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""" import argparse import json import os import re import torch from transformers import BloomConfig, BloomModel from transformers.file_utils import CONFIG_NAME, WEIGHTS_NAME from transformers.utils import logging logging.set_verbosity_info() __snake_case = [ "word_embeddings_layernorm.weight", "word_embeddings_layernorm.bias", "input_layernorm.weight", "input_layernorm.bias", "post_attention_layernorm.weight", "post_attention_layernorm.bias", "self_attention.dense.bias", "mlp.dense_4h_to_h.bias", "ln_f.weight", "ln_f.bias", ] __snake_case = [ "mlp.dense_4h_to_h.weight", "self_attention.dense.weight", ] def A_ ( _lowerCAmelCase : Any, _lowerCAmelCase : List[Any] ): """simple docstring""" _a = { '''word_embeddings.weight''': '''word_embeddings.weight''', '''word_embeddings.norm.weight''': '''word_embeddings_layernorm.weight''', '''word_embeddings.norm.bias''': '''word_embeddings_layernorm.bias''', '''weight''': '''ln_f.weight''', '''bias''': '''ln_f.bias''', } if key in layer_rename_map: return layer_rename_map[key] # Handle transformer blocks _a = int(re.match(R'''.*layer_(\d*).*''', __lowerCAmelCase )[1] ) layer_number -= 3 return f'h.{layer_number}.' + key def A_ ( _lowerCAmelCase : Dict ): """simple docstring""" if dtype == torch.bool: return 1 / 8 _a = re.search(R'''[^\d](\d+)$''', str(__lowerCAmelCase ) ) if bit_search is None: raise ValueError(f'`dtype` is not a valid dtype: {dtype}.' ) _a = int(bit_search.groups()[0] ) return bit_size // 8 def A_ ( _lowerCAmelCase : Optional[Any], _lowerCAmelCase : Union[str, Any], _lowerCAmelCase : List[Any], _lowerCAmelCase : List[Any], _lowerCAmelCase : Optional[Any] ): """simple docstring""" if bloom_config_file == "": _a = BloomConfig() else: _a = BloomConfig.from_json_file(__lowerCAmelCase ) if shard_model: _a = os.listdir(__lowerCAmelCase ) _a = sorted(filter(lambda _lowerCAmelCase : s.startswith('''layer''' ) and "model_00" in s, __lowerCAmelCase ) ) _a = {'''weight_map''': {}, '''metadata''': {}} _a = 0 _a = None _a = BloomConfig() for j, file in enumerate(__lowerCAmelCase ): print('''Processing file: {}'''.format(__lowerCAmelCase ) ) _a = None for i in range(__lowerCAmelCase ): # load all TP files _a = file.replace('''model_00''', f'model_0{i}' ) _a = torch.load(os.path.join(__lowerCAmelCase, __lowerCAmelCase ), map_location='''cpu''' ) # Rename keys in the transformers names _a = list(temp.keys() ) for key in keys: _a = temp.pop(__lowerCAmelCase ) if tensors is None: _a = temp else: for key in tensors.keys(): if any(key.endswith(__lowerCAmelCase ) for end in WEIGHTS_TO_AVERAGE_ENDSWITH ): # We average (sum and then divide) some weights accross TP ranks (see https://github.com/bigscience-workshop/Megatron-DeepSpeed/blob/olruwase/sync_layer_norms/megatron/training.py#L425) tensors[key] += temp[key] else: # Some weights are RowParallelLinear in Megatron-Deepspeed, others are ColumnParallel _a = 1 if any(text in key for text in WEIGHTS_WITH_ROW_PARALLELISM_CONTAIN ) else 0 # We concatenate these weights accross TP ranks _a = torch.cat([tensors[key], temp[key]], dim=__lowerCAmelCase ) # Divide by the number of TP the weights we want to average for key in tensors.keys(): if any(key.endswith(__lowerCAmelCase ) for end in WEIGHTS_TO_AVERAGE_ENDSWITH ): _a = tensors[key] / pretraining_tp torch.save( __lowerCAmelCase, os.path.join( __lowerCAmelCase, '''pytorch_model_{}-of-{}.bin'''.format(str(j + 1 ).zfill(5 ), str(len(__lowerCAmelCase ) ).zfill(5 ) ), ), ) for key in tensors.keys(): _a = tensors[key] total_size += value.numel() * get_dtype_size(value.dtype ) if key not in index_dict["weight_map"]: _a = '''pytorch_model_{}-of-{}.bin'''.format( str(j + 1 ).zfill(5 ), str(len(__lowerCAmelCase ) ).zfill(5 ) ) _a = BloomConfig() _a = pytorch_dump_folder_path + '''/''' + CONFIG_NAME _a = total_size with open(__lowerCAmelCase, '''w''', encoding='''utf-8''' ) as f: f.write(config.to_json_string() ) with open(os.path.join(__lowerCAmelCase, WEIGHTS_NAME + '''.index.json''' ), '''w''', encoding='''utf-8''' ) as f: _a = json.dumps(__lowerCAmelCase, indent=2, sort_keys=__lowerCAmelCase ) + '''\n''' f.write(__lowerCAmelCase ) else: _a = BloomModel(__lowerCAmelCase ) _a = os.listdir(__lowerCAmelCase ) _a = sorted(filter(lambda _lowerCAmelCase : s.startswith('''layer''' ) and "model_00" in s, __lowerCAmelCase ) ) _a = None for i, file in enumerate(__lowerCAmelCase ): _a = None for i in range(__lowerCAmelCase ): # load all TP files _a = file.replace('''model_00''', f'model_0{i}' ) _a = torch.load(os.path.join(__lowerCAmelCase, __lowerCAmelCase ), map_location='''cpu''' ) # Rename keys in the transformers names _a = list(temp.keys() ) for key in keys: _a = temp.pop(__lowerCAmelCase ) if tensors is None: _a = temp else: for key in tensors.keys(): # We average (sum and then divide) some weights accross TP ranks (see https://github.com/bigscience-workshop/Megatron-DeepSpeed/blob/olruwase/sync_layer_norms/megatron/training.py#L425) if any(key.endswith(__lowerCAmelCase ) for end in WEIGHTS_TO_AVERAGE_ENDSWITH ): tensors[key] += temp[key] else: # Some weights are RowParallelLinear in Megatron-Deepspeed, others are ColumnParallel _a = 1 if any(text in key for text in WEIGHTS_WITH_ROW_PARALLELISM_CONTAIN ) else 0 # We concatenate these weights accross TP ranks _a = torch.cat([tensors[key], temp[key]], dim=__lowerCAmelCase ) # Divide by the number of TP the weights we want to average for key in tensors.keys(): if any(key.endswith(__lowerCAmelCase ) for end in WEIGHTS_TO_AVERAGE_ENDSWITH ): _a = tensors[key] / pretraining_tp _a = model.load_state_dict(__lowerCAmelCase, strict=__lowerCAmelCase ) assert not other_keys.unexpected_keys, f'The keys {other_keys.unexpected_keys} are unexpected' if missing_keys is None: _a = set(other_keys.missing_keys ) else: _a = missing_keys.intersection(set(other_keys.missing_keys ) ) assert not missing_keys, f'The keys {missing_keys} are missing' # Save pytorch-model os.makedirs(__lowerCAmelCase, exist_ok=__lowerCAmelCase ) _a = pytorch_dump_folder_path + '''/''' + WEIGHTS_NAME _a = pytorch_dump_folder_path + '''/''' + CONFIG_NAME print(f'Save PyTorch model to {pytorch_weights_dump_path} with dtype {config.torch_dtype}' ) if config.torch_dtype is not None: _a = model.to(config.torch_dtype ) torch.save(model.state_dict(), __lowerCAmelCase ) print(f'Save configuration file to {pytorch_config_dump_path}' ) with open(__lowerCAmelCase, '''w''', encoding='''utf-8''' ) as f: f.write(config.to_json_string() ) if __name__ == "__main__": __snake_case = argparse.ArgumentParser() # Required parameters parser.add_argument( '''--bloom_checkpoint_path''', default=None, type=str, required=True, help='''Path to the Megatron-LM checkpoint path.''', ) parser.add_argument( '''--pytorch_dump_folder_path''', default=None, type=str, required=True, help='''Path to the output PyTorch model.''' ) parser.add_argument( '''--bloom_config_file''', default='''''', type=str, help=( '''An optional config json file corresponding to the pre-trained model. \n''' '''This specifies the model architecture.''' ), ) parser.add_argument( '''--shard_model''', action='''store_true''', help='''An optional setting to shard the output model \nThis enables sharding the converted checkpoint''', ) parser.add_argument( '''--pretraining_tp''', default=4, type=int, help='''Pretraining TP rank that has been used when training the model in Megatron-LM \n''', ) __snake_case = parser.parse_args() convert_bloom_checkpoint_to_pytorch( args.bloom_checkpoint_path, args.bloom_config_file, args.pytorch_dump_folder_path, args.shard_model, args.pretraining_tp, )
320
import json import os import subprocess import unittest from ast import literal_eval import pytest from parameterized import parameterized, parameterized_class from . import is_sagemaker_available if is_sagemaker_available(): from sagemaker import Session, TrainingJobAnalytics from sagemaker.huggingface import HuggingFace @pytest.mark.skipif( literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , ) @pytest.mark.usefixtures("""sm_env""" ) @parameterized_class( [ { """framework""": """pytorch""", """script""": """run_glue_model_parallelism.py""", """model_name_or_path""": """roberta-large""", """instance_type""": """ml.p3dn.24xlarge""", """results""": {"""train_runtime""": 1_600, """eval_accuracy""": 0.3, """eval_loss""": 1.2}, }, { """framework""": """pytorch""", """script""": """run_glue.py""", """model_name_or_path""": """roberta-large""", """instance_type""": """ml.p3dn.24xlarge""", """results""": {"""train_runtime""": 1_600, """eval_accuracy""": 0.3, """eval_loss""": 1.2}, }, ] ) class lowerCAmelCase__ ( unittest.TestCase ): def __A ( self : Optional[int] ) -> List[Any]: if self.framework == "pytorch": subprocess.run( f'''cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py'''.split() , encoding='''utf-8''' , check=SCREAMING_SNAKE_CASE__ , ) assert hasattr(self , '''env''' ) def __A ( self : List[str] , SCREAMING_SNAKE_CASE__ : List[Any] ) -> Optional[Any]: # configuration for running training on smdistributed Model Parallel __lowerCamelCase = { '''enabled''': True, '''processes_per_host''': 8, } __lowerCamelCase = { '''enabled''': True, '''parameters''': { '''microbatches''': 4, '''placement_strategy''': '''spread''', '''pipeline''': '''interleaved''', '''optimize''': '''speed''', '''partitions''': 4, '''ddp''': True, }, } __lowerCamelCase = {'''smdistributed''': {'''modelparallel''': smp_options}, '''mpi''': mpi_options} __lowerCamelCase = '''trainer''' if self.script == '''run_glue.py''' else '''smtrainer''' # creates estimator return HuggingFace( entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f'''{self.env.base_job_name}-{instance_count}-smp-{name_extension}''' , instance_count=SCREAMING_SNAKE_CASE__ , instance_type=self.instance_type , debugger_hook_config=SCREAMING_SNAKE_CASE__ , hyperparameters={ **self.env.hyperparameters, '''model_name_or_path''': self.model_name_or_path, '''max_steps''': 5_00, } , metric_definitions=self.env.metric_definitions , distribution=SCREAMING_SNAKE_CASE__ , py_version='''py36''' , ) def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : List[Any] ) -> List[str]: TrainingJobAnalytics(SCREAMING_SNAKE_CASE__ ).export_csv(f'''{self.env.test_path}/{job_name}_metrics.csv''' ) @parameterized.expand([(1,)] ) def __A ( self : Optional[Any] , SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Optional[int]: # create estimator __lowerCamelCase = self.create_estimator(SCREAMING_SNAKE_CASE__ ) # run training estimator.fit() # result dataframe __lowerCamelCase = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe() # extract kpis __lowerCamelCase = list(result_metrics_df[result_metrics_df.metric_name == '''eval_accuracy''']['''value'''] ) __lowerCamelCase = list(result_metrics_df[result_metrics_df.metric_name == '''eval_loss''']['''value'''] ) # get train time from SageMaker job, this includes starting, preprocessing, stopping __lowerCamelCase = ( Session().describe_training_job(estimator.latest_training_job.name ).get('''TrainingTimeInSeconds''' , 99_99_99 ) ) # assert kpis assert train_runtime <= self.results["train_runtime"] assert all(t >= self.results['''eval_accuracy'''] for t in eval_accuracy ) assert all(t <= self.results['''eval_loss'''] for t in eval_loss ) # dump tests result into json file to share in PR with open(f'''{estimator.latest_training_job.name}.json''' , '''w''' ) as outfile: json.dump({'''train_time''': train_runtime, '''eval_accuracy''': eval_accuracy, '''eval_loss''': eval_loss} , SCREAMING_SNAKE_CASE__ )
270
0
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, )
305
from __future__ import annotations from fractions import Fraction def __magic_name__ ( __lowerCAmelCase : int , __lowerCAmelCase : int ) -> bool: return ( num != den and num % 10 == den // 10 and (num // 10) / (den % 10) == num / den ) def __magic_name__ ( __lowerCAmelCase : int ) -> list[str]: __lowerCamelCase = [] __lowerCamelCase = 11 __lowerCamelCase = int('''1''' + '''0''' * digit_len ) for num in range(__lowerCAmelCase , __lowerCAmelCase ): while den <= 99: if (num != den) and (num % 10 == den // 10) and (den % 10 != 0): if is_digit_cancelling(__lowerCAmelCase , __lowerCAmelCase ): solutions.append(f'''{num}/{den}''' ) den += 1 num += 1 __lowerCamelCase = 10 return solutions def __magic_name__ ( __lowerCAmelCase : int = 2 ) -> int: __lowerCamelCase = 1.0 for fraction in fraction_list(__lowerCAmelCase ): __lowerCamelCase = Fraction(__lowerCAmelCase ) result *= frac.denominator / frac.numerator return int(__lowerCAmelCase ) if __name__ == "__main__": print(solution())
270
0
'''simple docstring''' from collections import OrderedDict from typing import TYPE_CHECKING, Any, List, Mapping, Optional, Union from ...configuration_utils import PretrainedConfig from ...onnx import OnnxConfig from ...utils import TensorType, logging if TYPE_CHECKING: from ...onnx.config import PatchingSpec from ...tokenization_utils_base import PreTrainedTokenizerBase UpperCAmelCase_ = logging.get_logger(__name__) UpperCAmelCase_ = { "allenai/longformer-base-4096": "https://huggingface.co/allenai/longformer-base-4096/resolve/main/config.json", "allenai/longformer-large-4096": "https://huggingface.co/allenai/longformer-large-4096/resolve/main/config.json", "allenai/longformer-large-4096-finetuned-triviaqa": ( "https://huggingface.co/allenai/longformer-large-4096-finetuned-triviaqa/resolve/main/config.json" ), "allenai/longformer-base-4096-extra.pos.embd.only": ( "https://huggingface.co/allenai/longformer-base-4096-extra.pos.embd.only/resolve/main/config.json" ), "allenai/longformer-large-4096-extra.pos.embd.only": ( "https://huggingface.co/allenai/longformer-large-4096-extra.pos.embd.only/resolve/main/config.json" ), } class lowerCAmelCase_ ( __lowercase ): '''simple docstring''' lowerCAmelCase_ : List[str] = """longformer""" def __init__( self : Tuple , _UpperCAmelCase : Union[List[int], int] = 5_12 , _UpperCAmelCase : int = 2 , _UpperCAmelCase : int = 1 , _UpperCAmelCase : int = 0 , _UpperCAmelCase : int = 2 , _UpperCAmelCase : int = 3_05_22 , _UpperCAmelCase : int = 7_68 , _UpperCAmelCase : int = 12 , _UpperCAmelCase : int = 12 , _UpperCAmelCase : int = 30_72 , _UpperCAmelCase : str = "gelu" , _UpperCAmelCase : float = 0.1 , _UpperCAmelCase : float = 0.1 , _UpperCAmelCase : int = 5_12 , _UpperCAmelCase : int = 2 , _UpperCAmelCase : float = 0.02 , _UpperCAmelCase : float = 1E-12 , _UpperCAmelCase : bool = False , **_UpperCAmelCase : Dict , ): """simple docstring""" super().__init__(pad_token_id=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) UpperCAmelCase__ = attention_window UpperCAmelCase__ = sep_token_id UpperCAmelCase__ = bos_token_id UpperCAmelCase__ = eos_token_id UpperCAmelCase__ = vocab_size UpperCAmelCase__ = hidden_size UpperCAmelCase__ = num_hidden_layers UpperCAmelCase__ = num_attention_heads UpperCAmelCase__ = hidden_act UpperCAmelCase__ = intermediate_size UpperCAmelCase__ = hidden_dropout_prob UpperCAmelCase__ = attention_probs_dropout_prob UpperCAmelCase__ = max_position_embeddings UpperCAmelCase__ = type_vocab_size UpperCAmelCase__ = initializer_range UpperCAmelCase__ = layer_norm_eps UpperCAmelCase__ = onnx_export class lowerCAmelCase_ ( __lowercase ): '''simple docstring''' def __init__( self : int , _UpperCAmelCase : "PretrainedConfig" , _UpperCAmelCase : str = "default" , _UpperCAmelCase : "List[PatchingSpec]" = None ): """simple docstring""" super().__init__(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) UpperCAmelCase__ = True @property def SCREAMING_SNAKE_CASE__ ( self : List[str] ): """simple docstring""" if self.task == "multiple-choice": UpperCAmelCase__ = {0: """batch""", 1: """choice""", 2: """sequence"""} else: UpperCAmelCase__ = {0: """batch""", 1: """sequence"""} return OrderedDict( [ ("""input_ids""", dynamic_axis), ("""attention_mask""", dynamic_axis), ("""global_attention_mask""", dynamic_axis), ] ) @property def SCREAMING_SNAKE_CASE__ ( self : Dict ): """simple docstring""" UpperCAmelCase__ = super().outputs if self.task == "default": UpperCAmelCase__ = {0: """batch"""} return outputs @property def SCREAMING_SNAKE_CASE__ ( self : int ): """simple docstring""" return 1E-4 @property def SCREAMING_SNAKE_CASE__ ( self : Optional[Any] ): """simple docstring""" return max(super().default_onnx_opset , 14 ) def SCREAMING_SNAKE_CASE__ ( self : Union[str, Any] , _UpperCAmelCase : "PreTrainedTokenizerBase" , _UpperCAmelCase : int = -1 , _UpperCAmelCase : int = -1 , _UpperCAmelCase : bool = False , _UpperCAmelCase : Optional[TensorType] = None , ): """simple docstring""" UpperCAmelCase__ = super().generate_dummy_inputs( preprocessor=SCREAMING_SNAKE_CASE__ , batch_size=SCREAMING_SNAKE_CASE__ , seq_length=SCREAMING_SNAKE_CASE__ , is_pair=SCREAMING_SNAKE_CASE__ , framework=SCREAMING_SNAKE_CASE__ ) import torch # for some reason, replacing this code by inputs["global_attention_mask"] = torch.randint(2, inputs["input_ids"].shape, dtype=torch.int64) # makes the export fail randomly UpperCAmelCase__ = torch.zeros_like(inputs["""input_ids"""] ) # make every second token global UpperCAmelCase__ = 1 return inputs
346
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_tf_available, is_torch_available, is_vision_available, ) SCREAMING_SNAKE_CASE__ : List[str] = { "configuration_blip": [ "BLIP_PRETRAINED_CONFIG_ARCHIVE_MAP", "BlipConfig", "BlipTextConfig", "BlipVisionConfig", ], "processing_blip": ["BlipProcessor"], } try: if not is_vision_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : str = ["BlipImageProcessor"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : int = [ "BLIP_PRETRAINED_MODEL_ARCHIVE_LIST", "BlipModel", "BlipPreTrainedModel", "BlipForConditionalGeneration", "BlipForQuestionAnswering", "BlipVisionModel", "BlipTextModel", "BlipForImageTextRetrieval", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : Any = [ "TF_BLIP_PRETRAINED_MODEL_ARCHIVE_LIST", "TFBlipModel", "TFBlipPreTrainedModel", "TFBlipForConditionalGeneration", "TFBlipForQuestionAnswering", "TFBlipVisionModel", "TFBlipTextModel", "TFBlipForImageTextRetrieval", ] if TYPE_CHECKING: from .configuration_blip import BLIP_PRETRAINED_CONFIG_ARCHIVE_MAP, BlipConfig, BlipTextConfig, BlipVisionConfig from .processing_blip import BlipProcessor try: if not is_vision_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .image_processing_blip import BlipImageProcessor try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_blip import ( BLIP_PRETRAINED_MODEL_ARCHIVE_LIST, BlipForConditionalGeneration, BlipForImageTextRetrieval, BlipForQuestionAnswering, BlipModel, BlipPreTrainedModel, BlipTextModel, BlipVisionModel, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_blip import ( TF_BLIP_PRETRAINED_MODEL_ARCHIVE_LIST, TFBlipForConditionalGeneration, TFBlipForImageTextRetrieval, TFBlipForQuestionAnswering, TFBlipModel, TFBlipPreTrainedModel, TFBlipTextModel, TFBlipVisionModel, ) else: import sys SCREAMING_SNAKE_CASE__ : Tuple = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
270
0
def lowercase_ ( _A : str ): """simple docstring""" if not all(x.isalpha() for x in string ): raise ValueError("String must only contain alphabetic characters." ) lowerCamelCase__ : Any = sorted(string.lower() ) return len(__lowerCAmelCase ) == len(set(__lowerCAmelCase ) ) if __name__ == "__main__": A : Optional[Any] = input("Enter a string ").strip() A : Dict = is_isogram(input_str) print(f'{input_str} is {"an" if isogram else "not an"} isogram.')
184
from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available SCREAMING_SNAKE_CASE__ : Optional[Any] = {"configuration_wavlm": ["WAVLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "WavLMConfig"]} try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : int = [ "WAVLM_PRETRAINED_MODEL_ARCHIVE_LIST", "WavLMForAudioFrameClassification", "WavLMForCTC", "WavLMForSequenceClassification", "WavLMForXVector", "WavLMModel", "WavLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_wavlm import WAVLM_PRETRAINED_CONFIG_ARCHIVE_MAP, WavLMConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_wavlm import ( WAVLM_PRETRAINED_MODEL_ARCHIVE_LIST, WavLMForAudioFrameClassification, WavLMForCTC, WavLMForSequenceClassification, WavLMForXVector, WavLMModel, WavLMPreTrainedModel, ) else: import sys SCREAMING_SNAKE_CASE__ : int = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
270
0
"""simple docstring""" import argparse import io import requests import torch from omegaconf import OmegaConf from diffusers import AutoencoderKL from diffusers.pipelines.stable_diffusion.convert_from_ckpt import ( assign_to_checkpoint, conv_attn_to_linear, create_vae_diffusers_config, renew_vae_attention_paths, renew_vae_resnet_paths, ) def lowerCAmelCase_ ( snake_case_ : List[str] , snake_case_ : Optional[int] ) ->Any: lowerCamelCase__ : Optional[int] =checkpoint lowerCamelCase__ : List[Any] ={} lowerCamelCase__ : Tuple =vae_state_dict['encoder.conv_in.weight'] lowerCamelCase__ : Union[str, Any] =vae_state_dict['encoder.conv_in.bias'] lowerCamelCase__ : Tuple =vae_state_dict['encoder.conv_out.weight'] lowerCamelCase__ : List[Any] =vae_state_dict['encoder.conv_out.bias'] lowerCamelCase__ : Any =vae_state_dict['encoder.norm_out.weight'] lowerCamelCase__ : str =vae_state_dict['encoder.norm_out.bias'] lowerCamelCase__ : Optional[int] =vae_state_dict['decoder.conv_in.weight'] lowerCamelCase__ : int =vae_state_dict['decoder.conv_in.bias'] lowerCamelCase__ : List[Any] =vae_state_dict['decoder.conv_out.weight'] lowerCamelCase__ : Dict =vae_state_dict['decoder.conv_out.bias'] lowerCamelCase__ : List[str] =vae_state_dict['decoder.norm_out.weight'] lowerCamelCase__ : Optional[int] =vae_state_dict['decoder.norm_out.bias'] lowerCamelCase__ : Optional[Any] =vae_state_dict['quant_conv.weight'] lowerCamelCase__ : int =vae_state_dict['quant_conv.bias'] lowerCamelCase__ : str =vae_state_dict['post_quant_conv.weight'] lowerCamelCase__ : str =vae_state_dict['post_quant_conv.bias'] # Retrieves the keys for the encoder down blocks only lowerCamelCase__ : Optional[Any] =len({'.'.join(layer.split('.' )[:3] ) for layer in vae_state_dict if 'encoder.down' in layer} ) lowerCamelCase__ : Tuple ={ layer_id: [key for key in vae_state_dict if f"""down.{layer_id}""" in key] for layer_id in range(__lowerCAmelCase ) } # Retrieves the keys for the decoder up blocks only lowerCamelCase__ : int =len({'.'.join(layer.split('.' )[:3] ) for layer in vae_state_dict if 'decoder.up' in layer} ) lowerCamelCase__ : Optional[int] ={ layer_id: [key for key in vae_state_dict if f"""up.{layer_id}""" in key] for layer_id in range(__lowerCAmelCase ) } for i in range(__lowerCAmelCase ): lowerCamelCase__ : List[Any] =[key for key in down_blocks[i] if f"""down.{i}""" in key and f"""down.{i}.downsample""" not in key] if f"""encoder.down.{i}.downsample.conv.weight""" in vae_state_dict: lowerCamelCase__ : Any =vae_state_dict.pop( f"""encoder.down.{i}.downsample.conv.weight""" ) lowerCamelCase__ : Optional[Any] =vae_state_dict.pop( f"""encoder.down.{i}.downsample.conv.bias""" ) lowerCamelCase__ : Tuple =renew_vae_resnet_paths(__lowerCAmelCase ) lowerCamelCase__ : List[Any] ={'old': f"""down.{i}.block""", 'new': f"""down_blocks.{i}.resnets"""} assign_to_checkpoint(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , additional_replacements=[meta_path] , config=__lowerCAmelCase ) lowerCamelCase__ : Dict =[key for key in vae_state_dict if 'encoder.mid.block' in key] lowerCamelCase__ : Dict =2 for i in range(1 , num_mid_res_blocks + 1 ): lowerCamelCase__ : Optional[int] =[key for key in mid_resnets if f"""encoder.mid.block_{i}""" in key] lowerCamelCase__ : str =renew_vae_resnet_paths(__lowerCAmelCase ) lowerCamelCase__ : int ={'old': f"""mid.block_{i}""", 'new': f"""mid_block.resnets.{i - 1}"""} assign_to_checkpoint(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , additional_replacements=[meta_path] , config=__lowerCAmelCase ) lowerCamelCase__ : int =[key for key in vae_state_dict if 'encoder.mid.attn' in key] lowerCamelCase__ : int =renew_vae_attention_paths(__lowerCAmelCase ) lowerCamelCase__ : Any ={'old': 'mid.attn_1', 'new': 'mid_block.attentions.0'} assign_to_checkpoint(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , additional_replacements=[meta_path] , config=__lowerCAmelCase ) conv_attn_to_linear(__lowerCAmelCase ) for i in range(__lowerCAmelCase ): lowerCamelCase__ : Any =num_up_blocks - 1 - i lowerCamelCase__ : str =[ key for key in up_blocks[block_id] if f"""up.{block_id}""" in key and f"""up.{block_id}.upsample""" not in key ] if f"""decoder.up.{block_id}.upsample.conv.weight""" in vae_state_dict: lowerCamelCase__ : List[Any] =vae_state_dict[ f"""decoder.up.{block_id}.upsample.conv.weight""" ] lowerCamelCase__ : int =vae_state_dict[ f"""decoder.up.{block_id}.upsample.conv.bias""" ] lowerCamelCase__ : Any =renew_vae_resnet_paths(__lowerCAmelCase ) lowerCamelCase__ : List[Any] ={'old': f"""up.{block_id}.block""", 'new': f"""up_blocks.{i}.resnets"""} assign_to_checkpoint(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , additional_replacements=[meta_path] , config=__lowerCAmelCase ) lowerCamelCase__ : Dict =[key for key in vae_state_dict if 'decoder.mid.block' in key] lowerCamelCase__ : Union[str, Any] =2 for i in range(1 , num_mid_res_blocks + 1 ): lowerCamelCase__ : Optional[Any] =[key for key in mid_resnets if f"""decoder.mid.block_{i}""" in key] lowerCamelCase__ : Dict =renew_vae_resnet_paths(__lowerCAmelCase ) lowerCamelCase__ : List[str] ={'old': f"""mid.block_{i}""", 'new': f"""mid_block.resnets.{i - 1}"""} assign_to_checkpoint(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , additional_replacements=[meta_path] , config=__lowerCAmelCase ) lowerCamelCase__ : List[Any] =[key for key in vae_state_dict if 'decoder.mid.attn' in key] lowerCamelCase__ : Union[str, Any] =renew_vae_attention_paths(__lowerCAmelCase ) lowerCamelCase__ : List[str] ={'old': 'mid.attn_1', 'new': 'mid_block.attentions.0'} assign_to_checkpoint(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , additional_replacements=[meta_path] , config=__lowerCAmelCase ) conv_attn_to_linear(__lowerCAmelCase ) return new_checkpoint def lowerCAmelCase_ ( snake_case_ : str , snake_case_ : str , ) ->List[str]: # Only support V1 lowerCamelCase__ : int =requests.get( ' https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml' ) lowerCamelCase__ : Optional[Any] =io.BytesIO(r.content ) lowerCamelCase__ : Optional[int] =OmegaConf.load(__lowerCAmelCase ) lowerCamelCase__ : Any =5_1_2 lowerCamelCase__ : Optional[int] ='cuda' if torch.cuda.is_available() else 'cpu' if checkpoint_path.endswith('safetensors' ): from safetensors import safe_open lowerCamelCase__ : Tuple ={} with safe_open(__lowerCAmelCase , framework='pt' , device='cpu' ) as f: for key in f.keys(): lowerCamelCase__ : Any =f.get_tensor(__lowerCAmelCase ) else: lowerCamelCase__ : List[str] =torch.load(__lowerCAmelCase , map_location=__lowerCAmelCase )['state_dict'] # Convert the VAE model. lowerCamelCase__ : Any =create_vae_diffusers_config(__lowerCAmelCase , image_size=__lowerCAmelCase ) lowerCamelCase__ : str =custom_convert_ldm_vae_checkpoint(__lowerCAmelCase , __lowerCAmelCase ) lowerCamelCase__ : Dict =AutoencoderKL(**__lowerCAmelCase ) vae.load_state_dict(__lowerCAmelCase ) vae.save_pretrained(__lowerCAmelCase ) if __name__ == "__main__": lowerCAmelCase = argparse.ArgumentParser() parser.add_argument("""--vae_pt_path""", default=None, type=str, required=True, help="""Path to the VAE.pt to convert.""") parser.add_argument("""--dump_path""", default=None, type=str, required=True, help="""Path to the VAE.pt to convert.""") lowerCAmelCase = parser.parse_args() vae_pt_to_vae_diffuser(args.vae_pt_path, args.dump_path)
126
import pprint import requests SCREAMING_SNAKE_CASE__ : str = "https://zenquotes.io/api" def __magic_name__ ( ) -> list: return requests.get(API_ENDPOINT_URL + '''/today''' ).json() def __magic_name__ ( ) -> list: return requests.get(API_ENDPOINT_URL + '''/random''' ).json() if __name__ == "__main__": SCREAMING_SNAKE_CASE__ : int = random_quotes() pprint.pprint(response)
270
0
"""simple docstring""" from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available _lowercase : List[Any] = { "configuration_bridgetower": [ "BRIDGETOWER_PRETRAINED_CONFIG_ARCHIVE_MAP", "BridgeTowerConfig", "BridgeTowerTextConfig", "BridgeTowerVisionConfig", ], "processing_bridgetower": ["BridgeTowerProcessor"], } try: if not is_vision_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _lowercase : str = ["BridgeTowerImageProcessor"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _lowercase : Optional[Any] = [ "BRIDGETOWER_PRETRAINED_MODEL_ARCHIVE_LIST", "BridgeTowerForContrastiveLearning", "BridgeTowerForImageAndTextRetrieval", "BridgeTowerForMaskedLM", "BridgeTowerModel", "BridgeTowerPreTrainedModel", ] if TYPE_CHECKING: from .configuration_bridgetower import ( BRIDGETOWER_PRETRAINED_CONFIG_ARCHIVE_MAP, BridgeTowerConfig, BridgeTowerTextConfig, BridgeTowerVisionConfig, ) from .processing_bridgetower import BridgeTowerProcessor try: if not is_vision_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .image_processing_bridgetower import BridgeTowerImageProcessor try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_bridgetower import ( BRIDGETOWER_PRETRAINED_MODEL_ARCHIVE_LIST, BridgeTowerForContrastiveLearning, BridgeTowerForImageAndTextRetrieval, BridgeTowerForMaskedLM, BridgeTowerModel, BridgeTowerPreTrainedModel, ) else: import sys _lowercase : str = _LazyModule(__name__, globals()["__file__"], _import_structure)
238
from maths.is_square_free import is_square_free from maths.prime_factors import prime_factors def __magic_name__ ( __lowerCAmelCase : int ) -> int: __lowerCamelCase = prime_factors(__lowerCAmelCase ) if is_square_free(__lowerCAmelCase ): return -1 if len(__lowerCAmelCase ) % 2 else 1 return 0 if __name__ == "__main__": import doctest doctest.testmod()
270
0
# HF Trainer benchmarking tool # # This tool can be used to run and compare multiple dimensions of the HF Trainers args. # # It then prints a report once in github format with all the information that needs to be shared # with others and second time in a console-friendly format, so it's easier to use for tuning things up. # # The main idea is: # # ./trainer-benchmark.py --base-cmd '<cmd args that don't change>' \ # --variations '--tf32 0|--tf32 1' '--fp16 0|--fp16 1|--bf16 1' \ # --target-metric-key train_samples_per_second # # The variations can be any command line argument that you want to compare and not just dtype as in # the example. # # --variations allows you to compare variations in multiple dimensions. # # as the first dimention has 2 options and the second 3 in our example, this will run the trainer 6 # times adding one of: # # 1. --tf32 0 --fp16 0 # 2. --tf32 0 --fp16 1 # 3. --tf32 0 --bf16 1 # 4. --tf32 1 --fp16 0 # 5. --tf32 1 --fp16 1 # 6. --tf32 1 --bf16 1 # # and print the results. This is just a cartesian product - and more than 2 dimensions can be used. # # If you want to rely on defaults, this: # --variations '--tf32 0|--tf32 1' '--fp16 0|--fp16 1|--bf16 1' # is identical to this: # --variations '--tf32 0|--tf32 1' '|--fp16|--bf16' # # the leading empty variation in the 2nd dimension is a valid variation. # # So here we get the following 6 variations: # # 1. --tf32 0 # 2. --tf32 0 --fp16 # 3. --tf32 0 --bf16 # 4. --tf32 1 # 5. --tf32 1 --fp16 # 6. --tf32 1 --bf16 # # In this particular case we don't know what the default tf32 setting is as it's normally # pytorch-version dependent). That's why it's best to do an explicit setting of each variation: # `--tf32 0|--tf32 1` # # Here is a full example of a train: # # CUDA_VISIBLE_DEVICES=0 python ./scripts/benchmark/trainer-benchmark.py \ # --base-cmd \ # ' examples/pytorch/translation/run_translation.py --model_name_or_path t5-small \ # --output_dir output_dir --do_train --label_smoothing 0.1 --logging_strategy no \ # --save_strategy no --per_device_train_batch_size 32 --max_source_length 512 \ # --max_target_length 512 --num_train_epochs 1 --overwrite_output_dir \ # --source_lang en --target_lang ro --dataset_name wmt16 --dataset_config "ro-en" \ # --source_prefix "translate English to Romanian: " --warmup_steps 50 \ # --max_train_samples 20000 --dataloader_num_workers 2 ' \ # --target-metric-key train_samples_per_second --repeat-times 1 --variations \ # '|--fp16|--bf16' '--tf32 0|--tf32 1' --report-metric-keys train_loss \ # --repeat-times 1 --base-variation '--tf32 0' # # and here is a possible output: # # # | Variation | Train | Diff | Train | # | | samples | % | loss | # | | per | | | # | | second | | | # |:----------------|----------:|-------:|--------:| # | --tf32 0 | 285.11 | 0 | 2.51 | # | --tf32 1 | 342.09 | 20 | 2.51 | # | --fp16 --tf32 0 | 423.49 | 49 | 2.51 | # | --fp16 --tf32 1 | 423.13 | 48 | 2.51 | # | --bf16 --tf32 0 | 416.80 | 46 | 2.52 | # | --bf16 --tf32 1 | 415.87 | 46 | 2.52 | # # # So you can quickly compare the different outcomes. # # Typically running each experiment once is enough, but if the environment is unstable you can # re-run each multiple times, e.g., 3 using --repeat-times 3 and it will report the averaged results. # # By default it'll use the lowest result as the base line to use as 100% and then compare the rest to # it as can be seen from the table above, but you can also specify which combination is the one to use as # the baseline, e.g., to change to another entry use: --base-variation '--tf32 1 --fp16 0' # # --target-metric-key is there to tell the program which metrics to compare - the different metric keys are # inside output_dir/all_results.json. e.g., to measure eval performance instead of train use: # --target-metric-key eval_samples_per_second # but of course you will need to adjust the --base-cmd value in the example to perform evaluation as # well (as currently it doesn't) # import argparse import datetime import io import itertools import json import math import os import platform import re import shlex import subprocess import sys from pathlib import Path from statistics import fmean import pandas as pd import torch from tqdm import tqdm import transformers lowercase__ : Union[str, Any] = float("nan") class UpperCAmelCase : '''simple docstring''' def __init__( self : Optional[Any] , __lowercase : str ): """simple docstring""" snake_case_ = sys.stdout snake_case_ = open(SCREAMING_SNAKE_CASE__ , "a" ) def __getattr__( self : Dict , __lowercase : Any ): """simple docstring""" return getattr(self.stdout , SCREAMING_SNAKE_CASE__ ) def snake_case__ ( self : Optional[Any] , __lowercase : Dict ): """simple docstring""" self.stdout.write(SCREAMING_SNAKE_CASE__ ) # strip tqdm codes self.file.write(re.sub(r"^.*\r" , "" , SCREAMING_SNAKE_CASE__ , 0 , re.M ) ) def lowerCamelCase__ ( _A=80 , _A=False ): '''simple docstring''' snake_case_ = [] # deal with critical env vars snake_case_ = ["CUDA_VISIBLE_DEVICES"] for key in env_keys: snake_case_ = os.environ.get(__lowerCAmelCase , __lowerCAmelCase ) if val is not None: cmd.append(f"{key}={val}" ) # python executable (not always needed if the script is executable) snake_case_ = sys.executable if full_python_path else sys.executable.split("/" )[-1] cmd.append(__lowerCAmelCase ) # now the normal args cmd += list(map(shlex.quote , sys.argv ) ) # split up into up to MAX_WIDTH lines with shell multi-line escapes snake_case_ = [] snake_case_ = "" while len(__lowerCAmelCase ) > 0: current_line += f"{cmd.pop(0 )} " if len(__lowerCAmelCase ) == 0 or len(__lowerCAmelCase ) + len(cmd[0] ) + 1 > max_width - 1: lines.append(__lowerCAmelCase ) snake_case_ = "" return "\\\n".join(__lowerCAmelCase ) def lowerCamelCase__ ( _A , _A ): '''simple docstring''' snake_case_ = re.sub(R"[\\\n]+" , " " , args.base_cmd ) # remove --output_dir if any and set our own snake_case_ = re.sub("--output_dir\s+[^\s]+" , "" , args.base_cmd ) args.base_cmd += f" --output_dir {output_dir}" # ensure we have --overwrite_output_dir snake_case_ = re.sub("--overwrite_output_dir\s+" , "" , args.base_cmd ) args.base_cmd += " --overwrite_output_dir" return [sys.executable] + shlex.split(args.base_cmd ) def lowerCamelCase__ ( _A , _A , _A , _A , _A , _A , _A ): '''simple docstring''' if 0: import random from time import sleep sleep(0 ) return dict( {k: random.uniform(0 , 100 ) for k in metric_keys} , **{target_metric_key: random.choice([nan, 10.31, 1_00.2, 55.66_66, 2_22.22_22_22_22] )} , ) snake_case_ = subprocess.run(__lowerCAmelCase , capture_output=__lowerCAmelCase , text=__lowerCAmelCase ) if verbose: print("STDOUT" , result.stdout ) print("STDERR" , result.stderr ) # save the streams snake_case_ = variation.replace(" " , "-" ) with open(Path(__lowerCAmelCase ) / f"log.{prefix}.stdout.txt" , "w" ) as f: f.write(result.stdout ) with open(Path(__lowerCAmelCase ) / f"log.{prefix}.stderr.txt" , "w" ) as f: f.write(result.stderr ) if result.returncode != 0: if verbose: print("failed" ) return {target_metric_key: nan} with io.open(f"{output_dir}/all_results.json" , "r" , encoding="utf-8" ) as f: snake_case_ = json.load(__lowerCAmelCase ) # filter out just the keys we want return {k: v for k, v in metrics.items() if k in metric_keys} def lowerCamelCase__ ( _A , _A , _A , _A , _A , _A , _A , _A , _A , _A , ): '''simple docstring''' snake_case_ = [] snake_case_ = [] snake_case_ = f"{id}: {variation:<{longest_variation_len}}" snake_case_ = f"{preamble}: " snake_case_ = set(report_metric_keys + [target_metric_key] ) for i in tqdm(range(__lowerCAmelCase ) , desc=__lowerCAmelCase , leave=__lowerCAmelCase ): snake_case_ = process_run_single( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) snake_case_ = single_run_metrics[target_metric_key] if not math.isnan(__lowerCAmelCase ): metrics.append(__lowerCAmelCase ) results.append(__lowerCAmelCase ) outcome += "✓" else: outcome += "✘" snake_case_ = f"\33[2K\r{outcome}" if len(__lowerCAmelCase ) > 0: snake_case_ = {k: fmean([x[k] for x in metrics] ) for k in metrics[0].keys()} snake_case_ = round(mean_metrics[target_metric_key] , 2 ) snake_case_ = f"{outcome} {mean_target}" if len(__lowerCAmelCase ) > 1: results_str += f" {tuple(round(__lowerCAmelCase , 2 ) for x in results )}" print(__lowerCAmelCase ) snake_case_ = variation return mean_metrics else: print(__lowerCAmelCase ) return {variation_key: variation, target_metric_key: nan} def lowerCamelCase__ ( ): '''simple docstring''' snake_case_ = torch.cuda.get_device_properties(torch.device("cuda" ) ) return f"\nDatetime : {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S' )}\n\nSoftware:\ntransformers: {transformers.__version__}\ntorch : {torch.__version__}\ncuda : {torch.version.cuda}\npython : {platform.python_version()}\n\nHardware:\n{torch.cuda.device_count()} GPUs : {properties.name}, {properties.total_memory/2**30:0.2f}GB\n" def lowerCamelCase__ ( _A , _A , _A , _A , _A ): '''simple docstring''' snake_case_ = pd.DataFrame(__lowerCAmelCase ) snake_case_ = "variation" snake_case_ = "diff_%" snake_case_ = nan if base_variation is not None and len(df[df[variation_key] == base_variation] ): # this may still return nan snake_case_ = df.loc[df[variation_key] == base_variation][target_metric_key].item() if math.isnan(__lowerCAmelCase ): # as a fallback, use the minimal value as the sentinel snake_case_ = df.loc[df[target_metric_key] != nan][target_metric_key].min() # create diff column if possible if not math.isnan(__lowerCAmelCase ): snake_case_ = df.apply( lambda _A : round(100 * (r[target_metric_key] - sentinel_value) / sentinel_value ) if not math.isnan(r[target_metric_key] ) else 0 , axis="columns" , ) # re-order columns snake_case_ = [variation_key, target_metric_key, diff_key, *report_metric_keys] snake_case_ = df.reindex(__lowerCAmelCase , axis="columns" ) # reorder cols # capitalize snake_case_ = df.rename(str.capitalize , axis="columns" ) # make the cols as narrow as possible snake_case_ = df.rename(lambda _A : c.replace("_" , "<br>" ) , axis="columns" ) snake_case_ = df.rename(lambda _A : c.replace("_" , "\n" ) , axis="columns" ) snake_case_ = ["", "Copy between the cut-here-lines and paste as is to github or a forum"] report += ["----------8<-----------------8<--------"] report += ["*** Results:", df_github.to_markdown(index=__lowerCAmelCase , floatfmt=".2f" )] report += ["```"] report += ["*** Setup:", get_versions()] report += ["*** The benchmark command line was:", get_original_command()] report += ["```"] report += ["----------8<-----------------8<--------"] report += ["*** Results (console):", df_console.to_markdown(index=__lowerCAmelCase , floatfmt=".2f" )] print("\n\n".join(__lowerCAmelCase ) ) def lowerCamelCase__ ( ): '''simple docstring''' snake_case_ = argparse.ArgumentParser() parser.add_argument( "--base-cmd" , default=__lowerCAmelCase , type=__lowerCAmelCase , required=__lowerCAmelCase , help="Base cmd" , ) parser.add_argument( "--variations" , default=__lowerCAmelCase , type=__lowerCAmelCase , nargs="+" , required=__lowerCAmelCase , help="Multi-dimensional variations, example: \'|--fp16|--bf16\' \'|--tf32\'" , ) parser.add_argument( "--base-variation" , default=__lowerCAmelCase , type=__lowerCAmelCase , help="Baseline variation to compare to. if None the minimal target value will be used to compare against" , ) parser.add_argument( "--target-metric-key" , default=__lowerCAmelCase , type=__lowerCAmelCase , required=__lowerCAmelCase , help="Target metric key in output_dir/all_results.json, e.g., train_samples_per_second" , ) parser.add_argument( "--report-metric-keys" , default="" , type=__lowerCAmelCase , help="Report metric keys - other metric keys from output_dir/all_results.json to report, e.g., train_loss. Use a single argument e.g., \'train_loss train_samples" , ) parser.add_argument( "--repeat-times" , default=1 , type=__lowerCAmelCase , help="How many times to re-run each variation - an average will be reported" , ) parser.add_argument( "--output_dir" , default="output_benchmark" , type=__lowerCAmelCase , help="The output directory where all the benchmark reports will go to and additionally this directory will be used to override --output_dir in the script that is being benchmarked" , ) parser.add_argument( "--verbose" , default=__lowerCAmelCase , action="store_true" , help="Whether to show the outputs of each run or just the benchmark progress" , ) snake_case_ = parser.parse_args() snake_case_ = args.output_dir Path(__lowerCAmelCase ).mkdir(exist_ok=__lowerCAmelCase ) snake_case_ = get_base_command(__lowerCAmelCase , __lowerCAmelCase ) # split each dimension into its --foo variations snake_case_ = [list(map(str.strip , re.split(R"\|" , __lowerCAmelCase ) ) ) for x in args.variations] # build a cartesian product of dimensions and convert those back into cmd-line arg strings, # while stripping white space for inputs that were empty snake_case_ = list(map(str.strip , map(" ".join , itertools.product(*__lowerCAmelCase ) ) ) ) snake_case_ = max(len(__lowerCAmelCase ) for x in variations ) # split wanted keys snake_case_ = args.report_metric_keys.split() # capture prints into a log file for convenience snake_case_ = f"benchmark-report-{datetime.datetime.now().strftime('%Y-%m-%d-%H-%M-%S' )}.txt" print(f"\nNote: each run\'s output is also logged under {output_dir}/log.*.std*.txt" ) print(f"and this script\'s output is also piped into {report_fn}" ) snake_case_ = Tee(__lowerCAmelCase ) print(f"\n*** Running {len(__lowerCAmelCase )} benchmarks:" ) print(f"Base command: {' '.join(__lowerCAmelCase )}" ) snake_case_ = "variation" snake_case_ = [] for id, variation in enumerate(tqdm(__lowerCAmelCase , desc="Total completion: " , leave=__lowerCAmelCase ) ): snake_case_ = base_cmd + variation.split() results.append( process_run( id + 1 , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , args.target_metric_key , __lowerCAmelCase , args.repeat_times , __lowerCAmelCase , args.verbose , ) ) process_results(__lowerCAmelCase , args.target_metric_key , __lowerCAmelCase , args.base_variation , __lowerCAmelCase ) if __name__ == "__main__": main()
187
import argparse import collections import json import os import re import string import sys import numpy as np SCREAMING_SNAKE_CASE__ : Union[str, Any] = re.compile(r"\b(a|an|the)\b", re.UNICODE) SCREAMING_SNAKE_CASE__ : int = None def __magic_name__ ( ) -> str: __lowerCamelCase = argparse.ArgumentParser('''Official evaluation script for SQuAD version 2.0.''' ) parser.add_argument('''data_file''' , metavar='''data.json''' , help='''Input data JSON file.''' ) parser.add_argument('''pred_file''' , metavar='''pred.json''' , help='''Model predictions.''' ) parser.add_argument( '''--out-file''' , '''-o''' , metavar='''eval.json''' , help='''Write accuracy metrics to file (default is stdout).''' ) parser.add_argument( '''--na-prob-file''' , '''-n''' , metavar='''na_prob.json''' , help='''Model estimates of probability of no answer.''' ) parser.add_argument( '''--na-prob-thresh''' , '''-t''' , type=__lowerCAmelCase , default=1.0 , help='''Predict "" if no-answer probability exceeds this (default = 1.0).''' , ) parser.add_argument( '''--out-image-dir''' , '''-p''' , metavar='''out_images''' , default=__lowerCAmelCase , help='''Save precision-recall curves to directory.''' ) parser.add_argument('''--verbose''' , '''-v''' , action='''store_true''' ) if len(sys.argv ) == 1: parser.print_help() sys.exit(1 ) return parser.parse_args() def __magic_name__ ( __lowerCAmelCase : List[str] ) -> Union[str, Any]: __lowerCamelCase = {} for article in dataset: for p in article["paragraphs"]: for qa in p["qas"]: __lowerCamelCase = bool(qa['''answers''']['''text'''] ) return qid_to_has_ans def __magic_name__ ( __lowerCAmelCase : Dict ) -> Optional[Any]: def remove_articles(__lowerCAmelCase : Optional[int] ): return ARTICLES_REGEX.sub(''' ''' , __lowerCAmelCase ) def white_space_fix(__lowerCAmelCase : Optional[int] ): return " ".join(text.split() ) def remove_punc(__lowerCAmelCase : Union[str, Any] ): __lowerCamelCase = set(string.punctuation ) return "".join(ch for ch in text if ch not in exclude ) def lower(__lowerCAmelCase : Dict ): return text.lower() return white_space_fix(remove_articles(remove_punc(lower(__lowerCAmelCase ) ) ) ) def __magic_name__ ( __lowerCAmelCase : List[Any] ) -> Optional[int]: if not s: return [] return normalize_answer(__lowerCAmelCase ).split() def __magic_name__ ( __lowerCAmelCase : Tuple , __lowerCAmelCase : Tuple ) -> int: return int(normalize_answer(__lowerCAmelCase ) == normalize_answer(__lowerCAmelCase ) ) def __magic_name__ ( __lowerCAmelCase : Any , __lowerCAmelCase : Tuple ) -> str: __lowerCamelCase = get_tokens(__lowerCAmelCase ) __lowerCamelCase = get_tokens(__lowerCAmelCase ) __lowerCamelCase = collections.Counter(__lowerCAmelCase ) & collections.Counter(__lowerCAmelCase ) __lowerCamelCase = sum(common.values() ) if len(__lowerCAmelCase ) == 0 or len(__lowerCAmelCase ) == 0: # If either is no-answer, then F1 is 1 if they agree, 0 otherwise return int(gold_toks == pred_toks ) if num_same == 0: return 0 __lowerCamelCase = 1.0 * num_same / len(__lowerCAmelCase ) __lowerCamelCase = 1.0 * num_same / len(__lowerCAmelCase ) __lowerCamelCase = (2 * precision * recall) / (precision + recall) return fa def __magic_name__ ( __lowerCAmelCase : Dict , __lowerCAmelCase : List[str] ) -> Optional[Any]: __lowerCamelCase = {} __lowerCamelCase = {} for article in dataset: for p in article["paragraphs"]: for qa in p["qas"]: __lowerCamelCase = qa['''id'''] __lowerCamelCase = [t for t in qa['''answers''']['''text'''] if normalize_answer(__lowerCAmelCase )] if not gold_answers: # For unanswerable questions, only correct answer is empty string __lowerCamelCase = [''''''] if qid not in preds: print(f'''Missing prediction for {qid}''' ) continue __lowerCamelCase = preds[qid] # Take max over all gold answers __lowerCamelCase = max(compute_exact(__lowerCAmelCase , __lowerCAmelCase ) for a in gold_answers ) __lowerCamelCase = max(compute_fa(__lowerCAmelCase , __lowerCAmelCase ) for a in gold_answers ) return exact_scores, fa_scores def __magic_name__ ( __lowerCAmelCase : Tuple , __lowerCAmelCase : str , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : Union[str, Any] ) -> List[str]: __lowerCamelCase = {} for qid, s in scores.items(): __lowerCamelCase = na_probs[qid] > na_prob_thresh if pred_na: __lowerCamelCase = float(not qid_to_has_ans[qid] ) else: __lowerCamelCase = s return new_scores def __magic_name__ ( __lowerCAmelCase : List[Any] , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : Optional[Any]=None ) -> Union[str, Any]: if not qid_list: __lowerCamelCase = len(__lowerCAmelCase ) return collections.OrderedDict( [ ('''exact''', 100.0 * sum(exact_scores.values() ) / total), ('''f1''', 100.0 * sum(fa_scores.values() ) / total), ('''total''', total), ] ) else: __lowerCamelCase = len(__lowerCAmelCase ) return collections.OrderedDict( [ ('''exact''', 100.0 * sum(exact_scores[k] for k in qid_list ) / total), ('''f1''', 100.0 * sum(fa_scores[k] for k in qid_list ) / total), ('''total''', total), ] ) def __magic_name__ ( __lowerCAmelCase : Any , __lowerCAmelCase : str , __lowerCAmelCase : Optional[Any] ) -> int: for k in new_eval: __lowerCamelCase = new_eval[k] def __magic_name__ ( __lowerCAmelCase : Optional[int] , __lowerCAmelCase : List[Any] , __lowerCAmelCase : Dict , __lowerCAmelCase : Union[str, Any] ) -> Optional[Any]: plt.step(__lowerCAmelCase , __lowerCAmelCase , color='''b''' , alpha=0.2 , where='''post''' ) plt.fill_between(__lowerCAmelCase , __lowerCAmelCase , step='''post''' , alpha=0.2 , color='''b''' ) plt.xlabel('''Recall''' ) plt.ylabel('''Precision''' ) plt.xlim([0.0, 1.05] ) plt.ylim([0.0, 1.05] ) plt.title(__lowerCAmelCase ) plt.savefig(__lowerCAmelCase ) plt.clf() def __magic_name__ ( __lowerCAmelCase : Any , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : Tuple , __lowerCAmelCase : Tuple , __lowerCAmelCase : Optional[Any]=None , __lowerCAmelCase : Tuple=None ) -> int: __lowerCamelCase = sorted(__lowerCAmelCase , key=lambda __lowerCAmelCase : na_probs[k] ) __lowerCamelCase = 0.0 __lowerCamelCase = 1.0 __lowerCamelCase = 0.0 __lowerCamelCase = [1.0] __lowerCamelCase = [0.0] __lowerCamelCase = 0.0 for i, qid in enumerate(__lowerCAmelCase ): if qid_to_has_ans[qid]: true_pos += scores[qid] __lowerCamelCase = true_pos / float(i + 1 ) __lowerCamelCase = true_pos / float(__lowerCAmelCase ) if i == len(__lowerCAmelCase ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]: # i.e., if we can put a threshold after this point avg_prec += cur_p * (cur_r - recalls[-1]) precisions.append(__lowerCAmelCase ) recalls.append(__lowerCAmelCase ) if out_image: plot_pr_curve(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) return {"ap": 100.0 * avg_prec} def __magic_name__ ( __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : List[str] , __lowerCAmelCase : Tuple , __lowerCAmelCase : List[str] , __lowerCAmelCase : List[Any] , __lowerCAmelCase : List[Any] ) -> List[Any]: if out_image_dir and not os.path.exists(__lowerCAmelCase ): os.makedirs(__lowerCAmelCase ) __lowerCamelCase = sum(1 for v in qid_to_has_ans.values() if v ) if num_true_pos == 0: return __lowerCamelCase = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , '''pr_exact.png''' ) , title='''Precision-Recall curve for Exact Match score''' , ) __lowerCamelCase = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , '''pr_f1.png''' ) , title='''Precision-Recall curve for F1 score''' , ) __lowerCamelCase = {k: float(__lowerCAmelCase ) for k, v in qid_to_has_ans.items()} __lowerCamelCase = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , '''pr_oracle.png''' ) , title='''Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)''' , ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''pr_exact''' ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''pr_f1''' ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''pr_oracle''' ) def __magic_name__ ( __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : Any , __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : int ) -> Optional[Any]: if not qid_list: return __lowerCamelCase = [na_probs[k] for k in qid_list] __lowerCamelCase = np.ones_like(__lowerCAmelCase ) / float(len(__lowerCAmelCase ) ) plt.hist(__lowerCAmelCase , weights=__lowerCAmelCase , bins=20 , range=(0.0, 1.0) ) plt.xlabel('''Model probability of no-answer''' ) plt.ylabel('''Proportion of dataset''' ) plt.title(f'''Histogram of no-answer probability: {name}''' ) plt.savefig(os.path.join(__lowerCAmelCase , f'''na_prob_hist_{name}.png''' ) ) plt.clf() def __magic_name__ ( __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : List[str] , __lowerCAmelCase : Union[str, Any] ) -> Optional[int]: __lowerCamelCase = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] ) __lowerCamelCase = num_no_ans __lowerCamelCase = cur_score __lowerCamelCase = 0.0 __lowerCamelCase = sorted(__lowerCAmelCase , key=lambda __lowerCAmelCase : na_probs[k] ) for i, qid in enumerate(__lowerCAmelCase ): if qid not in scores: continue if qid_to_has_ans[qid]: __lowerCamelCase = scores[qid] else: if preds[qid]: __lowerCamelCase = -1 else: __lowerCamelCase = 0 cur_score += diff if cur_score > best_score: __lowerCamelCase = cur_score __lowerCamelCase = na_probs[qid] return 100.0 * best_score / len(__lowerCAmelCase ), best_thresh def __magic_name__ ( __lowerCAmelCase : Dict , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : int , __lowerCAmelCase : Union[str, Any] , __lowerCAmelCase : int , __lowerCAmelCase : Optional[Any] ) -> int: __lowerCamelCase , __lowerCamelCase = find_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) __lowerCamelCase , __lowerCamelCase = find_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) __lowerCamelCase = best_exact __lowerCamelCase = exact_thresh __lowerCamelCase = best_fa __lowerCamelCase = fa_thresh def __magic_name__ ( ) -> Optional[int]: with open(OPTS.data_file ) as f: __lowerCamelCase = json.load(__lowerCAmelCase ) __lowerCamelCase = dataset_json['''data'''] with open(OPTS.pred_file ) as f: __lowerCamelCase = json.load(__lowerCAmelCase ) if OPTS.na_prob_file: with open(OPTS.na_prob_file ) as f: __lowerCamelCase = json.load(__lowerCAmelCase ) else: __lowerCamelCase = {k: 0.0 for k in preds} __lowerCamelCase = make_qid_to_has_ans(__lowerCAmelCase ) # maps qid to True/False __lowerCamelCase = [k for k, v in qid_to_has_ans.items() if v] __lowerCamelCase = [k for k, v in qid_to_has_ans.items() if not v] __lowerCamelCase , __lowerCamelCase = get_raw_scores(__lowerCAmelCase , __lowerCAmelCase ) __lowerCamelCase = apply_no_ans_threshold(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.na_prob_thresh ) __lowerCamelCase = apply_no_ans_threshold(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.na_prob_thresh ) __lowerCamelCase = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase ) if has_ans_qids: __lowerCamelCase = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase , qid_list=__lowerCAmelCase ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''HasAns''' ) if no_ans_qids: __lowerCamelCase = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase , qid_list=__lowerCAmelCase ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''NoAns''' ) if OPTS.na_prob_file: find_all_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) if OPTS.na_prob_file and OPTS.out_image_dir: run_precision_recall_analysis(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir ) histogram_na_prob(__lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir , '''hasAns''' ) histogram_na_prob(__lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir , '''noAns''' ) if OPTS.out_file: with open(OPTS.out_file , '''w''' ) as f: json.dump(__lowerCAmelCase , __lowerCAmelCase ) else: print(json.dumps(__lowerCAmelCase , indent=2 ) ) if __name__ == "__main__": SCREAMING_SNAKE_CASE__ : Any = parse_args() if OPTS.out_image_dir: import matplotlib matplotlib.use("Agg") import matplotlib.pyplot as plt main()
270
0
