_case_ = name.replace('''layer1_rn''', '''convs.0''' )
if "layer2_rn" in name:
snake_case_ = name.replace('''layer2_rn''', '''convs.1''' )
if "layer3_rn" in name:
snake_case_ = name.replace('''layer3_rn''', '''convs.2''' )
if "layer4_rn" in name:
snake_case_ = name.replace('''layer4_rn''', '''convs.3''' )
if "refinenet" in name:
snake_case_ = int(name[len('''neck.refinenet''' ) : len('''neck.refinenet''' ) + 1] )
# tricky here: we need to map 4 to 0, 3 to 1, 2 to 2 and 1 to 3
snake_case_ = name.replace(F"refinenet{layer_idx}", F"fusion_stage.layers.{abs(layer_idx-4 )}" )
if "out_conv" in name:
snake_case_ = name.replace('''out_conv''', '''projection''' )
if "resConfUnit1" in name:
snake_case_ = name.replace('''resConfUnit1''', '''residual_layer1''' )
if "resConfUnit2" in name:
snake_case_ = name.replace('''resConfUnit2''', '''residual_layer2''' )
if "conv1" in name:
snake_case_ = name.replace('''conv1''', '''convolution1''' )
if "conv2" in name:
snake_case_ = name.replace('''conv2''', '''convolution2''' )
# readout blocks
if "pretrained.act_postprocess1.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.0.project.0''', '''neck.reassemble_stage.readout_projects.0.0''' )
if "pretrained.act_postprocess2.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.0.project.0''', '''neck.reassemble_stage.readout_projects.1.0''' )
if "pretrained.act_postprocess3.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.0.project.0''', '''neck.reassemble_stage.readout_projects.2.0''' )
if "pretrained.act_postprocess4.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.0.project.0''', '''neck.reassemble_stage.readout_projects.3.0''' )
# resize blocks
if "pretrained.act_postprocess1.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.3''', '''neck.reassemble_stage.layers.0.projection''' )
if "pretrained.act_postprocess1.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.4''', '''neck.reassemble_stage.layers.0.resize''' )
if "pretrained.act_postprocess2.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.3''', '''neck.reassemble_stage.layers.1.projection''' )
if "pretrained.act_postprocess2.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.4''', '''neck.reassemble_stage.layers.1.resize''' )
if "pretrained.act_postprocess3.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.3''', '''neck.reassemble_stage.layers.2.projection''' )
if "pretrained.act_postprocess4.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.3''', '''neck.reassemble_stage.layers.3.projection''' )
if "pretrained.act_postprocess4.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.4''', '''neck.reassemble_stage.layers.3.resize''' )
if "pretrained" in name:
snake_case_ = name.replace('''pretrained''', '''dpt''' )
if "bn" in name:
snake_case_ = name.replace('''bn''', '''batch_norm''' )
if "head" in name:
snake_case_ = name.replace('''head''', '''head.head''' )
if "encoder.norm" in name:
snake_case_ = name.replace('''encoder.norm''', '''layernorm''' )
if "auxlayer" in name:
snake_case_ = name.replace('''auxlayer''', '''auxiliary_head.head''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.weight" )
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.bias" )
# next, add query, keys and values (in that order) to the state dict
snake_case_ = in_proj_weight[: config.hidden_size, :]
snake_case_ = in_proj_bias[: config.hidden_size]
snake_case_ = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
snake_case_ = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
snake_case_ = in_proj_weight[
-config.hidden_size :, :
]
snake_case_ = in_proj_bias[-config.hidden_size :]
def __magic_name__ ( ) -> Any:
'''simple docstring'''
snake_case_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
snake_case_ = Image.open(requests.get(__UpperCAmelCase, stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dpt_config(__UpperCAmelCase )
# load original state_dict from URL
snake_case_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase, map_location='''cpu''' )
# remove certain keys
remove_ignore_keys_(__UpperCAmelCase )
# rename keys
for key in state_dict.copy().keys():
snake_case_ = state_dict.pop(__UpperCAmelCase )
snake_case_ = val
# read in qkv matrices
read_in_q_k_v(__UpperCAmelCase, __UpperCAmelCase )
# load HuggingFace model
snake_case_ = DPTForSemanticSegmentation(__UpperCAmelCase ) if '''ade''' in checkpoint_url else DPTForDepthEstimation(__UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
# Check outputs on an image
snake_case_ = 480 if '''ade''' in checkpoint_url else 384
snake_case_ = DPTImageProcessor(size=__UpperCAmelCase )
snake_case_ = prepare_img()
snake_case_ = image_processor(__UpperCAmelCase, return_tensors='''pt''' )
# forward pass
snake_case_ = model(**__UpperCAmelCase ).logits if '''ade''' in checkpoint_url else model(**__UpperCAmelCase ).predicted_depth
# Assert logits
snake_case_ = torch.tensor([[6.3_1_9_9, 6.3_6_2_9, 6.4_1_4_8], [6.3_8_5_0, 6.3_6_1_5, 6.4_1_6_6], [6.3_5_1_9, 6.3_1_7_6, 6.3_5_7_5]] )
if "ade" in checkpoint_url:
snake_case_ = torch.tensor([[4.0_4_8_0, 4.2_4_2_0, 4.4_3_6_0], [4.3_1_2_4, 4.5_6_9_3, 4.8_2_6_1], [4.5_7_6_8, 4.8_9_6_5, 5.2_1_6_3]] )
assert outputs.shape == torch.Size(__UpperCAmelCase )
assert (
torch.allclose(outputs[0, 0, :3, :3], __UpperCAmelCase, atol=1e-4 )
if "ade" in checkpoint_url
else torch.allclose(outputs[0, :3, :3], __UpperCAmelCase )
)
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
print(F"Saving model to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
print(F"Saving image processor to {pytorch_dump_folder_path}" )
image_processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print('''Pushing model to hub...''' )
model.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add model''', use_temp_dir=__UpperCAmelCase, )
image_processor.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add image processor''', use_temp_dir=__UpperCAmelCase, )
if __name__ == "__main__":
a : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--checkpoint_url',
default='https://github.com/intel-isl/DPT/releases/download/1_0/dpt_large-midas-2f21e586.pt',
type=str,
help='URL of the original DPT checkpoint you\'d like to convert.',
)
parser.add_argument(
'--pytorch_dump_folder_path',
default=None,
type=str,
required=True,
help='Path to the output PyTorch model directory.',
)
parser.add_argument(
'--push_to_hub',
action='store_true',
)
parser.add_argument(
'--model_name',
default='dpt-large',
type=str,
help='Name of the model, in case you\'re pushing to the hub.',
)
a : List[Any] = parser.parse_args()
convert_dpt_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 56
| 1
|
'''simple docstring'''
from random import randint, random
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = False, __UpperCAmelCase = False, __UpperCAmelCase = 5, ) -> list:
'''simple docstring'''
snake_case_ = [[-1] * number_of_cells] # Create a highway without any car
snake_case_ = 0
snake_case_ = max(__UpperCAmelCase, 0 )
while i < number_of_cells:
snake_case_ = (
randint(0, __UpperCAmelCase ) if random_speed else initial_speed
) # Place the cars
i += (
randint(1, max_speed * 2 ) if random_frequency else frequency
) # Arbitrary number, may need tuning
return highway
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 0
snake_case_ = highway_now[car_index + 1 :]
for cell in range(len(__UpperCAmelCase ) ): # May need a better name for this
if cells[cell] != -1: # If the cell is not empty then
return distance # we have the distance we wanted
distance += 1
# Here if the car is near the end of the highway
return distance + get_distance(__UpperCAmelCase, -1 )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> list:
'''simple docstring'''
snake_case_ = len(__UpperCAmelCase )
# Beforce calculations, the highway is empty
snake_case_ = [-1] * number_of_cells
for car_index in range(__UpperCAmelCase ):
if highway_now[car_index] != -1:
# Add 1 to the current speed of the car and cap the speed
snake_case_ = min(highway_now[car_index] + 1, __UpperCAmelCase )
# Number of empty cell before the next car
snake_case_ = get_distance(__UpperCAmelCase, __UpperCAmelCase ) - 1
# We can't have the car causing an accident
snake_case_ = min(next_highway[car_index], __UpperCAmelCase )
if random() < probability:
# Randomly, a driver will slow down
snake_case_ = max(next_highway[car_index] - 1, 0 )
return next_highway
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> list:
'''simple docstring'''
snake_case_ = len(highway[0] )
for i in range(__UpperCAmelCase ):
snake_case_ = update(highway[i], __UpperCAmelCase, __UpperCAmelCase )
snake_case_ = [-1] * number_of_cells
for car_index in range(__UpperCAmelCase ):
snake_case_ = next_speeds_calculated[car_index]
if speed != -1:
# Change the position based on the speed (with % to create the loop)
snake_case_ = (car_index + speed) % number_of_cells
# Commit the change of position
snake_case_ = speed
highway.append(__UpperCAmelCase )
return highway
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
import re
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
snake_case_ = re.compile(
r'''^(?:0|94|\+94|0{2}94)''' r'''7(0|1|2|4|5|6|7|8)''' r'''(-| |)''' r'''\d{7}$''' )
return bool(re.search(__UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
a : Any = '0094702343221'
print(is_sri_lankan_phone_number(phone))
| 56
| 1
|
'''simple docstring'''
# flake8: noqa
# Lint as: python3
from typing import Dict, List, Optional, Type
from .. import config
from ..utils import logging
from .formatting import (
ArrowFormatter,
CustomFormatter,
Formatter,
PandasFormatter,
PythonFormatter,
TensorFormatter,
format_table,
query_table,
)
from .np_formatter import NumpyFormatter
a : List[str] = logging.get_logger(__name__)
a : Dict[Optional[str], Type[Formatter]] = {}
a : Dict[Optional[str], str] = {}
a : Dict[Optional[str], Exception] = {}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = None, ) -> List[Any]:
'''simple docstring'''
snake_case_ = aliases if aliases is not None else []
if format_type in _FORMAT_TYPES:
logger.warning(
F"Overwriting format type '{format_type}' ({_FORMAT_TYPES[format_type].__name__} -> {formatter_cls.__name__})" )
snake_case_ = formatter_cls
for alias in set(aliases + [format_type] ):
if alias in _FORMAT_TYPES_ALIASES:
logger.warning(
F"Overwriting format type alias '{alias}' ({_FORMAT_TYPES_ALIASES[alias]} -> {format_type})" )
snake_case_ = format_type
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = None ) -> Any:
'''simple docstring'''
snake_case_ = aliases if aliases is not None else []
for alias in set(aliases + [format_type] ):
snake_case_ = unavailable_error
# Here we define all the available formatting functions that can be used by `Dataset.set_format`
_register_formatter(PythonFormatter, None, aliases=['python'])
_register_formatter(ArrowFormatter, 'arrow', aliases=['pa', 'pyarrow'])
_register_formatter(NumpyFormatter, 'numpy', aliases=['np'])
_register_formatter(PandasFormatter, 'pandas', aliases=['pd'])
_register_formatter(CustomFormatter, 'custom')
if config.TORCH_AVAILABLE:
from .torch_formatter import TorchFormatter
_register_formatter(TorchFormatter, 'torch', aliases=['pt', 'pytorch'])
else:
a : Tuple = ValueError('PyTorch needs to be installed to be able to return PyTorch tensors.')
_register_unavailable_formatter(_torch_error, 'torch', aliases=['pt', 'pytorch'])
if config.TF_AVAILABLE:
from .tf_formatter import TFFormatter
_register_formatter(TFFormatter, 'tensorflow', aliases=['tf'])
else:
a : List[str] = ValueError('Tensorflow needs to be installed to be able to return Tensorflow tensors.')
_register_unavailable_formatter(_tf_error, 'tensorflow', aliases=['tf'])
if config.JAX_AVAILABLE:
from .jax_formatter import JaxFormatter
_register_formatter(JaxFormatter, 'jax', aliases=[])
else:
a : Optional[int] = ValueError('JAX needs to be installed to be able to return JAX arrays.')
_register_unavailable_formatter(_jax_error, 'jax', aliases=[])
def __magic_name__ ( __UpperCAmelCase ) -> Optional[str]:
'''simple docstring'''
if format_type in _FORMAT_TYPES_ALIASES:
return _FORMAT_TYPES_ALIASES[format_type]
else:
return format_type
def __magic_name__ ( __UpperCAmelCase, **__UpperCAmelCase ) -> Formatter:
'''simple docstring'''
snake_case_ = get_format_type_from_alias(__UpperCAmelCase )
if format_type in _FORMAT_TYPES:
return _FORMAT_TYPES[format_type](**__UpperCAmelCase )
if format_type in _FORMAT_TYPES_ALIASES_UNAVAILABLE:
raise _FORMAT_TYPES_ALIASES_UNAVAILABLE[format_type]
else:
raise ValueError(
F"Return type should be None or selected in {list(type for type in _FORMAT_TYPES.keys() if type != None )}, but got '{format_type}'" )
| 56
|
'''simple docstring'''
import re
from filelock import FileLock
try:
import nltk
a : Union[str, Any] = True
except (ImportError, ModuleNotFoundError):
a : Any = False
if NLTK_AVAILABLE:
with FileLock('.lock') as lock:
nltk.download('punkt', quiet=True)
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
re.sub('''<n>''', '''''', __UpperCAmelCase ) # remove pegasus newline char
assert NLTK_AVAILABLE, "nltk must be installed to separate newlines between sentences. (pip install nltk)"
return "\n".join(nltk.sent_tokenize(__UpperCAmelCase ) )
| 56
| 1
|
'''simple docstring'''
import time
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_torch, torch_device
from ..test_modeling_common import ids_tensor
if is_torch_available():
import torch
from transformers.generation import (
MaxLengthCriteria,
MaxNewTokensCriteria,
MaxTimeCriteria,
StoppingCriteriaList,
validate_stopping_criteria,
)
@require_torch
class a ( unittest.TestCase ):
def A_ ( self : str , lowercase_ : List[str] ):
snake_case_ = 3
snake_case_ = 250
snake_case_ = ids_tensor((batch_size, length) , lowercase_ )
snake_case_ = torch.ones((batch_size, length) , device=lowercase_ , dtype=torch.float ) / length
return input_ids, scores
def A_ ( self : List[Any] ):
snake_case_ ,snake_case_ = self._get_tensors(5 )
snake_case_ = StoppingCriteriaList(
[
MaxLengthCriteria(max_length=10 ),
MaxTimeCriteria(max_time=0.1 ),
] )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ ,snake_case_ = self._get_tensors(9 )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ ,snake_case_ = self._get_tensors(10 )
self.assertTrue(criteria(lowercase_ , lowercase_ ) )
def A_ ( self : str ):
snake_case_ = MaxLengthCriteria(max_length=10 )
snake_case_ ,snake_case_ = self._get_tensors(5 )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ ,snake_case_ = self._get_tensors(9 )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ ,snake_case_ = self._get_tensors(10 )
self.assertTrue(criteria(lowercase_ , lowercase_ ) )
def A_ ( self : Optional[Any] ):
snake_case_ = MaxNewTokensCriteria(start_length=5 , max_new_tokens=5 )
snake_case_ ,snake_case_ = self._get_tensors(5 )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ ,snake_case_ = self._get_tensors(9 )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ ,snake_case_ = self._get_tensors(10 )
self.assertTrue(criteria(lowercase_ , lowercase_ ) )
snake_case_ = StoppingCriteriaList([criteria] )
self.assertEqual(criteria_list.max_length , 10 )
def A_ ( self : List[Any] ):
snake_case_ ,snake_case_ = self._get_tensors(5 )
snake_case_ = MaxTimeCriteria(max_time=0.1 )
self.assertFalse(criteria(lowercase_ , lowercase_ ) )
snake_case_ = MaxTimeCriteria(max_time=0.1 , initial_timestamp=time.time() - 0.2 )
self.assertTrue(criteria(lowercase_ , lowercase_ ) )
def A_ ( self : Dict ):
validate_stopping_criteria(StoppingCriteriaList([MaxLengthCriteria(10 )] ) , 10 )
with self.assertWarns(lowercase_ ):
validate_stopping_criteria(StoppingCriteriaList([MaxLengthCriteria(10 )] ) , 11 )
snake_case_ = validate_stopping_criteria(StoppingCriteriaList() , 11 )
self.assertEqual(len(lowercase_ ) , 1 )
| 56
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
a : Tuple = {
'configuration_llama': ['LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP', 'LlamaConfig'],
}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = ['LlamaTokenizer']
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : str = ['LlamaTokenizerFast']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = [
'LlamaForCausalLM',
'LlamaModel',
'LlamaPreTrainedModel',
'LlamaForSequenceClassification',
]
if TYPE_CHECKING:
from .configuration_llama import LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP, LlamaConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama import LlamaTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama_fast import LlamaTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_llama import LlamaForCausalLM, LlamaForSequenceClassification, LlamaModel, LlamaPreTrainedModel
else:
import sys
a : str = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> list:
'''simple docstring'''
snake_case_ = len(__UpperCAmelCase )
snake_case_ = []
for i in range(len(__UpperCAmelCase ) - pat_len + 1 ):
snake_case_ = True
for j in range(__UpperCAmelCase ):
if s[i + j] != pattern[j]:
snake_case_ = False
break
if match_found:
position.append(__UpperCAmelCase )
return position
if __name__ == "__main__":
assert naive_pattern_search('ABCDEFG', 'DE') == [3]
print(naive_pattern_search('ABAAABCDBBABCDDEBCABC', 'ABC'))
| 56
|
'''simple docstring'''
import re
from typing import Callable, List, Optional, Union
import tensorflow as tf
try:
from tensorflow.keras.optimizers.legacy import Adam
except ImportError:
from tensorflow.keras.optimizers import Adam
class a ( tf.keras.optimizers.schedules.LearningRateSchedule ):
def __init__( self : Optional[Any] , lowercase_ : float , lowercase_ : Callable , lowercase_ : int , lowercase_ : float = 1.0 , lowercase_ : str = None , ):
super().__init__()
snake_case_ = initial_learning_rate
snake_case_ = warmup_steps
snake_case_ = power
snake_case_ = decay_schedule_fn
snake_case_ = name
def __call__( self : Tuple , lowercase_ : str ):
with tf.name_scope(self.name or '''WarmUp''' ) as name:
# Implements polynomial warmup. i.e., if global_step < warmup_steps, the
# learning rate will be `global_step/num_warmup_steps * init_lr`.
snake_case_ = tf.cast(lowercase_ , tf.floataa )
snake_case_ = tf.cast(self.warmup_steps , tf.floataa )
snake_case_ = global_step_float / warmup_steps_float
snake_case_ = self.initial_learning_rate * tf.math.pow(lowercase_ , self.power )
return tf.cond(
global_step_float < warmup_steps_float , lambda: warmup_learning_rate , lambda: self.decay_schedule_fn(step - self.warmup_steps ) , name=lowercase_ , )
def A_ ( self : Any ):
return {
"initial_learning_rate": self.initial_learning_rate,
"decay_schedule_fn": self.decay_schedule_fn,
"warmup_steps": self.warmup_steps,
"power": self.power,
"name": self.name,
}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, __UpperCAmelCase = 0.9, __UpperCAmelCase = 0.9_9_9, __UpperCAmelCase = 1e-8, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = 0.0, __UpperCAmelCase = 1.0, __UpperCAmelCase = None, ) -> List[str]:
'''simple docstring'''
snake_case_ = tf.keras.optimizers.schedules.PolynomialDecay(
initial_learning_rate=__UpperCAmelCase, decay_steps=num_train_steps - num_warmup_steps, end_learning_rate=init_lr * min_lr_ratio, power=__UpperCAmelCase, )
if num_warmup_steps:
snake_case_ = WarmUp(
initial_learning_rate=__UpperCAmelCase, decay_schedule_fn=__UpperCAmelCase, warmup_steps=__UpperCAmelCase, )
if weight_decay_rate > 0.0:
snake_case_ = AdamWeightDecay(
learning_rate=__UpperCAmelCase, weight_decay_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, exclude_from_weight_decay=['''LayerNorm''', '''layer_norm''', '''bias'''], include_in_weight_decay=__UpperCAmelCase, )
else:
snake_case_ = tf.keras.optimizers.Adam(
learning_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, )
# We return the optimizer and the LR scheduler in order to better track the
# evolution of the LR independently of the optimizer.
return optimizer, lr_schedule
class a ( _lowerCamelCase ):
def __init__( self : Dict , lowercase_ : Union[float, tf.keras.optimizers.schedules.LearningRateSchedule] = 0.001 , lowercase_ : float = 0.9 , lowercase_ : float = 0.999 , lowercase_ : float = 1e-7 , lowercase_ : bool = False , lowercase_ : float = 0.0 , lowercase_ : Optional[List[str]] = None , lowercase_ : Optional[List[str]] = None , lowercase_ : str = "AdamWeightDecay" , **lowercase_ : Optional[int] , ):
super().__init__(lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
snake_case_ = weight_decay_rate
snake_case_ = include_in_weight_decay
snake_case_ = exclude_from_weight_decay
@classmethod
def A_ ( cls : Dict , lowercase_ : Union[str, Any] ):
snake_case_ = {'''WarmUp''': WarmUp}
return super(lowercase_ , cls ).from_config(lowercase_ , custom_objects=lowercase_ )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : Optional[int] ):
super(lowercase_ , self )._prepare_local(lowercase_ , lowercase_ , lowercase_ )
snake_case_ = tf.constant(
self.weight_decay_rate , name='''adam_weight_decay_rate''' )
def A_ ( self : str , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Any ):
snake_case_ = self._do_use_weight_decay(var.name )
if do_decay:
return var.assign_sub(
learning_rate * var * apply_state[(var.device, var.dtype.base_dtype)]['''weight_decay_rate'''] , use_locking=self._use_locking , )
return tf.no_op()
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : str=None , **lowercase_ : List[str] ):
snake_case_ ,snake_case_ = list(zip(*lowercase_ ) )
return super(lowercase_ , self ).apply_gradients(zip(lowercase_ , lowercase_ ) , name=lowercase_ , **lowercase_ )
def A_ ( self : List[Any] , lowercase_ : str , lowercase_ : str , lowercase_ : Any ):
if apply_state is None:
return self._decayed_lr_t[var_dtype], {}
snake_case_ = apply_state or {}
snake_case_ = apply_state.get((var_device, var_dtype) )
if coefficients is None:
snake_case_ = self._fallback_apply_state(lowercase_ , lowercase_ )
snake_case_ = coefficients
return coefficients["lr_t"], {"apply_state": apply_state}
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : Union[str, Any] , lowercase_ : Optional[int]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_dense(lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : int , lowercase_ : str , lowercase_ : List[Any]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_sparse(lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = super().get_config()
config.update({'''weight_decay_rate''': self.weight_decay_rate} )
return config
def A_ ( self : Optional[int] , lowercase_ : int ):
if self.weight_decay_rate == 0:
return False
if self._include_in_weight_decay:
for r in self._include_in_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return True
if self._exclude_from_weight_decay:
for r in self._exclude_from_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return False
return True
class a ( _lowerCamelCase ):
def __init__( self : List[Any] ):
snake_case_ = []
snake_case_ = None
@property
def A_ ( self : Union[str, Any] ):
if self._accum_steps is None:
snake_case_ = tf.Variable(
tf.constant(0 , dtype=tf.intaa ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
return self._accum_steps.value()
@property
def A_ ( self : Dict ):
if not self._gradients:
raise ValueError('''The accumulator should be called first to initialize the gradients''' )
return [gradient.value() if gradient is not None else gradient for gradient in self._gradients]
def __call__( self : Any , lowercase_ : int ):
if not self._gradients:
snake_case_ = self.step # Create the step variable.
self._gradients.extend(
[
tf.Variable(
tf.zeros_like(lowercase_ ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
if gradient is not None
else gradient
for gradient in gradients
] )
if len(lowercase_ ) != len(self._gradients ):
raise ValueError(F"Expected {len(self._gradients )} gradients, but got {len(lowercase_ )}" )
for accum_gradient, gradient in zip(self._gradients , lowercase_ ):
if accum_gradient is not None and gradient is not None:
accum_gradient.assign_add(lowercase_ )
self._accum_steps.assign_add(1 )
def A_ ( self : Optional[int] ):
if not self._gradients:
return
self._accum_steps.assign(0 )
for gradient in self._gradients:
if gradient is not None:
gradient.assign(tf.zeros_like(lowercase_ ) )
| 56
| 1
|
'''simple docstring'''
from collections import deque
from math import floor
from random import random
from time import time
class a :
def __init__( self : str ):
snake_case_ = {}
def A_ ( self : Dict , lowercase_ : Tuple , lowercase_ : Optional[int] , lowercase_ : Optional[Any]=1 ):
if self.graph.get(lowercase_ ):
if self.graph[u].count([w, v] ) == 0:
self.graph[u].append([w, v] )
else:
snake_case_ = [[w, v]]
if not self.graph.get(lowercase_ ):
snake_case_ = []
def A_ ( self : Any ):
return list(self.graph )
def A_ ( self : Any , lowercase_ : Any , lowercase_ : Optional[Any] ):
if self.graph.get(lowercase_ ):
for _ in self.graph[u]:
if _[1] == v:
self.graph[u].remove(lowercase_ )
def A_ ( self : int , lowercase_ : Optional[int]=-2 , lowercase_ : Optional[int]=-1 ):
if s == d:
return []
snake_case_ = []
snake_case_ = []
if s == -2:
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = s
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if visited.count(node[1] ) < 1:
if node[1] == d:
visited.append(lowercase_ )
return visited
else:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
stack.pop()
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return visited
def A_ ( self : Optional[int] , lowercase_ : str=-1 ):
if c == -1:
snake_case_ = floor(random() * 1_0000 ) + 10
for i in range(lowercase_ ):
# every vertex has max 100 edges
for _ in range(floor(random() * 102 ) + 1 ):
snake_case_ = floor(random() * c ) + 1
if n != i:
self.add_pair(lowercase_ , lowercase_ , 1 )
def A_ ( self : Optional[int] , lowercase_ : Optional[Any]=-2 ):
snake_case_ = deque()
snake_case_ = []
if s == -2:
snake_case_ = list(self.graph )[0]
d.append(lowercase_ )
visited.append(lowercase_ )
while d:
snake_case_ = d.popleft()
if len(self.graph[s] ) != 0:
for node in self.graph[s]:
if visited.count(node[1] ) < 1:
d.append(node[1] )
visited.append(node[1] )
return visited
def A_ ( self : Optional[int] , lowercase_ : Optional[int] ):
snake_case_ = 0
for x in self.graph:
for y in self.graph[x]:
if y[1] == u:
count += 1
return count
def A_ ( self : List[Any] , lowercase_ : Optional[Any] ):
return len(self.graph[u] )
def A_ ( self : Any , lowercase_ : List[Any]=-2 ):
snake_case_ = []
snake_case_ = []
if s == -2:
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = s
snake_case_ = []
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if visited.count(node[1] ) < 1:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
sorted_nodes.append(stack.pop() )
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return sorted_nodes
def A_ ( self : Tuple ):
snake_case_ = []
snake_case_ = []
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = -2
snake_case_ = []
snake_case_ = s
snake_case_ = False
snake_case_ = set()
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if (
visited.count(node[1] ) > 0
and node[1] != parent
and indirect_parents.count(node[1] ) > 0
and not on_the_way_back
):
snake_case_ = len(lowercase_ ) - 1
while len_stack >= 0:
if stack[len_stack] == node[1]:
anticipating_nodes.add(node[1] )
break
else:
anticipating_nodes.add(stack[len_stack] )
len_stack -= 1
if visited.count(node[1] ) < 1:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
stack.pop()
snake_case_ = True
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = False
indirect_parents.append(lowercase_ )
snake_case_ = s
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return list(lowercase_ )
def A_ ( self : Dict ):
snake_case_ = []
snake_case_ = []
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = -2
snake_case_ = []
snake_case_ = s
snake_case_ = False
snake_case_ = set()
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if (
visited.count(node[1] ) > 0
and node[1] != parent
and indirect_parents.count(node[1] ) > 0
and not on_the_way_back
):
snake_case_ = len(lowercase_ ) - 1
while len_stack_minus_one >= 0:
if stack[len_stack_minus_one] == node[1]:
anticipating_nodes.add(node[1] )
break
else:
return True
if visited.count(node[1] ) < 1:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
stack.pop()
snake_case_ = True
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = False
indirect_parents.append(lowercase_ )
snake_case_ = s
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return False
def A_ ( self : Optional[int] , lowercase_ : List[Any]=-2 , lowercase_ : str=-1 ):
snake_case_ = time()
self.dfs(lowercase_ , lowercase_ )
snake_case_ = time()
return end - begin
def A_ ( self : List[Any] , lowercase_ : List[str]=-2 ):
snake_case_ = time()
self.bfs(lowercase_ )
snake_case_ = time()
return end - begin
class a :
def __init__( self : List[Any] ):
snake_case_ = {}
def A_ ( self : int , lowercase_ : int , lowercase_ : Optional[int] , lowercase_ : List[Any]=1 ):
# check if the u exists
if self.graph.get(lowercase_ ):
# if there already is a edge
if self.graph[u].count([w, v] ) == 0:
self.graph[u].append([w, v] )
else:
# if u does not exist
snake_case_ = [[w, v]]
# add the other way
if self.graph.get(lowercase_ ):
# if there already is a edge
if self.graph[v].count([w, u] ) == 0:
self.graph[v].append([w, u] )
else:
# if u does not exist
snake_case_ = [[w, u]]
def A_ ( self : Union[str, Any] , lowercase_ : List[Any] , lowercase_ : int ):
if self.graph.get(lowercase_ ):
for _ in self.graph[u]:
if _[1] == v:
self.graph[u].remove(lowercase_ )
# the other way round
if self.graph.get(lowercase_ ):
for _ in self.graph[v]:
if _[1] == u:
self.graph[v].remove(lowercase_ )
def A_ ( self : List[Any] , lowercase_ : List[str]=-2 , lowercase_ : Tuple=-1 ):
if s == d:
return []
snake_case_ = []
snake_case_ = []
if s == -2:
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = s
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if visited.count(node[1] ) < 1:
if node[1] == d:
visited.append(lowercase_ )
return visited
else:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
stack.pop()
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return visited
def A_ ( self : List[Any] , lowercase_ : Optional[Any]=-1 ):
if c == -1:
snake_case_ = floor(random() * 1_0000 ) + 10
for i in range(lowercase_ ):
# every vertex has max 100 edges
for _ in range(floor(random() * 102 ) + 1 ):
snake_case_ = floor(random() * c ) + 1
if n != i:
self.add_pair(lowercase_ , lowercase_ , 1 )
def A_ ( self : Tuple , lowercase_ : Dict=-2 ):
snake_case_ = deque()
snake_case_ = []
if s == -2:
snake_case_ = list(self.graph )[0]
d.append(lowercase_ )
visited.append(lowercase_ )
while d:
snake_case_ = d.popleft()
if len(self.graph[s] ) != 0:
for node in self.graph[s]:
if visited.count(node[1] ) < 1:
d.append(node[1] )
visited.append(node[1] )
return visited
def A_ ( self : List[str] , lowercase_ : Tuple ):
return len(self.graph[u] )
def A_ ( self : str ):
snake_case_ = []
snake_case_ = []
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = -2
snake_case_ = []
snake_case_ = s
snake_case_ = False
snake_case_ = set()
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if (
visited.count(node[1] ) > 0
and node[1] != parent
and indirect_parents.count(node[1] ) > 0
and not on_the_way_back
):
snake_case_ = len(lowercase_ ) - 1
while len_stack >= 0:
if stack[len_stack] == node[1]:
anticipating_nodes.add(node[1] )
break
else:
anticipating_nodes.add(stack[len_stack] )
len_stack -= 1
if visited.count(node[1] ) < 1:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
stack.pop()
snake_case_ = True
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = False
indirect_parents.append(lowercase_ )
snake_case_ = s
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return list(lowercase_ )
def A_ ( self : Any ):
snake_case_ = []
snake_case_ = []
snake_case_ = list(self.graph )[0]
stack.append(lowercase_ )
visited.append(lowercase_ )
snake_case_ = -2
snake_case_ = []
snake_case_ = s
snake_case_ = False
snake_case_ = set()
while True:
# check if there is any non isolated nodes
if len(self.graph[s] ) != 0:
snake_case_ = s
for node in self.graph[s]:
if (
visited.count(node[1] ) > 0
and node[1] != parent
and indirect_parents.count(node[1] ) > 0
and not on_the_way_back
):
snake_case_ = len(lowercase_ ) - 1
while len_stack_minus_one >= 0:
if stack[len_stack_minus_one] == node[1]:
anticipating_nodes.add(node[1] )
break
else:
return True
if visited.count(node[1] ) < 1:
stack.append(node[1] )
visited.append(node[1] )
snake_case_ = node[1]
break
# check if all the children are visited
if s == ss:
stack.pop()
snake_case_ = True
if len(lowercase_ ) != 0:
snake_case_ = stack[len(lowercase_ ) - 1]
else:
snake_case_ = False
indirect_parents.append(lowercase_ )
snake_case_ = s
snake_case_ = ss
# check if se have reached the starting point
if len(lowercase_ ) == 0:
return False
def A_ ( self : Dict ):
return list(self.graph )
def A_ ( self : List[Any] , lowercase_ : List[Any]=-2 , lowercase_ : Any=-1 ):
snake_case_ = time()
self.dfs(lowercase_ , lowercase_ )
snake_case_ = time()
return end - begin
def A_ ( self : Optional[Any] , lowercase_ : int=-2 ):
snake_case_ = time()
self.bfs(lowercase_ )
snake_case_ = time()
return end - begin
| 56
|
'''simple docstring'''
import gc
import unittest
import torch
from parameterized import parameterized
from diffusers import AutoencoderKL
from diffusers.utils import floats_tensor, load_hf_numpy, require_torch_gpu, slow, torch_all_close, torch_device
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = AutoencoderKL
snake_case_ = "sample"
snake_case_ = 1e-2
@property
def A_ ( self : Dict ):
snake_case_ = 4
snake_case_ = 3
snake_case_ = (32, 32)
snake_case_ = floats_tensor((batch_size, num_channels) + sizes ).to(lowercase_ )
return {"sample": image}
@property
def A_ ( self : List[Any] ):
return (3, 32, 32)
@property
def A_ ( self : Dict ):
return (3, 32, 32)
def A_ ( self : Union[str, Any] ):
snake_case_ = {
'''block_out_channels''': [32, 64],
'''in_channels''': 3,
'''out_channels''': 3,
'''down_block_types''': ['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''],
'''up_block_types''': ['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''],
'''latent_channels''': 4,
}
snake_case_ = self.dummy_input
return init_dict, inputs_dict
def A_ ( self : Any ):
pass
def A_ ( self : str ):
pass
@unittest.skipIf(torch_device == '''mps''' , '''Gradient checkpointing skipped on MPS''' )
def A_ ( self : Dict ):
# enable deterministic behavior for gradient checkpointing
snake_case_ ,snake_case_ = self.prepare_init_args_and_inputs_for_common()
snake_case_ = self.model_class(**lowercase_ )
model.to(lowercase_ )
assert not model.is_gradient_checkpointing and model.training
snake_case_ = model(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model.zero_grad()
snake_case_ = torch.randn_like(lowercase_ )
snake_case_ = (out - labels).mean()
loss.backward()
# re-instantiate the model now enabling gradient checkpointing
snake_case_ = self.model_class(**lowercase_ )
# clone model
model_a.load_state_dict(model.state_dict() )
model_a.to(lowercase_ )
model_a.enable_gradient_checkpointing()
assert model_a.is_gradient_checkpointing and model_a.training
snake_case_ = model_a(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model_a.zero_grad()
snake_case_ = (out_a - labels).mean()
loss_a.backward()
# compare the output and parameters gradients
self.assertTrue((loss - loss_a).abs() < 1e-5 )
snake_case_ = dict(model.named_parameters() )
snake_case_ = dict(model_a.named_parameters() )
for name, param in named_params.items():
self.assertTrue(torch_all_close(param.grad.data , named_params_a[name].grad.data , atol=5e-5 ) )
def A_ ( self : Tuple ):
snake_case_ ,snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' , output_loading_info=lowercase_ )
self.assertIsNotNone(lowercase_ )
self.assertEqual(len(loading_info['''missing_keys'''] ) , 0 )
model.to(lowercase_ )
snake_case_ = model(**self.dummy_input )
assert image is not None, "Make sure output is not None"
def A_ ( self : Tuple ):
snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' )
snake_case_ = model.to(lowercase_ )
model.eval()
if torch_device == "mps":
snake_case_ = torch.manual_seed(0 )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(0 )
snake_case_ = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ = image.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , sample_posterior=lowercase_ , generator=lowercase_ ).sample
snake_case_ = output[0, -1, -3:, -3:].flatten().cpu()
# Since the VAE Gaussian prior's generator is seeded on the appropriate device,
# the expected output slices are not the same for CPU and GPU.
if torch_device == "mps":
snake_case_ = torch.tensor(
[
-4.0_078e-01,
-3.8_323e-04,
-1.2_681e-01,
-1.1_462e-01,
2.0_095e-01,
1.0_893e-01,
-8.8_247e-02,
-3.0_361e-01,
-9.8_644e-03,
] )
elif torch_device == "cpu":
snake_case_ = torch.tensor(
[-0.1352, 0.0878, 0.0419, -0.0818, -0.1069, 0.0688, -0.1458, -0.4446, -0.0026] )
else:
snake_case_ = torch.tensor(
[-0.2421, 0.4642, 0.2507, -0.0438, 0.0682, 0.3160, -0.2018, -0.0727, 0.2485] )
self.assertTrue(torch_all_close(lowercase_ , lowercase_ , rtol=1e-2 ) )
@slow
class a ( unittest.TestCase ):
def A_ ( self : Union[str, Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] ):
return F"gaussian_noise_s={seed}_shape={'_'.join([str(lowercase_ ) for s in shape] )}.npy"
def A_ ( self : Any ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Dict , lowercase_ : List[Any]=0 , lowercase_ : Union[str, Any]=(4, 3, 512, 512) , lowercase_ : Optional[Any]=False ):
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = torch.from_numpy(load_hf_numpy(self.get_file_format(lowercase_ , lowercase_ ) ) ).to(lowercase_ ).to(lowercase_ )
return image
def A_ ( self : Any , lowercase_ : Dict="CompVis/stable-diffusion-v1-4" , lowercase_ : List[str]=False ):
snake_case_ = '''fp16''' if fpaa else None
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = AutoencoderKL.from_pretrained(
lowercase_ , subfolder='''vae''' , torch_dtype=lowercase_ , revision=lowercase_ , )
model.to(lowercase_ ).eval()
return model
def A_ ( self : Any , lowercase_ : int=0 ):
if torch_device == "mps":
return torch.manual_seed(lowercase_ )
return torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
@parameterized.expand(
[
# fmt: off
[33, [-0.1603, 0.9878, -0.0495, -0.0790, -0.2709, 0.8375, -0.2060, -0.0824], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2376, 0.1168, 0.1332, -0.4840, -0.2508, -0.0791, -0.0493, -0.4089], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Union[str, Any] , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[33, [-0.0513, 0.0289, 1.3799, 0.2166, -0.2573, -0.0871, 0.5103, -0.0999]],
[47, [-0.4128, -0.1320, -0.3704, 0.1965, -0.4116, -0.2332, -0.3340, 0.2247]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Optional[int] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , fpaa=lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.1609, 0.9866, -0.0487, -0.0777, -0.2716, 0.8368, -0.2055, -0.0814], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2377, 0.1147, 0.1333, -0.4841, -0.2506, -0.0805, -0.0491, -0.4085], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Tuple , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[13, [-0.2051, -0.1803, -0.2311, -0.2114, -0.3292, -0.3574, -0.2953, -0.3323]],
[37, [-0.2632, -0.2625, -0.2199, -0.2741, -0.4539, -0.4990, -0.3720, -0.4925]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Dict , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-3 )
@parameterized.expand(
[
# fmt: off
[27, [-0.0369, 0.0207, -0.0776, -0.0682, -0.1747, -0.1930, -0.1465, -0.2039]],
[16, [-0.1628, -0.2134, -0.2747, -0.2642, -0.3774, -0.4404, -0.3687, -0.4277]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Union[str, Any] , lowercase_ : Any , lowercase_ : Optional[Any] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=5e-3 )
@parameterized.expand([(13,), (16,), (27,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : List[str] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-1 )
@parameterized.expand([(13,), (16,), (37,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.3001, 0.0918, -2.6984, -3.9720, -3.2099, -5.0353, 1.7338, -0.2065, 3.4267]],
[47, [-1.5030, -4.3871, -6.0355, -9.1157, -1.6661, -2.7853, 2.1607, -5.0823, 2.5633]],
# fmt: on
] )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model.encode(lowercase_ ).latent_dist
snake_case_ = dist.sample(generator=lowercase_ )
assert list(sample.shape ) == [image.shape[0], 4] + [i // 8 for i in image.shape[2:]]
snake_case_ = sample[0, -1, -3:, -3:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
snake_case_ = 3e-3 if torch_device != '''mps''' else 1e-2
assert torch_all_close(lowercase_ , lowercase_ , atol=lowercase_ )
| 56
| 1
|
'''simple docstring'''
import copy
import os
from typing import TYPE_CHECKING, List, Union
if TYPE_CHECKING:
pass
from ...configuration_utils import PretrainedConfig
from ...utils import logging
a : str = logging.get_logger(__name__)
a : str = {
'kakaobrain/align-base': 'https://huggingface.co/kakaobrain/align-base/resolve/main/config.json',
}
class a ( _lowerCamelCase ):
snake_case_ = "align_text_model"
def __init__( self : Any , lowercase_ : str=3_0522 , lowercase_ : int=768 , lowercase_ : Tuple=12 , lowercase_ : int=12 , lowercase_ : Dict=3072 , lowercase_ : List[Any]="gelu" , lowercase_ : str=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Any=512 , lowercase_ : Dict=2 , lowercase_ : Optional[int]=0.02 , lowercase_ : Union[str, Any]=1e-12 , lowercase_ : Optional[int]=0 , lowercase_ : Union[str, Any]="absolute" , lowercase_ : Tuple=True , **lowercase_ : List[str] , ):
super().__init__(**lowercase_ )
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = hidden_act
snake_case_ = intermediate_size
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = position_embedding_type
snake_case_ = use_cache
snake_case_ = pad_token_id
@classmethod
def A_ ( cls : int , lowercase_ : Union[str, os.PathLike] , **lowercase_ : Tuple ):
cls._set_token_in_kwargs(lowercase_ )
snake_case_ ,snake_case_ = cls.get_config_dict(lowercase_ , **lowercase_ )
# get the text config dict if we are loading from AlignConfig
if config_dict.get('''model_type''' ) == "align":
snake_case_ = config_dict['''text_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 ( _lowerCamelCase ):
snake_case_ = "align_vision_model"
def __init__( self : Any , lowercase_ : int = 3 , lowercase_ : int = 600 , lowercase_ : float = 2.0 , lowercase_ : float = 3.1 , lowercase_ : int = 8 , lowercase_ : List[int] = [3, 3, 5, 3, 5, 5, 3] , lowercase_ : List[int] = [32, 16, 24, 40, 80, 112, 192] , lowercase_ : List[int] = [16, 24, 40, 80, 112, 192, 320] , lowercase_ : List[int] = [] , lowercase_ : List[int] = [1, 2, 2, 2, 1, 2, 1] , lowercase_ : List[int] = [1, 2, 2, 3, 3, 4, 1] , lowercase_ : List[int] = [1, 6, 6, 6, 6, 6, 6] , lowercase_ : float = 0.25 , lowercase_ : str = "swish" , lowercase_ : int = 2560 , lowercase_ : str = "mean" , lowercase_ : float = 0.02 , lowercase_ : float = 0.001 , lowercase_ : float = 0.99 , lowercase_ : float = 0.2 , **lowercase_ : Any , ):
super().__init__(**lowercase_ )
snake_case_ = num_channels
snake_case_ = image_size
snake_case_ = width_coefficient
snake_case_ = depth_coefficient
snake_case_ = depth_divisor
snake_case_ = kernel_sizes
snake_case_ = in_channels
snake_case_ = out_channels
snake_case_ = depthwise_padding
snake_case_ = strides
snake_case_ = num_block_repeats
snake_case_ = expand_ratios
snake_case_ = squeeze_expansion_ratio
snake_case_ = hidden_act
snake_case_ = hidden_dim
snake_case_ = pooling_type
snake_case_ = initializer_range
snake_case_ = batch_norm_eps
snake_case_ = batch_norm_momentum
snake_case_ = drop_connect_rate
snake_case_ = sum(lowercase_ ) * 4
@classmethod
def A_ ( cls : Optional[int] , lowercase_ : Union[str, os.PathLike] , **lowercase_ : Any ):
cls._set_token_in_kwargs(lowercase_ )
snake_case_ ,snake_case_ = cls.get_config_dict(lowercase_ , **lowercase_ )
# get the vision config dict if we are loading from AlignConfig
if config_dict.get('''model_type''' ) == "align":
snake_case_ = 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 ( _lowerCamelCase ):
snake_case_ = "align"
snake_case_ = True
def __init__( self : List[Any] , lowercase_ : int=None , lowercase_ : List[Any]=None , lowercase_ : Dict=640 , lowercase_ : Optional[int]=1.0 , lowercase_ : Any=0.02 , **lowercase_ : Optional[int] , ):
super().__init__(**lowercase_ )
if text_config is None:
snake_case_ = {}
logger.info('''text_config is None. Initializing the AlignTextConfig with default values.''' )
if vision_config is None:
snake_case_ = {}
logger.info('''vision_config is None. Initializing the AlignVisionConfig with default values.''' )
snake_case_ = AlignTextConfig(**lowercase_ )
snake_case_ = AlignVisionConfig(**lowercase_ )
snake_case_ = projection_dim
snake_case_ = temperature_init_value
snake_case_ = initializer_range
@classmethod
def A_ ( cls : Optional[int] , lowercase_ : AlignTextConfig , lowercase_ : AlignVisionConfig , **lowercase_ : List[str] ):
return cls(text_config=text_config.to_dict() , vision_config=vision_config.to_dict() , **lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = copy.deepcopy(self.__dict__ )
snake_case_ = self.text_config.to_dict()
snake_case_ = self.vision_config.to_dict()
snake_case_ = self.__class__.model_type
return output
| 56
|
'''simple docstring'''
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 ( _lowerCamelCase ):
snake_case_ = 42
@flax_register_to_config
class a ( nn.Module , _lowerCamelCase , _lowerCamelCase ):
snake_case_ = 32
snake_case_ = 4
snake_case_ = 4
snake_case_ = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
snake_case_ = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
snake_case_ = False
snake_case_ = (320, 640, 1_280, 1_280)
snake_case_ = 2
snake_case_ = 8
snake_case_ = None
snake_case_ = 1_280
snake_case_ = 0.0
snake_case_ = False
snake_case_ = jnp.floataa
snake_case_ = True
snake_case_ = 0
snake_case_ = False
def A_ ( self : Optional[int] , lowercase_ : jax.random.KeyArray ):
# init input tensors
snake_case_ = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case_ = jnp.zeros(lowercase_ , dtype=jnp.floataa )
snake_case_ = jnp.ones((1,) , dtype=jnp.intaa )
snake_case_ = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case_ ,snake_case_ = jax.random.split(lowercase_ )
snake_case_ = {'''params''': params_rng, '''dropout''': dropout_rng}
return self.init(lowercase_ , lowercase_ , lowercase_ , lowercase_ )["params"]
def A_ ( self : List[str] ):
snake_case_ = self.block_out_channels
snake_case_ = 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.
snake_case_ = self.num_attention_heads or self.attention_head_dim
# input
snake_case_ = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case_ = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case_ = FlaxTimestepEmbedding(lowercase_ , dtype=self.dtype )
snake_case_ = self.only_cross_attention
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (only_cross_attention,) * len(self.down_block_types )
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case_ = []
snake_case_ = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case_ = output_channel
snake_case_ = block_out_channels[i]
snake_case_ = i == len(lowercase_ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = down_blocks
# mid
snake_case_ = 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
snake_case_ = []
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case_ = output_channel
snake_case_ = reversed_block_out_channels[i]
snake_case_ = reversed_block_out_channels[min(i + 1 , len(lowercase_ ) - 1 )]
snake_case_ = i == len(lowercase_ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = output_channel
snake_case_ = up_blocks
# out
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : int , lowercase_ : Any , lowercase_ : int=None , lowercase_ : Any=None , lowercase_ : bool = True , lowercase_ : bool = False , ):
# 1. time
if not isinstance(lowercase_ , jnp.ndarray ):
snake_case_ = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(lowercase_ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case_ = timesteps.astype(dtype=jnp.floataa )
snake_case_ = jnp.expand_dims(lowercase_ , 0 )
snake_case_ = self.time_proj(lowercase_ )
snake_case_ = self.time_embedding(lowercase_ )
# 2. pre-process
snake_case_ = jnp.transpose(lowercase_ , (0, 2, 3, 1) )
snake_case_ = self.conv_in(lowercase_ )
# 3. down
snake_case_ = (sample,)
for down_block in self.down_blocks:
if isinstance(lowercase_ , lowercase_ ):
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , lowercase_ , deterministic=not train )
else:
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case_ = ()
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,)
snake_case_ = new_down_block_res_samples
# 4. mid
snake_case_ = 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:
snake_case_ = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case_ = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = up_block(
lowercase_ , temb=lowercase_ , encoder_hidden_states=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train , )
else:
snake_case_ = up_block(lowercase_ , temb=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train )
# 6. post-process
snake_case_ = self.conv_norm_out(lowercase_ )
snake_case_ = nn.silu(lowercase_ )
snake_case_ = self.conv_out(lowercase_ )
snake_case_ = jnp.transpose(lowercase_ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=lowercase_ )
| 56
| 1
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class a ( metaclass=_lowerCamelCase ):
snake_case_ = ["transformers", "torch", "note_seq"]
def __init__( self : Union[str, Any] , *lowercase_ : Optional[int] , **lowercase_ : int ):
requires_backends(self , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : List[Any] , *lowercase_ : Optional[Any] , **lowercase_ : str ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : Tuple , *lowercase_ : Union[str, Any] , **lowercase_ : List[Any] ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
| 56
|
'''simple docstring'''
import glob
import os
import random
from string import ascii_lowercase, digits
import cva
import numpy as np
# Parrameters
a : Dict = (720, 1280) # Height, Width
a : Tuple = (0.4, 0.6) # if height or width lower than this scale, drop it.
a : Dict = 1 / 100
a : str = ''
a : Any = ''
a : Optional[int] = ''
a : List[str] = 250
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dataset(__UpperCAmelCase, __UpperCAmelCase )
for index in range(__UpperCAmelCase ):
snake_case_ = random.sample(range(len(__UpperCAmelCase ) ), 4 )
snake_case_ ,snake_case_ ,snake_case_ = update_image_and_anno(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, filter_scale=__UpperCAmelCase, )
# Get random string code: '7b7ad245cdff75241935e4dd860f3bad'
snake_case_ = random_chars(32 )
snake_case_ = path.split(os.sep )[-1].rsplit('''.''', 1 )[0]
snake_case_ = F"{OUTPUT_DIR}/{file_name}_MOSAIC_{letter_code}"
cva.imwrite(F"{file_root}.jpg", __UpperCAmelCase, [cva.IMWRITE_JPEG_QUALITY, 85] )
print(F"Succeeded {index+1}/{NUMBER_IMAGES} with {file_name}" )
snake_case_ = []
for anno in new_annos:
snake_case_ = anno[3] - anno[1]
snake_case_ = anno[4] - anno[2]
snake_case_ = anno[1] + width / 2
snake_case_ = anno[2] + height / 2
snake_case_ = F"{anno[0]} {x_center} {y_center} {width} {height}"
annos_list.append(__UpperCAmelCase )
with open(F"{file_root}.txt", '''w''' ) as outfile:
outfile.write('''\n'''.join(line for line in annos_list ) )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> tuple[list, list]:
'''simple docstring'''
snake_case_ = []
snake_case_ = []
for label_file in glob.glob(os.path.join(__UpperCAmelCase, '''*.txt''' ) ):
snake_case_ = label_file.split(os.sep )[-1].rsplit('''.''', 1 )[0]
with open(__UpperCAmelCase ) as in_file:
snake_case_ = in_file.readlines()
snake_case_ = os.path.join(__UpperCAmelCase, F"{label_name}.jpg" )
snake_case_ = []
for obj_list in obj_lists:
snake_case_ = obj_list.rstrip('''\n''' ).split(''' ''' )
snake_case_ = float(obj[1] ) - float(obj[3] ) / 2
snake_case_ = float(obj[2] ) - float(obj[4] ) / 2
snake_case_ = float(obj[1] ) + float(obj[3] ) / 2
snake_case_ = float(obj[2] ) + float(obj[4] ) / 2
boxes.append([int(obj[0] ), xmin, ymin, xmax, ymax] )
if not boxes:
continue
img_paths.append(__UpperCAmelCase )
labels.append(__UpperCAmelCase )
return img_paths, labels
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, ) -> tuple[list, list, str]:
'''simple docstring'''
snake_case_ = np.zeros([output_size[0], output_size[1], 3], dtype=np.uinta )
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = int(scale_x * output_size[1] )
snake_case_ = int(scale_y * output_size[0] )
snake_case_ = []
snake_case_ = []
for i, index in enumerate(__UpperCAmelCase ):
snake_case_ = all_img_list[index]
path_list.append(__UpperCAmelCase )
snake_case_ = all_annos[index]
snake_case_ = cva.imread(__UpperCAmelCase )
if i == 0: # top-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = bbox[2] * scale_y
snake_case_ = bbox[3] * scale_x
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 1: # top-right
snake_case_ = cva.resize(__UpperCAmelCase, (output_size[1] - divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = bbox[2] * scale_y
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 2: # bottom-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = bbox[3] * scale_x
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
else: # bottom-right
snake_case_ = cva.resize(
__UpperCAmelCase, (output_size[1] - divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
# Remove bounding box small than scale of filter
if filter_scale > 0:
snake_case_ = [
anno
for anno in new_anno
if filter_scale < (anno[3] - anno[1]) and filter_scale < (anno[4] - anno[2])
]
return output_img, new_anno, path_list[0]
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
assert number_char > 1, "The number of character should greater than 1"
snake_case_ = ascii_lowercase + digits
return "".join(random.choice(__UpperCAmelCase ) for _ in range(__UpperCAmelCase ) )
if __name__ == "__main__":
main()
print('DONE ✅')
| 56
| 1
|
'''simple docstring'''
import argparse
import json
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
VideoMAEConfig,
VideoMAEForPreTraining,
VideoMAEForVideoClassification,
VideoMAEImageProcessor,
)
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = VideoMAEConfig()
set_architecture_configs(__UpperCAmelCase, __UpperCAmelCase )
if "finetuned" not in model_name:
snake_case_ = False
if "finetuned" in model_name:
snake_case_ = '''huggingface/label-files'''
if "kinetics" in model_name:
snake_case_ = 400
snake_case_ = '''kinetics400-id2label.json'''
elif "ssv2" in model_name:
snake_case_ = 174
snake_case_ = '''something-something-v2-id2label.json'''
else:
raise ValueError('''Model name should either contain \'kinetics\' or \'ssv2\' in case it\'s fine-tuned.''' )
snake_case_ = json.load(open(hf_hub_download(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = idalabel
snake_case_ = {v: k for k, v in idalabel.items()}
return config
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
if "small" in model_name:
snake_case_ = 384
snake_case_ = 1536
snake_case_ = 12
snake_case_ = 16
snake_case_ = 12
snake_case_ = 3
snake_case_ = 192
snake_case_ = 768
elif "large" in model_name:
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 24
snake_case_ = 16
snake_case_ = 12
snake_case_ = 8
snake_case_ = 512
snake_case_ = 2048
elif "huge" in model_name:
snake_case_ = 1280
snake_case_ = 5120
snake_case_ = 32
snake_case_ = 16
snake_case_ = 12
snake_case_ = 8
snake_case_ = 640
snake_case_ = 2560
elif "base" not in model_name:
raise ValueError('''Model name should include either "small", "base", "large", or "huge"''' )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if "encoder." in name:
snake_case_ = name.replace('''encoder.''', '''''' )
if "cls_token" in name:
snake_case_ = name.replace('''cls_token''', '''videomae.embeddings.cls_token''' )
if "decoder_pos_embed" in name:
snake_case_ = name.replace('''decoder_pos_embed''', '''decoder.decoder_pos_embed''' )
if "pos_embed" in name and "decoder" not in name:
snake_case_ = name.replace('''pos_embed''', '''videomae.embeddings.position_embeddings''' )
if "patch_embed.proj" in name:
snake_case_ = name.replace('''patch_embed.proj''', '''videomae.embeddings.patch_embeddings.projection''' )
if "patch_embed.norm" in name:
snake_case_ = name.replace('''patch_embed.norm''', '''videomae.embeddings.norm''' )
if "decoder.blocks" in name:
snake_case_ = name.replace('''decoder.blocks''', '''decoder.decoder_layers''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''videomae.encoder.layer''' )
if "attn.proj" in name:
snake_case_ = name.replace('''attn.proj''', '''attention.output.dense''' )
if "attn" in name and "bias" not in name:
snake_case_ = name.replace('''attn''', '''attention.self''' )
if "attn" in name:
snake_case_ = name.replace('''attn''', '''attention.attention''' )
if "norm1" in name:
snake_case_ = name.replace('''norm1''', '''layernorm_before''' )
if "norm2" in name:
snake_case_ = name.replace('''norm2''', '''layernorm_after''' )
if "mlp.fc1" in name:
snake_case_ = name.replace('''mlp.fc1''', '''intermediate.dense''' )
if "mlp.fc2" in name:
snake_case_ = name.replace('''mlp.fc2''', '''output.dense''' )
if "decoder_embed" in name:
snake_case_ = name.replace('''decoder_embed''', '''decoder.decoder_embed''' )
if "decoder_norm" in name:
snake_case_ = name.replace('''decoder_norm''', '''decoder.decoder_norm''' )
if "decoder_pred" in name:
snake_case_ = name.replace('''decoder_pred''', '''decoder.decoder_pred''' )
if "norm.weight" in name and "decoder" not in name and "fc" not in name:
snake_case_ = name.replace('''norm.weight''', '''videomae.layernorm.weight''' )
if "norm.bias" in name and "decoder" not in name and "fc" not in name:
snake_case_ = name.replace('''norm.bias''', '''videomae.layernorm.bias''' )
if "head" in name and "decoder" not in name:
snake_case_ = name.replace('''head''', '''classifier''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ = orig_state_dict.pop(__UpperCAmelCase )
if key.startswith('''encoder.''' ):
snake_case_ = key.replace('''encoder.''', '''''' )
if "qkv" in key:
snake_case_ = key.split('''.''' )
if key.startswith('''decoder.blocks''' ):
snake_case_ = config.decoder_hidden_size
snake_case_ = int(key_split[2] )
snake_case_ = '''decoder.decoder_layers.'''
if "weight" in key:
snake_case_ = val[:dim, :]
snake_case_ = val[dim : dim * 2, :]
snake_case_ = val[-dim:, :]
else:
snake_case_ = config.hidden_size
snake_case_ = int(key_split[1] )
snake_case_ = '''videomae.encoder.layer.'''
if "weight" in key:
snake_case_ = val[:dim, :]
snake_case_ = val[dim : dim * 2, :]
snake_case_ = val[-dim:, :]
else:
snake_case_ = val
return orig_state_dict
def __magic_name__ ( ) -> List[Any]:
'''simple docstring'''
snake_case_ = hf_hub_download(
repo_id='''hf-internal-testing/spaghetti-video''', filename='''eating_spaghetti.npy''', repo_type='''dataset''' )
snake_case_ = np.load(__UpperCAmelCase )
return list(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = get_videomae_config(__UpperCAmelCase )
if "finetuned" in model_name:
snake_case_ = VideoMAEForVideoClassification(__UpperCAmelCase )
else:
snake_case_ = VideoMAEForPreTraining(__UpperCAmelCase )
# download original checkpoint, hosted on Google Drive
snake_case_ = '''pytorch_model.bin'''
gdown.cached_download(__UpperCAmelCase, __UpperCAmelCase, quiet=__UpperCAmelCase )
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )
if "model" in files:
snake_case_ = files['''model''']
else:
snake_case_ = files['''module''']
snake_case_ = convert_state_dict(__UpperCAmelCase, __UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
# verify model on basic input
snake_case_ = VideoMAEImageProcessor(image_mean=[0.5, 0.5, 0.5], image_std=[0.5, 0.5, 0.5] )
snake_case_ = prepare_video()
snake_case_ = image_processor(__UpperCAmelCase, return_tensors='''pt''' )
if "finetuned" not in model_name:
snake_case_ = hf_hub_download(repo_id='''hf-internal-testing/bool-masked-pos''', filename='''bool_masked_pos.pt''' )
snake_case_ = torch.load(__UpperCAmelCase )
snake_case_ = model(**__UpperCAmelCase )
snake_case_ = outputs.logits
snake_case_ = [
'''videomae-small-finetuned-kinetics''',
'''videomae-small-finetuned-ssv2''',
# Kinetics-400 checkpoints (short = pretrained only for 800 epochs instead of 1600)
'''videomae-base-short''',
'''videomae-base-short-finetuned-kinetics''',
'''videomae-base''',
'''videomae-base-finetuned-kinetics''',
'''videomae-large''',
'''videomae-large-finetuned-kinetics''',
'''videomae-huge-finetuned-kinetics''',
# Something-Something-v2 checkpoints (short = pretrained only for 800 epochs instead of 2400)
'''videomae-base-short-ssv2''',
'''videomae-base-short-finetuned-ssv2''',
'''videomae-base-ssv2''',
'''videomae-base-finetuned-ssv2''',
]
# NOTE: logits were tested with image_mean and image_std equal to [0.5, 0.5, 0.5] and [0.5, 0.5, 0.5]
if model_name == "videomae-small-finetuned-kinetics":
snake_case_ = torch.Size([1, 400] )
snake_case_ = torch.tensor([-0.9_2_9_1, -0.4_0_6_1, -0.9_3_0_7] )
elif model_name == "videomae-small-finetuned-ssv2":
snake_case_ = torch.Size([1, 174] )
snake_case_ = torch.tensor([0.2_6_7_1, -0.4_6_8_9, -0.8_2_3_5] )
elif model_name == "videomae-base":
snake_case_ = torch.Size([1, 1408, 1536] )
snake_case_ = torch.tensor([[0.7_7_3_9, 0.7_9_6_8, 0.7_0_8_9], [0.6_7_0_1, 0.7_4_8_7, 0.6_2_0_9], [0.4_2_8_7, 0.5_1_5_8, 0.4_7_7_3]] )
elif model_name == "videomae-base-short":
snake_case_ = torch.Size([1, 1408, 1536] )
snake_case_ = torch.tensor([[0.7_9_9_4, 0.9_6_1_2, 0.8_5_0_8], [0.7_4_0_1, 0.8_9_5_8, 0.8_3_0_2], [0.5_8_6_2, 0.7_4_6_8, 0.7_3_2_5]] )
# we verified the loss both for normalized and unnormalized targets for this one
snake_case_ = torch.tensor([0.5_1_4_2] ) if config.norm_pix_loss else torch.tensor([0.6_4_6_9] )
elif model_name == "videomae-large":
snake_case_ = torch.Size([1, 1408, 1536] )
snake_case_ = torch.tensor([[0.7_1_4_9, 0.7_9_9_7, 0.6_9_6_6], [0.6_7_6_8, 0.7_8_6_9, 0.6_9_4_8], [0.5_1_3_9, 0.6_2_2_1, 0.5_6_0_5]] )
elif model_name == "videomae-large-finetuned-kinetics":
snake_case_ = torch.Size([1, 400] )
snake_case_ = torch.tensor([0.0_7_7_1, 0.0_0_1_1, -0.3_6_2_5] )
elif model_name == "videomae-huge-finetuned-kinetics":
snake_case_ = torch.Size([1, 400] )
snake_case_ = torch.tensor([0.2_4_3_3, 0.1_6_3_2, -0.4_8_9_4] )
elif model_name == "videomae-base-short-finetuned-kinetics":
snake_case_ = torch.Size([1, 400] )
snake_case_ = torch.tensor([0.6_5_8_8, 0.0_9_9_0, -0.2_4_9_3] )
elif model_name == "videomae-base-finetuned-kinetics":
snake_case_ = torch.Size([1, 400] )
snake_case_ = torch.tensor([0.3_6_6_9, -0.0_6_8_8, -0.2_4_2_1] )
elif model_name == "videomae-base-short-ssv2":
snake_case_ = torch.Size([1, 1408, 1536] )
snake_case_ = torch.tensor([[0.4_7_1_2, 0.5_2_9_6, 0.5_7_8_6], [0.2_2_7_8, 0.2_7_2_9, 0.4_0_2_6], [0.0_3_5_2, 0.0_7_3_0, 0.2_5_0_6]] )
elif model_name == "videomae-base-short-finetuned-ssv2":
snake_case_ = torch.Size([1, 174] )
snake_case_ = torch.tensor([-0.0_5_3_7, -0.1_5_3_9, -0.3_2_6_6] )
elif model_name == "videomae-base-ssv2":
snake_case_ = torch.Size([1, 1408, 1536] )
snake_case_ = torch.tensor([[0.8_1_3_1, 0.8_7_2_7, 0.8_5_4_6], [0.7_3_6_6, 0.9_3_7_7, 0.8_8_7_0], [0.5_9_3_5, 0.8_8_7_4, 0.8_5_6_4]] )
elif model_name == "videomae-base-finetuned-ssv2":
snake_case_ = torch.Size([1, 174] )
snake_case_ = torch.tensor([0.1_9_6_1, -0.8_3_3_7, -0.6_3_8_9] )
else:
raise ValueError(F"Model name not supported. Should be one of {model_names}" )
# verify logits
assert logits.shape == expected_shape
if "finetuned" in model_name:
assert torch.allclose(logits[0, :3], __UpperCAmelCase, atol=1e-4 )
else:
print('''Logits:''', logits[0, :3, :3] )
assert torch.allclose(logits[0, :3, :3], __UpperCAmelCase, atol=1e-4 )
print('''Logits ok!''' )
# verify loss, if applicable
if model_name == "videomae-base-short":
snake_case_ = outputs.loss
assert torch.allclose(__UpperCAmelCase, __UpperCAmelCase, atol=1e-4 )
print('''Loss ok!''' )
if pytorch_dump_folder_path is not None:
print(F"Saving model and image processor to {pytorch_dump_folder_path}" )
image_processor.save_pretrained(__UpperCAmelCase )
model.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print('''Pushing to the hub...''' )
model.push_to_hub(__UpperCAmelCase, organization='''nielsr''' )
if __name__ == "__main__":
a : Optional[Any] = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--checkpoint_url',
default='https://drive.google.com/u/1/uc?id=1tEhLyskjb755TJ65ptsrafUG2llSwQE1&export=download&confirm=t&uuid=aa3276eb-fb7e-482a-adec-dc7171df14c4',
type=str,
help=(
'URL of the original PyTorch checkpoint (on Google Drive) you\'d like to convert. Should be a direct'
' download link.'