import argparse import os from pathlib import Path import torch from bark.generation import _load_model as _bark_load_model from huggingface_hub import hf_hub_download from transformers import EncodecConfig, EncodecModel, set_seed from transformers.models.bark.configuration_bark import ( BarkCoarseConfig, BarkConfig, BarkFineConfig, BarkSemanticConfig, ) from transformers.models.bark.generation_configuration_bark import ( BarkCoarseGenerationConfig, BarkFineGenerationConfig, BarkGenerationConfig, BarkSemanticGenerationConfig, ) from transformers.models.bark.modeling_bark import BarkCoarseModel, BarkFineModel, BarkModel, BarkSemanticModel from transformers.utils import logging logging.set_verbosity_info() UpperCamelCase = logging.get_logger(__name__) set_seed(770) UpperCamelCase = { "c_attn": "att_proj", "c_proj": "out_proj", "c_fc": "in_proj", "transformer.": "", "h.": "layers.", "ln_1": "layernorm_1", "ln_2": "layernorm_2", "ln_f": "layernorm_final", "wpe": "position_embeds_layer", "wte": "input_embeds_layer", } UpperCamelCase = { "text_small": { "repo_id": "suno/bark", "file_name": "text.pt", }, "coarse_small": { "repo_id": "suno/bark", "file_name": "coarse.pt", }, "fine_small": { "repo_id": "suno/bark", "file_name": "fine.pt", }, "text": { "repo_id": "suno/bark", "file_name": "text_2.pt", }, "coarse": { "repo_id": "suno/bark", "file_name": "coarse_2.pt", }, "fine": { "repo_id": "suno/bark", "file_name": "fine_2.pt", }, } UpperCamelCase = os.path.dirname(os.path.abspath(__file__)) UpperCamelCase = os.path.join(os.path.expanduser('''~'''), '''.cache''') UpperCamelCase = os.path.join(os.getenv('''XDG_CACHE_HOME''', default_cache_dir), '''suno''', '''bark_v0''') def __lowerCamelCase ( snake_case__ ,snake_case__=False ) -> Optional[Any]: """simple docstring""" _SCREAMING_SNAKE_CASE = model_type if use_small: key += "_small" return os.path.join(__lowerCAmelCase ,REMOTE_MODEL_PATHS[key]["""file_name"""] ) def __lowerCamelCase ( snake_case__ ,snake_case__ ) -> Union[str, Any]: """simple docstring""" os.makedirs(__lowerCAmelCase ,exist_ok=__lowerCAmelCase ) hf_hub_download(repo_id=__lowerCAmelCase ,filename=__lowerCAmelCase ,local_dir=__lowerCAmelCase ) def __lowerCamelCase ( snake_case__ ,snake_case__ ,snake_case__=False ,snake_case__="text" ) -> List[str]: """simple docstring""" if model_type == "text": _SCREAMING_SNAKE_CASE = BarkSemanticModel _SCREAMING_SNAKE_CASE = BarkSemanticConfig _SCREAMING_SNAKE_CASE = BarkSemanticGenerationConfig elif model_type == "coarse": _SCREAMING_SNAKE_CASE = BarkCoarseModel _SCREAMING_SNAKE_CASE = BarkCoarseConfig _SCREAMING_SNAKE_CASE = BarkCoarseGenerationConfig elif model_type == "fine": _SCREAMING_SNAKE_CASE = BarkFineModel _SCREAMING_SNAKE_CASE = BarkFineConfig _SCREAMING_SNAKE_CASE = BarkFineGenerationConfig else: raise NotImplementedError() _SCREAMING_SNAKE_CASE = F'{model_type}_small' if use_small else model_type _SCREAMING_SNAKE_CASE = REMOTE_MODEL_PATHS[model_key] if not os.path.exists(__lowerCAmelCase ): logger.info(F'{model_type} model not found, downloading into `{CACHE_DIR}`.' ) _download(model_info["""repo_id"""] ,model_info["""file_name"""] ) _SCREAMING_SNAKE_CASE = torch.load(__lowerCAmelCase ,map_location=__lowerCAmelCase ) # this is a hack _SCREAMING_SNAKE_CASE = checkpoint["""model_args"""] if "input_vocab_size" not in model_args: _SCREAMING_SNAKE_CASE = model_args["""vocab_size"""] _SCREAMING_SNAKE_CASE = model_args["""vocab_size"""] del model_args["vocab_size"] # convert Bark model arguments to HF Bark model arguments _SCREAMING_SNAKE_CASE = model_args.pop("""n_head""" ) _SCREAMING_SNAKE_CASE = model_args.pop("""n_embd""" ) _SCREAMING_SNAKE_CASE = model_args.pop("""n_layer""" ) _SCREAMING_SNAKE_CASE = ConfigClass(**checkpoint["""model_args"""] ) _SCREAMING_SNAKE_CASE = ModelClass(config=__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = GenerationConfigClass() _SCREAMING_SNAKE_CASE = model_generation_config _SCREAMING_SNAKE_CASE = checkpoint["""model"""] # fixup checkpoint _SCREAMING_SNAKE_CASE = """_orig_mod.""" for k, v in list(state_dict.items() ): if k.startswith(__lowerCAmelCase ): # replace part of the key with corresponding layer name in HF implementation _SCREAMING_SNAKE_CASE = k[len(__lowerCAmelCase ) :] for old_layer_name in new_layer_name_dict: _SCREAMING_SNAKE_CASE = new_k.replace(__lowerCAmelCase ,new_layer_name_dict[old_layer_name] ) _SCREAMING_SNAKE_CASE = state_dict.pop(__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = set(state_dict.keys() ) - set(model.state_dict().keys() ) _SCREAMING_SNAKE_CASE = {k for k in extra_keys if not k.endswith(""".attn.bias""" )} _SCREAMING_SNAKE_CASE = set(model.state_dict().keys() ) - set(state_dict.keys() ) _SCREAMING_SNAKE_CASE = {k for k in missing_keys if not k.endswith(""".attn.bias""" )} if len(__lowerCAmelCase ) != 0: raise ValueError(F'extra keys found: {extra_keys}' ) if len(__lowerCAmelCase ) != 0: raise ValueError(F'missing keys: {missing_keys}' ) model.load_state_dict(__lowerCAmelCase ,strict=__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = model.num_parameters(exclude_embeddings=__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = checkpoint["""best_val_loss"""].item() logger.info(F'model loaded: {round(n_params/1e6 ,1 )}M params, {round(__lowerCAmelCase ,3 )} loss' ) model.eval() model.to(__lowerCAmelCase ) del checkpoint, state_dict return model def __lowerCamelCase ( snake_case__ ,snake_case__=False ,snake_case__="text" ) -> List[str]: """simple docstring""" if model_type not in ("text", "coarse", "fine"): raise NotImplementedError() _SCREAMING_SNAKE_CASE = """cpu""" # do conversion on cpu _SCREAMING_SNAKE_CASE = _get_ckpt_path(__lowerCAmelCase ,use_small=__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = _load_model(__lowerCAmelCase ,__lowerCAmelCase ,model_type=__lowerCAmelCase ,use_small=__lowerCAmelCase ) # load bark initial model _SCREAMING_SNAKE_CASE = _bark_load_model(__lowerCAmelCase ,"""cpu""" ,model_type=__lowerCAmelCase ,use_small=__lowerCAmelCase ) if model_type == "text": _SCREAMING_SNAKE_CASE = bark_model["""model"""] if model.num_parameters(exclude_embeddings=__lowerCAmelCase ) != bark_model.get_num_params(): raise ValueError("""initial and new models don\'t have the same number of parameters""" ) # check if same output as the bark model _SCREAMING_SNAKE_CASE = 5 _SCREAMING_SNAKE_CASE = 10 if model_type in ["text", "coarse"]: _SCREAMING_SNAKE_CASE = torch.randint(2_56 ,(batch_size, sequence_length) ,dtype=torch.int ) _SCREAMING_SNAKE_CASE = bark_model(__lowerCAmelCase )[0] _SCREAMING_SNAKE_CASE = model(__lowerCAmelCase ) # take last logits _SCREAMING_SNAKE_CASE = output_new_model_total.logits[:, [-1], :] else: _SCREAMING_SNAKE_CASE = 3 _SCREAMING_SNAKE_CASE = 8 _SCREAMING_SNAKE_CASE = torch.randint(2_56 ,(batch_size, sequence_length, n_codes_total) ,dtype=torch.int ) _SCREAMING_SNAKE_CASE = model(__lowerCAmelCase ,__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = bark_model(__lowerCAmelCase ,__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = output_new_model_total.logits # output difference should come from the difference of self-attention implementation design if output_new_model.shape != output_old_model.shape: raise ValueError("""initial and new outputs don\'t have the same shape""" ) if (output_new_model - output_old_model).abs().max().item() > 1e-3: raise ValueError("""initial and new outputs are not equal""" ) Path(__lowerCAmelCase ).mkdir(exist_ok=__lowerCAmelCase ) model.save_pretrained(__lowerCAmelCase ) def __lowerCamelCase ( snake_case__ ,snake_case__ ,snake_case__ ,snake_case__ ,snake_case__ ,snake_case__ ,) -> str: """simple docstring""" _SCREAMING_SNAKE_CASE = os.path.join(__lowerCAmelCase ,__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = BarkSemanticConfig.from_pretrained(os.path.join(__lowerCAmelCase ,"""config.json""" ) ) _SCREAMING_SNAKE_CASE = BarkCoarseConfig.from_pretrained(os.path.join(__lowerCAmelCase ,"""config.json""" ) ) _SCREAMING_SNAKE_CASE = BarkFineConfig.from_pretrained(os.path.join(__lowerCAmelCase ,"""config.json""" ) ) _SCREAMING_SNAKE_CASE = EncodecConfig.from_pretrained("""facebook/encodec_24khz""" ) _SCREAMING_SNAKE_CASE = BarkSemanticModel.from_pretrained(__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = BarkCoarseModel.from_pretrained(__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = BarkFineModel.from_pretrained(__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = EncodecModel.from_pretrained("""facebook/encodec_24khz""" ) _SCREAMING_SNAKE_CASE = BarkConfig.from_sub_model_configs( __lowerCAmelCase ,__lowerCAmelCase ,__lowerCAmelCase ,__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = BarkGenerationConfig.from_sub_model_configs( semantic.generation_config ,coarseAcoustic.generation_config ,fineAcoustic.generation_config ) _SCREAMING_SNAKE_CASE = BarkModel(__lowerCAmelCase ) _SCREAMING_SNAKE_CASE = semantic _SCREAMING_SNAKE_CASE = coarseAcoustic _SCREAMING_SNAKE_CASE = fineAcoustic _SCREAMING_SNAKE_CASE = codec _SCREAMING_SNAKE_CASE = bark_generation_config Path(__lowerCAmelCase ).mkdir(exist_ok=__lowerCAmelCase ) bark.save_pretrained(__lowerCAmelCase ,repo_id=__lowerCAmelCase ,push_to_hub=__lowerCAmelCase ) if __name__ == "__main__": UpperCamelCase = argparse.ArgumentParser() # Required parameters parser.add_argument('''model_type''', type=str, help='''text, coarse or fine.''') parser.add_argument('''pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model.''') parser.add_argument('''--is_small''', action='''store_true''', help='''convert the small version instead of the large.''') UpperCamelCase = parser.parse_args() load_model(args.pytorch_dump_folder_path, model_type=args.model_type, use_small=args.is_small)
306
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 SCREAMING_SNAKE_CASE__ : Optional[int] = os.path.abspath(os.path.dirname(__file__)) with ExtendSysPath(F'{bindir}/../../examples/pytorch/translation'): from run_translation import main # noqa set_seed(42) SCREAMING_SNAKE_CASE__ : Any = "sshleifer/student_marian_en_ro_6_1" SCREAMING_SNAKE_CASE__ : Tuple = "sshleifer/tiny-mbart" @require_torch class lowerCAmelCase__ ( __lowercase ): def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : Tuple=False , SCREAMING_SNAKE_CASE__ : Optional[Any]=None , SCREAMING_SNAKE_CASE__ : Optional[Any]=True , SCREAMING_SNAKE_CASE__ : str=True , SCREAMING_SNAKE_CASE__ : Optional[Any]=True , SCREAMING_SNAKE_CASE__ : Optional[int]=True , ) -> Optional[int]: __lowerCamelCase = self.run_trainer( eval_steps=1 , max_len=12 , model_name=SCREAMING_SNAKE_CASE__ , num_train_epochs=1 , distributed=SCREAMING_SNAKE_CASE__ , extra_args_str=SCREAMING_SNAKE_CASE__ , predict_with_generate=SCREAMING_SNAKE_CASE__ , do_train=SCREAMING_SNAKE_CASE__ , do_eval=SCREAMING_SNAKE_CASE__ , do_predict=SCREAMING_SNAKE_CASE__ , ) __lowerCamelCase = TrainerState.load_from_json(os.path.join(SCREAMING_SNAKE_CASE__ , '''trainer_state.json''' ) ).log_history if not do_eval: return __lowerCamelCase = [log for log in logs if '''eval_loss''' in log.keys()] __lowerCamelCase = eval_metrics[0] if predict_with_generate: assert "eval_bleu" in first_step_stats __lowerCamelCase = eval_metrics[-1] assert isinstance(last_step_stats['''eval_bleu'''] , SCREAMING_SNAKE_CASE__ ) assert not math.isnan(float(last_step_stats['''eval_loss'''] ) ), "eval_loss must not be `nan`" @require_torch_non_multi_gpu def __A ( self : Optional[int] ) -> int: self.run_seqaseq_quick() @require_torch_multi_gpu def __A ( self : int ) -> List[str]: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ ) @require_torch_multi_gpu def __A ( self : Optional[Any] ) -> Tuple: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ ) @unittest.skip('''Requires an update of the env running those tests''' ) @require_torch_multi_gpu @require_fairscale def __A ( self : Dict ) -> Tuple: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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 : Optional[int] ) -> List[str]: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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 ) -> Any: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=SCREAMING_SNAKE_CASE__ ) @unittest.skip('''Requires an update of the env running those tests''' ) @require_torch_multi_gpu @require_fairscale def __A ( self : Dict ) -> Tuple: self.run_seqaseq_quick( distributed=SCREAMING_SNAKE_CASE__ , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=SCREAMING_SNAKE_CASE__ ) @require_apex @require_torch_gpu def __A ( self : Union[str, Any] ) -> List[str]: # XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same # program and it breaks other tests that run from the same pytest worker, therefore until this is # sorted out it must be run only in an external program, that is distributed=True in this # test and only under one or more gpus - if we want cpu will need to make a special test # # specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via # 2nd main() call it botches the future eval. # self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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=SCREAMING_SNAKE_CASE__ , extra_args_str='''--fp16 --fp16_backend=apex''' ) @parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''] ) @require_torch_multi_gpu def __A ( self : Any , SCREAMING_SNAKE_CASE__ : List[str] ) -> Optional[Any]: # as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout __lowerCamelCase = { # 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}, } __lowerCamelCase = experiments[experiment_id] __lowerCamelCase = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False} __lowerCamelCase = '''Running training''' with CaptureStderr() as cl: self.run_seqaseq_quick(**SCREAMING_SNAKE_CASE__ , extra_args_str=data['''extra_args_str'''] ) __lowerCamelCase = len(re.findall(SCREAMING_SNAKE_CASE__ , cl.err ) ) self.assertEqual(SCREAMING_SNAKE_CASE__ , data['''n_matches'''] ) @slow def __A ( self : Any ) -> Optional[Any]: __lowerCamelCase = self.run_trainer( eval_steps=2 , max_len=1_28 , model_name=SCREAMING_SNAKE_CASE__ , learning_rate=3e-4 , num_train_epochs=10 , distributed=SCREAMING_SNAKE_CASE__ , ) # Check metrics __lowerCamelCase = TrainerState.load_from_json(os.path.join(SCREAMING_SNAKE_CASE__ , '''trainer_state.json''' ) ).log_history __lowerCamelCase = [log for log in logs if '''eval_loss''' in log.keys()] __lowerCamelCase = eval_metrics[0] __lowerCamelCase = eval_metrics[-1] assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing" assert isinstance(last_step_stats['''eval_bleu'''] , SCREAMING_SNAKE_CASE__ ) # test if do_predict saves generations and metrics __lowerCamelCase = os.listdir(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = {os.path.basename(SCREAMING_SNAKE_CASE__ ) for p in contents} assert "generated_predictions.txt" in contents assert "predict_results.json" in contents @slow @require_bitsandbytes def __A ( self : Optional[int] ) -> str: from transformers.training_args import OptimizerNames def train_and_return_metrics(SCREAMING_SNAKE_CASE__ : str ) -> Tuple[int, float]: __lowerCamelCase = '''--skip_memory_metrics 0''' __lowerCamelCase = self.run_trainer( max_len=1_28 , model_name=SCREAMING_SNAKE_CASE__ , learning_rate=3e-4 , num_train_epochs=1 , optim=SCREAMING_SNAKE_CASE__ , distributed=SCREAMING_SNAKE_CASE__ , extra_args_str=SCREAMING_SNAKE_CASE__ , do_eval=SCREAMING_SNAKE_CASE__ , do_predict=SCREAMING_SNAKE_CASE__ , n_gpus_to_use=1 , ) # Check metrics __lowerCamelCase = TrainerState.load_from_json(Path(SCREAMING_SNAKE_CASE__ , '''trainer_state.json''' ) ).log_history __lowerCamelCase = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20 ) __lowerCamelCase = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20 ) __lowerCamelCase = logs[0]['''train_loss'''] return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss __lowerCamelCase , __lowerCamelCase , __lowerCamelCase = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value ) __lowerCamelCase , __lowerCamelCase , __lowerCamelCase = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value ) __lowerCamelCase = gpu_alloc_mem_orig - gpu_alloc_mem_bnb __lowerCamelCase = gpu_peak_mem_orig + gpu_alloc_mem_orig __lowerCamelCase = gpu_peak_mem_bnb + gpu_alloc_mem_bnb __lowerCamelCase = 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 __lowerCamelCase = 1_20 # 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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , '''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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , '''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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , f'''loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}''' ) def __A ( self : Dict , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : float = 3e-3 , SCREAMING_SNAKE_CASE__ : str = "adafactor" , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : str = None , SCREAMING_SNAKE_CASE__ : int = 0 , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : int = None , ) -> List[Any]: __lowerCamelCase = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro''' __lowerCamelCase = self.get_auto_remove_tmp_dir() __lowerCamelCase = f''' --model_name_or_path {model_name} --train_file {data_dir}/train.json --validation_file {data_dir}/val.json --test_file {data_dir}/test.json --output_dir {output_dir} --overwrite_output_dir --max_train_samples 8 --max_source_length {max_len} --max_target_length {max_len} --do_train --num_train_epochs {str(SCREAMING_SNAKE_CASE__ )} --per_device_train_batch_size 4 --learning_rate {learning_rate} --warmup_steps 8 --logging_steps 0 --logging_strategy no --save_steps {str(SCREAMING_SNAKE_CASE__ )} --group_by_length --label_smoothing_factor 0.1 --target_lang ro_RO --source_lang en_XX '''.split() __lowerCamelCase = f''' --do_eval --per_device_eval_batch_size 4 --max_eval_samples 8 --val_max_target_length {max_len} --evaluation_strategy steps --eval_steps {str(SCREAMING_SNAKE_CASE__ )} '''.split() __lowerCamelCase = ''' --do_predict '''.split() __lowerCamelCase = [] 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: __lowerCamelCase = get_gpu_count() __lowerCamelCase = get_torch_dist_unique_port() __lowerCamelCase = f''' -m torch.distributed.run --nproc_per_node={n_gpus_to_use} --master_port={master_port} {self.examples_dir_str}/pytorch/translation/run_translation.py '''.split() __lowerCamelCase = [sys.executable] + distributed_args + args # keep for quick debug # print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die execute_subprocess_async(SCREAMING_SNAKE_CASE__ , env=self.get_env() ) else: __lowerCamelCase = ['''run_translation.py'''] + args with patch.object(SCREAMING_SNAKE_CASE__ , '''argv''' , SCREAMING_SNAKE_CASE__ ): main() return output_dir
270
0
'''simple docstring''' import argparse import requests import torch from PIL import Image from transformers import SwinConfig, SwinForMaskedImageModeling, ViTImageProcessor def UpperCAmelCase_ ( __lowerCamelCase : Optional[Any] ): lowercase_ :Dict = SwinConfig(image_size=1_92 ) if "base" in model_name: lowercase_ :List[str] = 6 lowercase_ :List[str] = 1_28 lowercase_ :Optional[int] = (2, 2, 18, 2) lowercase_ :Union[str, Any] = (4, 8, 16, 32) elif "large" in model_name: lowercase_ :Union[str, Any] = 12 lowercase_ :int = 1_92 lowercase_ :Any = (2, 2, 18, 2) lowercase_ :Optional[int] = (6, 12, 24, 48) else: raise ValueError("Model not supported, only supports base and large variants" ) lowercase_ :Optional[Any] = window_size lowercase_ :Any = embed_dim lowercase_ :List[str] = depths lowercase_ :Optional[Any] = num_heads return config def UpperCAmelCase_ ( __lowerCamelCase : int ): if "encoder.mask_token" in name: lowercase_ :Tuple = name.replace("encoder.mask_token" ,"embeddings.mask_token" ) if "encoder.patch_embed.proj" in name: lowercase_ :List[str] = name.replace("encoder.patch_embed.proj" ,"embeddings.patch_embeddings.projection" ) if "encoder.patch_embed.norm" in name: lowercase_ :int = name.replace("encoder.patch_embed.norm" ,"embeddings.norm" ) if "attn.proj" in name: lowercase_ :Tuple = name.replace("attn.proj" ,"attention.output.dense" ) if "attn" in name: lowercase_ :Tuple = name.replace("attn" ,"attention.self" ) if "norm1" in name: lowercase_ :List[str] = name.replace("norm1" ,"layernorm_before" ) if "norm2" in name: lowercase_ :str = name.replace("norm2" ,"layernorm_after" ) if "mlp.fc1" in name: lowercase_ :Optional[Any] = name.replace("mlp.fc1" ,"intermediate.dense" ) if "mlp.fc2" in name: lowercase_ :List[str] = name.replace("mlp.fc2" ,"output.dense" ) if name == "encoder.norm.weight": lowercase_ :List[Any] = "layernorm.weight" if name == "encoder.norm.bias": lowercase_ :Dict = "layernorm.bias" if "decoder" in name: pass else: lowercase_ :Union[str, Any] = "swin." + name return name def UpperCAmelCase_ ( __lowerCamelCase : int ,__lowerCamelCase : Tuple ): for key in orig_state_dict.copy().keys(): lowercase_ :Union[str, Any] = orig_state_dict.pop(__lowerCAmelCase ) if "attn_mask" in key: pass elif "qkv" in key: lowercase_ :Dict = key.split("." ) lowercase_ :List[str] = int(key_split[2] ) lowercase_ :Dict = int(key_split[4] ) lowercase_ :Tuple = model.swin.encoder.layers[layer_num].blocks[block_num].attention.self.all_head_size if "weight" in key: lowercase_ :Dict = val[:dim, :] lowercase_ :str = val[ dim : dim * 2, : ] lowercase_ :Any = val[-dim:, :] else: lowercase_ :Tuple = val[ :dim ] lowercase_ :Optional[Any] = val[ dim : dim * 2 ] lowercase_ :List[str] = val[ -dim: ] else: lowercase_ :Optional[int] = val return orig_state_dict def UpperCAmelCase_ ( __lowerCamelCase : Optional[int] ,__lowerCamelCase : Optional[Any] ,__lowerCamelCase : Optional[Any] ,__lowerCamelCase : Tuple ): lowercase_ :Dict = torch.load(__lowerCAmelCase ,map_location="cpu" )["model"] lowercase_ :List[str] = get_swin_config(__lowerCAmelCase ) lowercase_ :int = SwinForMaskedImageModeling(__lowerCAmelCase ) model.eval() lowercase_ :List[Any] = convert_state_dict(__lowerCAmelCase ,__lowerCAmelCase ) model.load_state_dict(__lowerCAmelCase ) lowercase_ :Optional[Any] = "http://images.cocodataset.org/val2017/000000039769.jpg" lowercase_ :Union[str, Any] = ViTImageProcessor(size={"height": 1_92, "width": 1_92} ) lowercase_ :Optional[int] = Image.open(requests.get(__lowerCAmelCase ,stream=__lowerCAmelCase ).raw ) lowercase_ :List[str] = image_processor(images=__lowerCAmelCase ,return_tensors="pt" ) with torch.no_grad(): lowercase_ :Tuple = model(**__lowerCAmelCase ).logits print(outputs.keys() ) 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(__lowerCAmelCase ) print(F'Saving image processor to {pytorch_dump_folder_path}' ) image_processor.save_pretrained(__lowerCAmelCase ) if push_to_hub: print(F'Pushing model and image processor for {model_name} to hub' ) model.push_to_hub(F'microsoft/{model_name}' ) image_processor.push_to_hub(F'microsoft/{model_name}' ) if __name__ == "__main__": lowerCAmelCase : Tuple =argparse.ArgumentParser() # Required parameters parser.add_argument( '''--model_name''', default='''swin-base-simmim-window6-192''', type=str, choices=['''swin-base-simmim-window6-192''', '''swin-large-simmim-window12-192'''], help='''Name of the Swin SimMIM model you\'d like to convert.''', ) parser.add_argument( '''--checkpoint_path''', default='''/Users/nielsrogge/Documents/SwinSimMIM/simmim_pretrain__swin_base__img192_window6__100ep.pth''', type=str, help='''Path to the original PyTorch checkpoint (.pth file).''', ) 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.''' ) lowerCAmelCase : Tuple =parser.parse_args() convert_swin_checkpoint(args.model_name, args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub)
223
import copy from typing import Dict, List, Optional from ...configuration_utils import PretrainedConfig from ...utils import logging from ..auto import CONFIG_MAPPING SCREAMING_SNAKE_CASE__ : Dict = { "facebook/mask2former-swin-small-coco-instance": ( "https://huggingface.co/facebook/mask2former-swin-small-coco-instance/blob/main/config.json" ) # See all Mask2Former models at https://huggingface.co/models?filter=mask2former } SCREAMING_SNAKE_CASE__ : int = logging.get_logger(__name__) class lowerCAmelCase__ ( __lowercase ): a__ : Any = """mask2former""" a__ : Dict = ["""swin"""] a__ : Any = {"""hidden_size""": """hidden_dim"""} def __init__( self : Any , SCREAMING_SNAKE_CASE__ : Optional[Dict] = None , SCREAMING_SNAKE_CASE__ : int = 2_56 , SCREAMING_SNAKE_CASE__ : int = 2_56 , SCREAMING_SNAKE_CASE__ : int = 2_56 , SCREAMING_SNAKE_CASE__ : int = 10_24 , SCREAMING_SNAKE_CASE__ : str = "relu" , SCREAMING_SNAKE_CASE__ : int = 6 , SCREAMING_SNAKE_CASE__ : int = 10 , SCREAMING_SNAKE_CASE__ : int = 8 , SCREAMING_SNAKE_CASE__ : float = 0.0 , SCREAMING_SNAKE_CASE__ : int = 20_48 , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : int = 4 , SCREAMING_SNAKE_CASE__ : int = 2_55 , SCREAMING_SNAKE_CASE__ : int = 1_00 , SCREAMING_SNAKE_CASE__ : float = 0.1 , SCREAMING_SNAKE_CASE__ : float = 2.0 , SCREAMING_SNAKE_CASE__ : float = 5.0 , SCREAMING_SNAKE_CASE__ : float = 5.0 , SCREAMING_SNAKE_CASE__ : int = 1_25_44 , SCREAMING_SNAKE_CASE__ : float = 3.0 , SCREAMING_SNAKE_CASE__ : float = 0.75 , SCREAMING_SNAKE_CASE__ : float = 0.02 , SCREAMING_SNAKE_CASE__ : float = 1.0 , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : List[int] = [4, 8, 16, 32] , SCREAMING_SNAKE_CASE__ : bool = None , **SCREAMING_SNAKE_CASE__ : Tuple , ) -> str: if backbone_config is None: logger.info('''`backbone_config` is `None`. Initializing the config with the default `Swin` backbone.''' ) __lowerCamelCase = CONFIG_MAPPING['''swin''']( image_size=2_24 , in_channels=3 , patch_size=4 , embed_dim=96 , depths=[2, 2, 18, 2] , num_heads=[3, 6, 12, 24] , window_size=7 , drop_path_rate=0.3 , use_absolute_embeddings=SCREAMING_SNAKE_CASE__ , out_features=['''stage1''', '''stage2''', '''stage3''', '''stage4'''] , ) if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): __lowerCamelCase = backbone_config.pop('''model_type''' ) __lowerCamelCase = CONFIG_MAPPING[backbone_model_type] __lowerCamelCase = config_class.from_dict(SCREAMING_SNAKE_CASE__ ) # verify that the backbone is supported if backbone_config.model_type not in self.backbones_supported: logger.warning_once( f'''Backbone {backbone_config.model_type} is not a supported model and may not be compatible with Mask2Former. ''' f'''Supported model types: {','.join(self.backbones_supported )}''' ) __lowerCamelCase = backbone_config __lowerCamelCase = feature_size __lowerCamelCase = mask_feature_size __lowerCamelCase = hidden_dim __lowerCamelCase = encoder_feedforward_dim __lowerCamelCase = activation_function __lowerCamelCase = encoder_layers __lowerCamelCase = decoder_layers __lowerCamelCase = num_attention_heads __lowerCamelCase = dropout __lowerCamelCase = dim_feedforward __lowerCamelCase = pre_norm __lowerCamelCase = enforce_input_projection __lowerCamelCase = common_stride __lowerCamelCase = ignore_value __lowerCamelCase = num_queries __lowerCamelCase = no_object_weight __lowerCamelCase = class_weight __lowerCamelCase = mask_weight __lowerCamelCase = dice_weight __lowerCamelCase = train_num_points __lowerCamelCase = oversample_ratio __lowerCamelCase = importance_sample_ratio __lowerCamelCase = init_std __lowerCamelCase = init_xavier_std __lowerCamelCase = use_auxiliary_loss __lowerCamelCase = feature_strides __lowerCamelCase = output_auxiliary_logits __lowerCamelCase = decoder_layers super().__init__(**SCREAMING_SNAKE_CASE__ ) @classmethod def __A ( cls : Any , SCREAMING_SNAKE_CASE__ : PretrainedConfig , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> List[Any]: return cls( backbone_config=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ , ) def __A ( self : Any ) -> Dict[str, any]: __lowerCamelCase = copy.deepcopy(self.__dict__ ) __lowerCamelCase = self.backbone_config.to_dict() __lowerCamelCase = self.__class__.model_type return output
270
0
'''simple docstring''' import math from enum import Enum from typing import Optional, Union from torch.optim import Optimizer from torch.optim.lr_scheduler import LambdaLR from .utils import logging __lowerCAmelCase : Optional[Any] =logging.get_logger(__name__) class UpperCAmelCase ( __lowercase ): __lowercase = """linear""" __lowercase = """cosine""" __lowercase = """cosine_with_restarts""" __lowercase = """polynomial""" __lowercase = """constant""" __lowercase = """constant_with_warmup""" __lowercase = """piecewise_constant""" def UpperCamelCase ( _lowerCamelCase : Optimizer , _lowerCamelCase : int = -1 ): return LambdaLR(__lowerCAmelCase , lambda _lowerCamelCase : 1 , last_epoch=__lowerCAmelCase ) def UpperCamelCase ( _lowerCamelCase : Optimizer , _lowerCamelCase : int , _lowerCamelCase : int = -1 ): def lr_lambda(_lowerCamelCase : int ): if current_step < num_warmup_steps: return float(__lowerCAmelCase ) / float(max(1.0 , __lowerCAmelCase ) ) return 1.0 return LambdaLR(__lowerCAmelCase , __lowerCAmelCase , last_epoch=__lowerCAmelCase ) def UpperCamelCase ( _lowerCamelCase : Optimizer , _lowerCamelCase : str , _lowerCamelCase : int = -1 ): A__ = {} A__ = step_rules.split("," ) for rule_str in rule_list[:-1]: A__, A__ = rule_str.split(":" ) A__ = int(__lowerCAmelCase ) A__ = float(__lowerCAmelCase ) A__ = value A__ = float(rule_list[-1] ) def create_rules_function(_lowerCamelCase : int , _lowerCamelCase : List[str] ): def rule_func(_lowerCamelCase : int ) -> float: A__ = sorted(rules_dict.keys() ) for i, sorted_step in enumerate(__lowerCAmelCase ): if steps < sorted_step: return rules_dict[sorted_steps[i]] return last_lr_multiple return rule_func A__ = create_rules_function(__lowerCAmelCase , __lowerCAmelCase ) return LambdaLR(__lowerCAmelCase , __lowerCAmelCase , last_epoch=__lowerCAmelCase ) def UpperCamelCase ( _lowerCamelCase : Dict , _lowerCamelCase : str , _lowerCamelCase : Tuple , _lowerCamelCase : Optional[int]=-1 ): def lr_lambda(_lowerCamelCase : int ): if current_step < num_warmup_steps: return float(__lowerCAmelCase ) / float(max(1 , __lowerCAmelCase ) ) return max( 0.0 , float(num_training_steps - current_step ) / float(max(1 , num_training_steps - num_warmup_steps ) ) ) return LambdaLR(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) def UpperCamelCase ( _lowerCamelCase : Optimizer , _lowerCamelCase : int , _lowerCamelCase : int , _lowerCamelCase : float = 0.5 , _lowerCamelCase : int = -1 ): def lr_lambda(_lowerCamelCase : List[Any] ): if current_step < num_warmup_steps: return float(__lowerCAmelCase ) / float(max(1 , __lowerCAmelCase ) ) A__ = float(current_step - num_warmup_steps ) / float(max(1 , num_training_steps - num_warmup_steps ) ) return max(0.0 , 0.5 * (1.0 + math.cos(math.pi * float(__lowerCAmelCase ) * 2.0 * progress )) ) return LambdaLR(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) def UpperCamelCase ( _lowerCamelCase : Optimizer , _lowerCamelCase : int , _lowerCamelCase : int , _lowerCamelCase : int = 1 , _lowerCamelCase : int = -1 ): def lr_lambda(_lowerCamelCase : Dict ): if current_step < num_warmup_steps: return float(__lowerCAmelCase ) / float(max(1 , __lowerCAmelCase ) ) A__ = float(current_step - num_warmup_steps ) / float(max(1 , num_training_steps - num_warmup_steps ) ) if progress >= 1.0: return 0.0 return max(0.0 , 0.5 * (1.0 + math.cos(math.pi * ((float(__lowerCAmelCase ) * progress) % 1.0) )) ) return LambdaLR(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) def UpperCamelCase ( _lowerCamelCase : str , _lowerCamelCase : Union[str, Any] , _lowerCamelCase : List[Any] , _lowerCamelCase : int=1e-7 , _lowerCamelCase : Dict=1.0 , _lowerCamelCase : Union[str, Any]=-1 ): A__ = optimizer.defaults["lr"] if not (lr_init > lr_end): raise ValueError(F"lr_end ({lr_end}) must be be smaller than initial lr ({lr_init})" ) def lr_lambda(_lowerCamelCase : int ): if current_step < num_warmup_steps: return float(__lowerCAmelCase ) / float(max(1 , __lowerCAmelCase ) ) elif current_step > num_training_steps: return lr_end / lr_init # as LambdaLR multiplies by lr_init else: A__ = lr_init - lr_end A__ = num_training_steps - num_warmup_steps A__ = 1 - (current_step - num_warmup_steps) / decay_steps A__ = lr_range * pct_remaining**power + lr_end return decay / lr_init # as LambdaLR multiplies by lr_init return LambdaLR(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) __lowerCAmelCase : int ={ SchedulerType.LINEAR: get_linear_schedule_with_warmup, SchedulerType.COSINE: get_cosine_schedule_with_warmup, SchedulerType.COSINE_WITH_RESTARTS: get_cosine_with_hard_restarts_schedule_with_warmup, SchedulerType.POLYNOMIAL: get_polynomial_decay_schedule_with_warmup, SchedulerType.CONSTANT: get_constant_schedule, SchedulerType.CONSTANT_WITH_WARMUP: get_constant_schedule_with_warmup, SchedulerType.PIECEWISE_CONSTANT: get_piecewise_constant_schedule, } def UpperCamelCase ( _lowerCamelCase : Union[str, SchedulerType] , _lowerCamelCase : Optimizer , _lowerCamelCase : Optional[str] = None , _lowerCamelCase : Optional[int] = None , _lowerCamelCase : Optional[int] = None , _lowerCamelCase : int = 1 , _lowerCamelCase : float = 1.0 , _lowerCamelCase : int = -1 , ): A__ = SchedulerType(__lowerCAmelCase ) A__ = TYPE_TO_SCHEDULER_FUNCTION[name] if name == SchedulerType.CONSTANT: return schedule_func(__lowerCAmelCase , last_epoch=__lowerCAmelCase ) if name == SchedulerType.PIECEWISE_CONSTANT: return schedule_func(__lowerCAmelCase , step_rules=__lowerCAmelCase , last_epoch=__lowerCAmelCase ) # All other schedulers require `num_warmup_steps` if num_warmup_steps is None: raise ValueError(F"{name} requires `num_warmup_steps`, please provide that argument." ) if name == SchedulerType.CONSTANT_WITH_WARMUP: return schedule_func(__lowerCAmelCase , num_warmup_steps=__lowerCAmelCase , last_epoch=__lowerCAmelCase ) # All other schedulers require `num_training_steps` if num_training_steps is None: raise ValueError(F"{name} requires `num_training_steps`, please provide that argument." ) if name == SchedulerType.COSINE_WITH_RESTARTS: return schedule_func( __lowerCAmelCase , num_warmup_steps=__lowerCAmelCase , num_training_steps=__lowerCAmelCase , num_cycles=__lowerCAmelCase , last_epoch=__lowerCAmelCase , ) if name == SchedulerType.POLYNOMIAL: return schedule_func( __lowerCAmelCase , num_warmup_steps=__lowerCAmelCase , num_training_steps=__lowerCAmelCase , power=__lowerCAmelCase , last_epoch=__lowerCAmelCase , ) return schedule_func( __lowerCAmelCase , num_warmup_steps=__lowerCAmelCase , num_training_steps=__lowerCAmelCase , last_epoch=__lowerCAmelCase )
237
import os def __magic_name__ ( ) -> str: __lowerCamelCase = os.path.join(os.path.dirname(__lowerCAmelCase ) , '''num.txt''' ) with open(__lowerCAmelCase ) as file_hand: return str(sum(int(__lowerCAmelCase ) for line in file_hand ) )[:10] if __name__ == "__main__": print(solution())
270
0
import argparse import collections import json import os import re import string import sys import numpy as np lowerCAmelCase__ = re.compile(R"""\b(a|an|the)\b""", re.UNICODE) lowerCAmelCase__ = None def lowerCAmelCase__ ( ) -> str: '''simple docstring''' A__ = argparse.ArgumentParser("Official evaluation script for SQuAD version 2.0." ) parser.add_argument("data_file" , metavar="data.json" , help="Input data JSON file." ) parser.add_argument("pred_file" , metavar="pred.json" , help="Model predictions." ) parser.add_argument( "--out-file" , "-o" , metavar="eval.json" , help="Write accuracy metrics to file (default is stdout)." ) parser.add_argument( "--na-prob-file" , "-n" , metavar="na_prob.json" , help="Model estimates of probability of no answer." ) parser.add_argument( "--na-prob-thresh" , "-t" , type=__lowerCAmelCase , default=1.0 , help="Predict \"\" if no-answer probability exceeds this (default = 1.0)." , ) parser.add_argument( "--out-image-dir" , "-p" , metavar="out_images" , default=__lowerCAmelCase , help="Save precision-recall curves to directory." ) parser.add_argument("--verbose" , "-v" , action="store_true" ) if len(sys.argv ) == 1: parser.print_help() sys.exit(1 ) return parser.parse_args() def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: List[str] ) -> Union[str, Any]: '''simple docstring''' A__ = {} for article in dataset: for p in article["paragraphs"]: for qa in p["qas"]: A__ = bool(qa["answers"]["text"] ) return qid_to_has_ans def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Dict ) -> Optional[Any]: '''simple docstring''' def remove_articles(SCREAMING_SNAKE_CASE_: Optional[int] ): return ARTICLES_REGEX.sub(" " , __lowerCAmelCase ) def white_space_fix(SCREAMING_SNAKE_CASE_: Optional[int] ): return " ".join(text.split() ) def remove_punc(SCREAMING_SNAKE_CASE_: Union[str, Any] ): A__ = set(string.punctuation ) return "".join(ch for ch in text if ch not in exclude ) def lower(SCREAMING_SNAKE_CASE_: Dict ): return text.lower() return white_space_fix(remove_articles(remove_punc(lower(__lowerCAmelCase ) ) ) ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: List[Any] ) -> Optional[int]: '''simple docstring''' if not s: return [] return normalize_answer(__lowerCAmelCase ).split() def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: Tuple ) -> int: '''simple docstring''' return int(normalize_answer(__lowerCAmelCase ) == normalize_answer(__lowerCAmelCase ) ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Any , SCREAMING_SNAKE_CASE_: Tuple ) -> str: '''simple docstring''' A__ = get_tokens(__lowerCAmelCase ) A__ = get_tokens(__lowerCAmelCase ) A__ = collections.Counter(__lowerCAmelCase ) & collections.Counter(__lowerCAmelCase ) A__ = sum(common.values() ) if len(__lowerCAmelCase ) == 0 or len(__lowerCAmelCase ) == 0: # If either is no-answer, then F1 is 1 if they agree, 0 otherwise return int(gold_toks == pred_toks ) if num_same == 0: return 0 A__ = 1.0 * num_same / len(__lowerCAmelCase ) A__ = 1.0 * num_same / len(__lowerCAmelCase ) A__ = (2 * precision * recall) / (precision + recall) return fa def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Dict , SCREAMING_SNAKE_CASE_: List[str] ) -> Optional[Any]: '''simple docstring''' A__ = {} A__ = {} for article in dataset: for p in article["paragraphs"]: for qa in p["qas"]: A__ = qa["id"] A__ = [t for t in qa["answers"]["text"] if normalize_answer(__lowerCAmelCase )] if not gold_answers: # For unanswerable questions, only correct answer is empty string A__ = [""] if qid not in preds: print(F'Missing prediction for {qid}' ) continue A__ = preds[qid] # Take max over all gold answers A__ = max(compute_exact(__lowerCAmelCase , __lowerCAmelCase ) for a in gold_answers ) A__ = max(compute_fa(__lowerCAmelCase , __lowerCAmelCase ) for a in gold_answers ) return exact_scores, fa_scores def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: str , SCREAMING_SNAKE_CASE_: Optional[int] , SCREAMING_SNAKE_CASE_: Union[str, Any] ) -> List[str]: '''simple docstring''' A__ = {} for qid, s in scores.items(): A__ = na_probs[qid] > na_prob_thresh if pred_na: A__ = float(not qid_to_has_ans[qid] ) else: A__ = s return new_scores def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: List[Any] , SCREAMING_SNAKE_CASE_: Optional[int] , SCREAMING_SNAKE_CASE_: Optional[Any]=None ) -> Union[str, Any]: '''simple docstring''' if not qid_list: A__ = len(__lowerCAmelCase ) return collections.OrderedDict( [ ("exact", 100.0 * sum(exact_scores.values() ) / total), ("f1", 100.0 * sum(fa_scores.values() ) / total), ("total", total), ] ) else: A__ = len(__lowerCAmelCase ) return collections.OrderedDict( [ ("exact", 100.0 * sum(exact_scores[k] for k in qid_list ) / total), ("f1", 100.0 * sum(fa_scores[k] for k in qid_list ) / total), ("total", total), ] ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Any , SCREAMING_SNAKE_CASE_: str , SCREAMING_SNAKE_CASE_: Optional[Any] ) -> int: '''simple docstring''' for k in new_eval: A__ = new_eval[k] def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Optional[int] , SCREAMING_SNAKE_CASE_: List[Any] , SCREAMING_SNAKE_CASE_: Dict , SCREAMING_SNAKE_CASE_: Union[str, Any] ) -> Optional[Any]: '''simple docstring''' plt.step(__lowerCAmelCase , __lowerCAmelCase , color="b" , alpha=0.2 , where="post" ) plt.fill_between(__lowerCAmelCase , __lowerCAmelCase , step="post" , alpha=0.2 , color="b" ) plt.xlabel("Recall" ) plt.ylabel("Precision" ) plt.xlim([0.0, 1.05] ) plt.ylim([0.0, 1.05] ) plt.title(__lowerCAmelCase ) plt.savefig(__lowerCAmelCase ) plt.clf() def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Any , SCREAMING_SNAKE_CASE_: Optional[int] , SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: Optional[Any]=None , SCREAMING_SNAKE_CASE_: Tuple=None ) -> int: '''simple docstring''' A__ = sorted(__lowerCAmelCase , key=lambda SCREAMING_SNAKE_CASE_ : na_probs[k] ) A__ = 0.0 A__ = 1.0 A__ = 0.0 A__ = [1.0] A__ = [0.0] A__ = 0.0 for i, qid in enumerate(__lowerCAmelCase ): if qid_to_has_ans[qid]: true_pos += scores[qid] A__ = true_pos / float(i + 1 ) A__ = true_pos / float(__lowerCAmelCase ) if i == len(__lowerCAmelCase ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]: # i.e., if we can put a threshold after this point avg_prec += cur_p * (cur_r - recalls[-1]) precisions.append(__lowerCAmelCase ) recalls.append(__lowerCAmelCase ) if out_image: plot_pr_curve(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) return {"ap": 100.0 * avg_prec} def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: List[str] , SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: List[str] , SCREAMING_SNAKE_CASE_: List[Any] , SCREAMING_SNAKE_CASE_: List[Any] ) -> List[Any]: '''simple docstring''' if out_image_dir and not os.path.exists(__lowerCAmelCase ): os.makedirs(__lowerCAmelCase ) A__ = sum(1 for v in qid_to_has_ans.values() if v ) if num_true_pos == 0: return A__ = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , "pr_exact.png" ) , title="Precision-Recall curve for Exact Match score" , ) A__ = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , "pr_f1.png" ) , title="Precision-Recall curve for F1 score" , ) A__ = {k: float(__lowerCAmelCase ) for k, v in qid_to_has_ans.items()} A__ = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , "pr_oracle.png" ) , title="Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)" , ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , "pr_exact" ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , "pr_f1" ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , "pr_oracle" ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: Any , SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: int ) -> Optional[Any]: '''simple docstring''' if not qid_list: return A__ = [na_probs[k] for k in qid_list] A__ = np.ones_like(__lowerCAmelCase ) / float(len(__lowerCAmelCase ) ) plt.hist(__lowerCAmelCase , weights=__lowerCAmelCase , bins=2_0 , range=(0.0, 1.0) ) plt.xlabel("Model probability of no-answer" ) plt.ylabel("Proportion of dataset" ) plt.title(F'Histogram of no-answer probability: {name}' ) plt.savefig(os.path.join(__lowerCAmelCase , F'na_prob_hist_{name}.png' ) ) plt.clf() def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: List[str] , SCREAMING_SNAKE_CASE_: Union[str, Any] ) -> Optional[int]: '''simple docstring''' A__ = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] ) A__ = num_no_ans A__ = cur_score A__ = 0.0 A__ = sorted(__lowerCAmelCase , key=lambda SCREAMING_SNAKE_CASE_ : na_probs[k] ) for i, qid in enumerate(__lowerCAmelCase ): if qid not in scores: continue if qid_to_has_ans[qid]: A__ = scores[qid] else: if preds[qid]: A__ = -1 else: A__ = 0 cur_score += diff if cur_score > best_score: A__ = cur_score A__ = na_probs[qid] return 100.0 * best_score / len(__lowerCAmelCase ), best_thresh def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Dict , SCREAMING_SNAKE_CASE_: Optional[int] , SCREAMING_SNAKE_CASE_: int , SCREAMING_SNAKE_CASE_: Union[str, Any] , SCREAMING_SNAKE_CASE_: int , SCREAMING_SNAKE_CASE_: Optional[Any] ) -> int: '''simple docstring''' A__ , A__ = find_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) A__ , A__ = find_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) A__ = best_exact A__ = exact_thresh A__ = best_fa A__ = fa_thresh def lowerCAmelCase__ ( ) -> Optional[int]: '''simple docstring''' with open(OPTS.data_file ) as f: A__ = json.load(__lowerCAmelCase ) A__ = dataset_json["data"] with open(OPTS.pred_file ) as f: A__ = json.load(__lowerCAmelCase ) if OPTS.na_prob_file: with open(OPTS.na_prob_file ) as f: A__ = json.load(__lowerCAmelCase ) else: A__ = {k: 0.0 for k in preds} A__ = make_qid_to_has_ans(__lowerCAmelCase ) # maps qid to True/False A__ = [k for k, v in qid_to_has_ans.items() if v] A__ = [k for k, v in qid_to_has_ans.items() if not v] A__ , A__ = get_raw_scores(__lowerCAmelCase , __lowerCAmelCase ) A__ = apply_no_ans_threshold(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.na_prob_thresh ) A__ = apply_no_ans_threshold(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.na_prob_thresh ) A__ = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase ) if has_ans_qids: A__ = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase , qid_list=__lowerCAmelCase ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , "HasAns" ) if no_ans_qids: A__ = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase , qid_list=__lowerCAmelCase ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , "NoAns" ) if OPTS.na_prob_file: find_all_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) if OPTS.na_prob_file and OPTS.out_image_dir: run_precision_recall_analysis(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir ) histogram_na_prob(__lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir , "hasAns" ) histogram_na_prob(__lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir , "noAns" ) if OPTS.out_file: with open(OPTS.out_file , "w" ) as f: json.dump(__lowerCAmelCase , __lowerCAmelCase ) else: print(json.dumps(__lowerCAmelCase , indent=2 ) ) if __name__ == "__main__": lowerCAmelCase__ = parse_args() if OPTS.out_image_dir: import matplotlib matplotlib.use("""Agg""") import matplotlib.pyplot as plt main()
68
import logging import os from dataclasses import dataclass from enum import Enum from typing import List, Optional, Union from filelock import FileLock from transformers import PreTrainedTokenizer, is_tf_available, is_torch_available SCREAMING_SNAKE_CASE__ : Optional[int] = logging.getLogger(__name__) @dataclass class lowerCAmelCase__ : a__ : str a__ : List[str] a__ : Optional[List[str]] @dataclass class lowerCAmelCase__ : a__ : List[int] a__ : List[int] a__ : Optional[List[int]] = None a__ : Optional[List[int]] = None class lowerCAmelCase__ ( __lowercase ): a__ : Optional[Any] = """train""" a__ : Optional[int] = """dev""" a__ : Dict = """test""" class lowerCAmelCase__ : @staticmethod def __A ( SCREAMING_SNAKE_CASE__ : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Union[Split, str] ) -> List[InputExample]: raise NotImplementedError @staticmethod def __A ( SCREAMING_SNAKE_CASE__ : str ) -> List[str]: raise NotImplementedError @staticmethod def __A ( SCREAMING_SNAKE_CASE__ : List[InputExample] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : PreTrainedTokenizer , SCREAMING_SNAKE_CASE__ : Optional[Any]=False , SCREAMING_SNAKE_CASE__ : List[str]="[CLS]" , SCREAMING_SNAKE_CASE__ : Tuple=1 , SCREAMING_SNAKE_CASE__ : str="[SEP]" , SCREAMING_SNAKE_CASE__ : List[Any]=False , SCREAMING_SNAKE_CASE__ : Union[str, Any]=False , SCREAMING_SNAKE_CASE__ : Tuple=0 , SCREAMING_SNAKE_CASE__ : int=0 , SCREAMING_SNAKE_CASE__ : str=-1_00 , SCREAMING_SNAKE_CASE__ : Union[str, Any]=0 , SCREAMING_SNAKE_CASE__ : List[Any]=True , ) -> List[InputFeatures]: __lowerCamelCase = {label: i for i, label in enumerate(SCREAMING_SNAKE_CASE__ )} __lowerCamelCase = [] for ex_index, example in enumerate(SCREAMING_SNAKE_CASE__ ): if ex_index % 1_00_00 == 0: logger.info('''Writing example %d of %d''' , SCREAMING_SNAKE_CASE__ , len(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = [] __lowerCamelCase = [] for word, label in zip(example.words , example.labels ): __lowerCamelCase = tokenizer.tokenize(SCREAMING_SNAKE_CASE__ ) # bert-base-multilingual-cased sometimes output "nothing ([]) when calling tokenize with just a space. if len(SCREAMING_SNAKE_CASE__ ) > 0: tokens.extend(SCREAMING_SNAKE_CASE__ ) # Use the real label id for the first token of the word, and padding ids for the remaining tokens label_ids.extend([label_map[label]] + [pad_token_label_id] * (len(SCREAMING_SNAKE_CASE__ ) - 1) ) # Account for [CLS] and [SEP] with "- 2" and with "- 3" for RoBERTa. __lowerCamelCase = tokenizer.num_special_tokens_to_add() if len(SCREAMING_SNAKE_CASE__ ) > max_seq_length - special_tokens_count: __lowerCamelCase = tokens[: (max_seq_length - special_tokens_count)] __lowerCamelCase = label_ids[: (max_seq_length - special_tokens_count)] # The convention in BERT is: # (a) For sequence pairs: # tokens: [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP] # type_ids: 0 0 0 0 0 0 0 0 1 1 1 1 1 1 # (b) For single sequences: # tokens: [CLS] the dog is hairy . [SEP] # type_ids: 0 0 0 0 0 0 0 # # Where "type_ids" are used to indicate whether this is the first # sequence or the second sequence. The embedding vectors for `type=0` and # `type=1` were learned during pre-training and are added to the wordpiece # embedding vector (and position vector). This is not *strictly* necessary # since the [SEP] token unambiguously separates the sequences, but it makes # it easier for the model to learn the concept of sequences. # # For classification tasks, the first vector (corresponding to [CLS]) is # used as the "sentence vector". Note that this only makes sense because # the entire model is fine-tuned. tokens += [sep_token] label_ids += [pad_token_label_id] if sep_token_extra: # roberta uses an extra separator b/w pairs of sentences tokens += [sep_token] label_ids += [pad_token_label_id] __lowerCamelCase = [sequence_a_segment_id] * len(SCREAMING_SNAKE_CASE__ ) if cls_token_at_end: tokens += [cls_token] label_ids += [pad_token_label_id] segment_ids += [cls_token_segment_id] else: __lowerCamelCase = [cls_token] + tokens __lowerCamelCase = [pad_token_label_id] + label_ids __lowerCamelCase = [cls_token_segment_id] + segment_ids __lowerCamelCase = tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) # The mask has 1 for real tokens and 0 for padding tokens. Only real # tokens are attended to. __lowerCamelCase = [1 if mask_padding_with_zero else 0] * len(SCREAMING_SNAKE_CASE__ ) # Zero-pad up to the sequence length. __lowerCamelCase = max_seq_length - len(SCREAMING_SNAKE_CASE__ ) if pad_on_left: __lowerCamelCase = ([pad_token] * padding_length) + input_ids __lowerCamelCase = ([0 if mask_padding_with_zero else 1] * padding_length) + input_mask __lowerCamelCase = ([pad_token_segment_id] * padding_length) + segment_ids __lowerCamelCase = ([pad_token_label_id] * padding_length) + label_ids else: input_ids += [pad_token] * padding_length input_mask += [0 if mask_padding_with_zero else 1] * padding_length segment_ids += [pad_token_segment_id] * padding_length label_ids += [pad_token_label_id] * padding_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length if ex_index < 5: logger.info('''*** Example ***''' ) logger.info('''guid: %s''' , example.guid ) logger.info('''tokens: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in tokens] ) ) logger.info('''input_ids: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in input_ids] ) ) logger.info('''input_mask: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in input_mask] ) ) logger.info('''segment_ids: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in segment_ids] ) ) logger.info('''label_ids: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in label_ids] ) ) if "token_type_ids" not in tokenizer.model_input_names: __lowerCamelCase = None features.append( InputFeatures( input_ids=SCREAMING_SNAKE_CASE__ , attention_mask=SCREAMING_SNAKE_CASE__ , token_type_ids=SCREAMING_SNAKE_CASE__ , label_ids=SCREAMING_SNAKE_CASE__ ) ) return features if is_torch_available(): import torch from torch import nn from torch.utils.data import Dataset class lowerCAmelCase__ ( __lowercase ): a__ : List[InputFeatures] a__ : int = nn.CrossEntropyLoss().ignore_index def __init__( self : List[str] , SCREAMING_SNAKE_CASE__ : TokenClassificationTask , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : PreTrainedTokenizer , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : Optional[int] = None , SCREAMING_SNAKE_CASE__ : Union[str, Any]=False , SCREAMING_SNAKE_CASE__ : Split = Split.train , ) -> Union[str, Any]: # Load data features from cache or dataset file __lowerCamelCase = os.path.join( SCREAMING_SNAKE_CASE__ , '''cached_{}_{}_{}'''.format(mode.value , tokenizer.__class__.__name__ , str(SCREAMING_SNAKE_CASE__ ) ) , ) # Make sure only the first process in distributed training processes the dataset, # and the others will use the cache. __lowerCamelCase = cached_features_file + '''.lock''' with FileLock(SCREAMING_SNAKE_CASE__ ): if os.path.exists(SCREAMING_SNAKE_CASE__ ) and not overwrite_cache: logger.info(f'''Loading features from cached file {cached_features_file}''' ) __lowerCamelCase = torch.load(SCREAMING_SNAKE_CASE__ ) else: logger.info(f'''Creating features from dataset file at {data_dir}''' ) __lowerCamelCase = token_classification_task.read_examples_from_file(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # TODO clean up all this to leverage built-in features of tokenizers __lowerCamelCase = token_classification_task.convert_examples_to_features( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , cls_token_at_end=bool(model_type in ['''xlnet'''] ) , cls_token=tokenizer.cls_token , cls_token_segment_id=2 if model_type in ['''xlnet'''] else 0 , sep_token=tokenizer.sep_token , sep_token_extra=SCREAMING_SNAKE_CASE__ , pad_on_left=bool(tokenizer.padding_side == '''left''' ) , pad_token=tokenizer.pad_token_id , pad_token_segment_id=tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , ) logger.info(f'''Saving features into cached file {cached_features_file}''' ) torch.save(self.features , SCREAMING_SNAKE_CASE__ ) def __len__( self : Dict ) -> str: return len(self.features ) def __getitem__( self : Any , SCREAMING_SNAKE_CASE__ : Dict ) -> InputFeatures: return self.features[i] if is_tf_available(): import tensorflow as tf class lowerCAmelCase__ : a__ : List[InputFeatures] a__ : int = -100 def __init__( self : Optional[Any] , SCREAMING_SNAKE_CASE__ : TokenClassificationTask , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : PreTrainedTokenizer , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : Optional[int] = None , SCREAMING_SNAKE_CASE__ : Optional[Any]=False , SCREAMING_SNAKE_CASE__ : Split = Split.train , ) -> List[Any]: __lowerCamelCase = token_classification_task.read_examples_from_file(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # TODO clean up all this to leverage built-in features of tokenizers __lowerCamelCase = token_classification_task.convert_examples_to_features( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , cls_token_at_end=bool(model_type in ['''xlnet'''] ) , cls_token=tokenizer.cls_token , cls_token_segment_id=2 if model_type in ['''xlnet'''] else 0 , sep_token=tokenizer.sep_token , sep_token_extra=SCREAMING_SNAKE_CASE__ , pad_on_left=bool(tokenizer.padding_side == '''left''' ) , pad_token=tokenizer.pad_token_id , pad_token_segment_id=tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , ) def gen(): for ex in self.features: if ex.token_type_ids is None: yield ( {"input_ids": ex.input_ids, "attention_mask": ex.attention_mask}, ex.label_ids, ) else: yield ( { "input_ids": ex.input_ids, "attention_mask": ex.attention_mask, "token_type_ids": ex.token_type_ids, }, ex.label_ids, ) if "token_type_ids" not in tokenizer.model_input_names: __lowerCamelCase = tf.data.Dataset.from_generator( SCREAMING_SNAKE_CASE__ , ({'''input_ids''': tf.intaa, '''attention_mask''': tf.intaa}, tf.intaa) , ( {'''input_ids''': tf.TensorShape([None] ), '''attention_mask''': tf.TensorShape([None] )}, tf.TensorShape([None] ), ) , ) else: __lowerCamelCase = tf.data.Dataset.from_generator( SCREAMING_SNAKE_CASE__ , ({'''input_ids''': tf.intaa, '''attention_mask''': tf.intaa, '''token_type_ids''': tf.intaa}, tf.intaa) , ( { '''input_ids''': tf.TensorShape([None] ), '''attention_mask''': tf.TensorShape([None] ), '''token_type_ids''': tf.TensorShape([None] ), }, tf.TensorShape([None] ), ) , ) def __A ( self : Union[str, Any] ) -> Union[str, Any]: __lowerCamelCase = self.dataset.apply(tf.data.experimental.assert_cardinality(len(self.features ) ) ) return self.dataset def __len__( self : List[Any] ) -> Any: return len(self.features ) def __getitem__( self : List[str] , SCREAMING_SNAKE_CASE__ : Optional[int] ) -> InputFeatures: return self.features[i]
270
0
"""simple docstring""" from collections import defaultdict class __lowerCamelCase : '''simple docstring''' def __init__( self , __UpperCAmelCase , __UpperCAmelCase ) -> Union[str, Any]: _a = total # total no of tasks (N) # DP table will have a dimension of (2^M)*N # initially all values are set to -1 _a = [ [-1 for i in range(total + 1 )] for j in range(2 ** len(SCREAMING_SNAKE_CASE__ ) ) ] _a = defaultdict(SCREAMING_SNAKE_CASE__ ) # stores the list of persons for each task # final_mask is used to check if all persons are included by setting all bits # to 1 _a = (1 << len(SCREAMING_SNAKE_CASE__ )) - 1 def _UpperCAmelCase ( self , __UpperCAmelCase , __UpperCAmelCase ) -> str: # if mask == self.finalmask all persons are distributed tasks, return 1 if mask == self.final_mask: return 1 # if not everyone gets the task and no more tasks are available, return 0 if task_no > self.total_tasks: return 0 # if case already considered if self.dp[mask][task_no] != -1: return self.dp[mask][task_no] # Number of ways when we don't this task in the arrangement _a = self.count_ways_until(SCREAMING_SNAKE_CASE__ , task_no + 1 ) # now assign the tasks one by one to all possible persons and recursively # assign for the remaining tasks. if task_no in self.task: for p in self.task[task_no]: # if p is already given a task if mask & (1 << p): continue # assign this task to p and change the mask value. And recursively # assign tasks with the new mask value. total_ways_util += self.count_ways_until(mask | (1 << p) , task_no + 1 ) # save the value. _a = total_ways_util return self.dp[mask][task_no] def _UpperCAmelCase ( self , __UpperCAmelCase ) -> List[Any]: # Store the list of persons for each task for i in range(len(SCREAMING_SNAKE_CASE__ ) ): for j in task_performed[i]: self.task[j].append(SCREAMING_SNAKE_CASE__ ) # call the function to fill the DP table, final answer is stored in dp[0][1] return self.count_ways_until(0 , 1 ) if __name__ == "__main__": __snake_case = 5 # total no of tasks (the value of N) # the list of tasks that can be done by M persons. __snake_case = [[1, 3, 4], [1, 2, 5], [3, 4]] print( AssignmentUsingBitmask(task_performed, total_tasks).count_no_of_ways( task_performed ) )
320
from typing import Optional from .. import Features, NamedSplit from ..packaged_modules.text.text import Text from ..utils.typing import NestedDataStructureLike, PathLike from .abc import AbstractDatasetReader class lowerCAmelCase__ ( __lowercase ): def __init__( self : Tuple , SCREAMING_SNAKE_CASE__ : NestedDataStructureLike[PathLike] , SCREAMING_SNAKE_CASE__ : Optional[NamedSplit] = None , SCREAMING_SNAKE_CASE__ : Optional[Features] = None , SCREAMING_SNAKE_CASE__ : str = None , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : Optional[int] = None , **SCREAMING_SNAKE_CASE__ : str , ) -> Union[str, Any]: super().__init__( SCREAMING_SNAKE_CASE__ , split=SCREAMING_SNAKE_CASE__ , features=SCREAMING_SNAKE_CASE__ , cache_dir=SCREAMING_SNAKE_CASE__ , keep_in_memory=SCREAMING_SNAKE_CASE__ , streaming=SCREAMING_SNAKE_CASE__ , num_proc=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ , ) __lowerCamelCase = path_or_paths if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) else {self.split: path_or_paths} __lowerCamelCase = Text( cache_dir=SCREAMING_SNAKE_CASE__ , data_files=SCREAMING_SNAKE_CASE__ , features=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ , ) def __A ( self : int ) -> Dict: # Build iterable dataset if self.streaming: __lowerCamelCase = self.builder.as_streaming_dataset(split=self.split ) # Build regular (map-style) dataset else: __lowerCamelCase = None __lowerCamelCase = None __lowerCamelCase = None __lowerCamelCase = None self.builder.download_and_prepare( download_config=SCREAMING_SNAKE_CASE__ , download_mode=SCREAMING_SNAKE_CASE__ , verification_mode=SCREAMING_SNAKE_CASE__ , base_path=SCREAMING_SNAKE_CASE__ , num_proc=self.num_proc , ) __lowerCamelCase = self.builder.as_dataset( split=self.split , verification_mode=SCREAMING_SNAKE_CASE__ , in_memory=self.keep_in_memory ) return dataset
270
0
import argparse import evaluate import torch from datasets import load_dataset from torch.optim import AdamW from torch.utils.data import DataLoader from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed from accelerate import Accelerator, DistributedType ######################################################################## # This is a fully working simple example to use Accelerate # # This example trains a Bert base model on GLUE MRPC # in any of the following settings (with the same script): # - single CPU or single GPU # - multi GPUS (using PyTorch distributed mode) # - (multi) TPUs # - fp16 (mixed-precision) or fp32 (normal precision) # # To run it in each of these various modes, follow the instructions # in the readme for examples: # https://github.com/huggingface/accelerate/tree/main/examples # ######################################################################## A : Optional[int] = 1_6 A : Tuple = 3_2 def UpperCamelCase ( __magic_name__ : Accelerator , __magic_name__ : int = 16 ) -> Union[str, Any]: """simple docstring""" lowercase__ = AutoTokenizer.from_pretrained("""bert-base-cased""" ) lowercase__ = load_dataset("""glue""" , """mrpc""" ) def tokenize_function(__magic_name__ : Tuple ): # max_length=None => use the model max length (it's actually the default) lowercase__ = tokenizer(examples["""sentence1"""] , examples["""sentence2"""] , truncation=__lowerCAmelCase , max_length=__lowerCAmelCase ) return outputs # Apply the method we just defined to all the examples in all the splits of the dataset # starting with the main process first: with accelerator.main_process_first(): lowercase__ = datasets.map( __lowerCAmelCase , batched=__lowerCAmelCase , remove_columns=["""idx""", """sentence1""", """sentence2"""] , ) # We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the # transformers library lowercase__ = tokenized_datasets.rename_column("""label""" , """labels""" ) def collate_fn(__magic_name__ : Optional[int] ): # On TPU it's best to pad everything to the same length or training will be very slow. lowercase__ = 128 if accelerator.distributed_type == DistributedType.TPU else None # When using mixed precision we want round multiples of 8/16 if accelerator.mixed_precision == "fp8": lowercase__ = 16 elif accelerator.mixed_precision != "no": lowercase__ = 8 else: lowercase__ = None return tokenizer.pad( __lowerCAmelCase , padding="""longest""" , max_length=__lowerCAmelCase , pad_to_multiple_of=__lowerCAmelCase , return_tensors="""pt""" , ) # Instantiate dataloaders. lowercase__ = DataLoader( tokenized_datasets["""train"""] , shuffle=__lowerCAmelCase , collate_fn=__lowerCAmelCase , batch_size=__lowerCAmelCase , drop_last=__lowerCAmelCase ) lowercase__ = DataLoader( tokenized_datasets["""validation"""] , shuffle=__lowerCAmelCase , collate_fn=__lowerCAmelCase , batch_size=__lowerCAmelCase , drop_last=(accelerator.mixed_precision == """fp8""") , ) return train_dataloader, eval_dataloader def UpperCamelCase ( __magic_name__ : Optional[int] , __magic_name__ : Union[str, Any] ) -> Any: """simple docstring""" lowercase__ = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision ) # Sample hyper-parameters for learning rate, batch size, seed and a few other HPs lowercase__ = config["""lr"""] lowercase__ = int(config["""num_epochs"""] ) lowercase__ = int(config["""seed"""] ) lowercase__ = int(config["""batch_size"""] ) lowercase__ = evaluate.load("""glue""" , """mrpc""" ) # If the batch size is too big we use gradient accumulation lowercase__ = 1 if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU: lowercase__ = batch_size // MAX_GPU_BATCH_SIZE lowercase__ = MAX_GPU_BATCH_SIZE set_seed(__lowerCAmelCase ) lowercase__ , lowercase__ = get_dataloaders(__lowerCAmelCase , __lowerCAmelCase ) # Instantiate the model (we build the model here so that the seed also control new weights initialization) lowercase__ = AutoModelForSequenceClassification.from_pretrained("""bert-base-cased""" , return_dict=__lowerCAmelCase ) # We could avoid this line since the accelerator is set with `device_placement=True` (default value). # Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer # creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that). lowercase__ = model.to(accelerator.device ) # Instantiate optimizer lowercase__ = AdamW(params=model.parameters() , lr=__lowerCAmelCase ) # Instantiate scheduler lowercase__ = get_linear_schedule_with_warmup( optimizer=__lowerCAmelCase , num_warmup_steps=100 , num_training_steps=(len(__lowerCAmelCase ) * num_epochs) // gradient_accumulation_steps , ) # Prepare everything # There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the # prepare method. lowercase__ , lowercase__ , lowercase__ , lowercase__ , lowercase__ = accelerator.prepare( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) # Now we train the model for epoch in range(__lowerCAmelCase ): model.train() for step, batch in enumerate(__lowerCAmelCase ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) lowercase__ = model(**__lowerCAmelCase ) lowercase__ = outputs.loss lowercase__ = loss / gradient_accumulation_steps accelerator.backward(__lowerCAmelCase ) if step % gradient_accumulation_steps == 0: optimizer.step() lr_scheduler.step() optimizer.zero_grad() model.eval() for step, batch in enumerate(__lowerCAmelCase ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): lowercase__ = model(**__lowerCAmelCase ) lowercase__ = outputs.logits.argmax(dim=-1 ) lowercase__ , lowercase__ = accelerator.gather_for_metrics((predictions, batch["""labels"""]) ) metric.add_batch( predictions=__lowerCAmelCase , references=__lowerCAmelCase , ) lowercase__ = metric.compute() # Use accelerator.print to print only on the main process. accelerator.print(f'''epoch {epoch}:''' , __lowerCAmelCase ) def UpperCamelCase ( ) -> Any: """simple docstring""" lowercase__ = argparse.ArgumentParser(description="""Simple example of training script.""" ) parser.add_argument( """--mixed_precision""" , type=__lowerCAmelCase , default=__lowerCAmelCase , choices=["""no""", """fp16""", """bf16""", """fp8"""] , help="""Whether to use mixed precision. Choose""" """between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.""" """and an Nvidia Ampere GPU.""" , ) parser.add_argument("""--cpu""" , action="""store_true""" , help="""If passed, will train on the CPU.""" ) lowercase__ = parser.parse_args() lowercase__ = {"""lr""": 2E-5, """num_epochs""": 3, """seed""": 42, """batch_size""": 16} training_function(__lowerCAmelCase , __lowerCAmelCase ) if __name__ == "__main__": main()
305
from ...utils import ( OptionalDependencyNotAvailable, is_torch_available, is_transformers_available, is_transformers_version, ) try: if not (is_transformers_available() and is_torch_available() and is_transformers_version(">=", "4.25.0")): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: from ...utils.dummy_torch_and_transformers_objects import UnCLIPImageVariationPipeline, UnCLIPPipeline else: from .pipeline_unclip import UnCLIPPipeline from .pipeline_unclip_image_variation import UnCLIPImageVariationPipeline from .text_proj import UnCLIPTextProjModel
270
0
'''simple docstring''' from __future__ import annotations from statistics import mean def _UpperCamelCase ( SCREAMING_SNAKE_CASE__ : list[int] , SCREAMING_SNAKE_CASE__ : list[int] , SCREAMING_SNAKE_CASE__ : int ): '''simple docstring''' UpperCAmelCase__ = [0] * no_of_processes UpperCAmelCase__ = [0] * no_of_processes # Initialize remaining_time to waiting_time. for i in range(__lowerCAmelCase ): UpperCAmelCase__ = burst_time[i] UpperCAmelCase__ = [] UpperCAmelCase__ = 0 UpperCAmelCase__ = 0 # When processes are not completed, # A process whose arrival time has passed \ # and has remaining execution time is put into the ready_process. # The shortest process in the ready_process, target_process is executed. while completed != no_of_processes: UpperCAmelCase__ = [] UpperCAmelCase__ = -1 for i in range(__lowerCAmelCase ): if (arrival_time[i] <= total_time) and (remaining_time[i] > 0): ready_process.append(__lowerCAmelCase ) if len(__lowerCAmelCase ) > 0: UpperCAmelCase__ = ready_process[0] for i in ready_process: if remaining_time[i] < remaining_time[target_process]: UpperCAmelCase__ = i total_time += burst_time[target_process] completed += 1 UpperCAmelCase__ = 0 UpperCAmelCase__ = ( total_time - arrival_time[target_process] - burst_time[target_process] ) else: total_time += 1 return waiting_time def _UpperCamelCase ( SCREAMING_SNAKE_CASE__ : list[int] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : list[int] ): '''simple docstring''' UpperCAmelCase__ = [0] * no_of_processes for i in range(__lowerCAmelCase ): UpperCAmelCase__ = burst_time[i] + waiting_time[i] return turn_around_time if __name__ == "__main__": print('[TEST CASE 01]') UpperCAmelCase_ = 4 UpperCAmelCase_ = [2, 5, 3, 7] UpperCAmelCase_ = [0, 0, 0, 0] UpperCAmelCase_ = calculate_waitingtime(arrival_time, burst_time, no_of_processes) UpperCAmelCase_ = calculate_turnaroundtime( burst_time, no_of_processes, waiting_time ) # Printing the Result print('PID\tBurst Time\tArrival Time\tWaiting Time\tTurnaround Time') for i, process_id in enumerate(list(range(1, 5))): print( f"{process_id}\t{burst_time[i]}\t\t\t{arrival_time[i]}\t\t\t\t" f"{waiting_time[i]}\t\t\t\t{turn_around_time[i]}" ) print(f"\nAverage waiting time = {mean(waiting_time):.5f}") print(f"Average turnaround time = {mean(turn_around_time):.5f}")
346
import sacrebleu as scb from packaging import version from sacrebleu import CHRF import datasets SCREAMING_SNAKE_CASE__ : Union[str, Any] = "\\n@inproceedings{popovic-2015-chrf,\n title = \"chr{F}: character n-gram {F}-score for automatic {MT} evaluation\",\n author = \"Popovi{\'c}, Maja\",\n booktitle = \"Proceedings of the Tenth Workshop on Statistical Machine Translation\",\n month = sep,\n year = \"2015\",\n address = \"Lisbon, Portugal\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/W15-3049\",\n doi = \"10.18653/v1/W15-3049\",\n pages = \"392--395\",\n}\n@inproceedings{popovic-2017-chrf,\n title = \"chr{F}++: words helping character n-grams\",\n author = \"Popovi{\'c}, Maja\",\n booktitle = \"Proceedings of the Second Conference on Machine Translation\",\n month = sep,\n year = \"2017\",\n address = \"Copenhagen, Denmark\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/W17-4770\",\n doi = \"10.18653/v1/W17-4770\",\n pages = \"612--618\",\n}\n@inproceedings{post-2018-call,\n title = \"A Call for Clarity in Reporting {BLEU} Scores\",\n author = \"Post, Matt\",\n booktitle = \"Proceedings of the Third Conference on Machine Translation: Research Papers\",\n month = oct,\n year = \"2018\",\n address = \"Belgium, Brussels\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W18-6319\",\n pages = \"186--191\",\n}\n" SCREAMING_SNAKE_CASE__ : int = "\\nChrF and ChrF++ are two MT evaluation metrics. They both use the F-score statistic for character n-gram matches,\nand ChrF++ adds word n-grams as well which correlates more strongly with direct assessment. We use the implementation\nthat is already present in sacrebleu.\n\nThe implementation here is slightly different from sacrebleu in terms of the required input format. The length of\nthe references and hypotheses lists need to be the same, so you may need to transpose your references compared to\nsacrebleu's required input format. See https://github.com/huggingface/datasets/issues/3154#issuecomment-950746534\n\nSee the README.md file at https://github.com/mjpost/sacreBLEU#chrf--chrf for more information.\n" SCREAMING_SNAKE_CASE__ : Optional[Any] = "\nProduces ChrF(++) scores for hypotheses given reference translations.\n\nArgs:\n predictions (list of str): The predicted sentences.\n references (list of list of str): The references. There should be one reference sub-list for each prediction sentence.\n char_order (int): Character n-gram order. Defaults to `6`.\n word_order (int): Word n-gram order. If equals to `2`, the metric is referred to as chrF++. Defaults to `0`.\n beta (int): Determine the importance of recall w.r.t precision. Defaults to `2`.\n lowercase (bool): if `True`, enables case-insensitivity. Defaults to `False`.\n whitespace (bool): If `True`, include whitespaces when extracting character n-grams.\n eps_smoothing (bool): If `True`, applies epsilon smoothing similar\n to reference chrF++.py, NLTK and Moses implementations. If `False`,\n it takes into account effective match order similar to sacreBLEU < 2.0.0. Defaults to `False`.\n\nReturns:\n 'score' (float): The chrF (chrF++) score,\n 'char_order' (int): The character n-gram order,\n 'word_order' (int): The word n-gram order. If equals to 2, the metric is referred to as chrF++,\n 'beta' (int): Determine the importance of recall w.r.t precision\n\nExamples:\n Example 1--a simple example of calculating chrF:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction, references=reference)\n >>> print(results)\n {'score': 84.64214891738334, 'char_order': 6, 'word_order': 0, 'beta': 2}\n\n Example 2--the same example, but with the argument word_order=2, to calculate chrF++ instead of chrF:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction,\n ... references=reference,\n ... word_order=2)\n >>> print(results)\n {'score': 82.87263732906315, 'char_order': 6, 'word_order': 2, 'beta': 2}\n\n Example 3--the same chrF++ example as above, but with `lowercase=True` to normalize all case:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction,\n ... references=reference,\n ... word_order=2,\n ... lowercase=True)\n >>> print(results)\n {'score': 92.12853119829202, 'char_order': 6, 'word_order': 2, 'beta': 2}\n" @datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION ) class lowerCAmelCase__ ( datasets.Metric ): def __A ( self : Dict ) -> Tuple: if version.parse(scb.__version__ ) < version.parse('''1.4.12''' ): raise ImportWarning( '''To use `sacrebleu`, the module `sacrebleu>=1.4.12` is required, and the current version of `sacrebleu` doesn\'t match this condition.\n''' '''You can install it with `pip install "sacrebleu>=1.4.12"`.''' ) return datasets.MetricInfo( description=_DESCRIPTION , citation=_CITATION , homepage='''https://github.com/mjpost/sacreBLEU#chrf--chrf''' , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features( { '''predictions''': datasets.Value('''string''' , id='''sequence''' ), '''references''': datasets.Sequence(datasets.Value('''string''' , id='''sequence''' ) , id='''references''' ), } ) , codebase_urls=['''https://github.com/mjpost/sacreBLEU#chrf--chrf'''] , reference_urls=[ '''https://github.com/m-popovic/chrF''', ] , ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : int = CHRF.CHAR_ORDER , SCREAMING_SNAKE_CASE__ : int = CHRF.WORD_ORDER , SCREAMING_SNAKE_CASE__ : int = CHRF.BETA , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , ) -> Optional[Any]: __lowerCamelCase = len(references[0] ) if any(len(SCREAMING_SNAKE_CASE__ ) != references_per_prediction for refs in references ): raise ValueError('''Sacrebleu requires the same number of references for each prediction''' ) __lowerCamelCase = [[refs[i] for refs in references] for i in range(SCREAMING_SNAKE_CASE__ )] __lowerCamelCase = CHRF(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = sb_chrf.corpus_score(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return { "score": output.score, "char_order": output.char_order, "word_order": output.word_order, "beta": output.beta, }
270
0
from collections import Counter from timeit import timeit def lowercase_ ( _A : str = "" , ): """simple docstring""" return sum(c % 2 for c in Counter(input_str.replace(" " , "" ).lower() ).values() ) < 2 def lowercase_ ( _A : str = "" ): """simple docstring""" if len(__lowerCAmelCase ) == 0: return True lowerCamelCase__ : Tuple = input_str.replace(" " , "" ).lower() # character_freq_dict: Stores the frequency of every character in the input string lowerCamelCase__ : List[Any] = {} for character in lower_case_input_str: lowerCamelCase__ : Tuple = character_freq_dict.get(__lowerCAmelCase , 0 ) + 1 lowerCamelCase__ : Optional[int] = 0 for character_count in character_freq_dict.values(): if character_count % 2: odd_char += 1 if odd_char > 1: return False return True def lowercase_ ( _A : str = "" ): """simple docstring""" print("\nFor string = " , __lowerCAmelCase , ":" ) print( "> can_string_be_rearranged_as_palindrome_counter()" , "\tans =" , can_string_be_rearranged_as_palindrome_counter(__lowerCAmelCase ) , "\ttime =" , timeit( "z.can_string_be_rearranged_as_palindrome_counter(z.check_str)" , setup="import __main__ as z" , ) , "seconds" , ) print( "> can_string_be_rearranged_as_palindrome()" , "\tans =" , can_string_be_rearranged_as_palindrome(__lowerCAmelCase ) , "\ttime =" , timeit( "z.can_string_be_rearranged_as_palindrome(z.check_str)" , setup="import __main__ as z" , ) , "seconds" , ) if __name__ == "__main__": A : str = input( "Enter string to determine if it can be rearranged as a palindrome or not: " ).strip() benchmark(check_str) A : str = can_string_be_rearranged_as_palindrome_counter(check_str) print(f'{check_str} can {"" if status else "not "}be rearranged as a palindrome')
184
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_tf_available, is_torch_available, ) SCREAMING_SNAKE_CASE__ : List[Any] = { "configuration_resnet": ["RESNET_PRETRAINED_CONFIG_ARCHIVE_MAP", "ResNetConfig", "ResNetOnnxConfig"] } try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : int = [ "RESNET_PRETRAINED_MODEL_ARCHIVE_LIST", "ResNetForImageClassification", "ResNetModel", "ResNetPreTrainedModel", "ResNetBackbone", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : List[Any] = [ "TF_RESNET_PRETRAINED_MODEL_ARCHIVE_LIST", "TFResNetForImageClassification", "TFResNetModel", "TFResNetPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : Union[str, Any] = [ "FlaxResNetForImageClassification", "FlaxResNetModel", "FlaxResNetPreTrainedModel", ] if TYPE_CHECKING: from .configuration_resnet import RESNET_PRETRAINED_CONFIG_ARCHIVE_MAP, ResNetConfig, ResNetOnnxConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_resnet import ( RESNET_PRETRAINED_MODEL_ARCHIVE_LIST, ResNetBackbone, ResNetForImageClassification, ResNetModel, ResNetPreTrainedModel, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_resnet import ( TF_RESNET_PRETRAINED_MODEL_ARCHIVE_LIST, TFResNetForImageClassification, TFResNetModel, TFResNetPreTrainedModel, ) try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_resnet import FlaxResNetForImageClassification, FlaxResNetModel, FlaxResNetPreTrainedModel else: import sys SCREAMING_SNAKE_CASE__ : Any = _LazyModule(__name__, globals()["__file__"], _import_structure)
270
0
"""simple docstring""" import warnings from diffusers import StableDiffusionImgaImgPipeline # noqa F401 warnings.warn( """The `image_to_image.py` script is outdated. Please use directly `from diffusers import""" """ StableDiffusionImg2ImgPipeline` instead.""" )
126
import unittest import numpy as np import requests 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 from transformers.pytorch_utils import is_torch_greater_or_equal_than_1_11 else: SCREAMING_SNAKE_CASE__ : str = False if is_vision_available(): from PIL import Image from transformers import PixaStructImageProcessor class lowerCAmelCase__ ( unittest.TestCase ): def __init__( self : Tuple , SCREAMING_SNAKE_CASE__ : Union[str, Any] , SCREAMING_SNAKE_CASE__ : List[Any]=7 , SCREAMING_SNAKE_CASE__ : List[str]=3 , SCREAMING_SNAKE_CASE__ : Tuple=18 , SCREAMING_SNAKE_CASE__ : int=30 , SCREAMING_SNAKE_CASE__ : Any=4_00 , SCREAMING_SNAKE_CASE__ : Any=None , SCREAMING_SNAKE_CASE__ : str=True , SCREAMING_SNAKE_CASE__ : str=True , SCREAMING_SNAKE_CASE__ : Optional[Any]=None , ) -> Union[str, Any]: __lowerCamelCase = size if size is not None else {'''height''': 20, '''width''': 20} __lowerCamelCase = parent __lowerCamelCase = batch_size __lowerCamelCase = num_channels __lowerCamelCase = image_size __lowerCamelCase = min_resolution __lowerCamelCase = max_resolution __lowerCamelCase = size __lowerCamelCase = do_normalize __lowerCamelCase = do_convert_rgb __lowerCamelCase = [5_12, 10_24, 20_48, 40_96] __lowerCamelCase = patch_size if patch_size is not None else {'''height''': 16, '''width''': 16} def __A ( self : Union[str, Any] ) -> Union[str, Any]: return {"do_normalize": self.do_normalize, "do_convert_rgb": self.do_convert_rgb} def __A ( self : int ) -> Any: __lowerCamelCase = '''https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/australia.jpg''' __lowerCamelCase = Image.open(requests.get(SCREAMING_SNAKE_CASE__ , stream=SCREAMING_SNAKE_CASE__ ).raw ).convert('''RGB''' ) return raw_image @unittest.skipIf( not is_torch_greater_or_equal_than_1_11 , reason="""`Pix2StructImageProcessor` requires `torch>=1.11.0`.""" , ) @require_torch @require_vision class lowerCAmelCase__ ( __lowercase , unittest.TestCase ): a__ : Dict = PixaStructImageProcessor if is_vision_available() else None def __A ( self : Optional[Any] ) -> List[Any]: __lowerCamelCase = PixaStructImageProcessingTester(self ) @property def __A ( self : Optional[int] ) -> Optional[int]: return self.image_processor_tester.prepare_image_processor_dict() def __A ( self : Tuple ) -> Dict: __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_normalize''' ) ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_convert_rgb''' ) ) def __A ( self : int ) -> str: __lowerCamelCase = self.image_processor_tester.prepare_dummy_image() __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) __lowerCamelCase = 20_48 __lowerCamelCase = image_processor(SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ) self.assertTrue(torch.allclose(inputs.flattened_patches.mean() , torch.tensor(0.0606 ) , atol=1e-3 , rtol=1e-3 ) ) def __A ( self : Union[str, Any] ) -> Dict: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PIL images __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , Image.Image ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) def __A ( self : Dict ) -> str: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PIL images __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , Image.Image ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 __lowerCamelCase = True for max_patch in self.image_processor_tester.max_patches: # Test not batched input with self.assertRaises(SCREAMING_SNAKE_CASE__ ): __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches __lowerCamelCase = '''Hello''' __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ , header_text=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ , header_text=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) def __A ( self : List[str] ) -> Any: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random numpy tensors __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ , numpify=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , np.ndarray ) __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) def __A ( self : Tuple ) -> List[str]: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PyTorch tensors __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ , torchify=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , torch.Tensor ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) @unittest.skipIf( not is_torch_greater_or_equal_than_1_11 , reason="""`Pix2StructImageProcessor` requires `torch>=1.11.0`.""" , ) @require_torch @require_vision class lowerCAmelCase__ ( __lowercase , unittest.TestCase ): a__ : Any = PixaStructImageProcessor if is_vision_available() else None def __A ( self : Union[str, Any] ) -> Optional[Any]: __lowerCamelCase = PixaStructImageProcessingTester(self , num_channels=4 ) __lowerCamelCase = 3 @property def __A ( self : List[Any] ) -> Optional[Any]: return self.image_processor_tester.prepare_image_processor_dict() def __A ( self : Union[str, Any] ) -> List[str]: __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_normalize''' ) ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_convert_rgb''' ) ) def __A ( self : Optional[int] ) -> str: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PIL images __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , Image.Image ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * (self.image_processor_tester.num_channels - 1) ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
270
0
"""simple docstring""" import os from shutil import copyfile from typing import List, Optional, Tuple from ...tokenization_utils import AddedToken from ...tokenization_utils_fast import PreTrainedTokenizerFast from ...utils import is_sentencepiece_available, logging if is_sentencepiece_available(): from .tokenization_rembert import RemBertTokenizer else: _lowercase : int = None _lowercase : Any = logging.get_logger(__name__) _lowercase : List[Any] = {"vocab_file": "sentencepiece.model", "tokenizer_file": "tokenizer.json"} _lowercase : Dict = { "vocab_file": { "google/rembert": "https://huggingface.co/google/rembert/resolve/main/sentencepiece.model", }, "tokenizer_file": { "google/rembert": "https://huggingface.co/google/rembert/resolve/main/tokenizer.json", }, } _lowercase : int = { "google/rembert": 2_5_6, } _lowercase : Optional[int] = "▁" class __SCREAMING_SNAKE_CASE ( __lowercase ): '''simple docstring''' _a = VOCAB_FILES_NAMES _a = PRETRAINED_VOCAB_FILES_MAP _a = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES _a = RemBertTokenizer def __init__( self : Optional[Any], lowerCamelCase : Optional[Any]=None, lowerCamelCase : Any=None, lowerCamelCase : Tuple=True, lowerCamelCase : int=True, lowerCamelCase : int=False, lowerCamelCase : int="[CLS]", lowerCamelCase : str="[SEP]", lowerCamelCase : Any="<unk>", lowerCamelCase : Dict="[SEP]", lowerCamelCase : List[str]="<pad>", lowerCamelCase : str="[CLS]", lowerCamelCase : List[Any]="[MASK]", **lowerCamelCase : Union[str, Any], )-> Tuple: # Mask token behave like a normal word, i.e. include the space before it lowerCamelCase__ : Union[str, Any] =AddedToken(SCREAMING_SNAKE_CASE__, lstrip=SCREAMING_SNAKE_CASE__, rstrip=SCREAMING_SNAKE_CASE__ ) if isinstance(SCREAMING_SNAKE_CASE__, SCREAMING_SNAKE_CASE__ ) else mask_token super().__init__( SCREAMING_SNAKE_CASE__, tokenizer_file=SCREAMING_SNAKE_CASE__, do_lower_case=SCREAMING_SNAKE_CASE__, remove_space=SCREAMING_SNAKE_CASE__, keep_accents=SCREAMING_SNAKE_CASE__, bos_token=SCREAMING_SNAKE_CASE__, eos_token=SCREAMING_SNAKE_CASE__, unk_token=SCREAMING_SNAKE_CASE__, sep_token=SCREAMING_SNAKE_CASE__, pad_token=SCREAMING_SNAKE_CASE__, cls_token=SCREAMING_SNAKE_CASE__, mask_token=SCREAMING_SNAKE_CASE__, **SCREAMING_SNAKE_CASE__, ) lowerCamelCase__ : Dict =do_lower_case lowerCamelCase__ : Union[str, Any] =remove_space lowerCamelCase__ : Tuple =keep_accents lowerCamelCase__ : Tuple =vocab_file lowerCamelCase__ : Optional[Any] =False if not self.vocab_file else True def snake_case ( self : Any, lowerCamelCase : List[int], lowerCamelCase : Optional[List[int]] = None )-> List[int]: lowerCamelCase__ : List[str] =[self.sep_token_id] lowerCamelCase__ : Optional[Any] =[self.cls_token_id] if token_ids_a is None: return cls + token_ids_a + sep return cls + token_ids_a + sep + token_ids_a + sep def snake_case ( self : Union[str, Any], lowerCamelCase : List[int], lowerCamelCase : Optional[List[int]] = None, lowerCamelCase : bool = False )-> List[int]: if already_has_special_tokens: if token_ids_a is not None: raise ValueError( '''You should not supply a second sequence if the provided sequence of ''' '''ids is already formatted with special tokens for the model.''' ) return [1 if x in [self.sep_token_id, self.cls_token_id] else 0 for x in token_ids_a] if token_ids_a is not None: return [1] + ([0] * len(SCREAMING_SNAKE_CASE__ )) + [1] + ([0] * len(SCREAMING_SNAKE_CASE__ )) + [1] return [1] + ([0] * len(SCREAMING_SNAKE_CASE__ )) + [1] def snake_case ( self : List[str], lowerCamelCase : List[int], lowerCamelCase : Optional[List[int]] = None )-> List[int]: lowerCamelCase__ : int =[self.sep_token_id] lowerCamelCase__ : List[Any] =[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 ) * [0] + len(token_ids_a + sep ) * [1] def snake_case ( self : Dict, lowerCamelCase : str, lowerCamelCase : Optional[str] = None )-> Tuple[str]: if not os.path.isdir(SCREAMING_SNAKE_CASE__ ): logger.error('''Vocabulary path ({}) should be a directory'''.format(SCREAMING_SNAKE_CASE__ ) ) return lowerCamelCase__ : Union[str, Any] =os.path.join( SCREAMING_SNAKE_CASE__, (filename_prefix + '''-''' if filename_prefix else '''''') + VOCAB_FILES_NAMES['''vocab_file'''] ) if os.path.abspath(self.vocab_file ) != os.path.abspath(SCREAMING_SNAKE_CASE__ ): copyfile(self.vocab_file, SCREAMING_SNAKE_CASE__ ) return (out_vocab_file,)
238
import shutil import tempfile import unittest from unittest.mock import patch from transformers import ( DefaultFlowCallback, IntervalStrategy, PrinterCallback, ProgressCallback, Trainer, TrainerCallback, TrainingArguments, is_torch_available, ) from transformers.testing_utils import require_torch if is_torch_available(): from transformers.trainer import DEFAULT_CALLBACKS from .test_trainer import RegressionDataset, RegressionModelConfig, RegressionPreTrainedModel class lowerCAmelCase__ ( __lowercase ): def __init__( self : int ) -> Optional[int]: __lowerCamelCase = [] def __A ( self : Any , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : str , **SCREAMING_SNAKE_CASE__ : List[Any] ) -> Tuple: self.events.append('''on_init_end''' ) def __A ( self : Tuple , SCREAMING_SNAKE_CASE__ : Optional[int] , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : Optional[int] , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Any: self.events.append('''on_train_begin''' ) def __A ( self : str , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Any , SCREAMING_SNAKE_CASE__ : Dict , **SCREAMING_SNAKE_CASE__ : int ) -> Any: self.events.append('''on_train_end''' ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Dict , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : List[Any] , **SCREAMING_SNAKE_CASE__ : List[Any] ) -> int: self.events.append('''on_epoch_begin''' ) def __A ( self : Optional[Any] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Any , SCREAMING_SNAKE_CASE__ : Optional[Any] , **SCREAMING_SNAKE_CASE__ : Tuple ) -> Optional[Any]: self.events.append('''on_epoch_end''' ) def __A ( self : str , SCREAMING_SNAKE_CASE__ : Tuple , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : Union[str, Any] , **SCREAMING_SNAKE_CASE__ : str ) -> List[Any]: self.events.append('''on_step_begin''' ) def __A ( self : Any , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Any , **SCREAMING_SNAKE_CASE__ : Optional[int] ) -> Tuple: self.events.append('''on_step_end''' ) def __A ( self : int , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : int , **SCREAMING_SNAKE_CASE__ : List[Any] ) -> Union[str, Any]: self.events.append('''on_evaluate''' ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str , **SCREAMING_SNAKE_CASE__ : str ) -> str: self.events.append('''on_predict''' ) def __A ( self : Optional[int] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : Optional[int] , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Tuple: self.events.append('''on_save''' ) def __A ( self : Optional[int] , SCREAMING_SNAKE_CASE__ : Dict , SCREAMING_SNAKE_CASE__ : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Union[str, Any] , **SCREAMING_SNAKE_CASE__ : List[str] ) -> List[str]: self.events.append('''on_log''' ) def __A ( self : int , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : List[Any] , **SCREAMING_SNAKE_CASE__ : Optional[int] ) -> str: self.events.append('''on_prediction_step''' ) @require_torch class lowerCAmelCase__ ( unittest.TestCase ): def __A ( self : Union[str, Any] ) -> Optional[Any]: __lowerCamelCase = tempfile.mkdtemp() def __A ( self : int ) -> List[str]: shutil.rmtree(self.output_dir ) def __A ( self : Tuple , SCREAMING_SNAKE_CASE__ : List[Any]=0 , SCREAMING_SNAKE_CASE__ : Any=0 , SCREAMING_SNAKE_CASE__ : List[str]=64 , SCREAMING_SNAKE_CASE__ : Optional[int]=64 , SCREAMING_SNAKE_CASE__ : Optional[int]=None , SCREAMING_SNAKE_CASE__ : Optional[int]=False , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Union[str, Any]: # disable_tqdm in TrainingArguments has a flaky default since it depends on the level of logging. We make sure # its set to False since the tests later on depend on its value. __lowerCamelCase = RegressionDataset(length=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = RegressionDataset(length=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = RegressionModelConfig(a=SCREAMING_SNAKE_CASE__ , b=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = RegressionPreTrainedModel(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = TrainingArguments(self.output_dir , disable_tqdm=SCREAMING_SNAKE_CASE__ , report_to=[] , **SCREAMING_SNAKE_CASE__ ) return Trainer( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , train_dataset=SCREAMING_SNAKE_CASE__ , eval_dataset=SCREAMING_SNAKE_CASE__ , callbacks=SCREAMING_SNAKE_CASE__ , ) def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str ) -> Optional[int]: self.assertEqual(len(SCREAMING_SNAKE_CASE__ ) , len(SCREAMING_SNAKE_CASE__ ) ) # Order doesn't matter __lowerCamelCase = sorted(SCREAMING_SNAKE_CASE__ , key=lambda SCREAMING_SNAKE_CASE__ : cb.__name__ if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) else cb.__class__.__name__ ) __lowerCamelCase = sorted(SCREAMING_SNAKE_CASE__ , key=lambda SCREAMING_SNAKE_CASE__ : cb.__name__ if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) else cb.__class__.__name__ ) for cba, cba in zip(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) and isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): self.assertEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) elif isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) and not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): self.assertEqual(SCREAMING_SNAKE_CASE__ , cba.__class__ ) elif not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) and isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): self.assertEqual(cba.__class__ , SCREAMING_SNAKE_CASE__ ) else: self.assertEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Optional[int] ) -> Tuple: __lowerCamelCase = ['''on_init_end''', '''on_train_begin'''] __lowerCamelCase = 0 __lowerCamelCase = len(trainer.get_eval_dataloader() ) __lowerCamelCase = ['''on_prediction_step'''] * len(trainer.get_eval_dataloader() ) + ['''on_log''', '''on_evaluate'''] for _ in range(trainer.state.num_train_epochs ): expected_events.append('''on_epoch_begin''' ) for _ in range(SCREAMING_SNAKE_CASE__ ): step += 1 expected_events += ["on_step_begin", "on_step_end"] if step % trainer.args.logging_steps == 0: expected_events.append('''on_log''' ) if trainer.args.evaluation_strategy == IntervalStrategy.STEPS and step % trainer.args.eval_steps == 0: expected_events += evaluation_events.copy() if step % trainer.args.save_steps == 0: expected_events.append('''on_save''' ) expected_events.append('''on_epoch_end''' ) if trainer.args.evaluation_strategy == IntervalStrategy.EPOCH: expected_events += evaluation_events.copy() expected_events += ["on_log", "on_train_end"] return expected_events def __A ( self : Union[str, Any] ) -> int: __lowerCamelCase = self.get_trainer() __lowerCamelCase = DEFAULT_CALLBACKS.copy() + [ProgressCallback] self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) # Callbacks passed at init are added to the default callbacks __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] ) expected_callbacks.append(SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) # TrainingArguments.disable_tqdm controls if use ProgressCallback or PrinterCallback __lowerCamelCase = self.get_trainer(disable_tqdm=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = DEFAULT_CALLBACKS.copy() + [PrinterCallback] self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) def __A ( self : List[Any] ) -> str: __lowerCamelCase = DEFAULT_CALLBACKS.copy() + [ProgressCallback] __lowerCamelCase = self.get_trainer() # We can add, pop, or remove by class name trainer.remove_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.remove(SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = self.get_trainer() __lowerCamelCase = trainer.pop_callback(SCREAMING_SNAKE_CASE__ ) self.assertEqual(cb.__class__ , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) trainer.add_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.insert(0 , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) # We can also add, pop, or remove by instance __lowerCamelCase = self.get_trainer() __lowerCamelCase = trainer.callback_handler.callbacks[0] trainer.remove_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.remove(SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = self.get_trainer() __lowerCamelCase = trainer.callback_handler.callbacks[0] __lowerCamelCase = trainer.pop_callback(SCREAMING_SNAKE_CASE__ ) self.assertEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) trainer.add_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.insert(0 , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) def __A ( self : Union[str, Any] ) -> Any: import warnings # XXX: for now ignore scatter_gather warnings in this test since it's not relevant to what's being tested warnings.simplefilter(action='''ignore''' , category=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) # Independent log/save/eval __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , logging_steps=5 ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , save_steps=5 ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , eval_steps=5 , evaluation_strategy='''steps''' ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , evaluation_strategy='''epoch''' ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) # A bit of everything __lowerCamelCase = self.get_trainer( callbacks=[MyTestTrainerCallback] , logging_steps=3 , save_steps=10 , eval_steps=5 , evaluation_strategy='''steps''' , ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) # warning should be emitted for duplicated callbacks with patch('''transformers.trainer_callback.logger.warning''' ) as warn_mock: __lowerCamelCase = self.get_trainer( callbacks=[MyTestTrainerCallback, MyTestTrainerCallback] , ) assert str(SCREAMING_SNAKE_CASE__ ) in warn_mock.call_args[0][0]
270
0
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_torch_available, ) lowercase__ : Union[str, Any] = { "configuration_encodec": [ "ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP", "EncodecConfig", ], "feature_extraction_encodec": ["EncodecFeatureExtractor"], } try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: lowercase__ : List[str] = [ "ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST", "EncodecModel", "EncodecPreTrainedModel", ] if TYPE_CHECKING: from .configuration_encodec import ( ENCODEC_PRETRAINED_CONFIG_ARCHIVE_MAP, EncodecConfig, ) from .feature_extraction_encodec import EncodecFeatureExtractor try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_encodec import ( ENCODEC_PRETRAINED_MODEL_ARCHIVE_LIST, EncodecModel, EncodecPreTrainedModel, ) else: import sys lowercase__ : Tuple = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
187
import numpy as np from cva import destroyAllWindows, imread, imshow, waitKey class lowerCAmelCase__ : def __init__( self : Any , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : int ) -> Dict: if dst_width < 0 or dst_height < 0: raise ValueError('''Destination width/height should be > 0''' ) __lowerCamelCase = img __lowerCamelCase = img.shape[1] __lowerCamelCase = img.shape[0] __lowerCamelCase = dst_width __lowerCamelCase = dst_height __lowerCamelCase = self.src_w / self.dst_w __lowerCamelCase = self.src_h / self.dst_h __lowerCamelCase = __lowerCamelCase = ( np.ones((self.dst_h, self.dst_w, 3) , np.uinta ) * 2_55 ) def __A ( self : List[Any] ) -> str: for i in range(self.dst_h ): for j in range(self.dst_w ): __lowerCamelCase = self.img[self.get_y(SCREAMING_SNAKE_CASE__ )][self.get_x(SCREAMING_SNAKE_CASE__ )] def __A ( self : str , SCREAMING_SNAKE_CASE__ : int ) -> int: return int(self.ratio_x * x ) def __A ( self : str , SCREAMING_SNAKE_CASE__ : int ) -> int: return int(self.ratio_y * y ) if __name__ == "__main__": SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ : List[str] = 800, 600 SCREAMING_SNAKE_CASE__ : int = imread("image_data/lena.jpg", 1) SCREAMING_SNAKE_CASE__ : Union[str, Any] = NearestNeighbour(im, dst_w, dst_h) n.process() imshow( F'Image resized from: {im.shape[1]}x{im.shape[0]} to {dst_w}x{dst_h}', n.output ) waitKey(0) destroyAllWindows()
270
0
def __lowerCamelCase ( snake_case__ ) -> list: """simple docstring""" _SCREAMING_SNAKE_CASE = len(__lowerCAmelCase ) for i in range(1 ,__lowerCAmelCase ): _SCREAMING_SNAKE_CASE = collection[i] _SCREAMING_SNAKE_CASE = 0 _SCREAMING_SNAKE_CASE = i - 1 while low <= high: _SCREAMING_SNAKE_CASE = (low + high) // 2 if val < collection[mid]: _SCREAMING_SNAKE_CASE = mid - 1 else: _SCREAMING_SNAKE_CASE = mid + 1 for j in range(__lowerCAmelCase ,__lowerCAmelCase ,-1 ): _SCREAMING_SNAKE_CASE = collection[j - 1] _SCREAMING_SNAKE_CASE = val return collection if __name__ == "__main__": UpperCamelCase = input('''Enter numbers separated by a comma:\n''').strip() UpperCamelCase = [int(item) for item in user_input.split(''',''')] print(binary_insertion_sort(unsorted))
306
from queue import PriorityQueue from typing import Any import numpy as np def __magic_name__ ( __lowerCAmelCase : dict , __lowerCAmelCase : str , __lowerCAmelCase : set , __lowerCAmelCase : set , __lowerCAmelCase : dict , __lowerCAmelCase : dict , __lowerCAmelCase : PriorityQueue , __lowerCAmelCase : dict , __lowerCAmelCase : float | int , ) -> float | int: for nxt, d in graph[v]: if nxt in visited_forward: continue __lowerCamelCase = cst_fwd.get(__lowerCAmelCase , np.inf ) __lowerCamelCase = cst_fwd[v] + d if new_cost_f < old_cost_f: queue.put((new_cost_f, nxt) ) __lowerCamelCase = new_cost_f __lowerCamelCase = v if nxt in visited_backward: if cst_fwd[v] + d + cst_bwd[nxt] < shortest_distance: __lowerCamelCase = cst_fwd[v] + d + cst_bwd[nxt] return shortest_distance def __magic_name__ ( __lowerCAmelCase : str , __lowerCAmelCase : str , __lowerCAmelCase : dict , __lowerCAmelCase : dict ) -> int: __lowerCamelCase = -1 __lowerCamelCase = set() __lowerCamelCase = set() __lowerCamelCase = {source: 0} __lowerCamelCase = {destination: 0} __lowerCamelCase = {source: None} __lowerCamelCase = {destination: None} __lowerCamelCase = PriorityQueue() __lowerCamelCase = PriorityQueue() __lowerCamelCase = np.inf queue_forward.put((0, source) ) queue_backward.put((0, destination) ) if source == destination: return 0 while not queue_forward.empty() and not queue_backward.empty(): __lowerCamelCase , __lowerCamelCase = queue_forward.get() visited_forward.add(__lowerCAmelCase ) __lowerCamelCase , __lowerCamelCase = queue_backward.get() visited_backward.add(__lowerCAmelCase ) __lowerCamelCase = pass_and_relaxation( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , ) __lowerCamelCase = pass_and_relaxation( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , ) if cst_fwd[v_fwd] + cst_bwd[v_bwd] >= shortest_distance: break if shortest_distance != np.inf: __lowerCamelCase = shortest_distance return shortest_path_distance SCREAMING_SNAKE_CASE__ : List[Any] = { "B": [["C", 1]], "C": [["D", 1]], "D": [["F", 1]], "E": [["B", 1], ["G", 2]], "F": [], "G": [["F", 1]], } SCREAMING_SNAKE_CASE__ : Optional[int] = { "B": [["E", 1]], "C": [["B", 1]], "D": [["C", 1]], "F": [["D", 1], ["G", 1]], "E": [[None, np.inf]], "G": [["E", 2]], } if __name__ == "__main__": import doctest doctest.testmod()
270
0
'''simple docstring''' import inspect import jax import jax.lax as lax import jax.numpy as jnp from ..utils import add_start_docstrings from ..utils.logging import get_logger lowerCAmelCase : Dict =get_logger(__name__) lowerCAmelCase : List[str] =r"\n Args:\n input_ids (`jnp.ndarray` of shape `(batch_size, sequence_length)`):\n Indices of input sequence tokens in the vocabulary.\n\n Indices can be obtained using [`PreTrainedTokenizer`]. See [`PreTrainedTokenizer.encode`] and\n [`PreTrainedTokenizer.__call__`] for details.\n\n [What are input IDs?](../glossary#input-ids)\n scores (`jnp.ndarray` of shape `(batch_size, config.vocab_size)`):\n Prediction scores of a language modeling head. These can be logits for each vocabulary when not using beam\n search or log softmax for each vocabulary token when using beam search\n kwargs (`Dict[str, Any]`, *optional*):\n Additional logits processor specific kwargs.\n\n Return:\n `jnp.ndarray` of shape `(batch_size, config.vocab_size)`: The processed prediction scores.\n\n" class a_ : @add_start_docstrings(SCREAMING_SNAKE_CASE__ ) def __call__( self : int , lowercase : jnp.ndarray , lowercase : jnp.ndarray ): """simple docstring""" raise NotImplementedError( F'{self.__class__} is an abstract class. Only classes inheriting this class can be called.' ) class a_ : @add_start_docstrings(SCREAMING_SNAKE_CASE__ ) def __call__( self : Optional[Any] , lowercase : jnp.ndarray , lowercase : jnp.ndarray ): """simple docstring""" raise NotImplementedError( F'{self.__class__} is an abstract class. Only classes inheriting this class can be called.' ) class a_ ( __lowercase ): @add_start_docstrings(SCREAMING_SNAKE_CASE__ ) def __call__( self : Union[str, Any] , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int , **lowercase : Union[str, Any] ): """simple docstring""" for processor in self: lowercase_ :Optional[int] = inspect.signature(processor.__call__ ).parameters if len(SCREAMING_SNAKE_CASE__ ) > 3: if not all(arg in kwargs for arg in list(function_args.keys() )[2:] ): raise ValueError( F'Make sure that all the required parameters: {list(function_args.keys() )} for ' F'{processor.__class__} are passed to the logits processor.' ) lowercase_ :int = processor(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) else: lowercase_ :str = processor(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return scores class a_ ( __lowercase ): def __init__( self : int , lowercase : float ): """simple docstring""" if not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) or not (temperature > 0): raise ValueError(F'`temperature` has to be a strictly positive float, but is {temperature}' ) lowercase_ :int = temperature def __call__( self : List[str] , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ :Optional[int] = scores / self.temperature return scores class a_ ( __lowercase ): def __init__( self : Tuple , lowercase : float , lowercase : float = -float("Inf" ) , lowercase : int = 1 ): """simple docstring""" if not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) or (top_p < 0 or top_p > 1.0): raise ValueError(F'`top_p` has to be a float > 0 and < 1, but is {top_p}' ) if not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) or (min_tokens_to_keep < 1): raise ValueError(F'`min_tokens_to_keep` has to be a positive integer, but is {min_tokens_to_keep}' ) lowercase_ :Optional[int] = top_p lowercase_ :List[Any] = filter_value lowercase_ :Union[str, Any] = min_tokens_to_keep def __call__( self : Tuple , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ , lowercase_ :Optional[Any] = lax.top_k(SCREAMING_SNAKE_CASE__ , scores.shape[-1] ) lowercase_ :List[Any] = jnp.full_like(SCREAMING_SNAKE_CASE__ , self.filter_value ) lowercase_ :List[str] = jax.nn.softmax(SCREAMING_SNAKE_CASE__ , axis=-1 ).cumsum(axis=-1 ) lowercase_ :Any = cumulative_probs < self.top_p # include the token that is higher than top_p as well lowercase_ :List[Any] = jnp.roll(SCREAMING_SNAKE_CASE__ , 1 ) score_mask |= score_mask.at[:, 0].set(SCREAMING_SNAKE_CASE__ ) # min tokens to keep lowercase_ :Any = score_mask.at[:, : self.min_tokens_to_keep].set(SCREAMING_SNAKE_CASE__ ) lowercase_ :Optional[Any] = jnp.where(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :List[Any] = jax.lax.sort_key_val(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )[-1] return next_scores class a_ ( __lowercase ): def __init__( self : Optional[Any] , lowercase : int , lowercase : float = -float("Inf" ) , lowercase : int = 1 ): """simple docstring""" if not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) or top_k <= 0: raise ValueError(F'`top_k` has to be a strictly positive integer, but is {top_k}' ) lowercase_ :Union[str, Any] = max(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :List[Any] = filter_value def __call__( self : Optional[Any] , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ , lowercase_ :Union[str, Any] = scores.shape lowercase_ :List[Any] = jnp.full(batch_size * vocab_size , self.filter_value ) lowercase_ :Dict = min(self.top_k , scores.shape[-1] ) # Safety check lowercase_ , lowercase_ :List[str] = lax.top_k(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :Dict = jnp.broadcast_to((jnp.arange(SCREAMING_SNAKE_CASE__ ) * vocab_size)[:, None] , (batch_size, topk) ).flatten() lowercase_ :List[Any] = topk_scores.flatten() lowercase_ :Tuple = topk_indices.flatten() + shift lowercase_ :str = next_scores_flat.at[topk_indices_flat].set(SCREAMING_SNAKE_CASE__ ) lowercase_ :Any = next_scores_flat.reshape(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return next_scores class a_ ( __lowercase ): def __init__( self : Any , lowercase : int ): """simple docstring""" lowercase_ :Optional[Any] = bos_token_id def __call__( self : Dict , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ :Optional[Any] = jnp.full(scores.shape , -float("inf" ) ) lowercase_ :Any = 1 - jnp.bool_(cur_len - 1 ) lowercase_ :Any = jnp.where(SCREAMING_SNAKE_CASE__ , new_scores.at[:, self.bos_token_id].set(0 ) , SCREAMING_SNAKE_CASE__ ) return scores class a_ ( __lowercase ): def __init__( self : Optional[Any] , lowercase : int , lowercase : int ): """simple docstring""" lowercase_ :List[str] = max_length lowercase_ :Any = eos_token_id def __call__( self : List[str] , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ :List[Any] = jnp.full(scores.shape , -float("inf" ) ) lowercase_ :Any = 1 - jnp.bool_(cur_len - self.max_length + 1 ) lowercase_ :str = jnp.where(SCREAMING_SNAKE_CASE__ , new_scores.at[:, self.eos_token_id].set(0 ) , SCREAMING_SNAKE_CASE__ ) return scores class a_ ( __lowercase ): def __init__( self : str , lowercase : int , lowercase : int ): """simple docstring""" if not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) or min_length < 0: raise ValueError(F'`min_length` has to be a positive integer, but is {min_length}' ) if not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) or eos_token_id < 0: raise ValueError(F'`eos_token_id` has to be a positive integer, but is {eos_token_id}' ) lowercase_ :Optional[int] = min_length lowercase_ :List[Any] = eos_token_id def __call__( self : int , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ :Tuple = 1 - jnp.clip(cur_len - self.min_length , 0 , 1 ) lowercase_ :Optional[int] = jnp.where(SCREAMING_SNAKE_CASE__ , scores.at[:, self.eos_token_id].set(-float("inf" ) ) , SCREAMING_SNAKE_CASE__ ) return scores class a_ ( __lowercase ): def __init__( self : Any , lowercase : List[str] , lowercase : List[str] ): """simple docstring""" lowercase_ :Optional[int] = list(SCREAMING_SNAKE_CASE__ ) lowercase_ :int = begin_index def __call__( self : Optional[Any] , lowercase : str , lowercase : int , lowercase : int ): """simple docstring""" lowercase_ :str = 1 - jnp.bool_(cur_len - self.begin_index ) lowercase_ :Optional[Any] = jnp.where(SCREAMING_SNAKE_CASE__ , scores.at[:, self.begin_suppress_tokens].set(-float("inf" ) ) , SCREAMING_SNAKE_CASE__ ) return scores class a_ ( __lowercase ): def __init__( self : Any , lowercase : list ): """simple docstring""" lowercase_ :Any = list(SCREAMING_SNAKE_CASE__ ) def __call__( self : Dict , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" lowercase_ :Tuple = scores.at[..., self.suppress_tokens].set(-float("inf" ) ) return scores class a_ ( __lowercase ): def __init__( self : Tuple , lowercase : Tuple ): """simple docstring""" lowercase_ :Tuple = dict(SCREAMING_SNAKE_CASE__ ) # Converts the dictionary of format {index: token} containing the tokens to be forced to an array, where the # index of the array corresponds to the index of the token to be forced, for XLA compatibility. # Indexes without forced tokens will have a negative value. lowercase_ :Optional[Any] = jnp.ones((max(force_token_map.keys() ) + 1) , dtype=jnp.intaa ) * -1 for index, token in force_token_map.items(): if token is not None: lowercase_ :Optional[Any] = force_token_array.at[index].set(SCREAMING_SNAKE_CASE__ ) lowercase_ :Tuple = jnp.intaa(SCREAMING_SNAKE_CASE__ ) def __call__( self : int , lowercase : jnp.ndarray , lowercase : jnp.ndarray , lowercase : int ): """simple docstring""" def _force_token(lowercase : Optional[int] ): lowercase_ :Optional[int] = scores.shape[0] lowercase_ :List[Any] = self.force_token_array[generation_idx] lowercase_ :str = jnp.ones_like(SCREAMING_SNAKE_CASE__ , dtype=scores.dtype ) * -float("inf" ) lowercase_ :List[str] = jnp.zeros((batch_size, 1) , dtype=scores.dtype ) lowercase_ :Any = lax.dynamic_update_slice(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , (0, current_token) ) return new_scores lowercase_ :str = lax.cond( cur_len >= self.force_token_array.shape[0] , lambda: scores , lambda: lax.cond( self.force_token_array[cur_len] >= 0 , lambda: _force_token(SCREAMING_SNAKE_CASE__ ) , lambda: scores , ) , ) return scores class a_ ( __lowercase ): def __init__( self : Union[str, Any] , lowercase : Optional[Any] , lowercase : int , lowercase : int ): """simple docstring""" lowercase_ :Optional[Any] = generate_config.eos_token_id lowercase_ :int = generate_config.no_timestamps_token_id lowercase_ :List[str] = generate_config.no_timestamps_token_id + 1 lowercase_ :List[Any] = decoder_input_length + 1 if generate_config.is_multilingual: # room for language token and task token self.begin_index += 2 if hasattr(SCREAMING_SNAKE_CASE__ , "max_initial_timestamp_index" ): lowercase_ :Optional[int] = generate_config.max_initial_timestamp_index else: lowercase_ :Optional[int] = model_config.vocab_size if self.max_initial_timestamp_index is None: lowercase_ :str = model_config.vocab_size def __call__( self : List[str] , lowercase : List[str] , lowercase : List[Any] , lowercase : List[Any] ): """simple docstring""" lowercase_ :Dict = scores.at[:, self.no_timestamps_token_id].set(-float("inf" ) ) def handle_pairs(lowercase : Union[str, Any] , lowercase : Tuple ): lowercase_ :Tuple = jnp.where((cur_len - self.begin_index) >= 1 , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :Optional[int] = jnp.where( input_ids_k[cur_len - 1] >= self.timestamp_begin , True and last_was_timestamp , SCREAMING_SNAKE_CASE__ , ) lowercase_ :Dict = jnp.where((cur_len - self.begin_index) < 2 , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :str = jnp.where( input_ids_k[cur_len - 2] >= self.timestamp_begin , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , ) return jnp.where( SCREAMING_SNAKE_CASE__ , jnp.where( penultimate_was_timestamp > 0 , scores_k.at[self.timestamp_begin :].set(-float("inf" ) ) , scores_k.at[: self.eos_token_id].set(-float("inf" ) ) , ) , SCREAMING_SNAKE_CASE__ , ) lowercase_ :List[str] = jax.vmap(SCREAMING_SNAKE_CASE__ )(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :int = jnp.where(cur_len == self.begin_index , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) lowercase_ :str = jnp.where( self.max_initial_timestamp_index is not None , True and apply_max_initial_timestamp , SCREAMING_SNAKE_CASE__ , ) lowercase_ :Optional[int] = self.timestamp_begin + self.max_initial_timestamp_index lowercase_ :Any = jnp.where( SCREAMING_SNAKE_CASE__ , scores.at[:, last_allowed + 1 :].set(-float("inf" ) ) , SCREAMING_SNAKE_CASE__ , ) # if sum of probability over timestamps is above any other token, sample timestamp lowercase_ :Optional[Any] = jax.nn.log_softmax(SCREAMING_SNAKE_CASE__ , axis=-1 ) def handle_cumulative_probs(lowercase : Dict , lowercase : str ): lowercase_ :Optional[int] = jax.nn.logsumexp(logprobs_k[self.timestamp_begin :] , axis=-1 ) lowercase_ :str = jnp.max(logprobs_k[: self.timestamp_begin] ) return jnp.where( timestamp_logprob > max_text_token_logprob , scores_k.at[: self.timestamp_begin].set(-float("inf" ) ) , SCREAMING_SNAKE_CASE__ , ) lowercase_ :List[Any] = jax.vmap(SCREAMING_SNAKE_CASE__ )(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return scores
223
from __future__ import annotations from sys import maxsize from typing import Generic, TypeVar SCREAMING_SNAKE_CASE__ : List[Any] = TypeVar("T") def __magic_name__ ( __lowerCAmelCase : int ) -> int: return (position - 1) // 2 def __magic_name__ ( __lowerCAmelCase : int ) -> int: return (2 * position) + 1 def __magic_name__ ( __lowerCAmelCase : int ) -> int: return (2 * position) + 2 class lowerCAmelCase__ ( Generic[T] ): def __init__( self : List[str] ) -> None: __lowerCamelCase = [] __lowerCamelCase = {} __lowerCamelCase = 0 def __len__( self : Optional[int] ) -> int: return self.elements def __repr__( self : Optional[int] ) -> str: return str(self.heap ) def __A ( self : Union[str, Any] ) -> bool: # Check if the priority queue is empty return self.elements == 0 def __A ( self : str , SCREAMING_SNAKE_CASE__ : T , SCREAMING_SNAKE_CASE__ : int ) -> None: # Add an element with given priority to the queue self.heap.append((elem, weight) ) __lowerCamelCase = self.elements self.elements += 1 self._bubble_up(SCREAMING_SNAKE_CASE__ ) def __A ( self : Any ) -> T: # Remove and return the element with lowest weight (highest priority) if self.elements > 1: self._swap_nodes(0 , self.elements - 1 ) __lowerCamelCase , __lowerCamelCase = self.heap.pop() del self.position_map[elem] self.elements -= 1 if self.elements > 0: __lowerCamelCase , __lowerCamelCase = self.heap[0] self._bubble_down(SCREAMING_SNAKE_CASE__ ) return elem def __A ( self : Dict , SCREAMING_SNAKE_CASE__ : T , SCREAMING_SNAKE_CASE__ : int ) -> None: # Update the weight of the given key __lowerCamelCase = self.position_map[elem] __lowerCamelCase = (elem, weight) if position > 0: __lowerCamelCase = get_parent_position(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase , __lowerCamelCase = self.heap[parent_position] if parent_weight > weight: self._bubble_up(SCREAMING_SNAKE_CASE__ ) else: self._bubble_down(SCREAMING_SNAKE_CASE__ ) else: self._bubble_down(SCREAMING_SNAKE_CASE__ ) def __A ( self : Optional[int] , SCREAMING_SNAKE_CASE__ : T ) -> None: # Place a node at the proper position (upward movement) [to be used internally # only] __lowerCamelCase = self.position_map[elem] if curr_pos == 0: return None __lowerCamelCase = get_parent_position(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase , __lowerCamelCase = self.heap[curr_pos] __lowerCamelCase , __lowerCamelCase = self.heap[parent_position] if parent_weight > weight: self._swap_nodes(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return self._bubble_up(SCREAMING_SNAKE_CASE__ ) return None def __A ( self : Optional[int] , SCREAMING_SNAKE_CASE__ : T ) -> None: # Place a node at the proper position (downward movement) [to be used # internally only] __lowerCamelCase = self.position_map[elem] __lowerCamelCase , __lowerCamelCase = self.heap[curr_pos] __lowerCamelCase = get_child_left_position(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = get_child_right_position(SCREAMING_SNAKE_CASE__ ) if child_left_position < self.elements and child_right_position < self.elements: __lowerCamelCase , __lowerCamelCase = self.heap[child_left_position] __lowerCamelCase , __lowerCamelCase = self.heap[child_right_position] if child_right_weight < child_left_weight and child_right_weight < weight: self._swap_nodes(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return self._bubble_down(SCREAMING_SNAKE_CASE__ ) if child_left_position < self.elements: __lowerCamelCase , __lowerCamelCase = self.heap[child_left_position] if child_left_weight < weight: self._swap_nodes(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return self._bubble_down(SCREAMING_SNAKE_CASE__ ) else: return None if child_right_position < self.elements: __lowerCamelCase , __lowerCamelCase = self.heap[child_right_position] if child_right_weight < weight: self._swap_nodes(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return self._bubble_down(SCREAMING_SNAKE_CASE__ ) return None def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : int ) -> None: # Swap the nodes at the given positions __lowerCamelCase = self.heap[nodea_pos][0] __lowerCamelCase = self.heap[nodea_pos][0] __lowerCamelCase , __lowerCamelCase = ( self.heap[nodea_pos], self.heap[nodea_pos], ) __lowerCamelCase = nodea_pos __lowerCamelCase = nodea_pos class lowerCAmelCase__ ( Generic[T] ): def __init__( self : Tuple ) -> None: __lowerCamelCase = {} __lowerCamelCase = 0 def __repr__( self : Optional[int] ) -> str: return str(self.connections ) def __len__( self : List[str] ) -> int: return self.nodes def __A ( self : List[str] , SCREAMING_SNAKE_CASE__ : T ) -> None: # Add a node in the graph if it is not in the graph if node not in self.connections: __lowerCamelCase = {} self.nodes += 1 def __A ( self : Dict , SCREAMING_SNAKE_CASE__ : T , SCREAMING_SNAKE_CASE__ : T , SCREAMING_SNAKE_CASE__ : int ) -> None: # Add an edge between 2 nodes in the graph self.add_node(SCREAMING_SNAKE_CASE__ ) self.add_node(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = weight __lowerCamelCase = weight def __magic_name__ ( __lowerCAmelCase : GraphUndirectedWeighted[T] , ) -> tuple[dict[T, int], dict[T, T | None]]: __lowerCamelCase = {node: maxsize for node in graph.connections} __lowerCamelCase = {node: None for node in graph.connections} __lowerCamelCase = MinPriorityQueue() for node, weight in dist.items(): priority_queue.push(__lowerCAmelCase , __lowerCAmelCase ) if priority_queue.is_empty(): return dist, parent # initialization __lowerCamelCase = priority_queue.extract_min() __lowerCamelCase = 0 for neighbour in graph.connections[node]: if dist[neighbour] > dist[node] + graph.connections[node][neighbour]: __lowerCamelCase = dist[node] + graph.connections[node][neighbour] priority_queue.update_key(__lowerCAmelCase , dist[neighbour] ) __lowerCamelCase = node # running prim's algorithm while not priority_queue.is_empty(): __lowerCamelCase = priority_queue.extract_min() for neighbour in graph.connections[node]: if dist[neighbour] > dist[node] + graph.connections[node][neighbour]: __lowerCamelCase = dist[node] + graph.connections[node][neighbour] priority_queue.update_key(__lowerCAmelCase , dist[neighbour] ) __lowerCamelCase = node return dist, parent
270
0