),
)
parser.add_argument(
'--pytorch_dump_folder_path',
default='/Users/nielsrogge/Documents/VideoMAE/Test',
type=str,
help='Path to the output PyTorch model directory.',
)
parser.add_argument('--model_name', default='videomae-base', type=str, help='Name of the model.')
parser.add_argument(
'--push_to_hub', action='store_true', help='Whether or not to push the converted model to the 🤗 hub.'
)
a : Union[str, Any] = parser.parse_args()
convert_videomae_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 56
|
'''simple docstring'''
import unittest
from transformers import (
MODEL_FOR_OBJECT_DETECTION_MAPPING,
AutoFeatureExtractor,
AutoModelForObjectDetection,
ObjectDetectionPipeline,
is_vision_available,
pipeline,
)
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_pytesseract,
require_tf,
require_timm,
require_torch,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_vision_available():
from PIL import Image
else:
class a :
@staticmethod
def A_ ( *lowercase_ : int , **lowercase_ : str ):
pass
@is_pipeline_test
@require_vision
@require_timm
@require_torch
class a ( unittest.TestCase ):
snake_case_ = MODEL_FOR_OBJECT_DETECTION_MAPPING
def A_ ( self : Any , lowercase_ : List[Any] , lowercase_ : Optional[int] , lowercase_ : List[str] ):
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , image_processor=lowercase_ )
return object_detector, ["./tests/fixtures/tests_samples/COCO/000000039769.png"]
def A_ ( self : Dict , lowercase_ : List[Any] , lowercase_ : int ):
snake_case_ = object_detector('''./tests/fixtures/tests_samples/COCO/000000039769.png''' , threshold=0.0 )
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
import datasets
snake_case_ = datasets.load_dataset('''hf-internal-testing/fixtures_image_utils''' , '''image''' , split='''test''' )
snake_case_ = [
Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' ),
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
# RGBA
dataset[0]['''file'''],
# LA
dataset[1]['''file'''],
# L
dataset[2]['''file'''],
]
snake_case_ = object_detector(lowercase_ , threshold=0.0 )
self.assertEqual(len(lowercase_ ) , len(lowercase_ ) )
for outputs in batch_outputs:
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
@require_tf
@unittest.skip('''Object detection not implemented in TF''' )
def A_ ( self : int ):
pass
@require_torch
def A_ ( self : Tuple ):
snake_case_ = '''hf-internal-testing/tiny-detr-mobilenetsv3'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=0.0 )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] , threshold=0.0 , )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
] , )
@require_torch
@slow
def A_ ( self : Optional[int] ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : Tuple ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : str ):
snake_case_ = 0.9985
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=lowercase_ )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
@require_torch
@require_pytesseract
@slow
def A_ ( self : Dict ):
snake_case_ = '''Narsil/layoutlmv3-finetuned-funsd'''
snake_case_ = 0.9993
snake_case_ = pipeline('''object-detection''' , model=lowercase_ , threshold=lowercase_ )
snake_case_ = object_detector(
'''https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
] , )
| 56
| 1
|
'''simple docstring'''
# tests directory-specific settings - this file is run automatically
# by pytest before any tests are run
import doctest
import sys
import warnings
from os.path import abspath, dirname, join
import _pytest
from transformers.testing_utils import HfDoctestModule, HfDocTestParser
# allow having multiple repository checkouts and not needing to remember to rerun
# 'pip install -e .[dev]' when switching between checkouts and running tests.
a : int = abspath(join(dirname(__file__), 'src'))
sys.path.insert(1, git_repo_path)
# silence FutureWarning warnings in tests since often we can't act on them until
# they become normal warnings - i.e. the tests still need to test the current functionality
warnings.simplefilter(action='ignore', category=FutureWarning)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
config.addinivalue_line(
'''markers''', '''is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested''' )
config.addinivalue_line(
'''markers''', '''is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested''' )
config.addinivalue_line('''markers''', '''is_pipeline_test: mark test to run only when pipelines are tested''' )
config.addinivalue_line('''markers''', '''is_staging_test: mark test to run only in the staging environment''' )
config.addinivalue_line('''markers''', '''accelerate_tests: mark test that require accelerate''' )
config.addinivalue_line('''markers''', '''tool_tests: mark the tool tests that are run on their specific schedule''' )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from transformers.testing_utils import pytest_addoption_shared
pytest_addoption_shared(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
from transformers.testing_utils import pytest_terminal_summary_main
snake_case_ = terminalreporter.config.getoption('''--make-reports''' )
if make_reports:
pytest_terminal_summary_main(__UpperCAmelCase, id=__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if exitstatus == 5:
snake_case_ = 0
# Doctest custom flag to ignore output.
a : Union[str, Any] = doctest.register_optionflag('IGNORE_RESULT')
a : Optional[int] = doctest.OutputChecker
class a ( _lowerCamelCase ):
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int] ):
if IGNORE_RESULT & optionflags:
return True
return OutputChecker.check_output(self , lowercase_ , lowercase_ , lowercase_ )
a : List[Any] = CustomOutputChecker
a : Optional[int] = HfDoctestModule
a : Tuple = HfDocTestParser
| 56
|
'''simple docstring'''
import unittest
from transformers import MPNetConfig, is_torch_available
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 (
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
class a :
def __init__( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Any=13 , lowercase_ : Optional[Any]=7 , lowercase_ : Optional[Any]=True , lowercase_ : Dict=True , lowercase_ : Tuple=False , lowercase_ : Optional[Any]=True , lowercase_ : Any=99 , lowercase_ : Union[str, Any]=64 , lowercase_ : str=5 , lowercase_ : int=4 , lowercase_ : List[Any]=64 , lowercase_ : Dict="gelu" , lowercase_ : Optional[int]=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Tuple=512 , lowercase_ : List[Any]=16 , lowercase_ : str=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[Any]=3 , lowercase_ : Optional[Any]=4 , lowercase_ : List[Any]=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_input_mask
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = scope
def A_ ( self : List[str] ):
return MPNetConfig.from_pretrained('''microsoft/mpnet-base''' )
def A_ ( self : str ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_input_mask:
snake_case_ = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = ids_tensor([self.batch_size] , self.num_choices )
snake_case_ = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def A_ ( self : Tuple ):
return MPNetConfig(
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 , initializer_range=self.initializer_range , )
def A_ ( self : Any , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Optional[int] ):
snake_case_ = MPNetModel(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , lowercase_ )
snake_case_ = 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 : str , lowercase_ : Optional[Any] , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = MPNetForQuestionAnswering(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Tuple , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : str , lowercase_ : Optional[int] , lowercase_ : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = MPNetForSequenceClassification(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , attention_mask=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : Any , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.num_choices
snake_case_ = MPNetForMultipleChoice(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = model(
lowercase_ , attention_mask=lowercase_ , labels=lowercase_ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def A_ ( self : Union[str, Any] , lowercase_ : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] , lowercase_ : str , lowercase_ : int , lowercase_ : List[str] ):
snake_case_ = self.num_labels
snake_case_ = MPNetForTokenClassification(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Union[str, Any] ):
snake_case_ = self.prepare_config_and_inputs()
((snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_)) = config_and_inputs
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
if is_torch_available()
else ()
)
snake_case_ = (
{
"feature-extraction": MPNetModel,
"fill-mask": MPNetForMaskedLM,
"question-answering": MPNetForQuestionAnswering,
"text-classification": MPNetForSequenceClassification,
"token-classification": MPNetForTokenClassification,
"zero-shot": MPNetForSequenceClassification,
}
if is_torch_available()
else {}
)
snake_case_ = False
snake_case_ = True
def A_ ( self : Tuple ):
snake_case_ = MPNetModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , hidden_size=37 )
def A_ ( self : Union[str, Any] ):
self.config_tester.run_common_tests()
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_model(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_sequence_classification(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_multiple_choice(*lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_token_classification(*lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_question_answering(*lowercase_ )
@require_torch
class a ( unittest.TestCase ):
@slow
def A_ ( self : List[Any] ):
snake_case_ = MPNetModel.from_pretrained('''microsoft/mpnet-base''' )
snake_case_ = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]] )
snake_case_ = model(lowercase_ )[0]
snake_case_ = torch.Size((1, 11, 768) )
self.assertEqual(output.shape , lowercase_ )
snake_case_ = torch.tensor(
[[[-0.0550, 0.1943, -0.0740], [-0.0562, 0.2211, -0.0579], [-0.0437, 0.3337, -0.0641]]] )
# compare the actual values for a slice.
self.assertTrue(torch.allclose(output[:, :3, :3] , lowercase_ , atol=1e-4 ) )
| 56
| 1
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
a : int = {
'configuration_altclip': [
'ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP',
'AltCLIPConfig',
'AltCLIPTextConfig',
'AltCLIPVisionConfig',
],
'processing_altclip': ['AltCLIPProcessor'],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : str = [
'ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST',
'AltCLIPPreTrainedModel',
'AltCLIPModel',
'AltCLIPTextModel',
'AltCLIPVisionModel',
]
if TYPE_CHECKING:
from .configuration_altclip import (
ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP,
AltCLIPConfig,
AltCLIPTextConfig,
AltCLIPVisionConfig,
)
from .processing_altclip import AltCLIPProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_altclip import (
ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST,
AltCLIPModel,
AltCLIPPreTrainedModel,
AltCLIPTextModel,
AltCLIPVisionModel,
)
else:
import sys
a : List[Any] = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
|
'''simple docstring'''
import json
import os
import pickle
import shutil
import tempfile
from unittest import TestCase
from unittest.mock import patch
import numpy as np
from datasets import Dataset
from transformers import is_faiss_available
from transformers.models.bart.configuration_bart import BartConfig
from transformers.models.bart.tokenization_bart import BartTokenizer
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES as DPR_VOCAB_FILES_NAMES
from transformers.models.dpr.configuration_dpr import DPRConfig
from transformers.models.dpr.tokenization_dpr import DPRContextEncoderTokenizer, DPRQuestionEncoderTokenizer
from transformers.models.rag.configuration_rag import RagConfig
from transformers.models.rag.retrieval_rag import CustomHFIndex, RagRetriever
from transformers.models.roberta.tokenization_roberta import VOCAB_FILES_NAMES as BART_VOCAB_FILES_NAMES
from transformers.testing_utils import require_faiss, require_sentencepiece, require_tokenizers, require_torch
if is_faiss_available():
import faiss
@require_faiss
class a ( _lowerCamelCase ):
def A_ ( self : str ):
snake_case_ = tempfile.mkdtemp()
snake_case_ = 8
# DPR tok
snake_case_ = [
'''[UNK]''',
'''[CLS]''',
'''[SEP]''',
'''[PAD]''',
'''[MASK]''',
'''want''',
'''##want''',
'''##ed''',
'''wa''',
'''un''',
'''runn''',
'''##ing''',
''',''',
'''low''',
'''lowest''',
]
snake_case_ = os.path.join(self.tmpdirname , '''dpr_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , DPR_VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
# BART tok
snake_case_ = [
'''l''',
'''o''',
'''w''',
'''e''',
'''r''',
'''s''',
'''t''',
'''i''',
'''d''',
'''n''',
'''\u0120''',
'''\u0120l''',
'''\u0120n''',
'''\u0120lo''',
'''\u0120low''',
'''er''',
'''\u0120lowest''',
'''\u0120newer''',
'''\u0120wider''',
'''<unk>''',
]
snake_case_ = dict(zip(lowercase_ , range(len(lowercase_ ) ) ) )
snake_case_ = ['''#version: 0.2''', '''\u0120 l''', '''\u0120l o''', '''\u0120lo w''', '''e r''', '''''']
snake_case_ = {'''unk_token''': '''<unk>'''}
snake_case_ = os.path.join(self.tmpdirname , '''bart_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''merges_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write(json.dumps(lowercase_ ) + '''\n''' )
with open(self.merges_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write('''\n'''.join(lowercase_ ) )
def A_ ( self : Union[str, Any] ):
return DPRQuestionEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : Union[str, Any] ):
return DPRContextEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : int ):
return BartTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''bart_tokenizer''' ) )
def A_ ( self : str ):
shutil.rmtree(self.tmpdirname )
def A_ ( self : str ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size ), 2 * np.ones(self.retrieval_vector_size )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
return dataset
def A_ ( self : str ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , )
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
return retriever
def A_ ( self : str , lowercase_ : bool ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''custom''' , )
if from_disk:
snake_case_ = os.path.join(self.tmpdirname , '''dataset''' )
snake_case_ = os.path.join(self.tmpdirname , '''index.faiss''' )
dataset.get_index('''embeddings''' ).save(os.path.join(self.tmpdirname , '''index.faiss''' ) )
dataset.drop_index('''embeddings''' )
dataset.save_to_disk(os.path.join(self.tmpdirname , '''dataset''' ) )
del dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
else:
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , index=CustomHFIndex(config.retrieval_vector_size , lowercase_ ) , )
return retriever
def A_ ( self : Tuple ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size + 1 ), 2 * np.ones(self.retrieval_vector_size + 1 )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
snake_case_ = os.path.join(self.tmpdirname , '''hf_bert_base.hnswSQ8_correct_phi_128.c_index''' )
dataset.save_faiss_index('''embeddings''' , index_file_name + '''.index.dpr''' )
pickle.dump(dataset['''id'''] , open(index_file_name + '''.index_meta.dpr''' , '''wb''' ) )
snake_case_ = os.path.join(self.tmpdirname , '''psgs_w100.tsv.pkl''' )
snake_case_ = {sample['''id''']: [sample['''text'''], sample['''title''']] for sample in dataset}
pickle.dump(lowercase_ , open(lowercase_ , '''wb''' ) )
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''legacy''' , index_path=self.tmpdirname , )
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() )
return retriever
def A_ ( self : Optional[Any] ):
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : str ):
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = self.get_dummy_dataset()
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : int ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : str ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : Any ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : Any ):
snake_case_ = 1
snake_case_ = self.get_dummy_legacy_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''text'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''text'''][0] , '''bar''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''text'''][0] , '''foo''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_legacy_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : List[str] ):
import torch
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
snake_case_ ,snake_case_ ,snake_case_ = (
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , np.ndarray )
snake_case_ = retriever(
lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ , return_tensors='''pt''' , )
snake_case_ ,snake_case_ ,snake_case_ ,snake_case_ = ( # noqa: F841
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
out['''doc_ids'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : Tuple ):
snake_case_ = self.get_dpr_ctx_encoder_tokenizer()
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
retriever.set_ctx_encoder_tokenizer(lowercase_ )
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
self.assertEqual(
len(lowercase_ ) , 6 ) # check whether the retriever output consist of 6 attributes including tokenized docs
self.assertEqual(
all(k in out for k in ('''tokenized_doc_ids''', '''tokenized_doc_attention_mask''') ) , lowercase_ ) # check for doc token related keys in dictionary.
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
import unittest
from transformers import DebertaVaConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import (
TFDebertaVaForMaskedLM,
TFDebertaVaForQuestionAnswering,
TFDebertaVaForSequenceClassification,
TFDebertaVaForTokenClassification,
TFDebertaVaModel,
)
class a :
def __init__( self : Any , lowercase_ : Optional[int] , lowercase_ : List[Any]=13 , lowercase_ : Dict=7 , lowercase_ : Any=True , lowercase_ : str=True , lowercase_ : int=True , lowercase_ : Any=True , lowercase_ : Any=99 , lowercase_ : List[Any]=32 , lowercase_ : List[Any]=2 , lowercase_ : Optional[int]=4 , lowercase_ : Any=37 , lowercase_ : List[str]="gelu" , lowercase_ : Dict=0.1 , lowercase_ : List[Any]=0.1 , lowercase_ : List[str]=512 , lowercase_ : int=16 , lowercase_ : Dict=2 , lowercase_ : Tuple=0.02 , lowercase_ : Any=False , lowercase_ : int=True , lowercase_ : Union[str, Any]="None" , lowercase_ : str=3 , lowercase_ : Dict=4 , lowercase_ : Any=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_input_mask
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = relative_attention
snake_case_ = position_biased_input
snake_case_ = pos_att_type
snake_case_ = scope
def A_ ( self : Optional[Any] ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_input_mask:
snake_case_ = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ = None
if self.use_token_type_ids:
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = DebertaVaConfig(
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 , relative_attention=self.relative_attention , position_biased_input=self.position_biased_input , initializer_range=self.initializer_range , return_dict=lowercase_ , )
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def A_ ( self : Dict , lowercase_ : Any , lowercase_ : List[Any] , lowercase_ : List[str] , lowercase_ : Union[str, Any] , lowercase_ : Any , lowercase_ : Any , lowercase_ : Tuple ):
snake_case_ = TFDebertaVaModel(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = [input_ids, input_mask]
snake_case_ = model(lowercase_ )
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def A_ ( self : Dict , lowercase_ : Dict , lowercase_ : List[str] , lowercase_ : int , lowercase_ : List[Any] , lowercase_ : Optional[Any] , lowercase_ : List[Any] , lowercase_ : Any ):
snake_case_ = TFDebertaVaForMaskedLM(config=lowercase_ )
snake_case_ = {
'''input_ids''': input_ids,
'''attention_mask''': input_mask,
'''token_type_ids''': token_type_ids,
}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def A_ ( self : Optional[int] , lowercase_ : Optional[int] , lowercase_ : Optional[Any] , lowercase_ : List[str] , lowercase_ : Tuple , lowercase_ : Dict , lowercase_ : Optional[int] , lowercase_ : int ):
snake_case_ = self.num_labels
snake_case_ = TFDebertaVaForSequenceClassification(config=lowercase_ )
snake_case_ = {
'''input_ids''': input_ids,
'''attention_mask''': input_mask,
'''token_type_ids''': token_type_ids,
}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : List[str] , lowercase_ : Union[str, Any] , lowercase_ : List[str] , lowercase_ : Tuple , lowercase_ : Dict , lowercase_ : str , lowercase_ : List[Any] , lowercase_ : str ):
snake_case_ = self.num_labels
snake_case_ = TFDebertaVaForTokenClassification(config=lowercase_ )
snake_case_ = {
'''input_ids''': input_ids,
'''attention_mask''': input_mask,
'''token_type_ids''': token_type_ids,
}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def A_ ( self : Dict , lowercase_ : Optional[int] , lowercase_ : Any , lowercase_ : Optional[int] , lowercase_ : Any , lowercase_ : Dict , lowercase_ : Union[str, Any] , lowercase_ : Any ):
snake_case_ = TFDebertaVaForQuestionAnswering(config=lowercase_ )
snake_case_ = {
'''input_ids''': input_ids,
'''attention_mask''': input_mask,
'''token_type_ids''': token_type_ids,
}
snake_case_ = model(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 : Optional[int] ):
snake_case_ = self.prepare_config_and_inputs()
(
(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,
) = config_and_inputs
snake_case_ = {'''input_ids''': input_ids, '''token_type_ids''': token_type_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_tf
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
TFDebertaVaModel,
TFDebertaVaForMaskedLM,
TFDebertaVaForQuestionAnswering,
TFDebertaVaForSequenceClassification,
TFDebertaVaForTokenClassification,
)
if is_tf_available()
else ()
)
snake_case_ = (
{
"feature-extraction": TFDebertaVaModel,
"fill-mask": TFDebertaVaForMaskedLM,
"question-answering": TFDebertaVaForQuestionAnswering,
"text-classification": TFDebertaVaForSequenceClassification,
"token-classification": TFDebertaVaForTokenClassification,
"zero-shot": TFDebertaVaForSequenceClassification,
}
if is_tf_available()
else {}
)
snake_case_ = False
snake_case_ = False
def A_ ( self : Dict ):
snake_case_ = TFDebertaVaModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , hidden_size=37 )
def A_ ( self : Dict ):
self.config_tester.run_common_tests()
def A_ ( self : int ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*lowercase_ )
def A_ ( self : int ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*lowercase_ )
def A_ ( self : Optional[int] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*lowercase_ )
@slow
def A_ ( self : List[str] ):
snake_case_ = TFDebertaVaModel.from_pretrained('''kamalkraj/deberta-v2-xlarge''' )
self.assertIsNotNone(lowercase_ )
@require_tf
class a ( unittest.TestCase ):
@unittest.skip(reason='''Model not available yet''' )
def A_ ( self : Optional[Any] ):
pass
@slow
def A_ ( self : Any ):
snake_case_ = TFDebertaVaModel.from_pretrained('''kamalkraj/deberta-v2-xlarge''' )
snake_case_ = tf.constant([[0, 3_1414, 232, 328, 740, 1140, 1_2695, 69, 4_6078, 1588, 2]] )
snake_case_ = tf.constant([[0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]] )
snake_case_ = model(lowercase_ , attention_mask=lowercase_ )[0]
snake_case_ = tf.constant(
[[[0.2356, 0.1948, 0.0369], [-0.1063, 0.3586, -0.5152], [-0.6399, -0.0259, -0.2525]]] )
tf.debugging.assert_near(output[:, 1:4, 1:4] , lowercase_ , atol=1e-4 )
| 56
|
'''simple docstring'''
import os
import re
import warnings
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_ta import TaTokenizer
else:
a : Dict = None
a : List[Any] = logging.get_logger(__name__)
a : List[Any] = {'vocab_file': 'spiece.model', 'tokenizer_file': 'tokenizer.json'}
a : str = {
'vocab_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/spiece.model',
't5-base': 'https://huggingface.co/t5-base/resolve/main/spiece.model',
't5-large': 'https://huggingface.co/t5-large/resolve/main/spiece.model',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/spiece.model',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/spiece.model',
},
'tokenizer_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/tokenizer.json',
't5-base': 'https://huggingface.co/t5-base/resolve/main/tokenizer.json',
't5-large': 'https://huggingface.co/t5-large/resolve/main/tokenizer.json',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/tokenizer.json',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/tokenizer.json',
},
}
# TODO(PVP) - this should be removed in Transformers v5
a : List[Any] = {
't5-small': 512,
't5-base': 512,
't5-large': 512,
't5-3b': 512,
't5-11b': 512,
}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
snake_case_ = TaTokenizer
snake_case_ = []
def __init__( self : List[Any] , lowercase_ : int=None , lowercase_ : Dict=None , lowercase_ : Dict="</s>" , lowercase_ : List[Any]="<unk>" , lowercase_ : int="<pad>" , lowercase_ : int=100 , lowercase_ : List[Any]=None , **lowercase_ : List[str] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
snake_case_ = [F"<extra_id_{i}>" for i in range(lowercase_ )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra special tokens
snake_case_ = len(set(filter(lambda lowercase_ : bool('''extra_id_''' in str(lowercase_ ) ) , lowercase_ ) ) )
if extra_tokens != extra_ids:
raise ValueError(
F"Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are"
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
super().__init__(
lowercase_ , tokenizer_file=lowercase_ , eos_token=lowercase_ , unk_token=lowercase_ , pad_token=lowercase_ , extra_ids=lowercase_ , additional_special_tokens=lowercase_ , **lowercase_ , )
snake_case_ = vocab_file
snake_case_ = False if not self.vocab_file else True
snake_case_ = extra_ids
@staticmethod
def A_ ( lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : int ):
if pretrained_model_name_or_path in TaTokenizerFast.max_model_input_sizes:
snake_case_ = TaTokenizerFast.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
F" {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this"
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
F" {pretrained_model_name_or_path} automatically truncating your input to"
F" {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences"
F" longer than {deprecated_max_model_length} you can either instantiate this tokenizer with"
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , lowercase_ , )
return max_model_length
def A_ ( self : Optional[Any] , lowercase_ : str , lowercase_ : Optional[str] = None ):
if not self.can_save_slow_tokenizer:
raise ValueError(
'''Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '''
'''tokenizer.''' )
if not os.path.isdir(lowercase_ ):
logger.error(F"Vocabulary path ({save_directory}) should be a directory" )
return
snake_case_ = 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_ ):
copyfile(self.vocab_file , lowercase_ )
logger.info(F"Copy vocab file to {out_vocab_file}" )
return (out_vocab_file,)
def A_ ( self : Optional[int] , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = token_ids_a + [self.eos_token_id]
if token_ids_a is None:
return self.prefix_tokens + token_ids_a
else:
snake_case_ = token_ids_a + [self.eos_token_id]
return self.prefix_tokens + token_ids_a + token_ids_a
def A_ ( self : int , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def A_ ( self : Dict ):
return list(
set(filter(lambda lowercase_ : bool(re.search(R'''<extra_id_\d+>''' , lowercase_ ) ) is not None , self.additional_special_tokens ) ) )
def A_ ( self : Any ):
return [self.convert_tokens_to_ids(lowercase_ ) for token in self.get_sentinel_tokens()]
| 56
| 1
|
'''simple docstring'''
from typing import Dict
from transformers import EvalPrediction, HfArgumentParser, TrainingArguments, is_torch_available
from transformers.testing_utils import (
TestCasePlus,
execute_subprocess_async,
get_torch_dist_unique_port,
require_torch_multi_gpu,
require_torch_neuroncore,
)
from transformers.training_args import ParallelMode
from transformers.utils import logging
a : List[str] = logging.get_logger(__name__)
if is_torch_available():
import torch
from torch import nn
from torch.utils.data import Dataset
from transformers import Trainer
class a ( _lowerCamelCase ):
def __init__( self : List[Any] , lowercase_ : int = 101 ):
snake_case_ = length
def __len__( self : Optional[Any] ):
return self.length
def __getitem__( self : Any , lowercase_ : Any ):
return i
class a :
def __call__( self : Dict , lowercase_ : str ):
return {"input_ids": torch.tensor(lowercase_ ), "labels": torch.tensor(lowercase_ )}
class a ( nn.Module ):
def __init__( self : Any ):
super().__init__()
# Add some (unused) params otherwise DDP will complain.
snake_case_ = nn.Linear(120 , 80 )
def A_ ( self : Tuple , lowercase_ : List[str] , lowercase_ : List[Any]=None ):
if labels is not None:
return torch.tensor(0.0 , device=input_ids.device ), input_ids
else:
return input_ids
class a ( _lowerCamelCase ):
@require_torch_neuroncore
def A_ ( self : List[Any] ):
snake_case_ = F"--nproc_per_node=2\n --master_port={get_torch_dist_unique_port()}\n {self.test_file_dir}/test_trainer_distributed.py\n ".split()
snake_case_ = self.get_auto_remove_tmp_dir()
snake_case_ = F"--output_dir {output_dir}".split()
snake_case_ = ['''torchrun'''] + distributed_args + args
execute_subprocess_async(lowercase_ , env=self.get_env() )
# successful return here == success - any errors would have caused an error in the sub-call
class a ( _lowerCamelCase ):
@require_torch_multi_gpu
def A_ ( self : List[str] ):
snake_case_ = F"--nproc_per_node={torch.cuda.device_count()}\n --master_port={get_torch_dist_unique_port()}\n {self.test_file_dir}/test_trainer_distributed.py\n ".split()
snake_case_ = self.get_auto_remove_tmp_dir()
snake_case_ = F"--output_dir {output_dir}".split()
snake_case_ = ['''torchrun'''] + distributed_args + args
execute_subprocess_async(lowercase_ , env=self.get_env() )
# successful return here == success - any errors would have caused an error in the sub-call
if __name__ == "__main__":
# The script below is meant to be run under torch.distributed, on a machine with multiple GPUs:
#
# PYTHONPATH="src" python -m torch.distributed.run --nproc_per_node 2 --output_dir output_dir ./tests/test_trainer_distributed.py
a : List[Any] = HfArgumentParser((TrainingArguments,))
a : Any = parser.parse_args_into_dataclasses()[0]
logger.warning(
f'''Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}, '''
f'''distributed training: {training_args.parallel_mode != ParallelMode.NOT_DISTRIBUTED}'''
)
# Essentially, what we want to verify in the distributed case is that we get all samples back,
# in the right order. (this is crucial for prediction for instance)
for dataset_length in [101, 40, 7]:
a : List[Any] = DummyDataset(dataset_length)
def __magic_name__ ( __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = list(range(len(__UpperCAmelCase ) ) )
snake_case_ = p.predictions.tolist() == sequential and p.label_ids.tolist() == sequential
if not success and training_args.local_rank == 0:
logger.warning(
'''Predictions and/or labels do not match expected results:\n - predictions: '''
F"{p.predictions.tolist()}\n - labels: {p.label_ids.tolist()}\n - expected: {sequential}" )
return {"success": success}
a : List[str] = Trainer(
model=DummyModel(),
args=training_args,
data_collator=DummyDataCollator(),
eval_dataset=dataset,
compute_metrics=compute_metrics,
)
a : str = trainer.evaluate()
logger.info(metrics)
if metrics["eval_success"] is not True:
logger.error(metrics)
exit(1)
a : str = trainer.predict(dataset)
logger.info(p.metrics)
if p.metrics["test_success"] is not True:
logger.error(p.metrics)
exit(1)
a : Any = 2
a : int = trainer.evaluate()
logger.info(metrics)
if metrics["eval_success"] is not True:
logger.error(metrics)
exit(1)
a : Optional[Any] = trainer.predict(dataset)
logger.info(p.metrics)
if p.metrics["test_success"] is not True:
logger.error(p.metrics)
exit(1)
a : Tuple = None
| 56
|
'''simple docstring'''
from __future__ import annotations
import math
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
if depth < 0:
raise ValueError('''Depth cannot be less than 0''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Scores cannot be empty''' )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
return min(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ = [90, 23, 6, 33, 21, 65, 123, 3_4423]
snake_case_ = math.log(len(__UpperCAmelCase ), 2 )
print('''Optimal value : ''', end='''''' )
print(minimax(0, 0, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 56
| 1
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm import create_model
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import BitConfig, BitForImageClassification, BitImageProcessor
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
a : List[Any] = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = '''huggingface/label-files'''
snake_case_ = '''imagenet-1k-id2label.json'''
snake_case_ = json.load(open(hf_hub_download(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = {v: k for k, v in idalabel.items()}
snake_case_ = '''std_conv''' if '''bit''' in model_name else False
# note that when using BiT as backbone for ViT-hybrid checkpoints,
# one needs to additionally set config.layer_type = "bottleneck", config.stem_type = "same",
# config.conv_layer = "std_conv_same"
snake_case_ = BitConfig(
conv_layer=__UpperCAmelCase, num_labels=1000, idalabel=__UpperCAmelCase, labelaid=__UpperCAmelCase, )
return config
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if "stem.conv" in name:
snake_case_ = name.replace('''stem.conv''', '''bit.embedder.convolution''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''layers''' )
if "head.fc" in name:
snake_case_ = name.replace('''head.fc''', '''classifier.1''' )
if name.startswith('''norm''' ):
snake_case_ = '''bit.''' + name
if "bit" not in name and "classifier" not in name:
snake_case_ = '''bit.encoder.''' + name
return name
def __magic_name__ ( ) -> Tuple:
'''simple docstring'''
snake_case_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
snake_case_ = Image.open(requests.get(__UpperCAmelCase, stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase=False ) -> str:
'''simple docstring'''
snake_case_ = get_config(__UpperCAmelCase )
# load original model from timm
snake_case_ = create_model(__UpperCAmelCase, pretrained=__UpperCAmelCase )
timm_model.eval()
# load state_dict of original model
snake_case_ = timm_model.state_dict()
for key in state_dict.copy().keys():
snake_case_ = state_dict.pop(__UpperCAmelCase )
snake_case_ = val.squeeze() if '''head''' in key else val
# load HuggingFace model
snake_case_ = BitForImageClassification(__UpperCAmelCase )
model.eval()
model.load_state_dict(__UpperCAmelCase )
# create image processor
snake_case_ = create_transform(**resolve_data_config({}, model=__UpperCAmelCase ) )
snake_case_ = transform.transforms
snake_case_ = {
'''bilinear''': PILImageResampling.BILINEAR,
'''bicubic''': PILImageResampling.BICUBIC,
'''nearest''': PILImageResampling.NEAREST,
}
snake_case_ = BitImageProcessor(
do_resize=__UpperCAmelCase, size={'''shortest_edge''': timm_transforms[0].size}, resample=pillow_resamplings[timm_transforms[0].interpolation.value], do_center_crop=__UpperCAmelCase, crop_size={'''height''': timm_transforms[1].size[0], '''width''': timm_transforms[1].size[1]}, do_normalize=__UpperCAmelCase, image_mean=timm_transforms[-1].mean.tolist(), image_std=timm_transforms[-1].std.tolist(), )
snake_case_ = prepare_img()
snake_case_ = transform(__UpperCAmelCase ).unsqueeze(0 )
snake_case_ = processor(__UpperCAmelCase, return_tensors='''pt''' ).pixel_values
# verify pixel values
assert torch.allclose(__UpperCAmelCase, __UpperCAmelCase )
# verify logits
with torch.no_grad():
snake_case_ = model(__UpperCAmelCase )
snake_case_ = outputs.logits
print('''Logits:''', logits[0, :3] )
print('''Predicted class:''', model.config.idalabel[logits.argmax(-1 ).item()] )
snake_case_ = timm_model(__UpperCAmelCase )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(__UpperCAmelCase, outputs.logits, atol=1e-3 )
print('''Looks ok!''' )
if pytorch_dump_folder_path is not None:
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
print(F"Saving model {model_name} and processor to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print(F"Pushing model {model_name} and processor to the hub" )
model.push_to_hub(F"ybelkada/{model_name}" )
processor.push_to_hub(F"ybelkada/{model_name}" )
if __name__ == "__main__":
a : Optional[Any] = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--model_name',
default='resnetv2_50x1_bitm',
type=str,
help='Name of the BiT timm 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 to push the model to the hub.',
)
a : Any = parser.parse_args()
convert_bit_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 56
|
'''simple docstring'''
import argparse
import torch
from torch import nn
from transformers import SpeechaTextConfig, SpeechaTextForConditionalGeneration
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = [
'''encoder.version''',
'''decoder.version''',
'''model.encoder.version''',
'''model.decoder.version''',
'''decoder.output_projection.weight''',
'''_float_tensor''',
'''encoder.embed_positions._float_tensor''',
'''decoder.embed_positions._float_tensor''',
]
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = list(s_dict.keys() )
for key in keys:
if "transformer_layers" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
elif "subsample" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ ,snake_case_ = emb.weight.shape
snake_case_ = nn.Linear(__UpperCAmelCase, __UpperCAmelCase, bias=__UpperCAmelCase )
snake_case_ = emb.weight.data
return lin_layer
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )
snake_case_ = mam_aaa['''args''']
snake_case_ = mam_aaa['''model''']
snake_case_ = state_dict['''decoder.output_projection.weight''']
remove_ignore_keys_(__UpperCAmelCase )
rename_keys(__UpperCAmelCase )
snake_case_ = state_dict['''decoder.embed_tokens.weight'''].shape[0]
snake_case_ = args.share_decoder_input_output_embed
snake_case_ = [int(__UpperCAmelCase ) for i in args.conv_kernel_sizes.split(''',''' )]
snake_case_ = SpeechaTextConfig(
vocab_size=__UpperCAmelCase, max_source_positions=args.max_source_positions, max_target_positions=args.max_target_positions, encoder_layers=args.encoder_layers, decoder_layers=args.decoder_layers, encoder_attention_heads=args.encoder_attention_heads, decoder_attention_heads=args.decoder_attention_heads, encoder_ffn_dim=args.encoder_ffn_embed_dim, decoder_ffn_dim=args.decoder_ffn_embed_dim, d_model=args.encoder_embed_dim, dropout=args.dropout, attention_dropout=args.attention_dropout, activation_dropout=args.activation_dropout, activation_function='''relu''', num_conv_layers=len(__UpperCAmelCase ), conv_channels=args.conv_channels, conv_kernel_sizes=__UpperCAmelCase, input_feat_per_channel=args.input_feat_per_channel, input_channels=args.input_channels, tie_word_embeddings=__UpperCAmelCase, num_beams=5, max_length=200, use_cache=__UpperCAmelCase, decoder_start_token_id=2, early_stopping=__UpperCAmelCase, )
snake_case_ = SpeechaTextForConditionalGeneration(__UpperCAmelCase )
snake_case_ ,snake_case_ = model.model.load_state_dict(__UpperCAmelCase, strict=__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0 and not set(__UpperCAmelCase ) <= {
"encoder.embed_positions.weights",
"decoder.embed_positions.weights",
}:
raise ValueError(
'''Only `encoder.embed_positions.weights` and `decoder.embed_positions.weights` are allowed to be missing,'''
F" but all the following weights are missing {missing}" )
if tie_embeds:
snake_case_ = make_linear_from_emb(model.model.decoder.embed_tokens )
else:
snake_case_ = lm_head_weights
model.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
a : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument('--fairseq_path', type=str, help='Path to the fairseq model (.pt) file.')
parser.add_argument('--pytorch_dump_folder_path', default=None, type=str, help='Path to the output PyTorch model.')
a : List[Any] = parser.parse_args()
convert_fairseq_sat_checkpoint_to_tfms(args.fairseq_path, args.pytorch_dump_folder_path)
| 56
| 1
|
'''simple docstring'''
# A Bipartite Graph is a graph whose vertices can be divided into two independent sets,
# U and V such that every edge (u, v) either connects a vertex from U to V or a vertex
# from V to U. In other words, for every edge (u, v), either u belongs to U and v to V,
# or u belongs to V and v to U. We can also say that there is no edge that connects
# vertices of same set.
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = [False] * len(__UpperCAmelCase )
snake_case_ = [-1] * len(__UpperCAmelCase )
def dfs(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ = True
snake_case_ = c
for u in graph[v]:
if not visited[u]:
dfs(__UpperCAmelCase, 1 - c )
for i in range(len(__UpperCAmelCase ) ):
if not visited[i]:
dfs(__UpperCAmelCase, 0 )
for i in range(len(__UpperCAmelCase ) ):
for j in graph[i]:
if color[i] == color[j]:
return False
return True
# Adjacency list of graph
a : Tuple = {0: [1, 3], 1: [0, 2], 2: [1, 3], 3: [0, 2], 4: []}
print(check_bipartite_dfs(graph))
| 56
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class a ( metaclass=_lowerCamelCase ):
snake_case_ = ["transformers", "torch", "note_seq"]
def __init__( self : Union[str, Any] , *lowercase_ : Optional[int] , **lowercase_ : int ):
requires_backends(self , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : List[Any] , *lowercase_ : Optional[Any] , **lowercase_ : str ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : Tuple , *lowercase_ : Union[str, Any] , **lowercase_ : List[Any] ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
| 56
| 1
|
'''simple docstring'''
import json
from typing import List, Optional, Tuple
from tokenizers import pre_tokenizers, processors
from ...tokenization_utils_base import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import logging
from .tokenization_mvp import MvpTokenizer
a : List[str] = logging.get_logger(__name__)
a : str = {'vocab_file': 'vocab.json', 'merges_file': 'merges.txt', 'tokenizer_file': 'tokenizer.json'}
# See all MVP models at https://huggingface.co/models?filter=mvp
a : Union[str, Any] = {
'vocab_file': {
'RUCAIBox/mvp': 'https://huggingface.co/RUCAIBox/mvp/resolve/main/vocab.json',
},
'added_tokens.json': {
'RUCAIBox/mvp': 'https://huggingface.co/RUCAIBox/mvp/resolve/main/added_tokens.json',
},
'merges_file': {
'RUCAIBox/mvp': 'https://huggingface.co/RUCAIBox/mvp/resolve/main/merges.txt',
},
'tokenizer_file': {
'RUCAIBox/mvp': 'https://huggingface.co/RUCAIBox/mvp/resolve/main/tokenizer.json',
},
}
a : int = {
'RUCAIBox/mvp': 1024,
}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
snake_case_ = MvpTokenizer
def __init__( self : str , lowercase_ : Union[str, Any]=None , lowercase_ : Any=None , lowercase_ : int=None , lowercase_ : str="replace" , lowercase_ : Any="<s>" , lowercase_ : Dict="</s>" , lowercase_ : Any="</s>" , lowercase_ : Optional[Any]="<s>" , lowercase_ : Optional[Any]="<unk>" , lowercase_ : int="<pad>" , lowercase_ : Optional[int]="<mask>" , lowercase_ : Dict=False , lowercase_ : Any=True , **lowercase_ : Tuple , ):
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_ , )
snake_case_ = json.loads(self.backend_tokenizer.pre_tokenizer.__getstate__() )
if pre_tok_state.get('''add_prefix_space''' , lowercase_ ) != add_prefix_space:
snake_case_ = getattr(lowercase_ , pre_tok_state.pop('''type''' ) )
snake_case_ = add_prefix_space
snake_case_ = pre_tok_class(**lowercase_ )
snake_case_ = add_prefix_space
# the pre_tokenizer is already updated in the GPT2TokenizerFast `__init__`
snake_case_ = '''post_processor'''
snake_case_ = getattr(self.backend_tokenizer , lowercase_ , lowercase_ )
if tokenizer_component_instance:
snake_case_ = 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:
snake_case_ = tuple(state['''sep'''] )
if "cls" in state:
snake_case_ = tuple(state['''cls'''] )
snake_case_ = False
if state.get('''add_prefix_space''' , lowercase_ ) != add_prefix_space:
snake_case_ = add_prefix_space
snake_case_ = True
if state.get('''trim_offsets''' , lowercase_ ) != trim_offsets:
snake_case_ = trim_offsets
snake_case_ = True
if changes_to_apply:
snake_case_ = getattr(lowercase_ , state.pop('''type''' ) )
snake_case_ = component_class(**lowercase_ )
setattr(self.backend_tokenizer , lowercase_ , lowercase_ )
@property
def A_ ( self : Union[str, Any] ):
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 : Optional[Any] , lowercase_ : Any ):
snake_case_ = AddedToken(lowercase_ , lstrip=lowercase_ , rstrip=lowercase_ ) if isinstance(lowercase_ , lowercase_ ) else value
snake_case_ = value
def A_ ( self : Any , *lowercase_ : Optional[int] , **lowercase_ : str ):
snake_case_ = 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 : str , *lowercase_ : int , **lowercase_ : Dict ):
snake_case_ = 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 : Optional[int] , lowercase_ : str , lowercase_ : Optional[str] = None ):
snake_case_ = self._tokenizer.model.save(lowercase_ , name=lowercase_ )
return tuple(lowercase_ )
def A_ ( self : Optional[int] , lowercase_ : int , lowercase_ : Optional[Any]=None ):
snake_case_ = [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 : str , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.sep_token_id]
snake_case_ = [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]
| 56
|
'''simple docstring'''
# tests directory-specific settings - this file is run automatically
# by pytest before any tests are run
import doctest
import sys
import warnings
from os.path import abspath, dirname, join
import _pytest
from transformers.testing_utils import HfDoctestModule, HfDocTestParser
# allow having multiple repository checkouts and not needing to remember to rerun
# 'pip install -e .[dev]' when switching between checkouts and running tests.
a : int = abspath(join(dirname(__file__), 'src'))
sys.path.insert(1, git_repo_path)
# silence FutureWarning warnings in tests since often we can't act on them until
# they become normal warnings - i.e. the tests still need to test the current functionality
warnings.simplefilter(action='ignore', category=FutureWarning)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
config.addinivalue_line(
'''markers''', '''is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested''' )
config.addinivalue_line(
'''markers''', '''is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested''' )
config.addinivalue_line('''markers''', '''is_pipeline_test: mark test to run only when pipelines are tested''' )
config.addinivalue_line('''markers''', '''is_staging_test: mark test to run only in the staging environment''' )
config.addinivalue_line('''markers''', '''accelerate_tests: mark test that require accelerate''' )
config.addinivalue_line('''markers''', '''tool_tests: mark the tool tests that are run on their specific schedule''' )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from transformers.testing_utils import pytest_addoption_shared
pytest_addoption_shared(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
from transformers.testing_utils import pytest_terminal_summary_main
snake_case_ = terminalreporter.config.getoption('''--make-reports''' )
if make_reports:
pytest_terminal_summary_main(__UpperCAmelCase, id=__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if exitstatus == 5:
snake_case_ = 0
# Doctest custom flag to ignore output.
a : Union[str, Any] = doctest.register_optionflag('IGNORE_RESULT')
a : Optional[int] = doctest.OutputChecker
class a ( _lowerCamelCase ):
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int] ):
if IGNORE_RESULT & optionflags:
return True
return OutputChecker.check_output(self , lowercase_ , lowercase_ , lowercase_ )
a : List[Any] = CustomOutputChecker
a : Optional[int] = HfDoctestModule
a : Tuple = HfDocTestParser
| 56
| 1
|
'''simple docstring'''
from copy import deepcopy
from typing import Optional, Union
import numpy as np
from ...processing_utils import ProcessorMixin
from ...tokenization_utils_base import BatchEncoding
from ...utils import TensorType, is_tf_available, is_torch_available
if is_torch_available():
import torch
if is_tf_available():
import tensorflow as tf
class a ( _lowerCamelCase ):
snake_case_ = ["image_processor"]
snake_case_ = "SamImageProcessor"
def __init__( self : List[Any] , lowercase_ : List[Any] ):
super().__init__(lowercase_ )
snake_case_ = self.image_processor
snake_case_ = -10
snake_case_ = self.image_processor.size['''longest_edge''']
def __call__( self : List[str] , lowercase_ : Tuple=None , lowercase_ : Tuple=None , lowercase_ : Optional[int]=None , lowercase_ : Dict=None , lowercase_ : Optional[Union[str, TensorType]] = None , **lowercase_ : Any , ):
snake_case_ = self.image_processor(
lowercase_ , return_tensors=lowercase_ , **lowercase_ , )
# pop arguments that are not used in the foward but used nevertheless
snake_case_ = encoding_image_processor['''original_sizes''']
if hasattr(lowercase_ , '''numpy''' ): # Checks if Torch or TF tensor
snake_case_ = original_sizes.numpy()
snake_case_ ,snake_case_ ,snake_case_ = self._check_and_preprocess_points(
input_points=lowercase_ , input_labels=lowercase_ , input_boxes=lowercase_ , )
snake_case_ = self._normalize_and_convert(
lowercase_ , lowercase_ , input_points=lowercase_ , input_labels=lowercase_ , input_boxes=lowercase_ , return_tensors=lowercase_ , )
return encoding_image_processor
def A_ ( self : Dict , lowercase_ : Union[str, Any] , lowercase_ : str , lowercase_ : Optional[Any]=None , lowercase_ : Tuple=None , lowercase_ : Optional[Any]=None , lowercase_ : Optional[Any]="pt" , ):
if input_points is not None:
if len(lowercase_ ) != len(lowercase_ ):
snake_case_ = [
self._normalize_coordinates(self.target_size , lowercase_ , original_sizes[0] ) for point in input_points
]
else:
snake_case_ = [
self._normalize_coordinates(self.target_size , lowercase_ , lowercase_ )
for point, original_size in zip(lowercase_ , lowercase_ )
]
# check that all arrays have the same shape
if not all(point.shape == input_points[0].shape for point in input_points ):
if input_labels is not None:
snake_case_ ,snake_case_ = self._pad_points_and_labels(lowercase_ , lowercase_ )
snake_case_ = np.array(lowercase_ )
if input_labels is not None:
snake_case_ = np.array(lowercase_ )
if input_boxes is not None:
if len(lowercase_ ) != len(lowercase_ ):
snake_case_ = [
self._normalize_coordinates(self.target_size , lowercase_ , original_sizes[0] , is_bounding_box=lowercase_ )
for box in input_boxes
]
else:
snake_case_ = [
self._normalize_coordinates(self.target_size , lowercase_ , lowercase_ , is_bounding_box=lowercase_ )
for box, original_size in zip(lowercase_ , lowercase_ )
]
snake_case_ = np.array(lowercase_ )
if input_boxes is not None:
if return_tensors == "pt":
snake_case_ = torch.from_numpy(lowercase_ )
# boxes batch size of 1 by default
snake_case_ = input_boxes.unsqueeze(1 ) if len(input_boxes.shape ) != 3 else input_boxes
elif return_tensors == "tf":
snake_case_ = tf.convert_to_tensor(lowercase_ )
# boxes batch size of 1 by default
snake_case_ = tf.expand_dims(lowercase_ , 1 ) if len(input_boxes.shape ) != 3 else input_boxes
encoding_image_processor.update({'''input_boxes''': input_boxes} )
if input_points is not None:
if return_tensors == "pt":
snake_case_ = torch.from_numpy(lowercase_ )
# point batch size of 1 by default
snake_case_ = input_points.unsqueeze(1 ) if len(input_points.shape ) != 4 else input_points
elif return_tensors == "tf":
snake_case_ = tf.convert_to_tensor(lowercase_ )
# point batch size of 1 by default
snake_case_ = tf.expand_dims(lowercase_ , 1 ) if len(input_points.shape ) != 4 else input_points
encoding_image_processor.update({'''input_points''': input_points} )
if input_labels is not None:
if return_tensors == "pt":
snake_case_ = torch.from_numpy(lowercase_ )
# point batch size of 1 by default
snake_case_ = input_labels.unsqueeze(1 ) if len(input_labels.shape ) != 3 else input_labels
elif return_tensors == "tf":
snake_case_ = tf.convert_to_tensor(lowercase_ )
# point batch size of 1 by default
snake_case_ = tf.expand_dims(lowercase_ , 1 ) if len(input_labels.shape ) != 3 else input_labels
encoding_image_processor.update({'''input_labels''': input_labels} )
return encoding_image_processor
def A_ ( self : Union[str, Any] , lowercase_ : Any , lowercase_ : Dict ):
snake_case_ = max([point.shape[0] for point in input_points] )
snake_case_ = []
for i, point in enumerate(lowercase_ ):
if point.shape[0] != expected_nb_points:
snake_case_ = np.concatenate(
[point, np.zeros((expected_nb_points - point.shape[0], 2) ) + self.point_pad_value] , axis=0 )
snake_case_ = np.append(input_labels[i] , [self.point_pad_value] )
processed_input_points.append(lowercase_ )
snake_case_ = processed_input_points
return input_points, input_labels
def A_ ( self : int , lowercase_ : int , lowercase_ : np.ndarray , lowercase_ : List[str] , lowercase_ : Tuple=False ):
snake_case_ ,snake_case_ = original_size
snake_case_ ,snake_case_ = self.image_processor._get_preprocess_shape(lowercase_ , longest_edge=lowercase_ )
snake_case_ = deepcopy(lowercase_ ).astype(lowercase_ )
if is_bounding_box:
snake_case_ = coords.reshape(-1 , 2 , 2 )
snake_case_ = coords[..., 0] * (new_w / old_w)
snake_case_ = coords[..., 1] * (new_h / old_h)
if is_bounding_box:
snake_case_ = coords.reshape(-1 , 4 )
return coords
def A_ ( self : Any , lowercase_ : int=None , lowercase_ : int=None , lowercase_ : Optional[Any]=None , ):
if input_points is not None:
if hasattr(lowercase_ , '''numpy''' ): # Checks for TF or Torch tensor
snake_case_ = input_points.numpy().tolist()
if not isinstance(lowercase_ , lowercase_ ) or not isinstance(input_points[0] , lowercase_ ):
raise ValueError('''Input points must be a list of list of floating points.''' )
snake_case_ = [np.array(lowercase_ ) for input_point in input_points]
else:
snake_case_ = None
if input_labels is not None:
if hasattr(lowercase_ , '''numpy''' ):
snake_case_ = input_labels.numpy().tolist()
if not isinstance(lowercase_ , lowercase_ ) or not isinstance(input_labels[0] , lowercase_ ):
raise ValueError('''Input labels must be a list of list integers.''' )
snake_case_ = [np.array(lowercase_ ) for label in input_labels]
else:
snake_case_ = None
if input_boxes is not None:
if hasattr(lowercase_ , '''numpy''' ):
snake_case_ = input_boxes.numpy().tolist()
if (
not isinstance(lowercase_ , lowercase_ )
or not isinstance(input_boxes[0] , lowercase_ )
or not isinstance(input_boxes[0][0] , lowercase_ )
):
raise ValueError('''Input boxes must be a list of list of list of floating points.''' )
snake_case_ = [np.array(lowercase_ ).astype(np.floataa ) for box in input_boxes]
else:
snake_case_ = None
return input_points, input_labels, input_boxes
@property
def A_ ( self : Optional[int] ):
snake_case_ = self.image_processor.model_input_names
return list(dict.fromkeys(lowercase_ ) )
def A_ ( self : Optional[Any] , *lowercase_ : Optional[Any] , **lowercase_ : List[Any] ):
return self.image_processor.post_process_masks(*lowercase_ , **lowercase_ )
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import TensorType, is_torch_available, logging
a : Dict = logging.get_logger(__name__)
a : List[str] = {
'Helsinki-NLP/opus-mt-en-de': 'https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/config.json',
# See all Marian models at https://huggingface.co/models?filter=marian
}
class a ( _lowerCamelCase ):
snake_case_ = "marian"
snake_case_ = ["past_key_values"]
snake_case_ = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self : List[Any] , lowercase_ : Optional[Any]=5_8101 , lowercase_ : Dict=None , lowercase_ : List[str]=1024 , lowercase_ : Optional[Any]=12 , lowercase_ : int=4096 , lowercase_ : Any=16 , lowercase_ : Optional[int]=12 , lowercase_ : str=4096 , lowercase_ : Union[str, Any]=16 , lowercase_ : Dict=0.0 , lowercase_ : Union[str, Any]=0.0 , lowercase_ : Optional[Any]=True , lowercase_ : Union[str, Any]=True , lowercase_ : int="gelu" , lowercase_ : Dict=1024 , lowercase_ : int=0.1 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.02 , lowercase_ : int=5_8100 , lowercase_ : Optional[Any]=False , lowercase_ : Any=5_8100 , lowercase_ : Optional[int]=0 , lowercase_ : Dict=0 , lowercase_ : List[str]=True , **lowercase_ : Any , ):
snake_case_ = vocab_size
snake_case_ = decoder_vocab_size or vocab_size
snake_case_ = max_position_embeddings
snake_case_ = d_model
snake_case_ = encoder_ffn_dim
snake_case_ = encoder_layers
snake_case_ = encoder_attention_heads
snake_case_ = decoder_ffn_dim
snake_case_ = decoder_layers
snake_case_ = decoder_attention_heads
snake_case_ = dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = activation_function
snake_case_ = init_std
snake_case_ = encoder_layerdrop
snake_case_ = decoder_layerdrop
snake_case_ = use_cache
snake_case_ = encoder_layers
snake_case_ = scale_embedding # scale factor will be sqrt(d_model) if True
snake_case_ = share_encoder_decoder_embeddings
super().__init__(
pad_token_id=lowercase_ , eos_token_id=lowercase_ , is_encoder_decoder=lowercase_ , decoder_start_token_id=lowercase_ , forced_eos_token_id=lowercase_ , **lowercase_ , )
class a ( _lowerCamelCase ):
@property
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.inputs
def A_ ( self : Union[str, Any] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ = {0: '''batch'''}
snake_case_ = {0: '''batch''', 1: '''past_decoder_sequence + sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''decoder_sequence'''}
snake_case_ = {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.