'''simple docstring''' import inspect import os import re from transformers.configuration_utils import PretrainedConfig from transformers.utils import direct_transformers_import # All paths are set with the intent you should run this script from the root of the repo with the command # python utils/check_config_docstrings.py __lowerCAmelCase : List[str] ="src/transformers" # This is to make sure the transformers module imported is the one in the repo. __lowerCAmelCase : Dict =direct_transformers_import(PATH_TO_TRANSFORMERS) __lowerCAmelCase : int =transformers.models.auto.configuration_auto.CONFIG_MAPPING __lowerCAmelCase : Dict ={ # used to compute the property `self.chunk_length` "EncodecConfig": ["overlap"], # used as `self.bert_model = BertModel(config, ...)` "DPRConfig": True, # not used in modeling files, but it's an important information "FSMTConfig": ["langs"], # used internally in the configuration class file "GPTNeoConfig": ["attention_types"], # used internally in the configuration class file "EsmConfig": ["is_folding_model"], # used during training (despite we don't have training script for these models yet) "Mask2FormerConfig": ["ignore_value"], # `ignore_value` used during training (despite we don't have training script for these models yet) # `norm` used in conversion script (despite not using in the modeling file) "OneFormerConfig": ["ignore_value", "norm"], # used during preprocessing and collation, see `collating_graphormer.py` "GraphormerConfig": ["spatial_pos_max"], # used internally in the configuration class file "T5Config": ["feed_forward_proj"], # used internally in the configuration class file # `tokenizer_class` get default value `T5Tokenizer` intentionally "MT5Config": ["feed_forward_proj", "tokenizer_class"], "UMT5Config": ["feed_forward_proj", "tokenizer_class"], # used internally in the configuration class file "LongT5Config": ["feed_forward_proj"], # used internally in the configuration class file "SwitchTransformersConfig": ["feed_forward_proj"], # having default values other than `1e-5` - we can't fix them without breaking "BioGptConfig": ["layer_norm_eps"], # having default values other than `1e-5` - we can't fix them without breaking "GLPNConfig": ["layer_norm_eps"], # having default values other than `1e-5` - we can't fix them without breaking "SegformerConfig": ["layer_norm_eps"], # having default values other than `1e-5` - we can't fix them without breaking "CvtConfig": ["layer_norm_eps"], # having default values other than `1e-5` - we can't fix them without breaking "PerceiverConfig": ["layer_norm_eps"], # used internally to calculate the feature size "InformerConfig": ["num_static_real_features", "num_time_features"], # used internally to calculate the feature size "TimeSeriesTransformerConfig": ["num_static_real_features", "num_time_features"], # used internally to calculate the feature size "AutoformerConfig": ["num_static_real_features", "num_time_features"], # used internally to calculate `mlp_dim` "SamVisionConfig": ["mlp_ratio"], # For (head) training, but so far not implemented "ClapAudioConfig": ["num_classes"], # Not used, but providing useful information to users "SpeechT5HifiGanConfig": ["sampling_rate"], } # TODO (ydshieh): Check the failing cases, try to fix them or move some cases to the above block once we are sure SPECIAL_CASES_TO_ALLOW.update( { "CLIPSegConfig": True, "DeformableDetrConfig": True, "DetaConfig": True, "DinatConfig": True, "DonutSwinConfig": True, "EfficientFormerConfig": True, "FSMTConfig": True, "JukeboxConfig": True, "LayoutLMv2Config": True, "MaskFormerSwinConfig": True, "MT5Config": True, "NatConfig": True, "OneFormerConfig": True, "PerceiverConfig": True, "RagConfig": True, "SpeechT5Config": True, "SwinConfig": True, "Swin2SRConfig": True, "Swinv2Config": True, "SwitchTransformersConfig": True, "TableTransformerConfig": True, "TapasConfig": True, "TransfoXLConfig": True, "UniSpeechConfig": True, "UniSpeechSatConfig": True, "WavLMConfig": True, "WhisperConfig": True, # TODO: @Arthur (for `alignment_head` and `alignment_layer`) "JukeboxPriorConfig": True, # TODO: @Younes (for `is_decoder`) "Pix2StructTextConfig": True, } ) def UpperCamelCase ( _lowerCamelCase : Any , _lowerCamelCase : Tuple , _lowerCamelCase : List[str] , _lowerCamelCase : Tuple ): A__ = False for attribute in attributes: for modeling_source in source_strings: # check if we can find `config.xxx`, `getattr(config, "xxx", ...)` or `getattr(self.config, "xxx", ...)` if ( F"config.{attribute}" in modeling_source or F"getattr(config, \"{attribute}\"" in modeling_source or F"getattr(self.config, \"{attribute}\"" in modeling_source ): A__ = True # Deal with multi-line cases elif ( re.search( rF"getattr[ \t\v\n\r\f]*\([ \t\v\n\r\f]*(self\.)?config,[ \t\v\n\r\f]*\"{attribute}\"" , __lowerCAmelCase , ) is not None ): A__ = True # `SequenceSummary` is called with `SequenceSummary(config)` elif attribute in [ "summary_type", "summary_use_proj", "summary_activation", "summary_last_dropout", "summary_proj_to_labels", "summary_first_dropout", ]: if "SequenceSummary" in modeling_source: A__ = True if attribute_used: break if attribute_used: break # common and important attributes, even if they do not always appear in the modeling files A__ = [ "bos_index", "eos_index", "pad_index", "unk_index", "mask_index", "image_size", "use_cache", "out_features", "out_indices", ] A__ = ["encoder_no_repeat_ngram_size"] # Special cases to be allowed A__ = True if not attribute_used: A__ = False for attribute in attributes: # Allow if the default value in the configuration class is different from the one in `PretrainedConfig` if attribute in ["is_encoder_decoder"] and default_value is True: A__ = True elif attribute in ["tie_word_embeddings"] and default_value is False: A__ = True # Allow cases without checking the default value in the configuration class elif attribute in attributes_to_allow + attributes_used_in_generation: A__ = True elif attribute.endswith("_token_id" ): A__ = True # configuration class specific cases if not case_allowed: A__ = SPECIAL_CASES_TO_ALLOW.get(config_class.__name__ , [] ) A__ = allowed_cases is True or attribute in allowed_cases return attribute_used or case_allowed def UpperCamelCase ( _lowerCamelCase : str ): A__ = dict(inspect.signature(config_class.__init__ ).parameters ) A__ = [x for x in list(signature.keys() ) if x not in ["self", "kwargs"]] A__ = [signature[param].default for param in parameter_names] # If `attribute_map` exists, an attribute can have different names to be used in the modeling files, and as long # as one variant is used, the test should pass A__ = {} if len(config_class.attribute_map ) > 0: A__ = {v: k for k, v in config_class.attribute_map.items()} # Get the path to modeling source files A__ = inspect.getsourcefile(__lowerCAmelCase ) A__ = os.path.dirname(__lowerCAmelCase ) # Let's check against all frameworks: as long as one framework uses an attribute, we are good. A__ = [os.path.join(__lowerCAmelCase , __lowerCAmelCase ) for fn in os.listdir(__lowerCAmelCase ) if fn.startswith("modeling_" )] # Get the source code strings A__ = [] for path in modeling_paths: if os.path.isfile(__lowerCAmelCase ): with open(__lowerCAmelCase ) as fp: modeling_sources.append(fp.read() ) A__ = [] for config_param, default_value in zip(__lowerCAmelCase , __lowerCAmelCase ): # `attributes` here is all the variant names for `config_param` A__ = [config_param] # some configuration classes have non-empty `attribute_map`, and both names could be used in the # corresponding modeling files. As long as one of them appears, it is fine. if config_param in reversed_attribute_map: attributes.append(reversed_attribute_map[config_param] ) if not check_attribute_being_used(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ): unused_attributes.append(attributes[0] ) return sorted(__lowerCAmelCase ) def UpperCamelCase ( ): A__ = {} for _config_class in list(CONFIG_MAPPING.values() ): # Skip deprecated models if "models.deprecated" in _config_class.__module__: continue # Some config classes are not in `CONFIG_MAPPING` (e.g. `CLIPVisionConfig`, `Blip2VisionConfig`, etc.) A__ = [ cls for name, cls in inspect.getmembers( inspect.getmodule(_config_class ) , lambda _lowerCamelCase : inspect.isclass(__lowerCAmelCase ) and issubclass(__lowerCAmelCase , __lowerCAmelCase ) and inspect.getmodule(__lowerCAmelCase ) == inspect.getmodule(_config_class ) , ) ] for config_class in config_classes_in_module: A__ = check_config_attributes_being_used(__lowerCAmelCase ) if len(__lowerCAmelCase ) > 0: A__ = unused_attributes if len(__lowerCAmelCase ) > 0: A__ = "The following configuration classes contain unused attributes in the corresponding modeling files:\n" for name, attributes in configs_with_unused_attributes.items(): error += F"{name}: {attributes}\n" raise ValueError(__lowerCAmelCase ) if __name__ == "__main__": check_config_attributes()
237
from __future__ import annotations from statistics import mean def __magic_name__ ( __lowerCAmelCase : list[int] , __lowerCAmelCase : list[int] , __lowerCAmelCase : int ) -> list[int]: __lowerCamelCase = [0] * no_of_processes __lowerCamelCase = [0] * no_of_processes # Initialize remaining_time to waiting_time. for i in range(__lowerCAmelCase ): __lowerCamelCase = burst_time[i] __lowerCamelCase = [] __lowerCamelCase = 0 __lowerCamelCase = 0 # When processes are not completed, # A process whose arrival time has passed \ # and has remaining execution time is put into the ready_process. # The shortest process in the ready_process, target_process is executed. while completed != no_of_processes: __lowerCamelCase = [] __lowerCamelCase = -1 for i in range(__lowerCAmelCase ): if (arrival_time[i] <= total_time) and (remaining_time[i] > 0): ready_process.append(__lowerCAmelCase ) if len(__lowerCAmelCase ) > 0: __lowerCamelCase = ready_process[0] for i in ready_process: if remaining_time[i] < remaining_time[target_process]: __lowerCamelCase = i total_time += burst_time[target_process] completed += 1 __lowerCamelCase = 0 __lowerCamelCase = ( total_time - arrival_time[target_process] - burst_time[target_process] ) else: total_time += 1 return waiting_time def __magic_name__ ( __lowerCAmelCase : list[int] , __lowerCAmelCase : int , __lowerCAmelCase : list[int] ) -> list[int]: __lowerCamelCase = [0] * no_of_processes for i in range(__lowerCAmelCase ): __lowerCamelCase = burst_time[i] + waiting_time[i] return turn_around_time if __name__ == "__main__": print("[TEST CASE 01]") SCREAMING_SNAKE_CASE__ : Tuple = 4 SCREAMING_SNAKE_CASE__ : Optional[int] = [2, 5, 3, 7] SCREAMING_SNAKE_CASE__ : List[str] = [0, 0, 0, 0] SCREAMING_SNAKE_CASE__ : str = calculate_waitingtime(arrival_time, burst_time, no_of_processes) SCREAMING_SNAKE_CASE__ : Union[str, Any] = calculate_turnaroundtime( burst_time, no_of_processes, waiting_time ) # Printing the Result print("PID\tBurst Time\tArrival Time\tWaiting Time\tTurnaround Time") for i, process_id in enumerate(list(range(1, 5))): print( F'{process_id}\t{burst_time[i]}\t\t\t{arrival_time[i]}\t\t\t\t' F'{waiting_time[i]}\t\t\t\t{turn_around_time[i]}' ) print(F'\nAverage waiting time = {mean(waiting_time):.5f}') print(F'Average turnaround time = {mean(turn_around_time):.5f}')
270
0
import warnings from ...processing_utils import ProcessorMixin from ...tokenization_utils_base import BatchEncoding class a__ ( __lowercase ): """simple docstring""" __lowerCamelCase = ["""image_processor""", """tokenizer"""] __lowerCamelCase = """ViTImageProcessor""" __lowerCamelCase = ("""CLIPTokenizer""", """CLIPTokenizerFast""") def __init__( self , lowercase=None , lowercase=None , **lowercase ) -> List[str]: '''simple docstring''' A__ = None if "feature_extractor" in kwargs: warnings.warn( "The `feature_extractor` argument is deprecated and will be removed in v5, use `image_processor`" " instead." , SCREAMING_SNAKE_CASE__ , ) A__ = kwargs.pop("feature_extractor" ) A__ = image_processor if image_processor is not None else feature_extractor if image_processor is None: raise ValueError("You need to specify an `image_processor`." ) if tokenizer is None: raise ValueError("You need to specify a `tokenizer`." ) super().__init__(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def __call__( self , lowercase=None , lowercase=None , lowercase=None , lowercase=None , **lowercase ) -> Optional[int]: '''simple docstring''' if text is None and visual_prompt is None and images is None: raise ValueError("You have to specify either text, visual prompt or images." ) if text is not None and visual_prompt is not None: raise ValueError("You have to specify exactly one type of prompt. Either text or visual prompt." ) if text is not None: A__ = self.tokenizer(SCREAMING_SNAKE_CASE__ , return_tensors=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) if visual_prompt is not None: A__ = self.image_processor(SCREAMING_SNAKE_CASE__ , return_tensors=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) if images is not None: A__ = self.image_processor(SCREAMING_SNAKE_CASE__ , return_tensors=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) if visual_prompt is not None and images is not None: A__ = { "pixel_values": image_features.pixel_values, "conditional_pixel_values": prompt_features.pixel_values, } return encoding elif text is not None and images is not None: A__ = image_features.pixel_values return encoding elif text is not None: return encoding elif visual_prompt is not None: A__ = { "conditional_pixel_values": prompt_features.pixel_values, } return encoding else: return BatchEncoding(data=dict(**SCREAMING_SNAKE_CASE__ ) , tensor_type=SCREAMING_SNAKE_CASE__ ) def UpperCamelCase ( self , *lowercase , **lowercase ) -> List[str]: '''simple docstring''' return self.tokenizer.batch_decode(*SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) def UpperCamelCase ( self , *lowercase , **lowercase ) -> Any: '''simple docstring''' return self.tokenizer.decode(*SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) @property def UpperCamelCase ( self ) -> List[Any]: '''simple docstring''' warnings.warn( "`feature_extractor_class` is deprecated and will be removed in v5. Use `image_processor_class` instead." , SCREAMING_SNAKE_CASE__ , ) return self.image_processor_class @property def UpperCamelCase ( self ) -> Any: '''simple docstring''' warnings.warn( "`feature_extractor` is deprecated and will be removed in v5. Use `image_processor` instead." , SCREAMING_SNAKE_CASE__ , ) return self.image_processor
68
from abc import ABC, abstractmethod from argparse import ArgumentParser class lowerCAmelCase__ ( __lowercase ): @staticmethod @abstractmethod def __A ( SCREAMING_SNAKE_CASE__ : ArgumentParser ) -> str: raise NotImplementedError() @abstractmethod def __A ( self : Optional[int] ) -> Union[str, Any]: raise NotImplementedError()
270
0
"""simple docstring""" import json import os import sys import tempfile import unittest from pathlib import Path from shutil import copyfile from huggingface_hub import HfFolder, Repository, create_repo, delete_repo from requests.exceptions import HTTPError import transformers from transformers import ( CONFIG_MAPPING, FEATURE_EXTRACTOR_MAPPING, PROCESSOR_MAPPING, TOKENIZER_MAPPING, AutoConfig, AutoFeatureExtractor, AutoProcessor, AutoTokenizer, BertTokenizer, ProcessorMixin, WavaVecaConfig, WavaVecaFeatureExtractor, WavaVecaProcessor, ) from transformers.testing_utils import TOKEN, USER, get_tests_dir, is_staging_test from transformers.tokenization_utils import TOKENIZER_CONFIG_FILE from transformers.utils import FEATURE_EXTRACTOR_NAME, is_tokenizers_available sys.path.append(str(Path(__file__).parent.parent.parent.parent / '''utils''')) from test_module.custom_configuration import CustomConfig # noqa E402 from test_module.custom_feature_extraction import CustomFeatureExtractor # noqa E402 from test_module.custom_processing import CustomProcessor # noqa E402 from test_module.custom_tokenization import CustomTokenizer # noqa E402 __snake_case = get_tests_dir('''fixtures/dummy_feature_extractor_config.json''') __snake_case = get_tests_dir('''fixtures/vocab.json''') __snake_case = get_tests_dir('''fixtures''') class __lowerCamelCase ( unittest.TestCase ): '''simple docstring''' A_ : str = ["""[UNK]""", """[CLS]""", """[SEP]""", """[PAD]""", """[MASK]""", """bla""", """blou"""] def _UpperCAmelCase ( self ) -> List[Any]: _a = 0 def _UpperCAmelCase ( self ) -> List[Any]: _a = AutoProcessor.from_pretrained('''facebook/wav2vec2-base-960h''' ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Union[str, Any]: with tempfile.TemporaryDirectory() as tmpdirname: _a = WavaVecaConfig() _a = AutoProcessor.from_pretrained('''facebook/wav2vec2-base-960h''' ) # save in new folder model_config.save_pretrained(SCREAMING_SNAKE_CASE__ ) processor.save_pretrained(SCREAMING_SNAKE_CASE__ ) _a = AutoProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Any: with tempfile.TemporaryDirectory() as tmpdirname: # copy relevant files copyfile(SCREAMING_SNAKE_CASE__ , os.path.join(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) ) copyfile(SCREAMING_SNAKE_CASE__ , os.path.join(SCREAMING_SNAKE_CASE__ , '''vocab.json''' ) ) _a = AutoProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Optional[Any]: with tempfile.TemporaryDirectory() as tmpdirname: _a = WavaVecaFeatureExtractor() _a = AutoTokenizer.from_pretrained('''facebook/wav2vec2-base-960h''' ) _a = WavaVecaProcessor(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # save in new folder processor.save_pretrained(SCREAMING_SNAKE_CASE__ ) # drop `processor_class` in tokenizer with open(os.path.join(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) , '''r''' ) as f: _a = json.load(SCREAMING_SNAKE_CASE__ ) config_dict.pop('''processor_class''' ) with open(os.path.join(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) , '''w''' ) as f: f.write(json.dumps(SCREAMING_SNAKE_CASE__ ) ) _a = AutoProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> str: with tempfile.TemporaryDirectory() as tmpdirname: _a = WavaVecaFeatureExtractor() _a = AutoTokenizer.from_pretrained('''facebook/wav2vec2-base-960h''' ) _a = WavaVecaProcessor(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # save in new folder processor.save_pretrained(SCREAMING_SNAKE_CASE__ ) # drop `processor_class` in feature extractor with open(os.path.join(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) , '''r''' ) as f: _a = json.load(SCREAMING_SNAKE_CASE__ ) config_dict.pop('''processor_class''' ) with open(os.path.join(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) , '''w''' ) as f: f.write(json.dumps(SCREAMING_SNAKE_CASE__ ) ) _a = AutoProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Dict: with tempfile.TemporaryDirectory() as tmpdirname: _a = WavaVecaConfig(processor_class='''Wav2Vec2Processor''' ) model_config.save_pretrained(SCREAMING_SNAKE_CASE__ ) # copy relevant files copyfile(SCREAMING_SNAKE_CASE__ , os.path.join(SCREAMING_SNAKE_CASE__ , '''vocab.json''' ) ) # create emtpy sample processor with open(os.path.join(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) , '''w''' ) as f: f.write('''{}''' ) _a = AutoProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Any: # If remote code is not set, we will time out when asking whether to load the model. with self.assertRaises(SCREAMING_SNAKE_CASE__ ): _a = AutoProcessor.from_pretrained('''hf-internal-testing/test_dynamic_processor''' ) # If remote code is disabled, we can't load this config. with self.assertRaises(SCREAMING_SNAKE_CASE__ ): _a = AutoProcessor.from_pretrained( '''hf-internal-testing/test_dynamic_processor''' , trust_remote_code=SCREAMING_SNAKE_CASE__ ) _a = AutoProcessor.from_pretrained('''hf-internal-testing/test_dynamic_processor''' , trust_remote_code=SCREAMING_SNAKE_CASE__ ) self.assertTrue(processor.special_attribute_present ) self.assertEqual(processor.__class__.__name__ , '''NewProcessor''' ) _a = processor.feature_extractor self.assertTrue(feature_extractor.special_attribute_present ) self.assertEqual(feature_extractor.__class__.__name__ , '''NewFeatureExtractor''' ) _a = processor.tokenizer self.assertTrue(tokenizer.special_attribute_present ) if is_tokenizers_available(): self.assertEqual(tokenizer.__class__.__name__ , '''NewTokenizerFast''' ) # Test we can also load the slow version _a = AutoProcessor.from_pretrained( '''hf-internal-testing/test_dynamic_processor''' , trust_remote_code=SCREAMING_SNAKE_CASE__ , use_fast=SCREAMING_SNAKE_CASE__ ) _a = new_processor.tokenizer self.assertTrue(new_tokenizer.special_attribute_present ) self.assertEqual(new_tokenizer.__class__.__name__ , '''NewTokenizer''' ) else: self.assertEqual(tokenizer.__class__.__name__ , '''NewTokenizer''' ) def _UpperCAmelCase ( self ) -> Tuple: try: AutoConfig.register('''custom''' , SCREAMING_SNAKE_CASE__ ) AutoFeatureExtractor.register(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) AutoTokenizer.register(SCREAMING_SNAKE_CASE__ , slow_tokenizer_class=SCREAMING_SNAKE_CASE__ ) AutoProcessor.register(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # Trying to register something existing in the Transformers library will raise an error with self.assertRaises(SCREAMING_SNAKE_CASE__ ): AutoProcessor.register(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # Now that the config is registered, it can be used as any other config with the auto-API _a = CustomFeatureExtractor.from_pretrained(SCREAMING_SNAKE_CASE__ ) with tempfile.TemporaryDirectory() as tmp_dir: _a = os.path.join(SCREAMING_SNAKE_CASE__ , '''vocab.txt''' ) with open(SCREAMING_SNAKE_CASE__ , '''w''' , encoding='''utf-8''' ) as vocab_writer: vocab_writer.write(''''''.join([x + '''\n''' for x in self.vocab_tokens] ) ) _a = CustomTokenizer(SCREAMING_SNAKE_CASE__ ) _a = CustomProcessor(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) with tempfile.TemporaryDirectory() as tmp_dir: processor.save_pretrained(SCREAMING_SNAKE_CASE__ ) _a = AutoProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) finally: if "custom" in CONFIG_MAPPING._extra_content: del CONFIG_MAPPING._extra_content["custom"] if CustomConfig in FEATURE_EXTRACTOR_MAPPING._extra_content: del FEATURE_EXTRACTOR_MAPPING._extra_content[CustomConfig] if CustomConfig in TOKENIZER_MAPPING._extra_content: del TOKENIZER_MAPPING._extra_content[CustomConfig] if CustomConfig in PROCESSOR_MAPPING._extra_content: del PROCESSOR_MAPPING._extra_content[CustomConfig] def _UpperCAmelCase ( self ) -> Dict: class __lowerCamelCase ( __lowercase ): '''simple docstring''' A_ : List[Any] = False class __lowerCamelCase ( __lowercase ): '''simple docstring''' A_ : List[str] = False class __lowerCamelCase ( __lowercase ): '''simple docstring''' A_ : Union[str, Any] = """AutoFeatureExtractor""" A_ : int = """AutoTokenizer""" A_ : int = False try: AutoConfig.register('''custom''' , SCREAMING_SNAKE_CASE__ ) AutoFeatureExtractor.register(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) AutoTokenizer.register(SCREAMING_SNAKE_CASE__ , slow_tokenizer_class=SCREAMING_SNAKE_CASE__ ) AutoProcessor.register(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # If remote code is not set, the default is to use local classes. _a = AutoProcessor.from_pretrained('''hf-internal-testing/test_dynamic_processor''' ) self.assertEqual(processor.__class__.__name__ , '''NewProcessor''' ) self.assertFalse(processor.special_attribute_present ) self.assertFalse(processor.feature_extractor.special_attribute_present ) self.assertFalse(processor.tokenizer.special_attribute_present ) # If remote code is disabled, we load the local ones. _a = AutoProcessor.from_pretrained( '''hf-internal-testing/test_dynamic_processor''' , trust_remote_code=SCREAMING_SNAKE_CASE__ ) self.assertEqual(processor.__class__.__name__ , '''NewProcessor''' ) self.assertFalse(processor.special_attribute_present ) self.assertFalse(processor.feature_extractor.special_attribute_present ) self.assertFalse(processor.tokenizer.special_attribute_present ) # If remote is enabled, we load from the Hub. _a = AutoProcessor.from_pretrained( '''hf-internal-testing/test_dynamic_processor''' , trust_remote_code=SCREAMING_SNAKE_CASE__ ) self.assertEqual(processor.__class__.__name__ , '''NewProcessor''' ) self.assertTrue(processor.special_attribute_present ) self.assertTrue(processor.feature_extractor.special_attribute_present ) self.assertTrue(processor.tokenizer.special_attribute_present ) finally: if "custom" in CONFIG_MAPPING._extra_content: del CONFIG_MAPPING._extra_content["custom"] if CustomConfig in FEATURE_EXTRACTOR_MAPPING._extra_content: del FEATURE_EXTRACTOR_MAPPING._extra_content[CustomConfig] if CustomConfig in TOKENIZER_MAPPING._extra_content: del TOKENIZER_MAPPING._extra_content[CustomConfig] if CustomConfig in PROCESSOR_MAPPING._extra_content: del PROCESSOR_MAPPING._extra_content[CustomConfig] def _UpperCAmelCase ( self ) -> Any: _a = AutoProcessor.from_pretrained('''hf-internal-testing/tiny-random-bert''' ) self.assertEqual(processor.__class__.__name__ , '''BertTokenizerFast''' ) def _UpperCAmelCase ( self ) -> Dict: _a = AutoProcessor.from_pretrained('''hf-internal-testing/tiny-random-convnext''' ) self.assertEqual(processor.__class__.__name__ , '''ConvNextImageProcessor''' ) @is_staging_test class __lowerCamelCase ( unittest.TestCase ): '''simple docstring''' A_ : Dict = ["""[UNK]""", """[CLS]""", """[SEP]""", """[PAD]""", """[MASK]""", """bla""", """blou"""] @classmethod def _UpperCAmelCase ( cls ) -> Dict: _a = TOKEN HfFolder.save_token(SCREAMING_SNAKE_CASE__ ) @classmethod def _UpperCAmelCase ( cls ) -> Union[str, Any]: try: delete_repo(token=cls._token , repo_id='''test-processor''' ) except HTTPError: pass try: delete_repo(token=cls._token , repo_id='''valid_org/test-processor-org''' ) except HTTPError: pass try: delete_repo(token=cls._token , repo_id='''test-dynamic-processor''' ) except HTTPError: pass def _UpperCAmelCase ( self ) -> Union[str, Any]: _a = WavaVecaProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) with tempfile.TemporaryDirectory() as tmp_dir: processor.save_pretrained( os.path.join(SCREAMING_SNAKE_CASE__ , '''test-processor''' ) , push_to_hub=SCREAMING_SNAKE_CASE__ , use_auth_token=self._token ) _a = WavaVecaProcessor.from_pretrained(F'{USER}/test-processor' ) for k, v in processor.feature_extractor.__dict__.items(): self.assertEqual(SCREAMING_SNAKE_CASE__ , getattr(new_processor.feature_extractor , SCREAMING_SNAKE_CASE__ ) ) self.assertDictEqual(new_processor.tokenizer.get_vocab() , processor.tokenizer.get_vocab() ) def _UpperCAmelCase ( self ) -> Dict: _a = WavaVecaProcessor.from_pretrained(SCREAMING_SNAKE_CASE__ ) with tempfile.TemporaryDirectory() as tmp_dir: processor.save_pretrained( os.path.join(SCREAMING_SNAKE_CASE__ , '''test-processor-org''' ) , push_to_hub=SCREAMING_SNAKE_CASE__ , use_auth_token=self._token , organization='''valid_org''' , ) _a = WavaVecaProcessor.from_pretrained('''valid_org/test-processor-org''' ) for k, v in processor.feature_extractor.__dict__.items(): self.assertEqual(SCREAMING_SNAKE_CASE__ , getattr(new_processor.feature_extractor , SCREAMING_SNAKE_CASE__ ) ) self.assertDictEqual(new_processor.tokenizer.get_vocab() , processor.tokenizer.get_vocab() ) def _UpperCAmelCase ( self ) -> str: CustomFeatureExtractor.register_for_auto_class() CustomTokenizer.register_for_auto_class() CustomProcessor.register_for_auto_class() _a = CustomFeatureExtractor.from_pretrained(SCREAMING_SNAKE_CASE__ ) with tempfile.TemporaryDirectory() as tmp_dir: _a = os.path.join(SCREAMING_SNAKE_CASE__ , '''vocab.txt''' ) with open(SCREAMING_SNAKE_CASE__ , '''w''' , encoding='''utf-8''' ) as vocab_writer: vocab_writer.write(''''''.join([x + '''\n''' for x in self.vocab_tokens] ) ) _a = CustomTokenizer(SCREAMING_SNAKE_CASE__ ) _a = CustomProcessor(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) with tempfile.TemporaryDirectory() as tmp_dir: create_repo(F'{USER}/test-dynamic-processor' , token=self._token ) _a = Repository(SCREAMING_SNAKE_CASE__ , clone_from=F'{USER}/test-dynamic-processor' , token=self._token ) processor.save_pretrained(SCREAMING_SNAKE_CASE__ ) # This has added the proper auto_map field to the feature extractor config self.assertDictEqual( processor.feature_extractor.auto_map , { '''AutoFeatureExtractor''': '''custom_feature_extraction.CustomFeatureExtractor''', '''AutoProcessor''': '''custom_processing.CustomProcessor''', } , ) # This has added the proper auto_map field to the tokenizer config with open(os.path.join(SCREAMING_SNAKE_CASE__ , '''tokenizer_config.json''' ) ) as f: _a = json.load(SCREAMING_SNAKE_CASE__ ) self.assertDictEqual( tokenizer_config['''auto_map'''] , { '''AutoTokenizer''': ['''custom_tokenization.CustomTokenizer''', None], '''AutoProcessor''': '''custom_processing.CustomProcessor''', } , ) # The code has been copied from fixtures self.assertTrue(os.path.isfile(os.path.join(SCREAMING_SNAKE_CASE__ , '''custom_feature_extraction.py''' ) ) ) self.assertTrue(os.path.isfile(os.path.join(SCREAMING_SNAKE_CASE__ , '''custom_tokenization.py''' ) ) ) self.assertTrue(os.path.isfile(os.path.join(SCREAMING_SNAKE_CASE__ , '''custom_processing.py''' ) ) ) repo.push_to_hub() _a = AutoProcessor.from_pretrained(F'{USER}/test-dynamic-processor' , trust_remote_code=SCREAMING_SNAKE_CASE__ ) # Can't make an isinstance check because the new_processor is from the CustomProcessor class of a dynamic module self.assertEqual(new_processor.__class__.__name__ , '''CustomProcessor''' )
320
import json import os import subprocess import unittest from ast import literal_eval import pytest from parameterized import parameterized, parameterized_class from . import is_sagemaker_available if is_sagemaker_available(): from sagemaker import Session, TrainingJobAnalytics from sagemaker.huggingface import HuggingFace @pytest.mark.skipif( literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , ) @pytest.mark.usefixtures("""sm_env""" ) @parameterized_class( [ { """framework""": """pytorch""", """script""": """run_glue_model_parallelism.py""", """model_name_or_path""": """roberta-large""", """instance_type""": """ml.p3dn.24xlarge""", """results""": {"""train_runtime""": 1_600, """eval_accuracy""": 0.3, """eval_loss""": 1.2}, }, { """framework""": """pytorch""", """script""": """run_glue.py""", """model_name_or_path""": """roberta-large""", """instance_type""": """ml.p3dn.24xlarge""", """results""": {"""train_runtime""": 1_600, """eval_accuracy""": 0.3, """eval_loss""": 1.2}, }, ] ) class lowerCAmelCase__ ( unittest.TestCase ): def __A ( self : Optional[int] ) -> List[Any]: if self.framework == "pytorch": subprocess.run( f'''cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py'''.split() , encoding='''utf-8''' , check=SCREAMING_SNAKE_CASE__ , ) assert hasattr(self , '''env''' ) def __A ( self : List[str] , SCREAMING_SNAKE_CASE__ : List[Any] ) -> Optional[Any]: # configuration for running training on smdistributed Model Parallel __lowerCamelCase = { '''enabled''': True, '''processes_per_host''': 8, } __lowerCamelCase = { '''enabled''': True, '''parameters''': { '''microbatches''': 4, '''placement_strategy''': '''spread''', '''pipeline''': '''interleaved''', '''optimize''': '''speed''', '''partitions''': 4, '''ddp''': True, }, } __lowerCamelCase = {'''smdistributed''': {'''modelparallel''': smp_options}, '''mpi''': mpi_options} __lowerCamelCase = '''trainer''' if self.script == '''run_glue.py''' else '''smtrainer''' # creates estimator return HuggingFace( entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f'''{self.env.base_job_name}-{instance_count}-smp-{name_extension}''' , instance_count=SCREAMING_SNAKE_CASE__ , instance_type=self.instance_type , debugger_hook_config=SCREAMING_SNAKE_CASE__ , hyperparameters={ **self.env.hyperparameters, '''model_name_or_path''': self.model_name_or_path, '''max_steps''': 5_00, } , metric_definitions=self.env.metric_definitions , distribution=SCREAMING_SNAKE_CASE__ , py_version='''py36''' , ) def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : List[Any] ) -> List[str]: TrainingJobAnalytics(SCREAMING_SNAKE_CASE__ ).export_csv(f'''{self.env.test_path}/{job_name}_metrics.csv''' ) @parameterized.expand([(1,)] ) def __A ( self : Optional[Any] , SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Optional[int]: # create estimator __lowerCamelCase = self.create_estimator(SCREAMING_SNAKE_CASE__ ) # run training estimator.fit() # result dataframe __lowerCamelCase = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe() # extract kpis __lowerCamelCase = list(result_metrics_df[result_metrics_df.metric_name == '''eval_accuracy''']['''value'''] ) __lowerCamelCase = list(result_metrics_df[result_metrics_df.metric_name == '''eval_loss''']['''value'''] ) # get train time from SageMaker job, this includes starting, preprocessing, stopping __lowerCamelCase = ( Session().describe_training_job(estimator.latest_training_job.name ).get('''TrainingTimeInSeconds''' , 99_99_99 ) ) # assert kpis assert train_runtime <= self.results["train_runtime"] assert all(t >= self.results['''eval_accuracy'''] for t in eval_accuracy ) assert all(t <= self.results['''eval_loss'''] for t in eval_loss ) # dump tests result into json file to share in PR with open(f'''{estimator.latest_training_job.name}.json''' , '''w''' ) as outfile: json.dump({'''train_time''': train_runtime, '''eval_accuracy''': eval_accuracy, '''eval_loss''': eval_loss} , SCREAMING_SNAKE_CASE__ )
270
0
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 : Optional[int] = os.path.abspath(os.path.dirname(__file__)) with ExtendSysPath(F'{bindir}/../../examples/pytorch/translation'): from run_translation import main # noqa set_seed(4_2) A : Any = "sshleifer/student_marian_en_ro_6_1" A : Tuple = "sshleifer/tiny-mbart" @require_torch class A ( __lowercase ): '''simple docstring''' def lowerCamelCase__ (self : List[Any] , _UpperCAmelCase : Tuple=False , _UpperCAmelCase : Optional[Any]=None , _UpperCAmelCase : Optional[Any]=True , _UpperCAmelCase : str=True , _UpperCAmelCase : Optional[Any]=True , _UpperCAmelCase : Optional[int]=True , ) -> Optional[int]: """simple docstring""" lowercase__ = self.run_trainer( eval_steps=1 , max_len=12 , model_name=SCREAMING_SNAKE_CASE__ , num_train_epochs=1 , distributed=SCREAMING_SNAKE_CASE__ , extra_args_str=SCREAMING_SNAKE_CASE__ , predict_with_generate=SCREAMING_SNAKE_CASE__ , do_train=SCREAMING_SNAKE_CASE__ , do_eval=SCREAMING_SNAKE_CASE__ , do_predict=SCREAMING_SNAKE_CASE__ , ) lowercase__ = TrainerState.load_from_json(os.path.join(SCREAMING_SNAKE_CASE__ , """trainer_state.json""" ) ).log_history if not do_eval: return lowercase__ = [log for log in logs if """eval_loss""" in log.keys()] lowercase__ = eval_metrics[0] if predict_with_generate: assert "eval_bleu" in first_step_stats lowercase__ = eval_metrics[-1] assert isinstance(last_step_stats["""eval_bleu"""] , SCREAMING_SNAKE_CASE__ ) assert not math.isnan(float(last_step_stats["""eval_loss"""] ) ), "eval_loss must not be `nan`" @require_torch_non_multi_gpu def lowerCamelCase__ (self : Optional[int] ) -> int: """simple docstring""" self.run_seqaseq_quick() @require_torch_multi_gpu def lowerCamelCase__ (self : int ) -> List[str]: """simple docstring""" self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ ) @require_torch_multi_gpu def lowerCamelCase__ (self : Optional[Any] ) -> Tuple: """simple docstring""" self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ ) @unittest.skip("""Requires an update of the env running those tests""" ) @require_torch_multi_gpu @require_fairscale def lowerCamelCase__ (self : Dict ) -> Tuple: """simple docstring""" self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , extra_args_str="""--sharded_ddp simple""" ) @unittest.skip("""Requires an update of the env running those tests""" ) @require_torch_multi_gpu @require_fairscale def lowerCamelCase__ (self : Optional[int] ) -> List[str]: """simple docstring""" self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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 lowerCamelCase__ (self : Tuple ) -> Any: """simple docstring""" self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , extra_args_str="""--sharded_ddp zero_dp_2""" , predict_with_generate=SCREAMING_SNAKE_CASE__ ) @unittest.skip("""Requires an update of the env running those tests""" ) @require_torch_multi_gpu @require_fairscale def lowerCamelCase__ (self : Dict ) -> Tuple: """simple docstring""" self.run_seqaseq_quick( distributed=SCREAMING_SNAKE_CASE__ , extra_args_str="""--sharded_ddp zero_dp_2 --fp16""" , predict_with_generate=SCREAMING_SNAKE_CASE__ ) @require_apex @require_torch_gpu def lowerCamelCase__ (self : Union[str, Any] ) -> List[str]: """simple docstring""" self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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=SCREAMING_SNAKE_CASE__ , extra_args_str="""--fp16 --fp16_backend=apex""" ) @parameterized.expand(["""base""", """low""", """high""", """mixed"""] ) @require_torch_multi_gpu def lowerCamelCase__ (self : Any , _UpperCAmelCase : List[str] ) -> Optional[Any]: """simple docstring""" lowercase__ = { # 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}, } lowercase__ = experiments[experiment_id] lowercase__ = {"""distributed""": True, """predict_with_generate""": False, """do_eval""": False, """do_predict""": False} lowercase__ = """Running training""" with CaptureStderr() as cl: self.run_seqaseq_quick(**SCREAMING_SNAKE_CASE__ , extra_args_str=data["""extra_args_str"""] ) lowercase__ = len(re.findall(SCREAMING_SNAKE_CASE__ , cl.err ) ) self.assertEqual(SCREAMING_SNAKE_CASE__ , data["""n_matches"""] ) @slow def lowerCamelCase__ (self : Any ) -> Optional[Any]: """simple docstring""" lowercase__ = self.run_trainer( eval_steps=2 , max_len=128 , model_name=SCREAMING_SNAKE_CASE__ , learning_rate=3E-4 , num_train_epochs=10 , distributed=SCREAMING_SNAKE_CASE__ , ) # Check metrics lowercase__ = TrainerState.load_from_json(os.path.join(SCREAMING_SNAKE_CASE__ , """trainer_state.json""" ) ).log_history lowercase__ = [log for log in logs if """eval_loss""" in log.keys()] lowercase__ = eval_metrics[0] lowercase__ = eval_metrics[-1] assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing" assert isinstance(last_step_stats["""eval_bleu"""] , SCREAMING_SNAKE_CASE__ ) # test if do_predict saves generations and metrics lowercase__ = os.listdir(SCREAMING_SNAKE_CASE__ ) lowercase__ = {os.path.basename(SCREAMING_SNAKE_CASE__ ) for p in contents} assert "generated_predictions.txt" in contents assert "predict_results.json" in contents @slow @require_bitsandbytes def lowerCamelCase__ (self : Optional[int] ) -> str: """simple docstring""" from transformers.training_args import OptimizerNames def train_and_return_metrics(_UpperCAmelCase : str ) -> Tuple[int, float]: lowercase__ = """--skip_memory_metrics 0""" lowercase__ = self.run_trainer( max_len=128 , model_name=SCREAMING_SNAKE_CASE__ , learning_rate=3E-4 , num_train_epochs=1 , optim=SCREAMING_SNAKE_CASE__ , distributed=SCREAMING_SNAKE_CASE__ , extra_args_str=SCREAMING_SNAKE_CASE__ , do_eval=SCREAMING_SNAKE_CASE__ , do_predict=SCREAMING_SNAKE_CASE__ , n_gpus_to_use=1 , ) # Check metrics lowercase__ = TrainerState.load_from_json(Path(SCREAMING_SNAKE_CASE__ , """trainer_state.json""" ) ).log_history lowercase__ = int(logs[0]["""train_mem_gpu_peaked_delta"""] / 2**20 ) lowercase__ = int(logs[0]["""train_mem_gpu_alloc_delta"""] / 2**20 ) lowercase__ = logs[0]["""train_loss"""] return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss lowercase__ , lowercase__ , lowercase__ = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value ) lowercase__ , lowercase__ , lowercase__ = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value ) lowercase__ = gpu_alloc_mem_orig - gpu_alloc_mem_bnb lowercase__ = gpu_peak_mem_orig + gpu_alloc_mem_orig lowercase__ = gpu_peak_mem_bnb + gpu_alloc_mem_bnb lowercase__ = 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 lowercase__ = 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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , """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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , """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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , f'''loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}''' ) def lowerCamelCase__ (self : Dict , _UpperCAmelCase : int , _UpperCAmelCase : str , _UpperCAmelCase : int , _UpperCAmelCase : float = 3E-3 , _UpperCAmelCase : str = "adafactor" , _UpperCAmelCase : bool = False , _UpperCAmelCase : str = None , _UpperCAmelCase : int = 0 , _UpperCAmelCase : bool = True , _UpperCAmelCase : bool = True , _UpperCAmelCase : bool = True , _UpperCAmelCase : bool = True , _UpperCAmelCase : int = None , ) -> List[Any]: """simple docstring""" lowercase__ = self.test_file_dir / """../fixtures/tests_samples/wmt_en_ro""" lowercase__ = self.get_auto_remove_tmp_dir() lowercase__ = f''' --model_name_or_path {model_name} --train_file {data_dir}/train.json --validation_file {data_dir}/val.json --test_file {data_dir}/test.json --output_dir {output_dir} --overwrite_output_dir --max_train_samples 8 --max_source_length {max_len} --max_target_length {max_len} --do_train --num_train_epochs {str(SCREAMING_SNAKE_CASE__ )} --per_device_train_batch_size 4 --learning_rate {learning_rate} --warmup_steps 8 --logging_steps 0 --logging_strategy no --save_steps {str(SCREAMING_SNAKE_CASE__ )} --group_by_length --label_smoothing_factor 0.1 --target_lang ro_RO --source_lang en_XX '''.split() lowercase__ = f''' --do_eval --per_device_eval_batch_size 4 --max_eval_samples 8 --val_max_target_length {max_len} --evaluation_strategy steps --eval_steps {str(SCREAMING_SNAKE_CASE__ )} '''.split() lowercase__ = """ --do_predict """.split() lowercase__ = [] 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: lowercase__ = get_gpu_count() lowercase__ = get_torch_dist_unique_port() lowercase__ = f''' -m torch.distributed.run --nproc_per_node={n_gpus_to_use} --master_port={master_port} {self.examples_dir_str}/pytorch/translation/run_translation.py '''.split() lowercase__ = [sys.executable] + distributed_args + args # keep for quick debug # print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die execute_subprocess_async(SCREAMING_SNAKE_CASE__ , env=self.get_env() ) else: lowercase__ = ["""run_translation.py"""] + args with patch.object(SCREAMING_SNAKE_CASE__ , """argv""" , SCREAMING_SNAKE_CASE__ ): main() return output_dir
305
from __future__ import annotations from fractions import Fraction def __magic_name__ ( __lowerCAmelCase : int , __lowerCAmelCase : int ) -> bool: return ( num != den and num % 10 == den // 10 and (num // 10) / (den % 10) == num / den ) def __magic_name__ ( __lowerCAmelCase : int ) -> list[str]: __lowerCamelCase = [] __lowerCamelCase = 11 __lowerCamelCase = int('''1''' + '''0''' * digit_len ) for num in range(__lowerCAmelCase , __lowerCAmelCase ): while den <= 99: if (num != den) and (num % 10 == den // 10) and (den % 10 != 0): if is_digit_cancelling(__lowerCAmelCase , __lowerCAmelCase ): solutions.append(f'''{num}/{den}''' ) den += 1 num += 1 __lowerCamelCase = 10 return solutions def __magic_name__ ( __lowerCAmelCase : int = 2 ) -> int: __lowerCamelCase = 1.0 for fraction in fraction_list(__lowerCAmelCase ): __lowerCamelCase = Fraction(__lowerCAmelCase ) result *= frac.denominator / frac.numerator return int(__lowerCAmelCase ) if __name__ == "__main__": print(solution())
270
0
'''simple docstring''' import functools from typing import Any def _UpperCamelCase ( SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : list[str] ): '''simple docstring''' if not isinstance(__lowerCAmelCase , __lowerCAmelCase ) or len(__lowerCAmelCase ) == 0: raise ValueError("""the string should be not empty string""" ) if not isinstance(__lowerCAmelCase , __lowerCAmelCase ) or not all( isinstance(__lowerCAmelCase , __lowerCAmelCase ) and len(__lowerCAmelCase ) > 0 for item in words ): raise ValueError("""the words should be a list of non-empty strings""" ) # Build trie UpperCAmelCase__ = {} UpperCAmelCase__ = """WORD_KEEPER""" for word in words: UpperCAmelCase__ = trie for c in word: if c not in trie_node: UpperCAmelCase__ = {} UpperCAmelCase__ = trie_node[c] UpperCAmelCase__ = True UpperCAmelCase__ = len(__lowerCAmelCase ) # Dynamic programming method @functools.cache def is_breakable(SCREAMING_SNAKE_CASE__ : int ) -> bool: if index == len_string: return True UpperCAmelCase__ = trie for i in range(__lowerCAmelCase , __lowerCAmelCase ): UpperCAmelCase__ = trie_node.get(string[i] , __lowerCAmelCase ) if trie_node is None: return False if trie_node.get(__lowerCAmelCase , __lowerCAmelCase ) and is_breakable(i + 1 ): return True return False return is_breakable(0 ) if __name__ == "__main__": import doctest doctest.testmod()
346
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_tf_available, is_torch_available, is_vision_available, ) SCREAMING_SNAKE_CASE__ : List[str] = { "configuration_blip": [ "BLIP_PRETRAINED_CONFIG_ARCHIVE_MAP", "BlipConfig", "BlipTextConfig", "BlipVisionConfig", ], "processing_blip": ["BlipProcessor"], } try: if not is_vision_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : str = ["BlipImageProcessor"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : int = [ "BLIP_PRETRAINED_MODEL_ARCHIVE_LIST", "BlipModel", "BlipPreTrainedModel", "BlipForConditionalGeneration", "BlipForQuestionAnswering", "BlipVisionModel", "BlipTextModel", "BlipForImageTextRetrieval", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : Any = [ "TF_BLIP_PRETRAINED_MODEL_ARCHIVE_LIST", "TFBlipModel", "TFBlipPreTrainedModel", "TFBlipForConditionalGeneration", "TFBlipForQuestionAnswering", "TFBlipVisionModel", "TFBlipTextModel", "TFBlipForImageTextRetrieval", ] if TYPE_CHECKING: from .configuration_blip import BLIP_PRETRAINED_CONFIG_ARCHIVE_MAP, BlipConfig, BlipTextConfig, BlipVisionConfig from .processing_blip import BlipProcessor try: if not is_vision_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .image_processing_blip import BlipImageProcessor try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_blip import ( BLIP_PRETRAINED_MODEL_ARCHIVE_LIST, BlipForConditionalGeneration, BlipForImageTextRetrieval, BlipForQuestionAnswering, BlipModel, BlipPreTrainedModel, BlipTextModel, BlipVisionModel, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_blip import ( TF_BLIP_PRETRAINED_MODEL_ARCHIVE_LIST, TFBlipForConditionalGeneration, TFBlipForImageTextRetrieval, TFBlipForQuestionAnswering, TFBlipModel, TFBlipPreTrainedModel, TFBlipTextModel, TFBlipVisionModel, ) else: import sys SCREAMING_SNAKE_CASE__ : Tuple = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
270
0
import argparse import os import pickle import sys import torch from transformers import TransfoXLConfig, TransfoXLLMHeadModel, load_tf_weights_in_transfo_xl from transformers.models.transfo_xl import tokenization_transfo_xl as data_utils from transformers.models.transfo_xl.tokenization_transfo_xl import CORPUS_NAME, VOCAB_FILES_NAMES from transformers.utils import CONFIG_NAME, WEIGHTS_NAME, logging logging.set_verbosity_info() # We do this to be able to load python 2 datasets pickles # See e.g. https://stackoverflow.com/questions/2121874/python-pickling-after-changing-a-modules-directory/2121918#2121918 A : Union[str, Any] = data_utils.TransfoXLTokenizer A : str = data_utils.TransfoXLCorpus A : List[Any] = data_utils A : Any = data_utils def lowercase_ ( _A : int , _A : str , _A : int , _A : Optional[int] ): """simple docstring""" if transfo_xl_dataset_file: # Convert a pre-processed corpus (see original TensorFlow repo) with open(__lowerCAmelCase , "rb" ) as fp: lowerCamelCase__ : List[Any] = pickle.load(__lowerCAmelCase , encoding="latin1" ) # Save vocabulary and dataset cache as Dictionaries (should be better than pickles for the long-term) lowerCamelCase__ : List[str] = pytorch_dump_folder_path + "/" + VOCAB_FILES_NAMES["pretrained_vocab_file"] print(F"Save vocabulary to {pytorch_vocab_dump_path}" ) lowerCamelCase__ : str = corpus.vocab.__dict__ torch.save(__lowerCAmelCase , __lowerCAmelCase ) lowerCamelCase__ : Optional[Any] = corpus.__dict__ corpus_dict_no_vocab.pop("vocab" , __lowerCAmelCase ) lowerCamelCase__ : str = pytorch_dump_folder_path + "/" + CORPUS_NAME print(F"Save dataset to {pytorch_dataset_dump_path}" ) torch.save(__lowerCAmelCase , __lowerCAmelCase ) if tf_checkpoint_path: # Convert a pre-trained TensorFlow model lowerCamelCase__ : List[Any] = os.path.abspath(__lowerCAmelCase ) lowerCamelCase__ : List[Any] = os.path.abspath(__lowerCAmelCase ) print(F"Converting Transformer XL checkpoint from {tf_path} with config at {config_path}." ) # Initialise PyTorch model if transfo_xl_config_file == "": lowerCamelCase__ : Optional[int] = TransfoXLConfig() else: lowerCamelCase__ : str = TransfoXLConfig.from_json_file(__lowerCAmelCase ) print(F"Building PyTorch model from configuration: {config}" ) lowerCamelCase__ : int = TransfoXLLMHeadModel(__lowerCAmelCase ) lowerCamelCase__ : int = load_tf_weights_in_transfo_xl(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) # Save pytorch-model lowerCamelCase__ : Any = os.path.join(__lowerCAmelCase , __lowerCAmelCase ) lowerCamelCase__ : Optional[Any] = os.path.join(__lowerCAmelCase , __lowerCAmelCase ) print(F"Save PyTorch model to {os.path.abspath(__lowerCAmelCase )}" ) torch.save(model.state_dict() , __lowerCAmelCase ) print(F"Save configuration file to {os.path.abspath(__lowerCAmelCase )}" ) with open(__lowerCAmelCase , "w" , encoding="utf-8" ) as f: f.write(config.to_json_string() ) if __name__ == "__main__": A : Any = argparse.ArgumentParser() parser.add_argument( "--pytorch_dump_folder_path", default=None, type=str, required=True, help="Path to the folder to store the PyTorch model or dataset/vocab.", ) parser.add_argument( "--tf_checkpoint_path", default="", type=str, help="An optional path to a TensorFlow checkpoint path to be converted.", ) parser.add_argument( "--transfo_xl_config_file", default="", type=str, help=( "An optional config json file corresponding to the pre-trained BERT model. \n" "This specifies the model architecture." ), ) parser.add_argument( "--transfo_xl_dataset_file", default="", type=str, help="An optional dataset file to be converted in a vocabulary.", ) A : List[Any] = parser.parse_args() convert_transfo_xl_checkpoint_to_pytorch( args.tf_checkpoint_path, args.transfo_xl_config_file, args.pytorch_dump_folder_path, args.transfo_xl_dataset_file, )
184
from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available SCREAMING_SNAKE_CASE__ : Optional[Any] = {"configuration_wavlm": ["WAVLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "WavLMConfig"]} try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : int = [ "WAVLM_PRETRAINED_MODEL_ARCHIVE_LIST", "WavLMForAudioFrameClassification", "WavLMForCTC", "WavLMForSequenceClassification", "WavLMForXVector", "WavLMModel", "WavLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_wavlm import WAVLM_PRETRAINED_CONFIG_ARCHIVE_MAP, WavLMConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_wavlm import ( WAVLM_PRETRAINED_MODEL_ARCHIVE_LIST, WavLMForAudioFrameClassification, WavLMForCTC, WavLMForSequenceClassification, WavLMForXVector, WavLMModel, WavLMPreTrainedModel, ) else: import sys SCREAMING_SNAKE_CASE__ : int = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
270
0
"""simple docstring""" from dataclasses import dataclass from typing import Optional, Tuple, Union import numpy as np import torch from ..configuration_utils import ConfigMixin, register_to_config from ..utils import BaseOutput, randn_tensor from .scheduling_utils import SchedulerMixin @dataclass class A_ ( __lowercase ): """simple docstring""" SCREAMING_SNAKE_CASE_ = 42 SCREAMING_SNAKE_CASE_ = 42 SCREAMING_SNAKE_CASE_ = None class A_ ( __lowercase , __lowercase ): """simple docstring""" SCREAMING_SNAKE_CASE_ = 2 @register_to_config def __init__( self :Optional[Any] , lowerCamelCase_ :float = 0.02 , lowerCamelCase_ :float = 100 , lowerCamelCase_ :float = 1.0_07 , lowerCamelCase_ :float = 80 , lowerCamelCase_ :float = 0.05 , lowerCamelCase_ :float = 50 , ): """simple docstring""" lowerCamelCase__ : List[str] =sigma_max # setable values lowerCamelCase__ : Tuple =None lowerCamelCase__ : Optional[int] =None lowerCamelCase__ : int =None # sigma(t_i) def UpperCAmelCase__ ( self :Dict , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :Optional[int] = None ): """simple docstring""" return sample def UpperCAmelCase__ ( self :List[str] , lowerCamelCase_ :int , lowerCamelCase_ :Union[str, torch.device] = None ): """simple docstring""" lowerCamelCase__ : List[Any] =num_inference_steps lowerCamelCase__ : List[str] =np.arange(0 , self.num_inference_steps )[::-1].copy() lowerCamelCase__ : Optional[Any] =torch.from_numpy(SCREAMING_SNAKE_CASE__ ).to(SCREAMING_SNAKE_CASE__ ) lowerCamelCase__ : str =[ ( self.config.sigma_max**2 * (self.config.sigma_min**2 / self.config.sigma_max**2) ** (i / (num_inference_steps - 1)) ) for i in self.timesteps ] lowerCamelCase__ : Union[str, Any] =torch.tensor(SCREAMING_SNAKE_CASE__ , dtype=torch.floataa , device=SCREAMING_SNAKE_CASE__ ) def UpperCAmelCase__ ( self :Tuple , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :float , lowerCamelCase_ :Optional[torch.Generator] = None ): """simple docstring""" if self.config.s_min <= sigma <= self.config.s_max: lowerCamelCase__ : str =min(self.config.s_churn / self.num_inference_steps , 2**0.5 - 1 ) else: lowerCamelCase__ : Dict =0 # sample eps ~ N(0, S_noise^2 * I) lowerCamelCase__ : Union[str, Any] =self.config.s_noise * randn_tensor(sample.shape , generator=SCREAMING_SNAKE_CASE__ ).to(sample.device ) lowerCamelCase__ : Any =sigma + gamma * sigma lowerCamelCase__ : Tuple =sample + ((sigma_hat**2 - sigma**2) ** 0.5 * eps) return sample_hat, sigma_hat def UpperCAmelCase__ ( self :Union[str, Any] , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :float , lowerCamelCase_ :float , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :bool = True , ): """simple docstring""" lowerCamelCase__ : str =sample_hat + sigma_hat * model_output lowerCamelCase__ : Optional[int] =(sample_hat - pred_original_sample) / sigma_hat lowerCamelCase__ : int =sample_hat + (sigma_prev - sigma_hat) * derivative if not return_dict: return (sample_prev, derivative) return KarrasVeOutput( prev_sample=SCREAMING_SNAKE_CASE__ , derivative=SCREAMING_SNAKE_CASE__ , pred_original_sample=SCREAMING_SNAKE_CASE__ ) def UpperCAmelCase__ ( self :Optional[Any] , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :float , lowerCamelCase_ :float , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :torch.FloatTensor , lowerCamelCase_ :bool = True , ): """simple docstring""" lowerCamelCase__ : List[Any] =sample_prev + sigma_prev * model_output lowerCamelCase__ : Tuple =(sample_prev - pred_original_sample) / sigma_prev lowerCamelCase__ : int =sample_hat + (sigma_prev - sigma_hat) * (0.5 * derivative + 0.5 * derivative_corr) if not return_dict: return (sample_prev, derivative) return KarrasVeOutput( prev_sample=SCREAMING_SNAKE_CASE__ , derivative=SCREAMING_SNAKE_CASE__ , pred_original_sample=SCREAMING_SNAKE_CASE__ ) def UpperCAmelCase__ ( self :Optional[Any] , lowerCamelCase_ :Optional[Any] , lowerCamelCase_ :Optional[int] , lowerCamelCase_ :List[Any] ): """simple docstring""" raise NotImplementedError()
126
import pprint import requests SCREAMING_SNAKE_CASE__ : str = "https://zenquotes.io/api" def __magic_name__ ( ) -> list: return requests.get(API_ENDPOINT_URL + '''/today''' ).json() def __magic_name__ ( ) -> list: return requests.get(API_ENDPOINT_URL + '''/random''' ).json() if __name__ == "__main__": SCREAMING_SNAKE_CASE__ : int = random_quotes() pprint.pprint(response)
270
0
"""simple docstring""" import warnings from .generation import TFGenerationMixin class __SCREAMING_SNAKE_CASE ( __lowercase ): '''simple docstring''' warnings.warn( 'Importing `TFGenerationMixin` from `src/transformers/generation_tf_utils.py` is deprecated and will ' 'be removed in Transformers v5. Import as `from transformers import TFGenerationMixin` instead.' , __lowercase , )
238
from maths.is_square_free import is_square_free from maths.prime_factors import prime_factors def __magic_name__ ( __lowerCAmelCase : int ) -> int: __lowerCamelCase = prime_factors(__lowerCAmelCase ) if is_square_free(__lowerCAmelCase ): return -1 if len(__lowerCAmelCase ) % 2 else 1 return 0 if __name__ == "__main__": import doctest doctest.testmod()
270
0
import os from pathlib import Path from unittest.mock import patch import pytest import zstandard as zstd from datasets.download.download_config import DownloadConfig from datasets.utils.file_utils import ( OfflineModeIsEnabled, cached_path, fsspec_get, fsspec_head, ftp_get, ftp_head, get_from_cache, http_get, http_head, ) lowercase__ : List[Any] = "\\n Text data.\n Second line of data." lowercase__ : List[str] = "file" @pytest.fixture(scope="session" ) def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = tmp_path_factory.mktemp("data" ) / (FILE_PATH + ".zstd") snake_case_ = bytes(__lowerCAmelCase , "utf-8" ) with zstd.open(__lowerCAmelCase , "wb" ) as f: f.write(__lowerCAmelCase ) return path @pytest.fixture def lowerCamelCase__ ( _A ): '''simple docstring''' with open(os.path.join(tmpfs.local_root_dir , __lowerCAmelCase ) , "w" ) as f: f.write(__lowerCAmelCase ) return FILE_PATH @pytest.mark.parametrize("compression_format" , ["gzip", "xz", "zstd"] ) def lowerCamelCase__ ( _A , _A , _A , _A , _A , _A ): '''simple docstring''' snake_case_ = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_path} snake_case_ = input_paths[compression_format] snake_case_ = tmp_path / "cache" snake_case_ = DownloadConfig(cache_dir=__lowerCAmelCase , extract_compressed_file=__lowerCAmelCase ) snake_case_ = cached_path(__lowerCAmelCase , download_config=__lowerCAmelCase ) with open(__lowerCAmelCase ) as f: snake_case_ = f.read() with open(__lowerCAmelCase ) as f: snake_case_ = f.read() assert extracted_file_content == expected_file_content @pytest.mark.parametrize("default_extracted" , [True, False] ) @pytest.mark.parametrize("default_cache_dir" , [True, False] ) def lowerCamelCase__ ( _A , _A , _A , _A , _A ): '''simple docstring''' snake_case_ = "custom_cache" snake_case_ = "custom_extracted_dir" snake_case_ = tmp_path / "custom_extracted_path" if default_extracted: snake_case_ = ("downloads" if default_cache_dir else custom_cache_dir, "extracted") else: monkeypatch.setattr("datasets.config.EXTRACTED_DATASETS_DIR" , __lowerCAmelCase ) monkeypatch.setattr("datasets.config.EXTRACTED_DATASETS_PATH" , str(__lowerCAmelCase ) ) snake_case_ = custom_extracted_path.parts[-2:] if default_cache_dir else (custom_cache_dir, custom_extracted_dir) snake_case_ = xz_file snake_case_ = ( DownloadConfig(extract_compressed_file=__lowerCAmelCase ) if default_cache_dir else DownloadConfig(cache_dir=tmp_path / custom_cache_dir , extract_compressed_file=__lowerCAmelCase ) ) snake_case_ = cached_path(__lowerCAmelCase , download_config=__lowerCAmelCase ) assert Path(__lowerCAmelCase ).parent.parts[-2:] == expected def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = str(Path(__lowerCAmelCase ).resolve() ) assert cached_path(__lowerCAmelCase ) == text_file # relative path snake_case_ = str(Path(__lowerCAmelCase ).resolve().relative_to(Path(os.getcwd() ) ) ) assert cached_path(__lowerCAmelCase ) == text_file def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = str(tmp_path.resolve() / "__missing_file__.txt" ) with pytest.raises(__lowerCAmelCase ): cached_path(__lowerCAmelCase ) # relative path snake_case_ = "./__missing_file__.txt" with pytest.raises(__lowerCAmelCase ): cached_path(__lowerCAmelCase ) def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = get_from_cache(f"tmp://{tmpfs_file}" ) with open(__lowerCAmelCase ) as f: snake_case_ = f.read() assert output_file_content == FILE_CONTENT @patch("datasets.config.HF_DATASETS_OFFLINE" , __lowerCAmelCase ) def lowerCamelCase__ ( ): '''simple docstring''' with pytest.raises(__lowerCAmelCase ): cached_path("https://huggingface.co" ) @patch("datasets.config.HF_DATASETS_OFFLINE" , __lowerCAmelCase ) def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = tmp_path_factory.mktemp("data" ) / "file.html" with pytest.raises(__lowerCAmelCase ): http_get("https://huggingface.co" , temp_file=__lowerCAmelCase ) with pytest.raises(__lowerCAmelCase ): http_head("https://huggingface.co" ) @patch("datasets.config.HF_DATASETS_OFFLINE" , __lowerCAmelCase ) def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = tmp_path_factory.mktemp("data" ) / "file.html" with pytest.raises(__lowerCAmelCase ): ftp_get("ftp://huggingface.co" , temp_file=__lowerCAmelCase ) with pytest.raises(__lowerCAmelCase ): ftp_head("ftp://huggingface.co" ) @patch("datasets.config.HF_DATASETS_OFFLINE" , __lowerCAmelCase ) def lowerCamelCase__ ( _A ): '''simple docstring''' snake_case_ = tmp_path_factory.mktemp("data" ) / "file.html" with pytest.raises(__lowerCAmelCase ): fsspec_get("s3://huggingface.co" , temp_file=__lowerCAmelCase ) with pytest.raises(__lowerCAmelCase ): fsspec_head("s3://huggingface.co" )
187
import argparse import collections import json import os import re import string import sys import numpy as np SCREAMING_SNAKE_CASE__ : Union[str, Any] = re.compile(r"\b(a|an|the)\b", re.UNICODE) SCREAMING_SNAKE_CASE__ : int = None def __magic_name__ ( ) -> str: __lowerCamelCase = argparse.ArgumentParser('''Official evaluation script for SQuAD version 2.0.''' ) parser.add_argument('''data_file''' , metavar='''data.json''' , help='''Input data JSON file.''' ) parser.add_argument('''pred_file''' , metavar='''pred.json''' , help='''Model predictions.''' ) parser.add_argument( '''--out-file''' , '''-o''' , metavar='''eval.json''' , help='''Write accuracy metrics to file (default is stdout).''' ) parser.add_argument( '''--na-prob-file''' , '''-n''' , metavar='''na_prob.json''' , help='''Model estimates of probability of no answer.''' ) parser.add_argument( '''--na-prob-thresh''' , '''-t''' , type=__lowerCAmelCase , default=1.0 , help='''Predict "" if no-answer probability exceeds this (default = 1.0).''' , ) parser.add_argument( '''--out-image-dir''' , '''-p''' , metavar='''out_images''' , default=__lowerCAmelCase , help='''Save precision-recall curves to directory.''' ) parser.add_argument('''--verbose''' , '''-v''' , action='''store_true''' ) if len(sys.argv ) == 1: parser.print_help() sys.exit(1 ) return parser.parse_args() def __magic_name__ ( __lowerCAmelCase : List[str] ) -> Union[str, Any]: __lowerCamelCase = {} for article in dataset: for p in article["paragraphs"]: for qa in p["qas"]: __lowerCamelCase = bool(qa['''answers''']['''text'''] ) return qid_to_has_ans def __magic_name__ ( __lowerCAmelCase : Dict ) -> Optional[Any]: def remove_articles(__lowerCAmelCase : Optional[int] ): return ARTICLES_REGEX.sub(''' ''' , __lowerCAmelCase ) def white_space_fix(__lowerCAmelCase : Optional[int] ): return " ".join(text.split() ) def remove_punc(__lowerCAmelCase : Union[str, Any] ): __lowerCamelCase = set(string.punctuation ) return "".join(ch for ch in text if ch not in exclude ) def lower(__lowerCAmelCase : Dict ): return text.lower() return white_space_fix(remove_articles(remove_punc(lower(__lowerCAmelCase ) ) ) ) def __magic_name__ ( __lowerCAmelCase : List[Any] ) -> Optional[int]: if not s: return [] return normalize_answer(__lowerCAmelCase ).split() def __magic_name__ ( __lowerCAmelCase : Tuple , __lowerCAmelCase : Tuple ) -> int: return int(normalize_answer(__lowerCAmelCase ) == normalize_answer(__lowerCAmelCase ) ) def __magic_name__ ( __lowerCAmelCase : Any , __lowerCAmelCase : Tuple ) -> str: __lowerCamelCase = get_tokens(__lowerCAmelCase ) __lowerCamelCase = get_tokens(__lowerCAmelCase ) __lowerCamelCase = collections.Counter(__lowerCAmelCase ) & collections.Counter(__lowerCAmelCase ) __lowerCamelCase = sum(common.values() ) if len(__lowerCAmelCase ) == 0 or len(__lowerCAmelCase ) == 0: # If either is no-answer, then F1 is 1 if they agree, 0 otherwise return int(gold_toks == pred_toks ) if num_same == 0: return 0 __lowerCamelCase = 1.0 * num_same / len(__lowerCAmelCase ) __lowerCamelCase = 1.0 * num_same / len(__lowerCAmelCase ) __lowerCamelCase = (2 * precision * recall) / (precision + recall) return fa def __magic_name__ ( __lowerCAmelCase : Dict , __lowerCAmelCase : List[str] ) -> Optional[Any]: __lowerCamelCase = {} __lowerCamelCase = {} for article in dataset: for p in article["paragraphs"]: for qa in p["qas"]: __lowerCamelCase = qa['''id'''] __lowerCamelCase = [t for t in qa['''answers''']['''text'''] if normalize_answer(__lowerCAmelCase )] if not gold_answers: # For unanswerable questions, only correct answer is empty string __lowerCamelCase = [''''''] if qid not in preds: print(f'''Missing prediction for {qid}''' ) continue __lowerCamelCase = preds[qid] # Take max over all gold answers __lowerCamelCase = max(compute_exact(__lowerCAmelCase , __lowerCAmelCase ) for a in gold_answers ) __lowerCamelCase = max(compute_fa(__lowerCAmelCase , __lowerCAmelCase ) for a in gold_answers ) return exact_scores, fa_scores def __magic_name__ ( __lowerCAmelCase : Tuple , __lowerCAmelCase : str , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : Union[str, Any] ) -> List[str]: __lowerCamelCase = {} for qid, s in scores.items(): __lowerCamelCase = na_probs[qid] > na_prob_thresh if pred_na: __lowerCamelCase = float(not qid_to_has_ans[qid] ) else: __lowerCamelCase = s return new_scores def __magic_name__ ( __lowerCAmelCase : List[Any] , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : Optional[Any]=None ) -> Union[str, Any]: if not qid_list: __lowerCamelCase = len(__lowerCAmelCase ) return collections.OrderedDict( [ ('''exact''', 100.0 * sum(exact_scores.values() ) / total), ('''f1''', 100.0 * sum(fa_scores.values() ) / total), ('''total''', total), ] ) else: __lowerCamelCase = len(__lowerCAmelCase ) return collections.OrderedDict( [ ('''exact''', 100.0 * sum(exact_scores[k] for k in qid_list ) / total), ('''f1''', 100.0 * sum(fa_scores[k] for k in qid_list ) / total), ('''total''', total), ] ) def __magic_name__ ( __lowerCAmelCase : Any , __lowerCAmelCase : str , __lowerCAmelCase : Optional[Any] ) -> int: for k in new_eval: __lowerCamelCase = new_eval[k] def __magic_name__ ( __lowerCAmelCase : Optional[int] , __lowerCAmelCase : List[Any] , __lowerCAmelCase : Dict , __lowerCAmelCase : Union[str, Any] ) -> Optional[Any]: plt.step(__lowerCAmelCase , __lowerCAmelCase , color='''b''' , alpha=0.2 , where='''post''' ) plt.fill_between(__lowerCAmelCase , __lowerCAmelCase , step='''post''' , alpha=0.2 , color='''b''' ) plt.xlabel('''Recall''' ) plt.ylabel('''Precision''' ) plt.xlim([0.0, 1.05] ) plt.ylim([0.0, 1.05] ) plt.title(__lowerCAmelCase ) plt.savefig(__lowerCAmelCase ) plt.clf() def __magic_name__ ( __lowerCAmelCase : Any , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : Tuple , __lowerCAmelCase : Tuple , __lowerCAmelCase : Optional[Any]=None , __lowerCAmelCase : Tuple=None ) -> int: __lowerCamelCase = sorted(__lowerCAmelCase , key=lambda __lowerCAmelCase : na_probs[k] ) __lowerCamelCase = 0.0 __lowerCamelCase = 1.0 __lowerCamelCase = 0.0 __lowerCamelCase = [1.0] __lowerCamelCase = [0.0] __lowerCamelCase = 0.0 for i, qid in enumerate(__lowerCAmelCase ): if qid_to_has_ans[qid]: true_pos += scores[qid] __lowerCamelCase = true_pos / float(i + 1 ) __lowerCamelCase = true_pos / float(__lowerCAmelCase ) if i == len(__lowerCAmelCase ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]: # i.e., if we can put a threshold after this point avg_prec += cur_p * (cur_r - recalls[-1]) precisions.append(__lowerCAmelCase ) recalls.append(__lowerCAmelCase ) if out_image: plot_pr_curve(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) return {"ap": 100.0 * avg_prec} def __magic_name__ ( __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : List[str] , __lowerCAmelCase : Tuple , __lowerCAmelCase : List[str] , __lowerCAmelCase : List[Any] , __lowerCAmelCase : List[Any] ) -> List[Any]: if out_image_dir and not os.path.exists(__lowerCAmelCase ): os.makedirs(__lowerCAmelCase ) __lowerCamelCase = sum(1 for v in qid_to_has_ans.values() if v ) if num_true_pos == 0: return __lowerCamelCase = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , '''pr_exact.png''' ) , title='''Precision-Recall curve for Exact Match score''' , ) __lowerCamelCase = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , '''pr_f1.png''' ) , title='''Precision-Recall curve for F1 score''' , ) __lowerCamelCase = {k: float(__lowerCAmelCase ) for k, v in qid_to_has_ans.items()} __lowerCamelCase = make_precision_recall_eval( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , out_image=os.path.join(__lowerCAmelCase , '''pr_oracle.png''' ) , title='''Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)''' , ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''pr_exact''' ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''pr_f1''' ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''pr_oracle''' ) def __magic_name__ ( __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : Any , __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : int ) -> Optional[Any]: if not qid_list: return __lowerCamelCase = [na_probs[k] for k in qid_list] __lowerCamelCase = np.ones_like(__lowerCAmelCase ) / float(len(__lowerCAmelCase ) ) plt.hist(__lowerCAmelCase , weights=__lowerCAmelCase , bins=20 , range=(0.0, 1.0) ) plt.xlabel('''Model probability of no-answer''' ) plt.ylabel('''Proportion of dataset''' ) plt.title(f'''Histogram of no-answer probability: {name}''' ) plt.savefig(os.path.join(__lowerCAmelCase , f'''na_prob_hist_{name}.png''' ) ) plt.clf() def __magic_name__ ( __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : Optional[Any] , __lowerCAmelCase : List[str] , __lowerCAmelCase : Union[str, Any] ) -> Optional[int]: __lowerCamelCase = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] ) __lowerCamelCase = num_no_ans __lowerCamelCase = cur_score __lowerCamelCase = 0.0 __lowerCamelCase = sorted(__lowerCAmelCase , key=lambda __lowerCAmelCase : na_probs[k] ) for i, qid in enumerate(__lowerCAmelCase ): if qid not in scores: continue if qid_to_has_ans[qid]: __lowerCamelCase = scores[qid] else: if preds[qid]: __lowerCamelCase = -1 else: __lowerCamelCase = 0 cur_score += diff if cur_score > best_score: __lowerCamelCase = cur_score __lowerCamelCase = na_probs[qid] return 100.0 * best_score / len(__lowerCAmelCase ), best_thresh def __magic_name__ ( __lowerCAmelCase : Dict , __lowerCAmelCase : Optional[int] , __lowerCAmelCase : int , __lowerCAmelCase : Union[str, Any] , __lowerCAmelCase : int , __lowerCAmelCase : Optional[Any] ) -> int: __lowerCamelCase , __lowerCamelCase = find_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) __lowerCamelCase , __lowerCamelCase = find_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) __lowerCamelCase = best_exact __lowerCamelCase = exact_thresh __lowerCamelCase = best_fa __lowerCamelCase = fa_thresh def __magic_name__ ( ) -> Optional[int]: with open(OPTS.data_file ) as f: __lowerCamelCase = json.load(__lowerCAmelCase ) __lowerCamelCase = dataset_json['''data'''] with open(OPTS.pred_file ) as f: __lowerCamelCase = json.load(__lowerCAmelCase ) if OPTS.na_prob_file: with open(OPTS.na_prob_file ) as f: __lowerCamelCase = json.load(__lowerCAmelCase ) else: __lowerCamelCase = {k: 0.0 for k in preds} __lowerCamelCase = make_qid_to_has_ans(__lowerCAmelCase ) # maps qid to True/False __lowerCamelCase = [k for k, v in qid_to_has_ans.items() if v] __lowerCamelCase = [k for k, v in qid_to_has_ans.items() if not v] __lowerCamelCase , __lowerCamelCase = get_raw_scores(__lowerCAmelCase , __lowerCAmelCase ) __lowerCamelCase = apply_no_ans_threshold(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.na_prob_thresh ) __lowerCamelCase = apply_no_ans_threshold(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.na_prob_thresh ) __lowerCamelCase = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase ) if has_ans_qids: __lowerCamelCase = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase , qid_list=__lowerCAmelCase ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''HasAns''' ) if no_ans_qids: __lowerCamelCase = make_eval_dict(__lowerCAmelCase , __lowerCAmelCase , qid_list=__lowerCAmelCase ) merge_eval(__lowerCAmelCase , __lowerCAmelCase , '''NoAns''' ) if OPTS.na_prob_file: find_all_best_thresh(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) if OPTS.na_prob_file and OPTS.out_image_dir: run_precision_recall_analysis(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir ) histogram_na_prob(__lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir , '''hasAns''' ) histogram_na_prob(__lowerCAmelCase , __lowerCAmelCase , OPTS.out_image_dir , '''noAns''' ) if OPTS.out_file: with open(OPTS.out_file , '''w''' ) as f: json.dump(__lowerCAmelCase , __lowerCAmelCase ) else: print(json.dumps(__lowerCAmelCase , indent=2 ) ) if __name__ == "__main__": SCREAMING_SNAKE_CASE__ : Any = parse_args() if OPTS.out_image_dir: import matplotlib matplotlib.use("Agg") import matplotlib.pyplot as plt main()
270
0
import re import warnings from contextlib import contextmanager from ...processing_utils import ProcessorMixin class __UpperCAmelCase (__lowercase ): __snake_case : Any = ["""image_processor""", """tokenizer"""] __snake_case : Any = """AutoImageProcessor""" __snake_case : Union[str, Any] = """AutoTokenizer""" def __init__( self: Tuple , UpperCAmelCase_: Optional[int]=None , UpperCAmelCase_: Optional[Any]=None , **UpperCAmelCase_: List[Any] ): '''simple docstring''' _SCREAMING_SNAKE_CASE = None if "feature_extractor" in kwargs: warnings.warn( """The `feature_extractor` argument is deprecated and will be removed in v5, use `image_processor`""" """ instead.""" , SCREAMING_SNAKE_CASE__ , ) _SCREAMING_SNAKE_CASE = kwargs.pop("""feature_extractor""" ) _SCREAMING_SNAKE_CASE = image_processor if image_processor is not None else feature_extractor if image_processor is None: raise ValueError("""You need to specify an `image_processor`.""" ) if tokenizer is None: raise ValueError("""You need to specify a `tokenizer`.""" ) super().__init__(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) _SCREAMING_SNAKE_CASE = self.image_processor _SCREAMING_SNAKE_CASE = False def __call__( self: Union[str, Any] , *UpperCAmelCase_: Any , **UpperCAmelCase_: Optional[Any] ): '''simple docstring''' if self._in_target_context_manager: return self.current_processor(*SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) _SCREAMING_SNAKE_CASE = kwargs.pop("""images""" , SCREAMING_SNAKE_CASE__ ) _SCREAMING_SNAKE_CASE = kwargs.pop("""text""" , SCREAMING_SNAKE_CASE__ ) if len(SCREAMING_SNAKE_CASE__ ) > 0: _SCREAMING_SNAKE_CASE = args[0] _SCREAMING_SNAKE_CASE = args[1:] if images is None and text is None: raise ValueError("""You need to specify either an `images` or `text` input to process.""" ) if images is not None: _SCREAMING_SNAKE_CASE = self.image_processor(SCREAMING_SNAKE_CASE__ , *SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) if text is not None: _SCREAMING_SNAKE_CASE = self.tokenizer(SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) if text is None: return inputs elif images is None: return encodings else: _SCREAMING_SNAKE_CASE = encodings["""input_ids"""] return inputs def UpperCamelCase ( self: Tuple , *UpperCAmelCase_: Any , **UpperCAmelCase_: Any ): '''simple docstring''' return self.tokenizer.batch_decode(*SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) def UpperCamelCase ( self: Optional[int] , *UpperCAmelCase_: Optional[int] , **UpperCAmelCase_: int ): '''simple docstring''' return self.tokenizer.decode(*SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) @contextmanager def UpperCamelCase ( self: List[str] ): '''simple docstring''' warnings.warn( """`as_target_processor` is deprecated and will be removed in v5 of Transformers. You can process your """ """labels by using the argument `text` of the regular `__call__` method (either in the same call as """ """your images inputs, or in a separate call.""" ) _SCREAMING_SNAKE_CASE = True _SCREAMING_SNAKE_CASE = self.tokenizer yield _SCREAMING_SNAKE_CASE = self.image_processor _SCREAMING_SNAKE_CASE = False def UpperCamelCase ( self: Any , UpperCAmelCase_: List[Any] , UpperCAmelCase_: List[Any]=False , UpperCAmelCase_: List[Any]=None ): '''simple docstring''' if added_vocab is None: _SCREAMING_SNAKE_CASE = self.tokenizer.get_added_vocab() _SCREAMING_SNAKE_CASE = {} while tokens: _SCREAMING_SNAKE_CASE = re.search(R"""<s_(.*?)>""" , SCREAMING_SNAKE_CASE__ , re.IGNORECASE ) if start_token is None: break _SCREAMING_SNAKE_CASE = start_token.group(1 ) _SCREAMING_SNAKE_CASE = re.search(RF'</s_{key}>' , SCREAMING_SNAKE_CASE__ , re.IGNORECASE ) _SCREAMING_SNAKE_CASE = start_token.group() if end_token is None: _SCREAMING_SNAKE_CASE = tokens.replace(SCREAMING_SNAKE_CASE__ , """""" ) else: _SCREAMING_SNAKE_CASE = end_token.group() _SCREAMING_SNAKE_CASE = re.escape(SCREAMING_SNAKE_CASE__ ) _SCREAMING_SNAKE_CASE = re.escape(SCREAMING_SNAKE_CASE__ ) _SCREAMING_SNAKE_CASE = re.search(F'{start_token_escaped}(.*?){end_token_escaped}' , SCREAMING_SNAKE_CASE__ , re.IGNORECASE ) if content is not None: _SCREAMING_SNAKE_CASE = content.group(1 ).strip() if r"<s_" in content and r"</s_" in content: # non-leaf node _SCREAMING_SNAKE_CASE = self.tokenajson(SCREAMING_SNAKE_CASE__ , is_inner_value=SCREAMING_SNAKE_CASE__ , added_vocab=SCREAMING_SNAKE_CASE__ ) if value: if len(SCREAMING_SNAKE_CASE__ ) == 1: _SCREAMING_SNAKE_CASE = value[0] _SCREAMING_SNAKE_CASE = value else: # leaf nodes _SCREAMING_SNAKE_CASE = [] for leaf in content.split(R"""<sep/>""" ): _SCREAMING_SNAKE_CASE = leaf.strip() if leaf in added_vocab and leaf[0] == "<" and leaf[-2:] == "/>": _SCREAMING_SNAKE_CASE = leaf[1:-2] # for categorical special tokens output[key].append(SCREAMING_SNAKE_CASE__ ) if len(output[key] ) == 1: _SCREAMING_SNAKE_CASE = output[key][0] _SCREAMING_SNAKE_CASE = tokens[tokens.find(SCREAMING_SNAKE_CASE__ ) + len(SCREAMING_SNAKE_CASE__ ) :].strip() if tokens[:6] == r"<sep/>": # non-leaf nodes return [output] + self.tokenajson(tokens[6:] , is_inner_value=SCREAMING_SNAKE_CASE__ , added_vocab=SCREAMING_SNAKE_CASE__ ) if len(SCREAMING_SNAKE_CASE__ ): return [output] if is_inner_value else output else: return [] if is_inner_value else {"text_sequence": tokens} @property def UpperCamelCase ( self: Optional[Any] ): '''simple docstring''' warnings.warn( """`feature_extractor_class` is deprecated and will be removed in v5. Use `image_processor_class` instead.""" , SCREAMING_SNAKE_CASE__ , ) return self.image_processor_class @property def UpperCamelCase ( self: Any ): '''simple docstring''' warnings.warn( """`feature_extractor` is deprecated and will be removed in v5. Use `image_processor` instead.""" , SCREAMING_SNAKE_CASE__ , ) return self.image_processor
306
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 SCREAMING_SNAKE_CASE__ : Optional[int] = os.path.abspath(os.path.dirname(__file__)) with ExtendSysPath(F'{bindir}/../../examples/pytorch/translation'): from run_translation import main # noqa set_seed(42) SCREAMING_SNAKE_CASE__ : Any = "sshleifer/student_marian_en_ro_6_1" SCREAMING_SNAKE_CASE__ : Tuple = "sshleifer/tiny-mbart" @require_torch class lowerCAmelCase__ ( __lowercase ): def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : Tuple=False , SCREAMING_SNAKE_CASE__ : Optional[Any]=None , SCREAMING_SNAKE_CASE__ : Optional[Any]=True , SCREAMING_SNAKE_CASE__ : str=True , SCREAMING_SNAKE_CASE__ : Optional[Any]=True , SCREAMING_SNAKE_CASE__ : Optional[int]=True , ) -> Optional[int]: __lowerCamelCase = self.run_trainer( eval_steps=1 , max_len=12 , model_name=SCREAMING_SNAKE_CASE__ , num_train_epochs=1 , distributed=SCREAMING_SNAKE_CASE__ , extra_args_str=SCREAMING_SNAKE_CASE__ , predict_with_generate=SCREAMING_SNAKE_CASE__ , do_train=SCREAMING_SNAKE_CASE__ , do_eval=SCREAMING_SNAKE_CASE__ , do_predict=SCREAMING_SNAKE_CASE__ , ) __lowerCamelCase = TrainerState.load_from_json(os.path.join(SCREAMING_SNAKE_CASE__ , '''trainer_state.json''' ) ).log_history if not do_eval: return __lowerCamelCase = [log for log in logs if '''eval_loss''' in log.keys()] __lowerCamelCase = eval_metrics[0] if predict_with_generate: assert "eval_bleu" in first_step_stats __lowerCamelCase = eval_metrics[-1] assert isinstance(last_step_stats['''eval_bleu'''] , SCREAMING_SNAKE_CASE__ ) assert not math.isnan(float(last_step_stats['''eval_loss'''] ) ), "eval_loss must not be `nan`" @require_torch_non_multi_gpu def __A ( self : Optional[int] ) -> int: self.run_seqaseq_quick() @require_torch_multi_gpu def __A ( self : int ) -> List[str]: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ ) @require_torch_multi_gpu def __A ( self : Optional[Any] ) -> Tuple: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ ) @unittest.skip('''Requires an update of the env running those tests''' ) @require_torch_multi_gpu @require_fairscale def __A ( self : Dict ) -> Tuple: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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 : Optional[int] ) -> List[str]: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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 ) -> Any: self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , extra_args_str='''--sharded_ddp zero_dp_2''' , predict_with_generate=SCREAMING_SNAKE_CASE__ ) @unittest.skip('''Requires an update of the env running those tests''' ) @require_torch_multi_gpu @require_fairscale def __A ( self : Dict ) -> Tuple: self.run_seqaseq_quick( distributed=SCREAMING_SNAKE_CASE__ , extra_args_str='''--sharded_ddp zero_dp_2 --fp16''' , predict_with_generate=SCREAMING_SNAKE_CASE__ ) @require_apex @require_torch_gpu def __A ( self : Union[str, Any] ) -> List[str]: # XXX: apex breaks the trainer if it's run twice e.g. run_seq2seq.main() from the same # program and it breaks other tests that run from the same pytest worker, therefore until this is # sorted out it must be run only in an external program, that is distributed=True in this # test and only under one or more gpus - if we want cpu will need to make a special test # # specifically to the problem traced it to self.optimizer.step() - if it's run 2nd time via # 2nd main() call it botches the future eval. # self.run_seqaseq_quick(distributed=SCREAMING_SNAKE_CASE__ , 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=SCREAMING_SNAKE_CASE__ , extra_args_str='''--fp16 --fp16_backend=apex''' ) @parameterized.expand(['''base''', '''low''', '''high''', '''mixed'''] ) @require_torch_multi_gpu def __A ( self : Any , SCREAMING_SNAKE_CASE__ : List[str] ) -> Optional[Any]: # as each sub-test is slow-ish split into multiple sub-tests to avoid CI timeout __lowerCamelCase = { # 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}, } __lowerCamelCase = experiments[experiment_id] __lowerCamelCase = {'''distributed''': True, '''predict_with_generate''': False, '''do_eval''': False, '''do_predict''': False} __lowerCamelCase = '''Running training''' with CaptureStderr() as cl: self.run_seqaseq_quick(**SCREAMING_SNAKE_CASE__ , extra_args_str=data['''extra_args_str'''] ) __lowerCamelCase = len(re.findall(SCREAMING_SNAKE_CASE__ , cl.err ) ) self.assertEqual(SCREAMING_SNAKE_CASE__ , data['''n_matches'''] ) @slow def __A ( self : Any ) -> Optional[Any]: __lowerCamelCase = self.run_trainer( eval_steps=2 , max_len=1_28 , model_name=SCREAMING_SNAKE_CASE__ , learning_rate=3e-4 , num_train_epochs=10 , distributed=SCREAMING_SNAKE_CASE__ , ) # Check metrics __lowerCamelCase = TrainerState.load_from_json(os.path.join(SCREAMING_SNAKE_CASE__ , '''trainer_state.json''' ) ).log_history __lowerCamelCase = [log for log in logs if '''eval_loss''' in log.keys()] __lowerCamelCase = eval_metrics[0] __lowerCamelCase = eval_metrics[-1] assert first_step_stats["eval_loss"] > last_step_stats["eval_loss"], "model learned nothing" assert isinstance(last_step_stats['''eval_bleu'''] , SCREAMING_SNAKE_CASE__ ) # test if do_predict saves generations and metrics __lowerCamelCase = os.listdir(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = {os.path.basename(SCREAMING_SNAKE_CASE__ ) for p in contents} assert "generated_predictions.txt" in contents assert "predict_results.json" in contents @slow @require_bitsandbytes def __A ( self : Optional[int] ) -> str: from transformers.training_args import OptimizerNames def train_and_return_metrics(SCREAMING_SNAKE_CASE__ : str ) -> Tuple[int, float]: __lowerCamelCase = '''--skip_memory_metrics 0''' __lowerCamelCase = self.run_trainer( max_len=1_28 , model_name=SCREAMING_SNAKE_CASE__ , learning_rate=3e-4 , num_train_epochs=1 , optim=SCREAMING_SNAKE_CASE__ , distributed=SCREAMING_SNAKE_CASE__ , extra_args_str=SCREAMING_SNAKE_CASE__ , do_eval=SCREAMING_SNAKE_CASE__ , do_predict=SCREAMING_SNAKE_CASE__ , n_gpus_to_use=1 , ) # Check metrics __lowerCamelCase = TrainerState.load_from_json(Path(SCREAMING_SNAKE_CASE__ , '''trainer_state.json''' ) ).log_history __lowerCamelCase = int(logs[0]['''train_mem_gpu_peaked_delta'''] / 2**20 ) __lowerCamelCase = int(logs[0]['''train_mem_gpu_alloc_delta'''] / 2**20 ) __lowerCamelCase = logs[0]['''train_loss'''] return gpu_peak_mem_mb, gpu_alloc_mem_mb, loss __lowerCamelCase , __lowerCamelCase , __lowerCamelCase = train_and_return_metrics(OptimizerNames.ADAMW_TORCH.value ) __lowerCamelCase , __lowerCamelCase , __lowerCamelCase = train_and_return_metrics(OptimizerNames.ADAMW_BNB.value ) __lowerCamelCase = gpu_alloc_mem_orig - gpu_alloc_mem_bnb __lowerCamelCase = gpu_peak_mem_orig + gpu_alloc_mem_orig __lowerCamelCase = gpu_peak_mem_bnb + gpu_alloc_mem_bnb __lowerCamelCase = 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 __lowerCamelCase = 1_20 # 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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , '''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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , '''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( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , f'''loss should be the same, but got loss_orig={loss_orig}, loss_bnb={loss_bnb}''' ) def __A ( self : Dict , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : float = 3e-3 , SCREAMING_SNAKE_CASE__ : str = "adafactor" , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : str = None , SCREAMING_SNAKE_CASE__ : int = 0 , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : int = None , ) -> List[Any]: __lowerCamelCase = self.test_file_dir / '''../fixtures/tests_samples/wmt_en_ro''' __lowerCamelCase = self.get_auto_remove_tmp_dir() __lowerCamelCase = f''' --model_name_or_path {model_name} --train_file {data_dir}/train.json --validation_file {data_dir}/val.json --test_file {data_dir}/test.json --output_dir {output_dir} --overwrite_output_dir --max_train_samples 8 --max_source_length {max_len} --max_target_length {max_len} --do_train --num_train_epochs {str(SCREAMING_SNAKE_CASE__ )} --per_device_train_batch_size 4 --learning_rate {learning_rate} --warmup_steps 8 --logging_steps 0 --logging_strategy no --save_steps {str(SCREAMING_SNAKE_CASE__ )} --group_by_length --label_smoothing_factor 0.1 --target_lang ro_RO --source_lang en_XX '''.split() __lowerCamelCase = f''' --do_eval --per_device_eval_batch_size 4 --max_eval_samples 8 --val_max_target_length {max_len} --evaluation_strategy steps --eval_steps {str(SCREAMING_SNAKE_CASE__ )} '''.split() __lowerCamelCase = ''' --do_predict '''.split() __lowerCamelCase = [] 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: __lowerCamelCase = get_gpu_count() __lowerCamelCase = get_torch_dist_unique_port() __lowerCamelCase = f''' -m torch.distributed.run --nproc_per_node={n_gpus_to_use} --master_port={master_port} {self.examples_dir_str}/pytorch/translation/run_translation.py '''.split() __lowerCamelCase = [sys.executable] + distributed_args + args # keep for quick debug # print(" ".join([f"\nPYTHONPATH={self.src_dir_str}"] +cmd)); die execute_subprocess_async(SCREAMING_SNAKE_CASE__ , env=self.get_env() ) else: __lowerCamelCase = ['''run_translation.py'''] + args with patch.object(SCREAMING_SNAKE_CASE__ , '''argv''' , SCREAMING_SNAKE_CASE__ ): main() return output_dir
270
0
'''simple docstring''' import logging from transformers.configuration_utils import PretrainedConfig lowerCAmelCase : Dict =logging.getLogger(__name__) class a_ ( __lowercase ): __A = """masked_bert""" def __init__( self : Union[str, Any] , lowercase : Optional[Any]=30_522 , lowercase : Any=768 , lowercase : List[Any]=12 , lowercase : Tuple=12 , lowercase : Tuple=3_072 , lowercase : List[str]="gelu" , lowercase : List[str]=0.1 , lowercase : Optional[Any]=0.1 , lowercase : str=512 , lowercase : List[Any]=2 , lowercase : List[str]=0.02 , lowercase : Dict=1e-1_2 , lowercase : int=0 , lowercase : Optional[int]="topK" , lowercase : str="constant" , lowercase : int=0.0 , **lowercase : Optional[Any] , ): """simple docstring""" super().__init__(pad_token_id=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) lowercase_ :Dict = vocab_size lowercase_ :Any = hidden_size lowercase_ :Optional[Any] = num_hidden_layers lowercase_ :List[Any] = num_attention_heads lowercase_ :Optional[int] = hidden_act lowercase_ :List[str] = intermediate_size lowercase_ :str = hidden_dropout_prob lowercase_ :List[str] = attention_probs_dropout_prob lowercase_ :Optional[int] = max_position_embeddings lowercase_ :Tuple = type_vocab_size lowercase_ :Tuple = initializer_range lowercase_ :str = layer_norm_eps lowercase_ :Dict = pruning_method lowercase_ :Tuple = mask_init lowercase_ :List[str] = mask_scale
223
import copy from typing import Dict, List, Optional from ...configuration_utils import PretrainedConfig from ...utils import logging from ..auto import CONFIG_MAPPING SCREAMING_SNAKE_CASE__ : Dict = { "facebook/mask2former-swin-small-coco-instance": ( "https://huggingface.co/facebook/mask2former-swin-small-coco-instance/blob/main/config.json" ) # See all Mask2Former models at https://huggingface.co/models?filter=mask2former } SCREAMING_SNAKE_CASE__ : int = logging.get_logger(__name__) class lowerCAmelCase__ ( __lowercase ): a__ : Any = """mask2former""" a__ : Dict = ["""swin"""] a__ : Any = {"""hidden_size""": """hidden_dim"""} def __init__( self : Any , SCREAMING_SNAKE_CASE__ : Optional[Dict] = None , SCREAMING_SNAKE_CASE__ : int = 2_56 , SCREAMING_SNAKE_CASE__ : int = 2_56 , SCREAMING_SNAKE_CASE__ : int = 2_56 , SCREAMING_SNAKE_CASE__ : int = 10_24 , SCREAMING_SNAKE_CASE__ : str = "relu" , SCREAMING_SNAKE_CASE__ : int = 6 , SCREAMING_SNAKE_CASE__ : int = 10 , SCREAMING_SNAKE_CASE__ : int = 8 , SCREAMING_SNAKE_CASE__ : float = 0.0 , SCREAMING_SNAKE_CASE__ : int = 20_48 , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : int = 4 , SCREAMING_SNAKE_CASE__ : int = 2_55 , SCREAMING_SNAKE_CASE__ : int = 1_00 , SCREAMING_SNAKE_CASE__ : float = 0.1 , SCREAMING_SNAKE_CASE__ : float = 2.0 , SCREAMING_SNAKE_CASE__ : float = 5.0 , SCREAMING_SNAKE_CASE__ : float = 5.0 , SCREAMING_SNAKE_CASE__ : int = 1_25_44 , SCREAMING_SNAKE_CASE__ : float = 3.0 , SCREAMING_SNAKE_CASE__ : float = 0.75 , SCREAMING_SNAKE_CASE__ : float = 0.02 , SCREAMING_SNAKE_CASE__ : float = 1.0 , SCREAMING_SNAKE_CASE__ : bool = True , SCREAMING_SNAKE_CASE__ : List[int] = [4, 8, 16, 32] , SCREAMING_SNAKE_CASE__ : bool = None , **SCREAMING_SNAKE_CASE__ : Tuple , ) -> str: if backbone_config is None: logger.info('''`backbone_config` is `None`. Initializing the config with the default `Swin` backbone.''' ) __lowerCamelCase = CONFIG_MAPPING['''swin''']( image_size=2_24 , in_channels=3 , patch_size=4 , embed_dim=96 , depths=[2, 2, 18, 2] , num_heads=[3, 6, 12, 24] , window_size=7 , drop_path_rate=0.3 , use_absolute_embeddings=SCREAMING_SNAKE_CASE__ , out_features=['''stage1''', '''stage2''', '''stage3''', '''stage4'''] , ) if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): __lowerCamelCase = backbone_config.pop('''model_type''' ) __lowerCamelCase = CONFIG_MAPPING[backbone_model_type] __lowerCamelCase = config_class.from_dict(SCREAMING_SNAKE_CASE__ ) # verify that the backbone is supported if backbone_config.model_type not in self.backbones_supported: logger.warning_once( f'''Backbone {backbone_config.model_type} is not a supported model and may not be compatible with Mask2Former. ''' f'''Supported model types: {','.join(self.backbones_supported )}''' ) __lowerCamelCase = backbone_config __lowerCamelCase = feature_size __lowerCamelCase = mask_feature_size __lowerCamelCase = hidden_dim __lowerCamelCase = encoder_feedforward_dim __lowerCamelCase = activation_function __lowerCamelCase = encoder_layers __lowerCamelCase = decoder_layers __lowerCamelCase = num_attention_heads __lowerCamelCase = dropout __lowerCamelCase = dim_feedforward __lowerCamelCase = pre_norm __lowerCamelCase = enforce_input_projection __lowerCamelCase = common_stride __lowerCamelCase = ignore_value __lowerCamelCase = num_queries __lowerCamelCase = no_object_weight __lowerCamelCase = class_weight __lowerCamelCase = mask_weight __lowerCamelCase = dice_weight __lowerCamelCase = train_num_points __lowerCamelCase = oversample_ratio __lowerCamelCase = importance_sample_ratio __lowerCamelCase = init_std __lowerCamelCase = init_xavier_std __lowerCamelCase = use_auxiliary_loss __lowerCamelCase = feature_strides __lowerCamelCase = output_auxiliary_logits __lowerCamelCase = decoder_layers super().__init__(**SCREAMING_SNAKE_CASE__ ) @classmethod def __A ( cls : Any , SCREAMING_SNAKE_CASE__ : PretrainedConfig , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> List[Any]: return cls( backbone_config=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ , ) def __A ( self : Any ) -> Dict[str, any]: __lowerCamelCase = copy.deepcopy(self.__dict__ ) __lowerCamelCase = self.backbone_config.to_dict() __lowerCamelCase = self.__class__.model_type return output
270
0
'''simple docstring''' import json import os import unittest from transformers import OpenAIGPTTokenizer, OpenAIGPTTokenizerFast from transformers.models.openai.tokenization_openai import VOCAB_FILES_NAMES from transformers.testing_utils import require_ftfy, require_spacy, require_tokenizers from ...test_tokenization_common import TokenizerTesterMixin @require_tokenizers class UpperCAmelCase ( __lowercase , unittest.TestCase ): __lowercase = OpenAIGPTTokenizer __lowercase = OpenAIGPTTokenizerFast __lowercase = True __lowercase = False def UpperCAmelCase_ ( self :Optional[Any] )-> int: super().setUp() # Adapted from Sennrich et al. 2015 and https://github.com/rsennrich/subword-nmt A__ = [ "l", "o", "w", "e", "r", "s", "t", "i", "d", "n", "w</w>", "r</w>", "t</w>", "lo", "low", "er</w>", "low</w>", "lowest</w>", "newer</w>", "wider</w>", "<unk>", ] A__ = dict(zip(SCREAMING_SNAKE_CASE__ , range(len(SCREAMING_SNAKE_CASE__ ) ) ) ) A__ = ["#version: 0.2", "l o", "lo w", "e r</w>", ""] A__ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["vocab_file"] ) A__ = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["merges_file"] ) with open(self.vocab_file , "w" ) as fp: fp.write(json.dumps(SCREAMING_SNAKE_CASE__ ) ) with open(self.merges_file , "w" ) as fp: fp.write("\n".join(SCREAMING_SNAKE_CASE__ ) ) def UpperCAmelCase_ ( self :Optional[int] , lowercase_ :Union[str, Any] )-> List[str]: return "lower newer", "lower newer" def UpperCAmelCase_ ( self :Optional[Any] )-> List[str]: A__ = OpenAIGPTTokenizer(self.vocab_file , self.merges_file ) A__ = "lower" A__ = ["low", "er</w>"] A__ = tokenizer.tokenize(SCREAMING_SNAKE_CASE__ ) self.assertListEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) A__ = tokens + ["<unk>"] A__ = [14, 15, 20] self.assertListEqual(tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) , SCREAMING_SNAKE_CASE__ ) def UpperCAmelCase_ ( self :Tuple , lowercase_ :Optional[int]=15 )-> Tuple: for tokenizer, pretrained_name, kwargs in self.tokenizers_list: with self.subTest(F"{tokenizer.__class__.__name__} ({pretrained_name})" ): A__ = self.rust_tokenizer_class.from_pretrained(SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ ) # Simple input A__ = "This is a simple input" A__ = ["This is a simple input 1", "This is a simple input 2"] A__ = ("This is a simple input", "This is a pair") A__ = [ ("This is a simple input 1", "This is a simple input 2"), ("This is a simple pair 1", "This is a simple pair 2"), ] # Simple input tests self.assertRaises(SCREAMING_SNAKE_CASE__ , tokenizer_r.encode , SCREAMING_SNAKE_CASE__ , max_length=SCREAMING_SNAKE_CASE__ , padding="max_length" ) # Simple input self.assertRaises(SCREAMING_SNAKE_CASE__ , tokenizer_r.encode_plus , SCREAMING_SNAKE_CASE__ , max_length=SCREAMING_SNAKE_CASE__ , padding="max_length" ) # Simple input self.assertRaises( SCREAMING_SNAKE_CASE__ , tokenizer_r.batch_encode_plus , SCREAMING_SNAKE_CASE__ , max_length=SCREAMING_SNAKE_CASE__ , padding="max_length" , ) # Pair input self.assertRaises(SCREAMING_SNAKE_CASE__ , tokenizer_r.encode , SCREAMING_SNAKE_CASE__ , max_length=SCREAMING_SNAKE_CASE__ , padding="max_length" ) # Pair input self.assertRaises(SCREAMING_SNAKE_CASE__ , tokenizer_r.encode_plus , SCREAMING_SNAKE_CASE__ , max_length=SCREAMING_SNAKE_CASE__ , padding="max_length" ) # Pair input self.assertRaises( SCREAMING_SNAKE_CASE__ , tokenizer_r.batch_encode_plus , SCREAMING_SNAKE_CASE__ , max_length=SCREAMING_SNAKE_CASE__ , padding="max_length" , ) def UpperCAmelCase_ ( self :Any )-> Tuple: pass @require_ftfy @require_spacy @require_tokenizers class UpperCAmelCase ( __lowercase ): pass
237
import os def __magic_name__ ( ) -> str: __lowerCamelCase = os.path.join(os.path.dirname(__lowerCAmelCase ) , '''num.txt''' ) with open(__lowerCAmelCase ) as file_hand: return str(sum(int(__lowerCAmelCase ) for line in file_hand ) )[:10] if __name__ == "__main__": print(solution())
270
0
import argparse import json import os import fairseq import torch from fairseq.data import Dictionary # Register SEW's fairseq modules from sew_asapp import tasks # noqa: F401 from transformers import ( SEWConfig, SEWForCTC, SEWModel, WavaVecaCTCTokenizer, WavaVecaFeatureExtractor, WavaVecaProcessor, logging, ) logging.set_verbosity_info() lowerCAmelCase__ = logging.get_logger(__name__) lowerCAmelCase__ = { "post_extract_proj": "feature_projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.upsample.0": "encoder.upsample.projection", "encoder.layer_norm": "encoder.layer_norm", "w2v_model.layer_norm": "layer_norm", "w2v_encoder.proj": "lm_head", "mask_emb": "masked_spec_embed", } def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: int , SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: List[Any] , SCREAMING_SNAKE_CASE_: Tuple ) -> Optional[int]: '''simple docstring''' for attribute in key.split("." ): A__ = getattr(__lowerCAmelCase , __lowerCAmelCase ) if weight_type is not None: A__ = getattr(__lowerCAmelCase , __lowerCAmelCase ).shape else: A__ = hf_pointer.shape assert hf_shape == value.shape, ( F'Shape of hf {key + "." + weight_type if weight_type is not None else ""} is {hf_shape}, but should be' F' {value.shape} for {full_name}' ) if weight_type == "weight": A__ = value elif weight_type == "weight_g": A__ = value elif weight_type == "weight_v": A__ = value elif weight_type == "bias": A__ = value else: A__ = value logger.info(F'{key + "." + weight_type if weight_type is not None else ""} was initialized from {full_name}.' ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: str , SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: List[Any] ) -> List[str]: '''simple docstring''' A__ = [] A__ = fairseq_model.state_dict() A__ = hf_model.sew.feature_extractor if is_finetuned else hf_model.feature_extractor for name, value in fairseq_dict.items(): A__ = False if "conv_layers" in name: load_conv_layer( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , hf_model.config.feat_extract_norm == "group" , ) A__ = True else: for key, mapped_key in MAPPING.items(): A__ = "sew." + mapped_key if (is_finetuned and mapped_key != "lm_head") else mapped_key if key in name or key.split("w2v_model." )[-1] == name.split("." )[0]: A__ = True if "*" in mapped_key: A__ = name.split(__lowerCAmelCase )[0].split("." )[-2] A__ = mapped_key.replace("*" , __lowerCAmelCase ) if "weight_g" in name: A__ = "weight_g" elif "weight_v" in name: A__ = "weight_v" elif "weight" in name: A__ = "weight" elif "bias" in name: A__ = "bias" else: A__ = None set_recursively(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) continue if not is_used: unused_weights.append(__lowerCAmelCase ) logger.warning(F'Unused weights: {unused_weights}' ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: Optional[Any] , SCREAMING_SNAKE_CASE_: Union[str, Any] , SCREAMING_SNAKE_CASE_: Tuple , SCREAMING_SNAKE_CASE_: List[Any] ) -> Optional[Any]: '''simple docstring''' A__ = full_name.split("conv_layers." )[-1] A__ = name.split("." ) A__ = int(items[0] ) A__ = int(items[1] ) if type_id == 0: if "bias" in name: assert value.shape == feature_extractor.conv_layers[layer_id].conv.bias.data.shape, ( F'{full_name} has size {value.shape}, but' F' {feature_extractor.conv_layers[layer_id].conv.bias.data.shape} was found.' ) A__ = value logger.info(F'Feat extract conv layer {layer_id} was initialized from {full_name}.' ) elif "weight" in name: assert value.shape == feature_extractor.conv_layers[layer_id].conv.weight.data.shape, ( F'{full_name} has size {value.shape}, but' F' {feature_extractor.conv_layers[layer_id].conv.weight.data.shape} was found.' ) A__ = value logger.info(F'Feat extract conv layer {layer_id} was initialized from {full_name}.' ) elif (type_id == 2 and not use_group_norm) or (type_id == 2 and layer_id == 0 and use_group_norm): if "bias" in name: assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape, ( F'{full_name} has size {value.shape}, but {feature_extractor[layer_id].layer_norm.bias.data.shape} was' " found." ) A__ = value logger.info(F'Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.' ) elif "weight" in name: assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape, ( F'{full_name} has size {value.shape}, but' F' {feature_extractor[layer_id].layer_norm.weight.data.shape} was found.' ) A__ = value logger.info(F'Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.' ) else: unused_weights.append(__lowerCAmelCase ) def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Optional[int] , SCREAMING_SNAKE_CASE_: List[str] ) -> List[Any]: '''simple docstring''' A__ = SEWConfig() if is_finetuned: A__ = model.wav_encoder.wav_model.cfg else: A__ = model.cfg A__ = fs_config.conv_bias A__ = eval(fs_config.conv_feature_layers ) A__ = [x[0] for x in conv_layers] A__ = [x[1] for x in conv_layers] A__ = [x[2] for x in conv_layers] A__ = "gelu" A__ = "layer" if fs_config.extractor_mode == "layer_norm" else "group" A__ = 0.0 A__ = fs_config.activation_fn.name A__ = fs_config.encoder_embed_dim A__ = 0.02 A__ = fs_config.encoder_ffn_embed_dim A__ = 1e-5 A__ = fs_config.encoder_layerdrop A__ = fs_config.encoder_attention_heads A__ = fs_config.conv_pos_groups A__ = fs_config.conv_pos A__ = len(__lowerCAmelCase ) A__ = fs_config.encoder_layers A__ = fs_config.squeeze_factor # take care of any params that are overridden by the Wav2VecCtc model if is_finetuned: A__ = model.cfg A__ = fs_config.final_dropout A__ = fs_config.layerdrop A__ = fs_config.activation_dropout A__ = fs_config.mask_prob > 0 or fs_config.mask_channel_prob > 0 A__ = fs_config.attention_dropout A__ = fs_config.dropout_input A__ = fs_config.dropout A__ = fs_config.mask_channel_length A__ = fs_config.mask_channel_prob A__ = fs_config.mask_length A__ = fs_config.mask_prob A__ = "Wav2Vec2FeatureExtractor" A__ = "Wav2Vec2CTCTokenizer" return config @torch.no_grad() def lowerCAmelCase__ ( SCREAMING_SNAKE_CASE_: Union[str, Any] , SCREAMING_SNAKE_CASE_: str , SCREAMING_SNAKE_CASE_: Optional[Any]=None , SCREAMING_SNAKE_CASE_: Union[str, Any]=None , SCREAMING_SNAKE_CASE_: Tuple=True ) -> int: '''simple docstring''' if is_finetuned: A__ , A__ , A__ = fairseq.checkpoint_utils.load_model_ensemble_and_task( [checkpoint_path] , arg_overrides={"data": "/".join(dict_path.split("/" )[:-1] )} ) else: A__ , A__ , A__ = fairseq.checkpoint_utils.load_model_ensemble_and_task([checkpoint_path] ) if config_path is not None: A__ = SEWConfig.from_pretrained(__lowerCAmelCase ) else: A__ = convert_config(model[0] , __lowerCAmelCase ) A__ = model[0].eval() A__ = True if config.feat_extract_norm == "layer" else False A__ = WavaVecaFeatureExtractor( feature_size=1 , sampling_rate=1_6_0_0_0 , padding_value=0 , do_normalize=__lowerCAmelCase , return_attention_mask=__lowerCAmelCase , ) if is_finetuned: if dict_path: A__ = Dictionary.load(__lowerCAmelCase ) # important change bos & pad token id since CTC symbol is <pad> and # not <s> as in fairseq A__ = target_dict.pad_index A__ = target_dict.bos_index A__ = target_dict.pad_index A__ = target_dict.bos_index A__ = target_dict.eos_index A__ = len(target_dict.symbols ) A__ = os.path.join(__lowerCAmelCase , "vocab.json" ) if not os.path.isdir(__lowerCAmelCase ): logger.error("--pytorch_dump_folder_path ({}) should be a directory".format(__lowerCAmelCase ) ) return os.makedirs(__lowerCAmelCase , exist_ok=__lowerCAmelCase ) with open(__lowerCAmelCase , "w" , encoding="utf-8" ) as vocab_handle: json.dump(target_dict.indices , __lowerCAmelCase ) A__ = WavaVecaCTCTokenizer( __lowerCAmelCase , unk_token=target_dict.unk_word , pad_token=target_dict.pad_word , bos_token=target_dict.bos_word , eos_token=target_dict.eos_word , word_delimiter_token="|" , do_lower_case=__lowerCAmelCase , ) A__ = WavaVecaProcessor(feature_extractor=__lowerCAmelCase , tokenizer=__lowerCAmelCase ) processor.save_pretrained(__lowerCAmelCase ) A__ = SEWForCTC(__lowerCAmelCase ) else: A__ = SEWModel(__lowerCAmelCase ) feature_extractor.save_pretrained(__lowerCAmelCase ) recursively_load_weights(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) hf_model.save_pretrained(__lowerCAmelCase ) if __name__ == "__main__": lowerCAmelCase__ = argparse.ArgumentParser() parser.add_argument("""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model.""") parser.add_argument("""--checkpoint_path""", default=None, type=str, help="""Path to fairseq checkpoint""") parser.add_argument("""--dict_path""", default=None, type=str, help="""Path to dict of fine-tuned model""") parser.add_argument("""--config_path""", default=None, type=str, help="""Path to hf config.json of model to convert""") parser.add_argument( """--is_finetuned""", action="""store_true""", help="""Whether the model to convert is a fine-tuned model or not""" ) lowerCAmelCase__ = parser.parse_args() convert_sew_checkpoint( args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, args.is_finetuned )
68
import logging import os from dataclasses import dataclass from enum import Enum from typing import List, Optional, Union from filelock import FileLock from transformers import PreTrainedTokenizer, is_tf_available, is_torch_available SCREAMING_SNAKE_CASE__ : Optional[int] = logging.getLogger(__name__) @dataclass class lowerCAmelCase__ : a__ : str a__ : List[str] a__ : Optional[List[str]] @dataclass class lowerCAmelCase__ : a__ : List[int] a__ : List[int] a__ : Optional[List[int]] = None a__ : Optional[List[int]] = None class lowerCAmelCase__ ( __lowercase ): a__ : Optional[Any] = """train""" a__ : Optional[int] = """dev""" a__ : Dict = """test""" class lowerCAmelCase__ : @staticmethod def __A ( SCREAMING_SNAKE_CASE__ : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Union[Split, str] ) -> List[InputExample]: raise NotImplementedError @staticmethod def __A ( SCREAMING_SNAKE_CASE__ : str ) -> List[str]: raise NotImplementedError @staticmethod def __A ( SCREAMING_SNAKE_CASE__ : List[InputExample] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : PreTrainedTokenizer , SCREAMING_SNAKE_CASE__ : Optional[Any]=False , SCREAMING_SNAKE_CASE__ : List[str]="[CLS]" , SCREAMING_SNAKE_CASE__ : Tuple=1 , SCREAMING_SNAKE_CASE__ : str="[SEP]" , SCREAMING_SNAKE_CASE__ : List[Any]=False , SCREAMING_SNAKE_CASE__ : Union[str, Any]=False , SCREAMING_SNAKE_CASE__ : Tuple=0 , SCREAMING_SNAKE_CASE__ : int=0 , SCREAMING_SNAKE_CASE__ : str=-1_00 , SCREAMING_SNAKE_CASE__ : Union[str, Any]=0 , SCREAMING_SNAKE_CASE__ : List[Any]=True , ) -> List[InputFeatures]: __lowerCamelCase = {label: i for i, label in enumerate(SCREAMING_SNAKE_CASE__ )} __lowerCamelCase = [] for ex_index, example in enumerate(SCREAMING_SNAKE_CASE__ ): if ex_index % 1_00_00 == 0: logger.info('''Writing example %d of %d''' , SCREAMING_SNAKE_CASE__ , len(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = [] __lowerCamelCase = [] for word, label in zip(example.words , example.labels ): __lowerCamelCase = tokenizer.tokenize(SCREAMING_SNAKE_CASE__ ) # bert-base-multilingual-cased sometimes output "nothing ([]) when calling tokenize with just a space. if len(SCREAMING_SNAKE_CASE__ ) > 0: tokens.extend(SCREAMING_SNAKE_CASE__ ) # Use the real label id for the first token of the word, and padding ids for the remaining tokens label_ids.extend([label_map[label]] + [pad_token_label_id] * (len(SCREAMING_SNAKE_CASE__ ) - 1) ) # Account for [CLS] and [SEP] with "- 2" and with "- 3" for RoBERTa. __lowerCamelCase = tokenizer.num_special_tokens_to_add() if len(SCREAMING_SNAKE_CASE__ ) > max_seq_length - special_tokens_count: __lowerCamelCase = tokens[: (max_seq_length - special_tokens_count)] __lowerCamelCase = label_ids[: (max_seq_length - special_tokens_count)] # The convention in BERT is: # (a) For sequence pairs: # tokens: [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP] # type_ids: 0 0 0 0 0 0 0 0 1 1 1 1 1 1 # (b) For single sequences: # tokens: [CLS] the dog is hairy . [SEP] # type_ids: 0 0 0 0 0 0 0 # # Where "type_ids" are used to indicate whether this is the first # sequence or the second sequence. The embedding vectors for `type=0` and # `type=1` were learned during pre-training and are added to the wordpiece # embedding vector (and position vector). This is not *strictly* necessary # since the [SEP] token unambiguously separates the sequences, but it makes # it easier for the model to learn the concept of sequences. # # For classification tasks, the first vector (corresponding to [CLS]) is # used as the "sentence vector". Note that this only makes sense because # the entire model is fine-tuned. tokens += [sep_token] label_ids += [pad_token_label_id] if sep_token_extra: # roberta uses an extra separator b/w pairs of sentences tokens += [sep_token] label_ids += [pad_token_label_id] __lowerCamelCase = [sequence_a_segment_id] * len(SCREAMING_SNAKE_CASE__ ) if cls_token_at_end: tokens += [cls_token] label_ids += [pad_token_label_id] segment_ids += [cls_token_segment_id] else: __lowerCamelCase = [cls_token] + tokens __lowerCamelCase = [pad_token_label_id] + label_ids __lowerCamelCase = [cls_token_segment_id] + segment_ids __lowerCamelCase = tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) # The mask has 1 for real tokens and 0 for padding tokens. Only real # tokens are attended to. __lowerCamelCase = [1 if mask_padding_with_zero else 0] * len(SCREAMING_SNAKE_CASE__ ) # Zero-pad up to the sequence length. __lowerCamelCase = max_seq_length - len(SCREAMING_SNAKE_CASE__ ) if pad_on_left: __lowerCamelCase = ([pad_token] * padding_length) + input_ids __lowerCamelCase = ([0 if mask_padding_with_zero else 1] * padding_length) + input_mask __lowerCamelCase = ([pad_token_segment_id] * padding_length) + segment_ids __lowerCamelCase = ([pad_token_label_id] * padding_length) + label_ids else: input_ids += [pad_token] * padding_length input_mask += [0 if mask_padding_with_zero else 1] * padding_length segment_ids += [pad_token_segment_id] * padding_length label_ids += [pad_token_label_id] * padding_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length assert len(SCREAMING_SNAKE_CASE__ ) == max_seq_length if ex_index < 5: logger.info('''*** Example ***''' ) logger.info('''guid: %s''' , example.guid ) logger.info('''tokens: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in tokens] ) ) logger.info('''input_ids: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in input_ids] ) ) logger.info('''input_mask: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in input_mask] ) ) logger.info('''segment_ids: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in segment_ids] ) ) logger.info('''label_ids: %s''' , ''' '''.join([str(SCREAMING_SNAKE_CASE__ ) for x in label_ids] ) ) if "token_type_ids" not in tokenizer.model_input_names: __lowerCamelCase = None features.append( InputFeatures( input_ids=SCREAMING_SNAKE_CASE__ , attention_mask=SCREAMING_SNAKE_CASE__ , token_type_ids=SCREAMING_SNAKE_CASE__ , label_ids=SCREAMING_SNAKE_CASE__ ) ) return features if is_torch_available(): import torch from torch import nn from torch.utils.data import Dataset class lowerCAmelCase__ ( __lowercase ): a__ : List[InputFeatures] a__ : int = nn.CrossEntropyLoss().ignore_index def __init__( self : List[str] , SCREAMING_SNAKE_CASE__ : TokenClassificationTask , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : PreTrainedTokenizer , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : Optional[int] = None , SCREAMING_SNAKE_CASE__ : Union[str, Any]=False , SCREAMING_SNAKE_CASE__ : Split = Split.train , ) -> Union[str, Any]: # Load data features from cache or dataset file __lowerCamelCase = os.path.join( SCREAMING_SNAKE_CASE__ , '''cached_{}_{}_{}'''.format(mode.value , tokenizer.__class__.__name__ , str(SCREAMING_SNAKE_CASE__ ) ) , ) # Make sure only the first process in distributed training processes the dataset, # and the others will use the cache. __lowerCamelCase = cached_features_file + '''.lock''' with FileLock(SCREAMING_SNAKE_CASE__ ): if os.path.exists(SCREAMING_SNAKE_CASE__ ) and not overwrite_cache: logger.info(f'''Loading features from cached file {cached_features_file}''' ) __lowerCamelCase = torch.load(SCREAMING_SNAKE_CASE__ ) else: logger.info(f'''Creating features from dataset file at {data_dir}''' ) __lowerCamelCase = token_classification_task.read_examples_from_file(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # TODO clean up all this to leverage built-in features of tokenizers __lowerCamelCase = token_classification_task.convert_examples_to_features( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , cls_token_at_end=bool(model_type in ['''xlnet'''] ) , cls_token=tokenizer.cls_token , cls_token_segment_id=2 if model_type in ['''xlnet'''] else 0 , sep_token=tokenizer.sep_token , sep_token_extra=SCREAMING_SNAKE_CASE__ , pad_on_left=bool(tokenizer.padding_side == '''left''' ) , pad_token=tokenizer.pad_token_id , pad_token_segment_id=tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , ) logger.info(f'''Saving features into cached file {cached_features_file}''' ) torch.save(self.features , SCREAMING_SNAKE_CASE__ ) def __len__( self : Dict ) -> str: return len(self.features ) def __getitem__( self : Any , SCREAMING_SNAKE_CASE__ : Dict ) -> InputFeatures: return self.features[i] if is_tf_available(): import tensorflow as tf class lowerCAmelCase__ : a__ : List[InputFeatures] a__ : int = -100 def __init__( self : Optional[Any] , SCREAMING_SNAKE_CASE__ : TokenClassificationTask , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : PreTrainedTokenizer , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str , SCREAMING_SNAKE_CASE__ : Optional[int] = None , SCREAMING_SNAKE_CASE__ : Optional[Any]=False , SCREAMING_SNAKE_CASE__ : Split = Split.train , ) -> List[Any]: __lowerCamelCase = token_classification_task.read_examples_from_file(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # TODO clean up all this to leverage built-in features of tokenizers __lowerCamelCase = token_classification_task.convert_examples_to_features( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , cls_token_at_end=bool(model_type in ['''xlnet'''] ) , cls_token=tokenizer.cls_token , cls_token_segment_id=2 if model_type in ['''xlnet'''] else 0 , sep_token=tokenizer.sep_token , sep_token_extra=SCREAMING_SNAKE_CASE__ , pad_on_left=bool(tokenizer.padding_side == '''left''' ) , pad_token=tokenizer.pad_token_id , pad_token_segment_id=tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , ) def gen(): for ex in self.features: if ex.token_type_ids is None: yield ( {"input_ids": ex.input_ids, "attention_mask": ex.attention_mask}, ex.label_ids, ) else: yield ( { "input_ids": ex.input_ids, "attention_mask": ex.attention_mask, "token_type_ids": ex.token_type_ids, }, ex.label_ids, ) if "token_type_ids" not in tokenizer.model_input_names: __lowerCamelCase = tf.data.Dataset.from_generator( SCREAMING_SNAKE_CASE__ , ({'''input_ids''': tf.intaa, '''attention_mask''': tf.intaa}, tf.intaa) , ( {'''input_ids''': tf.TensorShape([None] ), '''attention_mask''': tf.TensorShape([None] )}, tf.TensorShape([None] ), ) , ) else: __lowerCamelCase = tf.data.Dataset.from_generator( SCREAMING_SNAKE_CASE__ , ({'''input_ids''': tf.intaa, '''attention_mask''': tf.intaa, '''token_type_ids''': tf.intaa}, tf.intaa) , ( { '''input_ids''': tf.TensorShape([None] ), '''attention_mask''': tf.TensorShape([None] ), '''token_type_ids''': tf.TensorShape([None] ), }, tf.TensorShape([None] ), ) , ) def __A ( self : Union[str, Any] ) -> Union[str, Any]: __lowerCamelCase = self.dataset.apply(tf.data.experimental.assert_cardinality(len(self.features ) ) ) return self.dataset def __len__( self : List[Any] ) -> Any: return len(self.features ) def __getitem__( self : List[str] , SCREAMING_SNAKE_CASE__ : Optional[int] ) -> InputFeatures: return self.features[i]
270
0