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
else:
snake_case_ = 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
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.outputs
def A_ ( self : Dict ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super().outputs
else:
snake_case_ = super(lowercase_ , self ).outputs
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
return common_outputs
def A_ ( self : Dict , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
# Generate decoder inputs
snake_case_ = seq_length if not self.use_past else 1
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
snake_case_ = {F"decoder_{name}": tensor for name, tensor in decoder_inputs.items()}
snake_case_ = 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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
snake_case_ = common_inputs['''decoder_input_ids'''].shape[1]
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = decoder_seq_length + 3
snake_case_ = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
snake_case_ = torch.cat(
[common_inputs['''decoder_attention_mask'''], torch.ones(lowercase_ , lowercase_ )] , dim=1 )
snake_case_ = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
snake_case_ ,snake_case_ = self.num_layers
snake_case_ = min(lowercase_ , lowercase_ )
snake_case_ = max(lowercase_ , lowercase_ ) - min_num_layers
snake_case_ = '''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.
snake_case_ = 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 : Union[str, Any] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
# Not using the same length for past_key_values
snake_case_ = seqlen + 2
snake_case_ ,snake_case_ = self.num_layers
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = common_inputs['''attention_mask'''].dtype
snake_case_ = torch.cat(
[common_inputs['''attention_mask'''], torch.ones(lowercase_ , lowercase_ , dtype=lowercase_ )] , dim=1 )
snake_case_ = [
(torch.zeros(lowercase_ ), torch.zeros(lowercase_ )) for _ in range(lowercase_ )
]
return common_inputs
def A_ ( self : List[str] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
# 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
snake_case_ = 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
snake_case_ = tokenizer.num_special_tokens_to_add(lowercase_ )
snake_case_ = 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
snake_case_ = [''' '''.join([tokenizer.unk_token] ) * seq_length] * batch_size
snake_case_ = dict(tokenizer(lowercase_ , return_tensors=lowercase_ ) )
return common_inputs
def A_ ( self : Any , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
else:
snake_case_ = self._generate_dummy_inputs_for_causal_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
return common_inputs
def A_ ( self : Dict , lowercase_ : List[str] , lowercase_ : List[str] , lowercase_ : int , lowercase_ : List[str] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super()._flatten_past_key_values_(lowercase_ , lowercase_ , lowercase_ , lowercase_ )
else:
snake_case_ = super(lowercase_ , self )._flatten_past_key_values_(
lowercase_ , lowercase_ , lowercase_ , lowercase_ )
@property
def A_ ( self : List[str] ):
return 1e-4
| 56
| 1
|
'''simple docstring'''
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 : Optional[Any] , lowercase_ : Tuple , 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]]] = [0.4814_5466, 0.457_8275, 0.4082_1073] , lowercase_ : Optional[Union[float, List[float]]] = [0.2686_2954, 0.2613_0258, 0.2757_7711] , lowercase_ : bool = True , lowercase_ : Dict=7 , lowercase_ : Dict=30 , lowercase_ : List[str]=400 , lowercase_ : Optional[int]=3 , ):
snake_case_ = parent
snake_case_ = do_resize
snake_case_ = size if size is not None else {'''shortest_edge''': 288}
snake_case_ = size_divisor
snake_case_ = do_rescale
snake_case_ = rescale_factor
snake_case_ = do_normalize
snake_case_ = do_center_crop
snake_case_ = image_mean
snake_case_ = image_std
snake_case_ = do_pad
snake_case_ = batch_size
snake_case_ = num_channels
snake_case_ = min_resolution
snake_case_ = max_resolution
def A_ ( self : str ):
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 : Optional[int] , lowercase_ : Any , lowercase_ : List[Any]=False ):
if not batched:
snake_case_ = self.size['''shortest_edge''']
snake_case_ = image_inputs[0]
if isinstance(lowercase_ , Image.Image ):
snake_case_ ,snake_case_ = image.size
else:
snake_case_ ,snake_case_ = image.shape[1], image.shape[2]
snake_case_ = size / min(lowercase_ , lowercase_ )
if h < w:
snake_case_ ,snake_case_ = size, scale * w
else:
snake_case_ ,snake_case_ = scale * h, size
snake_case_ = int((1333 / 800) * size )
if max(lowercase_ , lowercase_ ) > max_size:
snake_case_ = max_size / max(lowercase_ , lowercase_ )
snake_case_ = newh * scale
snake_case_ = neww * scale
snake_case_ ,snake_case_ = int(newh + 0.5 ), int(neww + 0.5 )
snake_case_ ,snake_case_ = (
newh // self.size_divisor * self.size_divisor,
neww // self.size_divisor * self.size_divisor,
)
else:
snake_case_ = []
for image in image_inputs:
snake_case_ ,snake_case_ = self.get_expected_values([image] )
expected_values.append((expected_height, expected_width) )
snake_case_ = max(lowercase_ , key=lambda lowercase_ : item[0] )[0]
snake_case_ = max(lowercase_ , key=lambda lowercase_ : item[1] )[1]
return expected_height, expected_width
@require_torch
@require_vision
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = BridgeTowerImageProcessor if is_vision_available() else None
def A_ ( self : str ):
snake_case_ = BridgeTowerImageProcessingTester(self )
@property
def A_ ( self : Any ):
return self.image_processor_tester.prepare_image_processor_dict()
def A_ ( self : Union[str, Any] ):
snake_case_ = 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] ):
pass
def A_ ( self : List[Any] ):
# Initialize image processor
snake_case_ = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
snake_case_ = prepare_image_inputs(self.image_processor_tester , equal_resolution=lowercase_ )
for image in image_inputs:
self.assertIsInstance(lowercase_ , Image.Image )
# Test not batched input
snake_case_ = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
snake_case_ ,snake_case_ = 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
snake_case_ = image_processing(lowercase_ , return_tensors='''pt''' ).pixel_values
snake_case_ ,snake_case_ = 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 : Dict ):
# Initialize image processor
snake_case_ = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
snake_case_ = 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
snake_case_ = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
snake_case_ ,snake_case_ = 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
snake_case_ = image_processing(lowercase_ , return_tensors='''pt''' ).pixel_values
snake_case_ ,snake_case_ = 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 : Any ):
# Initialize image processor
snake_case_ = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
snake_case_ = 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
snake_case_ = image_processing(image_inputs[0] , return_tensors='''pt''' ).pixel_values
snake_case_ ,snake_case_ = 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
snake_case_ = image_processing(lowercase_ , return_tensors='''pt''' ).pixel_values
snake_case_ ,snake_case_ = 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,
) , )
| 56
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import AutoencoderKL, CycleDiffusionPipeline, DDIMScheduler, UNetaDConditionModel
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, skip_mps
from ..pipeline_params import (
IMAGE_TO_IMAGE_IMAGE_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = CycleDiffusionPipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {
"negative_prompt",
"height",
"width",
"negative_prompt_embeds",
}
snake_case_ = PipelineTesterMixin.required_optional_params - {"latents"}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"source_prompt"} )
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
def A_ ( self : Tuple ):
torch.manual_seed(0 )
snake_case_ = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=32 , )
snake_case_ = DDIMScheduler(
beta_start=0.0_0085 , beta_end=0.012 , beta_schedule='''scaled_linear''' , num_train_timesteps=1000 , clip_sample=lowercase_ , set_alpha_to_one=lowercase_ , )
torch.manual_seed(0 )
snake_case_ = 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 )
snake_case_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , )
snake_case_ = CLIPTextModel(lowercase_ )
snake_case_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
snake_case_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
'''safety_checker''': None,
'''feature_extractor''': None,
}
return components
def A_ ( self : Any , lowercase_ : int , lowercase_ : Optional[Any]=0 ):
snake_case_ = floats_tensor((1, 3, 32, 32) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
snake_case_ = image / 2 + 0.5
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''prompt''': '''An astronaut riding an elephant''',
'''source_prompt''': '''An astronaut riding a horse''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''eta''': 0.1,
'''strength''': 0.8,
'''guidance_scale''': 3,
'''source_guidance_scale''': 1,
'''output_type''': '''numpy''',
}
return inputs
def A_ ( self : Union[str, Any] ):
snake_case_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
snake_case_ = self.get_dummy_components()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.4459, 0.4943, 0.4544, 0.6643, 0.5474, 0.4327, 0.5701, 0.5959, 0.5179] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.get_dummy_components()
for name, module in components.items():
if hasattr(lowercase_ , '''half''' ):
snake_case_ = module.half()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.3506, 0.4543, 0.446, 0.4575, 0.5195, 0.4155, 0.5273, 0.518, 0.4116] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@skip_mps
def A_ ( self : Optional[int] ):
return super().test_save_load_local()
@unittest.skip('''non-deterministic pipeline''' )
def A_ ( self : List[Any] ):
return super().test_inference_batch_single_identical()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_dict_tuple_outputs_equivalent()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_save_load_optional_components()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_attention_slicing_forward_pass()
@slow
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : List[Any] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Union[str, Any] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car_fp16.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(
lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , torch_dtype=torch.floataa , revision='''fp16''' )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
# the values aren't exactly equal, but the images look the same visually
assert np.abs(image - expected_image ).max() < 5e-1
def A_ ( self : List[str] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
assert np.abs(image - expected_image ).max() < 2e-2
| 56
| 1
|
'''simple docstring'''
from collections import defaultdict
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 1
snake_case_ = True
for v in tree[start]:
if v not in visited:
ret += dfs(__UpperCAmelCase )
if ret % 2 == 0:
cuts.append(__UpperCAmelCase )
return ret
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
dfs(1 )
if __name__ == "__main__":
a ,a : Dict = 10, 9
a : Dict = defaultdict(list)
a : dict[int, bool] = {}
a : list[int] = []
a : Tuple = 0
a : str = [(2, 1), (3, 1), (4, 3), (5, 2), (6, 1), (7, 2), (8, 6), (9, 8), (10, 8)]
for u, v in edges:
tree[u].append(v)
tree[v].append(u)
even_tree()
print(len(cuts) - 1)
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : str = logging.get_logger(__name__)
a : str = {
'google/bigbird-roberta-base': 'https://huggingface.co/google/bigbird-roberta-base/resolve/main/config.json',
'google/bigbird-roberta-large': 'https://huggingface.co/google/bigbird-roberta-large/resolve/main/config.json',
'google/bigbird-base-trivia-itc': 'https://huggingface.co/google/bigbird-base-trivia-itc/resolve/main/config.json',
# See all BigBird models at https://huggingface.co/models?filter=big_bird
}
class a ( _lowerCamelCase ):
snake_case_ = "big_bird"
def __init__( self : Union[str, Any] , lowercase_ : List[Any]=5_0358 , lowercase_ : Tuple=768 , lowercase_ : Dict=12 , lowercase_ : str=12 , lowercase_ : Tuple=3072 , lowercase_ : Any="gelu_new" , lowercase_ : Optional[Any]=0.1 , lowercase_ : List[Any]=0.1 , lowercase_ : List[Any]=4096 , lowercase_ : List[Any]=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[int]=1e-12 , lowercase_ : Tuple=True , lowercase_ : Tuple=0 , lowercase_ : str=1 , lowercase_ : Union[str, Any]=2 , lowercase_ : Optional[Any]=66 , lowercase_ : Optional[int]="block_sparse" , lowercase_ : Any=True , lowercase_ : List[str]=False , lowercase_ : Any=64 , lowercase_ : Tuple=3 , lowercase_ : Tuple=None , **lowercase_ : Tuple , ):
super().__init__(
pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , sep_token_id=lowercase_ , **lowercase_ , )
snake_case_ = vocab_size
snake_case_ = max_position_embeddings
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = type_vocab_size
snake_case_ = layer_norm_eps
snake_case_ = use_cache
snake_case_ = rescale_embeddings
snake_case_ = attention_type
snake_case_ = use_bias
snake_case_ = block_size
snake_case_ = num_random_blocks
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : str ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
| 1
|
'''simple docstring'''
import pytest
from datasets.parallel import ParallelBackendConfig, parallel_backend
from datasets.utils.py_utils import map_nested
from .utils import require_dill_gt_0_3_2, require_joblibspark, require_not_windows
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]: # picklable for multiprocessing
'''simple docstring'''
return i + 1
@require_dill_gt_0_3_2
@require_joblibspark
@require_not_windows
def __magic_name__ ( ) -> List[str]:
'''simple docstring'''
with parallel_backend('''spark''' ):
assert ParallelBackendConfig.backend_name == "spark"
snake_case_ = [1, 2, 3]
with pytest.raises(__UpperCAmelCase ):
with parallel_backend('''unsupported backend''' ):
map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=2 )
with pytest.raises(__UpperCAmelCase ):
with parallel_backend('''unsupported backend''' ):
map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=-1 )
@require_dill_gt_0_3_2
@require_joblibspark
@require_not_windows
@pytest.mark.parametrize('''num_proc''', [2, -1] )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = [1, 2]
snake_case_ = {'''a''': 1, '''b''': 2}
snake_case_ = {'''a''': [1, 2], '''b''': [3, 4]}
snake_case_ = {'''a''': {'''1''': 1}, '''b''': 2}
snake_case_ = {'''a''': 1, '''b''': 2, '''c''': 3, '''d''': 4}
snake_case_ = [2, 3]
snake_case_ = {'''a''': 2, '''b''': 3}
snake_case_ = {'''a''': [2, 3], '''b''': [4, 5]}
snake_case_ = {'''a''': {'''1''': 2}, '''b''': 3}
snake_case_ = {'''a''': 2, '''b''': 3, '''c''': 4, '''d''': 5}
with parallel_backend('''spark''' ):
assert map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=__UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=__UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=__UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=__UpperCAmelCase ) == expected_map_nested_sa
assert map_nested(__UpperCAmelCase, __UpperCAmelCase, num_proc=__UpperCAmelCase ) == expected_map_nested_sa
| 56
|
'''simple docstring'''
import contextlib
import os
import sqlitea
import pytest
from datasets import Dataset, Features, Value
from datasets.io.sql import SqlDatasetReader, SqlDatasetWriter
from ..utils import assert_arrow_memory_doesnt_increase, assert_arrow_memory_increases, require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
assert isinstance(__UpperCAmelCase, __UpperCAmelCase )
assert dataset.num_rows == 4
assert dataset.num_columns == 3
assert dataset.column_names == ["col_1", "col_2", "col_3"]
for feature, expected_dtype in expected_features.items():
assert dataset.features[feature].dtype == expected_dtype
@require_sqlalchemy
@pytest.mark.parametrize('''keep_in_memory''', [False, True] )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase():
snake_case_ = SqlDatasetReader(
'''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase, keep_in_memory=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
@require_sqlalchemy
@pytest.mark.parametrize(
'''features''', [
None,
{'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''},
{'''col_1''': '''string''', '''col_2''': '''string''', '''col_3''': '''string'''},
{'''col_1''': '''int32''', '''col_2''': '''int32''', '''col_3''': '''int32'''},
{'''col_1''': '''float32''', '''col_2''': '''float32''', '''col_3''': '''float32'''},
], )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
snake_case_ = features.copy() if features else default_expected_features
snake_case_ = (
Features({feature: Value(__UpperCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None
)
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, features=__UpperCAmelCase, cache_dir=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
with contextlib.closing(sqlitea.connect(__UpperCAmelCase ) ) as con:
snake_case_ = con.cursor()
cur.execute('''SELECT * FROM dataset''' )
for row in cur:
yield row
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=1 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=2 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
with pytest.raises(__UpperCAmelCase ):
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=0 ).write()
| 56
| 1
|
'''simple docstring'''
import os
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 logging
a : Optional[Any] = logging.get_logger(__name__)
a : Union[str, Any] = '▁'
a : int = {'vocab_file': 'sentencepiece.bpe.model', 'monolingual_vocab_file': 'dict.txt'}
a : str = {
'vocab_file': {
'vinai/bartpho-syllable': 'https://huggingface.co/vinai/bartpho-syllable/resolve/main/sentencepiece.bpe.model',
},
'monolingual_vocab_file': {
'vinai/bartpho-syllable': 'https://huggingface.co/vinai/bartpho-syllable/resolve/main/dict.txt',
},
}
a : Optional[int] = {'vinai/bartpho-syllable': 1024}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
def __init__( self : Optional[int] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int]="<s>" , lowercase_ : Dict="</s>" , lowercase_ : str="</s>" , lowercase_ : Union[str, Any]="<s>" , lowercase_ : Optional[int]="<unk>" , lowercase_ : Optional[Any]="<pad>" , lowercase_ : Optional[Any]="<mask>" , lowercase_ : Optional[Dict[str, Any]] = None , **lowercase_ : List[Any] , ):
# Mask token behave like a normal word, i.e. include the space before it
snake_case_ = AddedToken(lowercase_ , lstrip=lowercase_ , rstrip=lowercase_ ) if isinstance(lowercase_ , lowercase_ ) else mask_token
snake_case_ = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
bos_token=lowercase_ , eos_token=lowercase_ , unk_token=lowercase_ , sep_token=lowercase_ , cls_token=lowercase_ , pad_token=lowercase_ , mask_token=lowercase_ , sp_model_kwargs=self.sp_model_kwargs , **lowercase_ , )
snake_case_ = vocab_file
snake_case_ = monolingual_vocab_file
snake_case_ = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(str(lowercase_ ) )
# Load the reduced vocab
# Keep order of special tokens for backward compatibility
snake_case_ = {}
snake_case_ = 0
for token in [bos_token, pad_token, eos_token, unk_token, sep_token, cls_token]:
if str(lowercase_ ) not in self.fairseq_tokens_to_ids:
snake_case_ = cnt
cnt += 1
with open(lowercase_ , '''r''' , encoding='''utf-8''' ) as f:
for line in f.readlines():
snake_case_ = line.strip().split()[0]
snake_case_ = len(self.fairseq_tokens_to_ids )
if str(lowercase_ ) not in self.fairseq_tokens_to_ids:
snake_case_ = len(self.fairseq_tokens_to_ids )
snake_case_ = {v: k for k, v in self.fairseq_tokens_to_ids.items()}
def __getstate__( self : Tuple ):
snake_case_ = self.__dict__.copy()
snake_case_ = None
snake_case_ = self.sp_model.serialized_model_proto()
return state
def __setstate__( self : Optional[Any] , lowercase_ : Optional[Any] ):
snake_case_ = d
# for backward compatibility
if not hasattr(self , '''sp_model_kwargs''' ):
snake_case_ = {}
snake_case_ = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.LoadFromSerializedProto(self.sp_model_proto )
def A_ ( self : Tuple , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
snake_case_ = [self.cls_token_id]
snake_case_ = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def A_ ( self : List[str] , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None , lowercase_ : bool = False ):
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 None:
return [1] + ([0] * len(lowercase_ )) + [1]
return [1] + ([0] * len(lowercase_ )) + [1, 1] + ([0] * len(lowercase_ )) + [1]
def A_ ( self : str , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.sep_token_id]
snake_case_ = [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]
@property
def A_ ( self : Any ):
return len(self.fairseq_ids_to_tokens )
def A_ ( self : Dict ):
snake_case_ = {self.convert_ids_to_tokens(lowercase_ ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def A_ ( self : Dict , lowercase_ : str ):
return self.sp_model.encode(lowercase_ , out_type=lowercase_ )
def A_ ( self : str , lowercase_ : List[Any] ):
if token in self.fairseq_tokens_to_ids:
return self.fairseq_tokens_to_ids[token]
else:
return self.unk_token_id
def A_ ( self : List[Any] , lowercase_ : int ):
return self.fairseq_ids_to_tokens[index]
def A_ ( self : List[Any] , lowercase_ : Optional[Any] ):
snake_case_ = ''''''.join(lowercase_ ).replace(lowercase_ , ''' ''' ).strip()
return out_string
def A_ ( self : Optional[int] , lowercase_ : str , lowercase_ : Optional[str] = None ):
if not os.path.isdir(lowercase_ ):
logger.error(F"Vocabulary path ({save_directory}) should be a directory" )
return
snake_case_ = os.path.join(
lowercase_ , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(
lowercase_ , (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''monolingual_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:
snake_case_ = self.sp_model.serialized_model_proto()
fi.write(lowercase_ )
if os.path.abspath(self.monolingual_vocab_file ) != os.path.abspath(
lowercase_ ) and os.path.isfile(self.monolingual_vocab_file ):
copyfile(self.monolingual_vocab_file , lowercase_ )
elif not os.path.isfile(self.monolingual_vocab_file ):
with open(lowercase_ , '''w''' , encoding='''utf-8''' ) as fp:
for token in self.fairseq_tokens_to_ids:
if token not in self.all_special_tokens:
fp.write(F"{str(lowercase_ )} \n" )
return out_vocab_file, out_monolingual_vocab_file
| 56
|
'''simple docstring'''
from collections import defaultdict
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 1
snake_case_ = True
for v in tree[start]:
if v not in visited:
ret += dfs(__UpperCAmelCase )
if ret % 2 == 0:
cuts.append(__UpperCAmelCase )
return ret
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
dfs(1 )
if __name__ == "__main__":
a ,a : Dict = 10, 9
a : Dict = defaultdict(list)
a : dict[int, bool] = {}
a : list[int] = []
a : Tuple = 0
a : str = [(2, 1), (3, 1), (4, 3), (5, 2), (6, 1), (7, 2), (8, 6), (9, 8), (10, 8)]
for u, v in edges:
tree[u].append(v)
tree[v].append(u)
even_tree()
print(len(cuts) - 1)
| 56
| 1
|
'''simple docstring'''
# Lint as: python3
import sys
from collections.abc import Mapping
from typing import TYPE_CHECKING
import numpy as np
import pyarrow as pa
from .. import config
from ..utils.py_utils import map_nested
from .formatting import TensorFormatter
if TYPE_CHECKING:
import torch
class a ( TensorFormatter[Mapping, "torch.Tensor", Mapping] ):
def __init__( self : Dict , lowercase_ : List[str]=None , **lowercase_ : Union[str, Any] ):
super().__init__(features=lowercase_ )
snake_case_ = torch_tensor_kwargs
import torch # noqa import torch at initialization
def A_ ( self : str , lowercase_ : str ):
import torch
if isinstance(lowercase_ , lowercase_ ) and column:
if all(
isinstance(lowercase_ , torch.Tensor ) and x.shape == column[0].shape and x.dtype == column[0].dtype
for x in column ):
return torch.stack(lowercase_ )
return column
def A_ ( self : Any , lowercase_ : Tuple ):
import torch
if isinstance(lowercase_ , (str, bytes, type(lowercase_ )) ):
return value
elif isinstance(lowercase_ , (np.character, np.ndarray) ) and np.issubdtype(value.dtype , np.character ):
return value.tolist()
snake_case_ = {}
if isinstance(lowercase_ , (np.number, np.ndarray) ) and np.issubdtype(value.dtype , np.integer ):
snake_case_ = {'''dtype''': torch.intaa}
elif isinstance(lowercase_ , (np.number, np.ndarray) ) and np.issubdtype(value.dtype , np.floating ):
snake_case_ = {'''dtype''': torch.floataa}
elif config.PIL_AVAILABLE and "PIL" in sys.modules:
import PIL.Image
if isinstance(lowercase_ , PIL.Image.Image ):
snake_case_ = np.asarray(lowercase_ )
return torch.tensor(lowercase_ , **{**default_dtype, **self.torch_tensor_kwargs} )
def A_ ( self : List[Any] , lowercase_ : Optional[int] ):
import torch
# support for torch, tf, jax etc.
if hasattr(lowercase_ , '''__array__''' ) and not isinstance(lowercase_ , torch.Tensor ):
snake_case_ = data_struct.__array__()
# support for nested types like struct of list of struct
if isinstance(lowercase_ , np.ndarray ):
if data_struct.dtype == object: # torch tensors cannot be instantied from an array of objects
return self._consolidate([self.recursive_tensorize(lowercase_ ) for substruct in data_struct] )
elif isinstance(lowercase_ , (list, tuple) ):
return self._consolidate([self.recursive_tensorize(lowercase_ ) for substruct in data_struct] )
return self._tensorize(lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : dict ):
return map_nested(self._recursive_tensorize , lowercase_ , map_list=lowercase_ )
def A_ ( self : Dict , lowercase_ : pa.Table ):
snake_case_ = self.numpy_arrow_extractor().extract_row(lowercase_ )
snake_case_ = self.python_features_decoder.decode_row(lowercase_ )
return self.recursive_tensorize(lowercase_ )
def A_ ( self : List[Any] , lowercase_ : pa.Table ):
snake_case_ = self.numpy_arrow_extractor().extract_column(lowercase_ )
snake_case_ = self.python_features_decoder.decode_column(lowercase_ , pa_table.column_names[0] )
snake_case_ = self.recursive_tensorize(lowercase_ )
snake_case_ = self._consolidate(lowercase_ )
return column
def A_ ( self : str , lowercase_ : pa.Table ):
snake_case_ = self.numpy_arrow_extractor().extract_batch(lowercase_ )
snake_case_ = self.python_features_decoder.decode_batch(lowercase_ )
snake_case_ = self.recursive_tensorize(lowercase_ )
for column_name in batch:
snake_case_ = self._consolidate(batch[column_name] )
return batch
| 56
|
'''simple docstring'''
import math
from collections.abc import Callable
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
snake_case_ = xa
snake_case_ = xa
while True:
if x_n == x_na or function(__UpperCAmelCase ) == function(__UpperCAmelCase ):
raise ZeroDivisionError('''float division by zero, could not find root''' )
snake_case_ = x_na - (
function(__UpperCAmelCase ) / ((function(__UpperCAmelCase ) - function(__UpperCAmelCase )) / (x_na - x_n))
)
if abs(x_na - x_na ) < 10**-5:
return x_na
snake_case_ = x_na
snake_case_ = x_na
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
return math.pow(__UpperCAmelCase, 3 ) - (2 * x) - 5
if __name__ == "__main__":
print(intersection(f, 3, 3.5))
| 56
| 1
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : Tuple = logging.get_logger(__name__)
a : Union[str, Any] = {
'camembert-base': 'https://huggingface.co/camembert-base/resolve/main/config.json',
'umberto-commoncrawl-cased-v1': (
'https://huggingface.co/Musixmatch/umberto-commoncrawl-cased-v1/resolve/main/config.json'
),
'umberto-wikipedia-uncased-v1': (
'https://huggingface.co/Musixmatch/umberto-wikipedia-uncased-v1/resolve/main/config.json'
),
}
class a ( _lowerCamelCase ):
snake_case_ = "camembert"
def __init__( self : List[Any] , lowercase_ : Dict=3_0522 , lowercase_ : Dict=768 , lowercase_ : str=12 , lowercase_ : Optional[Any]=12 , lowercase_ : Dict=3072 , lowercase_ : Union[str, Any]="gelu" , lowercase_ : int=0.1 , lowercase_ : str=0.1 , lowercase_ : Union[str, Any]=512 , lowercase_ : List[str]=2 , lowercase_ : Optional[int]=0.02 , lowercase_ : Union[str, Any]=1e-12 , lowercase_ : Dict=1 , lowercase_ : Any=0 , lowercase_ : int=2 , lowercase_ : List[str]="absolute" , lowercase_ : Any=True , lowercase_ : int=None , **lowercase_ : Union[str, Any] , ):
super().__init__(pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , **lowercase_ )
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = hidden_act
snake_case_ = intermediate_size
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = position_embedding_type
snake_case_ = use_cache
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : Any ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_url
from PIL import Image
from transformers import DPTConfig, DPTForDepthEstimation, DPTForSemanticSegmentation, DPTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
a : Any = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = DPTConfig()
if "large" in checkpoint_url:
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 24
snake_case_ = 16
snake_case_ = [5, 11, 17, 23]
snake_case_ = [256, 512, 1024, 1024]
snake_case_ = (1, 384, 384)
if "ade" in checkpoint_url:
snake_case_ = True
snake_case_ = 150
snake_case_ = '''huggingface/label-files'''
snake_case_ = '''ade20k-id2label.json'''
snake_case_ = json.load(open(cached_download(hf_hub_url(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ) ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = idalabel
snake_case_ = {v: k for k, v in idalabel.items()}
snake_case_ = [1, 150, 480, 480]
return config, expected_shape
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = ['''pretrained.model.head.weight''', '''pretrained.model.head.bias''']
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if (
"pretrained.model" in name
and "cls_token" not in name
and "pos_embed" not in name
and "patch_embed" not in name
):
snake_case_ = name.replace('''pretrained.model''', '''dpt.encoder''' )
if "pretrained.model" in name:
snake_case_ = name.replace('''pretrained.model''', '''dpt.embeddings''' )
if "patch_embed" in name:
snake_case_ = name.replace('''patch_embed''', '''patch_embeddings''' )
if "pos_embed" in name:
snake_case_ = name.replace('''pos_embed''', '''position_embeddings''' )
if "attn.proj" in name:
snake_case_ = name.replace('''attn.proj''', '''attention.output.dense''' )
if "proj" in name and "project" not in name:
snake_case_ = name.replace('''proj''', '''projection''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''layer''' )
if "mlp.fc1" in name:
snake_case_ = name.replace('''mlp.fc1''', '''intermediate.dense''' )
if "mlp.fc2" in name:
snake_case_ = name.replace('''mlp.fc2''', '''output.dense''' )
if "norm1" in name:
snake_case_ = name.replace('''norm1''', '''layernorm_before''' )
if "norm2" in name:
snake_case_ = name.replace('''norm2''', '''layernorm_after''' )
if "scratch.output_conv" in name:
snake_case_ = name.replace('''scratch.output_conv''', '''head''' )
if "scratch" in name:
snake_case_ = name.replace('''scratch''', '''neck''' )
if "layer1_rn" in name:
snake_case_ = name.replace('''layer1_rn''', '''convs.0''' )
if "layer2_rn" in name:
snake_case_ = name.replace('''layer2_rn''', '''convs.1''' )
if "layer3_rn" in name:
snake_case_ = name.replace('''layer3_rn''', '''convs.2''' )
if "layer4_rn" in name:
snake_case_ = name.replace('''layer4_rn''', '''convs.3''' )
if "refinenet" in name:
snake_case_ = int(name[len('''neck.refinenet''' ) : len('''neck.refinenet''' ) + 1] )
# tricky here: we need to map 4 to 0, 3 to 1, 2 to 2 and 1 to 3
snake_case_ = name.replace(F"refinenet{layer_idx}", F"fusion_stage.layers.{abs(layer_idx-4 )}" )
if "out_conv" in name:
snake_case_ = name.replace('''out_conv''', '''projection''' )
if "resConfUnit1" in name:
snake_case_ = name.replace('''resConfUnit1''', '''residual_layer1''' )
if "resConfUnit2" in name:
snake_case_ = name.replace('''resConfUnit2''', '''residual_layer2''' )
if "conv1" in name:
snake_case_ = name.replace('''conv1''', '''convolution1''' )
if "conv2" in name:
snake_case_ = name.replace('''conv2''', '''convolution2''' )
# readout blocks
if "pretrained.act_postprocess1.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.0.project.0''', '''neck.reassemble_stage.readout_projects.0.0''' )
if "pretrained.act_postprocess2.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.0.project.0''', '''neck.reassemble_stage.readout_projects.1.0''' )
if "pretrained.act_postprocess3.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.0.project.0''', '''neck.reassemble_stage.readout_projects.2.0''' )
if "pretrained.act_postprocess4.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.0.project.0''', '''neck.reassemble_stage.readout_projects.3.0''' )
# resize blocks
if "pretrained.act_postprocess1.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.3''', '''neck.reassemble_stage.layers.0.projection''' )
if "pretrained.act_postprocess1.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.4''', '''neck.reassemble_stage.layers.0.resize''' )
if "pretrained.act_postprocess2.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.3''', '''neck.reassemble_stage.layers.1.projection''' )
if "pretrained.act_postprocess2.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.4''', '''neck.reassemble_stage.layers.1.resize''' )
if "pretrained.act_postprocess3.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.3''', '''neck.reassemble_stage.layers.2.projection''' )
if "pretrained.act_postprocess4.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.3''', '''neck.reassemble_stage.layers.3.projection''' )
if "pretrained.act_postprocess4.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.4''', '''neck.reassemble_stage.layers.3.resize''' )
if "pretrained" in name:
snake_case_ = name.replace('''pretrained''', '''dpt''' )
if "bn" in name:
snake_case_ = name.replace('''bn''', '''batch_norm''' )
if "head" in name:
snake_case_ = name.replace('''head''', '''head.head''' )
if "encoder.norm" in name:
snake_case_ = name.replace('''encoder.norm''', '''layernorm''' )
if "auxlayer" in name:
snake_case_ = name.replace('''auxlayer''', '''auxiliary_head.head''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.weight" )
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.bias" )
# next, add query, keys and values (in that order) to the state dict
snake_case_ = in_proj_weight[: config.hidden_size, :]
snake_case_ = in_proj_bias[: config.hidden_size]
snake_case_ = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
snake_case_ = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
snake_case_ = in_proj_weight[
-config.hidden_size :, :
]
snake_case_ = in_proj_bias[-config.hidden_size :]
def __magic_name__ ( ) -> Any:
'''simple docstring'''
snake_case_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
snake_case_ = Image.open(requests.get(__UpperCAmelCase, stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dpt_config(__UpperCAmelCase )
# load original state_dict from URL
snake_case_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase, map_location='''cpu''' )
# remove certain keys
remove_ignore_keys_(__UpperCAmelCase )
# rename keys
for key in state_dict.copy().keys():
snake_case_ = state_dict.pop(__UpperCAmelCase )
snake_case_ = val
# read in qkv matrices
read_in_q_k_v(__UpperCAmelCase, __UpperCAmelCase )
# load HuggingFace model
snake_case_ = DPTForSemanticSegmentation(__UpperCAmelCase ) if '''ade''' in checkpoint_url else DPTForDepthEstimation(__UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
# Check outputs on an image
snake_case_ = 480 if '''ade''' in checkpoint_url else 384
snake_case_ = DPTImageProcessor(size=__UpperCAmelCase )
snake_case_ = prepare_img()
snake_case_ = image_processor(__UpperCAmelCase, return_tensors='''pt''' )
# forward pass
snake_case_ = model(**__UpperCAmelCase ).logits if '''ade''' in checkpoint_url else model(**__UpperCAmelCase ).predicted_depth
# Assert logits
snake_case_ = torch.tensor([[6.3_1_9_9, 6.3_6_2_9, 6.4_1_4_8], [6.3_8_5_0, 6.3_6_1_5, 6.4_1_6_6], [6.3_5_1_9, 6.3_1_7_6, 6.3_5_7_5]] )
if "ade" in checkpoint_url:
snake_case_ = torch.tensor([[4.0_4_8_0, 4.2_4_2_0, 4.4_3_6_0], [4.3_1_2_4, 4.5_6_9_3, 4.8_2_6_1], [4.5_7_6_8, 4.8_9_6_5, 5.2_1_6_3]] )
assert outputs.shape == torch.Size(__UpperCAmelCase )
assert (
torch.allclose(outputs[0, 0, :3, :3], __UpperCAmelCase, atol=1e-4 )
if "ade" in checkpoint_url
else torch.allclose(outputs[0, :3, :3], __UpperCAmelCase )
)
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
print(F"Saving model to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
print(F"Saving image processor to {pytorch_dump_folder_path}" )
image_processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print('''Pushing model to hub...''' )
model.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add model''', use_temp_dir=__UpperCAmelCase, )
image_processor.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add image processor''', use_temp_dir=__UpperCAmelCase, )
if __name__ == "__main__":
a : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--checkpoint_url',
default='https://github.com/intel-isl/DPT/releases/download/1_0/dpt_large-midas-2f21e586.pt',
type=str,
help='URL of the original DPT checkpoint you\'d like to convert.',
)
parser.add_argument(
'--pytorch_dump_folder_path',
default=None,
type=str,
required=True,
help='Path to the output PyTorch model directory.',
)
parser.add_argument(
'--push_to_hub',
action='store_true',
)
parser.add_argument(
'--model_name',
default='dpt-large',
type=str,
help='Name of the model, in case you\'re pushing to the hub.',
)
a : List[Any] = parser.parse_args()
convert_dpt_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 56
| 1
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : Optional[int] = logging.get_logger(__name__)
a : Optional[int] = {
'andreasmadsen/efficient_mlm_m0.40': (
'https://huggingface.co/andreasmadsen/efficient_mlm_m0.40/resolve/main/config.json'
),
}
class a ( _lowerCamelCase ):
snake_case_ = "roberta-prelayernorm"
def __init__( self : List[str] , lowercase_ : Union[str, Any]=5_0265 , lowercase_ : Any=768 , lowercase_ : Tuple=12 , lowercase_ : int=12 , lowercase_ : Dict=3072 , lowercase_ : Any="gelu" , lowercase_ : Dict=0.1 , lowercase_ : int=0.1 , lowercase_ : List[Any]=512 , lowercase_ : Any=2 , lowercase_ : str=0.02 , lowercase_ : Union[str, Any]=1e-12 , lowercase_ : Tuple=1 , lowercase_ : Optional[int]=0 , lowercase_ : List[str]=2 , lowercase_ : str="absolute" , lowercase_ : Tuple=True , lowercase_ : Any=None , **lowercase_ : Union[str, Any] , ):
super().__init__(pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , **lowercase_ )
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = hidden_act
snake_case_ = intermediate_size
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = position_embedding_type
snake_case_ = use_cache
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : Optional[int] ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
|
'''simple docstring'''
import re
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
snake_case_ = re.compile(
r'''^(?:0|94|\+94|0{2}94)''' r'''7(0|1|2|4|5|6|7|8)''' r'''(-| |)''' r'''\d{7}$''' )
return bool(re.search(__UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
a : Any = '0094702343221'
print(is_sri_lankan_phone_number(phone))
| 56
| 1
|
'''simple docstring'''
import os
import unittest
from transformers import LayoutLMTokenizer, LayoutLMTokenizerFast
from transformers.models.layoutlm.tokenization_layoutlm import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = LayoutLMTokenizer
snake_case_ = LayoutLMTokenizerFast
snake_case_ = True
snake_case_ = True
def A_ ( self : Optional[Any] ):
super().setUp()
snake_case_ = [
'''[UNK]''',
'''[CLS]''',
'''[SEP]''',
'''want''',
'''##want''',
'''##ed''',
'''wa''',
'''un''',
'''runn''',
'''##ing''',
''',''',
'''low''',
'''lowest''',
]
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
def A_ ( self : Any , **lowercase_ : Optional[Any] ):
return LayoutLMTokenizer.from_pretrained(self.tmpdirname , **lowercase_ )
def A_ ( self : Union[str, Any] , lowercase_ : Optional[int] ):
snake_case_ = '''UNwant\u00E9d,running'''
snake_case_ = '''unwanted, running'''
return input_text, output_text
def A_ ( self : Dict ):
snake_case_ = self.tokenizer_class(self.vocab_file )
snake_case_ = tokenizer.tokenize('''UNwant\u00E9d,running''' )
self.assertListEqual(lowercase_ , ['''un''', '''##want''', '''##ed''', ''',''', '''runn''', '''##ing'''] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(lowercase_ ) , [7, 4, 5, 10, 8, 9] )
def A_ ( self : Any ):
pass
| 56
|
'''simple docstring'''
import re
from filelock import FileLock
try:
import nltk
a : Union[str, Any] = True
except (ImportError, ModuleNotFoundError):
a : Any = False
if NLTK_AVAILABLE:
with FileLock('.lock') as lock:
nltk.download('punkt', quiet=True)
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
re.sub('''<n>''', '''''', __UpperCAmelCase ) # remove pegasus newline char
assert NLTK_AVAILABLE, "nltk must be installed to separate newlines between sentences. (pip install nltk)"
return "\n".join(nltk.sent_tokenize(__UpperCAmelCase ) )
| 56
| 1
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_tokenizers_available,
is_torch_available,
is_vision_available,
)
a : int = {
'configuration_layoutlmv3': [
'LAYOUTLMV3_PRETRAINED_CONFIG_ARCHIVE_MAP',
'LayoutLMv3Config',
'LayoutLMv3OnnxConfig',
],
'processing_layoutlmv3': ['LayoutLMv3Processor'],
'tokenization_layoutlmv3': ['LayoutLMv3Tokenizer'],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Any = ['LayoutLMv3TokenizerFast']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Dict = [
'LAYOUTLMV3_PRETRAINED_MODEL_ARCHIVE_LIST',
'LayoutLMv3ForQuestionAnswering',
'LayoutLMv3ForSequenceClassification',
'LayoutLMv3ForTokenClassification',
'LayoutLMv3Model',
'LayoutLMv3PreTrainedModel',
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[int] = [
'TF_LAYOUTLMV3_PRETRAINED_MODEL_ARCHIVE_LIST',
'TFLayoutLMv3ForQuestionAnswering',
'TFLayoutLMv3ForSequenceClassification',
'TFLayoutLMv3ForTokenClassification',
'TFLayoutLMv3Model',
'TFLayoutLMv3PreTrainedModel',
]
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = ['LayoutLMv3FeatureExtractor']
a : Any = ['LayoutLMv3ImageProcessor']
if TYPE_CHECKING:
from .configuration_layoutlmva import (
LAYOUTLMV3_PRETRAINED_CONFIG_ARCHIVE_MAP,
LayoutLMvaConfig,
LayoutLMvaOnnxConfig,
)
from .processing_layoutlmva import LayoutLMvaProcessor
from .tokenization_layoutlmva import LayoutLMvaTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_layoutlmva_fast import LayoutLMvaTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_layoutlmva import (
LAYOUTLMV3_PRETRAINED_MODEL_ARCHIVE_LIST,
LayoutLMvaForQuestionAnswering,
LayoutLMvaForSequenceClassification,
LayoutLMvaForTokenClassification,
LayoutLMvaModel,
LayoutLMvaPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_layoutlmva import (
TF_LAYOUTLMV3_PRETRAINED_MODEL_ARCHIVE_LIST,
TFLayoutLMvaForQuestionAnswering,
TFLayoutLMvaForSequenceClassification,
TFLayoutLMvaForTokenClassification,
TFLayoutLMvaModel,
TFLayoutLMvaPreTrainedModel,
)
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_layoutlmva import LayoutLMvaFeatureExtractor
from .image_processing_layoutlmva import LayoutLMvaImageProcessor
else:
import sys
a : Any = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
a : Tuple = {
'configuration_llama': ['LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP', 'LlamaConfig'],
}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = ['LlamaTokenizer']
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : str = ['LlamaTokenizerFast']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = [
'LlamaForCausalLM',
'LlamaModel',
'LlamaPreTrainedModel',
'LlamaForSequenceClassification',
]
if TYPE_CHECKING:
from .configuration_llama import LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP, LlamaConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama import LlamaTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama_fast import LlamaTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_llama import LlamaForCausalLM, LlamaForSequenceClassification, LlamaModel, LlamaPreTrainedModel
else:
import sys
a : str = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
| 1
|
'''simple docstring'''
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_torch, slow, torch_device
from ...generation.test_utils import GenerationTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
OPENAI_GPT_PRETRAINED_MODEL_ARCHIVE_LIST,
OpenAIGPTConfig,
OpenAIGPTDoubleHeadsModel,
OpenAIGPTForSequenceClassification,
OpenAIGPTLMHeadModel,
OpenAIGPTModel,
)
class a :
def __init__( self : Any , lowercase_ : str , lowercase_ : int=13 , lowercase_ : str=7 , lowercase_ : Tuple=True , lowercase_ : Union[str, Any]=True , lowercase_ : Dict=True , lowercase_ : Dict=99 , lowercase_ : int=32 , lowercase_ : Dict=5 , lowercase_ : str=4 , lowercase_ : Union[str, Any]=37 , lowercase_ : str="gelu" , lowercase_ : str=0.1 , lowercase_ : Any=0.1 , lowercase_ : List[str]=512 , lowercase_ : Optional[Any]=16 , lowercase_ : List[str]=2 , lowercase_ : Any=0.02 , lowercase_ : Any=3 , lowercase_ : int=4 , lowercase_ : Dict=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = scope
snake_case_ = self.vocab_size - 1
def A_ ( self : Dict ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_token_type_ids:
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = ids_tensor([self.batch_size] , self.num_choices )
snake_case_ = OpenAIGPTConfig(
vocab_size=self.vocab_size , n_embd=self.hidden_size , n_layer=self.num_hidden_layers , n_head=self.num_attention_heads , n_positions=self.max_position_embeddings , pad_token_id=self.pad_token_id , )
snake_case_ = ids_tensor([self.num_hidden_layers, self.num_attention_heads] , 2 )
return (
config,
input_ids,
head_mask,
token_type_ids,
sequence_labels,
token_labels,
choice_labels,
)
def A_ ( self : Dict , lowercase_ : Optional[Any] , lowercase_ : str , lowercase_ : List[str] , lowercase_ : Any , *lowercase_ : Dict ):
snake_case_ = OpenAIGPTModel(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , token_type_ids=lowercase_ , head_mask=lowercase_ )
snake_case_ = model(lowercase_ , token_type_ids=lowercase_ )
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : str , lowercase_ : str , lowercase_ : Dict , *lowercase_ : Union[str, Any] ):
snake_case_ = OpenAIGPTLMHeadModel(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , token_type_ids=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.loss.shape , () )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def A_ ( self : int , lowercase_ : List[Any] , lowercase_ : List[Any] , lowercase_ : int , lowercase_ : str , *lowercase_ : Optional[int] ):
snake_case_ = OpenAIGPTDoubleHeadsModel(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , token_type_ids=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.loss.shape , () )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def A_ ( self : Any , lowercase_ : str , lowercase_ : Any , lowercase_ : int , lowercase_ : Union[str, Any] , *lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = OpenAIGPTForSequenceClassification(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = model(lowercase_ , token_type_ids=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : Tuple ):
snake_case_ = self.prepare_config_and_inputs()
(
(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,
) = config_and_inputs
snake_case_ = {
'''input_ids''': input_ids,
'''token_type_ids''': token_type_ids,
'''head_mask''': head_mask,
}
return config, inputs_dict
@require_torch
class a ( _lowerCamelCase , _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(OpenAIGPTModel, OpenAIGPTLMHeadModel, OpenAIGPTDoubleHeadsModel, OpenAIGPTForSequenceClassification)
if is_torch_available()
else ()
)
snake_case_ = (
(OpenAIGPTLMHeadModel,) if is_torch_available() else ()
) # TODO (PVP): Add Double HeadsModel when generate() function is changed accordingly
snake_case_ = (
{
"feature-extraction": OpenAIGPTModel,
"text-classification": OpenAIGPTForSequenceClassification,
"text-generation": OpenAIGPTLMHeadModel,
"zero-shot": OpenAIGPTForSequenceClassification,
}
if is_torch_available()
else {}
)
def A_ ( self : Any , lowercase_ : List[str] , lowercase_ : str , lowercase_ : Any , lowercase_ : Optional[Any] , lowercase_ : int ):
if pipeline_test_casse_name == "ZeroShotClassificationPipelineTests":
# Get `tokenizer does not have a padding token` error for both fast/slow tokenizers.
# `OpenAIGPTConfig` was never used in pipeline tests, either because of a missing checkpoint or because a
# tiny config could not be created.
return True
return False
def A_ ( self : Optional[Any] , lowercase_ : Dict , lowercase_ : Any , lowercase_ : int=False ):
snake_case_ = super()._prepare_for_class(lowercase_ , lowercase_ , return_labels=lowercase_ )
if return_labels:
if model_class.__name__ == "OpenAIGPTDoubleHeadsModel":
snake_case_ = torch.zeros(
(self.model_tester.batch_size, self.model_tester.num_choices, self.model_tester.seq_length) , dtype=torch.long , device=lowercase_ , )
snake_case_ = inputs_dict['''labels''']
snake_case_ = inputs_dict['''labels''']
snake_case_ = torch.zeros(
(self.model_tester.batch_size, self.model_tester.num_choices) , dtype=torch.long , device=lowercase_ , )
snake_case_ = torch.zeros(
self.model_tester.batch_size , dtype=torch.long , device=lowercase_ )
return inputs_dict
def A_ ( self : List[str] ):
snake_case_ = OpenAIGPTModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , n_embd=37 )
def A_ ( self : List[Any] ):
self.config_tester.run_common_tests()
def A_ ( self : str ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_openai_gpt_model(*lowercase_ )
def A_ ( self : Dict ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_lm_head_model(*lowercase_ )
def A_ ( self : Optional[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_double_lm_head_model(*lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_openai_gpt_for_sequence_classification(*lowercase_ )
@slow
def A_ ( self : Optional[int] ):
for model_name in OPENAI_GPT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ = OpenAIGPTModel.from_pretrained(lowercase_ )
self.assertIsNotNone(lowercase_ )
@require_torch
class a ( unittest.TestCase ):
@slow
def A_ ( self : List[str] ):
snake_case_ = OpenAIGPTLMHeadModel.from_pretrained('''openai-gpt''' )
model.to(lowercase_ )
snake_case_ = torch.tensor([[481, 4735, 544]] , dtype=torch.long , device=lowercase_ ) # the president is
snake_case_ = [
481,
4735,
544,
246,
963,
870,
762,
239,
244,
4_0477,
244,
249,
719,
881,
487,
544,
240,
244,
603,
481,
] # the president is a very good man. " \n " i\'m sure he is, " said the
snake_case_ = model.generate(lowercase_ , do_sample=lowercase_ )
self.assertListEqual(output_ids[0].tolist() , lowercase_ )
| 56
|
'''simple docstring'''
import re
from typing import Callable, List, Optional, Union
import tensorflow as tf
try:
from tensorflow.keras.optimizers.legacy import Adam
except ImportError:
from tensorflow.keras.optimizers import Adam
class a ( tf.keras.optimizers.schedules.LearningRateSchedule ):
def __init__( self : Optional[Any] , lowercase_ : float , lowercase_ : Callable , lowercase_ : int , lowercase_ : float = 1.0 , lowercase_ : str = None , ):
super().__init__()
snake_case_ = initial_learning_rate
snake_case_ = warmup_steps
snake_case_ = power
snake_case_ = decay_schedule_fn
snake_case_ = name
def __call__( self : Tuple , lowercase_ : str ):
with tf.name_scope(self.name or '''WarmUp''' ) as name:
# Implements polynomial warmup. i.e., if global_step < warmup_steps, the
# learning rate will be `global_step/num_warmup_steps * init_lr`.
snake_case_ = tf.cast(lowercase_ , tf.floataa )
snake_case_ = tf.cast(self.warmup_steps , tf.floataa )
snake_case_ = global_step_float / warmup_steps_float
snake_case_ = self.initial_learning_rate * tf.math.pow(lowercase_ , self.power )
return tf.cond(
global_step_float < warmup_steps_float , lambda: warmup_learning_rate , lambda: self.decay_schedule_fn(step - self.warmup_steps ) , name=lowercase_ , )
def A_ ( self : Any ):
return {
"initial_learning_rate": self.initial_learning_rate,
"decay_schedule_fn": self.decay_schedule_fn,
"warmup_steps": self.warmup_steps,
"power": self.power,
"name": self.name,
}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, __UpperCAmelCase = 0.9, __UpperCAmelCase = 0.9_9_9, __UpperCAmelCase = 1e-8, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = 0.0, __UpperCAmelCase = 1.0, __UpperCAmelCase = None, ) -> List[str]:
'''simple docstring'''
snake_case_ = tf.keras.optimizers.schedules.PolynomialDecay(
initial_learning_rate=__UpperCAmelCase, decay_steps=num_train_steps - num_warmup_steps, end_learning_rate=init_lr * min_lr_ratio, power=__UpperCAmelCase, )
if num_warmup_steps:
snake_case_ = WarmUp(
initial_learning_rate=__UpperCAmelCase, decay_schedule_fn=__UpperCAmelCase, warmup_steps=__UpperCAmelCase, )
if weight_decay_rate > 0.0:
snake_case_ = AdamWeightDecay(
learning_rate=__UpperCAmelCase, weight_decay_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, exclude_from_weight_decay=['''LayerNorm''', '''layer_norm''', '''bias'''], include_in_weight_decay=__UpperCAmelCase, )
else:
snake_case_ = tf.keras.optimizers.Adam(
learning_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, )
# We return the optimizer and the LR scheduler in order to better track the
# evolution of the LR independently of the optimizer.
return optimizer, lr_schedule
class a ( _lowerCamelCase ):
def __init__( self : Dict , lowercase_ : Union[float, tf.keras.optimizers.schedules.LearningRateSchedule] = 0.001 , lowercase_ : float = 0.9 , lowercase_ : float = 0.999 , lowercase_ : float = 1e-7 , lowercase_ : bool = False , lowercase_ : float = 0.0 , lowercase_ : Optional[List[str]] = None , lowercase_ : Optional[List[str]] = None , lowercase_ : str = "AdamWeightDecay" , **lowercase_ : Optional[int] , ):
super().__init__(lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
snake_case_ = weight_decay_rate
snake_case_ = include_in_weight_decay
snake_case_ = exclude_from_weight_decay
@classmethod
def A_ ( cls : Dict , lowercase_ : Union[str, Any] ):
snake_case_ = {'''WarmUp''': WarmUp}
return super(lowercase_ , cls ).from_config(lowercase_ , custom_objects=lowercase_ )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : Optional[int] ):
super(lowercase_ , self )._prepare_local(lowercase_ , lowercase_ , lowercase_ )
snake_case_ = tf.constant(
self.weight_decay_rate , name='''adam_weight_decay_rate''' )
def A_ ( self : str , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Any ):
snake_case_ = self._do_use_weight_decay(var.name )
if do_decay:
return var.assign_sub(
learning_rate * var * apply_state[(var.device, var.dtype.base_dtype)]['''weight_decay_rate'''] , use_locking=self._use_locking , )
return tf.no_op()
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : str=None , **lowercase_ : List[str] ):
snake_case_ ,snake_case_ = list(zip(*lowercase_ ) )
return super(lowercase_ , self ).apply_gradients(zip(lowercase_ , lowercase_ ) , name=lowercase_ , **lowercase_ )
def A_ ( self : List[Any] , lowercase_ : str , lowercase_ : str , lowercase_ : Any ):
if apply_state is None:
return self._decayed_lr_t[var_dtype], {}
snake_case_ = apply_state or {}
snake_case_ = apply_state.get((var_device, var_dtype) )
if coefficients is None:
snake_case_ = self._fallback_apply_state(lowercase_ , lowercase_ )
snake_case_ = coefficients
return coefficients["lr_t"], {"apply_state": apply_state}
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : Union[str, Any] , lowercase_ : Optional[int]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_dense(lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : int , lowercase_ : str , lowercase_ : List[Any]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_sparse(lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = super().get_config()
config.update({'''weight_decay_rate''': self.weight_decay_rate} )
return config
def A_ ( self : Optional[int] , lowercase_ : int ):
if self.weight_decay_rate == 0:
return False
if self._include_in_weight_decay:
for r in self._include_in_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return True
if self._exclude_from_weight_decay:
for r in self._exclude_from_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return False
return True
class a ( _lowerCamelCase ):
def __init__( self : List[Any] ):
snake_case_ = []
snake_case_ = None
@property
def A_ ( self : Union[str, Any] ):
if self._accum_steps is None:
snake_case_ = tf.Variable(
tf.constant(0 , dtype=tf.intaa ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
return self._accum_steps.value()
@property
def A_ ( self : Dict ):
if not self._gradients:
raise ValueError('''The accumulator should be called first to initialize the gradients''' )
return [gradient.value() if gradient is not None else gradient for gradient in self._gradients]
def __call__( self : Any , lowercase_ : int ):
if not self._gradients:
snake_case_ = self.step # Create the step variable.
self._gradients.extend(
[
tf.Variable(
tf.zeros_like(lowercase_ ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
if gradient is not None
else gradient
for gradient in gradients
] )
if len(lowercase_ ) != len(self._gradients ):
raise ValueError(F"Expected {len(self._gradients )} gradients, but got {len(lowercase_ )}" )
for accum_gradient, gradient in zip(self._gradients , lowercase_ ):
if accum_gradient is not None and gradient is not None:
accum_gradient.assign_add(lowercase_ )
self._accum_steps.assign_add(1 )
def A_ ( self : Optional[int] ):
if not self._gradients:
return
self._accum_steps.assign(0 )
for gradient in self._gradients:
if gradient is not None:
gradient.assign(tf.zeros_like(lowercase_ ) )
| 56
| 1
|
'''simple docstring'''
import os
from pathlib import Path
def __magic_name__ ( ) -> str:
'''simple docstring'''
from torch.utils.cpp_extension import load
snake_case_ = Path(__UpperCAmelCase ).resolve().parent.parent.parent / '''kernels''' / '''deformable_detr'''
snake_case_ = [
root / filename
for filename in [
'''vision.cpp''',
os.path.join('''cpu''', '''ms_deform_attn_cpu.cpp''' ),
os.path.join('''cuda''', '''ms_deform_attn_cuda.cu''' ),
]
]
load(
'''MultiScaleDeformableAttention''', __UpperCAmelCase, with_cuda=__UpperCAmelCase, extra_include_paths=[str(__UpperCAmelCase )], extra_cflags=['''-DWITH_CUDA=1'''], extra_cuda_cflags=[
'''-DCUDA_HAS_FP16=1''',
'''-D__CUDA_NO_HALF_OPERATORS__''',
'''-D__CUDA_NO_HALF_CONVERSIONS__''',
'''-D__CUDA_NO_HALF2_OPERATORS__''',
], )
import MultiScaleDeformableAttention as MSDA
return MSDA
| 56
|
'''simple docstring'''
import gc
import unittest
import torch
from parameterized import parameterized
from diffusers import AutoencoderKL
from diffusers.utils import floats_tensor, load_hf_numpy, require_torch_gpu, slow, torch_all_close, torch_device
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = AutoencoderKL
snake_case_ = "sample"
snake_case_ = 1e-2
@property
def A_ ( self : Dict ):
snake_case_ = 4
snake_case_ = 3
snake_case_ = (32, 32)
snake_case_ = floats_tensor((batch_size, num_channels) + sizes ).to(lowercase_ )
return {"sample": image}
@property
def A_ ( self : List[Any] ):
return (3, 32, 32)
@property
def A_ ( self : Dict ):
return (3, 32, 32)
def A_ ( self : Union[str, Any] ):
snake_case_ = {
'''block_out_channels''': [32, 64],
'''in_channels''': 3,
'''out_channels''': 3,
'''down_block_types''': ['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''],
'''up_block_types''': ['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''],
'''latent_channels''': 4,
}
snake_case_ = self.dummy_input
return init_dict, inputs_dict
def A_ ( self : Any ):
pass
def A_ ( self : str ):
pass
@unittest.skipIf(torch_device == '''mps''' , '''Gradient checkpointing skipped on MPS''' )
def A_ ( self : Dict ):
# enable deterministic behavior for gradient checkpointing
snake_case_ ,snake_case_ = self.prepare_init_args_and_inputs_for_common()
snake_case_ = self.model_class(**lowercase_ )
model.to(lowercase_ )
assert not model.is_gradient_checkpointing and model.training
snake_case_ = model(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model.zero_grad()
snake_case_ = torch.randn_like(lowercase_ )
snake_case_ = (out - labels).mean()
loss.backward()
# re-instantiate the model now enabling gradient checkpointing
snake_case_ = self.model_class(**lowercase_ )
# clone model
model_a.load_state_dict(model.state_dict() )
model_a.to(lowercase_ )
model_a.enable_gradient_checkpointing()
assert model_a.is_gradient_checkpointing and model_a.training
snake_case_ = model_a(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model_a.zero_grad()
snake_case_ = (out_a - labels).mean()
loss_a.backward()
# compare the output and parameters gradients
self.assertTrue((loss - loss_a).abs() < 1e-5 )
snake_case_ = dict(model.named_parameters() )
snake_case_ = dict(model_a.named_parameters() )
for name, param in named_params.items():
self.assertTrue(torch_all_close(param.grad.data , named_params_a[name].grad.data , atol=5e-5 ) )
def A_ ( self : Tuple ):
snake_case_ ,snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' , output_loading_info=lowercase_ )
self.assertIsNotNone(lowercase_ )
self.assertEqual(len(loading_info['''missing_keys'''] ) , 0 )
model.to(lowercase_ )
snake_case_ = model(**self.dummy_input )
assert image is not None, "Make sure output is not None"
def A_ ( self : Tuple ):
snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' )
snake_case_ = model.to(lowercase_ )
model.eval()
if torch_device == "mps":
snake_case_ = torch.manual_seed(0 )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(0 )
snake_case_ = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ = image.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , sample_posterior=lowercase_ , generator=lowercase_ ).sample
snake_case_ = output[0, -1, -3:, -3:].flatten().cpu()
# Since the VAE Gaussian prior's generator is seeded on the appropriate device,
# the expected output slices are not the same for CPU and GPU.
if torch_device == "mps":
snake_case_ = torch.tensor(
[
-4.0_078e-01,
-3.8_323e-04,
-1.2_681e-01,
-1.1_462e-01,
2.0_095e-01,
1.0_893e-01,
-8.8_247e-02,
-3.0_361e-01,
-9.8_644e-03,
] )
elif torch_device == "cpu":
snake_case_ = torch.tensor(
[-0.1352, 0.0878, 0.0419, -0.0818, -0.1069, 0.0688, -0.1458, -0.4446, -0.0026] )
else:
snake_case_ = torch.tensor(
[-0.2421, 0.4642, 0.2507, -0.0438, 0.0682, 0.3160, -0.2018, -0.0727, 0.2485] )
self.assertTrue(torch_all_close(lowercase_ , lowercase_ , rtol=1e-2 ) )
@slow
class a ( unittest.TestCase ):
def A_ ( self : Union[str, Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] ):
return F"gaussian_noise_s={seed}_shape={'_'.join([str(lowercase_ ) for s in shape] )}.npy"
def A_ ( self : Any ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Dict , lowercase_ : List[Any]=0 , lowercase_ : Union[str, Any]=(4, 3, 512, 512) , lowercase_ : Optional[Any]=False ):
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = torch.from_numpy(load_hf_numpy(self.get_file_format(lowercase_ , lowercase_ ) ) ).to(lowercase_ ).to(lowercase_ )
return image
def A_ ( self : Any , lowercase_ : Dict="CompVis/stable-diffusion-v1-4" , lowercase_ : List[str]=False ):
snake_case_ = '''fp16''' if fpaa else None
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = AutoencoderKL.from_pretrained(
lowercase_ , subfolder='''vae''' , torch_dtype=lowercase_ , revision=lowercase_ , )
model.to(lowercase_ ).eval()
return model
def A_ ( self : Any , lowercase_ : int=0 ):
if torch_device == "mps":
return torch.manual_seed(lowercase_ )
return torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
@parameterized.expand(
[
# fmt: off
[33, [-0.1603, 0.9878, -0.0495, -0.0790, -0.2709, 0.8375, -0.2060, -0.0824], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2376, 0.1168, 0.1332, -0.4840, -0.2508, -0.0791, -0.0493, -0.4089], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Union[str, Any] , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[33, [-0.0513, 0.0289, 1.3799, 0.2166, -0.2573, -0.0871, 0.5103, -0.0999]],
[47, [-0.4128, -0.1320, -0.3704, 0.1965, -0.4116, -0.2332, -0.3340, 0.2247]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Optional[int] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , fpaa=lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.1609, 0.9866, -0.0487, -0.0777, -0.2716, 0.8368, -0.2055, -0.0814], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2377, 0.1147, 0.1333, -0.4841, -0.2506, -0.0805, -0.0491, -0.4085], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Tuple , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[13, [-0.2051, -0.1803, -0.2311, -0.2114, -0.3292, -0.3574, -0.2953, -0.3323]],
[37, [-0.2632, -0.2625, -0.2199, -0.2741, -0.4539, -0.4990, -0.3720, -0.4925]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Dict , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-3 )
@parameterized.expand(
[
# fmt: off
[27, [-0.0369, 0.0207, -0.0776, -0.0682, -0.1747, -0.1930, -0.1465, -0.2039]],
[16, [-0.1628, -0.2134, -0.2747, -0.2642, -0.3774, -0.4404, -0.3687, -0.4277]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Union[str, Any] , lowercase_ : Any , lowercase_ : Optional[Any] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=5e-3 )
@parameterized.expand([(13,), (16,), (27,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : List[str] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-1 )
@parameterized.expand([(13,), (16,), (37,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.3001, 0.0918, -2.6984, -3.9720, -3.2099, -5.0353, 1.7338, -0.2065, 3.4267]],
[47, [-1.5030, -4.3871, -6.0355, -9.1157, -1.6661, -2.7853, 2.1607, -5.0823, 2.5633]],
# fmt: on
] )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model.encode(lowercase_ ).latent_dist
snake_case_ = dist.sample(generator=lowercase_ )
assert list(sample.shape ) == [image.shape[0], 4] + [i // 8 for i in image.shape[2:]]
snake_case_ = sample[0, -1, -3:, -3:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
snake_case_ = 3e-3 if torch_device != '''mps''' else 1e-2
assert torch_all_close(lowercase_ , lowercase_ , atol=lowercase_ )
| 56
| 1
|
'''simple docstring'''
import itertools
import random
import unittest
import numpy as np
from transformers import WAV_2_VEC_2_PRETRAINED_MODEL_ARCHIVE_LIST, WavaVecaConfig, WavaVecaFeatureExtractor
from transformers.testing_utils import require_torch, slow
from ...test_sequence_feature_extraction_common import SequenceFeatureExtractionTestMixin
a : List[str] = random.Random()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=1.0, __UpperCAmelCase=None, __UpperCAmelCase=None ) -> int:
'''simple docstring'''
if rng is None:
snake_case_ = global_rng
snake_case_ = []
for batch_idx in range(shape[0] ):
values.append([] )
for _ in range(shape[1] ):
values[-1].append(rng.random() * scale )
return values
class a ( unittest.TestCase ):
def __init__( self : Tuple , lowercase_ : Dict , lowercase_ : Union[str, Any]=7 , lowercase_ : Any=400 , lowercase_ : Tuple=2000 , lowercase_ : int=1 , lowercase_ : Any=0.0 , lowercase_ : Optional[Any]=1_6000 , lowercase_ : List[str]=True , lowercase_ : Optional[Any]=True , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = min_seq_length
snake_case_ = max_seq_length
snake_case_ = (self.max_seq_length - self.min_seq_length) // (self.batch_size - 1)
snake_case_ = feature_size
snake_case_ = padding_value
snake_case_ = sampling_rate
snake_case_ = return_attention_mask
snake_case_ = do_normalize
def A_ ( self : int ):
return {
"feature_size": self.feature_size,
"padding_value": self.padding_value,
"sampling_rate": self.sampling_rate,
"return_attention_mask": self.return_attention_mask,
"do_normalize": self.do_normalize,
}
def A_ ( self : Tuple , lowercase_ : Any=False , lowercase_ : Dict=False ):
def _flatten(lowercase_ : Tuple ):
return list(itertools.chain(*lowercase_ ) )
if equal_length:
snake_case_ = floats_list((self.batch_size, self.max_seq_length) )
else:
# make sure that inputs increase in size
snake_case_ = [
_flatten(floats_list((x, self.feature_size) ) )
for x in range(self.min_seq_length , self.max_seq_length , self.seq_length_diff )
]
if numpify:
snake_case_ = [np.asarray(lowercase_ ) for x in speech_inputs]
return speech_inputs
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = WavaVecaFeatureExtractor
def A_ ( self : Any ):
snake_case_ = WavaVecaFeatureExtractionTester(self )
def A_ ( self : Dict , lowercase_ : Any ):
self.assertTrue(np.all(np.mean(lowercase_ , axis=0 ) < 1e-3 ) )
self.assertTrue(np.all(np.abs(np.var(lowercase_ , axis=0 ) - 1 ) < 1e-3 ) )
def A_ ( self : Union[str, Any] ):
# Tests that all call wrap to encode_plus and batch_encode_plus
snake_case_ = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
# create three inputs of length 800, 1000, and 1200
snake_case_ = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
snake_case_ = [np.asarray(lowercase_ ) for speech_input in speech_inputs]
# Test not batched input
snake_case_ = feat_extract(speech_inputs[0] , return_tensors='''np''' ).input_values
snake_case_ = feat_extract(np_speech_inputs[0] , return_tensors='''np''' ).input_values
self.assertTrue(np.allclose(lowercase_ , lowercase_ , atol=1e-3 ) )
# Test batched
snake_case_ = feat_extract(lowercase_ , return_tensors='''np''' ).input_values
snake_case_ = feat_extract(lowercase_ , return_tensors='''np''' ).input_values
for enc_seq_a, enc_seq_a in zip(lowercase_ , lowercase_ ):
self.assertTrue(np.allclose(lowercase_ , lowercase_ , atol=1e-3 ) )
# Test 2-D numpy arrays are batched.
snake_case_ = [floats_list((1, x) )[0] for x in (800, 800, 800)]
snake_case_ = np.asarray(lowercase_ )
snake_case_ = feat_extract(lowercase_ , return_tensors='''np''' ).input_values
snake_case_ = feat_extract(lowercase_ , return_tensors='''np''' ).input_values
for enc_seq_a, enc_seq_a in zip(lowercase_ , lowercase_ ):
self.assertTrue(np.allclose(lowercase_ , lowercase_ , atol=1e-3 ) )
def A_ ( self : Tuple ):
snake_case_ = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
snake_case_ = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
snake_case_ = ['''longest''', '''max_length''', '''do_not_pad''']
snake_case_ = [None, 1600, None]
for max_length, padding in zip(lowercase_ , lowercase_ ):
snake_case_ = feat_extract(lowercase_ , padding=lowercase_ , max_length=lowercase_ , return_tensors='''np''' )
snake_case_ = processed.input_values
self._check_zero_mean_unit_variance(input_values[0][:800] )
self.assertTrue(input_values[0][800:].sum() < 1e-6 )
self._check_zero_mean_unit_variance(input_values[1][:1000] )
self.assertTrue(input_values[0][1000:].sum() < 1e-6 )
self._check_zero_mean_unit_variance(input_values[2][:1200] )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
snake_case_ = range(800 , 1400 , 200 )
snake_case_ = [floats_list((1, x) )[0] for x in lengths]
snake_case_ = ['''longest''', '''max_length''', '''do_not_pad''']
snake_case_ = [None, 1600, None]
for max_length, padding in zip(lowercase_ , lowercase_ ):
snake_case_ = feat_extract(lowercase_ , max_length=lowercase_ , padding=lowercase_ )
snake_case_ = processed.input_values
self._check_zero_mean_unit_variance(input_values[0][:800] )
self._check_zero_mean_unit_variance(input_values[1][:1000] )
self._check_zero_mean_unit_variance(input_values[2][:1200] )
def A_ ( self : List[str] ):
snake_case_ = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
snake_case_ = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
snake_case_ = feat_extract(
lowercase_ , truncation=lowercase_ , max_length=1000 , padding='''max_length''' , return_tensors='''np''' )
snake_case_ = processed.input_values
self._check_zero_mean_unit_variance(input_values[0, :800] )
self._check_zero_mean_unit_variance(input_values[1] )
self._check_zero_mean_unit_variance(input_values[2] )
def A_ ( self : int ):
snake_case_ = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
snake_case_ = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
snake_case_ = feat_extract(
lowercase_ , truncation=lowercase_ , max_length=1000 , padding='''longest''' , return_tensors='''np''' )
snake_case_ = processed.input_values
self._check_zero_mean_unit_variance(input_values[0, :800] )
self._check_zero_mean_unit_variance(input_values[1, :1000] )
self._check_zero_mean_unit_variance(input_values[2] )
# make sure that if max_length < longest -> then pad to max_length
self.assertTrue(input_values.shape == (3, 1000) )
snake_case_ = [floats_list((1, x) )[0] for x in range(800 , 1400 , 200 )]
snake_case_ = feat_extract(
lowercase_ , truncation=lowercase_ , max_length=2000 , padding='''longest''' , return_tensors='''np''' )
snake_case_ = processed.input_values
self._check_zero_mean_unit_variance(input_values[0, :800] )
self._check_zero_mean_unit_variance(input_values[1, :1000] )
self._check_zero_mean_unit_variance(input_values[2] )
# make sure that if max_length > longest -> then pad to longest
self.assertTrue(input_values.shape == (3, 1200) )
@require_torch
def A_ ( self : Dict ):
import torch
snake_case_ = self.feature_extraction_class(**self.feat_extract_tester.prepare_feat_extract_dict() )
snake_case_ = np.random.rand(100 ).astype(np.floataa )
snake_case_ = np_speech_inputs.tolist()
for inputs in [py_speech_inputs, np_speech_inputs]:
snake_case_ = feature_extractor.pad([{'''input_values''': inputs}] , return_tensors='''np''' )
self.assertTrue(np_processed.input_values.dtype == np.floataa )
snake_case_ = feature_extractor.pad([{'''input_values''': inputs}] , return_tensors='''pt''' )
self.assertTrue(pt_processed.input_values.dtype == torch.floataa )
@slow
@require_torch
def A_ ( self : Optional[int] ):
# this test makes sure that models that are using
# group norm don't have their feature extractor return the
# attention_mask
for model_id in WAV_2_VEC_2_PRETRAINED_MODEL_ARCHIVE_LIST:
snake_case_ = WavaVecaConfig.from_pretrained(lowercase_ )
snake_case_ = WavaVecaFeatureExtractor.from_pretrained(lowercase_ )
# only "layer" feature extraction norm should make use of
# attention_mask
self.assertEqual(feat_extract.return_attention_mask , config.feat_extract_norm == '''layer''' )
| 56
|
'''simple docstring'''
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 ( _lowerCamelCase ):
snake_case_ = 42
@flax_register_to_config
class a ( nn.Module , _lowerCamelCase , _lowerCamelCase ):
snake_case_ = 32
snake_case_ = 4
snake_case_ = 4
snake_case_ = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
snake_case_ = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
snake_case_ = False
snake_case_ = (320, 640, 1_280, 1_280)
snake_case_ = 2
snake_case_ = 8
snake_case_ = None
snake_case_ = 1_280
snake_case_ = 0.0
snake_case_ = False
snake_case_ = jnp.floataa
snake_case_ = True
snake_case_ = 0
snake_case_ = False
def A_ ( self : Optional[int] , lowercase_ : jax.random.KeyArray ):
# init input tensors
snake_case_ = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case_ = jnp.zeros(lowercase_ , dtype=jnp.floataa )
snake_case_ = jnp.ones((1,) , dtype=jnp.intaa )
snake_case_ = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case_ ,snake_case_ = jax.random.split(lowercase_ )
snake_case_ = {'''params''': params_rng, '''dropout''': dropout_rng}
return self.init(lowercase_ , lowercase_ , lowercase_ , lowercase_ )["params"]
def A_ ( self : List[str] ):
snake_case_ = self.block_out_channels
snake_case_ = 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.
snake_case_ = self.num_attention_heads or self.attention_head_dim
# input
snake_case_ = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case_ = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case_ = FlaxTimestepEmbedding(lowercase_ , dtype=self.dtype )
snake_case_ = self.only_cross_attention
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (only_cross_attention,) * len(self.down_block_types )
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case_ = []
snake_case_ = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case_ = output_channel
snake_case_ = block_out_channels[i]
snake_case_ = i == len(lowercase_ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = down_blocks
# mid
snake_case_ = 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
snake_case_ = []
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case_ = output_channel
snake_case_ = reversed_block_out_channels[i]
snake_case_ = reversed_block_out_channels[min(i + 1 , len(lowercase_ ) - 1 )]
snake_case_ = i == len(lowercase_ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = output_channel
snake_case_ = up_blocks
# out
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : int , lowercase_ : Any , lowercase_ : int=None , lowercase_ : Any=None , lowercase_ : bool = True , lowercase_ : bool = False , ):
# 1. time
if not isinstance(lowercase_ , jnp.ndarray ):
snake_case_ = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(lowercase_ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case_ = timesteps.astype(dtype=jnp.floataa )
snake_case_ = jnp.expand_dims(lowercase_ , 0 )
snake_case_ = self.time_proj(lowercase_ )
snake_case_ = self.time_embedding(lowercase_ )
# 2. pre-process
snake_case_ = jnp.transpose(lowercase_ , (0, 2, 3, 1) )
snake_case_ = self.conv_in(lowercase_ )
# 3. down
snake_case_ = (sample,)
for down_block in self.down_blocks:
if isinstance(lowercase_ , lowercase_ ):
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , lowercase_ , deterministic=not train )
else:
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case_ = ()
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,)
snake_case_ = new_down_block_res_samples
# 4. mid
snake_case_ = 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:
snake_case_ = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case_ = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = up_block(
lowercase_ , temb=lowercase_ , encoder_hidden_states=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train , )
else:
snake_case_ = up_block(lowercase_ , temb=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train )
# 6. post-process
snake_case_ = self.conv_norm_out(lowercase_ )
snake_case_ = nn.silu(lowercase_ )
snake_case_ = self.conv_out(lowercase_ )
snake_case_ = jnp.transpose(lowercase_ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=lowercase_ )
| 56
| 1
|
'''simple docstring'''
import qiskit
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> qiskit.result.counts.Counts:
'''simple docstring'''
snake_case_ = qiskit.Aer.get_backend('''aer_simulator''' )
# Create a Quantum Circuit acting on the q register
snake_case_ = qiskit.QuantumCircuit(__UpperCAmelCase, __UpperCAmelCase )
# Map the quantum measurement to the classical bits
circuit.measure([0], [0] )
# Execute the circuit on the simulator
snake_case_ = qiskit.execute(__UpperCAmelCase, __UpperCAmelCase, shots=1000 )
# Return the histogram data of the results of the experiment.
return job.result().get_counts(__UpperCAmelCase )
if __name__ == "__main__":
print(f'''Total count for various states are: {single_qubit_measure(1, 1)}''')
| 56
|
'''simple docstring'''
import glob
import os
import random
from string import ascii_lowercase, digits
import cva
import numpy as np
# Parrameters
a : Dict = (720, 1280) # Height, Width
a : Tuple = (0.4, 0.6) # if height or width lower than this scale, drop it.
a : Dict = 1 / 100
a : str = ''
a : Any = ''
a : Optional[int] = ''
a : List[str] = 250
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dataset(__UpperCAmelCase, __UpperCAmelCase )
for index in range(__UpperCAmelCase ):
snake_case_ = random.sample(range(len(__UpperCAmelCase ) ), 4 )
snake_case_ ,snake_case_ ,snake_case_ = update_image_and_anno(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, filter_scale=__UpperCAmelCase, )
# Get random string code: '7b7ad245cdff75241935e4dd860f3bad'
snake_case_ = random_chars(32 )
snake_case_ = path.split(os.sep )[-1].rsplit('''.''', 1 )[0]
snake_case_ = F"{OUTPUT_DIR}/{file_name}_MOSAIC_{letter_code}"
cva.imwrite(F"{file_root}.jpg", __UpperCAmelCase, [cva.IMWRITE_JPEG_QUALITY, 85] )
print(F"Succeeded {index+1}/{NUMBER_IMAGES} with {file_name}" )
snake_case_ = []
for anno in new_annos:
snake_case_ = anno[3] - anno[1]
snake_case_ = anno[4] - anno[2]
snake_case_ = anno[1] + width / 2
snake_case_ = anno[2] + height / 2
snake_case_ = F"{anno[0]} {x_center} {y_center} {width} {height}"
annos_list.append(__UpperCAmelCase )
with open(F"{file_root}.txt", '''w''' ) as outfile:
outfile.write('''\n'''.join(line for line in annos_list ) )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> tuple[list, list]:
'''simple docstring'''
snake_case_ = []
snake_case_ = []
for label_file in glob.glob(os.path.join(__UpperCAmelCase, '''*.txt''' ) ):
snake_case_ = label_file.split(os.sep )[-1].rsplit('''.''', 1 )[0]
with open(__UpperCAmelCase ) as in_file:
snake_case_ = in_file.readlines()
snake_case_ = os.path.join(__UpperCAmelCase, F"{label_name}.jpg" )
snake_case_ = []
for obj_list in obj_lists:
snake_case_ = obj_list.rstrip('''\n''' ).split(''' ''' )
snake_case_ = float(obj[1] ) - float(obj[3] ) / 2
snake_case_ = float(obj[2] ) - float(obj[4] ) / 2
snake_case_ = float(obj[1] ) + float(obj[3] ) / 2
snake_case_ = float(obj[2] ) + float(obj[4] ) / 2
boxes.append([int(obj[0] ), xmin, ymin, xmax, ymax] )
if not boxes:
continue
img_paths.append(__UpperCAmelCase )
labels.append(__UpperCAmelCase )
return img_paths, labels
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, ) -> tuple[list, list, str]:
'''simple docstring'''
snake_case_ = np.zeros([output_size[0], output_size[1], 3], dtype=np.uinta )
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = int(scale_x * output_size[1] )
snake_case_ = int(scale_y * output_size[0] )
snake_case_ = []
snake_case_ = []
for i, index in enumerate(__UpperCAmelCase ):
snake_case_ = all_img_list[index]
path_list.append(__UpperCAmelCase )
snake_case_ = all_annos[index]
snake_case_ = cva.imread(__UpperCAmelCase )
if i == 0: # top-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = bbox[2] * scale_y
snake_case_ = bbox[3] * scale_x
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 1: # top-right
snake_case_ = cva.resize(__UpperCAmelCase, (output_size[1] - divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = bbox[2] * scale_y
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 2: # bottom-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = bbox[3] * scale_x
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
else: # bottom-right
snake_case_ = cva.resize(
__UpperCAmelCase, (output_size[1] - divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
# Remove bounding box small than scale of filter
if filter_scale > 0:
snake_case_ = [
anno
for anno in new_anno
if filter_scale < (anno[3] - anno[1]) and filter_scale < (anno[4] - anno[2])
]
return output_img, new_anno, path_list[0]
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
assert number_char > 1, "The number of character should greater than 1"
snake_case_ = ascii_lowercase + digits
return "".join(random.choice(__UpperCAmelCase ) for _ in range(__UpperCAmelCase ) )
if __name__ == "__main__":
main()
print('DONE ✅')
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
if number < 0:
raise ValueError('''number must not be negative''' )
return number & (number - 1) == 0
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
import unittest
from transformers import (
MODEL_FOR_OBJECT_DETECTION_MAPPING,
AutoFeatureExtractor,
AutoModelForObjectDetection,
ObjectDetectionPipeline,
is_vision_available,
pipeline,
)
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_pytesseract,
require_tf,
require_timm,
require_torch,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_vision_available():
from PIL import Image
else:
class a :
@staticmethod
def A_ ( *lowercase_ : int , **lowercase_ : str ):
pass
@is_pipeline_test
@require_vision
@require_timm
@require_torch
class a ( unittest.TestCase ):
snake_case_ = MODEL_FOR_OBJECT_DETECTION_MAPPING
def A_ ( self : Any , lowercase_ : List[Any] , lowercase_ : Optional[int] , lowercase_ : List[str] ):
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , image_processor=lowercase_ )
return object_detector, ["./tests/fixtures/tests_samples/COCO/000000039769.png"]
def A_ ( self : Dict , lowercase_ : List[Any] , lowercase_ : int ):
snake_case_ = object_detector('''./tests/fixtures/tests_samples/COCO/000000039769.png''' , threshold=0.0 )
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
import datasets
snake_case_ = datasets.load_dataset('''hf-internal-testing/fixtures_image_utils''' , '''image''' , split='''test''' )
snake_case_ = [
Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' ),
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
# RGBA
dataset[0]['''file'''],
# LA
dataset[1]['''file'''],
# L
dataset[2]['''file'''],
]
snake_case_ = object_detector(lowercase_ , threshold=0.0 )
self.assertEqual(len(lowercase_ ) , len(lowercase_ ) )
for outputs in batch_outputs:
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
@require_tf
@unittest.skip('''Object detection not implemented in TF''' )
def A_ ( self : int ):
pass
@require_torch
def A_ ( self : Tuple ):
snake_case_ = '''hf-internal-testing/tiny-detr-mobilenetsv3'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=0.0 )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] , threshold=0.0 , )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
] , )
@require_torch
@slow
def A_ ( self : Optional[int] ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : Tuple ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : str ):
snake_case_ = 0.9985
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=lowercase_ )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
@require_torch
@require_pytesseract
@slow
def A_ ( self : Dict ):
snake_case_ = '''Narsil/layoutlmv3-finetuned-funsd'''
snake_case_ = 0.9993
snake_case_ = pipeline('''object-detection''' , model=lowercase_ , threshold=lowercase_ )
snake_case_ = object_detector(
'''https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
] , )
| 56
| 1
|
'''simple docstring'''
import copy
from dataclasses import dataclass, field
from typing import ClassVar, Dict
from ..features import Audio, Features, Value
from .base import TaskTemplate
@dataclass(frozen=_lowerCamelCase )
class a ( _lowerCamelCase ):
snake_case_ = field(default="automatic-speech-recognition" , metadata={"include_in_asdict_even_if_is_default": True} )
snake_case_ = Features({"audio": Audio()} )
snake_case_ = Features({"transcription": Value("string" )} )
snake_case_ = "audio"
snake_case_ = "transcription"
def A_ ( self : Any , lowercase_ : Optional[int] ):
if self.audio_column not in features:
raise ValueError(F"Column {self.audio_column} is not present in features." )
if not isinstance(features[self.audio_column] , lowercase_ ):
raise ValueError(F"Column {self.audio_column} is not an Audio type." )
snake_case_ = copy.deepcopy(self )
snake_case_ = self.input_schema.copy()
snake_case_ = features[self.audio_column]
snake_case_ = input_schema
return task_template
@property
def A_ ( self : Any ):
return {self.audio_column: "audio", self.transcription_column: "transcription"}
| 56
|
'''simple docstring'''
import unittest
from transformers import MPNetConfig, is_torch_available
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 (
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
class a :
def __init__( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Any=13 , lowercase_ : Optional[Any]=7 , lowercase_ : Optional[Any]=True , lowercase_ : Dict=True , lowercase_ : Tuple=False , lowercase_ : Optional[Any]=True , lowercase_ : Any=99 , lowercase_ : Union[str, Any]=64 , lowercase_ : str=5 , lowercase_ : int=4 , lowercase_ : List[Any]=64 , lowercase_ : Dict="gelu" , lowercase_ : Optional[int]=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Tuple=512 , lowercase_ : List[Any]=16 , lowercase_ : str=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[Any]=3 , lowercase_ : Optional[Any]=4 , lowercase_ : List[Any]=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_input_mask
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = scope
def A_ ( self : List[str] ):
return MPNetConfig.from_pretrained('''microsoft/mpnet-base''' )
def A_ ( self : str ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_input_mask:
snake_case_ = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = ids_tensor([self.batch_size] , self.num_choices )
snake_case_ = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def A_ ( self : Tuple ):
return MPNetConfig(
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 , initializer_range=self.initializer_range , )
def A_ ( self : Any , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Optional[int] ):
snake_case_ = MPNetModel(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , lowercase_ )
snake_case_ = 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 : str , lowercase_ : Optional[Any] , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = MPNetForQuestionAnswering(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Tuple , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : str , lowercase_ : Optional[int] , lowercase_ : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = MPNetForSequenceClassification(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , attention_mask=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : Any , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.num_choices
snake_case_ = MPNetForMultipleChoice(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = model(
lowercase_ , attention_mask=lowercase_ , labels=lowercase_ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def A_ ( self : Union[str, Any] , lowercase_ : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] , lowercase_ : str , lowercase_ : int , lowercase_ : List[str] ):
snake_case_ = self.num_labels
snake_case_ = MPNetForTokenClassification(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Union[str, Any] ):
snake_case_ = self.prepare_config_and_inputs()
((snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_)) = config_and_inputs
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
if is_torch_available()
else ()
)
snake_case_ = (
{
"feature-extraction": MPNetModel,
"fill-mask": MPNetForMaskedLM,
"question-answering": MPNetForQuestionAnswering,
"text-classification": MPNetForSequenceClassification,
"token-classification": MPNetForTokenClassification,
"zero-shot": MPNetForSequenceClassification,
}
if is_torch_available()
else {}
)
snake_case_ = False
snake_case_ = True
def A_ ( self : Tuple ):
snake_case_ = MPNetModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , hidden_size=37 )
def A_ ( self : Union[str, Any] ):
self.config_tester.run_common_tests()
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_model(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_sequence_classification(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_multiple_choice(*lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_token_classification(*lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_question_answering(*lowercase_ )
@require_torch
class a ( unittest.TestCase ):
@slow
def A_ ( self : List[Any] ):
snake_case_ = MPNetModel.from_pretrained('''microsoft/mpnet-base''' )
snake_case_ = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]] )
snake_case_ = model(lowercase_ )[0]
snake_case_ = torch.Size((1, 11, 768) )
self.assertEqual(output.shape , lowercase_ )
snake_case_ = torch.tensor(
[[[-0.0550, 0.1943, -0.0740], [-0.0562, 0.2211, -0.0579], [-0.0437, 0.3337, -0.0641]]] )
# compare the actual values for a slice.
self.assertTrue(torch.allclose(output[:, :3, :3] , lowercase_ , atol=1e-4 ) )
| 56
| 1
|
'''simple docstring'''
import json
import os
import unittest
from transformers.models.roc_bert.tokenization_roc_bert import (
VOCAB_FILES_NAMES,
RoCBertBasicTokenizer,
RoCBertTokenizer,
RoCBertWordpieceTokenizer,
_is_control,
_is_punctuation,
_is_whitespace,
)
from transformers.testing_utils import require_tokenizers, slow
from ...test_tokenization_common import TokenizerTesterMixin, filter_non_english
@require_tokenizers
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = RoCBertTokenizer
snake_case_ = None
snake_case_ = False
snake_case_ = True
snake_case_ = filter_non_english
def A_ ( self : Union[str, Any] ):
super().setUp()
snake_case_ = ['''[UNK]''', '''[CLS]''', '''[SEP]''', '''[PAD]''', '''[MASK]''', '''你''', '''好''', '''是''', '''谁''', '''a''', '''b''', '''c''', '''d''']
snake_case_ = {}
snake_case_ = {}
for i, value in enumerate(lowercase_ ):
snake_case_ = i
snake_case_ = i
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''word_shape_file'''] )
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''word_pronunciation_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
with open(self.word_shape_file , '''w''' , encoding='''utf-8''' ) as word_shape_writer:
json.dump(lowercase_ , lowercase_ , ensure_ascii=lowercase_ )
with open(self.word_pronunciation_file , '''w''' , encoding='''utf-8''' ) as word_pronunciation_writer:
json.dump(lowercase_ , lowercase_ , ensure_ascii=lowercase_ )
def A_ ( self : List[str] ):
snake_case_ = self.tokenizer_class(self.vocab_file , self.word_shape_file , self.word_pronunciation_file )
snake_case_ = tokenizer.tokenize('''你好[SEP]你是谁''' )
self.assertListEqual(lowercase_ , ['''你''', '''好''', '''[SEP]''', '''你''', '''是''', '''谁'''] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(lowercase_ ) , [5, 6, 2, 5, 7, 8] )
self.assertListEqual(tokenizer.convert_tokens_to_shape_ids(lowercase_ ) , [5, 6, 2, 5, 7, 8] )
self.assertListEqual(tokenizer.convert_tokens_to_pronunciation_ids(lowercase_ ) , [5, 6, 2, 5, 7, 8] )
def A_ ( self : str ):
snake_case_ = RoCBertBasicTokenizer()
self.assertListEqual(tokenizer.tokenize('''ah\u535A\u63A8zz''' ) , ['''ah''', '''\u535A''', '''\u63A8''', '''zz'''] )
def A_ ( self : Tuple ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHeLLo!how \n Are yoU? ''' ) , ['''hello''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''hello'''] )
def A_ ( self : Union[str, Any] ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''hällo''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''h\u00E9llo'''] )
def A_ ( self : str ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''hallo''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''hello'''] )
def A_ ( self : List[str] ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''hallo''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''hello'''] )
def A_ ( self : int ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHeLLo!how \n Are yoU? ''' ) , ['''HeLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?'''] )
def A_ ( self : List[str] ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''HäLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?'''] )
def A_ ( self : Union[str, Any] ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''HaLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?'''] )
def A_ ( self : Union[str, Any] ):
snake_case_ = RoCBertBasicTokenizer(do_lower_case=lowercase_ , never_split=['''[UNK]'''] )
self.assertListEqual(
tokenizer.tokenize(''' \tHeLLo!how \n Are yoU? [UNK]''' ) , ['''HeLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?''', '''[UNK]'''] )
def A_ ( self : Union[str, Any] ):
snake_case_ = ['''[UNK]''', '''[CLS]''', '''[SEP]''', '''want''', '''##want''', '''##ed''', '''wa''', '''un''', '''runn''', '''##ing''']
snake_case_ = {}
for i, token in enumerate(lowercase_ ):
snake_case_ = i
snake_case_ = RoCBertWordpieceTokenizer(vocab=lowercase_ , unk_token='''[UNK]''' )
self.assertListEqual(tokenizer.tokenize('''''' ) , [] )
self.assertListEqual(tokenizer.tokenize('''unwanted running''' ) , ['''un''', '''##want''', '''##ed''', '''runn''', '''##ing'''] )
self.assertListEqual(tokenizer.tokenize('''unwantedX running''' ) , ['''[UNK]''', '''runn''', '''##ing'''] )
def A_ ( self : Optional[int] ):
self.assertTrue(_is_whitespace(''' ''' ) )
self.assertTrue(_is_whitespace('''\t''' ) )
self.assertTrue(_is_whitespace('''\r''' ) )
self.assertTrue(_is_whitespace('''\n''' ) )
self.assertTrue(_is_whitespace('''\u00A0''' ) )
self.assertFalse(_is_whitespace('''A''' ) )
self.assertFalse(_is_whitespace('''-''' ) )
def A_ ( self : List[Any] ):
self.assertTrue(_is_control('''\u0005''' ) )
self.assertFalse(_is_control('''A''' ) )
self.assertFalse(_is_control(''' ''' ) )
self.assertFalse(_is_control('''\t''' ) )
self.assertFalse(_is_control('''\r''' ) )
def A_ ( self : List[str] ):
self.assertTrue(_is_punctuation('''-''' ) )
self.assertTrue(_is_punctuation('''$''' ) )
self.assertTrue(_is_punctuation('''`''' ) )
self.assertTrue(_is_punctuation('''.''' ) )
self.assertFalse(_is_punctuation('''A''' ) )
self.assertFalse(_is_punctuation(''' ''' ) )
def A_ ( self : str ):
snake_case_ = self.get_tokenizer()
# Example taken from the issue https://github.com/huggingface/tokenizers/issues/340
self.assertListEqual([tokenizer.tokenize(lowercase_ ) for t in ['''Test''', '''\xad''', '''test''']] , [['''[UNK]'''], [], ['''[UNK]''']] )
if self.test_rust_tokenizer:
snake_case_ = self.get_rust_tokenizer()
self.assertListEqual(
[rust_tokenizer.tokenize(lowercase_ ) for t in ['''Test''', '''\xad''', '''test''']] , [['''[UNK]'''], [], ['''[UNK]''']] )
def A_ ( self : Dict ):
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(F"{tokenizer.__class__.__name__} ({pretrained_name})" ):
snake_case_ = self.rust_tokenizer_class.from_pretrained(lowercase_ , **lowercase_ )
snake_case_ = F"A, naïve {tokenizer_r.mask_token} AllenNLP sentence."
snake_case_ = tokenizer_r.encode_plus(
lowercase_ , return_attention_mask=lowercase_ , return_token_type_ids=lowercase_ , return_offsets_mapping=lowercase_ , add_special_tokens=lowercase_ , )
snake_case_ = tokenizer_r.do_lower_case if hasattr(lowercase_ , '''do_lower_case''' ) else False
snake_case_ = (
[
((0, 0), tokenizer_r.cls_token),
((0, 1), '''A'''),
((1, 2), ''','''),
((3, 5), '''na'''),
((5, 6), '''##ï'''),
((6, 8), '''##ve'''),
((9, 15), tokenizer_r.mask_token),
((16, 21), '''Allen'''),
((21, 23), '''##NL'''),
((23, 24), '''##P'''),
((25, 33), '''sentence'''),
((33, 34), '''.'''),
((0, 0), tokenizer_r.sep_token),
]
if not do_lower_case
else [
((0, 0), tokenizer_r.cls_token),
((0, 1), '''a'''),
((1, 2), ''','''),
((3, 8), '''naive'''),
((9, 15), tokenizer_r.mask_token),
((16, 21), '''allen'''),
((21, 23), '''##nl'''),
((23, 24), '''##p'''),
((25, 33), '''sentence'''),
((33, 34), '''.'''),
((0, 0), tokenizer_r.sep_token),
]
)
self.assertEqual(
[e[1] for e in expected_results] , tokenizer_r.convert_ids_to_tokens(tokens['''input_ids'''] ) )
self.assertEqual([e[0] for e in expected_results] , tokens['''offset_mapping'''] )
def A_ ( self : Union[str, Any] ):
snake_case_ = ['''的''', '''人''', '''有''']
snake_case_ = ''''''.join(lowercase_ )
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(F"{tokenizer.__class__.__name__} ({pretrained_name})" ):
snake_case_ = True
snake_case_ = self.tokenizer_class.from_pretrained(lowercase_ , **lowercase_ )
snake_case_ = self.rust_tokenizer_class.from_pretrained(lowercase_ , **lowercase_ )
snake_case_ = tokenizer_p.encode(lowercase_ , add_special_tokens=lowercase_ )
snake_case_ = tokenizer_r.encode(lowercase_ , add_special_tokens=lowercase_ )
snake_case_ = tokenizer_r.convert_ids_to_tokens(lowercase_ )
snake_case_ = tokenizer_p.convert_ids_to_tokens(lowercase_ )
# it is expected that each Chinese character is not preceded by "##"
self.assertListEqual(lowercase_ , lowercase_ )
self.assertListEqual(lowercase_ , lowercase_ )
snake_case_ = False
snake_case_ = self.rust_tokenizer_class.from_pretrained(lowercase_ , **lowercase_ )
snake_case_ = self.tokenizer_class.from_pretrained(lowercase_ , **lowercase_ )
snake_case_ = tokenizer_r.encode(lowercase_ , add_special_tokens=lowercase_ )
snake_case_ = tokenizer_p.encode(lowercase_ , add_special_tokens=lowercase_ )
snake_case_ = tokenizer_r.convert_ids_to_tokens(lowercase_ )
snake_case_ = tokenizer_p.convert_ids_to_tokens(lowercase_ )
# it is expected that only the first Chinese character is not preceded by "##".
snake_case_ = [
F"##{token}" if idx != 0 else token for idx, token in enumerate(lowercase_ )
]
self.assertListEqual(lowercase_ , lowercase_ )
self.assertListEqual(lowercase_ , lowercase_ )
@slow
def A_ ( self : Tuple ):
snake_case_ = self.tokenizer_class(self.vocab_file , self.word_shape_file , self.word_pronunciation_file )
snake_case_ = tokenizer.encode('''你好''' , add_special_tokens=lowercase_ )
snake_case_ = tokenizer.encode('''你是谁''' , add_special_tokens=lowercase_ )
snake_case_ = tokenizer.build_inputs_with_special_tokens(lowercase_ )
snake_case_ = tokenizer.build_inputs_with_special_tokens(lowercase_ , lowercase_ )
assert encoded_sentence == [1] + text + [2]
assert encoded_pair == [1] + text + [2] + text_a + [2]
def A_ ( self : List[Any] ):
snake_case_ = self.get_tokenizers(do_lower_case=lowercase_ )
for tokenizer in tokenizers:
with self.subTest(F"{tokenizer.__class__.__name__}" ):
snake_case_ = '''你好，你是谁'''
snake_case_ = tokenizer.tokenize(lowercase_ )
snake_case_ = tokenizer.convert_tokens_to_ids(lowercase_ )
snake_case_ = tokenizer.convert_tokens_to_shape_ids(lowercase_ )
snake_case_ = tokenizer.convert_tokens_to_pronunciation_ids(lowercase_ )
snake_case_ = tokenizer.prepare_for_model(
lowercase_ , lowercase_ , lowercase_ , add_special_tokens=lowercase_ )
snake_case_ = tokenizer.encode_plus(lowercase_ , add_special_tokens=lowercase_ )
self.assertEqual(lowercase_ , lowercase_ )
| 56
|
'''simple docstring'''
import json
import os
import pickle
import shutil
import tempfile
from unittest import TestCase
from unittest.mock import patch
import numpy as np
from datasets import Dataset
from transformers import is_faiss_available
from transformers.models.bart.configuration_bart import BartConfig
from transformers.models.bart.tokenization_bart import BartTokenizer
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES as DPR_VOCAB_FILES_NAMES
from transformers.models.dpr.configuration_dpr import DPRConfig
from transformers.models.dpr.tokenization_dpr import DPRContextEncoderTokenizer, DPRQuestionEncoderTokenizer
from transformers.models.rag.configuration_rag import RagConfig
from transformers.models.rag.retrieval_rag import CustomHFIndex, RagRetriever
from transformers.models.roberta.tokenization_roberta import VOCAB_FILES_NAMES as BART_VOCAB_FILES_NAMES
from transformers.testing_utils import require_faiss, require_sentencepiece, require_tokenizers, require_torch
if is_faiss_available():
import faiss
@require_faiss
class a ( _lowerCamelCase ):
def A_ ( self : str ):
snake_case_ = tempfile.mkdtemp()
snake_case_ = 8
# DPR tok
snake_case_ = [
'''[UNK]''',
'''[CLS]''',
'''[SEP]''',
'''[PAD]''',
'''[MASK]''',
'''want''',
'''##want''',
'''##ed''',
'''wa''',
'''un''',
'''runn''',
'''##ing''',
''',''',
'''low''',
'''lowest''',
]
snake_case_ = os.path.join(self.tmpdirname , '''dpr_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , DPR_VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
# BART tok
snake_case_ = [
'''l''',
'''o''',
'''w''',
'''e''',
'''r''',
'''s''',
'''t''',
'''i''',
'''d''',
'''n''',
'''\u0120''',
'''\u0120l''',
'''\u0120n''',
'''\u0120lo''',
'''\u0120low''',
'''er''',
'''\u0120lowest''',
'''\u0120newer''',
'''\u0120wider''',
'''<unk>''',
]
snake_case_ = dict(zip(lowercase_ , range(len(lowercase_ ) ) ) )
snake_case_ = ['''#version: 0.2''', '''\u0120 l''', '''\u0120l o''', '''\u0120lo w''', '''e r''', '''''']
snake_case_ = {'''unk_token''': '''<unk>'''}
snake_case_ = os.path.join(self.tmpdirname , '''bart_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''merges_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write(json.dumps(lowercase_ ) + '''\n''' )
with open(self.merges_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write('''\n'''.join(lowercase_ ) )
def A_ ( self : Union[str, Any] ):
return DPRQuestionEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : Union[str, Any] ):
return DPRContextEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : int ):
return BartTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''bart_tokenizer''' ) )
def A_ ( self : str ):
shutil.rmtree(self.tmpdirname )
def A_ ( self : str ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size ), 2 * np.ones(self.retrieval_vector_size )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
return dataset
def A_ ( self : str ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , )
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
return retriever
def A_ ( self : str , lowercase_ : bool ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''custom''' , )
if from_disk:
snake_case_ = os.path.join(self.tmpdirname , '''dataset''' )
snake_case_ = os.path.join(self.tmpdirname , '''index.faiss''' )
dataset.get_index('''embeddings''' ).save(os.path.join(self.tmpdirname , '''index.faiss''' ) )
dataset.drop_index('''embeddings''' )
dataset.save_to_disk(os.path.join(self.tmpdirname , '''dataset''' ) )
del dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
else:
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , index=CustomHFIndex(config.retrieval_vector_size , lowercase_ ) , )
return retriever
def A_ ( self : Tuple ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size + 1 ), 2 * np.ones(self.retrieval_vector_size + 1 )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
snake_case_ = os.path.join(self.tmpdirname , '''hf_bert_base.hnswSQ8_correct_phi_128.c_index''' )
dataset.save_faiss_index('''embeddings''' , index_file_name + '''.index.dpr''' )
pickle.dump(dataset['''id'''] , open(index_file_name + '''.index_meta.dpr''' , '''wb''' ) )
snake_case_ = os.path.join(self.tmpdirname , '''psgs_w100.tsv.pkl''' )
snake_case_ = {sample['''id''']: [sample['''text'''], sample['''title''']] for sample in dataset}
pickle.dump(lowercase_ , open(lowercase_ , '''wb''' ) )
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''legacy''' , index_path=self.tmpdirname , )
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() )
return retriever
def A_ ( self : Optional[Any] ):
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : str ):
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = self.get_dummy_dataset()
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : int ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : str ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : Any ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : Any ):
snake_case_ = 1
snake_case_ = self.get_dummy_legacy_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''text'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''text'''][0] , '''bar''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''text'''][0] , '''foo''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_legacy_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : List[str] ):
import torch
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
snake_case_ ,snake_case_ ,snake_case_ = (
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , np.ndarray )
snake_case_ = retriever(
lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ , return_tensors='''pt''' , )
snake_case_ ,snake_case_ ,snake_case_ ,snake_case_ = ( # noqa: F841
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
out['''doc_ids'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : Tuple ):
snake_case_ = self.get_dpr_ctx_encoder_tokenizer()
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
retriever.set_ctx_encoder_tokenizer(lowercase_ )
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
self.assertEqual(
len(lowercase_ ) , 6 ) # check whether the retriever output consist of 6 attributes including tokenized docs
self.assertEqual(
all(k in out for k in ('''tokenized_doc_ids''', '''tokenized_doc_attention_mask''') ) , lowercase_ ) # check for doc token related keys in dictionary.
| 56
| 1
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
import diffusers
from diffusers import (
AutoencoderKL,
EulerDiscreteScheduler,
StableDiffusionLatentUpscalePipeline,
StableDiffusionPipeline,
UNetaDConditionModel,
)
from diffusers.schedulers import KarrasDiffusionSchedulers
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS, TEXT_GUIDED_IMAGE_VARIATION_PARAMS
from ..test_pipelines_common import PipelineKarrasSchedulerTesterMixin, PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = [tensor.shape for tensor in tensor_list]
return all(shape == shapes[0] for shape in shapes[1:] )
class a ( _lowerCamelCase , _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = StableDiffusionLatentUpscalePipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {
"height",
"width",
"cross_attention_kwargs",
"negative_prompt_embeds",
"prompt_embeds",
}
snake_case_ = PipelineTesterMixin.required_optional_params - {"num_images_per_prompt"}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS
snake_case_ = frozenset(
[] ) # TO-DO: update image_params once pipeline is refactored with VaeImageProcessor.preprocess
snake_case_ = frozenset([] )
snake_case_ = True
@property
def A_ ( self : List[str] ):
snake_case_ = 1
snake_case_ = 4
snake_case_ = (16, 16)
snake_case_ = floats_tensor((batch_size, num_channels) + sizes , rng=random.Random(0 ) ).to(lowercase_ )
return image
def A_ ( self : Optional[int] ):
torch.manual_seed(0 )
snake_case_ = UNetaDConditionModel(
act_fn='''gelu''' , attention_head_dim=8 , norm_num_groups=lowercase_ , block_out_channels=[32, 32, 64, 64] , time_cond_proj_dim=160 , conv_in_kernel=1 , conv_out_kernel=1 , cross_attention_dim=32 , down_block_types=(
'''KDownBlock2D''',
'''KCrossAttnDownBlock2D''',
'''KCrossAttnDownBlock2D''',
'''KCrossAttnDownBlock2D''',
) , in_channels=8 , mid_block_type=lowercase_ , only_cross_attention=lowercase_ , out_channels=5 , resnet_time_scale_shift='''scale_shift''' , time_embedding_type='''fourier''' , timestep_post_act='''gelu''' , up_block_types=('''KCrossAttnUpBlock2D''', '''KCrossAttnUpBlock2D''', '''KCrossAttnUpBlock2D''', '''KUpBlock2D''') , )
snake_case_ = AutoencoderKL(
block_out_channels=[32, 32, 64, 64] , in_channels=3 , out_channels=3 , down_block_types=[
'''DownEncoderBlock2D''',
'''DownEncoderBlock2D''',
'''DownEncoderBlock2D''',
'''DownEncoderBlock2D''',
] , up_block_types=['''UpDecoderBlock2D''', '''UpDecoderBlock2D''', '''UpDecoderBlock2D''', '''UpDecoderBlock2D'''] , latent_channels=4 , )
snake_case_ = EulerDiscreteScheduler(prediction_type='''sample''' )
snake_case_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , hidden_act='''quick_gelu''' , projection_dim=512 , )
snake_case_ = CLIPTextModel(lowercase_ )
snake_case_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
snake_case_ = {
'''unet''': model.eval(),
'''vae''': vae.eval(),
'''scheduler''': scheduler,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
}
return components
def A_ ( self : Any , lowercase_ : int , lowercase_ : Dict=0 ):
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''prompt''': '''A painting of a squirrel eating a burger''',
'''image''': self.dummy_image.cpu(),
'''generator''': generator,
'''num_inference_steps''': 2,
'''output_type''': '''numpy''',
}
return inputs
def A_ ( self : Tuple ):
snake_case_ = '''cpu'''
snake_case_ = self.get_dummy_components()
snake_case_ = self.pipeline_class(**lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ ).images
snake_case_ = image[0, -3:, -3:, -1]
self.assertEqual(image.shape , (1, 256, 256, 3) )
snake_case_ = np.array(
[0.4722_2412, 0.4192_1633, 0.4471_7434, 0.4687_4192, 0.4258_8258, 0.4615_0726, 0.467_7534, 0.4558_3832, 0.4857_9055] )
snake_case_ = np.abs(image_slice.flatten() - expected_slice ).max()
self.assertLessEqual(lowercase_ , 1e-3 )
def A_ ( self : Optional[int] ):
super().test_attention_slicing_forward_pass(expected_max_diff=7e-3 )
def A_ ( self : str ):
super().test_cpu_offload_forward_pass(expected_max_diff=3e-3 )
def A_ ( self : str ):
super().test_dict_tuple_outputs_equivalent(expected_max_difference=3e-3 )
def A_ ( self : List[Any] ):
super().test_inference_batch_single_identical(expected_max_diff=7e-3 )
def A_ ( self : List[str] ):
super().test_pt_np_pil_outputs_equivalent(expected_max_diff=3e-3 )
def A_ ( self : Tuple ):
super().test_save_load_local(expected_max_difference=3e-3 )
def A_ ( self : int ):
super().test_save_load_optional_components(expected_max_difference=3e-3 )
def A_ ( self : int ):
snake_case_ = [
'''DDIMScheduler''',
'''DDPMScheduler''',
'''PNDMScheduler''',
'''HeunDiscreteScheduler''',
'''EulerAncestralDiscreteScheduler''',
'''KDPM2DiscreteScheduler''',
'''KDPM2AncestralDiscreteScheduler''',
'''DPMSolverSDEScheduler''',
]
snake_case_ = self.get_dummy_components()
snake_case_ = self.pipeline_class(**lowercase_ )
# make sure that PNDM does not need warm-up
pipe.scheduler.register_to_config(skip_prk_steps=lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = 2
snake_case_ = []
for scheduler_enum in KarrasDiffusionSchedulers:
if scheduler_enum.name in skip_schedulers:
# no sigma schedulers are not supported
# no schedulers
continue
snake_case_ = getattr(lowercase_ , scheduler_enum.name )
snake_case_ = scheduler_cls.from_config(pipe.scheduler.config )
snake_case_ = pipe(**lowercase_ )[0]
outputs.append(lowercase_ )
assert check_same_shape(lowercase_ )
@require_torch_gpu
@slow
class a ( unittest.TestCase ):
def A_ ( self : str ):
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Optional[int] ):
snake_case_ = torch.manual_seed(33 )
snake_case_ = StableDiffusionPipeline.from_pretrained('''CompVis/stable-diffusion-v1-4''' , torch_dtype=torch.floataa )
pipe.to('''cuda''' )
snake_case_ = StableDiffusionLatentUpscalePipeline.from_pretrained(
'''stabilityai/sd-x2-latent-upscaler''' , torch_dtype=torch.floataa )
upscaler.to('''cuda''' )
snake_case_ = '''a photo of an astronaut high resolution, unreal engine, ultra realistic'''
snake_case_ = pipe(lowercase_ , generator=lowercase_ , output_type='''latent''' ).images
snake_case_ = upscaler(
prompt=lowercase_ , image=lowercase_ , num_inference_steps=20 , guidance_scale=0 , generator=lowercase_ , output_type='''np''' , ).images[0]
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/latent-upscaler/astronaut_1024.npy''' )
assert np.abs((expected_image - image).mean() ) < 5e-2
def A_ ( self : List[str] ):
snake_case_ = torch.manual_seed(33 )
snake_case_ = StableDiffusionLatentUpscalePipeline.from_pretrained(
'''stabilityai/sd-x2-latent-upscaler''' , torch_dtype=torch.floataa )
upscaler.to('''cuda''' )
snake_case_ = '''the temple of fire by Ross Tran and Gerardo Dottori, oil on canvas'''
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/latent-upscaler/fire_temple_512.png''' )
snake_case_ = upscaler(
prompt=lowercase_ , image=lowercase_ , num_inference_steps=20 , guidance_scale=0 , generator=lowercase_ , output_type='''np''' , ).images[0]
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/latent-upscaler/fire_temple_1024.npy''' )
assert np.abs((expected_image - image).max() ) < 5e-2
| 56
|
'''simple docstring'''
import os
import re
import warnings
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_ta import TaTokenizer
else:
a : Dict = None
a : List[Any] = logging.get_logger(__name__)
a : List[Any] = {'vocab_file': 'spiece.model', 'tokenizer_file': 'tokenizer.json'}
a : str = {
'vocab_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/spiece.model',
't5-base': 'https://huggingface.co/t5-base/resolve/main/spiece.model',
't5-large': 'https://huggingface.co/t5-large/resolve/main/spiece.model',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/spiece.model',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/spiece.model',
},
'tokenizer_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/tokenizer.json',
't5-base': 'https://huggingface.co/t5-base/resolve/main/tokenizer.json',
't5-large': 'https://huggingface.co/t5-large/resolve/main/tokenizer.json',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/tokenizer.json',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/tokenizer.json',
},
}
# TODO(PVP) - this should be removed in Transformers v5
a : List[Any] = {
't5-small': 512,
't5-base': 512,
't5-large': 512,
't5-3b': 512,
't5-11b': 512,
}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
snake_case_ = TaTokenizer
snake_case_ = []
def __init__( self : List[Any] , lowercase_ : int=None , lowercase_ : Dict=None , lowercase_ : Dict="</s>" , lowercase_ : List[Any]="<unk>" , lowercase_ : int="<pad>" , lowercase_ : int=100 , lowercase_ : List[Any]=None , **lowercase_ : List[str] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
snake_case_ = [F"<extra_id_{i}>" for i in range(lowercase_ )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra special tokens
snake_case_ = len(set(filter(lambda lowercase_ : bool('''extra_id_''' in str(lowercase_ ) ) , lowercase_ ) ) )
if extra_tokens != extra_ids:
raise ValueError(
F"Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are"
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
super().__init__(
lowercase_ , tokenizer_file=lowercase_ , eos_token=lowercase_ , unk_token=lowercase_ , pad_token=lowercase_ , extra_ids=lowercase_ , additional_special_tokens=lowercase_ , **lowercase_ , )
snake_case_ = vocab_file
snake_case_ = False if not self.vocab_file else True
snake_case_ = extra_ids
@staticmethod
def A_ ( lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : int ):
if pretrained_model_name_or_path in TaTokenizerFast.max_model_input_sizes:
snake_case_ = TaTokenizerFast.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
F" {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this"
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
F" {pretrained_model_name_or_path} automatically truncating your input to"
F" {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences"
F" longer than {deprecated_max_model_length} you can either instantiate this tokenizer with"
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , lowercase_ , )
return max_model_length
def A_ ( self : Optional[Any] , lowercase_ : str , lowercase_ : Optional[str] = None ):
if not self.can_save_slow_tokenizer:
raise ValueError(
'''Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '''
'''tokenizer.''' )
if not os.path.isdir(lowercase_ ):
logger.error(F"Vocabulary path ({save_directory}) should be a directory" )
return
snake_case_ = 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_ ):
copyfile(self.vocab_file , lowercase_ )
logger.info(F"Copy vocab file to {out_vocab_file}" )
return (out_vocab_file,)
def A_ ( self : Optional[int] , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = token_ids_a + [self.eos_token_id]
if token_ids_a is None:
return self.prefix_tokens + token_ids_a
else:
snake_case_ = token_ids_a + [self.eos_token_id]
return self.prefix_tokens + token_ids_a + token_ids_a
def A_ ( self : int , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def A_ ( self : Dict ):
return list(
set(filter(lambda lowercase_ : bool(re.search(R'''<extra_id_\d+>''' , lowercase_ ) ) is not None , self.additional_special_tokens ) ) )
def A_ ( self : Any ):
return [self.convert_tokens_to_ids(lowercase_ ) for token in self.get_sentinel_tokens()]
| 56
| 1
|
'''simple docstring'''
from typing import Dict, Iterable, List, Optional, 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_DEFAULT_MEAN,
IMAGENET_DEFAULT_STD,
ChannelDimension,
ImageInput,
PILImageResampling,
make_list_of_images,
to_numpy_array,
valid_images,
)
from ...utils import TensorType, logging
a : Tuple = logging.get_logger(__name__)
class a ( _lowerCamelCase ):
snake_case_ = ["pixel_values"]
def __init__( self : List[str] , lowercase_ : bool = True , lowercase_ : Dict[str, int] = None , lowercase_ : PILImageResampling = PILImageResampling.BICUBIC , lowercase_ : bool = True , lowercase_ : Dict[str, int] = None , lowercase_ : bool = True , lowercase_ : Union[int, float] = 1 / 255 , lowercase_ : bool = True , lowercase_ : Optional[Union[float, Iterable[float]]] = IMAGENET_DEFAULT_MEAN , lowercase_ : Optional[Union[float, Iterable[float]]] = IMAGENET_DEFAULT_STD , **lowercase_ : Union[str, Any] , ):
super().__init__(**lowercase_ )
snake_case_ = size if size is not None else {'''shortest_edge''': 224}
snake_case_ = get_size_dict(lowercase_ , default_to_square=lowercase_ )
snake_case_ = crop_size if crop_size is not None else {'''height''': 224, '''width''': 224}
snake_case_ = get_size_dict(lowercase_ , param_name='''crop_size''' )
snake_case_ = do_resize
snake_case_ = size
snake_case_ = resample
snake_case_ = do_center_crop
snake_case_ = crop_size
snake_case_ = do_rescale
snake_case_ = rescale_factor
snake_case_ = do_normalize
snake_case_ = image_mean if image_mean is not None else IMAGENET_DEFAULT_MEAN
snake_case_ = image_std if image_std is not None else IMAGENET_DEFAULT_STD
def A_ ( self : Optional[Any] , lowercase_ : np.ndarray , lowercase_ : Dict[str, int] , lowercase_ : PILImageResampling = PILImageResampling.BICUBIC , lowercase_ : Optional[Union[str, ChannelDimension]] = None , **lowercase_ : Any , ):
snake_case_ = get_size_dict(lowercase_ , default_to_square=lowercase_ )
# size_dict is a dict with either keys "height" and "width" or "shortest_edge"
if "shortest_edge" in size:
snake_case_ = int((256 / 224) * size['''shortest_edge'''] )
snake_case_ = get_resize_output_image_size(lowercase_ , size=lowercase_ , default_to_square=lowercase_ )
snake_case_ = {'''height''': output_size[0], '''width''': output_size[1]}
if "height" not in size_dict or "width" not in size_dict:
raise ValueError(
F"Size dict must have keys 'height' and 'width' or 'shortest_edge'. Got {size_dict.keys()}" )
return resize(
lowercase_ , size=(size_dict['''height'''], size_dict['''width''']) , resample=lowercase_ , data_format=lowercase_ , **lowercase_ )
def A_ ( self : Tuple , lowercase_ : np.ndarray , lowercase_ : Dict[str, int] , lowercase_ : Optional[Union[str, ChannelDimension]] = None , **lowercase_ : Dict , ):
snake_case_ = get_size_dict(lowercase_ )
if "height" not in size or "width" not in size:
raise ValueError(F"Size dict must have keys 'height' and 'width'. Got {size.keys()}" )
return center_crop(lowercase_ , size=(size['''height'''], size['''width''']) , data_format=lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : np.ndarray , lowercase_ : Union[int, float] , lowercase_ : Optional[Union[str, ChannelDimension]] = None , **lowercase_ : Optional[int] , ):
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 , ):
return normalize(lowercase_ , mean=lowercase_ , std=lowercase_ , data_format=lowercase_ , **lowercase_ )
def A_ ( self : Optional[int] , lowercase_ : ImageInput , lowercase_ : Optional[bool] = None , lowercase_ : Optional[Dict[str, int]] = None , lowercase_ : PILImageResampling = None , lowercase_ : Optional[bool] = None , lowercase_ : Optional[Dict[str, int]] = None , lowercase_ : Optional[bool] = None , lowercase_ : Optional[float] = None , lowercase_ : Optional[bool] = None , lowercase_ : Optional[Union[float, Iterable[float]]] = None , lowercase_ : Optional[Union[float, Iterable[float]]] = None , lowercase_ : Optional[TensorType] = None , lowercase_ : ChannelDimension = ChannelDimension.FIRST , **lowercase_ : int , ):
snake_case_ = do_resize if do_resize is not None else self.do_resize
snake_case_ = resample if resample is not None else self.resample
snake_case_ = do_center_crop if do_center_crop is not None else self.do_center_crop
snake_case_ = do_rescale if do_rescale is not None else self.do_rescale
snake_case_ = rescale_factor if rescale_factor is not None else self.rescale_factor
snake_case_ = do_normalize if do_normalize is not None else self.do_normalize
snake_case_ = image_mean if image_mean is not None else self.image_mean
snake_case_ = image_std if image_std is not None else self.image_std
snake_case_ = size if size is not None else self.size
snake_case_ = get_size_dict(lowercase_ , default_to_square=lowercase_ )
snake_case_ = crop_size if crop_size is not None else self.crop_size
snake_case_ = get_size_dict(lowercase_ , param_name='''crop_size''' )
snake_case_ = 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.
snake_case_ = [to_numpy_array(lowercase_ ) for image in images]
if do_resize:
snake_case_ = [self.resize(lowercase_ , lowercase_ , lowercase_ ) for image in images]
if do_center_crop:
snake_case_ = [self.center_crop(lowercase_ , lowercase_ ) for image in images]
if do_rescale:
snake_case_ = [self.rescale(lowercase_ , lowercase_ ) for image in images]
if do_normalize:
snake_case_ = [self.normalize(lowercase_ , lowercase_ , lowercase_ ) for image in images]
snake_case_ = [to_channel_dimension_format(lowercase_ , lowercase_ ) for image in images]
snake_case_ = {'''pixel_values''': images}
return BatchFeature(data=lowercase_ , tensor_type=lowercase_ )
| 56
|
'''simple docstring'''
from __future__ import annotations
import math
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
if depth < 0:
raise ValueError('''Depth cannot be less than 0''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Scores cannot be empty''' )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
return min(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ = [90, 23, 6, 33, 21, 65, 123, 3_4423]
snake_case_ = math.log(len(__UpperCAmelCase ), 2 )
print('''Optimal value : ''', end='''''' )
print(minimax(0, 0, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
return int((input_a, input_a).count(0 ) != 0 )
def __magic_name__ ( ) -> None:
'''simple docstring'''
assert nand_gate(0, 0 ) == 1
assert nand_gate(0, 1 ) == 1
assert nand_gate(1, 0 ) == 1
assert nand_gate(1, 1 ) == 0
if __name__ == "__main__":
print(nand_gate(0, 0))
print(nand_gate(0, 1))
print(nand_gate(1, 0))
print(nand_gate(1, 1))
| 56
|
'''simple docstring'''
import argparse
import torch
from torch import nn
from transformers import SpeechaTextConfig, SpeechaTextForConditionalGeneration
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = [
'''encoder.version''',
'''decoder.version''',
'''model.encoder.version''',
'''model.decoder.version''',
'''decoder.output_projection.weight''',
'''_float_tensor''',
'''encoder.embed_positions._float_tensor''',
'''decoder.embed_positions._float_tensor''',
]
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = list(s_dict.keys() )
for key in keys:
if "transformer_layers" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
elif "subsample" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ ,snake_case_ = emb.weight.shape
snake_case_ = nn.Linear(__UpperCAmelCase, __UpperCAmelCase, bias=__UpperCAmelCase )
snake_case_ = emb.weight.data
return lin_layer
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )
snake_case_ = mam_aaa['''args''']
snake_case_ = mam_aaa['''model''']
snake_case_ = state_dict['''decoder.output_projection.weight''']
remove_ignore_keys_(__UpperCAmelCase )
rename_keys(__UpperCAmelCase )
snake_case_ = state_dict['''decoder.embed_tokens.weight'''].shape[0]
snake_case_ = args.share_decoder_input_output_embed
snake_case_ = [int(__UpperCAmelCase ) for i in args.conv_kernel_sizes.split(''',''' )]
snake_case_ = SpeechaTextConfig(
vocab_size=__UpperCAmelCase, max_source_positions=args.max_source_positions, max_target_positions=args.max_target_positions, encoder_layers=args.encoder_layers, decoder_layers=args.decoder_layers, encoder_attention_heads=args.encoder_attention_heads, decoder_attention_heads=args.decoder_attention_heads, encoder_ffn_dim=args.encoder_ffn_embed_dim, decoder_ffn_dim=args.decoder_ffn_embed_dim, d_model=args.encoder_embed_dim, dropout=args.dropout, attention_dropout=args.attention_dropout, activation_dropout=args.activation_dropout, activation_function='''relu''', num_conv_layers=len(__UpperCAmelCase ), conv_channels=args.conv_channels, conv_kernel_sizes=__UpperCAmelCase, input_feat_per_channel=args.input_feat_per_channel, input_channels=args.input_channels, tie_word_embeddings=__UpperCAmelCase, num_beams=5, max_length=200, use_cache=__UpperCAmelCase, decoder_start_token_id=2, early_stopping=__UpperCAmelCase, )
snake_case_ = SpeechaTextForConditionalGeneration(__UpperCAmelCase )
snake_case_ ,snake_case_ = model.model.load_state_dict(__UpperCAmelCase, strict=__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0 and not set(__UpperCAmelCase ) <= {
"encoder.embed_positions.weights",
"decoder.embed_positions.weights",
}:
raise ValueError(
'''Only `encoder.embed_positions.weights` and `decoder.embed_positions.weights` are allowed to be missing,'''
F" but all the following weights are missing {missing}" )
if tie_embeds:
snake_case_ = make_linear_from_emb(model.model.decoder.embed_tokens )
else:
snake_case_ = lm_head_weights
model.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
a : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument('--fairseq_path', type=str, help='Path to the fairseq model (.pt) file.')
parser.add_argument('--pytorch_dump_folder_path', default=None, type=str, help='Path to the output PyTorch model.')