"""simple docstring""" import inspect import unittest from transformers import MobileViTVaConfig from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device from transformers.utils import cached_property, is_torch_available, is_vision_available 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 MobileViTVaForImageClassification, MobileViTVaForSemanticSegmentation, MobileViTVaModel from transformers.models.mobilevitva.modeling_mobilevitva import ( MOBILEVITV2_PRETRAINED_MODEL_ARCHIVE_LIST, make_divisible, ) if is_vision_available(): from PIL import Image from transformers import MobileViTImageProcessor class __lowerCamelCase ( __lowercase ): '''simple docstring''' def _UpperCAmelCase ( self ) -> Union[str, Any]: _a = self.config_class(**self.inputs_dict ) self.parent.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''width_multiplier''' ) ) class __lowerCamelCase : '''simple docstring''' def __init__( self , __UpperCAmelCase , __UpperCAmelCase=13 , __UpperCAmelCase=64 , __UpperCAmelCase=2 , __UpperCAmelCase=3 , __UpperCAmelCase="swish" , __UpperCAmelCase=3 , __UpperCAmelCase=32 , __UpperCAmelCase=0.1 , __UpperCAmelCase=0.02 , __UpperCAmelCase=True , __UpperCAmelCase=True , __UpperCAmelCase=10 , __UpperCAmelCase=None , __UpperCAmelCase=0.25 , __UpperCAmelCase=0.0 , __UpperCAmelCase=0.0 , ) -> Optional[Any]: _a = parent _a = batch_size _a = image_size _a = patch_size _a = num_channels _a = make_divisible(512 * width_multiplier , divisor=8 ) _a = hidden_act _a = conv_kernel_size _a = output_stride _a = classifier_dropout_prob _a = use_labels _a = is_training _a = num_labels _a = initializer_range _a = scope _a = width_multiplier _a = ffn_dropout _a = attn_dropout def _UpperCAmelCase ( self ) -> Optional[Any]: _a = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] ) _a = None _a = None if self.use_labels: _a = ids_tensor([self.batch_size] , self.num_labels ) _a = ids_tensor([self.batch_size, self.image_size, self.image_size] , self.num_labels ) _a = self.get_config() return config, pixel_values, labels, pixel_labels def _UpperCAmelCase ( self ) -> int: return MobileViTVaConfig( image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_act=self.hidden_act , conv_kernel_size=self.conv_kernel_size , output_stride=self.output_stride , classifier_dropout_prob=self.classifier_dropout_prob , initializer_range=self.initializer_range , width_multiplier=self.width_multiplier , ffn_dropout=self.ffn_dropout_prob , attn_dropout=self.attn_dropout_prob , ) def _UpperCAmelCase ( self , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Any: _a = MobileViTVaModel(config=SCREAMING_SNAKE_CASE__ ) model.to(SCREAMING_SNAKE_CASE__ ) model.eval() _a = model(SCREAMING_SNAKE_CASE__ ) self.parent.assertEqual( result.last_hidden_state.shape , ( self.batch_size, self.last_hidden_size, self.image_size // self.output_stride, self.image_size // self.output_stride, ) , ) def _UpperCAmelCase ( self , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> List[Any]: _a = self.num_labels _a = MobileViTVaForImageClassification(SCREAMING_SNAKE_CASE__ ) model.to(SCREAMING_SNAKE_CASE__ ) model.eval() _a = model(SCREAMING_SNAKE_CASE__ , labels=SCREAMING_SNAKE_CASE__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) ) def _UpperCAmelCase ( self , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ) -> Any: _a = self.num_labels _a = MobileViTVaForSemanticSegmentation(SCREAMING_SNAKE_CASE__ ) model.to(SCREAMING_SNAKE_CASE__ ) model.eval() _a = model(SCREAMING_SNAKE_CASE__ ) self.parent.assertEqual( result.logits.shape , ( self.batch_size, self.num_labels, self.image_size // self.output_stride, self.image_size // self.output_stride, ) , ) _a = model(SCREAMING_SNAKE_CASE__ , labels=SCREAMING_SNAKE_CASE__ ) self.parent.assertEqual( result.logits.shape , ( self.batch_size, self.num_labels, self.image_size // self.output_stride, self.image_size // self.output_stride, ) , ) def _UpperCAmelCase ( self ) -> Dict: _a = self.prepare_config_and_inputs() _a , _a , _a , _a = config_and_inputs _a = {'''pixel_values''': pixel_values} return config, inputs_dict @require_torch class __lowerCamelCase ( __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' A_ : Optional[int] = ( (MobileViTVaModel, MobileViTVaForImageClassification, MobileViTVaForSemanticSegmentation) if is_torch_available() else () ) A_ : Any = ( { """feature-extraction""": MobileViTVaModel, """image-classification""": MobileViTVaForImageClassification, """image-segmentation""": MobileViTVaForSemanticSegmentation, } if is_torch_available() else {} ) A_ : List[str] = False A_ : Optional[Any] = False A_ : str = False A_ : Optional[Any] = False def _UpperCAmelCase ( self ) -> Optional[Any]: _a = MobileViTVaModelTester(self ) _a = MobileViTVaConfigTester(self , config_class=SCREAMING_SNAKE_CASE__ , has_text_modality=SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Tuple: self.config_tester.run_common_tests() @unittest.skip(reason='''MobileViTV2 does not use inputs_embeds''' ) def _UpperCAmelCase ( self ) -> str: pass @unittest.skip(reason='''MobileViTV2 does not support input and output embeddings''' ) def _UpperCAmelCase ( self ) -> Optional[Any]: pass @unittest.skip(reason='''MobileViTV2 does not output attentions''' ) def _UpperCAmelCase ( self ) -> Tuple: pass @require_torch_multi_gpu @unittest.skip(reason='''Got `CUDA error: misaligned address` for tests after this one being run.''' ) def _UpperCAmelCase ( self ) -> List[Any]: pass @unittest.skip('''Will be fixed soon by reducing the size of the model used for common tests.''' ) def _UpperCAmelCase ( self ) -> Dict: pass def _UpperCAmelCase ( self ) -> Any: _a , _a = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a = model_class(SCREAMING_SNAKE_CASE__ ) _a = inspect.signature(model.forward ) # signature.parameters is an OrderedDict => so arg_names order is deterministic _a = [*signature.parameters.keys()] _a = ['''pixel_values'''] self.assertListEqual(arg_names[:1] , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Any: _a = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_model(*SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Optional[Any]: def check_hidden_states_output(__UpperCAmelCase , __UpperCAmelCase , __UpperCAmelCase ): _a = model_class(SCREAMING_SNAKE_CASE__ ) model.to(SCREAMING_SNAKE_CASE__ ) model.eval() with torch.no_grad(): _a = model(**self._prepare_for_class(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) ) _a = outputs.hidden_states _a = 5 self.assertEqual(len(SCREAMING_SNAKE_CASE__ ) , SCREAMING_SNAKE_CASE__ ) # MobileViTV2's feature maps are of shape (batch_size, num_channels, height, width) # with the width and height being successively divided by 2. _a = 2 for i in range(len(SCREAMING_SNAKE_CASE__ ) ): self.assertListEqual( list(hidden_states[i].shape[-2:] ) , [self.model_tester.image_size // divisor, self.model_tester.image_size // divisor] , ) divisor *= 2 self.assertEqual(self.model_tester.output_stride , divisor // 2 ) _a , _a = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a = True check_hidden_states_output(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) # check that output_hidden_states also work using config del inputs_dict["output_hidden_states"] _a = True check_hidden_states_output(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Optional[Any]: _a = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_image_classification(*SCREAMING_SNAKE_CASE__ ) def _UpperCAmelCase ( self ) -> Tuple: _a = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_semantic_segmentation(*SCREAMING_SNAKE_CASE__ ) @slow def _UpperCAmelCase ( self ) -> Any: for model_name in MOBILEVITV2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]: _a = MobileViTVaModel.from_pretrained(SCREAMING_SNAKE_CASE__ ) self.assertIsNotNone(SCREAMING_SNAKE_CASE__ ) def A_ ( ): """simple docstring""" _a = Image.open('''./tests/fixtures/tests_samples/COCO/000000039769.png''' ) return image @require_torch @require_vision class __lowerCamelCase ( unittest.TestCase ): '''simple docstring''' @cached_property def _UpperCAmelCase ( self ) -> Dict: return ( MobileViTImageProcessor.from_pretrained('''apple/mobilevitv2-1.0-imagenet1k-256''' ) if is_vision_available() else None ) @slow def _UpperCAmelCase ( self ) -> Union[str, Any]: _a = MobileViTVaForImageClassification.from_pretrained('''apple/mobilevitv2-1.0-imagenet1k-256''' ).to( SCREAMING_SNAKE_CASE__ ) _a = self.default_image_processor _a = prepare_img() _a = image_processor(images=SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' ).to(SCREAMING_SNAKE_CASE__ ) # forward pass with torch.no_grad(): _a = model(**SCREAMING_SNAKE_CASE__ ) # verify the logits _a = torch.Size((1, 1000) ) self.assertEqual(outputs.logits.shape , SCREAMING_SNAKE_CASE__ ) _a = torch.tensor([-1.6_3_3_6e0_0, -7.3_2_0_4e-0_2, -5.1_8_8_3e-0_1] ).to(SCREAMING_SNAKE_CASE__ ) self.assertTrue(torch.allclose(outputs.logits[0, :3] , SCREAMING_SNAKE_CASE__ , atol=1e-4 ) ) @slow def _UpperCAmelCase ( self ) -> Any: _a = MobileViTVaForSemanticSegmentation.from_pretrained('''shehan97/mobilevitv2-1.0-voc-deeplabv3''' ) _a = model.to(SCREAMING_SNAKE_CASE__ ) _a = MobileViTImageProcessor.from_pretrained('''shehan97/mobilevitv2-1.0-voc-deeplabv3''' ) _a = prepare_img() _a = image_processor(images=SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' ).to(SCREAMING_SNAKE_CASE__ ) # forward pass with torch.no_grad(): _a = model(**SCREAMING_SNAKE_CASE__ ) _a = outputs.logits # verify the logits _a = torch.Size((1, 21, 32, 32) ) self.assertEqual(logits.shape , SCREAMING_SNAKE_CASE__ ) _a = torch.tensor( [ [[7.0863, 7.1525, 6.8201], [6.6931, 6.8770, 6.8933], [6.2978, 7.0366, 6.9636]], [[-3.7134, -3.6712, -3.6675], [-3.5825, -3.3549, -3.4777], [-3.3435, -3.3979, -3.2857]], [[-2.9329, -2.8003, -2.7369], [-3.0564, -2.4780, -2.0207], [-2.6889, -1.9298, -1.7640]], ] , device=SCREAMING_SNAKE_CASE__ , ) self.assertTrue(torch.allclose(logits[0, :3, :3, :3] , SCREAMING_SNAKE_CASE__ , atol=1e-4 ) ) @slow def _UpperCAmelCase ( self ) -> Optional[int]: _a = MobileViTVaForSemanticSegmentation.from_pretrained('''shehan97/mobilevitv2-1.0-voc-deeplabv3''' ) _a = model.to(SCREAMING_SNAKE_CASE__ ) _a = MobileViTImageProcessor.from_pretrained('''shehan97/mobilevitv2-1.0-voc-deeplabv3''' ) _a = prepare_img() _a = image_processor(images=SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' ).to(SCREAMING_SNAKE_CASE__ ) # forward pass with torch.no_grad(): _a = model(**SCREAMING_SNAKE_CASE__ ) _a = outputs.logits.detach().cpu() _a = image_processor.post_process_semantic_segmentation(outputs=SCREAMING_SNAKE_CASE__ , target_sizes=[(50, 60)] ) _a = torch.Size((50, 60) ) self.assertEqual(segmentation[0].shape , SCREAMING_SNAKE_CASE__ ) _a = image_processor.post_process_semantic_segmentation(outputs=SCREAMING_SNAKE_CASE__ ) _a = torch.Size((32, 32) ) self.assertEqual(segmentation[0].shape , SCREAMING_SNAKE_CASE__ )
320
from typing import Optional from .. import Features, NamedSplit from ..packaged_modules.text.text import Text from ..utils.typing import NestedDataStructureLike, PathLike from .abc import AbstractDatasetReader class lowerCAmelCase__ ( __lowercase ): def __init__( self : Tuple , SCREAMING_SNAKE_CASE__ : NestedDataStructureLike[PathLike] , SCREAMING_SNAKE_CASE__ : Optional[NamedSplit] = None , SCREAMING_SNAKE_CASE__ : Optional[Features] = None , SCREAMING_SNAKE_CASE__ : str = None , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : Optional[int] = None , **SCREAMING_SNAKE_CASE__ : str , ) -> Union[str, Any]: super().__init__( SCREAMING_SNAKE_CASE__ , split=SCREAMING_SNAKE_CASE__ , features=SCREAMING_SNAKE_CASE__ , cache_dir=SCREAMING_SNAKE_CASE__ , keep_in_memory=SCREAMING_SNAKE_CASE__ , streaming=SCREAMING_SNAKE_CASE__ , num_proc=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ , ) __lowerCamelCase = path_or_paths if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) else {self.split: path_or_paths} __lowerCamelCase = Text( cache_dir=SCREAMING_SNAKE_CASE__ , data_files=SCREAMING_SNAKE_CASE__ , features=SCREAMING_SNAKE_CASE__ , **SCREAMING_SNAKE_CASE__ , ) def __A ( self : int ) -> Dict: # Build iterable dataset if self.streaming: __lowerCamelCase = self.builder.as_streaming_dataset(split=self.split ) # Build regular (map-style) dataset else: __lowerCamelCase = None __lowerCamelCase = None __lowerCamelCase = None __lowerCamelCase = None self.builder.download_and_prepare( download_config=SCREAMING_SNAKE_CASE__ , download_mode=SCREAMING_SNAKE_CASE__ , verification_mode=SCREAMING_SNAKE_CASE__ , base_path=SCREAMING_SNAKE_CASE__ , num_proc=self.num_proc , ) __lowerCamelCase = self.builder.as_dataset( split=self.split , verification_mode=SCREAMING_SNAKE_CASE__ , in_memory=self.keep_in_memory ) return dataset
270
0
def UpperCamelCase ( __magic_name__ : Optional[Any] ) -> Optional[int]: """simple docstring""" lowercase__ = 1 lowercase__ = 2 while i * i <= n: lowercase__ = 0 while n % i == 0: n //= i multiplicity += 1 n_divisors *= multiplicity + 1 i += 1 if n > 1: n_divisors *= 2 return n_divisors def UpperCamelCase ( ) -> Optional[Any]: """simple docstring""" lowercase__ = 1 lowercase__ = 1 while True: i += 1 t_num += i if count_divisors(__lowerCAmelCase ) > 500: break return t_num if __name__ == "__main__": print(solution())
305
from ...utils import ( OptionalDependencyNotAvailable, is_torch_available, is_transformers_available, is_transformers_version, ) try: if not (is_transformers_available() and is_torch_available() and is_transformers_version(">=", "4.25.0")): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: from ...utils.dummy_torch_and_transformers_objects import UnCLIPImageVariationPipeline, UnCLIPPipeline else: from .pipeline_unclip import UnCLIPPipeline from .pipeline_unclip_image_variation import UnCLIPImageVariationPipeline from .text_proj import UnCLIPTextProjModel
270
0
'''simple docstring''' UpperCAmelCase_ = "\n# Installazione di Transformers\n! pip install transformers datasets\n# Per installare dalla fonte invece dell'ultima versione rilasciata, commenta il comando sopra e\n# rimuovi la modalità commento al comando seguente.\n# ! pip install git+https://github.com/huggingface/transformers.git\n" UpperCAmelCase_ = [{"type": "code", "content": INSTALL_CONTENT}] UpperCAmelCase_ = { "{processor_class}": "FakeProcessorClass", "{model_class}": "FakeModelClass", "{object_class}": "FakeObjectClass", }
346
import sacrebleu as scb from packaging import version from sacrebleu import CHRF import datasets SCREAMING_SNAKE_CASE__ : Union[str, Any] = "\\n@inproceedings{popovic-2015-chrf,\n title = \"chr{F}: character n-gram {F}-score for automatic {MT} evaluation\",\n author = \"Popovi{\'c}, Maja\",\n booktitle = \"Proceedings of the Tenth Workshop on Statistical Machine Translation\",\n month = sep,\n year = \"2015\",\n address = \"Lisbon, Portugal\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/W15-3049\",\n doi = \"10.18653/v1/W15-3049\",\n pages = \"392--395\",\n}\n@inproceedings{popovic-2017-chrf,\n title = \"chr{F}++: words helping character n-grams\",\n author = \"Popovi{\'c}, Maja\",\n booktitle = \"Proceedings of the Second Conference on Machine Translation\",\n month = sep,\n year = \"2017\",\n address = \"Copenhagen, Denmark\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/W17-4770\",\n doi = \"10.18653/v1/W17-4770\",\n pages = \"612--618\",\n}\n@inproceedings{post-2018-call,\n title = \"A Call for Clarity in Reporting {BLEU} Scores\",\n author = \"Post, Matt\",\n booktitle = \"Proceedings of the Third Conference on Machine Translation: Research Papers\",\n month = oct,\n year = \"2018\",\n address = \"Belgium, Brussels\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W18-6319\",\n pages = \"186--191\",\n}\n" SCREAMING_SNAKE_CASE__ : int = "\\nChrF and ChrF++ are two MT evaluation metrics. They both use the F-score statistic for character n-gram matches,\nand ChrF++ adds word n-grams as well which correlates more strongly with direct assessment. We use the implementation\nthat is already present in sacrebleu.\n\nThe implementation here is slightly different from sacrebleu in terms of the required input format. The length of\nthe references and hypotheses lists need to be the same, so you may need to transpose your references compared to\nsacrebleu's required input format. See https://github.com/huggingface/datasets/issues/3154#issuecomment-950746534\n\nSee the README.md file at https://github.com/mjpost/sacreBLEU#chrf--chrf for more information.\n" SCREAMING_SNAKE_CASE__ : Optional[Any] = "\nProduces ChrF(++) scores for hypotheses given reference translations.\n\nArgs:\n predictions (list of str): The predicted sentences.\n references (list of list of str): The references. There should be one reference sub-list for each prediction sentence.\n char_order (int): Character n-gram order. Defaults to `6`.\n word_order (int): Word n-gram order. If equals to `2`, the metric is referred to as chrF++. Defaults to `0`.\n beta (int): Determine the importance of recall w.r.t precision. Defaults to `2`.\n lowercase (bool): if `True`, enables case-insensitivity. Defaults to `False`.\n whitespace (bool): If `True`, include whitespaces when extracting character n-grams.\n eps_smoothing (bool): If `True`, applies epsilon smoothing similar\n to reference chrF++.py, NLTK and Moses implementations. If `False`,\n it takes into account effective match order similar to sacreBLEU < 2.0.0. Defaults to `False`.\n\nReturns:\n 'score' (float): The chrF (chrF++) score,\n 'char_order' (int): The character n-gram order,\n 'word_order' (int): The word n-gram order. If equals to 2, the metric is referred to as chrF++,\n 'beta' (int): Determine the importance of recall w.r.t precision\n\nExamples:\n Example 1--a simple example of calculating chrF:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction, references=reference)\n >>> print(results)\n {'score': 84.64214891738334, 'char_order': 6, 'word_order': 0, 'beta': 2}\n\n Example 2--the same example, but with the argument word_order=2, to calculate chrF++ instead of chrF:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction,\n ... references=reference,\n ... word_order=2)\n >>> print(results)\n {'score': 82.87263732906315, 'char_order': 6, 'word_order': 2, 'beta': 2}\n\n Example 3--the same chrF++ example as above, but with `lowercase=True` to normalize all case:\n >>> prediction = [\"The relationship between cats and dogs is not exactly friendly.\", \"a good bookshop is just a genteel black hole that knows how to read.\"]\n >>> reference = [[\"The relationship between dogs and cats is not exactly friendly.\"], [\"A good bookshop is just a genteel Black Hole that knows how to read.\"]]\n >>> chrf = datasets.load_metric(\"chrf\")\n >>> results = chrf.compute(predictions=prediction,\n ... references=reference,\n ... word_order=2,\n ... lowercase=True)\n >>> print(results)\n {'score': 92.12853119829202, 'char_order': 6, 'word_order': 2, 'beta': 2}\n" @datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION ) class lowerCAmelCase__ ( datasets.Metric ): def __A ( self : Dict ) -> Tuple: if version.parse(scb.__version__ ) < version.parse('''1.4.12''' ): raise ImportWarning( '''To use `sacrebleu`, the module `sacrebleu>=1.4.12` is required, and the current version of `sacrebleu` doesn\'t match this condition.\n''' '''You can install it with `pip install "sacrebleu>=1.4.12"`.''' ) return datasets.MetricInfo( description=_DESCRIPTION , citation=_CITATION , homepage='''https://github.com/mjpost/sacreBLEU#chrf--chrf''' , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features( { '''predictions''': datasets.Value('''string''' , id='''sequence''' ), '''references''': datasets.Sequence(datasets.Value('''string''' , id='''sequence''' ) , id='''references''' ), } ) , codebase_urls=['''https://github.com/mjpost/sacreBLEU#chrf--chrf'''] , reference_urls=[ '''https://github.com/m-popovic/chrF''', ] , ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : int = CHRF.CHAR_ORDER , SCREAMING_SNAKE_CASE__ : int = CHRF.WORD_ORDER , SCREAMING_SNAKE_CASE__ : int = CHRF.BETA , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , SCREAMING_SNAKE_CASE__ : bool = False , ) -> Optional[Any]: __lowerCamelCase = len(references[0] ) if any(len(SCREAMING_SNAKE_CASE__ ) != references_per_prediction for refs in references ): raise ValueError('''Sacrebleu requires the same number of references for each prediction''' ) __lowerCamelCase = [[refs[i] for refs in references] for i in range(SCREAMING_SNAKE_CASE__ )] __lowerCamelCase = CHRF(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = sb_chrf.corpus_score(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) return { "score": output.score, "char_order": output.char_order, "word_order": output.word_order, "beta": output.beta, }
270
0
import argparse import torch # Step 1. clone https://github.com/microsoft/unilm # Step 2. git checkout to https://github.com/microsoft/unilm/commit/b94ec76c36f02fb2b0bf0dcb0b8554a2185173cd # Step 3. cd unilm # Step 4. ln -s $(realpath wavlm/modules.py) ./ # create simlink # import classes from unilm.wavlm.WavLM import WavLM as WavLMOrig from unilm.wavlm.WavLM import WavLMConfig as WavLMConfigOrig from transformers import WavLMConfig, WavLMModel, logging logging.set_verbosity_info() A : Any = logging.get_logger(__name__) A : Tuple = { "post_extract_proj": "feature_projection.projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn.grep_linear": "encoder.layers.*.attention.gru_rel_pos_linear", "self_attn.relative_attention_bias": "encoder.layers.*.attention.rel_attn_embed", "self_attn.grep_a": "encoder.layers.*.attention.gru_rel_pos_const", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.layer_norm": "encoder.layer_norm", "w2v_model.layer_norm": "feature_projection.layer_norm", "quantizer.weight_proj": "quantizer.weight_proj", "quantizer.vars": "quantizer.codevectors", "project_q": "project_q", "final_proj": "project_hid", "w2v_encoder.proj": "ctc_proj", "mask_emb": "masked_spec_embed", } A : Tuple = [ "ctc_proj", "quantizer.weight_proj", "quantizer.codevectors", "project_q", "project_hid", ] def lowercase_ ( _A : Any , _A : Tuple , _A : List[Any] , _A : List[Any] , _A : List[Any] ): """simple docstring""" for attribute in key.split("." ): lowerCamelCase__ : List[Any] = getattr(__lowerCAmelCase , __lowerCAmelCase ) if weight_type is not None: lowerCamelCase__ : str = getattr(__lowerCAmelCase , __lowerCAmelCase ).shape else: lowerCamelCase__ : str = hf_pointer.shape assert hf_shape == value.shape, ( F"Shape of hf {key + '.' + weight_type if weight_type is not None else ''} is {hf_shape}, but should be" F" {value.shape} for {full_name}" ) if weight_type == "weight": lowerCamelCase__ : Tuple = value elif weight_type == "weight_g": lowerCamelCase__ : List[Any] = value elif weight_type == "weight_v": lowerCamelCase__ : int = value elif weight_type == "bias": lowerCamelCase__ : Dict = value else: lowerCamelCase__ : Any = value logger.info(F"{key + '.' + weight_type if weight_type is not None else ''} was initialized from {full_name}." ) def lowercase_ ( _A : Dict , _A : Union[str, Any] ): """simple docstring""" lowerCamelCase__ : Optional[int] = [] lowerCamelCase__ : List[Any] = fairseq_model.state_dict() lowerCamelCase__ : Optional[int] = hf_model.feature_extractor for name, value in fairseq_dict.items(): lowerCamelCase__ : int = False if "conv_layers" in name: load_conv_layer( __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , hf_model.config.feat_extract_norm == "group" , ) lowerCamelCase__ : str = True else: for key, mapped_key in MAPPING.items(): if key in name or key.split("w2v_model." )[-1] == name.split("." )[0]: lowerCamelCase__ : Any = True if "*" in mapped_key: lowerCamelCase__ : Union[str, Any] = name.split(__lowerCAmelCase )[0].split("." )[-2] lowerCamelCase__ : Tuple = mapped_key.replace("*" , __lowerCAmelCase ) if "weight_g" in name: lowerCamelCase__ : Any = "weight_g" elif "weight_v" in name: lowerCamelCase__ : List[Any] = "weight_v" elif "bias" in name and "relative_attention_bias" not in name: lowerCamelCase__ : Union[str, Any] = "bias" elif "weight" in name: # TODO: don't match quantizer.weight_proj lowerCamelCase__ : Dict = "weight" else: lowerCamelCase__ : str = None set_recursively(__lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase ) continue if not is_used: unused_weights.append(__lowerCAmelCase ) logger.warning(F"Unused weights: {unused_weights}" ) def lowercase_ ( _A : Tuple , _A : Any , _A : Optional[Any] , _A : Optional[int] , _A : List[Any] ): """simple docstring""" lowerCamelCase__ : List[str] = full_name.split("conv_layers." )[-1] lowerCamelCase__ : List[Any] = name.split("." ) lowerCamelCase__ : Optional[int] = int(items[0] ) lowerCamelCase__ : Tuple = int(items[1] ) if type_id == 0: if "bias" in name: assert value.shape == feature_extractor.conv_layers[layer_id].conv.bias.data.shape, ( F"{full_name} has size {value.shape}, but" F" {feature_extractor.conv_layers[layer_id].conv.bias.data.shape} was found." ) lowerCamelCase__ : int = value logger.info(F"Feat extract conv layer {layer_id} was initialized from {full_name}." ) elif "weight" in name: assert value.shape == feature_extractor.conv_layers[layer_id].conv.weight.data.shape, ( F"{full_name} has size {value.shape}, but" F" {feature_extractor.conv_layers[layer_id].conv.weight.data.shape} was found." ) lowerCamelCase__ : Tuple = value logger.info(F"Feat extract conv layer {layer_id} was initialized from {full_name}." ) elif (type_id == 2 and not use_group_norm) or (type_id == 2 and layer_id == 0 and use_group_norm): if "bias" in name: assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape, ( F"{full_name} has size {value.shape}, but {feature_extractor[layer_id].layer_norm.bias.data.shape} was" " found." ) lowerCamelCase__ : List[str] = value logger.info(F"Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}." ) elif "weight" in name: assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape, ( F"{full_name} has size {value.shape}, but" F" {feature_extractor[layer_id].layer_norm.weight.data.shape} was found." ) lowerCamelCase__ : Union[str, Any] = value logger.info(F"Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}." ) else: unused_weights.append(__lowerCAmelCase ) @torch.no_grad() def lowercase_ ( _A : Union[str, Any] , _A : Any , _A : List[str]=None ): """simple docstring""" lowerCamelCase__ : Tuple = torch.load(__lowerCAmelCase ) lowerCamelCase__ : List[str] = WavLMConfigOrig(checkpoint["cfg"] ) lowerCamelCase__ : Any = WavLMOrig(__lowerCAmelCase ) model.load_state_dict(checkpoint["model"] ) model.eval() if config_path is not None: lowerCamelCase__ : int = WavLMConfig.from_pretrained(__lowerCAmelCase ) else: lowerCamelCase__ : Union[str, Any] = WavLMConfig() lowerCamelCase__ : Optional[int] = WavLMModel(__lowerCAmelCase ) recursively_load_weights(__lowerCAmelCase , __lowerCAmelCase ) hf_wavlm.save_pretrained(__lowerCAmelCase ) if __name__ == "__main__": A : Optional[int] = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") A : Any = parser.parse_args() convert_wavlm_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path)
184
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_tf_available, is_torch_available, ) SCREAMING_SNAKE_CASE__ : List[Any] = { "configuration_resnet": ["RESNET_PRETRAINED_CONFIG_ARCHIVE_MAP", "ResNetConfig", "ResNetOnnxConfig"] } try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : int = [ "RESNET_PRETRAINED_MODEL_ARCHIVE_LIST", "ResNetForImageClassification", "ResNetModel", "ResNetPreTrainedModel", "ResNetBackbone", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : List[Any] = [ "TF_RESNET_PRETRAINED_MODEL_ARCHIVE_LIST", "TFResNetForImageClassification", "TFResNetModel", "TFResNetPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: SCREAMING_SNAKE_CASE__ : Union[str, Any] = [ "FlaxResNetForImageClassification", "FlaxResNetModel", "FlaxResNetPreTrainedModel", ] if TYPE_CHECKING: from .configuration_resnet import RESNET_PRETRAINED_CONFIG_ARCHIVE_MAP, ResNetConfig, ResNetOnnxConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_resnet import ( RESNET_PRETRAINED_MODEL_ARCHIVE_LIST, ResNetBackbone, ResNetForImageClassification, ResNetModel, ResNetPreTrainedModel, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_resnet import ( TF_RESNET_PRETRAINED_MODEL_ARCHIVE_LIST, TFResNetForImageClassification, TFResNetModel, TFResNetPreTrainedModel, ) try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_resnet import FlaxResNetForImageClassification, FlaxResNetModel, FlaxResNetPreTrainedModel else: import sys SCREAMING_SNAKE_CASE__ : Any = _LazyModule(__name__, globals()["__file__"], _import_structure)
270
0
"""simple docstring""" from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) lowerCAmelCase = { "configuration_rembert": ["REMBERT_PRETRAINED_CONFIG_ARCHIVE_MAP", "RemBertConfig", "RemBertOnnxConfig"] } try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: lowerCAmelCase = ["RemBertTokenizer"] try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: lowerCAmelCase = ["RemBertTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: lowerCAmelCase = [ "REMBERT_PRETRAINED_MODEL_ARCHIVE_LIST", "RemBertForCausalLM", "RemBertForMaskedLM", "RemBertForMultipleChoice", "RemBertForQuestionAnswering", "RemBertForSequenceClassification", "RemBertForTokenClassification", "RemBertLayer", "RemBertModel", "RemBertPreTrainedModel", "load_tf_weights_in_rembert", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: lowerCAmelCase = [ "TF_REMBERT_PRETRAINED_MODEL_ARCHIVE_LIST", "TFRemBertForCausalLM", "TFRemBertForMaskedLM", "TFRemBertForMultipleChoice", "TFRemBertForQuestionAnswering", "TFRemBertForSequenceClassification", "TFRemBertForTokenClassification", "TFRemBertLayer", "TFRemBertModel", "TFRemBertPreTrainedModel", ] if TYPE_CHECKING: from .configuration_rembert import REMBERT_PRETRAINED_CONFIG_ARCHIVE_MAP, RemBertConfig, RemBertOnnxConfig try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_rembert import RemBertTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_rembert_fast import RemBertTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_rembert import ( REMBERT_PRETRAINED_MODEL_ARCHIVE_LIST, RemBertForCausalLM, RemBertForMaskedLM, RemBertForMultipleChoice, RemBertForQuestionAnswering, RemBertForSequenceClassification, RemBertForTokenClassification, RemBertLayer, RemBertModel, RemBertPreTrainedModel, load_tf_weights_in_rembert, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_rembert import ( TF_REMBERT_PRETRAINED_MODEL_ARCHIVE_LIST, TFRemBertForCausalLM, TFRemBertForMaskedLM, TFRemBertForMultipleChoice, TFRemBertForQuestionAnswering, TFRemBertForSequenceClassification, TFRemBertForTokenClassification, TFRemBertLayer, TFRemBertModel, TFRemBertPreTrainedModel, ) else: import sys lowerCAmelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
126
import unittest import numpy as np import requests 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 from transformers.pytorch_utils import is_torch_greater_or_equal_than_1_11 else: SCREAMING_SNAKE_CASE__ : str = False if is_vision_available(): from PIL import Image from transformers import PixaStructImageProcessor class lowerCAmelCase__ ( unittest.TestCase ): def __init__( self : Tuple , SCREAMING_SNAKE_CASE__ : Union[str, Any] , SCREAMING_SNAKE_CASE__ : List[Any]=7 , SCREAMING_SNAKE_CASE__ : List[str]=3 , SCREAMING_SNAKE_CASE__ : Tuple=18 , SCREAMING_SNAKE_CASE__ : int=30 , SCREAMING_SNAKE_CASE__ : Any=4_00 , SCREAMING_SNAKE_CASE__ : Any=None , SCREAMING_SNAKE_CASE__ : str=True , SCREAMING_SNAKE_CASE__ : str=True , SCREAMING_SNAKE_CASE__ : Optional[Any]=None , ) -> Union[str, Any]: __lowerCamelCase = size if size is not None else {'''height''': 20, '''width''': 20} __lowerCamelCase = parent __lowerCamelCase = batch_size __lowerCamelCase = num_channels __lowerCamelCase = image_size __lowerCamelCase = min_resolution __lowerCamelCase = max_resolution __lowerCamelCase = size __lowerCamelCase = do_normalize __lowerCamelCase = do_convert_rgb __lowerCamelCase = [5_12, 10_24, 20_48, 40_96] __lowerCamelCase = patch_size if patch_size is not None else {'''height''': 16, '''width''': 16} def __A ( self : Union[str, Any] ) -> Union[str, Any]: return {"do_normalize": self.do_normalize, "do_convert_rgb": self.do_convert_rgb} def __A ( self : int ) -> Any: __lowerCamelCase = '''https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/australia.jpg''' __lowerCamelCase = Image.open(requests.get(SCREAMING_SNAKE_CASE__ , stream=SCREAMING_SNAKE_CASE__ ).raw ).convert('''RGB''' ) return raw_image @unittest.skipIf( not is_torch_greater_or_equal_than_1_11 , reason="""`Pix2StructImageProcessor` requires `torch>=1.11.0`.""" , ) @require_torch @require_vision class lowerCAmelCase__ ( __lowercase , unittest.TestCase ): a__ : Dict = PixaStructImageProcessor if is_vision_available() else None def __A ( self : Optional[Any] ) -> List[Any]: __lowerCamelCase = PixaStructImageProcessingTester(self ) @property def __A ( self : Optional[int] ) -> Optional[int]: return self.image_processor_tester.prepare_image_processor_dict() def __A ( self : Tuple ) -> Dict: __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_normalize''' ) ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_convert_rgb''' ) ) def __A ( self : int ) -> str: __lowerCamelCase = self.image_processor_tester.prepare_dummy_image() __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) __lowerCamelCase = 20_48 __lowerCamelCase = image_processor(SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ) self.assertTrue(torch.allclose(inputs.flattened_patches.mean() , torch.tensor(0.0606 ) , atol=1e-3 , rtol=1e-3 ) ) def __A ( self : Union[str, Any] ) -> Dict: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PIL images __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , Image.Image ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) def __A ( self : Dict ) -> str: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PIL images __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , Image.Image ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 __lowerCamelCase = True for max_patch in self.image_processor_tester.max_patches: # Test not batched input with self.assertRaises(SCREAMING_SNAKE_CASE__ ): __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches __lowerCamelCase = '''Hello''' __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ , header_text=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ , header_text=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) def __A ( self : List[str] ) -> Any: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random numpy tensors __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ , numpify=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , np.ndarray ) __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) def __A ( self : Tuple ) -> List[str]: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PyTorch tensors __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ , torchify=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , torch.Tensor ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * self.image_processor_tester.num_channels ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , ) @unittest.skipIf( not is_torch_greater_or_equal_than_1_11 , reason="""`Pix2StructImageProcessor` requires `torch>=1.11.0`.""" , ) @require_torch @require_vision class lowerCAmelCase__ ( __lowercase , unittest.TestCase ): a__ : Any = PixaStructImageProcessor if is_vision_available() else None def __A ( self : Union[str, Any] ) -> Optional[Any]: __lowerCamelCase = PixaStructImageProcessingTester(self , num_channels=4 ) __lowerCamelCase = 3 @property def __A ( self : List[Any] ) -> Optional[Any]: return self.image_processor_tester.prepare_image_processor_dict() def __A ( self : Union[str, Any] ) -> List[str]: __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_normalize''' ) ) self.assertTrue(hasattr(SCREAMING_SNAKE_CASE__ , '''do_convert_rgb''' ) ) def __A ( self : Optional[int] ) -> str: # Initialize image_processor __lowerCamelCase = self.image_processing_class(**self.image_processor_dict ) # create random PIL images __lowerCamelCase = prepare_image_inputs(self.image_processor_tester , equal_resolution=SCREAMING_SNAKE_CASE__ ) for image in image_inputs: self.assertIsInstance(SCREAMING_SNAKE_CASE__ , Image.Image ) # Test not batched input __lowerCamelCase = ( (self.image_processor_tester.patch_size['''height'''] * self.image_processor_tester.patch_size['''width''']) * (self.image_processor_tester.num_channels - 1) ) + 2 for max_patch in self.image_processor_tester.max_patches: # Test not batched input __lowerCamelCase = image_processor( image_inputs[0] , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (1, max_patch, expected_hidden_dim) , ) # Test batched __lowerCamelCase = image_processor( SCREAMING_SNAKE_CASE__ , return_tensors='''pt''' , max_patches=SCREAMING_SNAKE_CASE__ ).flattened_patches self.assertEqual( encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
270
0
"""simple docstring""" import math class __SCREAMING_SNAKE_CASE : '''simple docstring''' def __init__( self : Optional[int], lowerCamelCase : Any=0 )-> Any: # a graph with Node 0,1,...,N-1 lowerCamelCase__ : Optional[Any] =n lowerCamelCase__ : Any =[ [math.inf for j in range(0, SCREAMING_SNAKE_CASE__ )] for i in range(0, SCREAMING_SNAKE_CASE__ ) ] # adjacency matrix for weight lowerCamelCase__ : int =[ [math.inf for j in range(0, SCREAMING_SNAKE_CASE__ )] for i in range(0, SCREAMING_SNAKE_CASE__ ) ] # dp[i][j] stores minimum distance from i to j def snake_case ( self : Optional[Any], lowerCamelCase : Any, lowerCamelCase : List[Any], lowerCamelCase : Optional[int] )-> Optional[Any]: lowerCamelCase__ : List[Any] =w def snake_case ( self : Tuple )-> int: for k in range(0, self.n ): for i in range(0, self.n ): for j in range(0, self.n ): lowerCamelCase__ : Union[str, Any] =min(self.dp[i][j], self.dp[i][k] + self.dp[k][j] ) def snake_case ( self : Any, lowerCamelCase : Optional[int], lowerCamelCase : int )-> Dict: return self.dp[u][v] if __name__ == "__main__": _lowercase : Any = Graph(5) graph.add_edge(0, 2, 9) graph.add_edge(0, 4, 1_0) graph.add_edge(1, 3, 5) graph.add_edge(2, 3, 7) graph.add_edge(3, 0, 1_0) graph.add_edge(3, 1, 2) graph.add_edge(3, 2, 1) graph.add_edge(3, 4, 6) graph.add_edge(4, 1, 3) graph.add_edge(4, 2, 4) graph.add_edge(4, 3, 9) graph.floyd_warshall() graph.show_min(1, 4) graph.show_min(0, 3)
238
import shutil import tempfile import unittest from unittest.mock import patch from transformers import ( DefaultFlowCallback, IntervalStrategy, PrinterCallback, ProgressCallback, Trainer, TrainerCallback, TrainingArguments, is_torch_available, ) from transformers.testing_utils import require_torch if is_torch_available(): from transformers.trainer import DEFAULT_CALLBACKS from .test_trainer import RegressionDataset, RegressionModelConfig, RegressionPreTrainedModel class lowerCAmelCase__ ( __lowercase ): def __init__( self : int ) -> Optional[int]: __lowerCamelCase = [] def __A ( self : Any , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : str , **SCREAMING_SNAKE_CASE__ : List[Any] ) -> Tuple: self.events.append('''on_init_end''' ) def __A ( self : Tuple , SCREAMING_SNAKE_CASE__ : Optional[int] , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : Optional[int] , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Any: self.events.append('''on_train_begin''' ) def __A ( self : str , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Any , SCREAMING_SNAKE_CASE__ : Dict , **SCREAMING_SNAKE_CASE__ : int ) -> Any: self.events.append('''on_train_end''' ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Dict , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : List[Any] , **SCREAMING_SNAKE_CASE__ : List[Any] ) -> int: self.events.append('''on_epoch_begin''' ) def __A ( self : Optional[Any] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Any , SCREAMING_SNAKE_CASE__ : Optional[Any] , **SCREAMING_SNAKE_CASE__ : Tuple ) -> Optional[Any]: self.events.append('''on_epoch_end''' ) def __A ( self : str , SCREAMING_SNAKE_CASE__ : Tuple , SCREAMING_SNAKE_CASE__ : Optional[Any] , SCREAMING_SNAKE_CASE__ : Union[str, Any] , **SCREAMING_SNAKE_CASE__ : str ) -> List[Any]: self.events.append('''on_step_begin''' ) def __A ( self : Any , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : Any , **SCREAMING_SNAKE_CASE__ : Optional[int] ) -> Tuple: self.events.append('''on_step_end''' ) def __A ( self : int , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : int , **SCREAMING_SNAKE_CASE__ : List[Any] ) -> Union[str, Any]: self.events.append('''on_evaluate''' ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str , **SCREAMING_SNAKE_CASE__ : str ) -> str: self.events.append('''on_predict''' ) def __A ( self : Optional[int] , SCREAMING_SNAKE_CASE__ : int , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : Optional[int] , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Tuple: self.events.append('''on_save''' ) def __A ( self : Optional[int] , SCREAMING_SNAKE_CASE__ : Dict , SCREAMING_SNAKE_CASE__ : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Union[str, Any] , **SCREAMING_SNAKE_CASE__ : List[str] ) -> List[str]: self.events.append('''on_log''' ) def __A ( self : int , SCREAMING_SNAKE_CASE__ : List[Any] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : List[Any] , **SCREAMING_SNAKE_CASE__ : Optional[int] ) -> str: self.events.append('''on_prediction_step''' ) @require_torch class lowerCAmelCase__ ( unittest.TestCase ): def __A ( self : Union[str, Any] ) -> Optional[Any]: __lowerCamelCase = tempfile.mkdtemp() def __A ( self : int ) -> List[str]: shutil.rmtree(self.output_dir ) def __A ( self : Tuple , SCREAMING_SNAKE_CASE__ : List[Any]=0 , SCREAMING_SNAKE_CASE__ : Any=0 , SCREAMING_SNAKE_CASE__ : List[str]=64 , SCREAMING_SNAKE_CASE__ : Optional[int]=64 , SCREAMING_SNAKE_CASE__ : Optional[int]=None , SCREAMING_SNAKE_CASE__ : Optional[int]=False , **SCREAMING_SNAKE_CASE__ : Optional[Any] ) -> Union[str, Any]: # disable_tqdm in TrainingArguments has a flaky default since it depends on the level of logging. We make sure # its set to False since the tests later on depend on its value. __lowerCamelCase = RegressionDataset(length=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = RegressionDataset(length=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = RegressionModelConfig(a=SCREAMING_SNAKE_CASE__ , b=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = RegressionPreTrainedModel(SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = TrainingArguments(self.output_dir , disable_tqdm=SCREAMING_SNAKE_CASE__ , report_to=[] , **SCREAMING_SNAKE_CASE__ ) return Trainer( SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , train_dataset=SCREAMING_SNAKE_CASE__ , eval_dataset=SCREAMING_SNAKE_CASE__ , callbacks=SCREAMING_SNAKE_CASE__ , ) def __A ( self : List[Any] , SCREAMING_SNAKE_CASE__ : List[str] , SCREAMING_SNAKE_CASE__ : str ) -> Optional[int]: self.assertEqual(len(SCREAMING_SNAKE_CASE__ ) , len(SCREAMING_SNAKE_CASE__ ) ) # Order doesn't matter __lowerCamelCase = sorted(SCREAMING_SNAKE_CASE__ , key=lambda SCREAMING_SNAKE_CASE__ : cb.__name__ if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) else cb.__class__.__name__ ) __lowerCamelCase = sorted(SCREAMING_SNAKE_CASE__ , key=lambda SCREAMING_SNAKE_CASE__ : cb.__name__ if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) else cb.__class__.__name__ ) for cba, cba in zip(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): if isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) and isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): self.assertEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) elif isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) and not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): self.assertEqual(SCREAMING_SNAKE_CASE__ , cba.__class__ ) elif not isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) and isinstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ): self.assertEqual(cba.__class__ , SCREAMING_SNAKE_CASE__ ) else: self.assertEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) def __A ( self : Union[str, Any] , SCREAMING_SNAKE_CASE__ : Optional[int] ) -> Tuple: __lowerCamelCase = ['''on_init_end''', '''on_train_begin'''] __lowerCamelCase = 0 __lowerCamelCase = len(trainer.get_eval_dataloader() ) __lowerCamelCase = ['''on_prediction_step'''] * len(trainer.get_eval_dataloader() ) + ['''on_log''', '''on_evaluate'''] for _ in range(trainer.state.num_train_epochs ): expected_events.append('''on_epoch_begin''' ) for _ in range(SCREAMING_SNAKE_CASE__ ): step += 1 expected_events += ["on_step_begin", "on_step_end"] if step % trainer.args.logging_steps == 0: expected_events.append('''on_log''' ) if trainer.args.evaluation_strategy == IntervalStrategy.STEPS and step % trainer.args.eval_steps == 0: expected_events += evaluation_events.copy() if step % trainer.args.save_steps == 0: expected_events.append('''on_save''' ) expected_events.append('''on_epoch_end''' ) if trainer.args.evaluation_strategy == IntervalStrategy.EPOCH: expected_events += evaluation_events.copy() expected_events += ["on_log", "on_train_end"] return expected_events def __A ( self : Union[str, Any] ) -> int: __lowerCamelCase = self.get_trainer() __lowerCamelCase = DEFAULT_CALLBACKS.copy() + [ProgressCallback] self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) # Callbacks passed at init are added to the default callbacks __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] ) expected_callbacks.append(SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) # TrainingArguments.disable_tqdm controls if use ProgressCallback or PrinterCallback __lowerCamelCase = self.get_trainer(disable_tqdm=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = DEFAULT_CALLBACKS.copy() + [PrinterCallback] self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) def __A ( self : List[Any] ) -> str: __lowerCamelCase = DEFAULT_CALLBACKS.copy() + [ProgressCallback] __lowerCamelCase = self.get_trainer() # We can add, pop, or remove by class name trainer.remove_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.remove(SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = self.get_trainer() __lowerCamelCase = trainer.pop_callback(SCREAMING_SNAKE_CASE__ ) self.assertEqual(cb.__class__ , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) trainer.add_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.insert(0 , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) # We can also add, pop, or remove by instance __lowerCamelCase = self.get_trainer() __lowerCamelCase = trainer.callback_handler.callbacks[0] trainer.remove_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.remove(SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = self.get_trainer() __lowerCamelCase = trainer.callback_handler.callbacks[0] __lowerCamelCase = trainer.pop_callback(SCREAMING_SNAKE_CASE__ ) self.assertEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) trainer.add_callback(SCREAMING_SNAKE_CASE__ ) expected_callbacks.insert(0 , SCREAMING_SNAKE_CASE__ ) self.check_callbacks_equality(trainer.callback_handler.callbacks , SCREAMING_SNAKE_CASE__ ) def __A ( self : Union[str, Any] ) -> Any: import warnings # XXX: for now ignore scatter_gather warnings in this test since it's not relevant to what's being tested warnings.simplefilter(action='''ignore''' , category=SCREAMING_SNAKE_CASE__ ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) # Independent log/save/eval __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , logging_steps=5 ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , save_steps=5 ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , eval_steps=5 , evaluation_strategy='''steps''' ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) __lowerCamelCase = self.get_trainer(callbacks=[MyTestTrainerCallback] , evaluation_strategy='''epoch''' ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) # A bit of everything __lowerCamelCase = self.get_trainer( callbacks=[MyTestTrainerCallback] , logging_steps=3 , save_steps=10 , eval_steps=5 , evaluation_strategy='''steps''' , ) trainer.train() __lowerCamelCase = trainer.callback_handler.callbacks[-2].events self.assertEqual(SCREAMING_SNAKE_CASE__ , self.get_expected_events(SCREAMING_SNAKE_CASE__ ) ) # warning should be emitted for duplicated callbacks with patch('''transformers.trainer_callback.logger.warning''' ) as warn_mock: __lowerCamelCase = self.get_trainer( callbacks=[MyTestTrainerCallback, MyTestTrainerCallback] , ) assert str(SCREAMING_SNAKE_CASE__ ) in warn_mock.call_args[0][0]
270
0
'''simple docstring''' from __future__ import annotations import time __lowerCAmelCase = list[tuple[int, int]] __lowerCAmelCase = [ [0, 0, 0, 0, 0, 0, 0], [0, 1, 0, 0, 0, 0, 0], # 0 are free path whereas 1's are obstacles [0, 0, 0, 0, 0, 0, 0], [0, 0, 1, 0, 0, 0, 0], [1, 0, 1, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 1, 0, 0], ] __lowerCAmelCase = [[-1, 0], [0, -1], [1, 0], [0, 1]] # up, left, down, right class UpperCAmelCase__ : """simple docstring""" def __init__( self : Any ,_a : int ,_a : int ,_a : int ,_a : int ,_a : Node | None ): '''simple docstring''' _a : Union[str, Any] = pos_x _a : List[str] = pos_y _a : Optional[int] = (pos_y, pos_x) _a : Tuple = goal_x _a : Dict = goal_y _a : Tuple = parent class UpperCAmelCase__ : """simple docstring""" def __init__( self : Tuple ,_a : tuple[int, int] ,_a : tuple[int, int] ): '''simple docstring''' _a : Dict = Node(start[1] ,start[0] ,goal[1] ,goal[0] ,_a ) _a : Any = Node(goal[1] ,goal[0] ,goal[1] ,goal[0] ,_a ) _a : Dict = [self.start] _a : str = False def __lowercase ( self : Union[str, Any] ): '''simple docstring''' while self.node_queue: _a : Optional[int] = self.node_queue.pop(0 ) if current_node.pos == self.target.pos: _a : Dict = True return self.retrace_path(_a ) _a : Optional[Any] = self.get_successors(_a ) for node in successors: self.node_queue.append(_a ) if not self.reached: return [self.start.pos] return None def __lowercase ( self : List[Any] ,_a : Node ): '''simple docstring''' _a : str = [] for action in delta: _a : Any = parent.pos_x + action[1] _a : List[str] = parent.pos_y + action[0] if not (0 <= pos_x <= len(grid[0] ) - 1 and 0 <= pos_y <= len(_a ) - 1): continue if grid[pos_y][pos_x] != 0: continue successors.append( Node(_a ,_a ,self.target.pos_y ,self.target.pos_x ,_a ) ) return successors def __lowercase ( self : List[str] ,_a : Node | None ): '''simple docstring''' _a : Dict = node _a : Optional[Any] = [] while current_node is not None: path.append((current_node.pos_y, current_node.pos_x) ) _a : Union[str, Any] = current_node.parent path.reverse() return path class UpperCAmelCase__ : """simple docstring""" def __init__( self : int ,_a : str ,_a : Optional[int] ): '''simple docstring''' _a : Any = BreadthFirstSearch(_a ,_a ) _a : int = BreadthFirstSearch(_a ,_a ) _a : Dict = False def __lowercase ( self : Tuple ): '''simple docstring''' while self.fwd_bfs.node_queue or self.bwd_bfs.node_queue: _a : str = self.fwd_bfs.node_queue.pop(0 ) _a : Optional[int] = self.bwd_bfs.node_queue.pop(0 ) if current_bwd_node.pos == current_fwd_node.pos: _a : List[str] = True return self.retrace_bidirectional_path( _a ,_a ) _a : Tuple = current_bwd_node _a : Tuple = current_fwd_node _a : Tuple = { self.fwd_bfs: self.fwd_bfs.get_successors(_a ), self.bwd_bfs: self.bwd_bfs.get_successors(_a ), } for bfs in [self.fwd_bfs, self.bwd_bfs]: for node in successors[bfs]: bfs.node_queue.append(_a ) if not self.reached: return [self.fwd_bfs.start.pos] return None def __lowercase ( self : Dict ,_a : Node ,_a : Node ): '''simple docstring''' _a : List[Any] = self.fwd_bfs.retrace_path(_a ) _a : List[str] = self.bwd_bfs.retrace_path(_a ) bwd_path.pop() bwd_path.reverse() _a : List[str] = fwd_path + bwd_path return path if __name__ == "__main__": # all coordinates are given in format [y,x] import doctest doctest.testmod() __lowerCAmelCase = (0, 0) __lowerCAmelCase = (len(grid) - 1, len(grid[0]) - 1) for elem in grid: print(elem) __lowerCAmelCase = time.time() __lowerCAmelCase = BreadthFirstSearch(init, goal) __lowerCAmelCase = bfs.search() __lowerCAmelCase = time.time() - start_bfs_time print("""Unidirectional BFS computation time : """, bfs_time) __lowerCAmelCase = time.time() __lowerCAmelCase = BidirectionalBreadthFirstSearch(init, goal) __lowerCAmelCase = bd_bfs.search() __lowerCAmelCase = time.time() - start_bd_bfs_time print("""Bidirectional BFS computation time : """, bd_bfs_time)
271
'''simple docstring''' import unittest import numpy as np import torch from diffusers import PNDMPipeline, PNDMScheduler, UNetaDModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : List[str] ): '''simple docstring''' torch.manual_seed(0 ) _a : int = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : str = self.dummy_uncond_unet _a : int = PNDMScheduler() _a : str = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Any = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ,return_dict=_a )[0] _a : List[Any] = image[0, -3:, -3:, -1] _a : Any = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : List[Any] = np.array([1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2 assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1E-2 @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[str] = 'google/ddpm-cifar10-32' _a : str = UNetaDModel.from_pretrained(_a ) _a : Union[str, Any] = PNDMScheduler() _a : Tuple = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : str = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,output_type='numpy' ).images _a : Dict = image[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : Tuple = np.array([0.1564, 0.1_4645, 0.1406, 0.1_4715, 0.1_2425, 0.1_4045, 0.1_3115, 0.1_2175, 0.125] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
271
1
'''simple docstring''' import warnings from ...utils import logging from .image_processing_videomae import VideoMAEImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Tuple ,*_a : List[str] ,**_a : Any ): '''simple docstring''' warnings.warn( 'The class VideoMAEFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use VideoMAEImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
'''simple docstring''' import doctest import logging import os import unittest from pathlib import Path from typing import List, Union import transformers from transformers.testing_utils import require_tf, require_torch, slow __lowerCAmelCase = logging.getLogger() @unittest.skip('''Temporarily disable the doc tests.''' ) @require_torch @require_tf @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : str ,_a : Path ,_a : Union[str, None] = None ,_a : Union[List[str], None] = None ,_a : Union[str, List[str], None] = None ,_a : bool = True ,): '''simple docstring''' _a : Optional[int] = [file for file in os.listdir(_a ) if os.path.isfile(os.path.join(_a ,_a ) )] if identifier is not None: _a : List[str] = [file for file in files if identifier in file] if n_identifier is not None: if isinstance(_a ,_a ): for n_ in n_identifier: _a : Tuple = [file for file in files if n_ not in file] else: _a : Optional[Any] = [file for file in files if n_identifier not in file] _a : List[str] = ignore_files or [] ignore_files.append('__init__.py' ) _a : Tuple = [file for file in files if file not in ignore_files] for file in files: # Open all files print('Testing' ,_a ) if only_modules: _a : Any = file.split('.' )[0] try: _a : List[str] = getattr(_a ,_a ) _a : int = doctest.DocTestSuite(_a ) _a : Any = unittest.TextTestRunner().run(_a ) self.assertIs(len(result.failures ) ,0 ) except AttributeError: logger.info(F"""{module_identifier} is not a module.""" ) else: _a : Union[str, Any] = doctest.testfile(str('..' / directory / file ) ,optionflags=doctest.ELLIPSIS ) self.assertIs(result.failed ,0 ) def __lowercase ( self : Any ): '''simple docstring''' _a : int = Path('src/transformers' ) _a : List[Any] = 'modeling' _a : Optional[Any] = [ 'modeling_ctrl.py', 'modeling_tf_ctrl.py', ] self.analyze_directory(_a ,identifier=_a ,ignore_files=_a ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Optional[Any] = Path('src/transformers' ) _a : Optional[Any] = 'tokenization' self.analyze_directory(_a ,identifier=_a ) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Dict = Path('src/transformers' ) _a : str = 'configuration' self.analyze_directory(_a ,identifier=_a ) def __lowercase ( self : Any ): '''simple docstring''' _a : Tuple = Path('src/transformers' ) _a : List[Any] = ['configuration', 'modeling', 'tokenization'] self.analyze_directory(_a ,n_identifier=_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[Any] = Path('docs/source' ) _a : List[str] = ['favicon.ico'] self.analyze_directory(_a ,ignore_files=_a ,only_modules=_a )
271
1
'''simple docstring''' import copy from typing import TYPE_CHECKING, Any, Mapping, Optional, OrderedDict from packaging import version from ...configuration_utils import PretrainedConfig from ...onnx import OnnxConfig from ...utils import logging from ..auto.configuration_auto import AutoConfig if TYPE_CHECKING: from ... import PreTrainedTokenizerBase, TensorType __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : List[Any] = '''vision-encoder-decoder''' __UpperCAmelCase : str = True def __init__( self : Optional[int] ,**_a : Any ): '''simple docstring''' super().__init__(**_a ) if "encoder" not in kwargs or "decoder" not in kwargs: raise ValueError( F"""A configuraton of type {self.model_type} cannot be instantiated because """ F"""not both `encoder` and `decoder` sub-configurations are passed, but only {kwargs}""" ) _a : Any = kwargs.pop('encoder' ) _a : Tuple = encoder_config.pop('model_type' ) _a : int = kwargs.pop('decoder' ) _a : Dict = decoder_config.pop('model_type' ) _a : Optional[Any] = AutoConfig.for_model(_a ,**_a ) _a : Union[str, Any] = AutoConfig.for_model(_a ,**_a ) _a : Optional[Any] = True @classmethod def __lowercase ( cls : Optional[int] ,_a : PretrainedConfig ,_a : PretrainedConfig ,**_a : Any ): '''simple docstring''' logger.info('Setting `config.is_decoder=True` and `config.add_cross_attention=True` for decoder_config' ) _a : List[str] = True _a : Union[str, Any] = True return cls(encoder=encoder_config.to_dict() ,decoder=decoder_config.to_dict() ,**_a ) def __lowercase ( self : int ): '''simple docstring''' _a : Optional[Any] = copy.deepcopy(self.__dict__ ) _a : int = self.encoder.to_dict() _a : int = self.decoder.to_dict() _a : Any = self.__class__.model_type return output class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Union[str, Any] = version.parse('''1.11''' ) @property def __lowercase ( self : int ): '''simple docstring''' return OrderedDict( [ ('pixel_values', {0: 'batch', 1: 'num_channels', 2: 'height', 3: 'width'}), ] ) @property def __lowercase ( self : Optional[Any] ): '''simple docstring''' return 1E-4 @property def __lowercase ( self : int ): '''simple docstring''' return OrderedDict({'last_hidden_state': {0: 'batch', 1: 'encoder_sequence'}} ) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" @property def __lowercase ( self : List[str] ): '''simple docstring''' _a : List[str] = OrderedDict() _a : Optional[int] = {0: 'batch', 1: 'past_decoder_sequence + sequence'} _a : Optional[int] = {0: 'batch', 1: 'past_decoder_sequence + sequence'} _a : Any = {0: 'batch', 1: 'encoder_sequence'} return common_inputs def __lowercase ( self : Optional[int] ,_a : "PreTrainedTokenizerBase" ,_a : int = -1 ,_a : int = -1 ,_a : bool = False ,_a : Optional["TensorType"] = None ,): '''simple docstring''' import torch _a : List[str] = OrderedDict() _a : Dict = super().generate_dummy_inputs( _a ,batch_size=_a ,seq_length=_a ,is_pair=_a ,framework=_a ) _a, _a : str = dummy_input['input_ids'].shape _a : List[str] = (batch, encoder_sequence, self._config.encoder_hidden_size) _a : Optional[int] = dummy_input.pop('input_ids' ) _a : Union[str, Any] = dummy_input.pop('attention_mask' ) _a : str = torch.zeros(_a ) return common_inputs class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" @property def __lowercase ( self : str ): '''simple docstring''' pass def __lowercase ( self : Any ,_a : PretrainedConfig ): '''simple docstring''' return VisionEncoderDecoderEncoderOnnxConfig(_a ) def __lowercase ( self : Any ,_a : PretrainedConfig ,_a : PretrainedConfig ,_a : str = "default" ): '''simple docstring''' _a : Tuple = encoder_config.hidden_size return VisionEncoderDecoderDecoderOnnxConfig(_a ,_a )
271
'''simple docstring''' import argparse import pickle import numpy as np import torch from torch import nn from transformers import ReformerConfig, ReformerModelWithLMHead from transformers.utils import logging logging.set_verbosity_info() def UpperCAmelCase_ (__a : Optional[Any] , __a : str , __a : Optional[Any]=None ): """simple docstring""" assert torch_layer.weight.shape == weight.shape, f"""{torch_layer} layer.weight does not match""" _a : str = nn.Parameter(__a ) if bias is not None: assert torch_layer.bias.shape == bias.shape, f"""{torch_layer} layer.bias does not match""" _a : Any = nn.Parameter(__a ) def UpperCAmelCase_ (__a : int , __a : Optional[Any] , __a : int ): """simple docstring""" _a : Tuple = np.asarray(weights[0] ) _a : Union[str, Any] = np.asarray(weights[1] ) _a : Dict = np.asarray(weights[2] ) set_param( torch_layer.self_attention.query_key , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.value , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.output.dense , torch.tensor(__a ).view(-1 , __a ).contiguous().transpose(0 , 1 ) , ) def UpperCAmelCase_ (__a : Optional[Any] , __a : Optional[int] , __a : List[str] ): """simple docstring""" _a : Dict = np.asarray(weights[0] ) _a : Union[str, Any] = np.asarray(weights[1] ) _a : str = np.asarray(weights[2] ) _a : int = np.asarray(weights[3] ) set_param( torch_layer.self_attention.query , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.key , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.value , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.output.dense , torch.tensor(__a ).view(-1 , __a ).contiguous().transpose(0 , 1 ) , ) def UpperCAmelCase_ (__a : Any , __a : Any , __a : Optional[Any] ): """simple docstring""" _a : List[str] = weights[0][0][0] _a : List[Any] = np.asarray(layer_norm_a[0] ) _a : List[str] = np.asarray(layer_norm_a[1] ) set_param( torch_block.attention.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # lsh weights + output _a : List[str] = weights[0][1] if len(__a ) < 4: set_layer_weights_in_torch_lsh(__a , torch_block.attention , __a ) else: set_layer_weights_in_torch_local(__a , torch_block.attention , __a ) # intermediate weighs _a : Optional[Any] = weights[2][0][1][2] # Chunked Feed Forward if len(__a ) == 4: _a : Union[str, Any] = intermediate_weights[2] # layernorm 2 _a : Any = np.asarray(intermediate_weights[0][0] ) _a : List[Any] = np.asarray(intermediate_weights[0][1] ) set_param( torch_block.feed_forward.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # intermediate dense _a : Any = np.asarray(intermediate_weights[1][0] ) _a : Any = np.asarray(intermediate_weights[1][1] ) set_param( torch_block.feed_forward.dense.dense , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) # intermediate out _a : Optional[int] = np.asarray(intermediate_weights[4][0] ) _a : int = np.asarray(intermediate_weights[4][1] ) set_param( torch_block.feed_forward.output.dense , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) def UpperCAmelCase_ (__a : Dict , __a : Dict , __a : List[Any] ): """simple docstring""" _a : Optional[int] = torch_model.reformer # word embeds _a : Tuple = np.asarray(weights[1] ) set_param( torch_model_reformer.embeddings.word_embeddings , torch.tensor(__a ) , ) if isinstance(weights[3] , __a ): _a : Any = torch_model_reformer.embeddings.position_embeddings for emb_idx in range(len(position_embeddings.weights ) ): _a : List[Any] = np.asarray(weights[3][emb_idx][0] ) assert ( position_embeddings.weights[emb_idx].shape == emb_weights.shape ), f"""{position_embeddings[emb_idx]} emb does not match""" _a : Any = nn.Parameter(torch.tensor(__a ) ) _a : List[str] = weights[5] assert len(torch_model_reformer.encoder.layers ) * 4 == len( __a ), "HF and trax model do not have the same number of layers" for layer_idx, layer in enumerate(torch_model_reformer.encoder.layers ): _a : Tuple = trax_layer_weights[4 * layer_idx : 4 * (layer_idx + 1)] set_block_weights_in_torch(__a , __a , __a ) # output layer norm _a : Optional[Any] = np.asarray(weights[7][0] ) _a : int = np.asarray(weights[7][1] ) set_param( torch_model_reformer.encoder.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # output embeddings _a : List[str] = np.asarray(weights[9][0] ) _a : int = np.asarray(weights[9][1] ) set_param( torch_model.lm_head.decoder , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) def UpperCAmelCase_ (__a : Tuple , __a : Optional[Any] , __a : Dict ): """simple docstring""" _a : List[Any] = ReformerConfig.from_json_file(__a ) print(f"""Building PyTorch model from configuration: {config}""" ) _a : int = ReformerModelWithLMHead(__a ) with open(__a , 'rb' ) as f: _a : Optional[Any] = pickle.load(__a )['weights'] set_model_weights_in_torch(__a , __a , config.hidden_size ) # Save pytorch-model print(f"""Save PyTorch model to {pytorch_dump_path}""" ) torch.save(model.state_dict() , __a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser() # Required parameters parser.add_argument( """--trax_model_pkl_path""", default=None, type=str, required=True, help="""Path to the TensorFlow checkpoint path.""" ) parser.add_argument( """--config_file""", default=None, type=str, required=True, help=( """The config json file corresponding to the pre-trained Reformer model. \n""" """This specifies the model architecture.""" ), ) parser.add_argument( """--pytorch_dump_path""", default=None, type=str, required=True, help="""Path to the output PyTorch model.""" ) __lowerCAmelCase = parser.parse_args() convert_trax_checkpoint_to_pytorch(args.trax_model_pkl_path, args.config_file, args.pytorch_dump_path)
271
1
'''simple docstring''' import argparse import json from typing import List from ltp import LTP from transformers import BertTokenizer def UpperCAmelCase_ (__a : List[Any] ): """simple docstring""" if ( (cp >= 0x4E_00 and cp <= 0x9F_FF) or (cp >= 0x34_00 and cp <= 0x4D_BF) # or (cp >= 0x2_00_00 and cp <= 0x2_A6_DF) # or (cp >= 0x2_A7_00 and cp <= 0x2_B7_3F) # or (cp >= 0x2_B7_40 and cp <= 0x2_B8_1F) # or (cp >= 0x2_B8_20 and cp <= 0x2_CE_AF) # or (cp >= 0xF9_00 and cp <= 0xFA_FF) or (cp >= 0x2_F8_00 and cp <= 0x2_FA_1F) # ): # return True return False def UpperCAmelCase_ (__a : str ): """simple docstring""" for char in word: _a : Union[str, Any] = ord(__a ) if not _is_chinese_char(__a ): return 0 return 1 def UpperCAmelCase_ (__a : List[str] ): """simple docstring""" _a : Dict = set() for token in tokens: _a : str = len(__a ) > 1 and is_chinese(__a ) if chinese_word: word_set.add(__a ) _a : Optional[Any] = list(__a ) return word_list def UpperCAmelCase_ (__a : List[str] , __a : set() ): """simple docstring""" if not chinese_word_set: return bert_tokens _a : Optional[Any] = max([len(__a ) for w in chinese_word_set] ) _a : Optional[int] = bert_tokens _a, _a : Any = 0, len(__a ) while start < end: _a : Tuple = True if is_chinese(bert_word[start] ): _a : Union[str, Any] = min(end - start , __a ) for i in range(__a , 1 , -1 ): _a : Optional[Any] = ''.join(bert_word[start : start + i] ) if whole_word in chinese_word_set: for j in range(start + 1 , start + i ): _a : Any = '##' + bert_word[j] _a : Union[str, Any] = start + i _a : int = False break if single_word: start += 1 return bert_word def UpperCAmelCase_ (__a : List[str] , __a : LTP , __a : BertTokenizer ): """simple docstring""" _a : int = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : Union[str, Any] = ltp_tokenizer.seg(lines[i : i + 1_0_0] )[0] _a : Optional[Any] = [get_chinese_word(__a ) for r in res] ltp_res.extend(__a ) assert len(__a ) == len(__a ) _a : str = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : List[str] = bert_tokenizer(lines[i : i + 1_0_0] , add_special_tokens=__a , truncation=__a , max_length=5_1_2 ) bert_res.extend(res['input_ids'] ) assert len(__a ) == len(__a ) _a : List[str] = [] for input_ids, chinese_word in zip(__a , __a ): _a : int = [] for id in input_ids: _a : Optional[int] = bert_tokenizer._convert_id_to_token(__a ) input_tokens.append(__a ) _a : List[str] = add_sub_symbol(__a , __a ) _a : Tuple = [] # We only save pos of chinese subwords start with ##, which mean is part of a whole word. for i, token in enumerate(__a ): if token[:2] == "##": _a : str = token[2:] # save chinese tokens' pos if len(__a ) == 1 and _is_chinese_char(ord(__a ) ): ref_id.append(__a ) ref_ids.append(__a ) assert len(__a ) == len(__a ) return ref_ids def UpperCAmelCase_ (__a : Optional[Any] ): """simple docstring""" with open(args.file_name , 'r' , encoding='utf-8' ) as f: _a : Dict = f.readlines() _a : int = [line.strip() for line in data if len(__a ) > 0 and not line.isspace()] # avoid delimiter like '\u2029' _a : int = LTP(args.ltp ) # faster in GPU device _a : Tuple = BertTokenizer.from_pretrained(args.bert ) _a : int = prepare_ref(__a , __a , __a ) with open(args.save_path , 'w' , encoding='utf-8' ) as f: _a : Optional[Any] = [json.dumps(__a ) + '\n' for ref in ref_ids] f.writelines(__a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser(description="""prepare_chinese_ref""") parser.add_argument( """--file_name""", type=str, default="""./resources/chinese-demo.txt""", help="""file need process, same as training data in lm""", ) parser.add_argument( """--ltp""", type=str, default="""./resources/ltp""", help="""resources for LTP tokenizer, usually a path""" ) parser.add_argument("""--bert""", type=str, default="""./resources/robert""", help="""resources for Bert tokenizer""") parser.add_argument("""--save_path""", type=str, default="""./resources/ref.txt""", help="""path to save res""") __lowerCAmelCase = parser.parse_args() main(args)
271
'''simple docstring''' import unittest import numpy as np import torch from transformers import CLIPTextConfig, CLIPTextModel from diffusers import DDIMScheduler, LDMPipeline, UNetaDModel, VQModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' torch.manual_seed(0 ) _a : Any = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model @property def __lowercase ( self : Optional[int] ): '''simple docstring''' torch.manual_seed(0 ) _a : Union[str, Any] = VQModel( block_out_channels=[32, 64] ,in_channels=3 ,out_channels=3 ,down_block_types=['DownEncoderBlock2D', 'DownEncoderBlock2D'] ,up_block_types=['UpDecoderBlock2D', 'UpDecoderBlock2D'] ,latent_channels=3 ,) return model @property def __lowercase ( self : Optional[int] ): '''simple docstring''' torch.manual_seed(0 ) _a : Any = 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 ,) return CLIPTextModel(_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : Dict = self.dummy_uncond_unet _a : List[Any] = DDIMScheduler() _a : List[Any] = self.dummy_vq_model _a : str = LDMPipeline(unet=_a ,vqvae=_a ,scheduler=_a ) ldm.to(_a ) ldm.set_progress_bar_config(disable=_a ) _a : List[str] = torch.manual_seed(0 ) _a : List[str] = ldm(generator=_a ,num_inference_steps=2 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Union[str, Any] = ldm(generator=_a ,num_inference_steps=2 ,output_type='numpy' ,return_dict=_a )[0] _a : Tuple = image[0, -3:, -3:, -1] _a : Optional[Any] = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 64, 64, 3) _a : int = np.array([0.8512, 0.818, 0.6411, 0.6808, 0.4465, 0.5618, 0.46, 0.6231, 0.5172] ) _a : Any = 1E-2 if torch_device != 'mps' else 3E-2 assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < tolerance @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : List[str] = LDMPipeline.from_pretrained('CompVis/ldm-celebahq-256' ) ldm.to(_a ) ldm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Dict = ldm(generator=_a ,num_inference_steps=5 ,output_type='numpy' ).images _a : str = image[0, -3:, -3:, -1] assert image.shape == (1, 256, 256, 3) _a : Optional[int] = np.array([0.4399, 0.4_4975, 0.4_6825, 0.474, 0.4359, 0.4581, 0.4_5095, 0.4341, 0.4447] ) _a : int = 1E-2 if torch_device != 'mps' else 3E-2 assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
271
1
'''simple docstring''' import functools import operator from ...configuration_utils import PretrainedConfig from ...utils import logging __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """asapp/sew-d-tiny-100k""": """https://huggingface.co/asapp/sew-d-tiny-100k/resolve/main/config.json""", # See all SEW-D models at https://huggingface.co/models?filter=sew-d } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : str = '''sew-d''' def __init__( self : Any ,_a : Dict=32 ,_a : Optional[Any]=768 ,_a : Dict=12 ,_a : Union[str, Any]=12 ,_a : int=3072 ,_a : Optional[int]=2 ,_a : Optional[Any]=512 ,_a : Optional[int]=256 ,_a : Dict=True ,_a : List[Any]=True ,_a : Any=("p2c", "c2p") ,_a : Optional[Any]="layer_norm" ,_a : Dict="gelu_python" ,_a : Tuple=0.1 ,_a : Union[str, Any]=0.1 ,_a : Union[str, Any]=0.1 ,_a : Dict=0.0 ,_a : int=0.1 ,_a : Any=0.02 ,_a : int=1E-7 ,_a : Any=1E-5 ,_a : int="group" ,_a : int="gelu" ,_a : List[str]=(64, 128, 128, 128, 128, 256, 256, 256, 256, 512, 512, 512, 512) ,_a : List[str]=(5, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1) ,_a : Optional[int]=(10, 3, 1, 3, 1, 3, 1, 3, 1, 2, 1, 2, 1) ,_a : List[Any]=False ,_a : str=128 ,_a : List[str]=16 ,_a : List[Any]=True ,_a : Tuple=0.05 ,_a : Tuple=10 ,_a : List[str]=2 ,_a : Optional[int]=0.0 ,_a : List[Any]=10 ,_a : Union[str, Any]=0 ,_a : Optional[Any]="mean" ,_a : Any=False ,_a : Union[str, Any]=False ,_a : Union[str, Any]=256 ,_a : Optional[Any]=0 ,_a : str=1 ,_a : Optional[int]=2 ,**_a : Dict ,): '''simple docstring''' super().__init__(**_a ,pad_token_id=_a ,bos_token_id=_a ,eos_token_id=_a ) _a : Optional[Any] = hidden_size _a : str = feat_extract_norm _a : Dict = feat_extract_activation _a : Tuple = list(_a ) _a : int = list(_a ) _a : Tuple = list(_a ) _a : List[str] = conv_bias _a : Union[str, Any] = num_conv_pos_embeddings _a : Dict = num_conv_pos_embedding_groups _a : Dict = len(self.conv_dim ) _a : List[str] = num_hidden_layers _a : Union[str, Any] = intermediate_size _a : Tuple = squeeze_factor _a : Dict = max_position_embeddings _a : Any = position_buckets _a : Optional[int] = share_att_key _a : str = relative_attention _a : Dict = norm_rel_ebd _a : int = list(_a ) _a : Dict = hidden_act _a : Union[str, Any] = num_attention_heads _a : str = hidden_dropout _a : List[str] = attention_dropout _a : int = activation_dropout _a : Dict = feat_proj_dropout _a : Optional[int] = final_dropout _a : str = layer_norm_eps _a : List[Any] = feature_layer_norm_eps _a : int = initializer_range _a : Union[str, Any] = vocab_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)`,' F"""but is `len(config.conv_dim) = {len(self.conv_dim )}`, `len(config.conv_stride)""" F"""= {len(self.conv_stride )}`, `len(config.conv_kernel) = {len(self.conv_kernel )}`.""" ) # fine-tuning config parameters for SpecAugment: https://arxiv.org/abs/1904.08779 _a : List[str] = apply_spec_augment _a : Optional[int] = mask_time_prob _a : str = mask_time_length _a : Dict = mask_time_min_masks _a : int = mask_feature_prob _a : Tuple = mask_feature_length _a : Optional[Any] = mask_feature_min_masks # ctc loss _a : Optional[int] = ctc_loss_reduction _a : List[Any] = ctc_zero_infinity # sequence classification _a : Any = use_weighted_layer_sum _a : List[str] = classifier_proj_size @property def __lowercase ( self : Any ): '''simple docstring''' return functools.reduce(operator.mul ,self.conv_stride ,1 )
271
'''simple docstring''' import warnings from ...utils import logging from .image_processing_beit import BeitImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : int ,*_a : Optional[int] ,**_a : str ): '''simple docstring''' warnings.warn( 'The class BeitFeatureExtractor is deprecated and will be removed in version 5 of Transformers. Please' ' use BeitImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
1
'''simple docstring''' from ...configuration_utils import PretrainedConfig from ...utils import logging __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """naver-clova-ix/donut-base""": """https://huggingface.co/naver-clova-ix/donut-base/resolve/main/config.json""", # See all Donut models at https://huggingface.co/models?filter=donut-swin } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : List[str] = '''donut-swin''' __UpperCAmelCase : Optional[Any] = { '''num_attention_heads''': '''num_heads''', '''num_hidden_layers''': '''num_layers''', } def __init__( self : Dict ,_a : Optional[Any]=224 ,_a : List[str]=4 ,_a : Union[str, Any]=3 ,_a : Any=96 ,_a : str=[2, 2, 6, 2] ,_a : str=[3, 6, 12, 24] ,_a : Union[str, Any]=7 ,_a : int=4.0 ,_a : Optional[Any]=True ,_a : Optional[Any]=0.0 ,_a : List[Any]=0.0 ,_a : Dict=0.1 ,_a : List[Any]="gelu" ,_a : str=False ,_a : List[Any]=0.02 ,_a : Dict=1E-5 ,**_a : List[Any] ,): '''simple docstring''' super().__init__(**_a ) _a : Any = image_size _a : Any = patch_size _a : str = num_channels _a : int = embed_dim _a : Union[str, Any] = depths _a : Union[str, Any] = len(_a ) _a : Optional[int] = num_heads _a : Optional[int] = window_size _a : List[str] = mlp_ratio _a : str = qkv_bias _a : int = hidden_dropout_prob _a : Dict = attention_probs_dropout_prob _a : str = drop_path_rate _a : List[str] = hidden_act _a : List[str] = use_absolute_embeddings _a : int = layer_norm_eps _a : List[Any] = initializer_range # we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel # this indicates the channel dimension after the last stage of the model _a : Optional[int] = int(embed_dim * 2 ** (len(_a ) - 1) )
271
'''simple docstring''' from typing import Any, Dict, List, Optional, Tuple, Union import torch from torch import nn from torch.utils.data import DistributedSampler, RandomSampler from transformers import PreTrainedModel, Trainer, logging from transformers.integrations import is_fairscale_available from transformers.models.fsmt.configuration_fsmt import FSMTConfig from transformers.optimization import ( Adafactor, AdamW, get_constant_schedule, get_constant_schedule_with_warmup, get_cosine_schedule_with_warmup, get_cosine_with_hard_restarts_schedule_with_warmup, get_linear_schedule_with_warmup, get_polynomial_decay_schedule_with_warmup, ) from transformers.trainer_pt_utils import get_tpu_sampler from transformers.training_args import ParallelMode from transformers.utils import is_torch_tpu_available if is_fairscale_available(): from fairscale.optim import OSS __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """linear""": get_linear_schedule_with_warmup, """cosine""": get_cosine_schedule_with_warmup, """cosine_w_restarts""": get_cosine_with_hard_restarts_schedule_with_warmup, """polynomial""": get_polynomial_decay_schedule_with_warmup, """constant""": get_constant_schedule, """constant_w_warmup""": get_constant_schedule_with_warmup, } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Optional[int] ,_a : Optional[Any]=None ,_a : Dict=None ,*_a : int ,**_a : str ): '''simple docstring''' super().__init__(*_a ,**_a ) if config is None: assert isinstance(self.model ,_a ), ( "If no `config` is passed the model to be trained has to be of type `PreTrainedModel`, but is" F""" {self.model.__class__}""" ) _a : List[Any] = self.model.config else: _a : Optional[int] = config _a : List[str] = data_args _a : List[Any] = self.config.tgt_vocab_size if isinstance(self.config ,_a ) else self.config.vocab_size if self.args.label_smoothing != 0 or (self.data_args is not None and self.data_args.ignore_pad_token_for_loss): assert self.config.pad_token_id is not None, ( "Make sure that `config.pad_token_id` is correcly defined when ignoring `pad_token` for loss" " calculation or doing label smoothing." ) if self.config.pad_token_id is None and self.config.eos_token_id is not None: logger.warning( F"""The `config.pad_token_id` is `None`. Using `config.eos_token_id` = {self.config.eos_token_id} for""" ' padding..' ) if self.args.label_smoothing == 0: _a : List[str] = torch.nn.CrossEntropyLoss(ignore_index=self.config.pad_token_id ) else: # dynamically import label_smoothed_nll_loss from utils import label_smoothed_nll_loss _a : Tuple = label_smoothed_nll_loss def __lowercase ( self : List[str] ,_a : int ): '''simple docstring''' if self.optimizer is None: _a : Union[str, Any] = ['bias', 'LayerNorm.weight'] _a : Tuple = [ { 'params': [p for n, p in self.model.named_parameters() if not any(nd in n for nd in no_decay )], 'weight_decay': self.args.weight_decay, }, { 'params': [p for n, p in self.model.named_parameters() if any(nd in n for nd in no_decay )], 'weight_decay': 0.0, }, ] _a : Optional[int] = Adafactor if self.args.adafactor else AdamW if self.args.adafactor: _a : Any = Adafactor _a : Dict = {'scale_parameter': False, 'relative_step': False} else: _a : Union[str, Any] = AdamW _a : str = { 'betas': (self.args.adam_betaa, self.args.adam_betaa), 'eps': self.args.adam_epsilon, } _a : Union[str, Any] = self.args.learning_rate if self.sharded_ddp: _a : str = OSS( params=_a ,optim=_a ,**_a ,) else: _a : Tuple = optimizer_cls(_a ,**_a ) if self.lr_scheduler is None: _a : List[Any] = self._get_lr_scheduler(_a ) else: # ignoring --lr_scheduler logger.warning('scheduler is passed to `Seq2SeqTrainer`, `--lr_scheduler` arg is ignored.' ) def __lowercase ( self : List[Any] ,_a : List[Any] ): '''simple docstring''' _a : str = arg_to_scheduler[self.args.lr_scheduler] if self.args.lr_scheduler == "constant": _a : int = schedule_func(self.optimizer ) elif self.args.lr_scheduler == "constant_w_warmup": _a : List[str] = schedule_func(self.optimizer ,num_warmup_steps=self.args.warmup_steps ) else: _a : Optional[int] = schedule_func( self.optimizer ,num_warmup_steps=self.args.warmup_steps ,num_training_steps=_a ) return scheduler def __lowercase ( self : Tuple ): '''simple docstring''' if isinstance(self.train_dataset ,torch.utils.data.IterableDataset ): return None elif is_torch_tpu_available(): return get_tpu_sampler(self.train_dataset ) else: if self.args.sortish_sampler: self.train_dataset.make_sortish_sampler( self.args.per_device_train_batch_size ,distributed=(self.args.parallel_mode == ParallelMode.DISTRIBUTED) ,) return ( RandomSampler(self.train_dataset ) if self.args.local_rank == -1 else DistributedSampler(self.train_dataset ) ) def __lowercase ( self : Dict ,_a : Dict ,_a : Any ,_a : Dict ): '''simple docstring''' if self.args.label_smoothing == 0: if self.data_args is not None and self.data_args.ignore_pad_token_for_loss: # force training to ignore pad token _a : List[Any] = model(**_a ,use_cache=_a )[0] _a : Union[str, Any] = self.loss_fn(logits.view(-1 ,logits.shape[-1] ) ,labels.view(-1 ) ) else: # compute usual loss via models _a, _a : Union[str, Any] = model(**_a ,labels=_a ,use_cache=_a )[:2] else: # compute label smoothed loss _a : List[Any] = model(**_a ,use_cache=_a )[0] _a : Any = torch.nn.functional.log_softmax(_a ,dim=-1 ) _a, _a : List[str] = self.loss_fn(_a ,_a ,self.args.label_smoothing ,ignore_index=self.config.pad_token_id ) return loss, logits def __lowercase ( self : Optional[int] ,_a : Union[str, Any] ,_a : List[Any] ): '''simple docstring''' _a : Optional[int] = inputs.pop('labels' ) _a, _a : int = self._compute_loss(_a ,_a ,_a ) return loss def __lowercase ( self : Optional[Any] ,_a : nn.Module ,_a : Dict[str, Union[torch.Tensor, Any]] ,_a : bool ,_a : Optional[List[str]] = None ,): '''simple docstring''' _a : int = self._prepare_inputs(_a ) _a : Any = { 'max_length': self.data_args.val_max_target_length if self.data_args is not None else self.config.max_length, 'num_beams': self.data_args.eval_beams if self.data_args is not None else self.config.num_beams, } if self.args.predict_with_generate and not self.args.prediction_loss_only: _a : int = self.model.generate( inputs['input_ids'] ,attention_mask=inputs['attention_mask'] ,**_a ,) # in case the batch is shorter than max length, the output should be padded if generated_tokens.shape[-1] < gen_kwargs["max_length"]: _a : int = self._pad_tensors_to_max_len(_a ,gen_kwargs['max_length'] ) _a : Union[str, Any] = inputs.pop('labels' ) with torch.no_grad(): # compute loss on predict data _a, _a : Optional[int] = self._compute_loss(_a ,_a ,_a ) _a : Optional[Any] = loss.mean().detach() if self.args.prediction_loss_only: return (loss, None, None) _a : Optional[Any] = generated_tokens if self.args.predict_with_generate else logits if labels.shape[-1] < gen_kwargs["max_length"]: _a : Dict = self._pad_tensors_to_max_len(_a ,gen_kwargs['max_length'] ) return (loss, logits, labels) def __lowercase ( self : str ,_a : Tuple ,_a : Tuple ): '''simple docstring''' _a : List[Any] = self.config.pad_token_id if self.config.pad_token_id is not None else self.config.eos_token_id if pad_token_id is None: raise ValueError( 'Make sure that either `config.pad_token_id` or `config.eos_token_id` is defined if tensor has to be' F""" padded to `max_length`={max_length}""" ) _a : int = pad_token_id * torch.ones( (tensor.shape[0], max_length) ,dtype=tensor.dtype ,device=tensor.device ) _a : Union[str, Any] = tensor return padded_tensor
271
1
'''simple docstring''' import warnings from pathlib import Path from typing import List, Tuple, Union import fire from torch import nn from transformers import AutoModelForSeqaSeqLM, AutoTokenizer, PreTrainedModel from transformers.utils import logging __lowerCAmelCase = logging.get_logger(__name__) def UpperCAmelCase_ (__a : nn.ModuleList , __a : nn.ModuleList , __a : List[int] ): """simple docstring""" _a : Optional[int] = nn.ModuleList([src_layers[i] for i in layers_to_copy] ) assert len(__a ) == len(__a ), f"""{len(__a )} != {len(__a )}""" dest_layers.load_state_dict(layers_to_copy.state_dict() ) __lowerCAmelCase = { # maps num layers in teacher -> num_layers in student -> which teacher layers to copy. # 12: bart, 16: pegasus, 6: marian/Helsinki-NLP 1_2: { 1: [0], # This says that if the teacher has 12 layers and the student has 1, copy layer 0 of the teacher 2: [0, 6], 3: [0, 6, 1_1], 4: [0, 4, 8, 1_1], 6: [0, 2, 4, 7, 9, 1_1], 9: [0, 1, 2, 4, 5, 7, 9, 1_0, 1_1], 1_2: list(range(1_2)), }, 1_6: { # maps num layers in student -> which teacher layers to copy 1: [0], 2: [0, 1_5], 3: [0, 8, 1_5], 4: [0, 5, 1_0, 1_5], 6: [0, 3, 6, 9, 1_2, 1_5], 8: [0, 2, 4, 6, 8, 1_0, 1_2, 1_5], 9: [0, 1, 3, 5, 7, 9, 1_1, 1_3, 1_5], 1_2: [0, 1, 2, 3, 4, 5, 6, 7, 9, 1_1, 1_3, 1_5], 1_6: list(range(1_6)), }, 6: {1: [0], 2: [0, 5], 3: [0, 2, 5], 4: [0, 1, 3, 5], 6: list(range(6))}, } __lowerCAmelCase = { # maps num layers in student -> which teacher layers to copy. 6: {1: [5], 2: [3, 5], 3: [1, 4, 5], 4: [1, 2, 4, 5]}, 1_2: {1: [1_1], 2: [5, 1_1], 3: [3, 7, 1_1], 6: [1, 3, 5, 8, 1_0, 1_1]}, 1_6: {1: [1_5], 4: [4, 9, 1_2, 1_5], 8: [1, 3, 5, 7, 9, 1_1, 1_3, 1_5]}, } def UpperCAmelCase_ (__a : List[Any] , __a : Optional[Any] ): """simple docstring""" try: _a : Tuple = LAYERS_TO_COPY[n_teacher][n_student] return val except KeyError: if n_student != n_teacher: warnings.warn( f"""no hardcoded layers to copy for teacher {n_teacher} -> student {n_student}, defaulting to first""" f""" {n_student}""" ) return list(range(__a ) ) def UpperCAmelCase_ (__a : Any , __a : Any ): """simple docstring""" if n_student > n_teacher: raise ValueError(f"""Cannot perform intermediate supervision for student {n_student} > teacher {n_teacher}""" ) elif n_teacher == n_student: return list(range(__a ) ) elif n_student == 1: return [n_teacher - 1] else: return LAYERS_TO_SUPERVISE[n_teacher][n_student] def UpperCAmelCase_ (__a : Union[str, PreTrainedModel] , __a : Union[str, Path] = "student" , __a : Union[int, None] = None , __a : Union[int, None] = None , __a : str=False , __a : Union[str, Any]=None , __a : int=None , **__a : Optional[Any] , ): """simple docstring""" _a : str = 'encoder_layers and decoder_layers cannot be both None-- you would just have an identical teacher.' assert (e is not None) or (d is not None), _msg if isinstance(__a , __a ): AutoTokenizer.from_pretrained(__a ).save_pretrained(__a ) # purely for convenience _a : str = AutoModelForSeqaSeqLM.from_pretrained(__a ).eval() else: assert isinstance(__a , __a ), f"""teacher must be a model or string got type {type(__a )}""" _a : Any = teacher.config.to_diff_dict() try: _a, _a : Optional[Any] = teacher.config.encoder_layers, teacher.config.decoder_layers if e is None: _a : str = teacher_e if d is None: _a : Union[str, Any] = teacher_d init_kwargs.update({'encoder_layers': e, 'decoder_layers': d} ) except AttributeError: # T5 if hasattr(teacher.config , 'num_encoder_layers' ): _a, _a : int = teacher.config.num_encoder_layers, teacher.config.num_decoder_layers else: _a, _a : Any = teacher.config.num_layers, teacher.config.num_decoder_layers if e is None: _a : int = teacher_e if d is None: _a : Tuple = teacher_d if hasattr(teacher.config , 'num_encoder_layers' ): init_kwargs.update({'num_encoder_layers': e, 'num_decoder_layers': d} ) else: init_kwargs.update({'num_layers': e, 'num_decoder_layers': d} ) # Kwargs to instantiate student: teacher kwargs with updated layer numbers + **extra_config_kwargs init_kwargs.update(__a ) # Copy weights _a : Tuple = teacher.config_class(**__a ) _a : Tuple = AutoModelForSeqaSeqLM.from_config(__a ) # Start by copying the full teacher state dict this will copy the first N teacher layers to the student. _a : Optional[Any] = student.load_state_dict(teacher.state_dict() , strict=__a ) assert info.missing_keys == [], info.missing_keys # every student key should have a teacher keys. if copy_first_teacher_layers: # Our copying is done. We just log and save _a, _a : List[str] = list(range(__a ) ), list(range(__a ) ) logger.info( f"""Copied encoder layers {e_layers_to_copy} and decoder layers {d_layers_to_copy}. Saving them to""" f""" {save_path}""" ) student.save_pretrained(__a ) return student, e_layers_to_copy, d_layers_to_copy # Decide which layers of the teacher to copy. Not exactly alternating -- we try to keep first and last layer. if e_layers_to_copy is None: _a : List[int] = pick_layers_to_copy(__a , __a ) if d_layers_to_copy is None: _a : List[int] = pick_layers_to_copy(__a , __a ) try: if hasattr( __a , 'prophetnet' ): # For ProphetNet, student.model.encoder.layers is called student.prophetnet.encoder.layers copy_layers(teacher.prophetnet.encoder.layers , student.prophetnet.encoder.layers , __a ) copy_layers(teacher.prophetnet.decoder.layers , student.prophetnet.decoder.layers , __a ) else: copy_layers(teacher.model.encoder.layers , student.model.encoder.layers , __a ) copy_layers(teacher.model.decoder.layers , student.model.decoder.layers , __a ) except AttributeError: # For t5, student.model.encoder.layers is called student.encoder.block copy_layers(teacher.encoder.block , student.encoder.block , __a ) copy_layers(teacher.decoder.block , student.decoder.block , __a ) logger.info( f"""Copied encoder layers {e_layers_to_copy} and decoder layers {d_layers_to_copy}. Saving them to {save_path}""" ) _a : Dict = { 'teacher_type': teacher.config.model_type, 'copied_encoder_layers': e_layers_to_copy, 'copied_decoder_layers': d_layers_to_copy, } student.save_pretrained(__a ) # Save information about copying for easier reproducibility return student, e_layers_to_copy, d_layers_to_copy if __name__ == "__main__": fire.Fire(create_student_by_copying_alternating_layers)
271
'''simple docstring''' import gzip import hashlib import json import multiprocessing import os import re import shutil import time from pathlib import Path import numpy as np from arguments import PreprocessingArguments from datasets import load_dataset from minhash_deduplication import deduplicate_dataset from transformers import AutoTokenizer, HfArgumentParser __lowerCAmelCase = re.compile(r"""\s+""") def UpperCAmelCase_ (__a : Any ): """simple docstring""" return {"hash": hashlib.mda(re.sub(__a , '' , example['content'] ).encode('utf-8' ) ).hexdigest()} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : List[str] = [len(__a ) for line in example['content'].splitlines()] return {"line_mean": np.mean(__a ), "line_max": max(__a )} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Union[str, Any] = np.mean([c.isalnum() for c in example['content']] ) return {"alpha_frac": alpha_frac} def UpperCAmelCase_ (__a : Optional[int] , __a : Any ): """simple docstring""" if example["hash"] in uniques: uniques.remove(example['hash'] ) return True else: return False def UpperCAmelCase_ (__a : int , __a : Union[str, Any]=5 ): """simple docstring""" _a : Optional[int] = ['auto-generated', 'autogenerated', 'automatically generated'] _a : List[str] = example['content'].splitlines() for _, line in zip(range(__a ) , __a ): for keyword in keywords: if keyword in line.lower(): return {"autogenerated": True} else: return {"autogenerated": False} def UpperCAmelCase_ (__a : List[str] , __a : Dict=5 , __a : Tuple=0.05 ): """simple docstring""" _a : Optional[int] = ['unit tests', 'test file', 'configuration file'] _a : int = example['content'].splitlines() _a : int = 0 _a : Dict = 0 # first test for _, line in zip(range(__a ) , __a ): for keyword in keywords: if keyword in line.lower(): return {"config_or_test": True} # second test _a : int = example['content'].count('\n' ) _a : int = int(coeff * nlines ) for line in lines: count_config += line.lower().count('config' ) count_test += line.lower().count('test' ) if count_config > threshold or count_test > threshold: return {"config_or_test": True} return {"config_or_test": False} def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" _a : List[str] = ['def ', 'class ', 'for ', 'while '] _a : str = example['content'].splitlines() for line in lines: for keyword in keywords: if keyword in line.lower(): return {"has_no_keywords": False} return {"has_no_keywords": True} def UpperCAmelCase_ (__a : int , __a : Any=4 ): """simple docstring""" _a : List[str] = example['content'].splitlines() _a : Dict = 0 for line in lines: counter += line.lower().count('=' ) if counter > minimum: return {"has_few_assignments": False} return {"has_few_assignments": True} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Optional[Any] = tokenizer(example['content'] , truncation=__a )['input_ids'] _a : Optional[int] = len(example['content'] ) / len(__a ) return {"ratio": ratio} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Dict = {} results.update(get_hash(__a ) ) results.update(line_stats(__a ) ) results.update(alpha_stats(__a ) ) results.update(char_token_ratio(__a ) ) results.update(is_autogenerated(__a ) ) results.update(is_config_or_test(__a ) ) results.update(has_no_keywords(__a ) ) results.update(has_few_assignments(__a ) ) return results def UpperCAmelCase_ (__a : Any , __a : Any , __a : str ): """simple docstring""" if not check_uniques(__a , __a ): return False elif example["autogenerated"]: return False elif example["line_max"] > args.line_max: return False elif example["line_mean"] > args.line_mean: return False elif example["alpha_frac"] < args.alpha_frac: return False elif example["ratio"] < args.min_token_ratio: return False elif example["config_or_test"] and np.random.rand() <= args.filter_proba: return False elif example["has_no_keywords"] and np.random.rand() <= args.filter_proba: return False elif example["has_few_assignments"]: return False else: return True def UpperCAmelCase_ (__a : Union[str, Any] ): """simple docstring""" with open(__a , 'rb' ) as f_in: with gzip.open(str(__a ) + '.gz' , 'wb' , compresslevel=6 ) as f_out: shutil.copyfileobj(__a , __a ) os.unlink(__a ) # Settings __lowerCAmelCase = HfArgumentParser(PreprocessingArguments) __lowerCAmelCase = parser.parse_args() if args.num_workers is None: __lowerCAmelCase = multiprocessing.cpu_count() __lowerCAmelCase = AutoTokenizer.from_pretrained(args.tokenizer_dir) # Load dataset __lowerCAmelCase = time.time() __lowerCAmelCase = load_dataset(args.dataset_name, split="""train""") print(f'''Time to load dataset: {time.time()-t_start:.2f}''') # Run preprocessing __lowerCAmelCase = time.time() __lowerCAmelCase = ds.map(preprocess, num_proc=args.num_workers) print(f'''Time to preprocess dataset: {time.time()-t_start:.2f}''') # Deduplicate hashes __lowerCAmelCase = set(ds.unique("""hash""")) __lowerCAmelCase = len(uniques) / len(ds) print(f'''Fraction of duplicates: {1-frac:.2%}''') # Deduplicate data and apply heuristics __lowerCAmelCase = time.time() __lowerCAmelCase = ds.filter(filter, fn_kwargs={"""uniques""": uniques, """args""": args}) print(f'''Time to filter dataset: {time.time()-t_start:.2f}''') print(f'''Size of filtered dataset: {len(ds_filter)}''') # Deduplicate with minhash and jaccard similarity if args.near_deduplication: __lowerCAmelCase = time.time() __lowerCAmelCase , __lowerCAmelCase = deduplicate_dataset(ds_filter, args.jaccard_threshold) print(f'''Time to deduplicate dataset: {time.time()-t_start:.2f}''') print(f'''Size of deduplicate dataset: {len(ds_filter)}''') # Save data in batches of samples_per_file __lowerCAmelCase = Path(args.output_dir) output_dir.mkdir(exist_ok=True) # save duplicate_clusters in the output_dir as artifacts # not sure it is the right place the save it if args.near_deduplication: with open(output_dir / """duplicate_clusters.json""", """w""") as f: json.dump(duplicate_clusters, f) __lowerCAmelCase = output_dir / """data""" data_dir.mkdir(exist_ok=True) __lowerCAmelCase = time.time() for file_number, index in enumerate(range(0, len(ds_filter), args.samples_per_file)): __lowerCAmelCase = str(data_dir / f'''file-{file_number+1:012}.json''') __lowerCAmelCase = min(len(ds_filter), index + args.samples_per_file) ds_filter.select(list(range(index, end_index))).to_json(file_path) compress_file(file_path) print(f'''Time to save dataset: {time.time()-t_start:.2f}''')
271
1
'''simple docstring''' import unittest from transformers import PegasusTokenizer, PegasusTokenizerFast from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow from transformers.utils import cached_property from ...test_tokenization_common import TokenizerTesterMixin __lowerCAmelCase = get_tests_dir("""fixtures/test_sentencepiece_no_bos.model""") @require_sentencepiece @require_tokenizers class UpperCAmelCase__ ( lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[Any] = PegasusTokenizer __UpperCAmelCase : int = PegasusTokenizerFast __UpperCAmelCase : Tuple = True __UpperCAmelCase : Optional[Any] = True def __lowercase ( self : Any ): '''simple docstring''' super().setUp() # We have a SentencePiece fixture for testing _a : Optional[int] = PegasusTokenizer(_a ) tokenizer.save_pretrained(self.tmpdirname ) @cached_property def __lowercase ( self : Optional[int] ): '''simple docstring''' return PegasusTokenizer.from_pretrained('google/pegasus-large' ) def __lowercase ( self : str ,**_a : str ): '''simple docstring''' return PegasusTokenizer.from_pretrained(self.tmpdirname ,**_a ) def __lowercase ( self : Union[str, Any] ,_a : Dict ): '''simple docstring''' return ("This is a test", "This is a test") def __lowercase ( self : List[Any] ): '''simple docstring''' _a : Any = '</s>' _a : Optional[int] = 1 self.assertEqual(self.get_tokenizer()._convert_token_to_id(_a ) ,_a ) self.assertEqual(self.get_tokenizer()._convert_id_to_token(_a ) ,_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[str] = list(self.get_tokenizer().get_vocab().keys() ) self.assertEqual(vocab_keys[0] ,'<pad>' ) self.assertEqual(vocab_keys[1] ,'</s>' ) self.assertEqual(vocab_keys[-1] ,'v' ) self.assertEqual(len(_a ) ,1103 ) def __lowercase ( self : Dict ): '''simple docstring''' self.assertEqual(self.get_tokenizer().vocab_size ,1103 ) def __lowercase ( self : Any ): '''simple docstring''' _a : Union[str, Any] = self.rust_tokenizer_class.from_pretrained(self.tmpdirname ) _a : Tuple = self.tokenizer_class.from_pretrained(self.tmpdirname ) _a : Union[str, Any] = ( 'Let\'s see which <unk> is the better <unk_token_11> one <mask_1> It seems like this <mask_2> was important' ' </s> <pad> <pad> <pad>' ) _a : Any = rust_tokenizer([raw_input_str] ,return_tensors=_a ,add_special_tokens=_a ).input_ids[0] _a : List[str] = py_tokenizer([raw_input_str] ,return_tensors=_a ,add_special_tokens=_a ).input_ids[0] self.assertListEqual(_a ,_a ) def __lowercase ( self : str ): '''simple docstring''' _a : Union[str, Any] = self._large_tokenizer # <mask_1> masks whole sentence while <mask_2> masks single word _a : List[Any] = '<mask_1> To ensure a <mask_2> flow of bank resolutions.' _a : List[Any] = [2, 413, 615, 114, 3, 1971, 113, 1679, 1_0710, 107, 1] _a : Union[str, Any] = tokenizer([raw_input_str] ,return_tensors=_a ).input_ids[0] self.assertListEqual(_a ,_a ) def __lowercase ( self : Any ): '''simple docstring''' _a : str = self._large_tokenizer # The tracebacks for the following asserts are **better** without messages or self.assertEqual assert tokenizer.vocab_size == 9_6103 assert tokenizer.pad_token_id == 0 assert tokenizer.eos_token_id == 1 assert tokenizer.offset == 103 assert tokenizer.unk_token_id == tokenizer.offset + 2 == 105 assert tokenizer.unk_token == "<unk>" assert tokenizer.model_max_length == 1024 _a : Any = 'To ensure a smooth flow of bank resolutions.' _a : int = [413, 615, 114, 2291, 1971, 113, 1679, 1_0710, 107, 1] _a : Union[str, Any] = tokenizer([raw_input_str] ,return_tensors=_a ).input_ids[0] self.assertListEqual(_a ,_a ) assert tokenizer.convert_ids_to_tokens([0, 1, 2, 3] ) == ["<pad>", "</s>", "<mask_1>", "<mask_2>"] @require_torch def __lowercase ( self : str ): '''simple docstring''' _a : Tuple = ['This is going to be way too long.' * 150, 'short example'] _a : List[str] = ['not super long but more than 5 tokens', 'tiny'] _a : str = self._large_tokenizer(_a ,padding=_a ,truncation=_a ,return_tensors='pt' ) _a : int = self._large_tokenizer( text_target=_a ,max_length=5 ,padding=_a ,truncation=_a ,return_tensors='pt' ) assert batch.input_ids.shape == (2, 1024) assert batch.attention_mask.shape == (2, 1024) assert targets["input_ids"].shape == (2, 5) assert len(_a ) == 2 # input_ids, attention_mask. @slow def __lowercase ( self : List[Any] ): '''simple docstring''' _a : Tuple = {'input_ids': [[3_8979, 143, 1_8485, 606, 130, 2_6669, 8_7686, 121, 5_4189, 1129, 111, 2_6669, 8_7686, 121, 9114, 1_4787, 121, 1_3249, 158, 592, 956, 121, 1_4621, 3_1576, 143, 6_2613, 108, 9688, 930, 4_3430, 1_1562, 6_2613, 304, 108, 1_1443, 897, 108, 9314, 1_7415, 6_3399, 108, 1_1443, 7614, 1_8316, 118, 4284, 7148, 1_2430, 143, 1400, 2_5703, 158, 111, 4284, 7148, 1_1772, 143, 2_1297, 1064, 158, 122, 204, 3506, 1754, 1133, 1_4787, 1581, 115, 3_3224, 4482, 111, 1355, 110, 2_9173, 317, 5_0833, 108, 2_0147, 9_4665, 111, 7_7198, 107, 1], [110, 6_2613, 117, 638, 112, 1133, 121, 2_0098, 1355, 7_9050, 1_3872, 135, 1596, 5_3541, 1352, 141, 1_3039, 5542, 124, 302, 518, 111, 268, 2956, 115, 149, 4427, 107, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [139, 1235, 2799, 1_8289, 1_7780, 204, 109, 9474, 1296, 107, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], 'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501 # fmt: on self.tokenizer_integration_test_util( expected_encoding=_a ,model_name='google/bigbird-pegasus-large-arxiv' ,revision='ba85d0851d708441f91440d509690f1ab6353415' ,) @require_sentencepiece @require_tokenizers class UpperCAmelCase__ ( lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : List[str] = PegasusTokenizer __UpperCAmelCase : Optional[Any] = PegasusTokenizerFast __UpperCAmelCase : List[Any] = True __UpperCAmelCase : Any = True def __lowercase ( self : List[Any] ): '''simple docstring''' super().setUp() # We have a SentencePiece fixture for testing _a : Tuple = PegasusTokenizer(_a ,offset=0 ,mask_token_sent=_a ,mask_token='[MASK]' ) tokenizer.save_pretrained(self.tmpdirname ) @cached_property def __lowercase ( self : List[Any] ): '''simple docstring''' return PegasusTokenizer.from_pretrained('google/bigbird-pegasus-large-arxiv' ) def __lowercase ( self : List[Any] ,**_a : Optional[Any] ): '''simple docstring''' return PegasusTokenizer.from_pretrained(self.tmpdirname ,**_a ) def __lowercase ( self : Dict ,_a : Union[str, Any] ): '''simple docstring''' return ("This is a test", "This is a test") def __lowercase ( self : List[Any] ): '''simple docstring''' _a : str = self.rust_tokenizer_class.from_pretrained(self.tmpdirname ) _a : List[str] = self.tokenizer_class.from_pretrained(self.tmpdirname ) _a : Optional[int] = ( 'Let\'s see which <unk> is the better <unk_token> one [MASK] It seems like this [MASK] was important </s>' ' <pad> <pad> <pad>' ) _a : int = rust_tokenizer([raw_input_str] ,return_tensors=_a ,add_special_tokens=_a ).input_ids[0] _a : List[Any] = py_tokenizer([raw_input_str] ,return_tensors=_a ,add_special_tokens=_a ).input_ids[0] self.assertListEqual(_a ,_a ) @require_torch def __lowercase ( self : List[Any] ): '''simple docstring''' _a : Tuple = ['This is going to be way too long.' * 1000, 'short example'] _a : Optional[Any] = ['not super long but more than 5 tokens', 'tiny'] _a : int = self._large_tokenizer(_a ,padding=_a ,truncation=_a ,return_tensors='pt' ) _a : str = self._large_tokenizer( text_target=_a ,max_length=5 ,padding=_a ,truncation=_a ,return_tensors='pt' ) assert batch.input_ids.shape == (2, 4096) assert batch.attention_mask.shape == (2, 4096) assert targets["input_ids"].shape == (2, 5) assert len(_a ) == 2 # input_ids, attention_mask. def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Optional[int] = ( 'This is an example string that is used to test the original TF implementation against the HF' ' implementation' ) _a : Optional[Any] = self._large_tokenizer(_a ).input_ids self.assertListEqual( _a ,[182, 117, 142, 587, 4211, 120, 117, 263, 112, 804, 109, 856, 2_5016, 3137, 464, 109, 2_6955, 3137, 1] ,)
271
'''simple docstring''' import argparse from typing import List import evaluate import numpy as np import torch from datasets import DatasetDict, load_dataset # New Code # # We'll be using StratifiedKFold for this example from sklearn.model_selection import StratifiedKFold from torch.optim import AdamW from torch.utils.data import DataLoader from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed from accelerate import Accelerator, DistributedType ######################################################################## # This is a fully working simple example to use Accelerate, # specifically showcasing how to perform Cross Validation, # and builds off the `nlp_example.py` script. # # This example trains a Bert base model on GLUE MRPC # in any of the following settings (with the same script): # - single CPU or single GPU # - multi GPUS (using PyTorch distributed mode) # - (multi) TPUs # - fp16 (mixed-precision) or fp32 (normal precision) # # To help focus on the differences in the code, building `DataLoaders` # was refactored into its own function. # New additions from the base script can be found quickly by # looking for the # New Code # tags # # To run it in each of these various modes, follow the instructions # in the readme for examples: # https://github.com/huggingface/accelerate/tree/main/examples # ######################################################################## __lowerCAmelCase = 1_6 __lowerCAmelCase = 3_2 def UpperCAmelCase_ (__a : Accelerator , __a : DatasetDict , __a : List[int] , __a : List[int] , __a : int = 1_6 ): """simple docstring""" _a : Union[str, Any] = AutoTokenizer.from_pretrained('bert-base-cased' ) _a : str = DatasetDict( { 'train': dataset['train'].select(__a ), 'validation': dataset['train'].select(__a ), 'test': dataset['validation'], } ) def tokenize_function(__a : List[Any] ): # max_length=None => use the model max length (it's actually the default) _a : Optional[int] = tokenizer(examples['sentence1'] , examples['sentence2'] , truncation=__a , max_length=__a ) return outputs # Apply the method we just defined to all the examples in all the splits of the dataset # starting with the main process first: with accelerator.main_process_first(): _a : List[str] = datasets.map( __a , batched=__a , remove_columns=['idx', 'sentence1', 'sentence2'] , ) # We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the # transformers library _a : List[Any] = tokenized_datasets.rename_column('label' , 'labels' ) def collate_fn(__a : int ): # On TPU it's best to pad everything to the same length or training will be very slow. _a : Dict = 1_2_8 if accelerator.distributed_type == DistributedType.TPU else None # When using mixed precision we want round multiples of 8/16 if accelerator.mixed_precision == "fp8": _a : Tuple = 1_6 elif accelerator.mixed_precision != "no": _a : List[Any] = 8 else: _a : List[Any] = None return tokenizer.pad( __a , padding='longest' , max_length=__a , pad_to_multiple_of=__a , return_tensors='pt' , ) # Instantiate dataloaders. _a : Any = DataLoader( tokenized_datasets['train'] , shuffle=__a , collate_fn=__a , batch_size=__a ) _a : Optional[int] = DataLoader( tokenized_datasets['validation'] , shuffle=__a , collate_fn=__a , batch_size=__a ) _a : Optional[Any] = DataLoader( tokenized_datasets['test'] , shuffle=__a , collate_fn=__a , batch_size=__a ) return train_dataloader, eval_dataloader, test_dataloader def UpperCAmelCase_ (__a : Any , __a : Union[str, Any] ): """simple docstring""" _a : Dict = [] # Download the dataset _a : Tuple = load_dataset('glue' , 'mrpc' ) # Create our splits _a : Union[str, Any] = StratifiedKFold(n_splits=int(args.num_folds ) ) # Initialize accelerator _a : Any = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision ) # Sample hyper-parameters for learning rate, batch size, seed and a few other HPs _a : Optional[Any] = config['lr'] _a : Optional[int] = int(config['num_epochs'] ) _a : Dict = int(config['seed'] ) _a : Dict = int(config['batch_size'] ) _a : Optional[int] = evaluate.load('glue' , 'mrpc' ) # If the batch size is too big we use gradient accumulation _a : List[Any] = 1 if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU: _a : Any = batch_size // MAX_GPU_BATCH_SIZE _a : List[str] = MAX_GPU_BATCH_SIZE set_seed(__a ) # New Code # # Create our folds: _a : int = kfold.split(np.zeros(datasets['train'].num_rows ) , datasets['train']['label'] ) _a : Any = [] # Iterate over them for i, (train_idxs, valid_idxs) in enumerate(__a ): _a, _a, _a : Optional[Any] = get_fold_dataloaders( __a , __a , __a , __a , ) # Instantiate the model (we build the model here so that the seed also control new weights initialization) _a : Dict = AutoModelForSequenceClassification.from_pretrained('bert-base-cased' , return_dict=__a ) # We could avoid this line since the accelerator is set with `device_placement=True` (default value). # Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer # creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that). _a : List[Any] = model.to(accelerator.device ) # Instantiate optimizer _a : List[str] = AdamW(params=model.parameters() , lr=__a ) # Instantiate scheduler _a : List[Any] = get_linear_schedule_with_warmup( optimizer=__a , num_warmup_steps=1_0_0 , num_training_steps=(len(__a ) * num_epochs) // gradient_accumulation_steps , ) # Prepare everything # There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the # prepare method. _a, _a, _a, _a, _a : Union[str, Any] = accelerator.prepare( __a , __a , __a , __a , __a ) # Now we train the model for epoch in range(__a ): model.train() for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) _a : Dict = model(**__a ) _a : int = outputs.loss _a : Any = loss / gradient_accumulation_steps accelerator.backward(__a ) if step % gradient_accumulation_steps == 0: optimizer.step() lr_scheduler.step() optimizer.zero_grad() model.eval() for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): _a : Union[str, Any] = model(**__a ) _a : Tuple = outputs.logits.argmax(dim=-1 ) _a, _a : Any = accelerator.gather_for_metrics((predictions, batch['labels']) ) metric.add_batch( predictions=__a , references=__a , ) _a : List[Any] = metric.compute() # Use accelerator.print to print only on the main process. accelerator.print(f"""epoch {epoch}:""" , __a ) # New Code # # We also run predictions on the test set at the very end _a : Any = [] for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): _a : Tuple = model(**__a ) _a : Dict = outputs.logits _a, _a : Optional[int] = accelerator.gather_for_metrics((predictions, batch['labels']) ) fold_predictions.append(predictions.cpu() ) if i == 0: # We need all of the test predictions test_references.append(references.cpu() ) # Use accelerator.print to print only on the main process. test_predictions.append(torch.cat(__a , dim=0 ) ) # We now need to release all our memory and get rid of the current model, optimizer, etc accelerator.free_memory() # New Code # # Finally we check the accuracy of our folded results: _a : Dict = torch.cat(__a , dim=0 ) _a : Any = torch.stack(__a , dim=0 ).sum(dim=0 ).div(int(args.num_folds ) ).argmax(dim=-1 ) _a : str = metric.compute(predictions=__a , references=__a ) accelerator.print('Average test metrics from all folds:' , __a ) def UpperCAmelCase_ (): """simple docstring""" _a : Any = argparse.ArgumentParser(description='Simple example of training script.' ) parser.add_argument( '--mixed_precision' , type=__a , default=__a , choices=['no', 'fp16', 'bf16', 'fp8'] , help='Whether to use mixed precision. Choose' 'between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.' 'and an Nvidia Ampere GPU.' , ) parser.add_argument('--cpu' , action='store_true' , help='If passed, will train on the CPU.' ) # New Code # parser.add_argument('--num_folds' , type=__a , default=3 , help='The number of splits to perform across the dataset' ) _a : Any = parser.parse_args() _a : int = {'lr': 2e-5, 'num_epochs': 3, 'seed': 4_2, 'batch_size': 1_6} training_function(__a , __a ) if __name__ == "__main__": main()
271
1
'''simple docstring''' from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available __lowerCAmelCase = {"""tokenization_herbert""": ["""HerbertTokenizer"""]} try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""HerbertTokenizerFast"""] if TYPE_CHECKING: from .tokenization_herbert import HerbertTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_herbert_fast import HerbertTokenizerFast else: import sys __lowerCAmelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
271
'''simple docstring''' from __future__ import annotations __lowerCAmelCase = [-1_0, -5, 0, 5, 5.1, 1_1, 1_3, 2_1, 3, 4, -2_1, -1_0, -5, -1, 0] __lowerCAmelCase = [-5, 0, 5, 5.1, 1_1, 1_3, 2_1, -1, 4, -1, -1_0, -5, -1, 0, -1] def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : Optional[int] = [] _a : int = len(__a ) for i in range(__a ): _a : float = -1 for j in range(i + 1 , __a ): if arr[i] < arr[j]: _a : Any = arr[j] break result.append(__a ) return result def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : Tuple = [] for i, outer in enumerate(__a ): _a : float = -1 for inner in arr[i + 1 :]: if outer < inner: _a : Dict = inner break result.append(__a ) return result def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : int = len(__a ) _a : list[float] = [] _a : list[float] = [-1] * arr_size for index in reversed(range(__a ) ): if stack: while stack[-1] <= arr[index]: stack.pop() if not stack: break if stack: _a : Dict = stack[-1] stack.append(arr[index] ) return result if __name__ == "__main__": from doctest import testmod from timeit import timeit testmod() print(next_greatest_element_slow(arr)) print(next_greatest_element_fast(arr)) print(next_greatest_element(arr)) __lowerCAmelCase = ( """from __main__ import arr, next_greatest_element_slow, """ """next_greatest_element_fast, next_greatest_element""" ) print( """next_greatest_element_slow():""", timeit("""next_greatest_element_slow(arr)""", setup=setup), ) print( """next_greatest_element_fast():""", timeit("""next_greatest_element_fast(arr)""", setup=setup), ) print( """ next_greatest_element():""", timeit("""next_greatest_element(arr)""", setup=setup), )
271
1
'''simple docstring''' import multiprocessing import os from typing import BinaryIO, Optional, Union import fsspec from .. import Dataset, Features, NamedSplit, config from ..formatting import query_table from ..packaged_modules.json.json import Json from ..utils import logging from ..utils.typing import NestedDataStructureLike, PathLike from .abc import AbstractDatasetReader class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Tuple ,_a : NestedDataStructureLike[PathLike] ,_a : Optional[NamedSplit] = None ,_a : Optional[Features] = None ,_a : str = None ,_a : bool = False ,_a : bool = False ,_a : Optional[str] = None ,_a : Optional[int] = None ,**_a : Union[str, Any] ,): '''simple docstring''' super().__init__( _a ,split=_a ,features=_a ,cache_dir=_a ,keep_in_memory=_a ,streaming=_a ,num_proc=_a ,**_a ,) _a : List[Any] = field _a : List[Any] = path_or_paths if isinstance(_a ,_a ) else {self.split: path_or_paths} _a : int = Json( cache_dir=_a ,data_files=_a ,features=_a ,field=_a ,**_a ,) def __lowercase ( self : str ): '''simple docstring''' if self.streaming: _a : Tuple = self.builder.as_streaming_dataset(split=self.split ) # Build regular (map-style) dataset else: _a : Any = None _a : Optional[Any] = None _a : Dict = None _a : Dict = None self.builder.download_and_prepare( download_config=_a ,download_mode=_a ,verification_mode=_a ,base_path=_a ,num_proc=self.num_proc ,) _a : List[Any] = self.builder.as_dataset( split=self.split ,verification_mode=_a ,in_memory=self.keep_in_memory ) return dataset class UpperCAmelCase__ : """simple docstring""" def __init__( self : Any ,_a : Dataset ,_a : Union[PathLike, BinaryIO] ,_a : Optional[int] = None ,_a : Optional[int] = None ,**_a : Any ,): '''simple docstring''' if num_proc is not None and num_proc <= 0: raise ValueError(F"""num_proc {num_proc} must be an integer > 0.""" ) _a : Union[str, Any] = dataset _a : Optional[Any] = path_or_buf _a : Optional[int] = batch_size if batch_size else config.DEFAULT_MAX_BATCH_SIZE _a : Any = num_proc _a : Optional[Any] = 'utf-8' _a : Any = to_json_kwargs def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : List[str] = self.to_json_kwargs.pop('path_or_buf' ,_a ) _a : List[str] = self.to_json_kwargs.pop('orient' ,'records' ) _a : Dict = self.to_json_kwargs.pop('lines' ,True if orient == 'records' else False ) _a : Optional[Any] = self.to_json_kwargs.pop('index' ,False if orient in ['split', 'table'] else True ) _a : Union[str, Any] = self.to_json_kwargs.pop('compression' ,_a ) if compression not in [None, "infer", "gzip", "bz2", "xz"]: raise NotImplementedError(F"""`datasets` currently does not support {compression} compression""" ) if isinstance(self.path_or_buf ,(str, bytes, os.PathLike) ): with fsspec.open(self.path_or_buf ,'wb' ,compression=_a ) as buffer: _a : str = self._write(file_obj=_a ,orient=_a ,lines=_a ,index=_a ,**self.to_json_kwargs ) else: if compression: raise NotImplementedError( F"""The compression parameter is not supported when writing to a buffer, but compression={compression}""" ' was passed. Please provide a local path instead.' ) _a : List[Any] = self._write( file_obj=self.path_or_buf ,orient=_a ,lines=_a ,index=_a ,**self.to_json_kwargs ) return written def __lowercase ( self : Dict ,_a : Optional[Any] ): '''simple docstring''' _a, _a, _a, _a, _a : Optional[Any] = args _a : Any = query_table( table=self.dataset.data ,key=slice(_a ,offset + self.batch_size ) ,indices=self.dataset._indices ,) _a : List[Any] = batch.to_pandas().to_json( path_or_buf=_a ,orient=_a ,lines=_a ,index=_a ,**_a ) if not json_str.endswith('\n' ): json_str += "\n" return json_str.encode(self.encoding ) def __lowercase ( self : Any ,_a : BinaryIO ,_a : Optional[int] ,_a : Any ,_a : Union[str, Any] ,**_a : str ,): '''simple docstring''' _a : Dict = 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 json from Arrow format' ,): _a : Any = self._batch_json((offset, orient, lines, index, to_json_kwargs) ) written += file_obj.write(_a ) else: _a, _a : Optional[int] = len(self.dataset ), self.batch_size with multiprocessing.Pool(self.num_proc ) as pool: for json_str in logging.tqdm( pool.imap( self._batch_json ,[(offset, orient, lines, index, to_json_kwargs) for offset in range(0 ,_a ,_a )] ,) ,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 json from Arrow format' ,): written += file_obj.write(_a ) return written
271
'''simple docstring''' import os from huggingface_hub.constants import HUGGINGFACE_HUB_CACHE, hf_cache_home __lowerCAmelCase = HUGGINGFACE_HUB_CACHE __lowerCAmelCase = """config.json""" __lowerCAmelCase = """diffusion_pytorch_model.bin""" __lowerCAmelCase = """diffusion_flax_model.msgpack""" __lowerCAmelCase = """model.onnx""" __lowerCAmelCase = """diffusion_pytorch_model.safetensors""" __lowerCAmelCase = """weights.pb""" __lowerCAmelCase = """https://huggingface.co""" __lowerCAmelCase = default_cache_path __lowerCAmelCase = """diffusers_modules""" __lowerCAmelCase = os.getenv("""HF_MODULES_CACHE""", os.path.join(hf_cache_home, """modules""")) __lowerCAmelCase = ["""fp16""", """non-ema"""] __lowerCAmelCase = """.self_attn"""
271
1
'''simple docstring''' import logging import os from dataclasses import dataclass from enum import Enum from typing import List, Optional, Union from filelock import FileLock from transformers import PreTrainedTokenizer, is_tf_available, is_torch_available __lowerCAmelCase = logging.getLogger(__name__) @dataclass class UpperCAmelCase__ : """simple docstring""" __UpperCAmelCase : str __UpperCAmelCase : List[str] __UpperCAmelCase : Optional[List[str]] @dataclass class UpperCAmelCase__ : """simple docstring""" __UpperCAmelCase : List[int] __UpperCAmelCase : List[int] __UpperCAmelCase : Optional[List[int]] = None __UpperCAmelCase : Optional[List[int]] = None class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Dict = '''train''' __UpperCAmelCase : Optional[int] = '''dev''' __UpperCAmelCase : Optional[Any] = '''test''' class UpperCAmelCase__ : """simple docstring""" @staticmethod def __lowercase ( _a : int ,_a : Union[Split, str] ): '''simple docstring''' raise NotImplementedError @staticmethod def __lowercase ( _a : str ): '''simple docstring''' raise NotImplementedError @staticmethod def __lowercase ( _a : List[InputExample] ,_a : List[str] ,_a : int ,_a : PreTrainedTokenizer ,_a : List[str]=False ,_a : Tuple="[CLS]" ,_a : Optional[int]=1 ,_a : List[Any]="[SEP]" ,_a : List[str]=False ,_a : Optional[int]=False ,_a : Optional[Any]=0 ,_a : List[str]=0 ,_a : Tuple=-100 ,_a : Optional[int]=0 ,_a : int=True ,): '''simple docstring''' _a : Optional[Any] = {label: i for i, label in enumerate(_a )} _a : str = [] for ex_index, example in enumerate(_a ): if ex_index % 1_0000 == 0: logger.info('Writing example %d of %d' ,_a ,len(_a ) ) _a : List[Any] = [] _a : List[str] = [] for word, label in zip(example.words ,example.labels ): _a : Dict = tokenizer.tokenize(_a ) # bert-base-multilingual-cased sometimes output "nothing ([]) when calling tokenize with just a space. if len(_a ) > 0: tokens.extend(_a ) # Use the real label id for the first token of the word, and padding ids for the remaining tokens label_ids.extend([label_map[label]] + [pad_token_label_id] * (len(_a ) - 1) ) # Account for [CLS] and [SEP] with "- 2" and with "- 3" for RoBERTa. _a : Dict = tokenizer.num_special_tokens_to_add() if len(_a ) > max_seq_length - special_tokens_count: _a : int = tokens[: (max_seq_length - special_tokens_count)] _a : Optional[Any] = label_ids[: (max_seq_length - special_tokens_count)] # The convention in BERT is: # (a) For sequence pairs: # tokens: [CLS] is this jack ##son ##ville ? [SEP] no it is not . [SEP] # type_ids: 0 0 0 0 0 0 0 0 1 1 1 1 1 1 # (b) For single sequences: # tokens: [CLS] the dog is hairy . [SEP] # type_ids: 0 0 0 0 0 0 0 # # Where "type_ids" are used to indicate whether this is the first # sequence or the second sequence. The embedding vectors for `type=0` and # `type=1` were learned during pre-training and are added to the wordpiece # embedding vector (and position vector). This is not *strictly* necessary # since the [SEP] token unambiguously separates the sequences, but it makes # it easier for the model to learn the concept of sequences. # # For classification tasks, the first vector (corresponding to [CLS]) is # used as the "sentence vector". Note that this only makes sense because # the entire model is fine-tuned. tokens += [sep_token] label_ids += [pad_token_label_id] if sep_token_extra: # roberta uses an extra separator b/w pairs of sentences tokens += [sep_token] label_ids += [pad_token_label_id] _a : Optional[Any] = [sequence_a_segment_id] * len(_a ) if cls_token_at_end: tokens += [cls_token] label_ids += [pad_token_label_id] segment_ids += [cls_token_segment_id] else: _a : str = [cls_token] + tokens _a : List[Any] = [pad_token_label_id] + label_ids _a : Optional[int] = [cls_token_segment_id] + segment_ids _a : List[Any] = tokenizer.convert_tokens_to_ids(_a ) # The mask has 1 for real tokens and 0 for padding tokens. Only real # tokens are attended to. _a : Optional[int] = [1 if mask_padding_with_zero else 0] * len(_a ) # Zero-pad up to the sequence length. _a : Tuple = max_seq_length - len(_a ) if pad_on_left: _a : Dict = ([pad_token] * padding_length) + input_ids _a : List[str] = ([0 if mask_padding_with_zero else 1] * padding_length) + input_mask _a : int = ([pad_token_segment_id] * padding_length) + segment_ids _a : List[Any] = ([pad_token_label_id] * padding_length) + label_ids else: input_ids += [pad_token] * padding_length input_mask += [0 if mask_padding_with_zero else 1] * padding_length segment_ids += [pad_token_segment_id] * padding_length label_ids += [pad_token_label_id] * padding_length assert len(_a ) == max_seq_length assert len(_a ) == max_seq_length assert len(_a ) == max_seq_length assert len(_a ) == max_seq_length if ex_index < 5: logger.info('*** Example ***' ) logger.info('guid: %s' ,example.guid ) logger.info('tokens: %s' ,' '.join([str(_a ) for x in tokens] ) ) logger.info('input_ids: %s' ,' '.join([str(_a ) for x in input_ids] ) ) logger.info('input_mask: %s' ,' '.join([str(_a ) for x in input_mask] ) ) logger.info('segment_ids: %s' ,' '.join([str(_a ) for x in segment_ids] ) ) logger.info('label_ids: %s' ,' '.join([str(_a ) for x in label_ids] ) ) if "token_type_ids" not in tokenizer.model_input_names: _a : Tuple = None features.append( InputFeatures( input_ids=_a ,attention_mask=_a ,token_type_ids=_a ,label_ids=_a ) ) return features if is_torch_available(): import torch from torch import nn from torch.utils.data import Dataset class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : List[InputFeatures] __UpperCAmelCase : int = nn.CrossEntropyLoss().ignore_index def __init__( self : int ,_a : TokenClassificationTask ,_a : str ,_a : PreTrainedTokenizer ,_a : List[str] ,_a : str ,_a : Optional[int] = None ,_a : Any=False ,_a : Split = Split.train ,): '''simple docstring''' _a : Optional[int] = os.path.join( _a ,'cached_{}_{}_{}'.format(mode.value ,tokenizer.__class__.__name__ ,str(_a ) ) ,) # Make sure only the first process in distributed training processes the dataset, # and the others will use the cache. _a : Dict = cached_features_file + '.lock' with FileLock(_a ): if os.path.exists(_a ) and not overwrite_cache: logger.info(F"""Loading features from cached file {cached_features_file}""" ) _a : Optional[Any] = torch.load(_a ) else: logger.info(F"""Creating features from dataset file at {data_dir}""" ) _a : Tuple = token_classification_task.read_examples_from_file(_a ,_a ) # TODO clean up all this to leverage built-in features of tokenizers _a : Dict = token_classification_task.convert_examples_to_features( _a ,_a ,_a ,_a ,cls_token_at_end=bool(model_type in ['xlnet'] ) ,cls_token=tokenizer.cls_token ,cls_token_segment_id=2 if model_type in ['xlnet'] else 0 ,sep_token=tokenizer.sep_token ,sep_token_extra=_a ,pad_on_left=bool(tokenizer.padding_side == 'left' ) ,pad_token=tokenizer.pad_token_id ,pad_token_segment_id=tokenizer.pad_token_type_id ,pad_token_label_id=self.pad_token_label_id ,) logger.info(F"""Saving features into cached file {cached_features_file}""" ) torch.save(self.features ,_a ) def __len__( self : List[Any] ): '''simple docstring''' return len(self.features ) def __getitem__( self : Tuple ,_a : Any ): '''simple docstring''' return self.features[i] if is_tf_available(): import tensorflow as tf class UpperCAmelCase__ : """simple docstring""" __UpperCAmelCase : List[InputFeatures] __UpperCAmelCase : int = -100 def __init__( self : Tuple ,_a : TokenClassificationTask ,_a : str ,_a : PreTrainedTokenizer ,_a : List[str] ,_a : str ,_a : Optional[int] = None ,_a : Union[str, Any]=False ,_a : Split = Split.train ,): '''simple docstring''' _a : Optional[int] = token_classification_task.read_examples_from_file(_a ,_a ) # TODO clean up all this to leverage built-in features of tokenizers _a : str = token_classification_task.convert_examples_to_features( _a ,_a ,_a ,_a ,cls_token_at_end=bool(model_type in ['xlnet'] ) ,cls_token=tokenizer.cls_token ,cls_token_segment_id=2 if model_type in ['xlnet'] else 0 ,sep_token=tokenizer.sep_token ,sep_token_extra=_a ,pad_on_left=bool(tokenizer.padding_side == 'left' ) ,pad_token=tokenizer.pad_token_id ,pad_token_segment_id=tokenizer.pad_token_type_id ,pad_token_label_id=self.pad_token_label_id ,) def gen(): for ex in self.features: if ex.token_type_ids is None: yield ( {"input_ids": ex.input_ids, "attention_mask": ex.attention_mask}, ex.label_ids, ) else: yield ( { "input_ids": ex.input_ids, "attention_mask": ex.attention_mask, "token_type_ids": ex.token_type_ids, }, ex.label_ids, ) if "token_type_ids" not in tokenizer.model_input_names: _a : Union[str, Any] = tf.data.Dataset.from_generator( _a ,({'input_ids': tf.intaa, 'attention_mask': tf.intaa}, tf.intaa) ,( {'input_ids': tf.TensorShape([None] ), 'attention_mask': tf.TensorShape([None] )}, tf.TensorShape([None] ), ) ,) else: _a : int = tf.data.Dataset.from_generator( _a ,({'input_ids': tf.intaa, 'attention_mask': tf.intaa, 'token_type_ids': tf.intaa}, tf.intaa) ,( { 'input_ids': tf.TensorShape([None] ), 'attention_mask': tf.TensorShape([None] ), 'token_type_ids': tf.TensorShape([None] ), }, tf.TensorShape([None] ), ) ,) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.dataset.apply(tf.data.experimental.assert_cardinality(len(self.features ) ) ) return self.dataset def __len__( self : Optional[int] ): '''simple docstring''' return len(self.features ) def __getitem__( self : Any ,_a : Any ): '''simple docstring''' return self.features[i]