a : List[Any] = parser.parse_args()
convert_fairseq_sat_checkpoint_to_tfms(args.fairseq_path, args.pytorch_dump_folder_path)
| 56
| 1
|
'''simple docstring'''
from collections import OrderedDict
from typing import Any, Mapping, Optional, Union
from ...configuration_utils import PretrainedConfig
from ...feature_extraction_utils import FeatureExtractionMixin
from ...onnx import OnnxConfig
from ...onnx.utils import compute_effective_axis_dimension
from ...tokenization_utils_base import PreTrainedTokenizerBase
from ...utils import TensorType, logging
a : Tuple = logging.get_logger(__name__)
a : List[str] = {
'deepmind/language-perceiver': 'https://huggingface.co/deepmind/language-perceiver/resolve/main/config.json',
# See all Perceiver models at https://huggingface.co/models?filter=perceiver
}
class a ( _lowerCamelCase ):
snake_case_ = "perceiver"
def __init__( self : Dict , lowercase_ : str=256 , lowercase_ : List[Any]=1280 , lowercase_ : Dict=768 , lowercase_ : str=1 , lowercase_ : Optional[int]=26 , lowercase_ : Any=8 , lowercase_ : Tuple=8 , lowercase_ : int=None , lowercase_ : List[str]=None , lowercase_ : Any="kv" , lowercase_ : str=1 , lowercase_ : int=1 , lowercase_ : List[str]="gelu" , lowercase_ : List[str]=0.1 , lowercase_ : Union[str, Any]=0.02 , lowercase_ : Optional[Any]=1e-12 , lowercase_ : Any=True , lowercase_ : List[Any]=262 , lowercase_ : List[str]=2048 , lowercase_ : str=56 , lowercase_ : int=[368, 496] , lowercase_ : Any=16 , lowercase_ : Optional[int]=1920 , lowercase_ : Optional[int]=16 , lowercase_ : Union[str, Any]=[1, 16, 224, 224] , **lowercase_ : List[str] , ):
super().__init__(**lowercase_ )
snake_case_ = num_latents
snake_case_ = d_latents
snake_case_ = d_model
snake_case_ = num_blocks
snake_case_ = num_self_attends_per_block
snake_case_ = num_self_attention_heads
snake_case_ = num_cross_attention_heads
snake_case_ = qk_channels
snake_case_ = v_channels
snake_case_ = cross_attention_shape_for_attention
snake_case_ = self_attention_widening_factor
snake_case_ = cross_attention_widening_factor
snake_case_ = hidden_act
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = use_query_residual
# masked language modeling attributes
snake_case_ = vocab_size
snake_case_ = max_position_embeddings
# image classification attributes
snake_case_ = image_size
# flow attributes
snake_case_ = train_size
# multimodal autoencoding attributes
snake_case_ = num_frames
snake_case_ = audio_samples_per_frame
snake_case_ = samples_per_patch
snake_case_ = output_shape
class a ( _lowerCamelCase ):
@property
def A_ ( self : Any ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''inputs''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
@property
def A_ ( self : Dict ):
return 1e-4
def A_ ( self : Optional[int] , lowercase_ : Union["PreTrainedTokenizerBase", "FeatureExtractionMixin"] , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , lowercase_ : int = 3 , lowercase_ : int = 40 , lowercase_ : int = 40 , ):
# copied from `transformers.onnx.config.OnnxConfig` and slightly altered/simplified
if isinstance(lowercase_ , lowercase_ ):
# If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX
snake_case_ = 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
snake_case_ = preprocessor.num_special_tokens_to_add(lowercase_ )
snake_case_ = 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
snake_case_ = [''' '''.join(['''a'''] ) * seq_length] * batch_size
snake_case_ = dict(preprocessor(lowercase_ , return_tensors=lowercase_ ) )
snake_case_ = inputs.pop('''input_ids''' )
return inputs
elif isinstance(lowercase_ , lowercase_ ) and preprocessor.model_input_names[0] == "pixel_values":
# If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX
snake_case_ = compute_effective_axis_dimension(lowercase_ , fixed_dimension=OnnxConfig.default_fixed_batch )
snake_case_ = self._generate_dummy_images(lowercase_ , lowercase_ , lowercase_ , lowercase_ )
snake_case_ = dict(preprocessor(images=lowercase_ , return_tensors=lowercase_ ) )
snake_case_ = inputs.pop('''pixel_values''' )
return inputs
else:
raise ValueError(
'''Unable to generate dummy inputs for the model. Please provide a tokenizer or a preprocessor.''' )
| 56
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class a ( metaclass=_lowerCamelCase ):
snake_case_ = ["transformers", "torch", "note_seq"]
def __init__( self : Union[str, Any] , *lowercase_ : Optional[int] , **lowercase_ : int ):
requires_backends(self , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : List[Any] , *lowercase_ : Optional[Any] , **lowercase_ : str ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : Tuple , *lowercase_ : Union[str, Any] , **lowercase_ : List[Any] ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
| 56
| 1
|
'''simple docstring'''
import argparse
import logging
from collections import namedtuple
import torch
from model_bertabs import BertAbsSummarizer
from models.model_builder import AbsSummarizer # The authors' implementation
from transformers import BertTokenizer
logging.basicConfig(level=logging.INFO)
a : Optional[int] = logging.getLogger(__name__)
a : Optional[Any] = 'Hello world! cécé herlolip'
a : List[str] = namedtuple(
'BertAbsConfig',
[
'temp_dir',
'large',
'use_bert_emb',
'finetune_bert',
'encoder',
'share_emb',
'max_pos',
'enc_layers',
'enc_hidden_size',
'enc_heads',
'enc_ff_size',
'enc_dropout',
'dec_layers',
'dec_hidden_size',
'dec_heads',
'dec_ff_size',
'dec_dropout',
],
)
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = BertAbsConfig(
temp_dir='''.''', finetune_bert=__UpperCAmelCase, large=__UpperCAmelCase, share_emb=__UpperCAmelCase, use_bert_emb=__UpperCAmelCase, encoder='''bert''', max_pos=512, enc_layers=6, enc_hidden_size=512, enc_heads=8, enc_ff_size=512, enc_dropout=0.2, dec_layers=6, dec_hidden_size=768, dec_heads=8, dec_ff_size=2048, dec_dropout=0.2, )
snake_case_ = torch.load(__UpperCAmelCase, lambda __UpperCAmelCase, __UpperCAmelCase : storage )
snake_case_ = AbsSummarizer(__UpperCAmelCase, torch.device('''cpu''' ), __UpperCAmelCase )
original.eval()
snake_case_ = BertAbsSummarizer(__UpperCAmelCase, torch.device('''cpu''' ) )
new_model.eval()
# -------------------
# Convert the weights
# -------------------
logging.info('''convert the model''' )
new_model.bert.load_state_dict(original.bert.state_dict() )
new_model.decoder.load_state_dict(original.decoder.state_dict() )
new_model.generator.load_state_dict(original.generator.state_dict() )
# ----------------------------------
# Make sure the outpus are identical
# ----------------------------------
logging.info('''Make sure that the models\' outputs are identical''' )
snake_case_ = BertTokenizer.from_pretrained('''bert-base-uncased''' )
# prepare the model inputs
snake_case_ = tokenizer.encode('''This is sample éàalj\'-.''' )
encoder_input_ids.extend([tokenizer.pad_token_id] * (512 - len(__UpperCAmelCase )) )
snake_case_ = torch.tensor(__UpperCAmelCase ).unsqueeze(0 )
snake_case_ = tokenizer.encode('''This is sample 3 éàalj\'-.''' )
decoder_input_ids.extend([tokenizer.pad_token_id] * (512 - len(__UpperCAmelCase )) )
snake_case_ = torch.tensor(__UpperCAmelCase ).unsqueeze(0 )
# failsafe to make sure the weights reset does not affect the
# loaded weights.
assert torch.max(torch.abs(original.generator[0].weight - new_model.generator[0].weight ) ) == 0
# forward pass
snake_case_ = encoder_input_ids
snake_case_ = decoder_input_ids
snake_case_ = snake_case_ = None
snake_case_ = None
snake_case_ = snake_case_ = None
snake_case_ = snake_case_ = None
snake_case_ = None
# The original model does not apply the geneator layer immediatly but rather in
# the beam search (where it combines softmax + linear layer). Since we already
# apply the softmax in our generation process we only apply the linear layer here.
# We make sure that the outputs of the full stack are identical
snake_case_ = original(__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )[0]
snake_case_ = original.generator(__UpperCAmelCase )
snake_case_ = new_model(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )[0]
snake_case_ = new_model.generator(__UpperCAmelCase )
snake_case_ = torch.max(torch.abs(output_converted_model - output_original_model ) ).item()
print('''Maximum absolute difference beween weights: {:.2f}'''.format(__UpperCAmelCase ) )
snake_case_ = torch.max(torch.abs(output_converted_generator - output_original_generator ) ).item()
print('''Maximum absolute difference beween weights: {:.2f}'''.format(__UpperCAmelCase ) )
snake_case_ = torch.allclose(__UpperCAmelCase, __UpperCAmelCase, atol=1e-3 )
if are_identical:
logging.info('''all weights are equal up to 1e-3''' )
else:
raise ValueError('''the weights are different. The new model is likely different from the original one.''' )
# The model has been saved with torch.save(model) and this is bound to the exact
# directory structure. We save the state_dict instead.
logging.info('''saving the model\'s state dictionary''' )
torch.save(
new_model.state_dict(), '''./bertabs-finetuned-cnndm-extractive-abstractive-summarization/pytorch_model.bin''' )
if __name__ == "__main__":
a : Union[str, Any] = argparse.ArgumentParser()
parser.add_argument(
'--bertabs_checkpoint_path',
default=None,
type=str,
required=True,
help='Path the official PyTorch dump.',
)
parser.add_argument(
'--pytorch_dump_folder_path',
default=None,
type=str,
required=True,
help='Path to the output PyTorch model.',
)
a : Dict = parser.parse_args()
convert_bertabs_checkpoints(
args.bertabs_checkpoint_path,
args.pytorch_dump_folder_path,
)
| 56
|
'''simple docstring'''
# tests directory-specific settings - this file is run automatically
# by pytest before any tests are run
import doctest
import sys
import warnings
from os.path import abspath, dirname, join
import _pytest
from transformers.testing_utils import HfDoctestModule, HfDocTestParser
# allow having multiple repository checkouts and not needing to remember to rerun
# 'pip install -e .[dev]' when switching between checkouts and running tests.
a : int = abspath(join(dirname(__file__), 'src'))
sys.path.insert(1, git_repo_path)
# silence FutureWarning warnings in tests since often we can't act on them until
# they become normal warnings - i.e. the tests still need to test the current functionality
warnings.simplefilter(action='ignore', category=FutureWarning)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
config.addinivalue_line(
'''markers''', '''is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested''' )
config.addinivalue_line(
'''markers''', '''is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested''' )
config.addinivalue_line('''markers''', '''is_pipeline_test: mark test to run only when pipelines are tested''' )
config.addinivalue_line('''markers''', '''is_staging_test: mark test to run only in the staging environment''' )
config.addinivalue_line('''markers''', '''accelerate_tests: mark test that require accelerate''' )
config.addinivalue_line('''markers''', '''tool_tests: mark the tool tests that are run on their specific schedule''' )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from transformers.testing_utils import pytest_addoption_shared
pytest_addoption_shared(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
from transformers.testing_utils import pytest_terminal_summary_main
snake_case_ = terminalreporter.config.getoption('''--make-reports''' )
if make_reports:
pytest_terminal_summary_main(__UpperCAmelCase, id=__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if exitstatus == 5:
snake_case_ = 0
# Doctest custom flag to ignore output.
a : Union[str, Any] = doctest.register_optionflag('IGNORE_RESULT')
a : Optional[int] = doctest.OutputChecker
class a ( _lowerCamelCase ):
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int] ):
if IGNORE_RESULT & optionflags:
return True
return OutputChecker.check_output(self , lowercase_ , lowercase_ , lowercase_ )
a : List[Any] = CustomOutputChecker
a : Optional[int] = HfDoctestModule
a : Tuple = HfDocTestParser
| 56
| 1
|
'''simple docstring'''
import os
import sys
a : str = os.path.join(os.path.dirname(__file__), 'src')
sys.path.append(SRC_DIR)
from transformers import (
AutoConfig,
AutoModel,
AutoModelForCausalLM,
AutoModelForMaskedLM,
AutoModelForQuestionAnswering,
AutoModelForSequenceClassification,
AutoTokenizer,
add_start_docstrings,
)
a : Optional[Any] = [
'torch',
'numpy',
'tokenizers',
'filelock',
'requests',
'tqdm',
'regex',
'sentencepiece',
'sacremoses',
'importlib_metadata',
'huggingface_hub',
]
@add_start_docstrings(AutoConfig.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
return AutoConfig.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
@add_start_docstrings(AutoTokenizer.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
return AutoTokenizer.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
@add_start_docstrings(AutoModel.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Any:
'''simple docstring'''
return AutoModel.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
@add_start_docstrings(AutoModelForCausalLM.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> List[str]:
'''simple docstring'''
return AutoModelForCausalLM.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
@add_start_docstrings(AutoModelForMaskedLM.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Tuple:
'''simple docstring'''
return AutoModelForMaskedLM.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
@add_start_docstrings(AutoModelForSequenceClassification.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
return AutoModelForSequenceClassification.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
@add_start_docstrings(AutoModelForQuestionAnswering.__doc__ )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Tuple:
'''simple docstring'''
return AutoModelForQuestionAnswering.from_pretrained(*__UpperCAmelCase, **__UpperCAmelCase )
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import TensorType, is_torch_available, logging
a : Dict = logging.get_logger(__name__)
a : List[str] = {
'Helsinki-NLP/opus-mt-en-de': 'https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/config.json',
# See all Marian models at https://huggingface.co/models?filter=marian
}
class a ( _lowerCamelCase ):
snake_case_ = "marian"
snake_case_ = ["past_key_values"]
snake_case_ = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self : List[Any] , lowercase_ : Optional[Any]=5_8101 , lowercase_ : Dict=None , lowercase_ : List[str]=1024 , lowercase_ : Optional[Any]=12 , lowercase_ : int=4096 , lowercase_ : Any=16 , lowercase_ : Optional[int]=12 , lowercase_ : str=4096 , lowercase_ : Union[str, Any]=16 , lowercase_ : Dict=0.0 , lowercase_ : Union[str, Any]=0.0 , lowercase_ : Optional[Any]=True , lowercase_ : Union[str, Any]=True , lowercase_ : int="gelu" , lowercase_ : Dict=1024 , lowercase_ : int=0.1 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.02 , lowercase_ : int=5_8100 , lowercase_ : Optional[Any]=False , lowercase_ : Any=5_8100 , lowercase_ : Optional[int]=0 , lowercase_ : Dict=0 , lowercase_ : List[str]=True , **lowercase_ : Any , ):
snake_case_ = vocab_size
snake_case_ = decoder_vocab_size or vocab_size
snake_case_ = max_position_embeddings
snake_case_ = d_model
snake_case_ = encoder_ffn_dim
snake_case_ = encoder_layers
snake_case_ = encoder_attention_heads
snake_case_ = decoder_ffn_dim
snake_case_ = decoder_layers
snake_case_ = decoder_attention_heads
snake_case_ = dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = activation_function
snake_case_ = init_std
snake_case_ = encoder_layerdrop
snake_case_ = decoder_layerdrop
snake_case_ = use_cache
snake_case_ = encoder_layers
snake_case_ = scale_embedding # scale factor will be sqrt(d_model) if True
snake_case_ = share_encoder_decoder_embeddings
super().__init__(
pad_token_id=lowercase_ , eos_token_id=lowercase_ , is_encoder_decoder=lowercase_ , decoder_start_token_id=lowercase_ , forced_eos_token_id=lowercase_ , **lowercase_ , )
class a ( _lowerCamelCase ):
@property
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.inputs
def A_ ( self : Union[str, Any] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ = {0: '''batch'''}
snake_case_ = {0: '''batch''', 1: '''past_decoder_sequence + sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''decoder_sequence'''}
snake_case_ = {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.
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
else:
snake_case_ = 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
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.outputs
def A_ ( self : Dict ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super().outputs
else:
snake_case_ = super(lowercase_ , self ).outputs
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
return common_outputs
def A_ ( self : Dict , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
# Generate decoder inputs
snake_case_ = seq_length if not self.use_past else 1
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
snake_case_ = {F"decoder_{name}": tensor for name, tensor in decoder_inputs.items()}
snake_case_ = 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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
snake_case_ = common_inputs['''decoder_input_ids'''].shape[1]
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = decoder_seq_length + 3
snake_case_ = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
snake_case_ = torch.cat(
[common_inputs['''decoder_attention_mask'''], torch.ones(lowercase_ , lowercase_ )] , dim=1 )
snake_case_ = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
snake_case_ ,snake_case_ = self.num_layers
snake_case_ = min(lowercase_ , lowercase_ )
snake_case_ = max(lowercase_ , lowercase_ ) - min_num_layers
snake_case_ = '''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.
snake_case_ = 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 : Union[str, Any] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
# Not using the same length for past_key_values
snake_case_ = seqlen + 2
snake_case_ ,snake_case_ = self.num_layers
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = common_inputs['''attention_mask'''].dtype
snake_case_ = torch.cat(
[common_inputs['''attention_mask'''], torch.ones(lowercase_ , lowercase_ , dtype=lowercase_ )] , dim=1 )
snake_case_ = [
(torch.zeros(lowercase_ ), torch.zeros(lowercase_ )) for _ in range(lowercase_ )
]
return common_inputs
def A_ ( self : List[str] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
# 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
snake_case_ = 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
snake_case_ = tokenizer.num_special_tokens_to_add(lowercase_ )
snake_case_ = 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
snake_case_ = [''' '''.join([tokenizer.unk_token] ) * seq_length] * batch_size
snake_case_ = dict(tokenizer(lowercase_ , return_tensors=lowercase_ ) )
return common_inputs
def A_ ( self : Any , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
else:
snake_case_ = self._generate_dummy_inputs_for_causal_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
return common_inputs
def A_ ( self : Dict , lowercase_ : List[str] , lowercase_ : List[str] , lowercase_ : int , lowercase_ : List[str] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super()._flatten_past_key_values_(lowercase_ , lowercase_ , lowercase_ , lowercase_ )
else:
snake_case_ = super(lowercase_ , self )._flatten_past_key_values_(
lowercase_ , lowercase_ , lowercase_ , lowercase_ )
@property
def A_ ( self : List[str] ):
return 1e-4
| 56
| 1
|
'''simple docstring'''
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 ( _lowerCamelCase ):
snake_case_ = 42
@flax_register_to_config
class a ( nn.Module , _lowerCamelCase , _lowerCamelCase ):
snake_case_ = 32
snake_case_ = 4
snake_case_ = 4
snake_case_ = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
snake_case_ = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
snake_case_ = False
snake_case_ = (320, 640, 1_280, 1_280)
snake_case_ = 2
snake_case_ = 8
snake_case_ = None
snake_case_ = 1_280
snake_case_ = 0.0
snake_case_ = False
snake_case_ = jnp.floataa
snake_case_ = True
snake_case_ = 0
snake_case_ = False
def A_ ( self : Optional[int] , lowercase_ : jax.random.KeyArray ):
# init input tensors
snake_case_ = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case_ = jnp.zeros(lowercase_ , dtype=jnp.floataa )
snake_case_ = jnp.ones((1,) , dtype=jnp.intaa )
snake_case_ = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case_ ,snake_case_ = jax.random.split(lowercase_ )
snake_case_ = {'''params''': params_rng, '''dropout''': dropout_rng}
return self.init(lowercase_ , lowercase_ , lowercase_ , lowercase_ )["params"]
def A_ ( self : List[str] ):
snake_case_ = self.block_out_channels
snake_case_ = 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.
snake_case_ = self.num_attention_heads or self.attention_head_dim
# input
snake_case_ = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case_ = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case_ = FlaxTimestepEmbedding(lowercase_ , dtype=self.dtype )
snake_case_ = self.only_cross_attention
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (only_cross_attention,) * len(self.down_block_types )
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case_ = []
snake_case_ = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case_ = output_channel
snake_case_ = block_out_channels[i]
snake_case_ = i == len(lowercase_ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = down_blocks
# mid
snake_case_ = 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
snake_case_ = []
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case_ = output_channel
snake_case_ = reversed_block_out_channels[i]
snake_case_ = reversed_block_out_channels[min(i + 1 , len(lowercase_ ) - 1 )]
snake_case_ = i == len(lowercase_ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = output_channel
snake_case_ = up_blocks
# out
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : int , lowercase_ : Any , lowercase_ : int=None , lowercase_ : Any=None , lowercase_ : bool = True , lowercase_ : bool = False , ):
# 1. time
if not isinstance(lowercase_ , jnp.ndarray ):
snake_case_ = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(lowercase_ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case_ = timesteps.astype(dtype=jnp.floataa )
snake_case_ = jnp.expand_dims(lowercase_ , 0 )
snake_case_ = self.time_proj(lowercase_ )
snake_case_ = self.time_embedding(lowercase_ )
# 2. pre-process
snake_case_ = jnp.transpose(lowercase_ , (0, 2, 3, 1) )
snake_case_ = self.conv_in(lowercase_ )
# 3. down
snake_case_ = (sample,)
for down_block in self.down_blocks:
if isinstance(lowercase_ , lowercase_ ):
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , lowercase_ , deterministic=not train )
else:
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case_ = ()
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,)
snake_case_ = new_down_block_res_samples
# 4. mid
snake_case_ = 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:
snake_case_ = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case_ = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = up_block(
lowercase_ , temb=lowercase_ , encoder_hidden_states=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train , )
else:
snake_case_ = up_block(lowercase_ , temb=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train )
# 6. post-process
snake_case_ = self.conv_norm_out(lowercase_ )
snake_case_ = nn.silu(lowercase_ )
snake_case_ = self.conv_out(lowercase_ )
snake_case_ = jnp.transpose(lowercase_ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=lowercase_ )
| 56
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import AutoencoderKL, CycleDiffusionPipeline, DDIMScheduler, UNetaDConditionModel
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, skip_mps
from ..pipeline_params import (
IMAGE_TO_IMAGE_IMAGE_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = CycleDiffusionPipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {
"negative_prompt",
"height",
"width",
"negative_prompt_embeds",
}
snake_case_ = PipelineTesterMixin.required_optional_params - {"latents"}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"source_prompt"} )
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
def A_ ( self : Tuple ):
torch.manual_seed(0 )
snake_case_ = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=32 , )
snake_case_ = DDIMScheduler(
beta_start=0.0_0085 , beta_end=0.012 , beta_schedule='''scaled_linear''' , num_train_timesteps=1000 , clip_sample=lowercase_ , set_alpha_to_one=lowercase_ , )
torch.manual_seed(0 )
snake_case_ = 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 )
snake_case_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , )
snake_case_ = CLIPTextModel(lowercase_ )
snake_case_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
snake_case_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
'''safety_checker''': None,
'''feature_extractor''': None,
}
return components
def A_ ( self : Any , lowercase_ : int , lowercase_ : Optional[Any]=0 ):
snake_case_ = floats_tensor((1, 3, 32, 32) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
snake_case_ = image / 2 + 0.5
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''prompt''': '''An astronaut riding an elephant''',
'''source_prompt''': '''An astronaut riding a horse''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''eta''': 0.1,
'''strength''': 0.8,
'''guidance_scale''': 3,
'''source_guidance_scale''': 1,
'''output_type''': '''numpy''',
}
return inputs
def A_ ( self : Union[str, Any] ):
snake_case_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
snake_case_ = self.get_dummy_components()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.4459, 0.4943, 0.4544, 0.6643, 0.5474, 0.4327, 0.5701, 0.5959, 0.5179] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.get_dummy_components()
for name, module in components.items():
if hasattr(lowercase_ , '''half''' ):
snake_case_ = module.half()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.3506, 0.4543, 0.446, 0.4575, 0.5195, 0.4155, 0.5273, 0.518, 0.4116] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@skip_mps
def A_ ( self : Optional[int] ):
return super().test_save_load_local()
@unittest.skip('''non-deterministic pipeline''' )
def A_ ( self : List[Any] ):
return super().test_inference_batch_single_identical()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_dict_tuple_outputs_equivalent()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_save_load_optional_components()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_attention_slicing_forward_pass()
@slow
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : List[Any] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Union[str, Any] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car_fp16.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(
lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , torch_dtype=torch.floataa , revision='''fp16''' )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
# the values aren't exactly equal, but the images look the same visually
assert np.abs(image - expected_image ).max() < 5e-1
def A_ ( self : List[str] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
assert np.abs(image - expected_image ).max() < 2e-2
| 56
| 1
|
'''simple docstring'''
import logging
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
import datasets
import numpy as np
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PreTrainedTokenizer,
TFAutoModelForSequenceClassification,
TFTrainer,
TFTrainingArguments,
)
from transformers.utils import logging as hf_logging
hf_logging.set_verbosity_info()
hf_logging.enable_default_handler()
hf_logging.enable_explicit_format()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = None, ) -> str:
'''simple docstring'''
snake_case_ = {}
if train_file is not None:
snake_case_ = [train_file]
if eval_file is not None:
snake_case_ = [eval_file]
if test_file is not None:
snake_case_ = [test_file]
snake_case_ = datasets.load_dataset('''csv''', data_files=__UpperCAmelCase )
snake_case_ = list(ds[list(files.keys() )[0]].features.keys() )
snake_case_ = features_name.pop(__UpperCAmelCase )
snake_case_ = list(set(ds[list(files.keys() )[0]][label_name] ) )
snake_case_ = {label: i for i, label in enumerate(__UpperCAmelCase )}
snake_case_ = tokenizer.model_input_names
snake_case_ = {}
if len(__UpperCAmelCase ) == 1:
for k in files.keys():
snake_case_ = ds[k].map(
lambda __UpperCAmelCase : tokenizer.batch_encode_plus(
example[features_name[0]], truncation=__UpperCAmelCase, max_length=__UpperCAmelCase, padding='''max_length''' ), batched=__UpperCAmelCase, )
elif len(__UpperCAmelCase ) == 2:
for k in files.keys():
snake_case_ = ds[k].map(
lambda __UpperCAmelCase : tokenizer.batch_encode_plus(
(example[features_name[0]], example[features_name[1]]), truncation=__UpperCAmelCase, max_length=__UpperCAmelCase, padding='''max_length''', ), batched=__UpperCAmelCase, )
def gen_train():
for ex in transformed_ds[datasets.Split.TRAIN]:
snake_case_ = {k: v for k, v in ex.items() if k in input_names}
snake_case_ = labelaid[ex[label_name]]
yield (d, label)
def gen_val():
for ex in transformed_ds[datasets.Split.VALIDATION]:
snake_case_ = {k: v for k, v in ex.items() if k in input_names}
snake_case_ = labelaid[ex[label_name]]
yield (d, label)
def gen_test():
for ex in transformed_ds[datasets.Split.TEST]:
snake_case_ = {k: v for k, v in ex.items() if k in input_names}
snake_case_ = labelaid[ex[label_name]]
yield (d, label)
snake_case_ = (
tf.data.Dataset.from_generator(
__UpperCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), )
if datasets.Split.TRAIN in transformed_ds
else None
)
if train_ds is not None:
snake_case_ = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) )
snake_case_ = (
tf.data.Dataset.from_generator(
__UpperCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), )
if datasets.Split.VALIDATION in transformed_ds
else None
)
if val_ds is not None:
snake_case_ = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) )
snake_case_ = (
tf.data.Dataset.from_generator(
__UpperCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), )
if datasets.Split.TEST in transformed_ds
else None
)
if test_ds is not None:
snake_case_ = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) )
return train_ds, val_ds, test_ds, labelaid
a : Tuple = logging.getLogger(__name__)
@dataclass
class a :
snake_case_ = field(metadata={"help": "Which column contains the label"} )
snake_case_ = field(default=_lowerCamelCase , metadata={"help": "The path of the training file"} )
snake_case_ = field(default=_lowerCamelCase , metadata={"help": "The path of the development file"} )
snake_case_ = field(default=_lowerCamelCase , metadata={"help": "The path of the test file"} )
snake_case_ = field(
default=128 , metadata={
"help": (
"The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
} , )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Overwrite the cached training and evaluation sets"} )
@dataclass
class a :
snake_case_ = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Pretrained config name or path if not the same as model_name"} )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"} )
snake_case_ = field(default=_lowerCamelCase , metadata={"help": "Set this flag to use fast tokenization."} )
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
# or just modify its tokenizer_config.json.
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} , )
def __magic_name__ ( ) -> Tuple:
'''simple docstring'''
snake_case_ = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) )
snake_case_ ,snake_case_ ,snake_case_ = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir )
and os.listdir(training_args.output_dir )
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
F"Output directory ({training_args.output_dir}) already exists and is not empty. Use"
''' --overwrite_output_dir to overcome.''' )
# Setup logging
logging.basicConfig(
format='''%(asctime)s - %(levelname)s - %(name)s - %(message)s''', datefmt='''%m/%d/%Y %H:%M:%S''', level=logging.INFO, )
logger.info(
F"n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, "
F"16-bits training: {training_args.fpaa}" )
logger.info(F"Training/evaluation parameters {training_args}" )
# Load pretrained model and tokenizer
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
snake_case_ = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path, cache_dir=model_args.cache_dir, )
snake_case_ ,snake_case_ ,snake_case_ ,snake_case_ = get_tfds(
train_file=data_args.train_file, eval_file=data_args.dev_file, test_file=data_args.test_file, tokenizer=__UpperCAmelCase, label_column_id=data_args.label_column_id, max_seq_length=data_args.max_seq_length, )
snake_case_ = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path, num_labels=len(__UpperCAmelCase ), labelaid=__UpperCAmelCase, idalabel={id: label for label, id in labelaid.items()}, finetuning_task='''text-classification''', cache_dir=model_args.cache_dir, )
with training_args.strategy.scope():
snake_case_ = TFAutoModelForSequenceClassification.from_pretrained(
model_args.model_name_or_path, from_pt=bool('''.bin''' in model_args.model_name_or_path ), config=__UpperCAmelCase, cache_dir=model_args.cache_dir, )
def compute_metrics(__UpperCAmelCase ) -> Dict:
snake_case_ = np.argmax(p.predictions, axis=1 )
return {"acc": (preds == p.label_ids).mean()}
# Initialize our Trainer
snake_case_ = TFTrainer(
model=__UpperCAmelCase, args=__UpperCAmelCase, train_dataset=__UpperCAmelCase, eval_dataset=__UpperCAmelCase, compute_metrics=__UpperCAmelCase, )
# Training
if training_args.do_train:
trainer.train()
trainer.save_model()
tokenizer.save_pretrained(training_args.output_dir )
# Evaluation
snake_case_ = {}
if training_args.do_eval:
logger.info('''*** Evaluate ***''' )
snake_case_ = trainer.evaluate()
snake_case_ = os.path.join(training_args.output_dir, '''eval_results.txt''' )
with open(__UpperCAmelCase, '''w''' ) as writer:
logger.info('''***** Eval results *****''' )
for key, value in result.items():
logger.info(F" {key} = {value}" )
writer.write(F"{key} = {value}\n" )
results.update(__UpperCAmelCase )
return results
if __name__ == "__main__":
main()
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : str = logging.get_logger(__name__)
a : str = {
'google/bigbird-roberta-base': 'https://huggingface.co/google/bigbird-roberta-base/resolve/main/config.json',
'google/bigbird-roberta-large': 'https://huggingface.co/google/bigbird-roberta-large/resolve/main/config.json',
'google/bigbird-base-trivia-itc': 'https://huggingface.co/google/bigbird-base-trivia-itc/resolve/main/config.json',
# See all BigBird models at https://huggingface.co/models?filter=big_bird
}
class a ( _lowerCamelCase ):
snake_case_ = "big_bird"
def __init__( self : Union[str, Any] , lowercase_ : List[Any]=5_0358 , lowercase_ : Tuple=768 , lowercase_ : Dict=12 , lowercase_ : str=12 , lowercase_ : Tuple=3072 , lowercase_ : Any="gelu_new" , lowercase_ : Optional[Any]=0.1 , lowercase_ : List[Any]=0.1 , lowercase_ : List[Any]=4096 , lowercase_ : List[Any]=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[int]=1e-12 , lowercase_ : Tuple=True , lowercase_ : Tuple=0 , lowercase_ : str=1 , lowercase_ : Union[str, Any]=2 , lowercase_ : Optional[Any]=66 , lowercase_ : Optional[int]="block_sparse" , lowercase_ : Any=True , lowercase_ : List[str]=False , lowercase_ : Any=64 , lowercase_ : Tuple=3 , lowercase_ : Tuple=None , **lowercase_ : Tuple , ):
super().__init__(
pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , sep_token_id=lowercase_ , **lowercase_ , )
snake_case_ = vocab_size
snake_case_ = max_position_embeddings
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = type_vocab_size
snake_case_ = layer_norm_eps
snake_case_ = use_cache
snake_case_ = rescale_embeddings
snake_case_ = attention_type
snake_case_ = use_bias
snake_case_ = block_size
snake_case_ = num_random_blocks
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : str ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
| 1
|
'''simple docstring'''
import multiprocessing
from typing import TYPE_CHECKING, Optional, Union
from .. import Dataset, Features, config
from ..formatting import query_table
from ..packaged_modules.sql.sql import Sql
from ..utils import logging
from .abc import AbstractDatasetInputStream
if TYPE_CHECKING:
import sqlitea
import sqlalchemy
class a ( _lowerCamelCase ):
def __init__( self : str , lowercase_ : Union[str, "sqlalchemy.sql.Selectable"] , lowercase_ : Union[str, "sqlalchemy.engine.Connection", "sqlalchemy.engine.Engine", "sqlite3.Connection"] , lowercase_ : Optional[Features] = None , lowercase_ : str = None , lowercase_ : bool = False , **lowercase_ : Optional[int] , ):
super().__init__(features=lowercase_ , cache_dir=lowercase_ , keep_in_memory=lowercase_ , **lowercase_ )
snake_case_ = Sql(
cache_dir=lowercase_ , features=lowercase_ , sql=lowercase_ , con=lowercase_ , **lowercase_ , )
def A_ ( self : List[Any] ):
snake_case_ = None
snake_case_ = None
snake_case_ = None
snake_case_ = None
self.builder.download_and_prepare(
download_config=lowercase_ , download_mode=lowercase_ , verification_mode=lowercase_ , base_path=lowercase_ , )
# Build dataset for splits
snake_case_ = self.builder.as_dataset(
split='''train''' , verification_mode=lowercase_ , in_memory=self.keep_in_memory )
return dataset
class a :
def __init__( self : Tuple , lowercase_ : Dataset , lowercase_ : str , lowercase_ : Union[str, "sqlalchemy.engine.Connection", "sqlalchemy.engine.Engine", "sqlite3.Connection"] , lowercase_ : Optional[int] = None , lowercase_ : Optional[int] = None , **lowercase_ : Optional[Any] , ):
if num_proc is not None and num_proc <= 0:
raise ValueError(F"num_proc {num_proc} must be an integer > 0." )
snake_case_ = dataset
snake_case_ = name
snake_case_ = con
snake_case_ = batch_size if batch_size else config.DEFAULT_MAX_BATCH_SIZE
snake_case_ = num_proc
snake_case_ = to_sql_kwargs
def A_ ( self : str ):
snake_case_ = self.to_sql_kwargs.pop('''sql''' , lowercase_ )
snake_case_ = self.to_sql_kwargs.pop('''con''' , lowercase_ )
snake_case_ = self.to_sql_kwargs.pop('''index''' , lowercase_ )
snake_case_ = self._write(index=lowercase_ , **self.to_sql_kwargs )
return written
def A_ ( self : List[str] , lowercase_ : Any ):
snake_case_ ,snake_case_ ,snake_case_ = args
snake_case_ = {**to_sql_kwargs, '''if_exists''': '''append'''} if offset > 0 else to_sql_kwargs
snake_case_ = query_table(
table=self.dataset.data , key=slice(lowercase_ , offset + self.batch_size ) , indices=self.dataset._indices , )
snake_case_ = batch.to_pandas()
snake_case_ = df.to_sql(self.name , self.con , index=lowercase_ , **lowercase_ )
return num_rows or len(lowercase_ )
def A_ ( self : Optional[int] , lowercase_ : Optional[int] , **lowercase_ : Union[str, Any] ):
snake_case_ = 0
if self.num_proc is None or self.num_proc == 1:
for offset in logging.tqdm(
range(0 , len(self.dataset ) , self.batch_size ) , unit='''ba''' , disable=not logging.is_progress_bar_enabled() , desc='''Creating SQL from Arrow format''' , ):
written += self._batch_sql((offset, index, to_sql_kwargs) )
else:
snake_case_ ,snake_case_ = len(self.dataset ), self.batch_size
with multiprocessing.Pool(self.num_proc ) as pool:
for num_rows in logging.tqdm(
pool.imap(
self._batch_sql , [(offset, index, to_sql_kwargs) for offset in range(0 , lowercase_ , lowercase_ )] , ) , total=(num_rows // batch_size) + 1 if num_rows % batch_size else num_rows // batch_size , unit='''ba''' , disable=not logging.is_progress_bar_enabled() , desc='''Creating SQL from Arrow format''' , ):
written += num_rows
return written
| 56
|
'''simple docstring'''
import contextlib
import os
import sqlitea
import pytest
from datasets import Dataset, Features, Value
from datasets.io.sql import SqlDatasetReader, SqlDatasetWriter
from ..utils import assert_arrow_memory_doesnt_increase, assert_arrow_memory_increases, require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
assert isinstance(__UpperCAmelCase, __UpperCAmelCase )
assert dataset.num_rows == 4
assert dataset.num_columns == 3
assert dataset.column_names == ["col_1", "col_2", "col_3"]
for feature, expected_dtype in expected_features.items():
assert dataset.features[feature].dtype == expected_dtype
@require_sqlalchemy
@pytest.mark.parametrize('''keep_in_memory''', [False, True] )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase():
snake_case_ = SqlDatasetReader(
'''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase, keep_in_memory=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
@require_sqlalchemy
@pytest.mark.parametrize(
'''features''', [
None,
{'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''},
{'''col_1''': '''string''', '''col_2''': '''string''', '''col_3''': '''string'''},
{'''col_1''': '''int32''', '''col_2''': '''int32''', '''col_3''': '''int32'''},
{'''col_1''': '''float32''', '''col_2''': '''float32''', '''col_3''': '''float32'''},
], )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
snake_case_ = features.copy() if features else default_expected_features
snake_case_ = (
Features({feature: Value(__UpperCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None
)
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, features=__UpperCAmelCase, cache_dir=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
with contextlib.closing(sqlitea.connect(__UpperCAmelCase ) ) as con:
snake_case_ = con.cursor()
cur.execute('''SELECT * FROM dataset''' )
for row in cur:
yield row
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=1 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=2 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
with pytest.raises(__UpperCAmelCase ):
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=0 ).write()
| 56
| 1
|
'''simple docstring'''
import warnings
from ...utils import logging
from .image_processing_owlvit import OwlViTImageProcessor
a : Dict = logging.get_logger(__name__)
class a ( _lowerCamelCase ):
def __init__( self : List[str] , *lowercase_ : Union[str, Any] , **lowercase_ : Tuple ):
warnings.warn(
'''The class OwlViTFeatureExtractor is deprecated and will be removed in version 5 of Transformers. Please'''
''' use OwlViTImageProcessor instead.''' , lowercase_ , )
super().__init__(*lowercase_ , **lowercase_ )
| 56
|
'''simple docstring'''
from collections import defaultdict
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 1
snake_case_ = True
for v in tree[start]:
if v not in visited:
ret += dfs(__UpperCAmelCase )
if ret % 2 == 0:
cuts.append(__UpperCAmelCase )
return ret
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
dfs(1 )
if __name__ == "__main__":
a ,a : Dict = 10, 9
a : Dict = defaultdict(list)
a : dict[int, bool] = {}
a : list[int] = []
a : Tuple = 0
a : str = [(2, 1), (3, 1), (4, 3), (5, 2), (6, 1), (7, 2), (8, 6), (9, 8), (10, 8)]
for u, v in edges:
tree[u].append(v)
tree[v].append(u)
even_tree()
print(len(cuts) - 1)
| 56
| 1
|
'''simple docstring'''
import math
import time
from transformers import Trainer, is_torch_tpu_available
from transformers.trainer_utils import PredictionOutput, speed_metrics
if is_torch_tpu_available(check_device=False):
import torch_xla.core.xla_model as xm
import torch_xla.debug.metrics as met
class a ( _lowerCamelCase ):
def __init__( self : List[str] , *lowercase_ : List[str] , lowercase_ : Optional[int]=None , lowercase_ : Optional[int]=None , **lowercase_ : Optional[int] ):
super().__init__(*lowercase_ , **lowercase_ )
snake_case_ = eval_examples
snake_case_ = post_process_function
def A_ ( self : Dict , lowercase_ : int=None , lowercase_ : Optional[Any]=None , lowercase_ : Optional[int]=None , lowercase_ : str = "eval" ):
snake_case_ = self.eval_dataset if eval_dataset is None else eval_dataset
snake_case_ = self.get_eval_dataloader(lowercase_ )
snake_case_ = self.eval_examples if eval_examples is None else eval_examples
# Temporarily disable metric computation, we will do it in the loop here.
snake_case_ = self.compute_metrics
snake_case_ = None
snake_case_ = self.prediction_loop if self.args.use_legacy_prediction_loop else self.evaluation_loop
snake_case_ = time.time()
try:
snake_case_ = eval_loop(
lowercase_ , description='''Evaluation''' , prediction_loss_only=True if compute_metrics is None else None , ignore_keys=lowercase_ , metric_key_prefix=lowercase_ , )
finally:
snake_case_ = compute_metrics
snake_case_ = self.args.eval_batch_size * self.args.world_size
if F"{metric_key_prefix}_jit_compilation_time" in output.metrics:
start_time += output.metrics[F"{metric_key_prefix}_jit_compilation_time"]
output.metrics.update(
speed_metrics(
lowercase_ , lowercase_ , num_samples=output.num_samples , num_steps=math.ceil(output.num_samples / total_batch_size ) , ) )
if self.post_process_function is not None and self.compute_metrics is not None and self.args.should_save:
# Only the main node write the results by default
snake_case_ = self.post_process_function(lowercase_ , lowercase_ , output.predictions )
snake_case_ = self.compute_metrics(lowercase_ )
# Prefix all keys with metric_key_prefix + '_'
for key in list(metrics.keys() ):
if not key.startswith(F"{metric_key_prefix}_" ):
snake_case_ = metrics.pop(lowercase_ )
metrics.update(output.metrics )
else:
snake_case_ = output.metrics
if self.args.should_log:
# Only the main node log the results by default
self.log(lowercase_ )
if self.args.tpu_metrics_debug or self.args.debug:
# tpu-comment: Logging debug metrics for PyTorch/XLA (compile, execute times, ops, etc.)
xm.master_print(met.metrics_report() )
snake_case_ = self.callback_handler.on_evaluate(self.args , self.state , self.control , lowercase_ )
return metrics
def A_ ( self : str , lowercase_ : Optional[Any] , lowercase_ : List[Any] , lowercase_ : Optional[int]=None , lowercase_ : str = "test" ):
snake_case_ = self.get_test_dataloader(lowercase_ )
# Temporarily disable metric computation, we will do it in the loop here.
snake_case_ = self.compute_metrics
snake_case_ = None
snake_case_ = self.prediction_loop if self.args.use_legacy_prediction_loop else self.evaluation_loop
snake_case_ = time.time()
try:
snake_case_ = eval_loop(
lowercase_ , description='''Prediction''' , prediction_loss_only=True if compute_metrics is None else None , ignore_keys=lowercase_ , metric_key_prefix=lowercase_ , )
finally:
snake_case_ = compute_metrics
snake_case_ = self.args.eval_batch_size * self.args.world_size
if F"{metric_key_prefix}_jit_compilation_time" in output.metrics:
start_time += output.metrics[F"{metric_key_prefix}_jit_compilation_time"]
output.metrics.update(
speed_metrics(
lowercase_ , lowercase_ , num_samples=output.num_samples , num_steps=math.ceil(output.num_samples / total_batch_size ) , ) )
if self.post_process_function is None or self.compute_metrics is None:
return output
snake_case_ = self.post_process_function(lowercase_ , lowercase_ , output.predictions , '''predict''' )
snake_case_ = self.compute_metrics(lowercase_ )
# Prefix all keys with metric_key_prefix + '_'
for key in list(metrics.keys() ):
if not key.startswith(F"{metric_key_prefix}_" ):
snake_case_ = metrics.pop(lowercase_ )
metrics.update(output.metrics )
return PredictionOutput(predictions=predictions.predictions , label_ids=predictions.label_ids , metrics=lowercase_ )
| 56
|
'''simple docstring'''
import math
from collections.abc import Callable
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
snake_case_ = xa
snake_case_ = xa
while True:
if x_n == x_na or function(__UpperCAmelCase ) == function(__UpperCAmelCase ):
raise ZeroDivisionError('''float division by zero, could not find root''' )
snake_case_ = x_na - (
function(__UpperCAmelCase ) / ((function(__UpperCAmelCase ) - function(__UpperCAmelCase )) / (x_na - x_n))
)
if abs(x_na - x_na ) < 10**-5:
return x_na
snake_case_ = x_na
snake_case_ = x_na
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
return math.pow(__UpperCAmelCase, 3 ) - (2 * x) - 5
if __name__ == "__main__":
print(intersection(f, 3, 3.5))
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
if principal <= 0:
raise Exception('''Principal borrowed must be > 0''' )
if rate_per_annum < 0:
raise Exception('''Rate of interest must be >= 0''' )
if years_to_repay <= 0 or not isinstance(__UpperCAmelCase, __UpperCAmelCase ):
raise Exception('''Years to repay must be an integer > 0''' )
# Yearly rate is divided by 12 to get monthly rate
snake_case_ = rate_per_annum / 12
# Years to repay is multiplied by 12 to get number of payments as payment is monthly
snake_case_ = years_to_repay * 12
return (
principal
* rate_per_month
* (1 + rate_per_month) ** number_of_payments
/ ((1 + rate_per_month) ** number_of_payments - 1)
)
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_url
from PIL import Image
from transformers import DPTConfig, DPTForDepthEstimation, DPTForSemanticSegmentation, DPTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
a : Any = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = DPTConfig()
if "large" in checkpoint_url:
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 24
snake_case_ = 16
snake_case_ = [5, 11, 17, 23]
snake_case_ = [256, 512, 1024, 1024]
snake_case_ = (1, 384, 384)
if "ade" in checkpoint_url:
snake_case_ = True
snake_case_ = 150
snake_case_ = '''huggingface/label-files'''
snake_case_ = '''ade20k-id2label.json'''
snake_case_ = json.load(open(cached_download(hf_hub_url(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ) ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = idalabel
snake_case_ = {v: k for k, v in idalabel.items()}
snake_case_ = [1, 150, 480, 480]
return config, expected_shape
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = ['''pretrained.model.head.weight''', '''pretrained.model.head.bias''']
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if (
"pretrained.model" in name
and "cls_token" not in name
and "pos_embed" not in name
and "patch_embed" not in name
):
snake_case_ = name.replace('''pretrained.model''', '''dpt.encoder''' )
if "pretrained.model" in name:
snake_case_ = name.replace('''pretrained.model''', '''dpt.embeddings''' )
if "patch_embed" in name:
snake_case_ = name.replace('''patch_embed''', '''patch_embeddings''' )
if "pos_embed" in name:
snake_case_ = name.replace('''pos_embed''', '''position_embeddings''' )
if "attn.proj" in name:
snake_case_ = name.replace('''attn.proj''', '''attention.output.dense''' )
if "proj" in name and "project" not in name:
snake_case_ = name.replace('''proj''', '''projection''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''layer''' )
if "mlp.fc1" in name:
snake_case_ = name.replace('''mlp.fc1''', '''intermediate.dense''' )
if "mlp.fc2" in name:
snake_case_ = name.replace('''mlp.fc2''', '''output.dense''' )
if "norm1" in name:
snake_case_ = name.replace('''norm1''', '''layernorm_before''' )
if "norm2" in name:
snake_case_ = name.replace('''norm2''', '''layernorm_after''' )
if "scratch.output_conv" in name:
snake_case_ = name.replace('''scratch.output_conv''', '''head''' )
if "scratch" in name:
snake_case_ = name.replace('''scratch''', '''neck''' )
if "layer1_rn" in name:
snake_case_ = name.replace('''layer1_rn''', '''convs.0''' )
if "layer2_rn" in name:
snake_case_ = name.replace('''layer2_rn''', '''convs.1''' )
if "layer3_rn" in name:
snake_case_ = name.replace('''layer3_rn''', '''convs.2''' )
if "layer4_rn" in name:
snake_case_ = name.replace('''layer4_rn''', '''convs.3''' )
if "refinenet" in name:
snake_case_ = int(name[len('''neck.refinenet''' ) : len('''neck.refinenet''' ) + 1] )
# tricky here: we need to map 4 to 0, 3 to 1, 2 to 2 and 1 to 3
snake_case_ = name.replace(F"refinenet{layer_idx}", F"fusion_stage.layers.{abs(layer_idx-4 )}" )
if "out_conv" in name:
snake_case_ = name.replace('''out_conv''', '''projection''' )
if "resConfUnit1" in name:
snake_case_ = name.replace('''resConfUnit1''', '''residual_layer1''' )
if "resConfUnit2" in name:
snake_case_ = name.replace('''resConfUnit2''', '''residual_layer2''' )
if "conv1" in name:
snake_case_ = name.replace('''conv1''', '''convolution1''' )
if "conv2" in name:
snake_case_ = name.replace('''conv2''', '''convolution2''' )
# readout blocks
if "pretrained.act_postprocess1.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.0.project.0''', '''neck.reassemble_stage.readout_projects.0.0''' )
if "pretrained.act_postprocess2.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.0.project.0''', '''neck.reassemble_stage.readout_projects.1.0''' )
if "pretrained.act_postprocess3.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.0.project.0''', '''neck.reassemble_stage.readout_projects.2.0''' )
if "pretrained.act_postprocess4.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.0.project.0''', '''neck.reassemble_stage.readout_projects.3.0''' )
# resize blocks
if "pretrained.act_postprocess1.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.3''', '''neck.reassemble_stage.layers.0.projection''' )
if "pretrained.act_postprocess1.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.4''', '''neck.reassemble_stage.layers.0.resize''' )
if "pretrained.act_postprocess2.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.3''', '''neck.reassemble_stage.layers.1.projection''' )
if "pretrained.act_postprocess2.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.4''', '''neck.reassemble_stage.layers.1.resize''' )
if "pretrained.act_postprocess3.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.3''', '''neck.reassemble_stage.layers.2.projection''' )
if "pretrained.act_postprocess4.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.3''', '''neck.reassemble_stage.layers.3.projection''' )
if "pretrained.act_postprocess4.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.4''', '''neck.reassemble_stage.layers.3.resize''' )
if "pretrained" in name:
snake_case_ = name.replace('''pretrained''', '''dpt''' )
if "bn" in name:
snake_case_ = name.replace('''bn''', '''batch_norm''' )
if "head" in name:
snake_case_ = name.replace('''head''', '''head.head''' )
if "encoder.norm" in name:
snake_case_ = name.replace('''encoder.norm''', '''layernorm''' )
if "auxlayer" in name:
snake_case_ = name.replace('''auxlayer''', '''auxiliary_head.head''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.weight" )
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.bias" )
# next, add query, keys and values (in that order) to the state dict
snake_case_ = in_proj_weight[: config.hidden_size, :]
snake_case_ = in_proj_bias[: config.hidden_size]
snake_case_ = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
snake_case_ = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
snake_case_ = in_proj_weight[
-config.hidden_size :, :
]
snake_case_ = in_proj_bias[-config.hidden_size :]
def __magic_name__ ( ) -> Any:
'''simple docstring'''
snake_case_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
snake_case_ = Image.open(requests.get(__UpperCAmelCase, stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dpt_config(__UpperCAmelCase )
# load original state_dict from URL
snake_case_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase, map_location='''cpu''' )
# remove certain keys
remove_ignore_keys_(__UpperCAmelCase )
# rename keys
for key in state_dict.copy().keys():
snake_case_ = state_dict.pop(__UpperCAmelCase )
snake_case_ = val
# read in qkv matrices
read_in_q_k_v(__UpperCAmelCase, __UpperCAmelCase )
# load HuggingFace model
snake_case_ = DPTForSemanticSegmentation(__UpperCAmelCase ) if '''ade''' in checkpoint_url else DPTForDepthEstimation(__UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
# Check outputs on an image
snake_case_ = 480 if '''ade''' in checkpoint_url else 384
snake_case_ = DPTImageProcessor(size=__UpperCAmelCase )
snake_case_ = prepare_img()
snake_case_ = image_processor(__UpperCAmelCase, return_tensors='''pt''' )
# forward pass
snake_case_ = model(**__UpperCAmelCase ).logits if '''ade''' in checkpoint_url else model(**__UpperCAmelCase ).predicted_depth
# Assert logits
snake_case_ = torch.tensor([[6.3_1_9_9, 6.3_6_2_9, 6.4_1_4_8], [6.3_8_5_0, 6.3_6_1_5, 6.4_1_6_6], [6.3_5_1_9, 6.3_1_7_6, 6.3_5_7_5]] )
if "ade" in checkpoint_url:
snake_case_ = torch.tensor([[4.0_4_8_0, 4.2_4_2_0, 4.4_3_6_0], [4.3_1_2_4, 4.5_6_9_3, 4.8_2_6_1], [4.5_7_6_8, 4.8_9_6_5, 5.2_1_6_3]] )
assert outputs.shape == torch.Size(__UpperCAmelCase )
assert (
torch.allclose(outputs[0, 0, :3, :3], __UpperCAmelCase, atol=1e-4 )
if "ade" in checkpoint_url
else torch.allclose(outputs[0, :3, :3], __UpperCAmelCase )
)
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
print(F"Saving model to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
print(F"Saving image processor to {pytorch_dump_folder_path}" )
image_processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print('''Pushing model to hub...''' )
model.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add model''', use_temp_dir=__UpperCAmelCase, )
image_processor.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add image processor''', use_temp_dir=__UpperCAmelCase, )
if __name__ == "__main__":
a : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--checkpoint_url',
default='https://github.com/intel-isl/DPT/releases/download/1_0/dpt_large-midas-2f21e586.pt',
type=str,
help='URL of the original DPT checkpoint you\'d like to convert.',
)
parser.add_argument(
'--pytorch_dump_folder_path',
default=None,
type=str,
required=True,
help='Path to the output PyTorch model directory.',
)
parser.add_argument(
'--push_to_hub',
action='store_true',
)
parser.add_argument(
'--model_name',
default='dpt-large',
type=str,
help='Name of the model, in case you\'re pushing to the hub.',
)
a : List[Any] = parser.parse_args()
convert_dpt_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 56
| 1
|
'''simple docstring'''
import inspect
import unittest
from transformers import SegformerConfig, is_torch_available, is_vision_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, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
MODEL_MAPPING,
SegformerForImageClassification,
SegformerForSemanticSegmentation,
SegformerModel,
)
from transformers.models.segformer.modeling_segformer import SEGFORMER_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import SegformerImageProcessor
class a ( _lowerCamelCase ):
def A_ ( self : Optional[int] ):
snake_case_ = self.config_class(**self.inputs_dict )
self.parent.assertTrue(hasattr(lowercase_ , '''hidden_sizes''' ) )
self.parent.assertTrue(hasattr(lowercase_ , '''num_attention_heads''' ) )
self.parent.assertTrue(hasattr(lowercase_ , '''num_encoder_blocks''' ) )
class a :
def __init__( self : Optional[int] , lowercase_ : Optional[Any] , lowercase_ : Any=13 , lowercase_ : Optional[Any]=64 , lowercase_ : Any=3 , lowercase_ : Optional[Any]=4 , lowercase_ : Dict=[2, 2, 2, 2] , lowercase_ : int=[8, 4, 2, 1] , lowercase_ : str=[16, 32, 64, 128] , lowercase_ : Optional[Any]=[1, 4, 8, 16] , lowercase_ : Any=[1, 2, 4, 8] , lowercase_ : Optional[int]=True , lowercase_ : str=True , lowercase_ : List[str]="gelu" , lowercase_ : List[str]=0.1 , lowercase_ : Any=0.1 , lowercase_ : Any=0.02 , lowercase_ : Union[str, Any]=3 , lowercase_ : Union[str, Any]=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = image_size
snake_case_ = num_channels
snake_case_ = num_encoder_blocks
snake_case_ = sr_ratios
snake_case_ = depths
snake_case_ = hidden_sizes
snake_case_ = downsampling_rates
snake_case_ = num_attention_heads
snake_case_ = is_training
snake_case_ = use_labels
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = scope
def A_ ( self : str ):
snake_case_ = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size, self.image_size, self.image_size] , self.num_labels )
snake_case_ = self.get_config()
return config, pixel_values, labels
def A_ ( self : int ):
return SegformerConfig(
image_size=self.image_size , num_channels=self.num_channels , num_encoder_blocks=self.num_encoder_blocks , depths=self.depths , hidden_sizes=self.hidden_sizes , num_attention_heads=self.num_attention_heads , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , initializer_range=self.initializer_range , )
def A_ ( self : Any , lowercase_ : Optional[int] , lowercase_ : Any , lowercase_ : Any ):
snake_case_ = SegformerModel(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ )
snake_case_ = snake_case_ = self.image_size // (self.downsampling_rates[-1] * 2)
self.parent.assertEqual(
result.last_hidden_state.shape , (self.batch_size, self.hidden_sizes[-1], expected_height, expected_width) )
def A_ ( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] , lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = SegformerForSemanticSegmentation(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_labels, self.image_size // 4, self.image_size // 4) )
snake_case_ = model(lowercase_ , labels=lowercase_ )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_labels, self.image_size // 4, self.image_size // 4) )
self.parent.assertGreater(result.loss , 0.0 )
def A_ ( self : Any , lowercase_ : Union[str, Any] , lowercase_ : Tuple , lowercase_ : str ):
snake_case_ = 1
snake_case_ = SegformerForSemanticSegmentation(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = torch.randint(0 , 1 , (self.batch_size, self.image_size, self.image_size) ).to(lowercase_ )
snake_case_ = model(lowercase_ , labels=lowercase_ )
self.parent.assertGreater(result.loss , 0.0 )
def A_ ( self : List[str] ):
snake_case_ = self.prepare_config_and_inputs()
snake_case_ ,snake_case_ ,snake_case_ = config_and_inputs
snake_case_ = {'''pixel_values''': pixel_values}
return config, inputs_dict
@require_torch
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
SegformerModel,
SegformerForSemanticSegmentation,
SegformerForImageClassification,
)
if is_torch_available()
else ()
)
snake_case_ = (
{
"feature-extraction": SegformerModel,
"image-classification": SegformerForImageClassification,
"image-segmentation": SegformerForSemanticSegmentation,
}
if is_torch_available()
else {}
)
snake_case_ = True
snake_case_ = False
snake_case_ = False
snake_case_ = False
def A_ ( self : Any ):
snake_case_ = SegformerModelTester(self )
snake_case_ = SegformerConfigTester(self , config_class=lowercase_ )
def A_ ( self : Tuple ):
self.config_tester.run_common_tests()
def A_ ( self : int ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*lowercase_ )
def A_ ( self : int ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_binary_image_segmentation(*lowercase_ )
def A_ ( self : Any ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_segmentation(*lowercase_ )
@unittest.skip('''SegFormer does not use inputs_embeds''' )
def A_ ( self : str ):
pass
@unittest.skip('''SegFormer does not have get_input_embeddings method and get_output_embeddings methods''' )
def A_ ( self : List[Any] ):
pass
def A_ ( self : Optional[Any] ):
snake_case_ ,snake_case_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ = model_class(lowercase_ )
snake_case_ = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ = [*signature.parameters.keys()]
snake_case_ = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , lowercase_ )
def A_ ( self : str ):
snake_case_ ,snake_case_ = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ = True
for model_class in self.all_model_classes:
snake_case_ = True
snake_case_ = False
snake_case_ = True
snake_case_ = model_class(lowercase_ )
model.to(lowercase_ )
model.eval()
with torch.no_grad():
snake_case_ = model(**self._prepare_for_class(lowercase_ , lowercase_ ) )
snake_case_ = outputs.attentions
snake_case_ = sum(self.model_tester.depths )
self.assertEqual(len(lowercase_ ) , lowercase_ )
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
snake_case_ = True
snake_case_ = model_class(lowercase_ )
model.to(lowercase_ )
model.eval()
with torch.no_grad():
snake_case_ = model(**self._prepare_for_class(lowercase_ , lowercase_ ) )
snake_case_ = outputs.attentions
self.assertEqual(len(lowercase_ ) , lowercase_ )
# verify the first attentions (first block, first layer)
snake_case_ = (self.model_tester.image_size // 4) ** 2
snake_case_ = (self.model_tester.image_size // (4 * self.model_tester.sr_ratios[0])) ** 2
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads[0], expected_seq_len, expected_reduced_seq_len] , )
# verify the last attentions (last block, last layer)
snake_case_ = (self.model_tester.image_size // 32) ** 2
snake_case_ = (self.model_tester.image_size // (32 * self.model_tester.sr_ratios[-1])) ** 2
self.assertListEqual(
list(attentions[-1].shape[-3:] ) , [self.model_tester.num_attention_heads[-1], expected_seq_len, expected_reduced_seq_len] , )
snake_case_ = len(lowercase_ )
# Check attention is always last and order is fine
snake_case_ = True
snake_case_ = True
snake_case_ = model_class(lowercase_ )
model.to(lowercase_ )
model.eval()
with torch.no_grad():
snake_case_ = model(**self._prepare_for_class(lowercase_ , lowercase_ ) )
self.assertEqual(out_len + 1 , len(lowercase_ ) )
snake_case_ = outputs.attentions
self.assertEqual(len(lowercase_ ) , lowercase_ )
# verify the first attentions (first block, first layer)
snake_case_ = (self.model_tester.image_size // 4) ** 2
snake_case_ = (self.model_tester.image_size // (4 * self.model_tester.sr_ratios[0])) ** 2
self.assertListEqual(
list(self_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads[0], expected_seq_len, expected_reduced_seq_len] , )
def A_ ( self : str ):
def check_hidden_states_output(lowercase_ : str , lowercase_ : List[str] , lowercase_ : List[str] ):
snake_case_ = model_class(lowercase_ )
model.to(lowercase_ )
model.eval()
with torch.no_grad():
snake_case_ = model(**self._prepare_for_class(lowercase_ , lowercase_ ) )
snake_case_ = outputs.hidden_states
snake_case_ = self.model_tester.num_encoder_blocks
self.assertEqual(len(lowercase_ ) , lowercase_ )
# verify the first hidden states (first block)
self.assertListEqual(
list(hidden_states[0].shape[-3:] ) , [
self.model_tester.hidden_sizes[0],
self.model_tester.image_size // 4,
self.model_tester.image_size // 4,
] , )
snake_case_ ,snake_case_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ = True
check_hidden_states_output(lowercase_ , lowercase_ , lowercase_ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case_ = True
check_hidden_states_output(lowercase_ , lowercase_ , lowercase_ )
def A_ ( self : Union[str, Any] ):
if not self.model_tester.is_training:
return
snake_case_ ,snake_case_ = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ = True
for model_class in self.all_model_classes:
if model_class in get_values(lowercase_ ):
continue
snake_case_ = model_class(lowercase_ )
model.to(lowercase_ )
model.train()
snake_case_ = self._prepare_for_class(lowercase_ , lowercase_ , return_labels=lowercase_ )
snake_case_ = model(**lowercase_ ).loss
loss.backward()
@unittest.skip('''Will be fixed soon by reducing the size of the model used for common tests.''' )
def A_ ( self : List[Any] ):
pass
@slow
def A_ ( self : Dict ):
for model_name in SEGFORMER_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ = SegformerModel.from_pretrained(lowercase_ )
self.assertIsNotNone(lowercase_ )
def __magic_name__ ( ) -> Optional[int]:
'''simple docstring'''
snake_case_ = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_torch
class a ( unittest.TestCase ):
@slow
def A_ ( self : List[Any] ):
# only resize + normalize
snake_case_ = SegformerImageProcessor(
image_scale=(512, 512) , keep_ratio=lowercase_ , align=lowercase_ , do_random_crop=lowercase_ )
snake_case_ = SegformerForSemanticSegmentation.from_pretrained('''nvidia/segformer-b0-finetuned-ade-512-512''' ).to(
lowercase_ )
snake_case_ = prepare_img()
snake_case_ = image_processor(images=lowercase_ , return_tensors='''pt''' )
snake_case_ = encoded_inputs.pixel_values.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ )
snake_case_ = torch.Size((1, model.config.num_labels, 128, 128) )
self.assertEqual(outputs.logits.shape , lowercase_ )
snake_case_ = torch.tensor(
[
[[-4.6310, -5.5232, -6.2356], [-5.1921, -6.1444, -6.5996], [-5.4424, -6.2790, -6.7574]],
[[-12.1391, -13.3122, -13.9554], [-12.8732, -13.9352, -14.3563], [-12.9438, -13.8226, -14.2513]],
[[-12.5134, -13.4686, -14.4915], [-12.8669, -14.4343, -14.7758], [-13.2523, -14.5819, -15.0694]],
] ).to(lowercase_ )
self.assertTrue(torch.allclose(outputs.logits[0, :3, :3, :3] , lowercase_ , atol=1e-4 ) )
@slow
def A_ ( self : List[str] ):
# only resize + normalize
snake_case_ = SegformerImageProcessor(
image_scale=(512, 512) , keep_ratio=lowercase_ , align=lowercase_ , do_random_crop=lowercase_ )
snake_case_ = SegformerForSemanticSegmentation.from_pretrained(
'''nvidia/segformer-b1-finetuned-cityscapes-1024-1024''' ).to(lowercase_ )
snake_case_ = prepare_img()
snake_case_ = image_processor(images=lowercase_ , return_tensors='''pt''' )
snake_case_ = encoded_inputs.pixel_values.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ )
snake_case_ = torch.Size((1, model.config.num_labels, 128, 128) )
self.assertEqual(outputs.logits.shape , lowercase_ )
snake_case_ = torch.tensor(
[
[[-13.5748, -13.9111, -12.6500], [-14.3500, -15.3683, -14.2328], [-14.7532, -16.0424, -15.6087]],
[[-17.1651, -15.8725, -12.9653], [-17.2580, -17.3718, -14.8223], [-16.6058, -16.8783, -16.7452]],
[[-3.6456, -3.0209, -1.4203], [-3.0797, -3.1959, -2.0000], [-1.8757, -1.9217, -1.6997]],
] ).to(lowercase_ )
self.assertTrue(torch.allclose(outputs.logits[0, :3, :3, :3] , lowercase_ , atol=1e-1 ) )
@slow
def A_ ( self : str ):
# only resize + normalize
snake_case_ = SegformerImageProcessor(
image_scale=(512, 512) , keep_ratio=lowercase_ , align=lowercase_ , do_random_crop=lowercase_ )
snake_case_ = SegformerForSemanticSegmentation.from_pretrained('''nvidia/segformer-b0-finetuned-ade-512-512''' ).to(
lowercase_ )
snake_case_ = prepare_img()
snake_case_ = image_processor(images=lowercase_ , return_tensors='''pt''' )
snake_case_ = encoded_inputs.pixel_values.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ )
snake_case_ = outputs.logits.detach().cpu()
snake_case_ = image_processor.post_process_semantic_segmentation(outputs=lowercase_ , target_sizes=[(500, 300)] )
snake_case_ = torch.Size((500, 300) )
self.assertEqual(segmentation[0].shape , lowercase_ )
snake_case_ = image_processor.post_process_semantic_segmentation(outputs=lowercase_ )
snake_case_ = torch.Size((128, 128) )
self.assertEqual(segmentation[0].shape , lowercase_ )
| 56
|
'''simple docstring'''
import re
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
snake_case_ = re.compile(
r'''^(?:0|94|\+94|0{2}94)''' r'''7(0|1|2|4|5|6|7|8)''' r'''(-| |)''' r'''\d{7}$''' )
return bool(re.search(__UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
a : Any = '0094702343221'
print(is_sri_lankan_phone_number(phone))
| 56
| 1
|
'''simple docstring'''
import enum
import os
from hashlib import shaaaa
from typing import Optional
from .. import config
from .logging import get_logger
a : int = get_logger(__name__)
class a ( enum.Enum ):
snake_case_ = "all_checks"
snake_case_ = "basic_checks"
snake_case_ = "no_checks"
class a ( _lowerCamelCase ):
pass
class a ( _lowerCamelCase ):
pass
class a ( _lowerCamelCase ):
pass
class a ( _lowerCamelCase ):
pass
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase=None ) -> Dict:
'''simple docstring'''
if expected_checksums is None:
logger.info('''Unable to verify checksums.''' )
return
if len(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) > 0:
raise ExpectedMoreDownloadedFiles(str(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) )
if len(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) > 0:
raise UnexpectedDownloadedFile(str(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) )
snake_case_ = [url for url in expected_checksums if expected_checksums[url] != recorded_checksums[url]]
snake_case_ = ''' for ''' + verification_name if verification_name is not None else ''''''
if len(__UpperCAmelCase ) > 0:
raise NonMatchingChecksumError(
F"Checksums didn't match{for_verification_name}:\n"
F"{bad_urls}\n"
'''Set `verification_mode=\'no_checks\'` to skip checksums verification and ignore this error''' )
logger.info('''All the checksums matched successfully''' + for_verification_name )
class a ( _lowerCamelCase ):
pass
class a ( _lowerCamelCase ):
pass
class a ( _lowerCamelCase ):
pass
class a ( _lowerCamelCase ):
pass
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Tuple:
'''simple docstring'''
if expected_splits is None:
logger.info('''Unable to verify splits sizes.''' )
return
if len(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) > 0:
raise ExpectedMoreSplits(str(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) )
if len(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) > 0:
raise UnexpectedSplits(str(set(__UpperCAmelCase ) - set(__UpperCAmelCase ) ) )
snake_case_ = [
{'''expected''': expected_splits[name], '''recorded''': recorded_splits[name]}
for name in expected_splits
if expected_splits[name].num_examples != recorded_splits[name].num_examples
]
if len(__UpperCAmelCase ) > 0:
raise NonMatchingSplitsSizesError(str(__UpperCAmelCase ) )
logger.info('''All the splits matched successfully.''' )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase = True ) -> dict:
'''simple docstring'''
if record_checksum:
snake_case_ = shaaaa()
with open(__UpperCAmelCase, '''rb''' ) as f:
for chunk in iter(lambda: f.read(1 << 20 ), b'''''' ):
m.update(__UpperCAmelCase )
snake_case_ = m.hexdigest()
else:
snake_case_ = None
return {"num_bytes": os.path.getsize(__UpperCAmelCase ), "checksum": checksum}
def __magic_name__ ( __UpperCAmelCase ) -> Tuple:
'''simple docstring'''
if dataset_size and config.IN_MEMORY_MAX_SIZE:
return dataset_size < config.IN_MEMORY_MAX_SIZE
else:
return False
| 56
|
'''simple docstring'''
import re
from filelock import FileLock
try:
import nltk
a : Union[str, Any] = True
except (ImportError, ModuleNotFoundError):
a : Any = False
if NLTK_AVAILABLE:
with FileLock('.lock') as lock:
nltk.download('punkt', quiet=True)
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
re.sub('''<n>''', '''''', __UpperCAmelCase ) # remove pegasus newline char
assert NLTK_AVAILABLE, "nltk must be installed to separate newlines between sentences. (pip install nltk)"
return "\n".join(nltk.sent_tokenize(__UpperCAmelCase ) )
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
while b:
snake_case_ ,snake_case_ = b, a % b
return a
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
return a if b == 0 else euclidean_gcd_recursive(__UpperCAmelCase, a % b )
def __magic_name__ ( ) -> Any:
'''simple docstring'''
print(F"euclidean_gcd(3, 5) = {euclidean_gcd(3, 5 )}" )
print(F"euclidean_gcd(5, 3) = {euclidean_gcd(5, 3 )}" )
print(F"euclidean_gcd(1, 3) = {euclidean_gcd(1, 3 )}" )
print(F"euclidean_gcd(3, 6) = {euclidean_gcd(3, 6 )}" )
print(F"euclidean_gcd(6, 3) = {euclidean_gcd(6, 3 )}" )
print(F"euclidean_gcd_recursive(3, 5) = {euclidean_gcd_recursive(3, 5 )}" )
print(F"euclidean_gcd_recursive(5, 3) = {euclidean_gcd_recursive(5, 3 )}" )
print(F"euclidean_gcd_recursive(1, 3) = {euclidean_gcd_recursive(1, 3 )}" )
print(F"euclidean_gcd_recursive(3, 6) = {euclidean_gcd_recursive(3, 6 )}" )
print(F"euclidean_gcd_recursive(6, 3) = {euclidean_gcd_recursive(6, 3 )}" )
if __name__ == "__main__":
main()
| 56
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
a : Tuple = {
'configuration_llama': ['LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP', 'LlamaConfig'],
}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = ['LlamaTokenizer']
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : str = ['LlamaTokenizerFast']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
a : Optional[Any] = [
'LlamaForCausalLM',
'LlamaModel',
'LlamaPreTrainedModel',
'LlamaForSequenceClassification',
]
if TYPE_CHECKING:
from .configuration_llama import LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP, LlamaConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama import LlamaTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_llama_fast import LlamaTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_llama import LlamaForCausalLM, LlamaForSequenceClassification, LlamaModel, LlamaPreTrainedModel
else:
import sys
a : str = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 56
| 1
|
'''simple docstring'''
import tempfile
import unittest
from make_student import create_student_by_copying_alternating_layers
from transformers import AutoConfig
from transformers.file_utils import cached_property
from transformers.testing_utils import require_torch
a : List[Any] = 'sshleifer/bart-tiny-random'
a : Optional[Any] = 'patrickvonplaten/t5-tiny-random'
@require_torch
class a ( unittest.TestCase ):
@cached_property
def A_ ( self : Union[str, Any] ):
return AutoConfig.from_pretrained(lowercase_ )
def A_ ( self : str ):
snake_case_ ,*snake_case_ = create_student_by_copying_alternating_layers(lowercase_ , tempfile.mkdtemp() , e=1 , d=1 )
self.assertEqual(student.config.num_hidden_layers , 1 )
def A_ ( self : List[Any] ):
snake_case_ ,*snake_case_ = create_student_by_copying_alternating_layers(lowercase_ , tempfile.mkdtemp() , e=1 , d=lowercase_ )
def A_ ( self : List[str] ):
snake_case_ ,*snake_case_ = create_student_by_copying_alternating_layers(lowercase_ , tempfile.mkdtemp() , e=1 , d=lowercase_ )
self.assertEqual(student.config.encoder_layers , 1 )
self.assertEqual(student.config.decoder_layers , self.teacher_config.encoder_layers )
def A_ ( self : List[Any] ):
snake_case_ ,*snake_case_ = create_student_by_copying_alternating_layers(lowercase_ , tempfile.mkdtemp() , e=1 , d=1 )
self.assertEqual(student.config.encoder_layers , 1 )
self.assertEqual(student.config.decoder_layers , 1 )
def A_ ( self : Dict ):
with self.assertRaises(lowercase_ ):
create_student_by_copying_alternating_layers(lowercase_ , tempfile.mkdtemp() , e=lowercase_ , d=lowercase_ )
| 56
|
'''simple docstring'''
import re
from typing import Callable, List, Optional, Union
import tensorflow as tf
try:
from tensorflow.keras.optimizers.legacy import Adam
except ImportError:
from tensorflow.keras.optimizers import Adam
class a ( tf.keras.optimizers.schedules.LearningRateSchedule ):
def __init__( self : Optional[Any] , lowercase_ : float , lowercase_ : Callable , lowercase_ : int , lowercase_ : float = 1.0 , lowercase_ : str = None , ):
super().__init__()
snake_case_ = initial_learning_rate
snake_case_ = warmup_steps
snake_case_ = power
snake_case_ = decay_schedule_fn
snake_case_ = name
def __call__( self : Tuple , lowercase_ : str ):
with tf.name_scope(self.name or '''WarmUp''' ) as name:
# Implements polynomial warmup. i.e., if global_step < warmup_steps, the
# learning rate will be `global_step/num_warmup_steps * init_lr`.
snake_case_ = tf.cast(lowercase_ , tf.floataa )
snake_case_ = tf.cast(self.warmup_steps , tf.floataa )
snake_case_ = global_step_float / warmup_steps_float
snake_case_ = self.initial_learning_rate * tf.math.pow(lowercase_ , self.power )
return tf.cond(
global_step_float < warmup_steps_float , lambda: warmup_learning_rate , lambda: self.decay_schedule_fn(step - self.warmup_steps ) , name=lowercase_ , )
def A_ ( self : Any ):
return {
"initial_learning_rate": self.initial_learning_rate,
"decay_schedule_fn": self.decay_schedule_fn,
"warmup_steps": self.warmup_steps,
"power": self.power,
"name": self.name,
}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, __UpperCAmelCase = 0.9, __UpperCAmelCase = 0.9_9_9, __UpperCAmelCase = 1e-8, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = 0.0, __UpperCAmelCase = 1.0, __UpperCAmelCase = None, ) -> List[str]:
'''simple docstring'''
snake_case_ = tf.keras.optimizers.schedules.PolynomialDecay(
initial_learning_rate=__UpperCAmelCase, decay_steps=num_train_steps - num_warmup_steps, end_learning_rate=init_lr * min_lr_ratio, power=__UpperCAmelCase, )
if num_warmup_steps:
snake_case_ = WarmUp(
initial_learning_rate=__UpperCAmelCase, decay_schedule_fn=__UpperCAmelCase, warmup_steps=__UpperCAmelCase, )
if weight_decay_rate > 0.0:
snake_case_ = AdamWeightDecay(
learning_rate=__UpperCAmelCase, weight_decay_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, exclude_from_weight_decay=['''LayerNorm''', '''layer_norm''', '''bias'''], include_in_weight_decay=__UpperCAmelCase, )
else:
snake_case_ = tf.keras.optimizers.Adam(
learning_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, )
# We return the optimizer and the LR scheduler in order to better track the
# evolution of the LR independently of the optimizer.
return optimizer, lr_schedule
class a ( _lowerCamelCase ):
def __init__( self : Dict , lowercase_ : Union[float, tf.keras.optimizers.schedules.LearningRateSchedule] = 0.001 , lowercase_ : float = 0.9 , lowercase_ : float = 0.999 , lowercase_ : float = 1e-7 , lowercase_ : bool = False , lowercase_ : float = 0.0 , lowercase_ : Optional[List[str]] = None , lowercase_ : Optional[List[str]] = None , lowercase_ : str = "AdamWeightDecay" , **lowercase_ : Optional[int] , ):
super().__init__(lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
snake_case_ = weight_decay_rate
snake_case_ = include_in_weight_decay
snake_case_ = exclude_from_weight_decay
@classmethod
def A_ ( cls : Dict , lowercase_ : Union[str, Any] ):
snake_case_ = {'''WarmUp''': WarmUp}
return super(lowercase_ , cls ).from_config(lowercase_ , custom_objects=lowercase_ )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : Optional[int] ):
super(lowercase_ , self )._prepare_local(lowercase_ , lowercase_ , lowercase_ )
snake_case_ = tf.constant(
self.weight_decay_rate , name='''adam_weight_decay_rate''' )
def A_ ( self : str , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Any ):
snake_case_ = self._do_use_weight_decay(var.name )
if do_decay:
return var.assign_sub(
learning_rate * var * apply_state[(var.device, var.dtype.base_dtype)]['''weight_decay_rate'''] , use_locking=self._use_locking , )
return tf.no_op()
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : str=None , **lowercase_ : List[str] ):
snake_case_ ,snake_case_ = list(zip(*lowercase_ ) )
return super(lowercase_ , self ).apply_gradients(zip(lowercase_ , lowercase_ ) , name=lowercase_ , **lowercase_ )
def A_ ( self : List[Any] , lowercase_ : str , lowercase_ : str , lowercase_ : Any ):
if apply_state is None:
return self._decayed_lr_t[var_dtype], {}
snake_case_ = apply_state or {}
snake_case_ = apply_state.get((var_device, var_dtype) )
if coefficients is None:
snake_case_ = self._fallback_apply_state(lowercase_ , lowercase_ )
snake_case_ = coefficients
return coefficients["lr_t"], {"apply_state": apply_state}
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : Union[str, Any] , lowercase_ : Optional[int]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_dense(lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : int , lowercase_ : str , lowercase_ : List[Any]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_sparse(lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = super().get_config()
config.update({'''weight_decay_rate''': self.weight_decay_rate} )
return config
def A_ ( self : Optional[int] , lowercase_ : int ):
if self.weight_decay_rate == 0:
return False
if self._include_in_weight_decay:
for r in self._include_in_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return True
if self._exclude_from_weight_decay:
for r in self._exclude_from_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return False
return True
class a ( _lowerCamelCase ):
def __init__( self : List[Any] ):
snake_case_ = []
snake_case_ = None
@property
def A_ ( self : Union[str, Any] ):
if self._accum_steps is None:
snake_case_ = tf.Variable(
tf.constant(0 , dtype=tf.intaa ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
return self._accum_steps.value()
@property
def A_ ( self : Dict ):
if not self._gradients:
raise ValueError('''The accumulator should be called first to initialize the gradients''' )
return [gradient.value() if gradient is not None else gradient for gradient in self._gradients]
def __call__( self : Any , lowercase_ : int ):
if not self._gradients:
snake_case_ = self.step # Create the step variable.
self._gradients.extend(
[
tf.Variable(
tf.zeros_like(lowercase_ ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
if gradient is not None
else gradient
for gradient in gradients
] )
if len(lowercase_ ) != len(self._gradients ):
raise ValueError(F"Expected {len(self._gradients )} gradients, but got {len(lowercase_ )}" )
for accum_gradient, gradient in zip(self._gradients , lowercase_ ):
if accum_gradient is not None and gradient is not None:
accum_gradient.assign_add(lowercase_ )
self._accum_steps.assign_add(1 )
def A_ ( self : Optional[int] ):
if not self._gradients:
return
self._accum_steps.assign(0 )
for gradient in self._gradients:
if gradient is not None:
gradient.assign(tf.zeros_like(lowercase_ ) )
| 56
| 1
|
'''simple docstring'''
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
a : Union[str, Any] = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f'''{bindir}/../../examples/pytorch/translation'''):
from run_translation import main # noqa
set_seed(42)
a : List[str] = 'sshleifer/student_marian_en_ro_6_1'
a : Dict = 'sshleifer/tiny-mbart'
@require_torch
class a ( _lowerCamelCase ):
def A_ ( self : int , lowercase_ : Any=False , lowercase_ : int=None , lowercase_ : List[str]=True , lowercase_ : Union[str, Any]=True , lowercase_ : Tuple=True , lowercase_ : str=True , ):
snake_case_ = 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_ , )
snake_case_ = TrainerState.load_from_json(os.path.join(lowercase_ , '''trainer_state.json''' ) ).log_history
if not do_eval:
return
snake_case_ = [log for log in logs if '''eval_loss''' in log.keys()]
snake_case_ = eval_metrics[0]
if predict_with_generate:
assert "eval_bleu" in first_step_stats
snake_case_ = 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 : Any ):
self.run_seqaseq_quick()
@require_torch_multi_gpu
def A_ ( self : Optional[int] ):
self.run_seqaseq_quick(distributed=lowercase_ )
@require_torch_multi_gpu
def A_ ( self : str ):
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 : int ):
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 : 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 : Tuple ):
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 ):
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 : int ):
# 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 : Optional[int] , lowercase_ : Dict ):
# as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout
snake_case_ = {
# 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},
}
snake_case_ = experiments[experiment_id]
snake_case_ = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False}
snake_case_ = '''Running training'''
with CaptureStderr() as cl:
self.run_seqaseq_quick(**lowercase_ , extra_args_str=data['''extra_args_str'''] )
snake_case_ = len(re.findall(lowercase_ , cl.err ) )
self.assertEqual(lowercase_ , data['''n_matches'''] )
@slow
def A_ ( self : Optional[Any] ):
snake_case_ = self.run_trainer(
eval_steps=2 , max_len=128 , model_name=lowercase_ , learning_rate=3e-4 , num_train_epochs=10 , distributed=lowercase_ , )
# Check metrics
snake_case_ = TrainerState.load_from_json(os.path.join(lowercase_ , '''trainer_state.json''' ) ).log_history
snake_case_ = [log for log in logs if '''eval_loss''' in log.keys()]
snake_case_ = eval_metrics[0]
snake_case_ = 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
snake_case_ = os.listdir(lowercase_ )
snake_case_ = {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 : Tuple ):
from transformers.training_args import OptimizerNames
def train_and_return_metrics(lowercase_ : str ) -> Tuple[int, float]:
snake_case_ = '''--skip_memory_metrics 0'''
snake_case_ = 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
snake_case_ = TrainerState.load_from_json(Path(lowercase_ , '''trainer_state.json''' ) ).log_history
snake_case_ = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20 )
snake_case_ = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20 )
snake_case_ = logs[0]['''train_loss''']
return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss
snake_case_ ,snake_case_ ,snake_case_ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value )
snake_case_ ,snake_case_ ,snake_case_ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value )
snake_case_ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb
snake_case_ = gpu_peak_mem_orig + gpu_alloc_mem_orig
snake_case_ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb
snake_case_ = 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
snake_case_ = 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 : int , 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 , ):
snake_case_ = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro'''
snake_case_ = self.get_auto_remove_tmp_dir()
snake_case_ = 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()
snake_case_ = 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()
snake_case_ = '''
--do_predict
'''.split()
snake_case_ = []
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:
snake_case_ = get_gpu_count()
snake_case_ = get_torch_dist_unique_port()
snake_case_ = 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()
snake_case_ = [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:
snake_case_ = ['''run_translation.py'''] + args
with patch.object(lowercase_ , '''argv''' , lowercase_ ):
main()
return output_dir
| 56
|
'''simple docstring'''
import gc
import unittest
import torch
from parameterized import parameterized
from diffusers import AutoencoderKL
from diffusers.utils import floats_tensor, load_hf_numpy, require_torch_gpu, slow, torch_all_close, torch_device
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = AutoencoderKL
snake_case_ = "sample"
snake_case_ = 1e-2
@property
def A_ ( self : Dict ):
snake_case_ = 4
snake_case_ = 3
snake_case_ = (32, 32)
snake_case_ = floats_tensor((batch_size, num_channels) + sizes ).to(lowercase_ )
return {"sample": image}
@property
def A_ ( self : List[Any] ):
return (3, 32, 32)
@property
def A_ ( self : Dict ):
return (3, 32, 32)
def A_ ( self : Union[str, Any] ):
snake_case_ = {
'''block_out_channels''': [32, 64],
'''in_channels''': 3,
'''out_channels''': 3,
'''down_block_types''': ['''DownEncoderBlock2D''', '''DownEncoderBlock2D'''],
'''up_block_types''': ['''UpDecoderBlock2D''', '''UpDecoderBlock2D'''],
'''latent_channels''': 4,
}
snake_case_ = self.dummy_input
return init_dict, inputs_dict
def A_ ( self : Any ):
pass
def A_ ( self : str ):
pass
@unittest.skipIf(torch_device == '''mps''' , '''Gradient checkpointing skipped on MPS''' )
def A_ ( self : Dict ):
# enable deterministic behavior for gradient checkpointing
snake_case_ ,snake_case_ = self.prepare_init_args_and_inputs_for_common()
snake_case_ = self.model_class(**lowercase_ )
model.to(lowercase_ )
assert not model.is_gradient_checkpointing and model.training
snake_case_ = model(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model.zero_grad()
snake_case_ = torch.randn_like(lowercase_ )
snake_case_ = (out - labels).mean()
loss.backward()
# re-instantiate the model now enabling gradient checkpointing
snake_case_ = self.model_class(**lowercase_ )
# clone model
model_a.load_state_dict(model.state_dict() )
model_a.to(lowercase_ )
model_a.enable_gradient_checkpointing()
assert model_a.is_gradient_checkpointing and model_a.training
snake_case_ = model_a(**lowercase_ ).sample
# run the backwards pass on the model. For backwards pass, for simplicity purpose,
# we won't calculate the loss and rather backprop on out.sum()
model_a.zero_grad()
snake_case_ = (out_a - labels).mean()
loss_a.backward()
# compare the output and parameters gradients
self.assertTrue((loss - loss_a).abs() < 1e-5 )
snake_case_ = dict(model.named_parameters() )
snake_case_ = dict(model_a.named_parameters() )
for name, param in named_params.items():
self.assertTrue(torch_all_close(param.grad.data , named_params_a[name].grad.data , atol=5e-5 ) )
def A_ ( self : Tuple ):
snake_case_ ,snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' , output_loading_info=lowercase_ )
self.assertIsNotNone(lowercase_ )
self.assertEqual(len(loading_info['''missing_keys'''] ) , 0 )
model.to(lowercase_ )
snake_case_ = model(**self.dummy_input )
assert image is not None, "Make sure output is not None"
def A_ ( self : Tuple ):
snake_case_ = AutoencoderKL.from_pretrained('''fusing/autoencoder-kl-dummy''' )
snake_case_ = model.to(lowercase_ )
model.eval()
if torch_device == "mps":
snake_case_ = torch.manual_seed(0 )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(0 )
snake_case_ = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ = image.to(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , sample_posterior=lowercase_ , generator=lowercase_ ).sample
snake_case_ = output[0, -1, -3:, -3:].flatten().cpu()
# Since the VAE Gaussian prior's generator is seeded on the appropriate device,
# the expected output slices are not the same for CPU and GPU.
if torch_device == "mps":
snake_case_ = torch.tensor(
[
-4.0_078e-01,
-3.8_323e-04,
-1.2_681e-01,
-1.1_462e-01,
2.0_095e-01,
1.0_893e-01,
-8.8_247e-02,
-3.0_361e-01,
-9.8_644e-03,
] )
elif torch_device == "cpu":
snake_case_ = torch.tensor(
[-0.1352, 0.0878, 0.0419, -0.0818, -0.1069, 0.0688, -0.1458, -0.4446, -0.0026] )
else:
snake_case_ = torch.tensor(
[-0.2421, 0.4642, 0.2507, -0.0438, 0.0682, 0.3160, -0.2018, -0.0727, 0.2485] )
self.assertTrue(torch_all_close(lowercase_ , lowercase_ , rtol=1e-2 ) )
@slow
class a ( unittest.TestCase ):
def A_ ( self : Union[str, Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] ):
return F"gaussian_noise_s={seed}_shape={'_'.join([str(lowercase_ ) for s in shape] )}.npy"
def A_ ( self : Any ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Dict , lowercase_ : List[Any]=0 , lowercase_ : Union[str, Any]=(4, 3, 512, 512) , lowercase_ : Optional[Any]=False ):
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = torch.from_numpy(load_hf_numpy(self.get_file_format(lowercase_ , lowercase_ ) ) ).to(lowercase_ ).to(lowercase_ )
return image
def A_ ( self : Any , lowercase_ : Dict="CompVis/stable-diffusion-v1-4" , lowercase_ : List[str]=False ):
snake_case_ = '''fp16''' if fpaa else None
snake_case_ = torch.floataa if fpaa else torch.floataa
snake_case_ = AutoencoderKL.from_pretrained(
lowercase_ , subfolder='''vae''' , torch_dtype=lowercase_ , revision=lowercase_ , )
model.to(lowercase_ ).eval()
return model
def A_ ( self : Any , lowercase_ : int=0 ):
if torch_device == "mps":
return torch.manual_seed(lowercase_ )
return torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
@parameterized.expand(
[
# fmt: off
[33, [-0.1603, 0.9878, -0.0495, -0.0790, -0.2709, 0.8375, -0.2060, -0.0824], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2376, 0.1168, 0.1332, -0.4840, -0.2508, -0.0791, -0.0493, -0.4089], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Union[str, Any] , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[33, [-0.0513, 0.0289, 1.3799, 0.2166, -0.2573, -0.0871, 0.5103, -0.0999]],
[47, [-0.4128, -0.1320, -0.3704, 0.1965, -0.4116, -0.2332, -0.3340, 0.2247]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Optional[int] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , fpaa=lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ , generator=lowercase_ , sample_posterior=lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.1609, 0.9866, -0.0487, -0.0777, -0.2716, 0.8368, -0.2055, -0.0814], [-0.2395, 0.0098, 0.0102, -0.0709, -0.2840, -0.0274, -0.0718, -0.1824]],
[47, [-0.2377, 0.1147, 0.1333, -0.4841, -0.2506, -0.0805, -0.0491, -0.4085], [0.0350, 0.0847, 0.0467, 0.0344, -0.0842, -0.0547, -0.0633, -0.1131]],
# fmt: on
] )
def A_ ( self : Tuple , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
with torch.no_grad():
snake_case_ = model(lowercase_ ).sample
assert sample.shape == image.shape
snake_case_ = sample[-1, -2:, -2:, :2].flatten().float().cpu()
snake_case_ = torch.tensor(expected_slice_mps if torch_device == '''mps''' else expected_slice )
assert torch_all_close(lowercase_ , lowercase_ , atol=3e-3 )
@parameterized.expand(
[
# fmt: off
[13, [-0.2051, -0.1803, -0.2311, -0.2114, -0.3292, -0.3574, -0.2953, -0.3323]],
[37, [-0.2632, -0.2625, -0.2199, -0.2741, -0.4539, -0.4990, -0.3720, -0.4925]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Dict , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-3 )
@parameterized.expand(
[
# fmt: off
[27, [-0.0369, 0.0207, -0.0776, -0.0682, -0.1747, -0.1930, -0.1465, -0.2039]],
[16, [-0.1628, -0.2134, -0.2747, -0.2642, -0.3774, -0.4404, -0.3687, -0.4277]],
# fmt: on
] )
@require_torch_gpu
def A_ ( self : Union[str, Any] , lowercase_ : Any , lowercase_ : Optional[Any] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
snake_case_ = sample[-1, -2:, :2, -2:].flatten().float().cpu()
snake_case_ = torch.tensor(lowercase_ )
assert torch_all_close(lowercase_ , lowercase_ , atol=5e-3 )
@parameterized.expand([(13,), (16,), (27,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : List[str] ):
snake_case_ = self.get_sd_vae_model(fpaa=lowercase_ )
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) , fpaa=lowercase_ )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-1 )
@parameterized.expand([(13,), (16,), (37,)] )
@require_torch_gpu
@unittest.skipIf(not is_xformers_available() , reason='''xformers is not required when using PyTorch 2.0.''' )
def A_ ( self : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ , shape=(3, 4, 64, 64) )
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
model.enable_xformers_memory_efficient_attention()
with torch.no_grad():
snake_case_ = model.decode(lowercase_ ).sample
assert list(sample.shape ) == [3, 3, 512, 512]
assert torch_all_close(lowercase_ , lowercase_ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[33, [-0.3001, 0.0918, -2.6984, -3.9720, -3.2099, -5.0353, 1.7338, -0.2065, 3.4267]],
[47, [-1.5030, -4.3871, -6.0355, -9.1157, -1.6661, -2.7853, 2.1607, -5.0823, 2.5633]],
# fmt: on
] )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : Tuple ):
snake_case_ = self.get_sd_vae_model()
snake_case_ = self.get_sd_image(lowercase_ )
snake_case_ = self.get_generator(lowercase_ )
with torch.no_grad():
snake_case_ = model.encode(lowercase_ ).latent_dist
snake_case_ = dist.sample(generator=lowercase_ )
assert list(sample.shape ) == [image.shape[0], 4] + [i // 8 for i in image.shape[2:]]
snake_case_ = sample[0, -1, -3:, -3:].flatten().cpu()
snake_case_ = torch.tensor(lowercase_ )
snake_case_ = 3e-3 if torch_device != '''mps''' else 1e-2
assert torch_all_close(lowercase_ , lowercase_ , atol=lowercase_ )
| 56
| 1
|
'''simple docstring'''
# this script reports modified .py files under the desired list of top-level sub-dirs passed as a list of arguments, e.g.:
# python ./utils/get_modified_files.py utils src tests examples
#
# it uses git to find the forking point and which files were modified - i.e. files not under git won't be considered
# since the output of this script is fed into Makefile commands it doesn't print a newline after the results
import re
import subprocess
import sys
a : Tuple = subprocess.check_output('git merge-base main HEAD'.split()).decode('utf-8')
a : Dict = (
subprocess.check_output(f'''git diff --diff-filter=d --name-only {fork_point_sha}'''.split()).decode('utf-8').split()
)
a : Optional[Any] = '|'.join(sys.argv[1:])
a : List[Any] = re.compile(rf'''^({joined_dirs}).*?\.py$''')
a : Tuple = [x for x in modified_files if regex.match(x)]
print(' '.join(relevant_modified_files), end='')
| 56
|
'''simple docstring'''
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 ( _lowerCamelCase ):
snake_case_ = 42
@flax_register_to_config
class a ( nn.Module , _lowerCamelCase , _lowerCamelCase ):
snake_case_ = 32
snake_case_ = 4
snake_case_ = 4
snake_case_ = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
snake_case_ = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
snake_case_ = False
snake_case_ = (320, 640, 1_280, 1_280)
snake_case_ = 2
snake_case_ = 8
snake_case_ = None
snake_case_ = 1_280
snake_case_ = 0.0
snake_case_ = False
snake_case_ = jnp.floataa
snake_case_ = True
snake_case_ = 0
snake_case_ = False
def A_ ( self : Optional[int] , lowercase_ : jax.random.KeyArray ):
# init input tensors
snake_case_ = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case_ = jnp.zeros(lowercase_ , dtype=jnp.floataa )
snake_case_ = jnp.ones((1,) , dtype=jnp.intaa )
snake_case_ = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case_ ,snake_case_ = jax.random.split(lowercase_ )
snake_case_ = {'''params''': params_rng, '''dropout''': dropout_rng}
return self.init(lowercase_ , lowercase_ , lowercase_ , lowercase_ )["params"]
def A_ ( self : List[str] ):
snake_case_ = self.block_out_channels
snake_case_ = 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.
snake_case_ = self.num_attention_heads or self.attention_head_dim
# input
snake_case_ = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case_ = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case_ = FlaxTimestepEmbedding(lowercase_ , dtype=self.dtype )
snake_case_ = self.only_cross_attention
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (only_cross_attention,) * len(self.down_block_types )
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case_ = []
snake_case_ = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case_ = output_channel
snake_case_ = block_out_channels[i]
snake_case_ = i == len(lowercase_ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = down_blocks
# mid
snake_case_ = 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
snake_case_ = []
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = list(reversed(lowercase_ ) )
snake_case_ = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case_ = output_channel
snake_case_ = reversed_block_out_channels[i]
snake_case_ = reversed_block_out_channels[min(i + 1 , len(lowercase_ ) - 1 )]
snake_case_ = i == len(lowercase_ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case_ = 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:
snake_case_ = 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_ )
snake_case_ = output_channel
snake_case_ = up_blocks
# out
snake_case_ = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : int , lowercase_ : Any , lowercase_ : int=None , lowercase_ : Any=None , lowercase_ : bool = True , lowercase_ : bool = False , ):
# 1. time
if not isinstance(lowercase_ , jnp.ndarray ):
snake_case_ = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(lowercase_ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case_ = timesteps.astype(dtype=jnp.floataa )
snake_case_ = jnp.expand_dims(lowercase_ , 0 )
snake_case_ = self.time_proj(lowercase_ )
snake_case_ = self.time_embedding(lowercase_ )
# 2. pre-process
snake_case_ = jnp.transpose(lowercase_ , (0, 2, 3, 1) )
snake_case_ = self.conv_in(lowercase_ )
# 3. down
snake_case_ = (sample,)
for down_block in self.down_blocks:
if isinstance(lowercase_ , lowercase_ ):
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , lowercase_ , deterministic=not train )
else:
snake_case_ ,snake_case_ = down_block(lowercase_ , lowercase_ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case_ = ()
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,)
snake_case_ = new_down_block_res_samples
# 4. mid
snake_case_ = 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:
snake_case_ = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case_ = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(lowercase_ , lowercase_ ):
snake_case_ = up_block(
lowercase_ , temb=lowercase_ , encoder_hidden_states=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train , )
else:
snake_case_ = up_block(lowercase_ , temb=lowercase_ , res_hidden_states_tuple=lowercase_ , deterministic=not train )
# 6. post-process
snake_case_ = self.conv_norm_out(lowercase_ )
snake_case_ = nn.silu(lowercase_ )
snake_case_ = self.conv_out(lowercase_ )
snake_case_ = jnp.transpose(lowercase_ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=lowercase_ )
| 56
| 1
|
'''simple docstring'''
def __magic_name__ ( __UpperCAmelCase ) -> bool:
'''simple docstring'''
if not isinstance(__UpperCAmelCase, __UpperCAmelCase ):
raise ValueError('''Input series is not valid, valid series - [2, 4, 6]''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Input list must be a non empty list''' )
if len(__UpperCAmelCase ) == 1:
return True
snake_case_ = series[1] - series[0]
for index in range(len(__UpperCAmelCase ) - 1 ):
if series[index + 1] - series[index] != common_diff:
return False
return True
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
if not isinstance(__UpperCAmelCase, __UpperCAmelCase ):
raise ValueError('''Input series is not valid, valid series - [2, 4, 6]''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Input list must be a non empty list''' )
snake_case_ = 0
for val in series:
answer += val
return answer / len(__UpperCAmelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 56
|
'''simple docstring'''
import glob
import os
import random
from string import ascii_lowercase, digits
import cva
import numpy as np
# Parrameters
a : Dict = (720, 1280) # Height, Width
a : Tuple = (0.4, 0.6) # if height or width lower than this scale, drop it.
a : Dict = 1 / 100
a : str = ''
a : Any = ''
a : Optional[int] = ''
a : List[str] = 250
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dataset(__UpperCAmelCase, __UpperCAmelCase )
for index in range(__UpperCAmelCase ):
snake_case_ = random.sample(range(len(__UpperCAmelCase ) ), 4 )
snake_case_ ,snake_case_ ,snake_case_ = update_image_and_anno(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, filter_scale=__UpperCAmelCase, )
# Get random string code: '7b7ad245cdff75241935e4dd860f3bad'
snake_case_ = random_chars(32 )
snake_case_ = path.split(os.sep )[-1].rsplit('''.''', 1 )[0]
snake_case_ = F"{OUTPUT_DIR}/{file_name}_MOSAIC_{letter_code}"
cva.imwrite(F"{file_root}.jpg", __UpperCAmelCase, [cva.IMWRITE_JPEG_QUALITY, 85] )
print(F"Succeeded {index+1}/{NUMBER_IMAGES} with {file_name}" )
snake_case_ = []
for anno in new_annos:
snake_case_ = anno[3] - anno[1]
snake_case_ = anno[4] - anno[2]
snake_case_ = anno[1] + width / 2
snake_case_ = anno[2] + height / 2
snake_case_ = F"{anno[0]} {x_center} {y_center} {width} {height}"
annos_list.append(__UpperCAmelCase )
with open(F"{file_root}.txt", '''w''' ) as outfile:
outfile.write('''\n'''.join(line for line in annos_list ) )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> tuple[list, list]:
'''simple docstring'''
snake_case_ = []
snake_case_ = []
for label_file in glob.glob(os.path.join(__UpperCAmelCase, '''*.txt''' ) ):
snake_case_ = label_file.split(os.sep )[-1].rsplit('''.''', 1 )[0]
with open(__UpperCAmelCase ) as in_file:
snake_case_ = in_file.readlines()
snake_case_ = os.path.join(__UpperCAmelCase, F"{label_name}.jpg" )
snake_case_ = []
for obj_list in obj_lists:
snake_case_ = obj_list.rstrip('''\n''' ).split(''' ''' )
snake_case_ = float(obj[1] ) - float(obj[3] ) / 2
snake_case_ = float(obj[2] ) - float(obj[4] ) / 2
snake_case_ = float(obj[1] ) + float(obj[3] ) / 2
snake_case_ = float(obj[2] ) + float(obj[4] ) / 2
boxes.append([int(obj[0] ), xmin, ymin, xmax, ymax] )
if not boxes:
continue
img_paths.append(__UpperCAmelCase )
labels.append(__UpperCAmelCase )
return img_paths, labels
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, ) -> tuple[list, list, str]:
'''simple docstring'''
snake_case_ = np.zeros([output_size[0], output_size[1], 3], dtype=np.uinta )
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = scale_range[0] + random.random() * (scale_range[1] - scale_range[0])
snake_case_ = int(scale_x * output_size[1] )
snake_case_ = int(scale_y * output_size[0] )
snake_case_ = []
snake_case_ = []
for i, index in enumerate(__UpperCAmelCase ):
snake_case_ = all_img_list[index]
path_list.append(__UpperCAmelCase )
snake_case_ = all_annos[index]
snake_case_ = cva.imread(__UpperCAmelCase )
if i == 0: # top-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = bbox[2] * scale_y
snake_case_ = bbox[3] * scale_x
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 1: # top-right
snake_case_ = cva.resize(__UpperCAmelCase, (output_size[1] - divid_point_x, divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = bbox[2] * scale_y
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = bbox[4] * scale_y
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
elif i == 2: # bottom-left
snake_case_ = cva.resize(__UpperCAmelCase, (divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = bbox[1] * scale_x
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = bbox[3] * scale_x
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
else: # bottom-right
snake_case_ = cva.resize(
__UpperCAmelCase, (output_size[1] - divid_point_x, output_size[0] - divid_point_y) )
snake_case_ = img
for bbox in img_annos:
snake_case_ = scale_x + bbox[1] * (1 - scale_x)
snake_case_ = scale_y + bbox[2] * (1 - scale_y)
snake_case_ = scale_x + bbox[3] * (1 - scale_x)
snake_case_ = scale_y + bbox[4] * (1 - scale_y)
new_anno.append([bbox[0], xmin, ymin, xmax, ymax] )
# Remove bounding box small than scale of filter
if filter_scale > 0:
snake_case_ = [
anno
for anno in new_anno
if filter_scale < (anno[3] - anno[1]) and filter_scale < (anno[4] - anno[2])
]
return output_img, new_anno, path_list[0]
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
assert number_char > 1, "The number of character should greater than 1"
snake_case_ = ascii_lowercase + digits
return "".join(random.choice(__UpperCAmelCase ) for _ in range(__UpperCAmelCase ) )
if __name__ == "__main__":
main()
print('DONE ✅')
| 56
| 1
|
'''simple docstring'''
import os
import sys
import tempfile
import torch
from .state import AcceleratorState
from .utils import PrecisionType, PrepareForLaunch, is_mps_available, patch_environment
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=(), __UpperCAmelCase=None, __UpperCAmelCase="no", __UpperCAmelCase="29500" ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = False
snake_case_ = False
if any(key.startswith('''KAGGLE''' ) for key in os.environ.keys() ):
snake_case_ = True
elif "IPython" in sys.modules:
snake_case_ = '''google.colab''' in str(sys.modules['''IPython'''].get_ipython() )
try:
snake_case_ = 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:
snake_case_ = 8
snake_case_ = 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 ):
snake_case_ = 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():
snake_case_ = '''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 __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=(), __UpperCAmelCase=2 ) -> 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''', ):
snake_case_ = PrepareForLaunch(__UpperCAmelCase, debug=__UpperCAmelCase )
start_processes(__UpperCAmelCase, args=__UpperCAmelCase, nprocs=__UpperCAmelCase, start_method='''fork''' )
| 56
|
'''simple docstring'''
import unittest
from transformers import (
MODEL_FOR_OBJECT_DETECTION_MAPPING,
AutoFeatureExtractor,
AutoModelForObjectDetection,
ObjectDetectionPipeline,
is_vision_available,
pipeline,
)
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_pytesseract,
require_tf,
require_timm,
require_torch,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_vision_available():
from PIL import Image
else:
class a :
@staticmethod
def A_ ( *lowercase_ : int , **lowercase_ : str ):
pass
@is_pipeline_test
@require_vision
@require_timm
@require_torch
class a ( unittest.TestCase ):
snake_case_ = MODEL_FOR_OBJECT_DETECTION_MAPPING
def A_ ( self : Any , lowercase_ : List[Any] , lowercase_ : Optional[int] , lowercase_ : List[str] ):
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , image_processor=lowercase_ )
return object_detector, ["./tests/fixtures/tests_samples/COCO/000000039769.png"]
def A_ ( self : Dict , lowercase_ : List[Any] , lowercase_ : int ):
snake_case_ = object_detector('''./tests/fixtures/tests_samples/COCO/000000039769.png''' , threshold=0.0 )
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
import datasets
snake_case_ = datasets.load_dataset('''hf-internal-testing/fixtures_image_utils''' , '''image''' , split='''test''' )
snake_case_ = [
Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' ),
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
# RGBA
dataset[0]['''file'''],
# LA
dataset[1]['''file'''],
# L
dataset[2]['''file'''],
]
snake_case_ = object_detector(lowercase_ , threshold=0.0 )
self.assertEqual(len(lowercase_ ) , len(lowercase_ ) )
for outputs in batch_outputs:
self.assertGreater(len(lowercase_ ) , 0 )
for detected_object in outputs:
self.assertEqual(
lowercase_ , {
'''score''': ANY(lowercase_ ),
'''label''': ANY(lowercase_ ),
'''box''': {'''xmin''': ANY(lowercase_ ), '''ymin''': ANY(lowercase_ ), '''xmax''': ANY(lowercase_ ), '''ymax''': ANY(lowercase_ )},
} , )
@require_tf
@unittest.skip('''Object detection not implemented in TF''' )
def A_ ( self : int ):
pass
@require_torch
def A_ ( self : Tuple ):
snake_case_ = '''hf-internal-testing/tiny-detr-mobilenetsv3'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=0.0 )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] , threshold=0.0 , )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
[
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
{'''score''': 0.3376, '''label''': '''LABEL_0''', '''box''': {'''xmin''': 159, '''ymin''': 120, '''xmax''': 480, '''ymax''': 359}},
],
] , )
@require_torch
@slow
def A_ ( self : Optional[int] ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = AutoModelForObjectDetection.from_pretrained(lowercase_ )
snake_case_ = AutoFeatureExtractor.from_pretrained(lowercase_ )
snake_case_ = ObjectDetectionPipeline(model=lowercase_ , feature_extractor=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : Tuple ):
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
snake_case_ = object_detector(
[
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
'''http://images.cocodataset.org/val2017/000000039769.jpg''',
] )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
[
{'''score''': 0.9982, '''label''': '''remote''', '''box''': {'''xmin''': 40, '''ymin''': 70, '''xmax''': 175, '''ymax''': 117}},
{'''score''': 0.9960, '''label''': '''remote''', '''box''': {'''xmin''': 333, '''ymin''': 72, '''xmax''': 368, '''ymax''': 187}},
{'''score''': 0.9955, '''label''': '''couch''', '''box''': {'''xmin''': 0, '''ymin''': 1, '''xmax''': 639, '''ymax''': 473}},
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
],
] , )
@require_torch
@slow
def A_ ( self : str ):
snake_case_ = 0.9985
snake_case_ = '''facebook/detr-resnet-50'''
snake_case_ = pipeline('''object-detection''' , model=lowercase_ )
snake_case_ = object_detector('''http://images.cocodataset.org/val2017/000000039769.jpg''' , threshold=lowercase_ )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9988, '''label''': '''cat''', '''box''': {'''xmin''': 13, '''ymin''': 52, '''xmax''': 314, '''ymax''': 470}},
{'''score''': 0.9987, '''label''': '''cat''', '''box''': {'''xmin''': 345, '''ymin''': 23, '''xmax''': 640, '''ymax''': 368}},
] , )
@require_torch
@require_pytesseract
@slow
def A_ ( self : Dict ):
snake_case_ = '''Narsil/layoutlmv3-finetuned-funsd'''
snake_case_ = 0.9993
snake_case_ = pipeline('''object-detection''' , model=lowercase_ , threshold=lowercase_ )
snake_case_ = object_detector(
'''https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png''' )
self.assertEqual(
nested_simplify(lowercase_ , decimals=4 ) , [
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
{'''score''': 0.9993, '''label''': '''I-ANSWER''', '''box''': {'''xmin''': 294, '''ymin''': 254, '''xmax''': 343, '''ymax''': 264}},
] , )
| 56
| 1
|
'''simple docstring'''
import os
import unittest
from transformers import BatchEncoding
from transformers.models.bert.tokenization_bert import (
BasicTokenizer,
WordpieceTokenizer,
_is_control,
_is_punctuation,
_is_whitespace,
)
from transformers.models.prophetnet.tokenization_prophetnet import VOCAB_FILES_NAMES, ProphetNetTokenizer
from transformers.testing_utils import require_torch, slow
from ...test_tokenization_common import TokenizerTesterMixin
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = ProphetNetTokenizer
snake_case_ = False
def A_ ( self : Optional[Any] ):
super().setUp()
snake_case_ = [
'''[UNK]''',
'''[CLS]''',
'''[SEP]''',
'''[PAD]''',
'''[MASK]''',
'''want''',
'''##want''',
'''##ed''',
'''wa''',
'''un''',
'''runn''',
'''##ing''',
''',''',
'''low''',
'''lowest''',
]
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
def A_ ( self : List[Any] , lowercase_ : Tuple ):
snake_case_ = '''UNwant\u00E9d,running'''
snake_case_ = '''unwanted, running'''
return input_text, output_text
def A_ ( self : Any ):
snake_case_ = self.tokenizer_class(self.vocab_file )
snake_case_ = tokenizer.tokenize('''UNwant\u00E9d,running''' )
self.assertListEqual(lowercase_ , ['''un''', '''##want''', '''##ed''', ''',''', '''runn''', '''##ing'''] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(lowercase_ ) , [9, 6, 7, 12, 10, 11] )
def A_ ( self : str ):
snake_case_ = BasicTokenizer()
self.assertListEqual(tokenizer.tokenize('''ah\u535A\u63A8zz''' ) , ['''ah''', '''\u535A''', '''\u63A8''', '''zz'''] )
def A_ ( self : str ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHeLLo!how \n Are yoU? ''' ) , ['''hello''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''hello'''] )
def A_ ( self : Dict ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''hällo''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''h\u00E9llo'''] )
def A_ ( self : Any ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''hallo''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''hello'''] )
def A_ ( self : str ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''hallo''', '''!''', '''how''', '''are''', '''you''', '''?'''] )
self.assertListEqual(tokenizer.tokenize('''H\u00E9llo''' ) , ['''hello'''] )
def A_ ( self : List[Any] ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHeLLo!how \n Are yoU? ''' ) , ['''HeLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?'''] )
def A_ ( self : Any ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''HäLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?'''] )
def A_ ( self : Union[str, Any] ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ , strip_accents=lowercase_ )
self.assertListEqual(
tokenizer.tokenize(''' \tHäLLo!how \n Are yoU? ''' ) , ['''HaLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?'''] )
def A_ ( self : Any ):
snake_case_ = BasicTokenizer(do_lower_case=lowercase_ , never_split=['''[UNK]'''] )
self.assertListEqual(
tokenizer.tokenize(''' \tHeLLo!how \n Are yoU? [UNK]''' ) , ['''HeLLo''', '''!''', '''how''', '''Are''', '''yoU''', '''?''', '''[UNK]'''] )
def A_ ( self : Dict ):
snake_case_ = ['''[UNK]''', '''[CLS]''', '''[SEP]''', '''want''', '''##want''', '''##ed''', '''wa''', '''un''', '''runn''', '''##ing''']
snake_case_ = {}
for i, token in enumerate(lowercase_ ):
snake_case_ = i
snake_case_ = WordpieceTokenizer(vocab=lowercase_ , unk_token='''[UNK]''' )
self.assertListEqual(tokenizer.tokenize('''''' ) , [] )
self.assertListEqual(tokenizer.tokenize('''unwanted running''' ) , ['''un''', '''##want''', '''##ed''', '''runn''', '''##ing'''] )
self.assertListEqual(tokenizer.tokenize('''unwantedX running''' ) , ['''[UNK]''', '''runn''', '''##ing'''] )
@require_torch
def A_ ( self : List[Any] ):
snake_case_ = self.tokenizer_class.from_pretrained('''microsoft/prophetnet-large-uncased''' )
snake_case_ = ['''A long paragraph for summarization.''', '''Another paragraph for summarization.''']
snake_case_ = [1037, 2146, 2_0423, 2005, 7680, 7849, 3989, 1012, 102]
snake_case_ = tokenizer(lowercase_ , padding=lowercase_ , return_tensors='''pt''' )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = list(batch.input_ids.numpy()[0] )
self.assertListEqual(lowercase_ , lowercase_ )
self.assertEqual((2, 9) , batch.input_ids.shape )
self.assertEqual((2, 9) , batch.attention_mask.shape )
def A_ ( self : Tuple ):
self.assertTrue(_is_whitespace(''' ''' ) )
self.assertTrue(_is_whitespace('''\t''' ) )
self.assertTrue(_is_whitespace('''\r''' ) )
self.assertTrue(_is_whitespace('''\n''' ) )
self.assertTrue(_is_whitespace('''\u00A0''' ) )
self.assertFalse(_is_whitespace('''A''' ) )
self.assertFalse(_is_whitespace('''-''' ) )
def A_ ( self : int ):
self.assertTrue(_is_control('''\u0005''' ) )
self.assertFalse(_is_control('''A''' ) )
self.assertFalse(_is_control(''' ''' ) )
self.assertFalse(_is_control('''\t''' ) )
self.assertFalse(_is_control('''\r''' ) )
def A_ ( self : Optional[int] ):
self.assertTrue(_is_punctuation('''-''' ) )
self.assertTrue(_is_punctuation('''$''' ) )
self.assertTrue(_is_punctuation('''`''' ) )
self.assertTrue(_is_punctuation('''.''' ) )
self.assertFalse(_is_punctuation('''A''' ) )
self.assertFalse(_is_punctuation(''' ''' ) )
@slow
def A_ ( self : Optional[Any] ):
snake_case_ = self.tokenizer_class.from_pretrained('''microsoft/prophetnet-large-uncased''' )
snake_case_ = tokenizer.encode('''sequence builders''' , add_special_tokens=lowercase_ )
snake_case_ = tokenizer.encode('''multi-sequence build''' , add_special_tokens=lowercase_ )
snake_case_ = tokenizer.build_inputs_with_special_tokens(lowercase_ )
snake_case_ = tokenizer.build_inputs_with_special_tokens(lowercase_ , lowercase_ )
assert encoded_sentence == text + [102]
assert encoded_pair == text + [102] + text_a + [102]
| 56
|
'''simple docstring'''
import unittest
from transformers import MPNetConfig, is_torch_available
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 (
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
class a :
def __init__( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Any=13 , lowercase_ : Optional[Any]=7 , lowercase_ : Optional[Any]=True , lowercase_ : Dict=True , lowercase_ : Tuple=False , lowercase_ : Optional[Any]=True , lowercase_ : Any=99 , lowercase_ : Union[str, Any]=64 , lowercase_ : str=5 , lowercase_ : int=4 , lowercase_ : List[Any]=64 , lowercase_ : Dict="gelu" , lowercase_ : Optional[int]=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Tuple=512 , lowercase_ : List[Any]=16 , lowercase_ : str=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[Any]=3 , lowercase_ : Optional[Any]=4 , lowercase_ : List[Any]=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_input_mask
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = scope
def A_ ( self : List[str] ):
return MPNetConfig.from_pretrained('''microsoft/mpnet-base''' )
def A_ ( self : str ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_input_mask:
snake_case_ = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = ids_tensor([self.batch_size] , self.num_choices )
snake_case_ = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def A_ ( self : Tuple ):
return MPNetConfig(
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 , initializer_range=self.initializer_range , )
def A_ ( self : Any , lowercase_ : Dict , lowercase_ : str , lowercase_ : Optional[Any] , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Optional[int] ):
snake_case_ = MPNetModel(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , lowercase_ )
snake_case_ = 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 : str , lowercase_ : Optional[Any] , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Optional[int] ):
snake_case_ = MPNetForQuestionAnswering(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Tuple , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : str , lowercase_ : Optional[int] , lowercase_ : Optional[Any] , lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = MPNetForSequenceClassification(lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = model(lowercase_ , attention_mask=lowercase_ , labels=lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : Any , lowercase_ : Any , lowercase_ : str , lowercase_ : Tuple , lowercase_ : List[str] , lowercase_ : Union[str, Any] , lowercase_ : Dict ):
snake_case_ = self.num_choices
snake_case_ = MPNetForMultipleChoice(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case_ = model(
lowercase_ , attention_mask=lowercase_ , labels=lowercase_ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def A_ ( self : Union[str, Any] , lowercase_ : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[int] , lowercase_ : str , lowercase_ : int , lowercase_ : List[str] ):
snake_case_ = self.num_labels
snake_case_ = MPNetForTokenClassification(config=lowercase_ )
model.to(lowercase_ )
model.eval()
snake_case_ = 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 : Union[str, Any] ):
snake_case_ = self.prepare_config_and_inputs()
((snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_) ,(snake_case_)) = config_and_inputs
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask}
return config, inputs_dict
@require_torch
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
MPNetForMaskedLM,
MPNetForMultipleChoice,
MPNetForQuestionAnswering,
MPNetForSequenceClassification,
MPNetForTokenClassification,
MPNetModel,
)
if is_torch_available()
else ()
)
snake_case_ = (
{
"feature-extraction": MPNetModel,
"fill-mask": MPNetForMaskedLM,
"question-answering": MPNetForQuestionAnswering,
"text-classification": MPNetForSequenceClassification,
"token-classification": MPNetForTokenClassification,
"zero-shot": MPNetForSequenceClassification,
}
if is_torch_available()
else {}
)
snake_case_ = False
snake_case_ = True
def A_ ( self : Tuple ):
snake_case_ = MPNetModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , hidden_size=37 )
def A_ ( self : Union[str, Any] ):
self.config_tester.run_common_tests()
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_model(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_sequence_classification(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_multiple_choice(*lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_token_classification(*lowercase_ )
def A_ ( self : Tuple ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mpnet_for_question_answering(*lowercase_ )
@require_torch
class a ( unittest.TestCase ):
@slow
def A_ ( self : List[Any] ):
snake_case_ = MPNetModel.from_pretrained('''microsoft/mpnet-base''' )
snake_case_ = torch.tensor([[0, 345, 232, 328, 740, 140, 1695, 69, 6078, 1588, 2]] )
snake_case_ = model(lowercase_ )[0]
snake_case_ = torch.Size((1, 11, 768) )
self.assertEqual(output.shape , lowercase_ )
snake_case_ = torch.tensor(
[[[-0.0550, 0.1943, -0.0740], [-0.0562, 0.2211, -0.0579], [-0.0437, 0.3337, -0.0641]]] )
# compare the actual values for a slice.
self.assertTrue(torch.allclose(output[:, :3, :3] , lowercase_ , atol=1e-4 ) )
| 56
| 1
|
'''simple docstring'''
import importlib.metadata
import warnings
from copy import deepcopy
from packaging import version
from ..utils import logging
from .import_utils import is_accelerate_available, is_bitsandbytes_available
if is_bitsandbytes_available():
import bitsandbytes as bnb
import torch
import torch.nn as nn
from ..pytorch_utils import ConvaD
if is_accelerate_available():
from accelerate import init_empty_weights
from accelerate.utils import find_tied_parameters
a : Optional[Any] = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase=None, __UpperCAmelCase=None ) -> Dict:
'''simple docstring'''
if "." in tensor_name:
snake_case_ = tensor_name.split('''.''' )
for split in splits[:-1]:
snake_case_ = getattr(__UpperCAmelCase, __UpperCAmelCase )
if new_module is None:
raise ValueError(F"{module} has no attribute {split}." )
snake_case_ = new_module
snake_case_ = splits[-1]
if tensor_name not in module._parameters and tensor_name not in module._buffers:
raise ValueError(F"{module} does not have a parameter or a buffer named {tensor_name}." )
snake_case_ = tensor_name in module._buffers
snake_case_ = getattr(__UpperCAmelCase, __UpperCAmelCase )
if old_value.device == torch.device('''meta''' ) and device not in ["meta", torch.device('''meta''' )] and value is None:
raise ValueError(F"{tensor_name} is on the meta device, we need a `value` to put in on {device}." )
snake_case_ = False
snake_case_ = False
if is_buffer or not is_bitsandbytes_available():
snake_case_ = False
snake_case_ = False
else:
snake_case_ = hasattr(bnb.nn, '''Params4bit''' ) and isinstance(module._parameters[tensor_name], bnb.nn.Paramsabit )
snake_case_ = isinstance(module._parameters[tensor_name], bnb.nn.IntaParams )
if is_abit or is_abit:
snake_case_ = module._parameters[tensor_name]
if param.device.type != "cuda":
if value is None:
snake_case_ = old_value.to(__UpperCAmelCase )
elif isinstance(__UpperCAmelCase, torch.Tensor ):
snake_case_ = value.to('''cpu''' )
if value.dtype == torch.inta:
snake_case_ = version.parse(importlib.metadata.version('''bitsandbytes''' ) ) > version.parse(
'''0.37.2''' )
if not is_abit_serializable:
raise ValueError(
'''Detected int8 weights but the version of bitsandbytes is not compatible with int8 serialization. '''
'''Make sure to download the latest `bitsandbytes` version. `pip install --upgrade bitsandbytes`.''' )
else:
snake_case_ = torch.tensor(__UpperCAmelCase, device='''cpu''' )
# Support models using `Conv1D` in place of `nn.Linear` (e.g. gpt2) by transposing the weight matrix prior to quantization.
# Since weights are saved in the correct "orientation", we skip transposing when loading.
if issubclass(module.source_cls, __UpperCAmelCase ) and fpaa_statistics is None:
snake_case_ = new_value.T
snake_case_ = old_value.__dict__
if is_abit:
snake_case_ = bnb.nn.IntaParams(__UpperCAmelCase, requires_grad=__UpperCAmelCase, **__UpperCAmelCase ).to(__UpperCAmelCase )
elif is_abit:
snake_case_ = bnb.nn.Paramsabit(__UpperCAmelCase, requires_grad=__UpperCAmelCase, **__UpperCAmelCase ).to(__UpperCAmelCase )
snake_case_ = new_value
if fpaa_statistics is not None:
setattr(module.weight, '''SCB''', fpaa_statistics.to(__UpperCAmelCase ) )
else:
if value is None:
snake_case_ = old_value.to(__UpperCAmelCase )
elif isinstance(__UpperCAmelCase, torch.Tensor ):
snake_case_ = value.to(__UpperCAmelCase )
else:
snake_case_ = torch.tensor(__UpperCAmelCase, device=__UpperCAmelCase )
if is_buffer:
snake_case_ = new_value
else:
snake_case_ = nn.Parameter(__UpperCAmelCase, requires_grad=old_value.requires_grad )
snake_case_ = new_value
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=None, __UpperCAmelCase=None, __UpperCAmelCase=None, __UpperCAmelCase=False ) -> List[str]:
'''simple docstring'''
for name, module in model.named_children():
if current_key_name is None:
snake_case_ = []
current_key_name.append(__UpperCAmelCase )
if (isinstance(__UpperCAmelCase, nn.Linear ) or isinstance(__UpperCAmelCase, __UpperCAmelCase )) and name not in modules_to_not_convert:
# Check if the current key is not in the `modules_to_not_convert`
if not any(key in '''.'''.join(__UpperCAmelCase ) for key in modules_to_not_convert ):
with init_empty_weights():
if isinstance(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ ,snake_case_ = module.weight.shape
else:
snake_case_ = module.in_features
snake_case_ = module.out_features
if quantization_config.quantization_method() == "llm_int8":
snake_case_ = bnb.nn.LinearabitLt(
__UpperCAmelCase, __UpperCAmelCase, module.bias is not None, has_fpaa_weights=quantization_config.llm_inta_has_fpaa_weight, threshold=quantization_config.llm_inta_threshold, )
snake_case_ = True
else:
if (
quantization_config.llm_inta_skip_modules is not None
and name in quantization_config.llm_inta_skip_modules
):
pass
else:
snake_case_ = bnb.nn.Linearabit(
__UpperCAmelCase, __UpperCAmelCase, module.bias is not None, quantization_config.bnb_abit_compute_dtype, compress_statistics=quantization_config.bnb_abit_use_double_quant, quant_type=quantization_config.bnb_abit_quant_type, )
snake_case_ = True
# Store the module class in case we need to transpose the weight later
snake_case_ = type(__UpperCAmelCase )
# Force requires grad to False to avoid unexpected errors
model._modules[name].requires_grad_(__UpperCAmelCase )
if len(list(module.children() ) ) > 0:
snake_case_ ,snake_case_ = _replace_with_bnb_linear(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, has_been_replaced=__UpperCAmelCase, )
# Remove the last key for recursion
current_key_name.pop(-1 )
return model, has_been_replaced
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase=None, __UpperCAmelCase=None, __UpperCAmelCase=None ) -> str:
'''simple docstring'''
snake_case_ = ['''lm_head'''] if modules_to_not_convert is None else modules_to_not_convert
snake_case_ ,snake_case_ = _replace_with_bnb_linear(
__UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase )
if not has_been_replaced:
logger.warning(
'''You are loading your model in 8bit or 4bit but no linear modules were found in your model.'''
''' Please double check your model architecture, or submit an issue on github if you think this is'''
''' a bug.''' )
return model
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
warnings.warn(
'''`replace_8bit_linear` will be deprecated in a future version, please use `replace_with_bnb_linear` instead''', __UpperCAmelCase, )
return replace_with_bnb_linear(*__UpperCAmelCase, **__UpperCAmelCase )
def __magic_name__ ( *__UpperCAmelCase, **__UpperCAmelCase ) -> Dict:
'''simple docstring'''
warnings.warn(
'''`set_module_8bit_tensor_to_device` will be deprecated in a future version, please use `set_module_quantized_tensor_to_device` instead''', __UpperCAmelCase, )
return set_module_quantized_tensor_to_device(*__UpperCAmelCase, **__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = deepcopy(__UpperCAmelCase ) # this has 0 cost since it is done inside `init_empty_weights` context manager`
tied_model.tie_weights()
snake_case_ = find_tied_parameters(__UpperCAmelCase )
# For compatibility with Accelerate < 0.18
if isinstance(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ = sum(list(tied_params.values() ), [] ) + list(tied_params.keys() )
else:
snake_case_ = sum(__UpperCAmelCase, [] )
snake_case_ = len(__UpperCAmelCase ) > 0
# Check if it is a base model
snake_case_ = not hasattr(__UpperCAmelCase, model.base_model_prefix )
# Ignore this for base models (BertModel, GPT2Model, etc.)
if (not has_tied_params) and is_base_model:
return []
# otherwise they have an attached head
snake_case_ = list(model.named_children() )
snake_case_ = [list_modules[-1][0]]
# add last module together with tied weights
snake_case_ = set(__UpperCAmelCase ) - set(__UpperCAmelCase )
snake_case_ = list(set(__UpperCAmelCase ) ) + list(__UpperCAmelCase )
# remove ".weight" from the keys
snake_case_ = ['''.weight''', '''.bias''']
snake_case_ = []
for name in list_untouched:
for name_to_remove in names_to_remove:
if name_to_remove in name:
snake_case_ = name.replace(__UpperCAmelCase, '''''' )
filtered_module_names.append(__UpperCAmelCase )
return filtered_module_names
| 56
|
'''simple docstring'''
import json
import os
import pickle
import shutil
import tempfile
from unittest import TestCase
from unittest.mock import patch
import numpy as np
from datasets import Dataset
from transformers import is_faiss_available
from transformers.models.bart.configuration_bart import BartConfig
from transformers.models.bart.tokenization_bart import BartTokenizer
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES as DPR_VOCAB_FILES_NAMES
from transformers.models.dpr.configuration_dpr import DPRConfig
from transformers.models.dpr.tokenization_dpr import DPRContextEncoderTokenizer, DPRQuestionEncoderTokenizer
from transformers.models.rag.configuration_rag import RagConfig
from transformers.models.rag.retrieval_rag import CustomHFIndex, RagRetriever
from transformers.models.roberta.tokenization_roberta import VOCAB_FILES_NAMES as BART_VOCAB_FILES_NAMES
from transformers.testing_utils import require_faiss, require_sentencepiece, require_tokenizers, require_torch
if is_faiss_available():
import faiss
@require_faiss
class a ( _lowerCamelCase ):
def A_ ( self : str ):
snake_case_ = tempfile.mkdtemp()
snake_case_ = 8
# DPR tok
snake_case_ = [
'''[UNK]''',
'''[CLS]''',
'''[SEP]''',
'''[PAD]''',
'''[MASK]''',
'''want''',
'''##want''',
'''##ed''',
'''wa''',
'''un''',
'''runn''',
'''##ing''',
''',''',
'''low''',
'''lowest''',
]
snake_case_ = os.path.join(self.tmpdirname , '''dpr_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , DPR_VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as vocab_writer:
vocab_writer.write(''''''.join([x + '''\n''' for x in vocab_tokens] ) )
# BART tok
snake_case_ = [
'''l''',
'''o''',
'''w''',
'''e''',
'''r''',
'''s''',
'''t''',
'''i''',
'''d''',
'''n''',
'''\u0120''',
'''\u0120l''',
'''\u0120n''',
'''\u0120lo''',
'''\u0120low''',
'''er''',
'''\u0120lowest''',
'''\u0120newer''',
'''\u0120wider''',
'''<unk>''',
]
snake_case_ = dict(zip(lowercase_ , range(len(lowercase_ ) ) ) )
snake_case_ = ['''#version: 0.2''', '''\u0120 l''', '''\u0120l o''', '''\u0120lo w''', '''e r''', '''''']
snake_case_ = {'''unk_token''': '''<unk>'''}
snake_case_ = os.path.join(self.tmpdirname , '''bart_tokenizer''' )
os.makedirs(lowercase_ , exist_ok=lowercase_ )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(lowercase_ , BART_VOCAB_FILES_NAMES['''merges_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write(json.dumps(lowercase_ ) + '''\n''' )
with open(self.merges_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write('''\n'''.join(lowercase_ ) )
def A_ ( self : Union[str, Any] ):
return DPRQuestionEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : Union[str, Any] ):
return DPRContextEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''dpr_tokenizer''' ) )
def A_ ( self : int ):
return BartTokenizer.from_pretrained(os.path.join(self.tmpdirname , '''bart_tokenizer''' ) )
def A_ ( self : str ):
shutil.rmtree(self.tmpdirname )
def A_ ( self : str ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size ), 2 * np.ones(self.retrieval_vector_size )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
return dataset
def A_ ( self : str ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , )
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
return retriever
def A_ ( self : str , lowercase_ : bool ):
snake_case_ = self.get_dummy_dataset()
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''custom''' , )
if from_disk:
snake_case_ = os.path.join(self.tmpdirname , '''dataset''' )
snake_case_ = os.path.join(self.tmpdirname , '''index.faiss''' )
dataset.get_index('''embeddings''' ).save(os.path.join(self.tmpdirname , '''index.faiss''' ) )
dataset.drop_index('''embeddings''' )
dataset.save_to_disk(os.path.join(self.tmpdirname , '''dataset''' ) )
del dataset
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , )
else:
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , index=CustomHFIndex(config.retrieval_vector_size , lowercase_ ) , )
return retriever
def A_ ( self : Tuple ):
snake_case_ = Dataset.from_dict(
{
'''id''': ['''0''', '''1'''],
'''text''': ['''foo''', '''bar'''],
'''title''': ['''Foo''', '''Bar'''],
'''embeddings''': [np.ones(self.retrieval_vector_size + 1 ), 2 * np.ones(self.retrieval_vector_size + 1 )],
} )
dataset.add_faiss_index('''embeddings''' , string_factory='''Flat''' , metric_type=faiss.METRIC_INNER_PRODUCT )
snake_case_ = os.path.join(self.tmpdirname , '''hf_bert_base.hnswSQ8_correct_phi_128.c_index''' )
dataset.save_faiss_index('''embeddings''' , index_file_name + '''.index.dpr''' )
pickle.dump(dataset['''id'''] , open(index_file_name + '''.index_meta.dpr''' , '''wb''' ) )
snake_case_ = os.path.join(self.tmpdirname , '''psgs_w100.tsv.pkl''' )
snake_case_ = {sample['''id''']: [sample['''text'''], sample['''title''']] for sample in dataset}
pickle.dump(lowercase_ , open(lowercase_ , '''wb''' ) )
snake_case_ = RagConfig(
retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='''legacy''' , index_path=self.tmpdirname , )
snake_case_ = RagRetriever(
lowercase_ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() )
return retriever
def A_ ( self : Optional[Any] ):
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : str ):
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
with patch('''transformers.models.rag.retrieval_rag.load_dataset''' ) as mock_load_dataset:
snake_case_ = self.get_dummy_dataset()
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : int ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : str ):
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''embeddings''', '''id''', '''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''id'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''id'''][0] , '''1''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''id'''][0] , '''0''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : Any ):
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
def A_ ( self : Any ):
snake_case_ = 1
snake_case_ = self.get_dummy_legacy_index_retriever()
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ ,snake_case_ ,snake_case_ = retriever.retrieve(lowercase_ , n_docs=lowercase_ )
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertEqual(len(lowercase_ ) , 2 )
self.assertEqual(sorted(doc_dicts[0] ) , ['''text''', '''title'''] )
self.assertEqual(len(doc_dicts[0]['''text'''] ) , lowercase_ )
self.assertEqual(doc_dicts[0]['''text'''][0] , '''bar''' ) # max inner product is reached with second doc
self.assertEqual(doc_dicts[1]['''text'''][0] , '''foo''' ) # max inner product is reached with first doc
self.assertListEqual(doc_ids.tolist() , [[1], [0]] )
def A_ ( self : int ):
snake_case_ = self.get_dummy_legacy_index_retriever()
with tempfile.TemporaryDirectory() as tmp_dirname:
retriever.save_pretrained(lowercase_ )
snake_case_ = RagRetriever.from_pretrained(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever.retrieve(lowercase_ , n_docs=1 )
self.assertTrue(out is not None )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : List[str] ):
import torch
snake_case_ = 1
snake_case_ = self.get_dummy_canonical_hf_index_retriever()
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
snake_case_ ,snake_case_ ,snake_case_ = (
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertIsInstance(lowercase_ , np.ndarray )
snake_case_ = retriever(
lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ , return_tensors='''pt''' , )
snake_case_ ,snake_case_ ,snake_case_ ,snake_case_ = ( # noqa: F841
out['''context_input_ids'''],
out['''context_attention_mask'''],
out['''retrieved_doc_embeds'''],
out['''doc_ids'''],
)
self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
self.assertIsInstance(lowercase_ , torch.Tensor )
@require_torch
@require_tokenizers
@require_sentencepiece
def A_ ( self : Tuple ):
snake_case_ = self.get_dpr_ctx_encoder_tokenizer()
snake_case_ = 1
snake_case_ = self.get_dummy_custom_hf_index_retriever(from_disk=lowercase_ )
retriever.set_ctx_encoder_tokenizer(lowercase_ )
snake_case_ = [[5, 7], [10, 11]]
snake_case_ = np.array(
[np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa )
snake_case_ = retriever(lowercase_ , lowercase_ , prefix=retriever.config.generator.prefix , n_docs=lowercase_ )
self.assertEqual(
len(lowercase_ ) , 6 ) # check whether the retriever output consist of 6 attributes including tokenized docs
self.assertEqual(
all(k in out for k in ('''tokenized_doc_ids''', '''tokenized_doc_attention_mask''') ) , lowercase_ ) # check for doc token related keys in dictionary.
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
import unittest
from transformers import MobileBertConfig, is_tf_available
from transformers.models.auto import get_values
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import (
TF_MODEL_FOR_PRETRAINING_MAPPING,
TFMobileBertForMaskedLM,
TFMobileBertForMultipleChoice,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertModel,
)
@require_tf
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (
(
TFMobileBertModel,
TFMobileBertForMaskedLM,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertForMultipleChoice,
)
if is_tf_available()
else ()
)
snake_case_ = (
{
"feature-extraction": TFMobileBertModel,
"fill-mask": TFMobileBertForMaskedLM,
"question-answering": TFMobileBertForQuestionAnswering,
"text-classification": TFMobileBertForSequenceClassification,
"token-classification": TFMobileBertForTokenClassification,
"zero-shot": TFMobileBertForSequenceClassification,
}
if is_tf_available()
else {}
)
snake_case_ = False
snake_case_ = False
def A_ ( self : Dict , lowercase_ : Tuple , lowercase_ : Tuple , lowercase_ : Union[str, Any]=False ):
snake_case_ = super()._prepare_for_class(lowercase_ , lowercase_ , return_labels=lowercase_ )
if return_labels:
if model_class in get_values(lowercase_ ):
snake_case_ = tf.zeros(self.model_tester.batch_size , dtype=tf.intaa )
return inputs_dict
class a ( _lowerCamelCase ):
def __init__( self : Union[str, Any] , lowercase_ : str , lowercase_ : int=13 , lowercase_ : Union[str, Any]=7 , lowercase_ : Union[str, Any]=True , lowercase_ : int=True , lowercase_ : int=True , lowercase_ : List[Any]=True , lowercase_ : List[Any]=99 , lowercase_ : Tuple=32 , lowercase_ : Any=32 , lowercase_ : Tuple=2 , lowercase_ : List[Any]=4 , lowercase_ : List[str]=37 , lowercase_ : List[Any]="gelu" , lowercase_ : List[str]=0.1 , lowercase_ : Optional[Any]=0.1 , lowercase_ : Optional[Any]=512 , lowercase_ : Union[str, Any]=16 , lowercase_ : Optional[int]=2 , lowercase_ : Tuple=0.02 , lowercase_ : Optional[int]=3 , lowercase_ : Any=4 , lowercase_ : Dict=None , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = seq_length
snake_case_ = is_training
snake_case_ = use_input_mask
snake_case_ = use_token_type_ids
snake_case_ = use_labels
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = type_vocab_size
snake_case_ = type_sequence_label_size
snake_case_ = initializer_range
snake_case_ = num_labels
snake_case_ = num_choices
snake_case_ = scope
snake_case_ = embedding_size
def A_ ( self : Union[str, Any] ):
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ = None
if self.use_input_mask:
snake_case_ = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ = None
if self.use_token_type_ids:
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ = None
snake_case_ = None
snake_case_ = None
if self.use_labels:
snake_case_ = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case_ = ids_tensor([self.batch_size] , self.num_choices )
snake_case_ = MobileBertConfig(
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 , embedding_size=self.embedding_size , )
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : str , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : List[Any] , lowercase_ : Optional[int] ):
snake_case_ = TFMobileBertModel(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(lowercase_ )
snake_case_ = [input_ids, input_mask]
snake_case_ = model(lowercase_ )
snake_case_ = 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 : List[str] , lowercase_ : Union[str, Any] , lowercase_ : str , lowercase_ : Dict , lowercase_ : str , lowercase_ : int , lowercase_ : Optional[Any] , lowercase_ : Optional[int] ):
snake_case_ = TFMobileBertForMaskedLM(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def A_ ( self : Union[str, Any] , lowercase_ : Optional[int] , lowercase_ : int , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Any , lowercase_ : Optional[Any] , lowercase_ : List[str] ):
snake_case_ = TFMobileBertForNextSentencePrediction(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, 2) )
def A_ ( self : Dict , lowercase_ : int , lowercase_ : Optional[int] , lowercase_ : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : Any ):
snake_case_ = TFMobileBertForPreTraining(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(
result.prediction_logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
self.parent.assertEqual(result.seq_relationship_logits.shape , (self.batch_size, 2) )
def A_ ( self : Union[str, Any] , lowercase_ : Union[str, Any] , lowercase_ : str , lowercase_ : Optional[int] , lowercase_ : Dict , lowercase_ : Any , lowercase_ : Union[str, Any] , lowercase_ : Tuple ):
snake_case_ = self.num_labels
snake_case_ = TFMobileBertForSequenceClassification(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : List[Any] , lowercase_ : Optional[Any] , lowercase_ : Dict , lowercase_ : int , lowercase_ : Optional[int] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Any ):
snake_case_ = self.num_choices
snake_case_ = TFMobileBertForMultipleChoice(config=lowercase_ )
snake_case_ = tf.tile(tf.expand_dims(lowercase_ , 1 ) , (1, self.num_choices, 1) )
snake_case_ = tf.tile(tf.expand_dims(lowercase_ , 1 ) , (1, self.num_choices, 1) )
snake_case_ = tf.tile(tf.expand_dims(lowercase_ , 1 ) , (1, self.num_choices, 1) )
snake_case_ = {
'''input_ids''': multiple_choice_inputs_ids,
'''attention_mask''': multiple_choice_input_mask,
'''token_type_ids''': multiple_choice_token_type_ids,
}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : Dict , lowercase_ : Optional[Any] , lowercase_ : int , lowercase_ : Optional[Any] , lowercase_ : Union[str, Any] , lowercase_ : Any ):
snake_case_ = self.num_labels
snake_case_ = TFMobileBertForTokenClassification(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def A_ ( self : List[Any] , lowercase_ : Optional[int] , lowercase_ : int , lowercase_ : Any , lowercase_ : Any , lowercase_ : Optional[int] , lowercase_ : List[Any] , lowercase_ : Tuple ):
snake_case_ = TFMobileBertForQuestionAnswering(config=lowercase_ )
snake_case_ = {'''input_ids''': input_ids, '''attention_mask''': input_mask, '''token_type_ids''': token_type_ids}
snake_case_ = model(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 : int ):
snake_case_ = self.prepare_config_and_inputs()
(
(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,(
snake_case_
) ,
) = config_and_inputs
snake_case_ = {'''input_ids''': input_ids, '''token_type_ids''': token_type_ids, '''attention_mask''': input_mask}
return config, inputs_dict
def A_ ( self : str ):
snake_case_ = TFMobileBertModelTest.TFMobileBertModelTester(self )
snake_case_ = ConfigTester(self , config_class=lowercase_ , hidden_size=37 )
def A_ ( self : int ):
self.config_tester.run_common_tests()
def A_ ( self : str ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_model(*lowercase_ )
def A_ ( self : Dict ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_masked_lm(*lowercase_ )
def A_ ( self : Any ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_multiple_choice(*lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_next_sequence_prediction(*lowercase_ )
def A_ ( self : Any ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_pretraining(*lowercase_ )
def A_ ( self : List[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_question_answering(*lowercase_ )
def A_ ( self : Optional[Any] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_sequence_classification(*lowercase_ )
def A_ ( self : List[str] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_mobilebert_for_token_classification(*lowercase_ )
@slow
def A_ ( self : List[str] ):
# for model_name in TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["google/mobilebert-uncased"]:
snake_case_ = TFMobileBertModel.from_pretrained(lowercase_ )
self.assertIsNotNone(lowercase_ )
@require_tf
class a ( unittest.TestCase ):
@slow
def A_ ( self : Optional[Any] ):
snake_case_ = TFMobileBertForPreTraining.from_pretrained('''google/mobilebert-uncased''' )
snake_case_ = tf.constant([[0, 1, 2, 3, 4, 5]] )
snake_case_ = model(lowercase_ )[0]
snake_case_ = [1, 6, 3_0522]
self.assertEqual(output.shape , lowercase_ )
snake_case_ = tf.constant(
[
[
[-4.591_9547, -9.24_8295, -9.64_5256],
[-6.730_6175, -6.44_0284, -6.605_2837],
[-7.274_3506, -6.784_7915, -6.02_4673],
]
] )
tf.debugging.assert_near(output[:, :3, :3] , lowercase_ , atol=1e-4 )
| 56
|
'''simple docstring'''
import os
import re
import warnings
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_ta import TaTokenizer
else:
a : Dict = None
a : List[Any] = logging.get_logger(__name__)
a : List[Any] = {'vocab_file': 'spiece.model', 'tokenizer_file': 'tokenizer.json'}
a : str = {
'vocab_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/spiece.model',
't5-base': 'https://huggingface.co/t5-base/resolve/main/spiece.model',
't5-large': 'https://huggingface.co/t5-large/resolve/main/spiece.model',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/spiece.model',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/spiece.model',
},
'tokenizer_file': {
't5-small': 'https://huggingface.co/t5-small/resolve/main/tokenizer.json',
't5-base': 'https://huggingface.co/t5-base/resolve/main/tokenizer.json',
't5-large': 'https://huggingface.co/t5-large/resolve/main/tokenizer.json',
't5-3b': 'https://huggingface.co/t5-3b/resolve/main/tokenizer.json',
't5-11b': 'https://huggingface.co/t5-11b/resolve/main/tokenizer.json',
},
}
# TODO(PVP) - this should be removed in Transformers v5
a : List[Any] = {
't5-small': 512,
't5-base': 512,
't5-large': 512,
't5-3b': 512,
't5-11b': 512,
}
class a ( _lowerCamelCase ):
snake_case_ = VOCAB_FILES_NAMES
snake_case_ = PRETRAINED_VOCAB_FILES_MAP
snake_case_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
snake_case_ = ["input_ids", "attention_mask"]
snake_case_ = TaTokenizer
snake_case_ = []
def __init__( self : List[Any] , lowercase_ : int=None , lowercase_ : Dict=None , lowercase_ : Dict="</s>" , lowercase_ : List[Any]="<unk>" , lowercase_ : int="<pad>" , lowercase_ : int=100 , lowercase_ : List[Any]=None , **lowercase_ : List[str] , ):
# Add extra_ids to the special token list
if extra_ids > 0 and additional_special_tokens is None:
snake_case_ = [F"<extra_id_{i}>" for i in range(lowercase_ )]
elif extra_ids > 0 and additional_special_tokens is not None:
# Check that we have the right number of extra special tokens
snake_case_ = len(set(filter(lambda lowercase_ : bool('''extra_id_''' in str(lowercase_ ) ) , lowercase_ ) ) )
if extra_tokens != extra_ids:
raise ValueError(
F"Both extra_ids ({extra_ids}) and additional_special_tokens ({additional_special_tokens}) are"
''' provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids'''
''' tokens''' )
super().__init__(
lowercase_ , tokenizer_file=lowercase_ , eos_token=lowercase_ , unk_token=lowercase_ , pad_token=lowercase_ , extra_ids=lowercase_ , additional_special_tokens=lowercase_ , **lowercase_ , )
snake_case_ = vocab_file
snake_case_ = False if not self.vocab_file else True
snake_case_ = extra_ids
@staticmethod
def A_ ( lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : int ):
if pretrained_model_name_or_path in TaTokenizerFast.max_model_input_sizes:
snake_case_ = TaTokenizerFast.max_model_input_sizes[pretrained_model_name_or_path]
if init_max_model_length is not None and init_max_model_length != max_model_length:
return init_max_model_length
elif init_max_model_length is None:
warnings.warn(
'''This tokenizer was incorrectly instantiated with a model max length of'''
F" {deprecated_max_model_length} which will be corrected in Transformers v5.\nFor now, this"
''' behavior is kept to avoid breaking backwards compatibility when padding/encoding with'''
''' `truncation is True`.\n- Be aware that you SHOULD NOT rely on'''
F" {pretrained_model_name_or_path} automatically truncating your input to"
F" {deprecated_max_model_length} when padding/encoding.\n- If you want to encode/pad to sequences"
F" longer than {deprecated_max_model_length} you can either instantiate this tokenizer with"
''' `model_max_length` or pass `max_length` when encoding/padding.\n- To avoid this warning, please'''
''' instantiate this tokenizer with `model_max_length` set to your preferred value.''' , lowercase_ , )
return max_model_length
def A_ ( self : Optional[Any] , lowercase_ : str , lowercase_ : Optional[str] = None ):
if not self.can_save_slow_tokenizer:
raise ValueError(
'''Your fast tokenizer does not have the necessary information to save the vocabulary for a slow '''
'''tokenizer.''' )
if not os.path.isdir(lowercase_ ):
logger.error(F"Vocabulary path ({save_directory}) should be a directory" )
return
snake_case_ = 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_ ):
copyfile(self.vocab_file , lowercase_ )
logger.info(F"Copy vocab file to {out_vocab_file}" )
return (out_vocab_file,)
def A_ ( self : Optional[int] , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = token_ids_a + [self.eos_token_id]
if token_ids_a is None:
return self.prefix_tokens + token_ids_a
else:
snake_case_ = token_ids_a + [self.eos_token_id]
return self.prefix_tokens + token_ids_a + token_ids_a
def A_ ( self : int , lowercase_ : List[int] , lowercase_ : Optional[List[int]] = None ):
snake_case_ = [self.eos_token_id]
if token_ids_a is None:
return len(token_ids_a + eos ) * [0]
return len(token_ids_a + eos + token_ids_a + eos ) * [0]
def A_ ( self : Dict ):
return list(
set(filter(lambda lowercase_ : bool(re.search(R'''<extra_id_\d+>''' , lowercase_ ) ) is not None , self.additional_special_tokens ) ) )
def A_ ( self : Any ):
return [self.convert_tokens_to_ids(lowercase_ ) for token in self.get_sentinel_tokens()]
| 56
| 1
|
'''simple docstring'''
import copy
from dataclasses import dataclass, field
from typing import ClassVar, Dict
from ..features import ClassLabel, Features, Image
from .base import TaskTemplate
@dataclass(frozen=_lowerCamelCase )
class a ( _lowerCamelCase ):
snake_case_ = field(default="image-classification" , metadata={"include_in_asdict_even_if_is_default": True} )
snake_case_ = Features({"image": Image()} )
snake_case_ = Features({"labels": ClassLabel} )
snake_case_ = "image"
snake_case_ = "labels"
def A_ ( self : int , lowercase_ : Optional[int] ):
if self.label_column not in features:
raise ValueError(F"Column {self.label_column} is not present in features." )
if not isinstance(features[self.label_column] , lowercase_ ):
raise ValueError(F"Column {self.label_column} is not a ClassLabel." )
snake_case_ = copy.deepcopy(self )
snake_case_ = self.label_schema.copy()
snake_case_ = features[self.label_column]
snake_case_ = label_schema
return task_template
@property
def A_ ( self : Tuple ):
return {
self.image_column: "image",
self.label_column: "labels",
}
| 56
|
'''simple docstring'''
from __future__ import annotations
import math
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
if depth < 0:
raise ValueError('''Depth cannot be less than 0''' )
if len(__UpperCAmelCase ) == 0:
raise ValueError('''Scores cannot be empty''' )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
return min(
minimax(depth + 1, node_index * 2, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), minimax(depth + 1, node_index * 2 + 1, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ), )
def __magic_name__ ( ) -> None:
'''simple docstring'''
snake_case_ = [90, 23, 6, 33, 21, 65, 123, 3_4423]
snake_case_ = math.log(len(__UpperCAmelCase ), 2 )
print('''Optimal value : ''', end='''''' )
print(minimax(0, 0, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 56
| 1
|
'''simple docstring'''
import dataclasses
import re
import string
from typing import Any, Dict, Iterator, List, Mapping, Optional, Sequence, Tuple
import numpy as np
from . import residue_constants
a : Any = Mapping[str, np.ndarray]
a : Tuple = Mapping[str, Any] # Is a nested dict.
a : Union[str, Any] = 0.01
@dataclasses.dataclass(frozen=_lowerCamelCase )
class a :
snake_case_ = 42 # [num_res, num_atom_type, 3]
# Amino-acid type for each residue represented as an integer between 0 and
# 20, where 20 is 'X'.
snake_case_ = 42 # [num_res]
# Binary float mask to indicate presence of a particular atom. 1.0 if an atom
# is present and 0.0 if not. This should be used for loss masking.
snake_case_ = 42 # [num_res, num_atom_type]
# Residue index as used in PDB. It is not necessarily continuous or 0-indexed.
snake_case_ = 42 # [num_res]
# B-factors, or temperature factors, of each residue (in sq. angstroms units),
# representing the displacement of the residue from its ground truth mean
# value.
snake_case_ = 42 # [num_res, num_atom_type]
# Chain indices for multi-chain predictions
snake_case_ = None
# Optional remark about the protein. Included as a comment in output PDB
# files
snake_case_ = None
# Templates used to generate this protein (prediction-only)
snake_case_ = None
# Chain corresponding to each parent
snake_case_ = None
def __magic_name__ ( __UpperCAmelCase ) -> Protein:
'''simple docstring'''
snake_case_ = r'''(\[[A-Z]+\]\n)'''
snake_case_ = [tag.strip() for tag in re.split(__UpperCAmelCase, __UpperCAmelCase ) if len(__UpperCAmelCase ) > 0]
snake_case_ = zip(tags[0::2], [l.split('''\n''' ) for l in tags[1::2]] )
snake_case_ = ["N", "CA", "C"]
snake_case_ = None
snake_case_ = None
snake_case_ = None
for g in groups:
if "[PRIMARY]" == g[0]:
snake_case_ = g[1][0].strip()
for i in range(len(__UpperCAmelCase ) ):
if seq[i] not in residue_constants.restypes:
snake_case_ = '''X''' # FIXME: strings are immutable
snake_case_ = np.array(
[residue_constants.restype_order.get(__UpperCAmelCase, residue_constants.restype_num ) for res_symbol in seq] )
elif "[TERTIARY]" == g[0]:
snake_case_ = []
for axis in range(3 ):
tertiary.append(list(map(__UpperCAmelCase, g[1][axis].split() ) ) )
snake_case_ = np.array(__UpperCAmelCase )
snake_case_ = np.zeros((len(tertiary[0] ) // 3, residue_constants.atom_type_num, 3) ).astype(np.floataa )
for i, atom in enumerate(__UpperCAmelCase ):
snake_case_ = np.transpose(tertiary_np[:, i::3] )
atom_positions *= PICO_TO_ANGSTROM
elif "[MASK]" == g[0]:
snake_case_ = np.array(list(map({'''-''': 0, '''+''': 1}.get, g[1][0].strip() ) ) )
snake_case_ = np.zeros(
(
len(__UpperCAmelCase ),
residue_constants.atom_type_num,
) ).astype(np.floataa )
for i, atom in enumerate(__UpperCAmelCase ):
snake_case_ = 1
atom_mask *= mask[..., None]
assert aatype is not None
return Protein(
atom_positions=__UpperCAmelCase, atom_mask=__UpperCAmelCase, aatype=__UpperCAmelCase, residue_index=np.arange(len(__UpperCAmelCase ) ), b_factors=__UpperCAmelCase, )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase = 0 ) -> List[str]:
'''simple docstring'''
snake_case_ = []
snake_case_ = prot.remark
if remark is not None:
pdb_headers.append(F"REMARK {remark}" )
snake_case_ = prot.parents
snake_case_ = prot.parents_chain_index
if parents is not None and parents_chain_index is not None:
snake_case_ = [p for i, p in zip(__UpperCAmelCase, __UpperCAmelCase ) if i == chain_id]
if parents is None or len(__UpperCAmelCase ) == 0:
snake_case_ = ['''N/A''']
pdb_headers.append(F"PARENT {' '.join(__UpperCAmelCase )}" )
return pdb_headers
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = []
snake_case_ = pdb_str.split('''\n''' )
snake_case_ = prot.remark
if remark is not None:
out_pdb_lines.append(F"REMARK {remark}" )
snake_case_ = 42
if prot.parents is not None and len(prot.parents ) > 0:
snake_case_ = []
if prot.parents_chain_index is not None:
snake_case_ = {}
for p, i in zip(prot.parents, prot.parents_chain_index ):
parent_dict.setdefault(str(__UpperCAmelCase ), [] )
parent_dict[str(__UpperCAmelCase )].append(__UpperCAmelCase )
snake_case_ = max([int(__UpperCAmelCase ) for chain_idx in parent_dict] )
for i in range(max_idx + 1 ):
snake_case_ = parent_dict.get(str(__UpperCAmelCase ), ['''N/A'''] )
parents_per_chain.append(__UpperCAmelCase )
else:
parents_per_chain.append(list(prot.parents ) )
else:
snake_case_ = [['''N/A''']]
def make_parent_line(__UpperCAmelCase ) -> str:
return F"PARENT {' '.join(__UpperCAmelCase )}"
out_pdb_lines.append(make_parent_line(parents_per_chain[0] ) )
snake_case_ = 0
for i, l in enumerate(__UpperCAmelCase ):
if "PARENT" not in l and "REMARK" not in l:
out_pdb_lines.append(__UpperCAmelCase )
if "TER" in l and "END" not in lines[i + 1]:
chain_counter += 1
if not chain_counter >= len(__UpperCAmelCase ):
snake_case_ = parents_per_chain[chain_counter]
else:
snake_case_ = ['''N/A''']
out_pdb_lines.append(make_parent_line(__UpperCAmelCase ) )
return "\n".join(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = residue_constants.restypes + ['''X''']
def res_atoa(__UpperCAmelCase ) -> str:
return residue_constants.restype_atoa.get(restypes[r], '''UNK''' )
snake_case_ = residue_constants.atom_types
snake_case_ = []
snake_case_ = prot.atom_mask
snake_case_ = prot.aatype
snake_case_ = prot.atom_positions
snake_case_ = prot.residue_index.astype(np.intaa )
snake_case_ = prot.b_factors
snake_case_ = prot.chain_index
if np.any(aatype > residue_constants.restype_num ):
raise ValueError('''Invalid aatypes.''' )
snake_case_ = get_pdb_headers(__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0:
pdb_lines.extend(__UpperCAmelCase )
snake_case_ = aatype.shape[0]
snake_case_ = 1
snake_case_ = 0
snake_case_ = string.ascii_uppercase
snake_case_ = None
# Add all atom sites.
for i in range(__UpperCAmelCase ):
snake_case_ = res_atoa(aatype[i] )
for atom_name, pos, mask, b_factor in zip(__UpperCAmelCase, atom_positions[i], atom_mask[i], b_factors[i] ):
if mask < 0.5:
continue
snake_case_ = '''ATOM'''
snake_case_ = atom_name if len(__UpperCAmelCase ) == 4 else F" {atom_name}"
snake_case_ = ''''''
snake_case_ = ''''''
snake_case_ = 1.0_0
snake_case_ = atom_name[0] # Protein supports only C, N, O, S, this works.
snake_case_ = ''''''
snake_case_ = '''A'''
if chain_index is not None:
snake_case_ = chain_tags[chain_index[i]]
# PDB is a columnar format, every space matters here!
snake_case_ = (
F"{record_type:<6}{atom_index:>5} {name:<4}{alt_loc:>1}"
F"{res_name_a:>3} {chain_tag:>1}"
F"{residue_index[i]:>4}{insertion_code:>1} "
F"{pos[0]:>8.3f}{pos[1]:>8.3f}{pos[2]:>8.3f}"
F"{occupancy:>6.2f}{b_factor:>6.2f} "
F"{element:>2}{charge:>2}"
)
pdb_lines.append(__UpperCAmelCase )
atom_index += 1
snake_case_ = i == n - 1
if chain_index is not None:
if i != n - 1 and chain_index[i + 1] != prev_chain_index:
snake_case_ = True
snake_case_ = chain_index[i + 1]
if should_terminate:
# Close the chain.
snake_case_ = '''TER'''
snake_case_ = (
F"{chain_end:<6}{atom_index:>5} {res_atoa(aatype[i] ):>3} {chain_tag:>1}{residue_index[i]:>4}"
)
pdb_lines.append(__UpperCAmelCase )
atom_index += 1
if i != n - 1:
# "prev" is a misnomer here. This happens at the beginning of
# each new chain.
pdb_lines.extend(get_pdb_headers(__UpperCAmelCase, __UpperCAmelCase ) )
pdb_lines.append('''END''' )
pdb_lines.append('''''' )
return "\n".join(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> np.ndarray:
'''simple docstring'''
return residue_constants.STANDARD_ATOM_MASK[prot.aatype]
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = None, ) -> Protein:
'''simple docstring'''
return Protein(
aatype=features['''aatype'''], atom_positions=result['''final_atom_positions'''], atom_mask=result['''final_atom_mask'''], residue_index=features['''residue_index'''] + 1, b_factors=b_factors if b_factors is not None else np.zeros_like(result['''final_atom_mask'''] ), chain_index=__UpperCAmelCase, remark=__UpperCAmelCase, parents=__UpperCAmelCase, parents_chain_index=__UpperCAmelCase, )
| 56
|
'''simple docstring'''
import argparse
import torch
from torch import nn
from transformers import SpeechaTextConfig, SpeechaTextForConditionalGeneration
def __magic_name__ ( __UpperCAmelCase ) -> str:
'''simple docstring'''
snake_case_ = [
'''encoder.version''',
'''decoder.version''',
'''model.encoder.version''',
'''model.decoder.version''',
'''decoder.output_projection.weight''',
'''_float_tensor''',
'''encoder.embed_positions._float_tensor''',
'''decoder.embed_positions._float_tensor''',
]
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = list(s_dict.keys() )
for key in keys:
if "transformer_layers" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
elif "subsample" in key:
snake_case_ = s_dict.pop(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ ,snake_case_ = emb.weight.shape
snake_case_ = nn.Linear(__UpperCAmelCase, __UpperCAmelCase, bias=__UpperCAmelCase )
snake_case_ = emb.weight.data
return lin_layer
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = torch.load(__UpperCAmelCase, map_location='''cpu''' )
snake_case_ = mam_aaa['''args''']
snake_case_ = mam_aaa['''model''']
snake_case_ = state_dict['''decoder.output_projection.weight''']
remove_ignore_keys_(__UpperCAmelCase )
rename_keys(__UpperCAmelCase )
snake_case_ = state_dict['''decoder.embed_tokens.weight'''].shape[0]
snake_case_ = args.share_decoder_input_output_embed
snake_case_ = [int(__UpperCAmelCase ) for i in args.conv_kernel_sizes.split(''',''' )]
snake_case_ = SpeechaTextConfig(
vocab_size=__UpperCAmelCase, max_source_positions=args.max_source_positions, max_target_positions=args.max_target_positions, encoder_layers=args.encoder_layers, decoder_layers=args.decoder_layers, encoder_attention_heads=args.encoder_attention_heads, decoder_attention_heads=args.decoder_attention_heads, encoder_ffn_dim=args.encoder_ffn_embed_dim, decoder_ffn_dim=args.decoder_ffn_embed_dim, d_model=args.encoder_embed_dim, dropout=args.dropout, attention_dropout=args.attention_dropout, activation_dropout=args.activation_dropout, activation_function='''relu''', num_conv_layers=len(__UpperCAmelCase ), conv_channels=args.conv_channels, conv_kernel_sizes=__UpperCAmelCase, input_feat_per_channel=args.input_feat_per_channel, input_channels=args.input_channels, tie_word_embeddings=__UpperCAmelCase, num_beams=5, max_length=200, use_cache=__UpperCAmelCase, decoder_start_token_id=2, early_stopping=__UpperCAmelCase, )
snake_case_ = SpeechaTextForConditionalGeneration(__UpperCAmelCase )
snake_case_ ,snake_case_ = model.model.load_state_dict(__UpperCAmelCase, strict=__UpperCAmelCase )
if len(__UpperCAmelCase ) > 0 and not set(__UpperCAmelCase ) <= {
"encoder.embed_positions.weights",
"decoder.embed_positions.weights",
}:
raise ValueError(
'''Only `encoder.embed_positions.weights` and `decoder.embed_positions.weights` are allowed to be missing,'''
F" but all the following weights are missing {missing}" )
if tie_embeds:
snake_case_ = make_linear_from_emb(model.model.decoder.embed_tokens )
else:
snake_case_ = lm_head_weights
model.save_pretrained(__UpperCAmelCase )
if __name__ == "__main__":
a : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument('--fairseq_path', type=str, help='Path to the fairseq model (.pt) file.')
parser.add_argument('--pytorch_dump_folder_path', default=None, type=str, help='Path to the output PyTorch model.')
a : List[Any] = parser.parse_args()
convert_fairseq_sat_checkpoint_to_tfms(args.fairseq_path, args.pytorch_dump_folder_path)
| 56
| 1
|
'''simple docstring'''
import json
import os
import unittest
from transformers import MgpstrTokenizer
from transformers.models.mgp_str.tokenization_mgp_str import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class a ( _lowerCamelCase , unittest.TestCase ):
snake_case_ = MgpstrTokenizer
snake_case_ = False
snake_case_ = {}
snake_case_ = False
def A_ ( self : Dict ):
super().setUp()
# fmt: off
snake_case_ = ['''[GO]''', '''[s]''', '''0''', '''1''', '''2''', '''3''', '''4''', '''5''', '''6''', '''7''', '''8''', '''9''', '''a''', '''b''', '''c''', '''d''', '''e''', '''f''', '''g''', '''h''', '''i''', '''j''', '''k''', '''l''', '''m''', '''n''', '''o''', '''p''', '''q''', '''r''', '''s''', '''t''', '''u''', '''v''', '''w''', '''x''', '''y''', '''z''']
# fmt: on
snake_case_ = dict(zip(lowercase_ , range(len(lowercase_ ) ) ) )
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write(json.dumps(lowercase_ ) + '''\n''' )
def A_ ( self : List[str] , **lowercase_ : Optional[Any] ):
return MgpstrTokenizer.from_pretrained(self.tmpdirname , **lowercase_ )
def A_ ( self : int , lowercase_ : Any ):
snake_case_ = '''tester'''
snake_case_ = '''tester'''
return input_text, output_text
@unittest.skip('''MGP-STR always lower cases letters.''' )
def A_ ( self : Union[str, Any] ):
pass
def A_ ( self : int ):
snake_case_ = self.get_tokenizers(do_lower_case=lowercase_ )
for tokenizer in tokenizers:
with self.subTest(F"{tokenizer.__class__.__name__}" ):
snake_case_ = '''[SPECIAL_TOKEN]'''
tokenizer.add_special_tokens({'''cls_token''': special_token} )
snake_case_ = tokenizer.encode([special_token] , add_special_tokens=lowercase_ )
self.assertEqual(len(lowercase_ ) , 1 )
snake_case_ = tokenizer.decode(lowercase_ , skip_special_tokens=lowercase_ )
self.assertTrue(special_token not in decoded )
def A_ ( self : Any ):
snake_case_ = self.get_tokenizers()
for tokenizer in tokenizers:
with self.subTest(F"{tokenizer.__class__.__name__}" ):
snake_case_ ,snake_case_ = self.get_input_output_texts(lowercase_ )
snake_case_ = tokenizer.tokenize(lowercase_ )
snake_case_ = tokenizer.convert_tokens_to_ids(lowercase_ )
snake_case_ = tokenizer.encode(lowercase_ , add_special_tokens=lowercase_ )
self.assertListEqual(lowercase_ , lowercase_ )
snake_case_ = tokenizer.convert_ids_to_tokens(lowercase_ )
self.assertNotEqual(len(lowercase_ ) , 0 )
snake_case_ = tokenizer.decode(lowercase_ )
self.assertIsInstance(lowercase_ , lowercase_ )
self.assertEqual(text_a.replace(''' ''' , '''''' ) , lowercase_ )
@unittest.skip('''MGP-STR tokenizer only handles one sequence.''' )
def A_ ( self : Tuple ):
pass
@unittest.skip('''inputs cannot be pretokenized in MgpstrTokenizer''' )
def A_ ( self : Tuple ):
pass
| 56
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class a ( metaclass=_lowerCamelCase ):
snake_case_ = ["transformers", "torch", "note_seq"]
def __init__( self : Union[str, Any] , *lowercase_ : Optional[int] , **lowercase_ : int ):
requires_backends(self , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : List[Any] , *lowercase_ : Optional[Any] , **lowercase_ : str ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
@classmethod
def A_ ( cls : Tuple , *lowercase_ : Union[str, Any] , **lowercase_ : List[Any] ):
requires_backends(cls , ['''transformers''', '''torch''', '''note_seq'''] )
| 56
| 1
|
'''simple docstring'''
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers import CLIPTokenizer, CLIPTokenizerFast
from transformers.models.clip.tokenization_clip import VOCAB_FILES_NAMES
from transformers.testing_utils import require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import OwlViTImageProcessor, OwlViTProcessor
@require_vision
class a ( unittest.TestCase ):
def A_ ( self : Tuple ):
snake_case_ = tempfile.mkdtemp()
# fmt: off
snake_case_ = ['''''', '''l''', '''o''', '''w''', '''e''', '''r''', '''s''', '''t''', '''i''', '''d''', '''n''', '''lo''', '''l</w>''', '''w</w>''', '''r</w>''', '''t</w>''', '''low</w>''', '''er</w>''', '''lowest</w>''', '''newer</w>''', '''wider''', '''<unk>''', '''<|startoftext|>''', '''<|endoftext|>''']
# fmt: on
snake_case_ = dict(zip(lowercase_ , range(len(lowercase_ ) ) ) )
snake_case_ = ['''#version: 0.2''', '''l o''', '''lo w</w>''', '''e r</w>''', '''''']
snake_case_ = {'''unk_token''': '''<unk>'''}
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''vocab_file'''] )
snake_case_ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''merges_file'''] )
with open(self.vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write(json.dumps(lowercase_ ) + '''\n''' )
with open(self.merges_file , '''w''' , encoding='''utf-8''' ) as fp:
fp.write('''\n'''.join(lowercase_ ) )
snake_case_ = {
'''do_resize''': True,
'''size''': 20,
'''do_center_crop''': True,
'''crop_size''': 18,
'''do_normalize''': True,
'''image_mean''': [0.4814_5466, 0.457_8275, 0.4082_1073],
'''image_std''': [0.2686_2954, 0.2613_0258, 0.2757_7711],
}
snake_case_ = os.path.join(self.tmpdirname , lowercase_ )
with open(self.image_processor_file , '''w''' , encoding='''utf-8''' ) as fp:
json.dump(lowercase_ , lowercase_ )
def A_ ( self : Tuple , **lowercase_ : Tuple ):
return CLIPTokenizer.from_pretrained(self.tmpdirname , pad_token='''!''' , **lowercase_ )
def A_ ( self : int , **lowercase_ : int ):
return CLIPTokenizerFast.from_pretrained(self.tmpdirname , pad_token='''!''' , **lowercase_ )
def A_ ( self : Dict , **lowercase_ : Optional[Any] ):
return OwlViTImageProcessor.from_pretrained(self.tmpdirname , **lowercase_ )
def A_ ( self : Dict ):
shutil.rmtree(self.tmpdirname )
def A_ ( self : Tuple ):
snake_case_ = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta )]
snake_case_ = [Image.fromarray(np.moveaxis(lowercase_ , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def A_ ( self : Dict ):
snake_case_ = self.get_tokenizer()
snake_case_ = self.get_rust_tokenizer()
snake_case_ = self.get_image_processor()
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
processor_slow.save_pretrained(self.tmpdirname )
snake_case_ = OwlViTProcessor.from_pretrained(self.tmpdirname , use_fast=lowercase_ )
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
processor_fast.save_pretrained(self.tmpdirname )
snake_case_ = OwlViTProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor_slow.tokenizer.get_vocab() , tokenizer_slow.get_vocab() )
self.assertEqual(processor_fast.tokenizer.get_vocab() , tokenizer_fast.get_vocab() )
self.assertEqual(tokenizer_slow.get_vocab() , tokenizer_fast.get_vocab() )
self.assertIsInstance(processor_slow.tokenizer , lowercase_ )
self.assertIsInstance(processor_fast.tokenizer , lowercase_ )
self.assertEqual(processor_slow.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertEqual(processor_fast.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor_slow.image_processor , lowercase_ )
self.assertIsInstance(processor_fast.image_processor , lowercase_ )
def A_ ( self : Optional[Any] ):
snake_case_ = OwlViTProcessor(tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
snake_case_ = self.get_tokenizer(bos_token='''(BOS)''' , eos_token='''(EOS)''' )
snake_case_ = self.get_image_processor(do_normalize=lowercase_ )
snake_case_ = OwlViTProcessor.from_pretrained(
self.tmpdirname , bos_token='''(BOS)''' , eos_token='''(EOS)''' , do_normalize=lowercase_ )
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_ )
def A_ ( self : str ):
snake_case_ = self.get_image_processor()
snake_case_ = self.get_tokenizer()
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
snake_case_ = self.prepare_image_inputs()
snake_case_ = image_processor(lowercase_ , return_tensors='''np''' )
snake_case_ = processor(images=lowercase_ , return_tensors='''np''' )
for key in input_image_proc.keys():
self.assertAlmostEqual(input_image_proc[key].sum() , input_processor[key].sum() , delta=1e-2 )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.get_image_processor()
snake_case_ = self.get_tokenizer()
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
snake_case_ = '''lower newer'''
snake_case_ = processor(text=lowercase_ , return_tensors='''np''' )
snake_case_ = tokenizer(lowercase_ , return_tensors='''np''' )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key][0].tolist() , encoded_processor[key][0].tolist() )
def A_ ( self : Any ):
snake_case_ = self.get_image_processor()
snake_case_ = self.get_tokenizer()
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
snake_case_ = '''lower newer'''
snake_case_ = self.prepare_image_inputs()
snake_case_ = processor(text=lowercase_ , images=lowercase_ )
self.assertListEqual(list(inputs.keys() ) , ['''input_ids''', '''attention_mask''', '''pixel_values'''] )
# test if it raises when no input is passed
with pytest.raises(lowercase_ ):
processor()
def A_ ( self : int ):
snake_case_ = '''google/owlvit-base-patch32'''
snake_case_ = OwlViTProcessor.from_pretrained(lowercase_ )
snake_case_ = ['''cat''', '''nasa badge''']
snake_case_ = processor(text=lowercase_ )
snake_case_ = 16
self.assertListEqual(list(inputs.keys() ) , ['''input_ids''', '''attention_mask'''] )
self.assertEqual(inputs['''input_ids'''].shape , (2, seq_length) )
# test if it raises when no input is passed
with pytest.raises(lowercase_ ):
processor()
def A_ ( self : str ):
snake_case_ = '''google/owlvit-base-patch32'''
snake_case_ = OwlViTProcessor.from_pretrained(lowercase_ )
snake_case_ = [['''cat''', '''nasa badge'''], ['''person''']]
snake_case_ = processor(text=lowercase_ )
snake_case_ = 16
snake_case_ = len(lowercase_ )
snake_case_ = max([len(lowercase_ ) for texts in input_texts] )
self.assertListEqual(list(inputs.keys() ) , ['''input_ids''', '''attention_mask'''] )
self.assertEqual(inputs['''input_ids'''].shape , (batch_size * num_max_text_queries, seq_length) )
# test if it raises when no input is passed
with pytest.raises(lowercase_ ):
processor()
def A_ ( self : Union[str, Any] ):
snake_case_ = '''google/owlvit-base-patch32'''
snake_case_ = OwlViTProcessor.from_pretrained(lowercase_ )
snake_case_ = ['''cat''', '''nasa badge''']
snake_case_ = processor(text=lowercase_ )
snake_case_ = 16
snake_case_ = inputs['''input_ids''']
snake_case_ = [
[4_9406, 2368, 4_9407, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[4_9406, 6841, 1_1301, 4_9407, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
]
self.assertListEqual(list(inputs.keys() ) , ['''input_ids''', '''attention_mask'''] )
self.assertEqual(inputs['''input_ids'''].shape , (2, seq_length) )
self.assertListEqual(list(input_ids[0] ) , predicted_ids[0] )
self.assertListEqual(list(input_ids[1] ) , predicted_ids[1] )
def A_ ( self : List[str] ):
snake_case_ = self.get_image_processor()
snake_case_ = self.get_tokenizer()
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
snake_case_ = self.prepare_image_inputs()
snake_case_ = self.prepare_image_inputs()
snake_case_ = processor(images=lowercase_ , query_images=lowercase_ )
self.assertListEqual(list(inputs.keys() ) , ['''query_pixel_values''', '''pixel_values'''] )
# test if it raises when no input is passed
with pytest.raises(lowercase_ ):
processor()
def A_ ( self : Any ):
snake_case_ = self.get_image_processor()
snake_case_ = self.get_tokenizer()
snake_case_ = OwlViTProcessor(tokenizer=lowercase_ , image_processor=lowercase_ )
snake_case_ = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
snake_case_ = processor.batch_decode(lowercase_ )
snake_case_ = tokenizer.batch_decode(lowercase_ )
self.assertListEqual(lowercase_ , lowercase_ )
| 56
|
'''simple docstring'''
# tests directory-specific settings - this file is run automatically
# by pytest before any tests are run
import doctest
import sys
import warnings
from os.path import abspath, dirname, join
import _pytest
from transformers.testing_utils import HfDoctestModule, HfDocTestParser
# allow having multiple repository checkouts and not needing to remember to rerun
# 'pip install -e .[dev]' when switching between checkouts and running tests.
a : int = abspath(join(dirname(__file__), 'src'))
sys.path.insert(1, git_repo_path)
# silence FutureWarning warnings in tests since often we can't act on them until
# they become normal warnings - i.e. the tests still need to test the current functionality
warnings.simplefilter(action='ignore', category=FutureWarning)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
config.addinivalue_line(
'''markers''', '''is_pt_tf_cross_test: mark test to run only when PT and TF interactions are tested''' )
config.addinivalue_line(
'''markers''', '''is_pt_flax_cross_test: mark test to run only when PT and FLAX interactions are tested''' )
config.addinivalue_line('''markers''', '''is_pipeline_test: mark test to run only when pipelines are tested''' )
config.addinivalue_line('''markers''', '''is_staging_test: mark test to run only in the staging environment''' )
config.addinivalue_line('''markers''', '''accelerate_tests: mark test that require accelerate''' )
config.addinivalue_line('''markers''', '''tool_tests: mark the tool tests that are run on their specific schedule''' )
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
from transformers.testing_utils import pytest_addoption_shared
pytest_addoption_shared(__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
from transformers.testing_utils import pytest_terminal_summary_main
snake_case_ = terminalreporter.config.getoption('''--make-reports''' )
if make_reports:
pytest_terminal_summary_main(__UpperCAmelCase, id=__UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
if exitstatus == 5:
snake_case_ = 0
# Doctest custom flag to ignore output.
a : Union[str, Any] = doctest.register_optionflag('IGNORE_RESULT')
a : Optional[int] = doctest.OutputChecker
class a ( _lowerCamelCase ):
def A_ ( self : List[Any] , lowercase_ : int , lowercase_ : Tuple , lowercase_ : Optional[int] ):
if IGNORE_RESULT & optionflags:
return True
return OutputChecker.check_output(self , lowercase_ , lowercase_ , lowercase_ )
a : List[Any] = CustomOutputChecker
a : Optional[int] = HfDoctestModule
a : Tuple = HfDocTestParser
| 56
| 1
|
'''simple docstring'''
import unicodedata
from dataclasses import dataclass
from typing import Optional, Union
import numpy as np
from transformers.data.data_collator import DataCollatorMixin
from transformers.file_utils import PaddingStrategy
from transformers.tokenization_utils_base import PreTrainedTokenizerBase
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if isinstance(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ = np.full((len(__UpperCAmelCase ), sequence_length, 2), __UpperCAmelCase )
else:
snake_case_ = np.full((len(__UpperCAmelCase ), sequence_length), __UpperCAmelCase )
for i, tensor in enumerate(__UpperCAmelCase ):
if padding_side == "right":
if isinstance(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ = tensor[:sequence_length]
else:
snake_case_ = tensor[:sequence_length]
else:
if isinstance(__UpperCAmelCase, __UpperCAmelCase ):
snake_case_ = tensor[:sequence_length]
else:
snake_case_ = tensor[:sequence_length]
return out_tensor.tolist()
def __magic_name__ ( __UpperCAmelCase ) -> Dict:
'''simple docstring'''
snake_case_ = ord(__UpperCAmelCase )
if (cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126):
return True
snake_case_ = unicodedata.category(__UpperCAmelCase )
if cat.startswith('''P''' ):
return True
return False
@dataclass
class a ( _lowerCamelCase ):
snake_case_ = 42
snake_case_ = True
snake_case_ = None
snake_case_ = None
snake_case_ = -100
snake_case_ = "pt"
def A_ ( self : int , lowercase_ : List[str] ):
import torch
snake_case_ = '''label''' if '''label''' in features[0].keys() else '''labels'''
snake_case_ = [feature[label_name] for feature in features] if label_name in features[0].keys() else None
snake_case_ = self.tokenizer.pad(
lowercase_ , padding=self.padding , max_length=self.max_length , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors='''pt''' if labels is None else None , )
if labels is None:
return batch
snake_case_ = torch.tensor(batch['''entity_ids'''] ).shape[1]
snake_case_ = self.tokenizer.padding_side
if padding_side == "right":
snake_case_ = [
list(lowercase_ ) + [self.label_pad_token_id] * (sequence_length - len(lowercase_ )) for label in labels
]
else:
snake_case_ = [
[self.label_pad_token_id] * (sequence_length - len(lowercase_ )) + list(lowercase_ ) for label in labels
]
snake_case_ = [feature['''ner_tags'''] for feature in features]
snake_case_ = padding_tensor(lowercase_ , -1 , lowercase_ , lowercase_ )
snake_case_ = [feature['''original_entity_spans'''] for feature in features]
snake_case_ = padding_tensor(lowercase_ , (-1, -1) , lowercase_ , lowercase_ )
snake_case_ = {k: torch.tensor(lowercase_ , dtype=torch.intaa ) for k, v in batch.items()}
return batch
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import TensorType, is_torch_available, logging
a : Dict = logging.get_logger(__name__)
a : List[str] = {
'Helsinki-NLP/opus-mt-en-de': 'https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/config.json',
# See all Marian models at https://huggingface.co/models?filter=marian
}
class a ( _lowerCamelCase ):
snake_case_ = "marian"
snake_case_ = ["past_key_values"]
snake_case_ = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self : List[Any] , lowercase_ : Optional[Any]=5_8101 , lowercase_ : Dict=None , lowercase_ : List[str]=1024 , lowercase_ : Optional[Any]=12 , lowercase_ : int=4096 , lowercase_ : Any=16 , lowercase_ : Optional[int]=12 , lowercase_ : str=4096 , lowercase_ : Union[str, Any]=16 , lowercase_ : Dict=0.0 , lowercase_ : Union[str, Any]=0.0 , lowercase_ : Optional[Any]=True , lowercase_ : Union[str, Any]=True , lowercase_ : int="gelu" , lowercase_ : Dict=1024 , lowercase_ : int=0.1 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.0 , lowercase_ : Tuple=0.02 , lowercase_ : int=5_8100 , lowercase_ : Optional[Any]=False , lowercase_ : Any=5_8100 , lowercase_ : Optional[int]=0 , lowercase_ : Dict=0 , lowercase_ : List[str]=True , **lowercase_ : Any , ):
snake_case_ = vocab_size
snake_case_ = decoder_vocab_size or vocab_size
snake_case_ = max_position_embeddings
snake_case_ = d_model
snake_case_ = encoder_ffn_dim
snake_case_ = encoder_layers
snake_case_ = encoder_attention_heads
snake_case_ = decoder_ffn_dim
snake_case_ = decoder_layers
snake_case_ = decoder_attention_heads
snake_case_ = dropout
snake_case_ = attention_dropout
snake_case_ = activation_dropout
snake_case_ = activation_function
snake_case_ = init_std
snake_case_ = encoder_layerdrop
snake_case_ = decoder_layerdrop
snake_case_ = use_cache
snake_case_ = encoder_layers
snake_case_ = scale_embedding # scale factor will be sqrt(d_model) if True
snake_case_ = share_encoder_decoder_embeddings
super().__init__(
pad_token_id=lowercase_ , eos_token_id=lowercase_ , is_encoder_decoder=lowercase_ , decoder_start_token_id=lowercase_ , forced_eos_token_id=lowercase_ , **lowercase_ , )
class a ( _lowerCamelCase ):
@property
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.inputs
def A_ ( self : Union[str, Any] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ = {0: '''batch'''}
snake_case_ = {0: '''batch''', 1: '''past_decoder_sequence + sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''decoder_sequence'''}
snake_case_ = {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.
snake_case_ = OrderedDict(
[
('''input_ids''', {0: '''batch''', 1: '''encoder_sequence'''}),
('''attention_mask''', {0: '''batch''', 1: '''encoder_sequence'''}),
] )
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
else:
snake_case_ = 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
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.outputs
def A_ ( self : Dict ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super().outputs
else:
snake_case_ = super(lowercase_ , self ).outputs
if self.use_past:
snake_case_ ,snake_case_ = self.num_layers
for i in range(lowercase_ ):
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
snake_case_ = {0: '''batch''', 2: '''past_sequence + sequence'''}
return common_outputs
def A_ ( self : Dict , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
# Generate decoder inputs
snake_case_ = seq_length if not self.use_past else 1
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ )
snake_case_ = {F"decoder_{name}": tensor for name, tensor in decoder_inputs.items()}
snake_case_ = 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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
snake_case_ = common_inputs['''decoder_input_ids'''].shape[1]
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = decoder_seq_length + 3
snake_case_ = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
snake_case_ = torch.cat(
[common_inputs['''decoder_attention_mask'''], torch.ones(lowercase_ , lowercase_ )] , dim=1 )
snake_case_ = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
snake_case_ ,snake_case_ = self.num_layers
snake_case_ = min(lowercase_ , lowercase_ )
snake_case_ = max(lowercase_ , lowercase_ ) - min_num_layers
snake_case_ = '''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.
snake_case_ = 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 : Union[str, Any] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
snake_case_ = self._generate_dummy_inputs_for_encoder_and_decoder(
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
snake_case_ ,snake_case_ = common_inputs['''input_ids'''].shape
# Not using the same length for past_key_values
snake_case_ = seqlen + 2
snake_case_ ,snake_case_ = self.num_layers
snake_case_ ,snake_case_ = self.num_attention_heads
snake_case_ = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case_ = common_inputs['''attention_mask'''].dtype
snake_case_ = torch.cat(
[common_inputs['''attention_mask'''], torch.ones(lowercase_ , lowercase_ , dtype=lowercase_ )] , dim=1 )
snake_case_ = [
(torch.zeros(lowercase_ ), torch.zeros(lowercase_ )) for _ in range(lowercase_ )
]
return common_inputs
def A_ ( self : List[str] , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
# 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
snake_case_ = 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
snake_case_ = tokenizer.num_special_tokens_to_add(lowercase_ )
snake_case_ = 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
snake_case_ = [''' '''.join([tokenizer.unk_token] ) * seq_length] * batch_size
snake_case_ = dict(tokenizer(lowercase_ , return_tensors=lowercase_ ) )
return common_inputs
def A_ ( self : Any , lowercase_ : PreTrainedTokenizer , lowercase_ : int = -1 , lowercase_ : int = -1 , lowercase_ : bool = False , lowercase_ : Optional[TensorType] = None , ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
else:
snake_case_ = self._generate_dummy_inputs_for_causal_lm(
lowercase_ , batch_size=lowercase_ , seq_length=lowercase_ , is_pair=lowercase_ , framework=lowercase_ )
return common_inputs
def A_ ( self : Dict , lowercase_ : List[str] , lowercase_ : List[str] , lowercase_ : int , lowercase_ : List[str] ):
if self.task in ["default", "seq2seq-lm"]:
snake_case_ = super()._flatten_past_key_values_(lowercase_ , lowercase_ , lowercase_ , lowercase_ )
else:
snake_case_ = super(lowercase_ , self )._flatten_past_key_values_(
lowercase_ , lowercase_ , lowercase_ , lowercase_ )
@property
def A_ ( self : List[str] ):
return 1e-4
| 56
| 1
|
'''simple docstring'''
import re
from typing import Callable, List, Optional, Union
import tensorflow as tf
try:
from tensorflow.keras.optimizers.legacy import Adam
except ImportError:
from tensorflow.keras.optimizers import Adam
class a ( tf.keras.optimizers.schedules.LearningRateSchedule ):
def __init__( self : Optional[Any] , lowercase_ : float , lowercase_ : Callable , lowercase_ : int , lowercase_ : float = 1.0 , lowercase_ : str = None , ):
super().__init__()
snake_case_ = initial_learning_rate
snake_case_ = warmup_steps
snake_case_ = power
snake_case_ = decay_schedule_fn
snake_case_ = name
def __call__( self : Tuple , lowercase_ : str ):
with tf.name_scope(self.name or '''WarmUp''' ) as name:
# Implements polynomial warmup. i.e., if global_step < warmup_steps, the
# learning rate will be `global_step/num_warmup_steps * init_lr`.
snake_case_ = tf.cast(lowercase_ , tf.floataa )
snake_case_ = tf.cast(self.warmup_steps , tf.floataa )
snake_case_ = global_step_float / warmup_steps_float
snake_case_ = self.initial_learning_rate * tf.math.pow(lowercase_ , self.power )
return tf.cond(
global_step_float < warmup_steps_float , lambda: warmup_learning_rate , lambda: self.decay_schedule_fn(step - self.warmup_steps ) , name=lowercase_ , )
def A_ ( self : Any ):
return {
"initial_learning_rate": self.initial_learning_rate,
"decay_schedule_fn": self.decay_schedule_fn,
"warmup_steps": self.warmup_steps,
"power": self.power,
"name": self.name,
}
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = 0.0, __UpperCAmelCase = 0.9, __UpperCAmelCase = 0.9_9_9, __UpperCAmelCase = 1e-8, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = 0.0, __UpperCAmelCase = 1.0, __UpperCAmelCase = None, ) -> List[str]:
'''simple docstring'''
snake_case_ = tf.keras.optimizers.schedules.PolynomialDecay(
initial_learning_rate=__UpperCAmelCase, decay_steps=num_train_steps - num_warmup_steps, end_learning_rate=init_lr * min_lr_ratio, power=__UpperCAmelCase, )
if num_warmup_steps:
snake_case_ = WarmUp(
initial_learning_rate=__UpperCAmelCase, decay_schedule_fn=__UpperCAmelCase, warmup_steps=__UpperCAmelCase, )
if weight_decay_rate > 0.0:
snake_case_ = AdamWeightDecay(
learning_rate=__UpperCAmelCase, weight_decay_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, exclude_from_weight_decay=['''LayerNorm''', '''layer_norm''', '''bias'''], include_in_weight_decay=__UpperCAmelCase, )
else:
snake_case_ = tf.keras.optimizers.Adam(
learning_rate=__UpperCAmelCase, beta_a=__UpperCAmelCase, beta_a=__UpperCAmelCase, epsilon=__UpperCAmelCase, clipnorm=__UpperCAmelCase, global_clipnorm=__UpperCAmelCase, )
# We return the optimizer and the LR scheduler in order to better track the
# evolution of the LR independently of the optimizer.
return optimizer, lr_schedule
class a ( _lowerCamelCase ):
def __init__( self : Dict , lowercase_ : Union[float, tf.keras.optimizers.schedules.LearningRateSchedule] = 0.001 , lowercase_ : float = 0.9 , lowercase_ : float = 0.999 , lowercase_ : float = 1e-7 , lowercase_ : bool = False , lowercase_ : float = 0.0 , lowercase_ : Optional[List[str]] = None , lowercase_ : Optional[List[str]] = None , lowercase_ : str = "AdamWeightDecay" , **lowercase_ : Optional[int] , ):
super().__init__(lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
snake_case_ = weight_decay_rate
snake_case_ = include_in_weight_decay
snake_case_ = exclude_from_weight_decay
@classmethod
def A_ ( cls : Dict , lowercase_ : Union[str, Any] ):
snake_case_ = {'''WarmUp''': WarmUp}
return super(lowercase_ , cls ).from_config(lowercase_ , custom_objects=lowercase_ )
def A_ ( self : str , lowercase_ : Optional[int] , lowercase_ : List[str] , lowercase_ : Optional[int] ):
super(lowercase_ , self )._prepare_local(lowercase_ , lowercase_ , lowercase_ )
snake_case_ = tf.constant(
self.weight_decay_rate , name='''adam_weight_decay_rate''' )
def A_ ( self : str , lowercase_ : Union[str, Any] , lowercase_ : Dict , lowercase_ : Any ):
snake_case_ = self._do_use_weight_decay(var.name )
if do_decay:
return var.assign_sub(
learning_rate * var * apply_state[(var.device, var.dtype.base_dtype)]['''weight_decay_rate'''] , use_locking=self._use_locking , )
return tf.no_op()
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : str=None , **lowercase_ : List[str] ):
snake_case_ ,snake_case_ = list(zip(*lowercase_ ) )
return super(lowercase_ , self ).apply_gradients(zip(lowercase_ , lowercase_ ) , name=lowercase_ , **lowercase_ )
def A_ ( self : List[Any] , lowercase_ : str , lowercase_ : str , lowercase_ : Any ):
if apply_state is None:
return self._decayed_lr_t[var_dtype], {}
snake_case_ = apply_state or {}
snake_case_ = apply_state.get((var_device, var_dtype) )
if coefficients is None:
snake_case_ = self._fallback_apply_state(lowercase_ , lowercase_ )
snake_case_ = coefficients
return coefficients["lr_t"], {"apply_state": apply_state}
def A_ ( self : Union[str, Any] , lowercase_ : int , lowercase_ : Union[str, Any] , lowercase_ : Optional[int]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_dense(lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Optional[Any] , lowercase_ : Optional[Any] , lowercase_ : int , lowercase_ : str , lowercase_ : List[Any]=None ):
snake_case_ ,snake_case_ = self._get_lr(var.device , var.dtype.base_dtype , lowercase_ )
snake_case_ = self._decay_weights_op(lowercase_ , lowercase_ , lowercase_ )
with tf.control_dependencies([decay] ):
return super(lowercase_ , self )._resource_apply_sparse(lowercase_ , lowercase_ , lowercase_ , **lowercase_ )
def A_ ( self : Union[str, Any] ):
snake_case_ = super().get_config()
config.update({'''weight_decay_rate''': self.weight_decay_rate} )
return config
def A_ ( self : Optional[int] , lowercase_ : int ):
if self.weight_decay_rate == 0:
return False
if self._include_in_weight_decay:
for r in self._include_in_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return True
if self._exclude_from_weight_decay:
for r in self._exclude_from_weight_decay:
if re.search(lowercase_ , lowercase_ ) is not None:
return False
return True
class a ( _lowerCamelCase ):
def __init__( self : List[Any] ):
snake_case_ = []
snake_case_ = None
@property
def A_ ( self : Union[str, Any] ):
if self._accum_steps is None:
snake_case_ = tf.Variable(
tf.constant(0 , dtype=tf.intaa ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
return self._accum_steps.value()
@property
def A_ ( self : Dict ):
if not self._gradients:
raise ValueError('''The accumulator should be called first to initialize the gradients''' )
return [gradient.value() if gradient is not None else gradient for gradient in self._gradients]
def __call__( self : Any , lowercase_ : int ):
if not self._gradients:
snake_case_ = self.step # Create the step variable.
self._gradients.extend(
[
tf.Variable(
tf.zeros_like(lowercase_ ) , trainable=lowercase_ , synchronization=tf.VariableSynchronization.ON_READ , aggregation=tf.VariableAggregation.ONLY_FIRST_REPLICA , )
if gradient is not None
else gradient
for gradient in gradients
] )
if len(lowercase_ ) != len(self._gradients ):
raise ValueError(F"Expected {len(self._gradients )} gradients, but got {len(lowercase_ )}" )
for accum_gradient, gradient in zip(self._gradients , lowercase_ ):
if accum_gradient is not None and gradient is not None:
accum_gradient.assign_add(lowercase_ )
self._accum_steps.assign_add(1 )
def A_ ( self : Optional[int] ):
if not self._gradients:
return
self._accum_steps.assign(0 )
for gradient in self._gradients:
if gradient is not None:
gradient.assign(tf.zeros_like(lowercase_ ) )
| 56
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import AutoencoderKL, CycleDiffusionPipeline, DDIMScheduler, UNetaDConditionModel
from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, skip_mps
from ..pipeline_params import (
IMAGE_TO_IMAGE_IMAGE_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS,
TEXT_GUIDED_IMAGE_VARIATION_PARAMS,
)
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = CycleDiffusionPipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {
"negative_prompt",
"height",
"width",
"negative_prompt_embeds",
}
snake_case_ = PipelineTesterMixin.required_optional_params - {"latents"}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({"source_prompt"} )
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
snake_case_ = IMAGE_TO_IMAGE_IMAGE_PARAMS
def A_ ( self : Tuple ):
torch.manual_seed(0 )
snake_case_ = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('''DownBlock2D''', '''CrossAttnDownBlock2D''') , up_block_types=('''CrossAttnUpBlock2D''', '''UpBlock2D''') , cross_attention_dim=32 , )
snake_case_ = DDIMScheduler(
beta_start=0.0_0085 , beta_end=0.012 , beta_schedule='''scaled_linear''' , num_train_timesteps=1000 , clip_sample=lowercase_ , set_alpha_to_one=lowercase_ , )
torch.manual_seed(0 )
snake_case_ = 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 )
snake_case_ = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , )
snake_case_ = CLIPTextModel(lowercase_ )
snake_case_ = CLIPTokenizer.from_pretrained('''hf-internal-testing/tiny-random-clip''' )
snake_case_ = {
'''unet''': unet,
'''scheduler''': scheduler,
'''vae''': vae,
'''text_encoder''': text_encoder,
'''tokenizer''': tokenizer,
'''safety_checker''': None,
'''feature_extractor''': None,
}
return components
def A_ ( self : Any , lowercase_ : int , lowercase_ : Optional[Any]=0 ):
snake_case_ = floats_tensor((1, 3, 32, 32) , rng=random.Random(lowercase_ ) ).to(lowercase_ )
snake_case_ = image / 2 + 0.5
if str(lowercase_ ).startswith('''mps''' ):
snake_case_ = torch.manual_seed(lowercase_ )
else:
snake_case_ = torch.Generator(device=lowercase_ ).manual_seed(lowercase_ )
snake_case_ = {
'''prompt''': '''An astronaut riding an elephant''',
'''source_prompt''': '''An astronaut riding a horse''',
'''image''': image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''eta''': 0.1,
'''strength''': 0.8,
'''guidance_scale''': 3,
'''source_guidance_scale''': 1,
'''output_type''': '''numpy''',
}
return inputs
def A_ ( self : Union[str, Any] ):
snake_case_ = '''cpu''' # ensure determinism for the device-dependent torch.Generator
snake_case_ = self.get_dummy_components()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.4459, 0.4943, 0.4544, 0.6643, 0.5474, 0.4327, 0.5701, 0.5959, 0.5179] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@unittest.skipIf(torch_device != '''cuda''' , '''This test requires a GPU''' )
def A_ ( self : Union[str, Any] ):
snake_case_ = self.get_dummy_components()
for name, module in components.items():
if hasattr(lowercase_ , '''half''' ):
snake_case_ = module.half()
snake_case_ = CycleDiffusionPipeline(**lowercase_ )
snake_case_ = pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = self.get_dummy_inputs(lowercase_ )
snake_case_ = pipe(**lowercase_ )
snake_case_ = output.images
snake_case_ = images[0, -3:, -3:, -1]
assert images.shape == (1, 32, 32, 3)
snake_case_ = np.array([0.3506, 0.4543, 0.446, 0.4575, 0.5195, 0.4155, 0.5273, 0.518, 0.4116] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
@skip_mps
def A_ ( self : Optional[int] ):
return super().test_save_load_local()
@unittest.skip('''non-deterministic pipeline''' )
def A_ ( self : List[Any] ):
return super().test_inference_batch_single_identical()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_dict_tuple_outputs_equivalent()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_save_load_optional_components()
@skip_mps
def A_ ( self : Union[str, Any] ):
return super().test_attention_slicing_forward_pass()
@slow
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : List[Any] ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : Union[str, Any] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car_fp16.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(
lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ , torch_dtype=torch.floataa , revision='''fp16''' )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
# the values aren't exactly equal, but the images look the same visually
assert np.abs(image - expected_image ).max() < 5e-1
def A_ ( self : List[str] ):
snake_case_ = load_image(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main'''
'''/cycle-diffusion/black_colored_car.png''' )
snake_case_ = load_numpy(
'''https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/cycle-diffusion/blue_colored_car.npy''' )
snake_case_ = init_image.resize((512, 512) )
snake_case_ = '''CompVis/stable-diffusion-v1-4'''
snake_case_ = DDIMScheduler.from_pretrained(lowercase_ , subfolder='''scheduler''' )
snake_case_ = CycleDiffusionPipeline.from_pretrained(lowercase_ , scheduler=lowercase_ , safety_checker=lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
pipe.enable_attention_slicing()
snake_case_ = '''A black colored car'''
snake_case_ = '''A blue colored car'''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , source_prompt=lowercase_ , image=lowercase_ , num_inference_steps=100 , eta=0.1 , strength=0.85 , guidance_scale=3 , source_guidance_scale=1 , generator=lowercase_ , output_type='''np''' , )
snake_case_ = output.images
assert np.abs(image - expected_image ).max() < 2e-2
| 56
| 1
|
'''simple docstring'''
from __future__ import annotations
import inspect
import unittest
from math import floor
import numpy as np
from transformers import CvtConfig
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 TFCvtForImageClassification, TFCvtModel
from transformers.models.cvt.modeling_tf_cvt import TF_CVT_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class a ( _lowerCamelCase ):
def A_ ( self : str ):
snake_case_ = self.config_class(**self.inputs_dict )
self.parent.assertTrue(hasattr(lowercase_ , '''embed_dim''' ) )
self.parent.assertTrue(hasattr(lowercase_ , '''num_heads''' ) )
class a :
def __init__( self : int , lowercase_ : Any , lowercase_ : Dict=13 , lowercase_ : Optional[int]=64 , lowercase_ : Optional[int]=3 , lowercase_ : Union[str, Any]=[16, 48, 96] , lowercase_ : List[str]=[1, 3, 6] , lowercase_ : Optional[Any]=[1, 2, 10] , lowercase_ : List[Any]=[7, 3, 3] , lowercase_ : List[Any]=[4, 2, 2] , lowercase_ : Union[str, Any]=[2, 1, 1] , lowercase_ : Tuple=[2, 2, 2] , lowercase_ : Union[str, Any]=[False, False, True] , lowercase_ : str=[0.0, 0.0, 0.0] , lowercase_ : Optional[Any]=0.02 , lowercase_ : Optional[Any]=1e-12 , lowercase_ : int=True , lowercase_ : Union[str, Any]=True , lowercase_ : Any=2 , ):
snake_case_ = parent
snake_case_ = batch_size
snake_case_ = image_size
snake_case_ = patch_sizes
snake_case_ = patch_stride
snake_case_ = patch_padding
snake_case_ = is_training
snake_case_ = use_labels
snake_case_ = num_labels
snake_case_ = num_channels
snake_case_ = embed_dim
snake_case_ = num_heads
snake_case_ = stride_kv
snake_case_ = depth
snake_case_ = cls_token
snake_case_ = attention_drop_rate
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
def A_ ( self : List[Any] ):
snake_case_ = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ = None
if self.use_labels:
# create a random int32 tensor of given shape
snake_case_ = ids_tensor([self.batch_size] , self.num_labels )
snake_case_ = self.get_config()
return config, pixel_values, labels
def A_ ( self : Tuple ):
return CvtConfig(
image_size=self.image_size , num_labels=self.num_labels , num_channels=self.num_channels , embed_dim=self.embed_dim , num_heads=self.num_heads , patch_sizes=self.patch_sizes , patch_padding=self.patch_padding , patch_stride=self.patch_stride , stride_kv=self.stride_kv , depth=self.depth , cls_token=self.cls_token , attention_drop_rate=self.attention_drop_rate , initializer_range=self.initializer_range , )
def A_ ( self : int , lowercase_ : Union[str, Any] , lowercase_ : Any , lowercase_ : Any ):
snake_case_ = TFCvtModel(config=lowercase_ )
snake_case_ = model(lowercase_ , training=lowercase_ )
snake_case_ = (self.image_size, self.image_size)
snake_case_ ,snake_case_ = image_size[0], image_size[1]
for i in range(len(self.depth ) ):
snake_case_ = floor(((height + 2 * self.patch_padding[i] - self.patch_sizes[i]) / self.patch_stride[i]) + 1 )
snake_case_ = floor(((width + 2 * self.patch_padding[i] - self.patch_sizes[i]) / self.patch_stride[i]) + 1 )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.embed_dim[-1], height, width) )
def A_ ( self : Any , lowercase_ : Dict , lowercase_ : Optional[Any] , lowercase_ : Optional[Any] ):
snake_case_ = self.num_labels
snake_case_ = TFCvtForImageClassification(lowercase_ )
snake_case_ = model(lowercase_ , labels=lowercase_ , training=lowercase_ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def A_ ( self : str ):
snake_case_ = self.prepare_config_and_inputs()
snake_case_ ,snake_case_ ,snake_case_ = config_and_inputs
snake_case_ = {'''pixel_values''': pixel_values}
return config, inputs_dict
@require_tf
class a ( _lowerCamelCase , _lowerCamelCase , unittest.TestCase ):
snake_case_ = (TFCvtModel, TFCvtForImageClassification) if is_tf_available() else ()
snake_case_ = (
{"feature-extraction": TFCvtModel, "image-classification": TFCvtForImageClassification}
if is_tf_available()
else {}
)
snake_case_ = False
snake_case_ = False
snake_case_ = False
snake_case_ = False
snake_case_ = False
def A_ ( self : Dict ):
snake_case_ = TFCvtModelTester(self )
snake_case_ = TFCvtConfigTester(self , config_class=lowercase_ , has_text_modality=lowercase_ , hidden_size=37 )
def A_ ( self : List[Any] ):
self.config_tester.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
@unittest.skip(reason='''Cvt does not output attentions''' )
def A_ ( self : Any ):
pass
@unittest.skip(reason='''Cvt does not use inputs_embeds''' )
def A_ ( self : Any ):
pass
@unittest.skip(reason='''Cvt does not support input and output embeddings''' )
def A_ ( self : int ):
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.''' , )
def A_ ( self : List[Any] ):
super().test_dataset_conversion()
@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 : Union[str, Any] ):
super().test_keras_fit()
@unittest.skip(reason='''Get `Failed to determine best cudnn convolution algo.` error after using TF 2.12+cuda 11.8''' )
def A_ ( self : Any ):
snake_case_ = tf.keras.mixed_precision.Policy('''mixed_float16''' )
tf.keras.mixed_precision.set_global_policy(lowercase_ )
super().test_keras_fit()
tf.keras.mixed_precision.set_global_policy('''float32''' )
def A_ ( self : Tuple ):
snake_case_ ,snake_case_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ = model_class(lowercase_ )
snake_case_ = inspect.signature(model.call )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ = [*signature.parameters.keys()]
snake_case_ = ['''pixel_values''']
self.assertListEqual(arg_names[:1] , lowercase_ )
def A_ ( self : List[str] ):
def check_hidden_states_output(lowercase_ : Any , lowercase_ : Tuple , lowercase_ : int ):
snake_case_ = model_class(lowercase_ )
snake_case_ = model(**self._prepare_for_class(lowercase_ , lowercase_ ) )
snake_case_ = outputs.hidden_states
snake_case_ = len(self.model_tester.depth )
self.assertEqual(len(lowercase_ ) , lowercase_ )
# verify the first hidden states (first block)
self.assertListEqual(
list(hidden_states[0].shape[-3:] ) , [
self.model_tester.embed_dim[0],
self.model_tester.image_size // 4,
self.model_tester.image_size // 4,
] , )
snake_case_ ,snake_case_ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ = True
check_hidden_states_output(lowercase_ , lowercase_ , lowercase_ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case_ = True
check_hidden_states_output(lowercase_ , lowercase_ , lowercase_ )
def A_ ( self : str ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*lowercase_ )
def A_ ( self : List[str] ):
snake_case_ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*lowercase_ )
@slow
def A_ ( self : Optional[int] ):
for model_name in TF_CVT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ = TFCvtModel.from_pretrained(lowercase_ )
self.assertIsNotNone(lowercase_ )
def __magic_name__ ( ) -> List[str]:
'''simple docstring'''
snake_case_ = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' )
return image
@require_tf
@require_vision
class a ( unittest.TestCase ):
@cached_property
def A_ ( self : Any ):
return AutoImageProcessor.from_pretrained(TF_CVT_PRETRAINED_MODEL_ARCHIVE_LIST[0] )
@slow
def A_ ( self : Optional[Any] ):
snake_case_ = TFCvtForImageClassification.from_pretrained(TF_CVT_PRETRAINED_MODEL_ARCHIVE_LIST[0] )
snake_case_ = self.default_image_processor
snake_case_ = prepare_img()
snake_case_ = image_processor(images=lowercase_ , return_tensors='''tf''' )
# forward pass
snake_case_ = model(**lowercase_ )
# verify the logits
snake_case_ = tf.TensorShape((1, 1000) )
self.assertEqual(outputs.logits.shape , lowercase_ )
snake_case_ = tf.constant([0.9285, 0.9015, -0.3150] )
self.assertTrue(np.allclose(outputs.logits[0, :3].numpy() , lowercase_ , atol=1e-4 ) )
| 56
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
a : str = logging.get_logger(__name__)
a : str = {
'google/bigbird-roberta-base': 'https://huggingface.co/google/bigbird-roberta-base/resolve/main/config.json',
'google/bigbird-roberta-large': 'https://huggingface.co/google/bigbird-roberta-large/resolve/main/config.json',
'google/bigbird-base-trivia-itc': 'https://huggingface.co/google/bigbird-base-trivia-itc/resolve/main/config.json',
# See all BigBird models at https://huggingface.co/models?filter=big_bird
}
class a ( _lowerCamelCase ):
snake_case_ = "big_bird"
def __init__( self : Union[str, Any] , lowercase_ : List[Any]=5_0358 , lowercase_ : Tuple=768 , lowercase_ : Dict=12 , lowercase_ : str=12 , lowercase_ : Tuple=3072 , lowercase_ : Any="gelu_new" , lowercase_ : Optional[Any]=0.1 , lowercase_ : List[Any]=0.1 , lowercase_ : List[Any]=4096 , lowercase_ : List[Any]=2 , lowercase_ : List[str]=0.02 , lowercase_ : Optional[int]=1e-12 , lowercase_ : Tuple=True , lowercase_ : Tuple=0 , lowercase_ : str=1 , lowercase_ : Union[str, Any]=2 , lowercase_ : Optional[Any]=66 , lowercase_ : Optional[int]="block_sparse" , lowercase_ : Any=True , lowercase_ : List[str]=False , lowercase_ : Any=64 , lowercase_ : Tuple=3 , lowercase_ : Tuple=None , **lowercase_ : Tuple , ):
super().__init__(
pad_token_id=lowercase_ , bos_token_id=lowercase_ , eos_token_id=lowercase_ , sep_token_id=lowercase_ , **lowercase_ , )
snake_case_ = vocab_size
snake_case_ = max_position_embeddings
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = initializer_range
snake_case_ = type_vocab_size
snake_case_ = layer_norm_eps
snake_case_ = use_cache
snake_case_ = rescale_embeddings
snake_case_ = attention_type
snake_case_ = use_bias
snake_case_ = block_size
snake_case_ = num_random_blocks
snake_case_ = classifier_dropout
class a ( _lowerCamelCase ):
@property
def A_ ( self : str ):
if self.task == "multiple-choice":
snake_case_ = {0: '''batch''', 1: '''choice''', 2: '''sequence'''}
else:
snake_case_ = {0: '''batch''', 1: '''sequence'''}
return OrderedDict(
[
('''input_ids''', dynamic_axis),
('''attention_mask''', dynamic_axis),
] )
| 56
| 1
|
'''simple docstring'''
import logging
import sys
from dataclasses import dataclass, field
from typing import Any, Dict, List, Optional, Union
import librosa
import torch
from datasets import DatasetDict, load_dataset
from packaging import version
from torch import nn
from transformers import (
HfArgumentParser,
Trainer,
TrainingArguments,
WavaVecaConfig,
WavaVecaFeatureExtractor,
WavaVecaForPreTraining,
is_apex_available,
trainer_utils,
)
from transformers.models.wavaveca.modeling_wavaveca import _compute_mask_indices
if is_apex_available():
from apex import amp
if version.parse(version.parse(torch.__version__).base_version) >= version.parse('1.6'):
a : Dict = True
from torch.cuda.amp import autocast
a : List[str] = logging.getLogger(__name__)
@dataclass
class a :
snake_case_ = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} , )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Whether to freeze the feature extractor layers of the model."} )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Whether to log verbose messages or not."} , )
snake_case_ = field(
default=2.0 , metadata={"help": "Maximum temperature for gumbel softmax."} )
snake_case_ = field(
default=0.5 , metadata={"help": "Minimum temperature for gumbel softmax."} )
snake_case_ = field(
default=0.999_995 , metadata={"help": "Decay of gumbel temperature during training."} )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
logging.basicConfig(
format='''%(asctime)s - %(levelname)s - %(name)s - %(message)s''', datefmt='''%m/%d/%Y %H:%M:%S''', handlers=[logging.StreamHandler(sys.stdout )], )
snake_case_ = logging.WARNING
if model_args.verbose_logging:
snake_case_ = logging.DEBUG
elif trainer_utils.is_main_process(training_args.local_rank ):
snake_case_ = logging.INFO
logger.setLevel(__UpperCAmelCase )
@dataclass
class a :
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "The name of the dataset to use (via the datasets library)."} )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "The configuration name of the dataset to use (via the datasets library)."} )
snake_case_ = field(
default="train" , metadata={
"help": "The name of the training data set split to use (via the datasets library). Defaults to 'train'"
} , )
snake_case_ = field(
default="validation" , metadata={
"help": (
"The name of the validation data set split to use (via the datasets library). Defaults to 'validation'"
)
} , )
snake_case_ = field(
default="file" , metadata={"help": "Column in the dataset that contains speech file path. Defaults to 'file'"} , )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "Overwrite the cached preprocessed datasets or not."} )
snake_case_ = field(
default=1 , metadata={
"help": "The percentage of the train set used as validation set in case there's no validation split"
} , )
snake_case_ = field(
default=_lowerCamelCase , metadata={"help": "The number of processes to use for the preprocessing."} , )
snake_case_ = field(
default=20.0 , metadata={"help": "Filter audio files that are longer than `max_duration_in_seconds` seconds"} )
@dataclass
class a :
snake_case_ = 42
snake_case_ = 42
snake_case_ = "longest"
snake_case_ = None
snake_case_ = None
def __call__( self : str , lowercase_ : List[Dict[str, Union[List[int], torch.Tensor]]] ):
# reformat list to dict and set to pytorch format
snake_case_ = self.feature_extractor.pad(
lowercase_ , max_length=self.max_length , padding=self.padding , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors='''pt''' , )
snake_case_ = self.model._get_feat_extract_output_lengths(batch['''input_values'''].shape[-1] )
snake_case_ = batch['''input_values'''].shape[0]
# make sure that no loss is computed on padded inputs
if batch["attention_mask"] is not None:
# compute real output lengths according to convolution formula
snake_case_ = self.model._get_feat_extract_output_lengths(batch['''attention_mask'''].sum(-1 ) ).to(
torch.long )
snake_case_ = torch.zeros(
(batch_size, mask_indices_seq_length) , dtype=torch.long , device=batch['''input_values'''].device )
# these two operations makes sure that all values
# before the output lengths indices are attended to
snake_case_ = 1
snake_case_ = attention_mask.flip([-1] ).cumsum(-1 ).flip([-1] ).bool()
# sample randomly masked indices
snake_case_ = _compute_mask_indices(
(batch_size, mask_indices_seq_length) , self.model.config.mask_time_prob , self.model.config.mask_time_length , attention_mask=lowercase_ , min_masks=2 , )
return batch
class a ( _lowerCamelCase ):
def __init__( self : Dict , *lowercase_ : Optional[Any] , lowercase_ : Tuple=1 , lowercase_ : Dict=0 , lowercase_ : Dict=1.0 , **lowercase_ : Optional[Any] ):
super().__init__(*lowercase_ , **lowercase_ )
snake_case_ = 0
snake_case_ = max_gumbel_temp
snake_case_ = min_gumbel_temp
snake_case_ = gumbel_temp_decay
def A_ ( self : Optional[Any] , lowercase_ : nn.Module , lowercase_ : Dict[str, Union[torch.Tensor, Any]] ):
model.train()
snake_case_ = self._prepare_inputs(lowercase_ )
if self.use_amp:
with autocast():
snake_case_ = self.compute_loss(lowercase_ , lowercase_ )
else:
snake_case_ = self.compute_loss(lowercase_ , lowercase_ )
if self.args.n_gpu > 1 or self.deepspeed:
if model.module.config.ctc_loss_reduction == "mean":
snake_case_ = loss.mean()
elif model.module.config.ctc_loss_reduction == "sum":
snake_case_ = loss.sum() / (inputs['''mask_time_indices''']).sum()
else:
raise ValueError(F"{model.config.ctc_loss_reduction} is not valid. Choose one of ['mean', 'sum']" )
if self.args.gradient_accumulation_steps > 1:
snake_case_ = loss / self.args.gradient_accumulation_steps
if self.use_amp:
self.scaler.scale(lowercase_ ).backward()
elif self.use_apex:
with amp.scale_loss(lowercase_ , self.optimizer ) as scaled_loss:
scaled_loss.backward()
elif self.deepspeed:
self.deepspeed.backward(lowercase_ )
else:
loss.backward()
self.num_update_step += 1
# make sure gumbel softmax temperature is decayed
if self.args.n_gpu > 1 or self.deepspeed:
model.module.set_gumbel_temperature(
max(self.max_gumbel_temp * self.gumbel_temp_decay**self.num_update_step , self.min_gumbel_temp ) )
else:
model.set_gumbel_temperature(
max(self.max_gumbel_temp * self.gumbel_temp_decay**self.num_update_step , self.min_gumbel_temp ) )
return loss.detach()
def __magic_name__ ( ) -> Dict:
'''simple docstring'''
snake_case_ = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments) )
snake_case_ ,snake_case_ ,snake_case_ = parser.parse_args_into_dataclasses()
configure_logger(__UpperCAmelCase, __UpperCAmelCase )
# Downloading and loading a dataset from the hub.
snake_case_ = load_dataset(data_args.dataset_name, data_args.dataset_config_name, cache_dir=model_args.cache_dir )
if "validation" not in datasets.keys():
# make sure only "validation" and "train" keys remain"
snake_case_ = DatasetDict()
snake_case_ = load_dataset(
data_args.dataset_name, data_args.dataset_config_name, split=F"{data_args.train_split_name}[:{data_args.validation_split_percentage}%]", cache_dir=model_args.cache_dir, )
snake_case_ = load_dataset(
data_args.dataset_name, data_args.dataset_config_name, split=F"{data_args.train_split_name}[{data_args.validation_split_percentage}%:]", cache_dir=model_args.cache_dir, )
else:
# make sure only "validation" and "train" keys remain"
snake_case_ = DatasetDict()
snake_case_ = load_dataset(
data_args.dataset_name, data_args.dataset_config_name, split='''validation''', cache_dir=model_args.cache_dir, )
snake_case_ = load_dataset(
data_args.dataset_name, data_args.dataset_config_name, split=F"{data_args.train_split_name}", cache_dir=model_args.cache_dir, )
# only normalized-inputs-training is supported
snake_case_ = WavaVecaFeatureExtractor.from_pretrained(
model_args.model_name_or_path, cache_dir=model_args.cache_dir, do_normalize=__UpperCAmelCase )
def prepare_dataset(__UpperCAmelCase ):
# check that all files have the correct sampling rate
snake_case_ ,snake_case_ = librosa.load(batch[data_args.speech_file_column], sr=feature_extractor.sampling_rate )
return batch
# load audio files into numpy arrays
snake_case_ = datasets.map(
__UpperCAmelCase, num_proc=data_args.preprocessing_num_workers, remove_columns=datasets['''train'''].column_names )
# filter audio files that are too long
snake_case_ = vectorized_datasets.filter(
lambda __UpperCAmelCase : len(data['''speech'''] ) < int(data_args.max_duration_in_seconds * feature_extractor.sampling_rate ) )
def normalize(__UpperCAmelCase ):
return feature_extractor(batch['''speech'''], sampling_rate=feature_extractor.sampling_rate )
# normalize and transform to `BatchFeatures`
snake_case_ = vectorized_datasets.map(
__UpperCAmelCase, batched=__UpperCAmelCase, num_proc=data_args.preprocessing_num_workers, load_from_cache_file=not data_args.overwrite_cache, remove_columns=vectorized_datasets['''train'''].column_names, )
# pretraining is only supported for "newer" stable layer norm architecture
# apply_spec_augment has to be True, mask_feature_prob has to be 0.0
snake_case_ = WavaVecaConfig.from_pretrained(
model_args.model_name_or_path, cache_dir=model_args.cache_dir, gradient_checkpointing=training_args.gradient_checkpointing, )
if not config.do_stable_layer_norm or config.feat_extract_norm != "layer":
raise ValueError(
'''PreTraining is only supported for ``config.do_stable_layer_norm=True`` and'''
''' ``config.feat_extract_norm=\'layer\'''' )
snake_case_ = WavaVecaForPreTraining(__UpperCAmelCase )
snake_case_ = DataCollatorForWavaVecaPretraining(model=__UpperCAmelCase, feature_extractor=__UpperCAmelCase )
snake_case_ = WavaVecaPreTrainer(
model=__UpperCAmelCase, data_collator=__UpperCAmelCase, args=__UpperCAmelCase, train_dataset=vectorized_datasets['''train'''], eval_dataset=vectorized_datasets['''validation'''], tokenizer=__UpperCAmelCase, max_gumbel_temp=model_args.max_gumbel_temperature, min_gumbel_temp=model_args.min_gumbel_temperature, gumbel_temp_decay=model_args.gumbel_temperature_decay, )
trainer.train()
if __name__ == "__main__":
main()
| 56
|
'''simple docstring'''
import contextlib
import os
import sqlitea
import pytest
from datasets import Dataset, Features, Value
from datasets.io.sql import SqlDatasetReader, SqlDatasetWriter
from ..utils import assert_arrow_memory_doesnt_increase, assert_arrow_memory_increases, require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> str:
'''simple docstring'''
assert isinstance(__UpperCAmelCase, __UpperCAmelCase )
assert dataset.num_rows == 4
assert dataset.num_columns == 3
assert dataset.column_names == ["col_1", "col_2", "col_3"]
for feature, expected_dtype in expected_features.items():
assert dataset.features[feature].dtype == expected_dtype
@require_sqlalchemy
@pytest.mark.parametrize('''keep_in_memory''', [False, True] )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase():
snake_case_ = SqlDatasetReader(
'''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase, keep_in_memory=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
@require_sqlalchemy
@pytest.mark.parametrize(
'''features''', [
None,
{'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''},
{'''col_1''': '''string''', '''col_2''': '''string''', '''col_3''': '''string'''},
{'''col_1''': '''int32''', '''col_2''': '''int32''', '''col_3''': '''int32'''},
{'''col_1''': '''float32''', '''col_2''': '''float32''', '''col_3''': '''float32'''},
], )
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = {'''col_1''': '''string''', '''col_2''': '''int64''', '''col_3''': '''float64'''}
snake_case_ = features.copy() if features else default_expected_features
snake_case_ = (
Features({feature: Value(__UpperCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None
)
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, features=__UpperCAmelCase, cache_dir=__UpperCAmelCase ).read()
_check_sql_dataset(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
with contextlib.closing(sqlitea.connect(__UpperCAmelCase ) ) as con:
snake_case_ = con.cursor()
cur.execute('''SELECT * FROM dataset''' )
for row in cur:
yield row
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=1 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Any:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=2 ).write()
snake_case_ = iter_sql_file(__UpperCAmelCase )
snake_case_ = iter_sql_file(__UpperCAmelCase )
for rowa, rowa in zip(__UpperCAmelCase, __UpperCAmelCase ):
assert rowa == rowa
@require_sqlalchemy
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> List[str]:
'''simple docstring'''
snake_case_ = tmp_path / '''cache'''
snake_case_ = os.path.join(__UpperCAmelCase, '''tmp.sql''' )
snake_case_ = SqlDatasetReader('''dataset''', '''sqlite:///''' + sqlite_path, cache_dir=__UpperCAmelCase ).read()
with pytest.raises(__UpperCAmelCase ):
SqlDatasetWriter(__UpperCAmelCase, '''dataset''', '''sqlite:///''' + output_sqlite_path, num_proc=0 ).write()
| 56
| 1
|
'''simple docstring'''
import datetime
import platform
import subprocess
from typing import Optional, Tuple, Union
import numpy as np
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> np.array:
'''simple docstring'''
snake_case_ = F"{sampling_rate}"
snake_case_ = '''1'''
snake_case_ = '''f32le'''
snake_case_ = [
'''ffmpeg''',
'''-i''',
'''pipe:0''',
'''-ac''',
ac,
'''-ar''',
ar,
'''-f''',
format_for_conversion,
'''-hide_banner''',
'''-loglevel''',
'''quiet''',
'''pipe:1''',
]
try:
with subprocess.Popen(__UpperCAmelCase, stdin=subprocess.PIPE, stdout=subprocess.PIPE ) as ffmpeg_process:
snake_case_ = ffmpeg_process.communicate(__UpperCAmelCase )
except FileNotFoundError as error:
raise ValueError('''ffmpeg was not found but is required to load audio files from filename''' ) from error
snake_case_ = output_stream[0]
snake_case_ = np.frombuffer(__UpperCAmelCase, np.floataa )
if audio.shape[0] == 0:
raise ValueError('''Malformed soundfile''' )
return audio
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = "f32le", ) -> List[Any]:
'''simple docstring'''
snake_case_ = F"{sampling_rate}"
snake_case_ = '''1'''
if format_for_conversion == "s16le":
snake_case_ = 2
elif format_for_conversion == "f32le":
snake_case_ = 4
else:
raise ValueError(F"Unhandled format `{format_for_conversion}`. Please use `s16le` or `f32le`" )
snake_case_ = platform.system()
if system == "Linux":
snake_case_ = '''alsa'''
snake_case_ = '''default'''
elif system == "Darwin":
snake_case_ = '''avfoundation'''
snake_case_ = ''':0'''
elif system == "Windows":
snake_case_ = '''dshow'''
snake_case_ = '''default'''
snake_case_ = [
'''ffmpeg''',
'''-f''',
format_,
'''-i''',
input_,
'''-ac''',
ac,
'''-ar''',
ar,
'''-f''',
format_for_conversion,
'''-fflags''',
'''nobuffer''',
'''-hide_banner''',
'''-loglevel''',
'''quiet''',
'''pipe:1''',
]
snake_case_ = int(round(sampling_rate * chunk_length_s ) ) * size_of_sample
snake_case_ = _ffmpeg_stream(__UpperCAmelCase, __UpperCAmelCase )
for item in iterator:
yield item
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = None, __UpperCAmelCase = None, __UpperCAmelCase = "f32le", ) -> Dict:
'''simple docstring'''
if stream_chunk_s is not None:
snake_case_ = stream_chunk_s
else:
snake_case_ = chunk_length_s
snake_case_ = ffmpeg_microphone(__UpperCAmelCase, __UpperCAmelCase, format_for_conversion=__UpperCAmelCase )
if format_for_conversion == "s16le":
snake_case_ = np.intaa
snake_case_ = 2
elif format_for_conversion == "f32le":
snake_case_ = np.floataa
snake_case_ = 4
else:
raise ValueError(F"Unhandled format `{format_for_conversion}`. Please use `s16le` or `f32le`" )
if stride_length_s is None:
snake_case_ = chunk_length_s / 6
snake_case_ = int(round(sampling_rate * chunk_length_s ) ) * size_of_sample
if isinstance(__UpperCAmelCase, (int, float) ):
snake_case_ = [stride_length_s, stride_length_s]
snake_case_ = int(round(sampling_rate * stride_length_s[0] ) ) * size_of_sample
snake_case_ = int(round(sampling_rate * stride_length_s[1] ) ) * size_of_sample
snake_case_ = datetime.datetime.now()
snake_case_ = datetime.timedelta(seconds=__UpperCAmelCase )
for item in chunk_bytes_iter(__UpperCAmelCase, __UpperCAmelCase, stride=(stride_left, stride_right), stream=__UpperCAmelCase ):
# Put everything back in numpy scale
snake_case_ = np.frombuffer(item['''raw'''], dtype=__UpperCAmelCase )
snake_case_ = (
item['''stride'''][0] // size_of_sample,
item['''stride'''][1] // size_of_sample,
)
snake_case_ = sampling_rate
audio_time += delta
if datetime.datetime.now() > audio_time + 10 * delta:
# We're late !! SKIP
continue
yield item
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase = False ) -> Any:
'''simple docstring'''
snake_case_ = b''''''
snake_case_ ,snake_case_ = stride
if stride_left + stride_right >= chunk_len:
raise ValueError(
F"Stride needs to be strictly smaller than chunk_len: ({stride_left}, {stride_right}) vs {chunk_len}" )
snake_case_ = 0
for raw in iterator:
acc += raw
if stream and len(__UpperCAmelCase ) < chunk_len:
snake_case_ = (_stride_left, 0)
yield {"raw": acc[:chunk_len], "stride": stride, "partial": True}
else:
while len(__UpperCAmelCase ) >= chunk_len:
# We are flushing the accumulator
snake_case_ = (_stride_left, stride_right)
snake_case_ = {'''raw''': acc[:chunk_len], '''stride''': stride}
if stream:
snake_case_ = False
yield item
snake_case_ = stride_left
snake_case_ = acc[chunk_len - stride_left - stride_right :]
# Last chunk
if len(__UpperCAmelCase ) > stride_left:
snake_case_ = {'''raw''': acc, '''stride''': (_stride_left, 0)}
if stream:
snake_case_ = False
yield item
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Optional[int]:
'''simple docstring'''
snake_case_ = 2**24 # 16Mo
try:
with subprocess.Popen(__UpperCAmelCase, stdout=subprocess.PIPE, bufsize=__UpperCAmelCase ) as ffmpeg_process:
while True:
snake_case_ = ffmpeg_process.stdout.read(__UpperCAmelCase )
if raw == b"":
break
yield raw
except FileNotFoundError as error:
raise ValueError('''ffmpeg was not found but is required to stream audio files from filename''' ) from error
| 56
|
'''simple docstring'''
from collections import defaultdict
def __magic_name__ ( __UpperCAmelCase ) -> int:
'''simple docstring'''
snake_case_ = 1
snake_case_ = True
for v in tree[start]:
if v not in visited:
ret += dfs(__UpperCAmelCase )
if ret % 2 == 0:
cuts.append(__UpperCAmelCase )
return ret
def __magic_name__ ( ) -> Union[str, Any]:
'''simple docstring'''
dfs(1 )
if __name__ == "__main__":
a ,a : Dict = 10, 9
a : Dict = defaultdict(list)
a : dict[int, bool] = {}
a : list[int] = []
a : Tuple = 0
a : str = [(2, 1), (3, 1), (4, 3), (5, 2), (6, 1), (7, 2), (8, 6), (9, 8), (10, 8)]
for u, v in edges:
tree[u].append(v)
tree[v].append(u)
even_tree()
print(len(cuts) - 1)
| 56
| 1
|
'''simple docstring'''
import gc
import tempfile
import unittest
import numpy as np
import torch
from diffusers import VersatileDiffusionTextToImagePipeline
from diffusers.utils.testing_utils import nightly, require_torch_gpu, torch_device
a : List[Any] = False
class a ( unittest.TestCase ):
pass
@nightly
@require_torch_gpu
class a ( unittest.TestCase ):
def A_ ( self : Tuple ):
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def A_ ( self : int ):
snake_case_ = VersatileDiffusionTextToImagePipeline.from_pretrained('''shi-labs/versatile-diffusion''' )
# remove text_unet
pipe.remove_unused_weights()
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = '''A painting of a squirrel eating a burger '''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , generator=lowercase_ , guidance_scale=7.5 , num_inference_steps=2 , output_type='''numpy''' ).images
with tempfile.TemporaryDirectory() as tmpdirname:
pipe.save_pretrained(lowercase_ )
snake_case_ = VersatileDiffusionTextToImagePipeline.from_pretrained(lowercase_ )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = generator.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , generator=lowercase_ , guidance_scale=7.5 , num_inference_steps=2 , output_type='''numpy''' ).images
assert np.abs(image - new_image ).sum() < 1e-5, "Models don't have the same forward pass"
def A_ ( self : Tuple ):
snake_case_ = VersatileDiffusionTextToImagePipeline.from_pretrained(
'''shi-labs/versatile-diffusion''' , torch_dtype=torch.floataa )
pipe.to(lowercase_ )
pipe.set_progress_bar_config(disable=lowercase_ )
snake_case_ = '''A painting of a squirrel eating a burger '''
snake_case_ = torch.manual_seed(0 )
snake_case_ = pipe(
prompt=lowercase_ , generator=lowercase_ , guidance_scale=7.5 , num_inference_steps=50 , output_type='''numpy''' ).images
snake_case_ = image[0, 253:256, 253:256, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ = np.array([0.3367, 0.3169, 0.2656, 0.3870, 0.4790, 0.3796, 0.4009, 0.4878, 0.4778] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2
| 56
|
'''simple docstring'''
import math
from collections.abc import Callable
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
snake_case_ = xa
snake_case_ = xa
while True:
if x_n == x_na or function(__UpperCAmelCase ) == function(__UpperCAmelCase ):
raise ZeroDivisionError('''float division by zero, could not find root''' )
snake_case_ = x_na - (
function(__UpperCAmelCase ) / ((function(__UpperCAmelCase ) - function(__UpperCAmelCase )) / (x_na - x_n))
)
if abs(x_na - x_na ) < 10**-5:
return x_na
snake_case_ = x_na
snake_case_ = x_na
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
return math.pow(__UpperCAmelCase, 3 ) - (2 * x) - 5
if __name__ == "__main__":
print(intersection(f, 3, 3.5))
| 56
| 1
|
'''simple docstring'''
import math
from collections.abc import Callable
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> float:
'''simple docstring'''
snake_case_ = xa
snake_case_ = xa
while True:
if x_n == x_na or function(__UpperCAmelCase ) == function(__UpperCAmelCase ):
raise ZeroDivisionError('''float division by zero, could not find root''' )
snake_case_ = x_na - (
function(__UpperCAmelCase ) / ((function(__UpperCAmelCase ) - function(__UpperCAmelCase )) / (x_na - x_n))
)
if abs(x_na - x_na ) < 10**-5:
return x_na
snake_case_ = x_na
snake_case_ = x_na
def __magic_name__ ( __UpperCAmelCase ) -> float:
'''simple docstring'''
return math.pow(__UpperCAmelCase, 3 ) - (2 * x) - 5
if __name__ == "__main__":
print(intersection(f, 3, 3.5))
| 56
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_url
from PIL import Image
from transformers import DPTConfig, DPTForDepthEstimation, DPTForSemanticSegmentation, DPTImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
a : Any = logging.get_logger(__name__)
def __magic_name__ ( __UpperCAmelCase ) -> Optional[Any]:
'''simple docstring'''
snake_case_ = DPTConfig()
if "large" in checkpoint_url:
snake_case_ = 1024
snake_case_ = 4096
snake_case_ = 24
snake_case_ = 16
snake_case_ = [5, 11, 17, 23]
snake_case_ = [256, 512, 1024, 1024]
snake_case_ = (1, 384, 384)
if "ade" in checkpoint_url:
snake_case_ = True
snake_case_ = 150
snake_case_ = '''huggingface/label-files'''
snake_case_ = '''ade20k-id2label.json'''
snake_case_ = json.load(open(cached_download(hf_hub_url(__UpperCAmelCase, __UpperCAmelCase, repo_type='''dataset''' ) ), '''r''' ) )
snake_case_ = {int(__UpperCAmelCase ): v for k, v in idalabel.items()}
snake_case_ = idalabel
snake_case_ = {v: k for k, v in idalabel.items()}
snake_case_ = [1, 150, 480, 480]
return config, expected_shape
def __magic_name__ ( __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ = ['''pretrained.model.head.weight''', '''pretrained.model.head.bias''']
for k in ignore_keys:
state_dict.pop(__UpperCAmelCase, __UpperCAmelCase )
def __magic_name__ ( __UpperCAmelCase ) -> List[Any]:
'''simple docstring'''
if (
"pretrained.model" in name
and "cls_token" not in name
and "pos_embed" not in name
and "patch_embed" not in name
):
snake_case_ = name.replace('''pretrained.model''', '''dpt.encoder''' )
if "pretrained.model" in name:
snake_case_ = name.replace('''pretrained.model''', '''dpt.embeddings''' )
if "patch_embed" in name:
snake_case_ = name.replace('''patch_embed''', '''patch_embeddings''' )
if "pos_embed" in name:
snake_case_ = name.replace('''pos_embed''', '''position_embeddings''' )
if "attn.proj" in name:
snake_case_ = name.replace('''attn.proj''', '''attention.output.dense''' )
if "proj" in name and "project" not in name:
snake_case_ = name.replace('''proj''', '''projection''' )
if "blocks" in name:
snake_case_ = name.replace('''blocks''', '''layer''' )
if "mlp.fc1" in name:
snake_case_ = name.replace('''mlp.fc1''', '''intermediate.dense''' )
if "mlp.fc2" in name:
snake_case_ = name.replace('''mlp.fc2''', '''output.dense''' )
if "norm1" in name:
snake_case_ = name.replace('''norm1''', '''layernorm_before''' )
if "norm2" in name:
snake_case_ = name.replace('''norm2''', '''layernorm_after''' )
if "scratch.output_conv" in name:
snake_case_ = name.replace('''scratch.output_conv''', '''head''' )
if "scratch" in name:
snake_case_ = name.replace('''scratch''', '''neck''' )
if "layer1_rn" in name:
snake_case_ = name.replace('''layer1_rn''', '''convs.0''' )
if "layer2_rn" in name:
snake_case_ = name.replace('''layer2_rn''', '''convs.1''' )
if "layer3_rn" in name:
snake_case_ = name.replace('''layer3_rn''', '''convs.2''' )
if "layer4_rn" in name:
snake_case_ = name.replace('''layer4_rn''', '''convs.3''' )
if "refinenet" in name:
snake_case_ = int(name[len('''neck.refinenet''' ) : len('''neck.refinenet''' ) + 1] )
# tricky here: we need to map 4 to 0, 3 to 1, 2 to 2 and 1 to 3
snake_case_ = name.replace(F"refinenet{layer_idx}", F"fusion_stage.layers.{abs(layer_idx-4 )}" )
if "out_conv" in name:
snake_case_ = name.replace('''out_conv''', '''projection''' )
if "resConfUnit1" in name:
snake_case_ = name.replace('''resConfUnit1''', '''residual_layer1''' )
if "resConfUnit2" in name:
snake_case_ = name.replace('''resConfUnit2''', '''residual_layer2''' )
if "conv1" in name:
snake_case_ = name.replace('''conv1''', '''convolution1''' )
if "conv2" in name:
snake_case_ = name.replace('''conv2''', '''convolution2''' )
# readout blocks
if "pretrained.act_postprocess1.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.0.project.0''', '''neck.reassemble_stage.readout_projects.0.0''' )
if "pretrained.act_postprocess2.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.0.project.0''', '''neck.reassemble_stage.readout_projects.1.0''' )
if "pretrained.act_postprocess3.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.0.project.0''', '''neck.reassemble_stage.readout_projects.2.0''' )
if "pretrained.act_postprocess4.0.project.0" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.0.project.0''', '''neck.reassemble_stage.readout_projects.3.0''' )
# resize blocks
if "pretrained.act_postprocess1.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.3''', '''neck.reassemble_stage.layers.0.projection''' )
if "pretrained.act_postprocess1.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess1.4''', '''neck.reassemble_stage.layers.0.resize''' )
if "pretrained.act_postprocess2.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.3''', '''neck.reassemble_stage.layers.1.projection''' )
if "pretrained.act_postprocess2.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess2.4''', '''neck.reassemble_stage.layers.1.resize''' )
if "pretrained.act_postprocess3.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess3.3''', '''neck.reassemble_stage.layers.2.projection''' )
if "pretrained.act_postprocess4.3" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.3''', '''neck.reassemble_stage.layers.3.projection''' )
if "pretrained.act_postprocess4.4" in name:
snake_case_ = name.replace('''pretrained.act_postprocess4.4''', '''neck.reassemble_stage.layers.3.resize''' )
if "pretrained" in name:
snake_case_ = name.replace('''pretrained''', '''dpt''' )
if "bn" in name:
snake_case_ = name.replace('''bn''', '''batch_norm''' )
if "head" in name:
snake_case_ = name.replace('''head''', '''head.head''' )
if "encoder.norm" in name:
snake_case_ = name.replace('''encoder.norm''', '''layernorm''' )
if "auxlayer" in name:
snake_case_ = name.replace('''auxlayer''', '''auxiliary_head.head''' )
return name
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase ) -> Dict:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.weight" )
snake_case_ = state_dict.pop(F"dpt.encoder.layer.{i}.attn.qkv.bias" )
# next, add query, keys and values (in that order) to the state dict
snake_case_ = in_proj_weight[: config.hidden_size, :]
snake_case_ = in_proj_bias[: config.hidden_size]
snake_case_ = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
snake_case_ = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
snake_case_ = in_proj_weight[
-config.hidden_size :, :
]
snake_case_ = in_proj_bias[-config.hidden_size :]
def __magic_name__ ( ) -> Any:
'''simple docstring'''
snake_case_ = '''http://images.cocodataset.org/val2017/000000039769.jpg'''
snake_case_ = Image.open(requests.get(__UpperCAmelCase, stream=__UpperCAmelCase ).raw )
return im
@torch.no_grad()
def __magic_name__ ( __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase, __UpperCAmelCase ) -> Union[str, Any]:
'''simple docstring'''
snake_case_ ,snake_case_ = get_dpt_config(__UpperCAmelCase )
# load original state_dict from URL
snake_case_ = torch.hub.load_state_dict_from_url(__UpperCAmelCase, map_location='''cpu''' )
# remove certain keys
remove_ignore_keys_(__UpperCAmelCase )
# rename keys
for key in state_dict.copy().keys():
snake_case_ = state_dict.pop(__UpperCAmelCase )
snake_case_ = val
# read in qkv matrices
read_in_q_k_v(__UpperCAmelCase, __UpperCAmelCase )
# load HuggingFace model
snake_case_ = DPTForSemanticSegmentation(__UpperCAmelCase ) if '''ade''' in checkpoint_url else DPTForDepthEstimation(__UpperCAmelCase )
model.load_state_dict(__UpperCAmelCase )
model.eval()
# Check outputs on an image
snake_case_ = 480 if '''ade''' in checkpoint_url else 384
snake_case_ = DPTImageProcessor(size=__UpperCAmelCase )
snake_case_ = prepare_img()
snake_case_ = image_processor(__UpperCAmelCase, return_tensors='''pt''' )
# forward pass
snake_case_ = model(**__UpperCAmelCase ).logits if '''ade''' in checkpoint_url else model(**__UpperCAmelCase ).predicted_depth
# Assert logits
snake_case_ = torch.tensor([[6.3_1_9_9, 6.3_6_2_9, 6.4_1_4_8], [6.3_8_5_0, 6.3_6_1_5, 6.4_1_6_6], [6.3_5_1_9, 6.3_1_7_6, 6.3_5_7_5]] )
if "ade" in checkpoint_url:
snake_case_ = torch.tensor([[4.0_4_8_0, 4.2_4_2_0, 4.4_3_6_0], [4.3_1_2_4, 4.5_6_9_3, 4.8_2_6_1], [4.5_7_6_8, 4.8_9_6_5, 5.2_1_6_3]] )
assert outputs.shape == torch.Size(__UpperCAmelCase )
assert (
torch.allclose(outputs[0, 0, :3, :3], __UpperCAmelCase, atol=1e-4 )
if "ade" in checkpoint_url
else torch.allclose(outputs[0, :3, :3], __UpperCAmelCase )
)
Path(__UpperCAmelCase ).mkdir(exist_ok=__UpperCAmelCase )
print(F"Saving model to {pytorch_dump_folder_path}" )
model.save_pretrained(__UpperCAmelCase )
print(F"Saving image processor to {pytorch_dump_folder_path}" )
image_processor.save_pretrained(__UpperCAmelCase )
if push_to_hub:
print('''Pushing model to hub...''' )
model.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add model''', use_temp_dir=__UpperCAmelCase, )
image_processor.push_to_hub(
repo_path_or_name=Path(__UpperCAmelCase, __UpperCAmelCase ), organization='''nielsr''', commit_message='''Add image processor''', use_temp_dir=__UpperCAmelCase, )
if __name__ == "__main__":
a : Tuple = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'--checkpoint_url',
default='https://github.com/intel-isl/DPT/releases/download/1_0/dpt_large-midas-2f21e586.pt',
type=str,
help='URL of the original DPT checkpoint you\'d like to convert.',
)
parser.add_argument(
'--pytorch_dump_folder_path',
default=None,
type=str,
required=True,
help='Path to the output PyTorch model directory.',
)
parser.add_argument(
'--push_to_hub',
action='store_true',
)
parser.add_argument(
'--model_name',
default='dpt-large',
type=str,
help='Name of the model, in case you\'re pushing to the hub.',
)
a : List[Any] = parser.parse_args()
convert_dpt_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 56
| 1
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.