271
'''simple docstring''' import inspect import unittest import numpy as np from tests.test_modeling_common import floats_tensor from transformers import MaskaFormerConfig, is_torch_available, is_vision_available from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device from transformers.utils import cached_property from ...test_configuration_common import ConfigTester from ...test_modeling_common import ModelTesterMixin from ...test_pipeline_mixin import PipelineTesterMixin if is_torch_available(): import torch from transformers import MaskaFormerForUniversalSegmentation, MaskaFormerModel if is_vision_available(): from transformers import MaskaFormerImageProcessor if is_vision_available(): from PIL import Image class UpperCAmelCase__ : """simple docstring""" def __init__( self : int ,_a : Any ,_a : Optional[int]=2 ,_a : Optional[Any]=True ,_a : Dict=False ,_a : Dict=10 ,_a : Any=3 ,_a : str=32 * 8 ,_a : Optional[int]=32 * 8 ,_a : int=4 ,_a : str=64 ,): '''simple docstring''' _a : Dict = parent _a : Union[str, Any] = batch_size _a : Tuple = is_training _a : List[str] = use_auxiliary_loss _a : Optional[Any] = num_queries _a : str = num_channels _a : List[str] = min_size _a : int = max_size _a : Optional[int] = num_labels _a : List[str] = hidden_dim _a : int = hidden_dim def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Tuple = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to( _a ) _a : Optional[Any] = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=_a ) _a : Union[str, Any] = ( torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=_a ) > 0.5 ).float() _a : Tuple = (torch.rand((self.batch_size, self.num_labels) ,device=_a ) > 0.5).long() _a : Dict = self.get_config() return config, pixel_values, pixel_mask, mask_labels, class_labels def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : int = MaskaFormerConfig( hidden_size=self.hidden_dim ,) _a : str = self.num_queries _a : Union[str, Any] = self.num_labels _a : Tuple = [1, 1, 1, 1] _a : Dict = self.num_channels _a : str = 64 _a : Tuple = 128 _a : Optional[Any] = self.hidden_dim _a : Union[str, Any] = self.hidden_dim _a : List[Any] = self.hidden_dim return config def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a, _a, _a, _a, _a : Optional[Any] = self.prepare_config_and_inputs() _a : str = {'pixel_values': pixel_values, 'pixel_mask': pixel_mask} return config, inputs_dict def __lowercase ( self : List[str] ,_a : Optional[Any] ,_a : str ): '''simple docstring''' _a : str = output.encoder_hidden_states _a : Any = output.pixel_decoder_hidden_states _a : Optional[Any] = output.transformer_decoder_hidden_states self.parent.assertTrue(len(_a ) ,len(config.backbone_config.depths ) ) self.parent.assertTrue(len(_a ) ,len(config.backbone_config.depths ) ) self.parent.assertTrue(len(_a ) ,config.decoder_layers ) def __lowercase ( self : List[str] ,_a : str ,_a : List[Any] ,_a : Any ,_a : Union[str, Any]=False ): '''simple docstring''' with torch.no_grad(): _a : str = MaskaFormerModel(config=_a ) model.to(_a ) model.eval() _a : Any = model(pixel_values=_a ,pixel_mask=_a ) _a : Optional[Any] = model(_a ,output_hidden_states=_a ) self.parent.assertEqual( output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.hidden_dim) ,) # let's ensure the other two hidden state exists self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None ) self.parent.assertTrue(output.encoder_last_hidden_state is not None ) if output_hidden_states: self.check_output_hidden_state(_a ,_a ) def __lowercase ( self : Tuple ,_a : List[Any] ,_a : Union[str, Any] ,_a : Tuple ,_a : List[str] ,_a : Any ): '''simple docstring''' _a : int = MaskaFormerForUniversalSegmentation(config=_a ) model.to(_a ) model.eval() def comm_check_on_output(_a : Any ): # let's still check that all the required stuff is there self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None ) self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None ) self.parent.assertTrue(result.encoder_last_hidden_state is not None ) # okay, now we need to check the logits shape # due to the encoder compression, masks have a //4 spatial size self.parent.assertEqual( result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,) # + 1 for null class self.parent.assertEqual( result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) ) with torch.no_grad(): _a : Any = model(pixel_values=_a ,pixel_mask=_a ) _a : Optional[int] = model(_a ) comm_check_on_output(_a ) _a : List[str] = model( pixel_values=_a ,pixel_mask=_a ,mask_labels=_a ,class_labels=_a ) comm_check_on_output(_a ) self.parent.assertTrue(result.loss is not None ) self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) ) @require_torch class UpperCAmelCase__ ( lowercase__ , lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[int] = (MaskaFormerModel, MaskaFormerForUniversalSegmentation) if is_torch_available() else () __UpperCAmelCase : Dict = {'''feature-extraction''': MaskaFormerModel} if is_torch_available() else {} __UpperCAmelCase : Dict = False __UpperCAmelCase : Tuple = False __UpperCAmelCase : Dict = False __UpperCAmelCase : List[Any] = False def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Union[str, Any] = MaskaFormerModelTester(self ) _a : Dict = ConfigTester(self ,config_class=_a ,has_text_modality=_a ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' self.config_tester.run_common_tests() def __lowercase ( self : Optional[int] ): '''simple docstring''' _a, _a : List[str] = self.model_tester.prepare_config_and_inputs_for_common() self.model_tester.create_and_check_maskaformer_model(_a ,**_a ,output_hidden_states=_a ) def __lowercase ( self : str ): '''simple docstring''' _a : Dict = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_maskaformer_instance_segmentation_head_model(*_a ) @unittest.skip(reason='Mask2Former does not use inputs_embeds' ) def __lowercase ( self : Any ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former does not have a get_input_embeddings method' ) def __lowercase ( self : str ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former is not a generative model' ) def __lowercase ( self : List[Any] ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former does not use token embeddings' ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' pass @require_torch_multi_gpu @unittest.skip( reason='Mask2Former has some layers using `add_module` which doesn\'t work well with `nn.DataParallel`' ) def __lowercase ( self : Dict ): '''simple docstring''' pass @unittest.skip('Will be fixed soon by reducing the size of the model used for common tests.' ) def __lowercase ( self : List[Any] ): '''simple docstring''' pass def __lowercase ( self : int ): '''simple docstring''' _a, _a : Any = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a : Union[str, Any] = model_class(_a ) _a : List[str] = inspect.signature(model.forward ) # signature.parameters is an OrderedDict => so arg_names order is deterministic _a : Optional[Any] = [*signature.parameters.keys()] _a : List[Any] = ['pixel_values'] self.assertListEqual(arg_names[:1] ,_a ) @slow def __lowercase ( self : List[str] ): '''simple docstring''' for model_name in ["facebook/mask2former-swin-small-coco-instance"]: _a : Dict = MaskaFormerModel.from_pretrained(_a ) self.assertIsNotNone(_a ) def __lowercase ( self : List[Any] ): '''simple docstring''' _a : int = (self.model_tester.min_size,) * 2 _a : Any = { 'pixel_values': torch.randn((2, 3, *size) ,device=_a ), 'mask_labels': torch.randn((2, 10, *size) ,device=_a ), 'class_labels': torch.zeros(2 ,10 ,device=_a ).long(), } _a : List[Any] = self.model_tester.get_config() _a : int = MaskaFormerForUniversalSegmentation(_a ).to(_a ) _a : str = model(**_a ) self.assertTrue(outputs.loss is not None ) def __lowercase ( self : List[str] ): '''simple docstring''' _a, _a : List[Any] = self.model_tester.prepare_config_and_inputs_for_common() self.model_tester.create_and_check_maskaformer_model(_a ,**_a ,output_hidden_states=_a ) def __lowercase ( self : int ): '''simple docstring''' _a, _a : Dict = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a : Any = model_class(_a ).to(_a ) _a : Optional[int] = model(**_a ,output_attentions=_a ) self.assertTrue(outputs.attentions is not None ) def __lowercase ( self : Tuple ): '''simple docstring''' if not self.model_tester.is_training: return _a : List[str] = self.all_model_classes[1] _a, _a, _a, _a, _a : List[str] = self.model_tester.prepare_config_and_inputs() _a : Any = model_class(_a ) model.to(_a ) model.train() _a : Union[str, Any] = model(_a ,mask_labels=_a ,class_labels=_a ).loss loss.backward() def __lowercase ( self : int ): '''simple docstring''' _a : int = self.all_model_classes[1] _a, _a, _a, _a, _a : List[Any] = self.model_tester.prepare_config_and_inputs() _a : str = True _a : str = True _a : List[str] = model_class(_a ).to(_a ) model.train() _a : Optional[int] = model(_a ,mask_labels=_a ,class_labels=_a ) _a : Tuple = outputs.encoder_hidden_states[0] encoder_hidden_states.retain_grad() _a : str = outputs.pixel_decoder_hidden_states[0] pixel_decoder_hidden_states.retain_grad() _a : Dict = outputs.transformer_decoder_hidden_states[0] transformer_decoder_hidden_states.retain_grad() _a : List[str] = outputs.attentions[0] attentions.retain_grad() outputs.loss.backward(retain_graph=_a ) self.assertIsNotNone(encoder_hidden_states.grad ) self.assertIsNotNone(pixel_decoder_hidden_states.grad ) self.assertIsNotNone(transformer_decoder_hidden_states.grad ) self.assertIsNotNone(attentions.grad ) __lowerCAmelCase = 1e-4 def UpperCAmelCase_ (): """simple docstring""" _a : int = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' ) return image @require_vision @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @cached_property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' return "facebook/mask2former-swin-small-coco-instance" @cached_property def __lowercase ( self : Any ): '''simple docstring''' return MaskaFormerImageProcessor.from_pretrained(self.model_checkpoints ) if is_vision_available() else None def __lowercase ( self : Any ): '''simple docstring''' _a : List[str] = MaskaFormerModel.from_pretrained(self.model_checkpoints ).to(_a ) _a : int = self.default_image_processor _a : Tuple = prepare_img() _a : Any = image_processor(_a ,return_tensors='pt' ).to(_a ) _a : Union[str, Any] = inputs['pixel_values'].shape # check size is divisible by 32 self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 ) # check size self.assertEqual(_a ,(1, 3, 384, 384) ) with torch.no_grad(): _a : Optional[Any] = model(**_a ) _a : List[Any] = torch.tensor( [[-0.2790, -1.0717, -1.1668], [-0.5128, -0.3128, -0.4987], [-0.5832, 0.1971, -0.0197]] ).to(_a ) self.assertTrue( torch.allclose( outputs.encoder_last_hidden_state[0, 0, :3, :3] ,_a ,atol=_a ) ) _a : str = torch.tensor( [[0.8973, 1.1847, 1.1776], [1.1934, 1.5040, 1.5128], [1.1153, 1.4486, 1.4951]] ).to(_a ) self.assertTrue( torch.allclose( outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,_a ,atol=_a ) ) _a : Any = torch.tensor( [[2.1152, 1.7000, -0.8603], [1.5808, 1.8004, -0.9353], [1.6043, 1.7495, -0.5999]] ).to(_a ) self.assertTrue( torch.allclose( outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,_a ,atol=_a ) ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[Any] = MaskaFormerForUniversalSegmentation.from_pretrained(self.model_checkpoints ).to(_a ).eval() _a : Optional[Any] = self.default_image_processor _a : List[Any] = prepare_img() _a : str = image_processor(_a ,return_tensors='pt' ).to(_a ) _a : Any = inputs['pixel_values'].shape # check size is divisible by 32 self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 ) # check size self.assertEqual(_a ,(1, 3, 384, 384) ) with torch.no_grad(): _a : Optional[int] = model(**_a ) # masks_queries_logits _a : Dict = outputs.masks_queries_logits self.assertEqual( masks_queries_logits.shape ,(1, model.config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ) _a : Dict = [ [-8.7839, -9.0056, -8.8121], [-7.4104, -7.0313, -6.5401], [-6.6105, -6.3427, -6.4675], ] _a : Optional[Any] = torch.tensor(_a ).to(_a ) self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,_a ,atol=_a ) ) # class_queries_logits _a : str = outputs.class_queries_logits self.assertEqual(class_queries_logits.shape ,(1, model.config.num_queries, model.config.num_labels + 1) ) _a : str = torch.tensor( [ [1.8324, -8.0835, -4.1922], [0.8450, -9.0050, -3.6053], [0.3045, -7.7293, -3.0275], ] ).to(_a ) self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,_a ,atol=_a ) ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : Any = MaskaFormerForUniversalSegmentation.from_pretrained(self.model_checkpoints ).to(_a ).eval() _a : Tuple = self.default_image_processor _a : Tuple = image_processor( [np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors='pt' ,) _a : str = inputs['pixel_values'].to(_a ) _a : str = [el.to(_a ) for el in inputs['mask_labels']] _a : Dict = [el.to(_a ) for el in inputs['class_labels']] with torch.no_grad(): _a : List[str] = model(**_a ) self.assertTrue(outputs.loss is not None )
271
1
'''simple docstring''' def UpperCAmelCase_ (__a : int , __a : int ): """simple docstring""" return numa ^ numa < 0 if __name__ == "__main__": import doctest doctest.testmod()
271
'''simple docstring''' import argparse import json from typing import List from ltp import LTP from transformers import BertTokenizer def UpperCAmelCase_ (__a : List[Any] ): """simple docstring""" if ( (cp >= 0x4E_00 and cp <= 0x9F_FF) or (cp >= 0x34_00 and cp <= 0x4D_BF) # or (cp >= 0x2_00_00 and cp <= 0x2_A6_DF) # or (cp >= 0x2_A7_00 and cp <= 0x2_B7_3F) # or (cp >= 0x2_B7_40 and cp <= 0x2_B8_1F) # or (cp >= 0x2_B8_20 and cp <= 0x2_CE_AF) # or (cp >= 0xF9_00 and cp <= 0xFA_FF) or (cp >= 0x2_F8_00 and cp <= 0x2_FA_1F) # ): # return True return False def UpperCAmelCase_ (__a : str ): """simple docstring""" for char in word: _a : Union[str, Any] = ord(__a ) if not _is_chinese_char(__a ): return 0 return 1 def UpperCAmelCase_ (__a : List[str] ): """simple docstring""" _a : Dict = set() for token in tokens: _a : str = len(__a ) > 1 and is_chinese(__a ) if chinese_word: word_set.add(__a ) _a : Optional[Any] = list(__a ) return word_list def UpperCAmelCase_ (__a : List[str] , __a : set() ): """simple docstring""" if not chinese_word_set: return bert_tokens _a : Optional[Any] = max([len(__a ) for w in chinese_word_set] ) _a : Optional[int] = bert_tokens _a, _a : Any = 0, len(__a ) while start < end: _a : Tuple = True if is_chinese(bert_word[start] ): _a : Union[str, Any] = min(end - start , __a ) for i in range(__a , 1 , -1 ): _a : Optional[Any] = ''.join(bert_word[start : start + i] ) if whole_word in chinese_word_set: for j in range(start + 1 , start + i ): _a : Any = '##' + bert_word[j] _a : Union[str, Any] = start + i _a : int = False break if single_word: start += 1 return bert_word def UpperCAmelCase_ (__a : List[str] , __a : LTP , __a : BertTokenizer ): """simple docstring""" _a : int = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : Union[str, Any] = ltp_tokenizer.seg(lines[i : i + 1_0_0] )[0] _a : Optional[Any] = [get_chinese_word(__a ) for r in res] ltp_res.extend(__a ) assert len(__a ) == len(__a ) _a : str = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : List[str] = bert_tokenizer(lines[i : i + 1_0_0] , add_special_tokens=__a , truncation=__a , max_length=5_1_2 ) bert_res.extend(res['input_ids'] ) assert len(__a ) == len(__a ) _a : List[str] = [] for input_ids, chinese_word in zip(__a , __a ): _a : int = [] for id in input_ids: _a : Optional[int] = bert_tokenizer._convert_id_to_token(__a ) input_tokens.append(__a ) _a : List[str] = add_sub_symbol(__a , __a ) _a : Tuple = [] # We only save pos of chinese subwords start with ##, which mean is part of a whole word. for i, token in enumerate(__a ): if token[:2] == "##": _a : str = token[2:] # save chinese tokens' pos if len(__a ) == 1 and _is_chinese_char(ord(__a ) ): ref_id.append(__a ) ref_ids.append(__a ) assert len(__a ) == len(__a ) return ref_ids def UpperCAmelCase_ (__a : Optional[Any] ): """simple docstring""" with open(args.file_name , 'r' , encoding='utf-8' ) as f: _a : Dict = f.readlines() _a : int = [line.strip() for line in data if len(__a ) > 0 and not line.isspace()] # avoid delimiter like '\u2029' _a : int = LTP(args.ltp ) # faster in GPU device _a : Tuple = BertTokenizer.from_pretrained(args.bert ) _a : int = prepare_ref(__a , __a , __a ) with open(args.save_path , 'w' , encoding='utf-8' ) as f: _a : Optional[Any] = [json.dumps(__a ) + '\n' for ref in ref_ids] f.writelines(__a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser(description="""prepare_chinese_ref""") parser.add_argument( """--file_name""", type=str, default="""./resources/chinese-demo.txt""", help="""file need process, same as training data in lm""", ) parser.add_argument( """--ltp""", type=str, default="""./resources/ltp""", help="""resources for LTP tokenizer, usually a path""" ) parser.add_argument("""--bert""", type=str, default="""./resources/robert""", help="""resources for Bert tokenizer""") parser.add_argument("""--save_path""", type=str, default="""./resources/ref.txt""", help="""path to save res""") __lowerCAmelCase = parser.parse_args() main(args)
271
1
'''simple docstring''' from typing import Optional import torch import torch.utils.checkpoint from torch import Tensor, nn from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss from ...activations import ACTaFN from ...file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward from ...modeling_outputs import ( BaseModelOutputWithNoAttention, BaseModelOutputWithPoolingAndNoAttention, ImageClassifierOutputWithNoAttention, ) from ...modeling_utils import PreTrainedModel from ...utils import logging from .configuration_regnet import RegNetConfig __lowerCAmelCase = logging.get_logger(__name__) # General docstring __lowerCAmelCase = """RegNetConfig""" # Base docstring __lowerCAmelCase = """facebook/regnet-y-040""" __lowerCAmelCase = [1, 1_0_8_8, 7, 7] # Image classification docstring __lowerCAmelCase = """facebook/regnet-y-040""" __lowerCAmelCase = """tabby, tabby cat""" __lowerCAmelCase = [ """facebook/regnet-y-040""", # See all regnet models at https://huggingface.co/models?filter=regnet ] class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : List[str] ,_a : int ,_a : int ,_a : int = 3 ,_a : int = 1 ,_a : int = 1 ,_a : Optional[str] = "relu" ,): '''simple docstring''' super().__init__() _a : Any = nn.Convad( _a ,_a ,kernel_size=_a ,stride=_a ,padding=kernel_size // 2 ,groups=_a ,bias=_a ,) _a : str = nn.BatchNormad(_a ) _a : Any = ACTaFN[activation] if activation is not None else nn.Identity() def __lowercase ( self : List[str] ,_a : int ): '''simple docstring''' _a : Dict = self.convolution(_a ) _a : int = self.normalization(_a ) _a : int = self.activation(_a ) return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : List[str] ,_a : RegNetConfig ): '''simple docstring''' super().__init__() _a : str = RegNetConvLayer( config.num_channels ,config.embedding_size ,kernel_size=3 ,stride=2 ,activation=config.hidden_act ) _a : Union[str, Any] = config.num_channels def __lowercase ( self : List[Any] ,_a : int ): '''simple docstring''' _a : Optional[int] = pixel_values.shape[1] if num_channels != self.num_channels: raise ValueError( 'Make sure that the channel dimension of the pixel values match with the one set in the configuration.' ) _a : Any = self.embedder(_a ) return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : List[Any] ,_a : int ,_a : int ,_a : int = 2 ): '''simple docstring''' super().__init__() _a : Dict = nn.Convad(_a ,_a ,kernel_size=1 ,stride=_a ,bias=_a ) _a : Dict = nn.BatchNormad(_a ) def __lowercase ( self : str ,_a : Tensor ): '''simple docstring''' _a : int = self.convolution(_a ) _a : Optional[Any] = self.normalization(_a ) return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : Tuple ,_a : int ,_a : int ): '''simple docstring''' super().__init__() _a : str = nn.AdaptiveAvgPoolad((1, 1) ) _a : Union[str, Any] = nn.Sequential( nn.Convad(_a ,_a ,kernel_size=1 ) ,nn.ReLU() ,nn.Convad(_a ,_a ,kernel_size=1 ) ,nn.Sigmoid() ,) def __lowercase ( self : Union[str, Any] ,_a : str ): '''simple docstring''' _a : Tuple = self.pooler(_a ) _a : Tuple = self.attention(_a ) _a : Optional[Any] = hidden_state * attention return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : Any ,_a : RegNetConfig ,_a : int ,_a : int ,_a : int = 1 ): '''simple docstring''' super().__init__() _a : str = in_channels != out_channels or stride != 1 _a : Tuple = max(1 ,out_channels // config.groups_width ) _a : Union[str, Any] = ( RegNetShortCut(_a ,_a ,stride=_a ) if should_apply_shortcut else nn.Identity() ) _a : Tuple = nn.Sequential( RegNetConvLayer(_a ,_a ,kernel_size=1 ,activation=config.hidden_act ) ,RegNetConvLayer(_a ,_a ,stride=_a ,groups=_a ,activation=config.hidden_act ) ,RegNetConvLayer(_a ,_a ,kernel_size=1 ,activation=_a ) ,) _a : Optional[Any] = ACTaFN[config.hidden_act] def __lowercase ( self : Any ,_a : int ): '''simple docstring''' _a : List[Any] = hidden_state _a : int = self.layer(_a ) _a : Any = self.shortcut(_a ) hidden_state += residual _a : int = self.activation(_a ) return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : Dict ,_a : RegNetConfig ,_a : int ,_a : int ,_a : int = 1 ): '''simple docstring''' super().__init__() _a : Optional[int] = in_channels != out_channels or stride != 1 _a : int = max(1 ,out_channels // config.groups_width ) _a : Union[str, Any] = ( RegNetShortCut(_a ,_a ,stride=_a ) if should_apply_shortcut else nn.Identity() ) _a : Optional[int] = nn.Sequential( RegNetConvLayer(_a ,_a ,kernel_size=1 ,activation=config.hidden_act ) ,RegNetConvLayer(_a ,_a ,stride=_a ,groups=_a ,activation=config.hidden_act ) ,RegNetSELayer(_a ,reduced_channels=int(round(in_channels / 4 ) ) ) ,RegNetConvLayer(_a ,_a ,kernel_size=1 ,activation=_a ) ,) _a : List[Any] = ACTaFN[config.hidden_act] def __lowercase ( self : Dict ,_a : Union[str, Any] ): '''simple docstring''' _a : Any = hidden_state _a : Tuple = self.layer(_a ) _a : int = self.shortcut(_a ) hidden_state += residual _a : List[str] = self.activation(_a ) return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : List[str] ,_a : RegNetConfig ,_a : int ,_a : int ,_a : int = 2 ,_a : int = 2 ,): '''simple docstring''' super().__init__() _a : List[str] = RegNetXLayer if config.layer_type == 'x' else RegNetYLayer _a : Optional[Any] = nn.Sequential( # downsampling is done in the first layer with stride of 2 layer( _a ,_a ,_a ,stride=_a ,) ,*[layer(_a ,_a ,_a ) for _ in range(depth - 1 )] ,) def __lowercase ( self : Any ,_a : Dict ): '''simple docstring''' _a : Optional[int] = self.layers(_a ) return hidden_state class UpperCAmelCase__ ( nn.Module ): """simple docstring""" def __init__( self : Any ,_a : RegNetConfig ): '''simple docstring''' super().__init__() _a : Dict = nn.ModuleList([] ) # based on `downsample_in_first_stage`, the first layer of the first stage may or may not downsample the input self.stages.append( RegNetStage( _a ,config.embedding_size ,config.hidden_sizes[0] ,stride=2 if config.downsample_in_first_stage else 1 ,depth=config.depths[0] ,) ) _a : Dict = zip(config.hidden_sizes ,config.hidden_sizes[1:] ) for (in_channels, out_channels), depth in zip(_a ,config.depths[1:] ): self.stages.append(RegNetStage(_a ,_a ,_a ,depth=_a ) ) def __lowercase ( self : str ,_a : Tensor ,_a : bool = False ,_a : bool = True ): '''simple docstring''' _a : List[str] = () if output_hidden_states else None for stage_module in self.stages: if output_hidden_states: _a : Union[str, Any] = hidden_states + (hidden_state,) _a : int = stage_module(_a ) if output_hidden_states: _a : Any = hidden_states + (hidden_state,) if not return_dict: return tuple(v for v in [hidden_state, hidden_states] if v is not None ) return BaseModelOutputWithNoAttention(last_hidden_state=_a ,hidden_states=_a ) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : int = RegNetConfig __UpperCAmelCase : Any = '''regnet''' __UpperCAmelCase : Optional[Any] = '''pixel_values''' __UpperCAmelCase : List[str] = True def __lowercase ( self : str ,_a : Optional[int] ): '''simple docstring''' if isinstance(_a ,nn.Convad ): nn.init.kaiming_normal_(module.weight ,mode='fan_out' ,nonlinearity='relu' ) elif isinstance(_a ,(nn.BatchNormad, nn.GroupNorm) ): nn.init.constant_(module.weight ,1 ) nn.init.constant_(module.bias ,0 ) def __lowercase ( self : Union[str, Any] ,_a : int ,_a : List[str]=False ): '''simple docstring''' if isinstance(_a ,_a ): _a : Optional[int] = value __lowerCAmelCase = r""" This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior. Parameters: config ([`RegNetConfig`]): Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights. """ __lowerCAmelCase = r""" Args: pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`): Pixel values. Pixel values can be obtained using [`AutoImageProcessor`]. See [`ConvNextImageProcessor.__call__`] for details. output_hidden_states (`bool`, *optional*): Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for more detail. return_dict (`bool`, *optional*): Whether or not to return a [`~file_utils.ModelOutput`] instead of a plain tuple. """ @add_start_docstrings( '''The bare RegNet model outputting raw features without any specific head on top.''' , lowercase__ , ) # Copied from transformers.models.resnet.modeling_resnet.ResNetModel with RESNET->REGNET,ResNet->RegNet class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : int ,_a : int ): '''simple docstring''' super().__init__(_a ) _a : Union[str, Any] = config _a : Union[str, Any] = RegNetEmbeddings(_a ) _a : str = RegNetEncoder(_a ) _a : Optional[int] = nn.AdaptiveAvgPoolad((1, 1) ) # Initialize weights and apply final processing self.post_init() @add_start_docstrings_to_model_forward(_a ) @add_code_sample_docstrings( checkpoint=_CHECKPOINT_FOR_DOC ,output_type=_a ,config_class=_CONFIG_FOR_DOC ,modality='vision' ,expected_output=_EXPECTED_OUTPUT_SHAPE ,) def __lowercase ( self : Union[str, Any] ,_a : Tensor ,_a : Optional[bool] = None ,_a : Optional[bool] = None ): '''simple docstring''' _a : List[str] = ( output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states ) _a : Optional[Any] = return_dict if return_dict is not None else self.config.use_return_dict _a : Optional[Any] = self.embedder(_a ) _a : Dict = self.encoder( _a ,output_hidden_states=_a ,return_dict=_a ) _a : Tuple = encoder_outputs[0] _a : Dict = self.pooler(_a ) if not return_dict: return (last_hidden_state, pooled_output) + encoder_outputs[1:] return BaseModelOutputWithPoolingAndNoAttention( last_hidden_state=_a ,pooler_output=_a ,hidden_states=encoder_outputs.hidden_states ,) @add_start_docstrings( ''' RegNet Model with an image classification head on top (a linear layer on top of the pooled features), e.g. for ImageNet. ''' , lowercase__ , ) # Copied from transformers.models.resnet.modeling_resnet.ResNetForImageClassification with RESNET->REGNET,ResNet->RegNet,resnet->regnet class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : str ,_a : Optional[Any] ): '''simple docstring''' super().__init__(_a ) _a : Dict = config.num_labels _a : List[Any] = RegNetModel(_a ) # classification head _a : Optional[int] = nn.Sequential( nn.Flatten() ,nn.Linear(config.hidden_sizes[-1] ,config.num_labels ) if config.num_labels > 0 else nn.Identity() ,) # initialize weights and apply final processing self.post_init() @add_start_docstrings_to_model_forward(_a ) @add_code_sample_docstrings( checkpoint=_IMAGE_CLASS_CHECKPOINT ,output_type=_a ,config_class=_CONFIG_FOR_DOC ,expected_output=_IMAGE_CLASS_EXPECTED_OUTPUT ,) def __lowercase ( self : List[Any] ,_a : Optional[torch.FloatTensor] = None ,_a : Optional[torch.LongTensor] = None ,_a : Optional[bool] = None ,_a : Optional[bool] = None ,): '''simple docstring''' _a : Any = return_dict if return_dict is not None else self.config.use_return_dict _a : int = self.regnet(_a ,output_hidden_states=_a ,return_dict=_a ) _a : str = outputs.pooler_output if return_dict else outputs[1] _a : Optional[Any] = self.classifier(_a ) _a : Dict = None if labels is not None: if self.config.problem_type is None: if self.num_labels == 1: _a : Tuple = 'regression' elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int): _a : Tuple = 'single_label_classification' else: _a : Optional[int] = 'multi_label_classification' if self.config.problem_type == "regression": _a : Tuple = MSELoss() if self.num_labels == 1: _a : Union[str, Any] = loss_fct(logits.squeeze() ,labels.squeeze() ) else: _a : List[Any] = loss_fct(_a ,_a ) elif self.config.problem_type == "single_label_classification": _a : Tuple = CrossEntropyLoss() _a : List[str] = loss_fct(logits.view(-1 ,self.num_labels ) ,labels.view(-1 ) ) elif self.config.problem_type == "multi_label_classification": _a : List[str] = BCEWithLogitsLoss() _a : str = loss_fct(_a ,_a ) if not return_dict: _a : int = (logits,) + outputs[2:] return (loss,) + output if loss is not None else output return ImageClassifierOutputWithNoAttention(loss=_a ,logits=_a ,hidden_states=outputs.hidden_states )
271
'''simple docstring''' import warnings from ...utils import logging from .image_processing_videomae import VideoMAEImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Tuple ,*_a : List[str] ,**_a : Any ): '''simple docstring''' warnings.warn( 'The class VideoMAEFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use VideoMAEImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
1
'''simple docstring''' import os import unittest from transformers import FunnelTokenizer, FunnelTokenizerFast from transformers.models.funnel.tokenization_funnel import VOCAB_FILES_NAMES from transformers.testing_utils import require_tokenizers from ...test_tokenization_common import TokenizerTesterMixin @require_tokenizers class UpperCAmelCase__ ( lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Any = FunnelTokenizer __UpperCAmelCase : List[str] = FunnelTokenizerFast __UpperCAmelCase : Optional[Any] = True __UpperCAmelCase : Tuple = True def __lowercase ( self : Tuple ): '''simple docstring''' super().setUp() _a : Any = [ '<unk>', '<cls>', '<sep>', 'want', '##want', '##ed', 'wa', 'un', 'runn', '##ing', ',', 'low', 'lowest', ] _a : Optional[Any] = 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 __lowercase ( self : int ,**_a : Dict ): '''simple docstring''' return FunnelTokenizer.from_pretrained(self.tmpdirname ,**_a ) def __lowercase ( self : List[str] ,**_a : List[Any] ): '''simple docstring''' return FunnelTokenizerFast.from_pretrained(self.tmpdirname ,**_a ) def __lowercase ( self : Dict ,_a : str ): '''simple docstring''' _a : Union[str, Any] = 'UNwant\u00E9d,running' _a : int = 'unwanted, running' return input_text, output_text def __lowercase ( self : Any ): '''simple docstring''' _a : List[Any] = self.tokenizer_class(self.vocab_file ) _a : str = tokenizer.tokenize('UNwant\u00E9d,running' ) self.assertListEqual(_a ,['un', '##want', '##ed', ',', 'runn', '##ing'] ) self.assertListEqual(tokenizer.convert_tokens_to_ids(_a ) ,[7, 4, 5, 10, 8, 9] ) def __lowercase ( self : Any ): '''simple docstring''' _a : Tuple = self.get_tokenizers(do_lower_case=_a ) for tokenizer in tokenizers: _a : str = tokenizer('UNwant\u00E9d,running' ) _a : Optional[int] = len(inputs['input_ids'] ) - 1 self.assertListEqual(inputs['token_type_ids'] ,[2] + [0] * sentence_len ) _a : Any = tokenizer('UNwant\u00E9d,running' ,'UNwant\u00E9d,running' ) self.assertListEqual(inputs['token_type_ids'] ,[2] + [0] * sentence_len + [1] * sentence_len )
271
'''simple docstring''' from __future__ import annotations from random import choice def UpperCAmelCase_ (__a : str ): """simple docstring""" return choice(__a ) def UpperCAmelCase_ (__a : list[int] , __a : int ): """simple docstring""" _a : Dict = random_pivot(__a ) # partition based on pivot # linear time _a : Optional[int] = [e for e in lst if e < pivot] _a : List[str] = [e for e in lst if e > pivot] # if we get lucky, pivot might be the element we want. # we can easily see this: # small (elements smaller than k) # + pivot (kth element) # + big (elements larger than k) if len(__a ) == k - 1: return pivot # pivot is in elements bigger than k elif len(__a ) < k - 1: return kth_number(__a , k - len(__a ) - 1 ) # pivot is in elements smaller than k else: return kth_number(__a , __a ) if __name__ == "__main__": import doctest doctest.testmod()
271
1
'''simple docstring''' def UpperCAmelCase_ (__a : int , __a : int ): """simple docstring""" return "\n".join( f"""{number} * {i} = {number * i}""" for i in range(1 , number_of_terms + 1 ) ) if __name__ == "__main__": print(multiplication_table(number=5, number_of_terms=1_0))
271
'''simple docstring''' class UpperCAmelCase__ : """simple docstring""" def __init__( self : Dict ): '''simple docstring''' _a : Dict = {} def __lowercase ( self : Union[str, Any] ): '''simple docstring''' print(self.vertex ) for i in self.vertex: print(_a ,' -> ' ,' -> '.join([str(_a ) for j in self.vertex[i]] ) ) def __lowercase ( self : Dict ,_a : int ,_a : int ): '''simple docstring''' if from_vertex in self.vertex: self.vertex[from_vertex].append(_a ) else: # else make a new vertex _a : int = [to_vertex] def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Tuple = [False] * len(self.vertex ) # call the recursive helper function for i in range(len(self.vertex ) ): if not visited[i]: self.dfs_recursive(_a ,_a ) def __lowercase ( self : Union[str, Any] ,_a : int ,_a : list ): '''simple docstring''' _a : List[Any] = True print(_a ,end=' ' ) # Recur for all the vertices that are adjacent to this node for i in self.vertex: if not visited[i]: self.dfs_recursive(_a ,_a ) if __name__ == "__main__": __lowerCAmelCase = Graph() g.add_edge(0, 1) g.add_edge(0, 2) g.add_edge(1, 2) g.add_edge(2, 0) g.add_edge(2, 3) g.add_edge(3, 3) g.print_graph() print("""DFS:""") g.dfs() # OUTPUT: # 0 -> 1 -> 2 # 1 -> 2 # 2 -> 0 -> 3 # 3 -> 3 # DFS: # 0 1 2 3
271
1
'''simple docstring''' import warnings from ...utils import logging from .image_processing_beit import BeitImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : int ,*_a : Optional[int] ,**_a : str ): '''simple docstring''' warnings.warn( 'The class BeitFeatureExtractor is deprecated and will be removed in version 5 of Transformers. Please' ' use BeitImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
'''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 __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = """▁""" __lowerCAmelCase = {"""vocab_file""": """sentencepiece.bpe.model""", """monolingual_vocab_file""": """dict.txt"""} __lowerCAmelCase = { """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""", }, } __lowerCAmelCase = {"""vinai/bartpho-syllable""": 1_0_2_4} class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[Any] = VOCAB_FILES_NAMES __UpperCAmelCase : Dict = PRETRAINED_VOCAB_FILES_MAP __UpperCAmelCase : Optional[int] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES __UpperCAmelCase : Dict = ['''input_ids''', '''attention_mask'''] def __init__( self : str ,_a : str ,_a : Any ,_a : Any="<s>" ,_a : Dict="</s>" ,_a : int="</s>" ,_a : Union[str, Any]="<s>" ,_a : List[Any]="<unk>" ,_a : Optional[Any]="<pad>" ,_a : List[str]="<mask>" ,_a : Optional[Dict[str, Any]] = None ,**_a : int ,): '''simple docstring''' _a : Any = AddedToken(_a ,lstrip=_a ,rstrip=_a ) if isinstance(_a ,_a ) else mask_token _a : Optional[Any] = {} if sp_model_kwargs is None else sp_model_kwargs super().__init__( bos_token=_a ,eos_token=_a ,unk_token=_a ,sep_token=_a ,cls_token=_a ,pad_token=_a ,mask_token=_a ,sp_model_kwargs=self.sp_model_kwargs ,**_a ,) _a : Optional[int] = vocab_file _a : Union[str, Any] = monolingual_vocab_file _a : List[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.Load(str(_a ) ) # Load the reduced vocab # Keep order of special tokens for backward compatibility _a : Union[str, Any] = {} _a : int = 0 for token in [bos_token, pad_token, eos_token, unk_token, sep_token, cls_token]: if str(_a ) not in self.fairseq_tokens_to_ids: _a : int = cnt cnt += 1 with open(_a ,'r' ,encoding='utf-8' ) as f: for line in f.readlines(): _a : str = line.strip().split()[0] _a : Tuple = len(self.fairseq_tokens_to_ids ) if str(_a ) not in self.fairseq_tokens_to_ids: _a : List[str] = len(self.fairseq_tokens_to_ids ) _a : Tuple = {v: k for k, v in self.fairseq_tokens_to_ids.items()} def __getstate__( self : Union[str, Any] ): '''simple docstring''' _a : int = self.__dict__.copy() _a : str = None _a : Optional[Any] = self.sp_model.serialized_model_proto() return state def __setstate__( self : Tuple ,_a : Tuple ): '''simple docstring''' _a : Tuple = d # for backward compatibility if not hasattr(self ,'sp_model_kwargs' ): _a : List[str] = {} _a : List[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.LoadFromSerializedProto(self.sp_model_proto ) def __lowercase ( self : Dict ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' if token_ids_a is None: return [self.cls_token_id] + token_ids_a + [self.sep_token_id] _a : Dict = [self.cls_token_id] _a : int = [self.sep_token_id] return cls + token_ids_a + sep + sep + token_ids_a + sep def __lowercase ( self : int ,_a : List[int] ,_a : Optional[List[int]] = None ,_a : bool = False ): '''simple docstring''' if already_has_special_tokens: return super().get_special_tokens_mask( token_ids_a=_a ,token_ids_a=_a ,already_has_special_tokens=_a ) if token_ids_a is None: return [1] + ([0] * len(_a )) + [1] return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1] def __lowercase ( self : Tuple ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' _a : List[str] = [self.sep_token_id] _a : List[str] = [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 __lowercase ( self : Dict ): '''simple docstring''' return len(self.fairseq_ids_to_tokens ) def __lowercase ( self : Dict ): '''simple docstring''' _a : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )} vocab.update(self.added_tokens_encoder ) return vocab def __lowercase ( self : Tuple ,_a : str ): '''simple docstring''' return self.sp_model.encode(_a ,out_type=_a ) def __lowercase ( self : Union[str, Any] ,_a : Union[str, Any] ): '''simple docstring''' if token in self.fairseq_tokens_to_ids: return self.fairseq_tokens_to_ids[token] else: return self.unk_token_id def __lowercase ( self : Any ,_a : int ): '''simple docstring''' return self.fairseq_ids_to_tokens[index] def __lowercase ( self : Tuple ,_a : Union[str, Any] ): '''simple docstring''' _a : str = ''.join(_a ).replace(_a ,' ' ).strip() return out_string def __lowercase ( self : Union[str, Any] ,_a : str ,_a : Optional[str] = None ): '''simple docstring''' if not os.path.isdir(_a ): logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" ) return _a : int = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) _a : int = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['monolingual_vocab_file'] ,) if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ): copyfile(self.vocab_file ,_a ) elif not os.path.isfile(self.vocab_file ): with open(_a ,'wb' ) as fi: _a : List[Any] = self.sp_model.serialized_model_proto() fi.write(_a ) if os.path.abspath(self.monolingual_vocab_file ) != os.path.abspath( _a ) and os.path.isfile(self.monolingual_vocab_file ): copyfile(self.monolingual_vocab_file ,_a ) elif not os.path.isfile(self.monolingual_vocab_file ): with open(_a ,'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(_a )} \n""" ) return out_vocab_file, out_monolingual_vocab_file
271
1
'''simple docstring''' from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_tf_available, is_torch_available, ) __lowerCAmelCase = { """configuration_vision_text_dual_encoder""": ["""VisionTextDualEncoderConfig"""], """processing_vision_text_dual_encoder""": ["""VisionTextDualEncoderProcessor"""], } try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""VisionTextDualEncoderModel"""] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""FlaxVisionTextDualEncoderModel"""] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""TFVisionTextDualEncoderModel"""] if TYPE_CHECKING: from .configuration_vision_text_dual_encoder import VisionTextDualEncoderConfig from .processing_vision_text_dual_encoder import VisionTextDualEncoderProcessor try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_vision_text_dual_encoder import VisionTextDualEncoderModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_vision_text_dual_encoder import FlaxVisionTextDualEncoderModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_vision_text_dual_encoder import TFVisionTextDualEncoderModel else: import sys __lowerCAmelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure)
271
'''simple docstring''' import numpy as np from transformers import BatchFeature from transformers.testing_utils import require_tf, require_torch from .test_feature_extraction_common import FeatureExtractionSavingTestMixin class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Union[str, Any] = None __UpperCAmelCase : List[Any] = None @property def __lowercase ( self : Dict ): '''simple docstring''' return self.feat_extract_tester.prepare_feat_extract_dict() def __lowercase ( self : str ): '''simple docstring''' _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) self.assertTrue(hasattr(_a ,'feature_size' ) ) self.assertTrue(hasattr(_a ,'sampling_rate' ) ) self.assertTrue(hasattr(_a ,'padding_value' ) ) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.feat_extract_tester.prepare_inputs_for_common() _a : str = self.feature_extraction_class(**self.feat_extract_dict ) _a : int = feat_extract.model_input_names[0] _a : List[Any] = BatchFeature({input_name: speech_inputs} ) self.assertTrue(all(len(_a ) == len(_a ) for x, y in zip(_a ,processed_features[input_name] ) ) ) _a : Any = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Union[str, Any] = BatchFeature({input_name: speech_inputs} ,tensor_type='np' ) _a : Union[str, Any] = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : Optional[int] = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) @require_torch def __lowercase ( self : Any ): '''simple docstring''' _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : int = feat_extract.model_input_names[0] _a : str = BatchFeature({input_name: speech_inputs} ,tensor_type='pt' ) _a : str = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : str = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) @require_tf def __lowercase ( self : int ): '''simple docstring''' _a : int = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : Tuple = feat_extract.model_input_names[0] _a : int = BatchFeature({input_name: speech_inputs} ,tensor_type='tf' ) _a : Optional[int] = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : Optional[Any] = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) def __lowercase ( self : Dict ,_a : Any=False ): '''simple docstring''' def _inputs_have_equal_length(_a : Tuple ): _a : Tuple = len(input[0] ) for input_slice in input[1:]: if len(_a ) != length: return False return True def _inputs_are_equal(_a : Optional[Any] ,_a : Union[str, Any] ): if len(_a ) != len(_a ): return False for input_slice_a, input_slice_a in zip(_a ,_a ): if not np.allclose(np.asarray(_a ) ,np.asarray(_a ) ,atol=1E-3 ): return False return True _a : int = self.feature_extraction_class(**self.feat_extract_dict ) _a : Tuple = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a ) _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Tuple = BatchFeature({input_name: speech_inputs} ) _a : str = self.feat_extract_tester.seq_length_diff _a : Dict = self.feat_extract_tester.max_seq_length + pad_diff _a : Dict = self.feat_extract_tester.min_seq_length _a : Optional[Any] = self.feat_extract_tester.batch_size _a : Tuple = self.feat_extract_tester.feature_size # test padding for List[int] + numpy _a : int = feat_extract.pad(_a ,padding=_a ) _a : List[Any] = input_a[input_name] _a : Tuple = feat_extract.pad(_a ,padding='longest' ) _a : Any = input_a[input_name] _a : Optional[Any] = feat_extract.pad(_a ,padding='max_length' ,max_length=len(speech_inputs[-1] ) ) _a : List[str] = input_a[input_name] _a : List[str] = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' ) _a : str = input_a[input_name] # max_length parameter has to be provided when setting `padding="max_length"` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='max_length' )[input_name] _a : int = feat_extract.pad( _a ,padding='max_length' ,max_length=_a ,return_tensors='np' ) _a : Optional[int] = input_a[input_name] self.assertFalse(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) self.assertTrue(len(input_a[0] ) == pad_min_length ) self.assertTrue(len(input_a[1] ) == pad_min_length + pad_diff ) self.assertTrue(input_a.shape[:2] == (batch_size, len(input_a[0] )) ) self.assertTrue(input_a.shape[:2] == (batch_size, pad_max_length) ) if feature_size > 1: self.assertTrue(input_a.shape[2] == input_a.shape[2] == feature_size ) # test padding for `pad_to_multiple_of` for List[int] + numpy _a : Tuple = feat_extract.pad(_a ,pad_to_multiple_of=10 ) _a : List[str] = input_a[input_name] _a : str = feat_extract.pad(_a ,padding='longest' ,pad_to_multiple_of=10 ) _a : Tuple = input_a[input_name] _a : Optional[int] = feat_extract.pad( _a ,padding='max_length' ,pad_to_multiple_of=10 ,max_length=_a ) _a : Any = input_a[input_name] _a : Optional[int] = feat_extract.pad( _a ,padding='max_length' ,pad_to_multiple_of=10 ,max_length=_a ,return_tensors='np' ,) _a : Dict = input_a[input_name] self.assertTrue(all(len(_a ) % 10 == 0 for x in input_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) _a : List[str] = pad_max_length if pad_max_length % 10 == 0 else (pad_max_length // 10 + 1) * 10 self.assertTrue(all(len(_a ) == expected_mult_pad_length for x in input_a ) ) self.assertEqual(input_a.shape[:2] ,(batch_size, expected_mult_pad_length) ) if feature_size > 1: self.assertTrue(input_a.shape[2] == feature_size ) # Check padding value is correct _a : Any = (np.ones(self.feat_extract_tester.feature_size ) * feat_extract.padding_value).sum() self.assertTrue( abs(np.asarray(input_a[0] )[pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1E-3 ) self.assertTrue( abs( np.asarray(input_a[1] )[pad_min_length + pad_diff :].sum() - padding_vector_sum * (pad_max_length - pad_min_length - pad_diff) ) < 1E-3 ) self.assertTrue( abs( np.asarray(input_a[2] )[pad_min_length + 2 * pad_diff :].sum() - padding_vector_sum * (pad_max_length - pad_min_length - 2 * pad_diff) ) < 1E-3 ) self.assertTrue( abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1E-3 ) self.assertTrue( abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (expected_mult_pad_length - pad_min_length) ) < 1E-3 ) def __lowercase ( self : List[Any] ,_a : Optional[int]=False ): '''simple docstring''' def _inputs_have_equal_length(_a : List[str] ): _a : Union[str, Any] = len(input[0] ) for input_slice in input[1:]: if len(_a ) != length: return False return True def _inputs_are_equal(_a : List[str] ,_a : List[str] ): if len(_a ) != len(_a ): return False for input_slice_a, input_slice_a in zip(_a ,_a ): if not np.allclose(np.asarray(_a ) ,np.asarray(_a ) ,atol=1E-3 ): return False return True _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : str = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a ) _a : Any = feat_extract.model_input_names[0] _a : List[Any] = BatchFeature({input_name: speech_inputs} ) # truncate to smallest _a : Union[str, Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,truncation=_a ) _a : str = input_a[input_name] _a : List[str] = feat_extract.pad(_a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ) _a : Tuple = input_a[input_name] self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertFalse(_inputs_have_equal_length(_a ) ) # truncate to smallest with np _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,return_tensors='np' ,truncation=_a ,) _a : Any = input_a[input_name] _a : List[Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,return_tensors='np' ) _a : int = input_a[input_name] self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(input_a.shape[1] == len(speech_inputs[0] ) ) # since truncation forces padding to be smaller than longest input # function can't return `np.ndarray`, but has to return list self.assertFalse(_inputs_have_equal_length(_a ) ) # truncate to middle _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,truncation=_a ,return_tensors='np' ,) _a : List[Any] = input_a[input_name] _a : Tuple = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,truncation=_a ) _a : Tuple = input_a[input_name] _a : Tuple = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,return_tensors='np' ) _a : Dict = input_a[input_name] self.assertTrue(input_a.shape[1] == len(speech_inputs[1] ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) # since truncation forces padding to be smaller than longest input # function can't return `np.ndarray`, but has to return list self.assertFalse(_inputs_have_equal_length(_a ) ) self.assertTrue(len(input_a[-1] ) == len(speech_inputs[-1] ) ) # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,truncation=_a )[input_name] # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='longest' ,truncation=_a )[input_name] # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='longest' ,truncation=_a )[input_name] # max_length parameter has to be provided when setting `truncation=True` and padding="max_length" with self.assertRaises(_a ): feat_extract.pad(_a ,padding='max_length' ,truncation=_a )[input_name] # test truncation for `pad_to_multiple_of` for List[int] + numpy _a : Optional[Any] = 12 _a : List[Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,pad_to_multiple_of=_a ,truncation=_a ,) _a : Tuple = input_a[input_name] _a : str = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,pad_to_multiple_of=_a ,) _a : List[Any] = input_a[input_name] # retrieve expected_length as multiple of pad_to_multiple_of _a : List[Any] = len(speech_inputs[0] ) if expected_length % pad_to_multiple_of != 0: _a : Union[str, Any] = ((len(speech_inputs[0] ) // pad_to_multiple_of) + 1) * pad_to_multiple_of self.assertTrue(len(input_a[0] ) == expected_length ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertFalse(_inputs_have_equal_length(_a ) ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' self._check_padding(numpify=_a ) def __lowercase ( self : Tuple ): '''simple docstring''' self._check_padding(numpify=_a ) def __lowercase ( self : Dict ): '''simple docstring''' self._check_truncation(numpify=_a ) def __lowercase ( self : str ): '''simple docstring''' self._check_truncation(numpify=_a ) @require_torch def __lowercase ( self : Dict ): '''simple docstring''' _a : Any = self.feature_extraction_class(**self.feat_extract_dict ) _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Optional[int] = BatchFeature({input_name: speech_inputs} ) _a : List[Any] = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' )[input_name] _a : List[str] = feat_extract.pad(_a ,padding='longest' ,return_tensors='pt' )[input_name] self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_pt.numpy().astype(np.floataa ).sum() ) < 1E-2 ) @require_tf def __lowercase ( self : int ): '''simple docstring''' _a : List[str] = self.feature_extraction_class(**self.feat_extract_dict ) _a : Optional[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Dict = feat_extract.model_input_names[0] _a : Optional[Any] = BatchFeature({input_name: speech_inputs} ) _a : Dict = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' )[input_name] _a : Any = feat_extract.pad(_a ,padding='longest' ,return_tensors='tf' )[input_name] self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_tf.numpy().astype(np.floataa ).sum() ) < 1E-2 ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : str = self.feat_extract_dict _a : List[Any] = True _a : Optional[int] = self.feature_extraction_class(**_a ) _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Tuple = [len(_a ) for x in speech_inputs] _a : int = feat_extract.model_input_names[0] _a : Optional[Any] = BatchFeature({input_name: speech_inputs} ) _a : str = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' ) self.assertIn('attention_mask' ,_a ) self.assertListEqual(list(processed.attention_mask.shape ) ,list(processed[input_name].shape[:2] ) ) self.assertListEqual(processed.attention_mask.sum(-1 ).tolist() ,_a ) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.feat_extract_dict _a : Tuple = True _a : Optional[int] = self.feature_extraction_class(**_a ) _a : Dict = self.feat_extract_tester.prepare_inputs_for_common() _a : Dict = [len(_a ) for x in speech_inputs] _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Any = BatchFeature({input_name: speech_inputs} ) _a : List[Any] = min(_a ) _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=_a ,truncation=_a ,return_tensors='np' ) self.assertIn('attention_mask' ,_a ) self.assertListEqual( list(processed_pad.attention_mask.shape ) ,[processed_pad[input_name].shape[0], max_length] ) self.assertListEqual( processed_pad.attention_mask[:, :max_length].sum(-1 ).tolist() ,[max_length for x in speech_inputs] )
271
1
'''simple docstring''' from __future__ import annotations import time import numpy as np __lowerCAmelCase = [8, 5, 9, 7] __lowerCAmelCase = [ [2, 0, 1, 1], [0, 1, 2, 1], [4, 0, 0, 3], [0, 2, 1, 0], [1, 0, 3, 0], ] __lowerCAmelCase = [ [3, 2, 1, 4], [0, 2, 5, 2], [5, 1, 0, 5], [1, 5, 3, 0], [3, 0, 3, 3], ] class UpperCAmelCase__ : """simple docstring""" def __init__( self : Tuple ,_a : list[int] ,_a : list[list[int]] ,_a : list[list[int]] ,): '''simple docstring''' _a : List[str] = claim_vector _a : Dict = allocated_resources_table _a : int = maximum_claim_table def __lowercase ( self : List[str] ): '''simple docstring''' return [ sum(p_item[i] for p_item in self.__allocated_resources_table ) for i in range(len(self.__allocated_resources_table[0] ) ) ] def __lowercase ( self : Optional[int] ): '''simple docstring''' return np.array(self.__claim_vector ) - np.array( self.__processes_resource_summation() ) def __lowercase ( self : int ): '''simple docstring''' return [ list(np.array(self.__maximum_claim_table[i] ) - np.array(_a ) ) for i, allocated_resource in enumerate(self.__allocated_resources_table ) ] def __lowercase ( self : Tuple ): '''simple docstring''' return {self.__need().index(_a ): i for i in self.__need()} def __lowercase ( self : Optional[Any] ,**_a : Any ): '''simple docstring''' _a : Tuple = self.__need() _a : Union[str, Any] = self.__allocated_resources_table _a : Tuple = self.__available_resources() _a : List[Any] = self.__need_index_manager() for kw, val in kwargs.items(): if kw and val is True: self.__pretty_data() print('_' * 50 + '\n' ) while need_list: _a : List[Any] = False for each_need in need_list: _a : Any = True for index, need in enumerate(_a ): if need > available_resources[index]: _a : Optional[int] = False break if execution: _a : List[Any] = True # get the original index of the process from ind_ctrl db for original_need_index, need_clone in need_index_manager.items(): if each_need == need_clone: _a : Union[str, Any] = original_need_index print(F"""Process {process_number + 1} is executing.""" ) # remove the process run from stack need_list.remove(_a ) # update available/freed resources stack _a : Union[str, Any] = np.array(_a ) + np.array( alloc_resources_table[process_number] ) print( 'Updated available resource stack for processes: ' + ' '.join([str(_a ) for x in available_resources] ) ) break if safe: print('The process is in a safe state.\n' ) else: print('System in unsafe state. Aborting...\n' ) break def __lowercase ( self : Tuple ): '''simple docstring''' print(' ' * 9 + 'Allocated Resource Table' ) for item in self.__allocated_resources_table: print( F"""P{self.__allocated_resources_table.index(_a ) + 1}""" + ' '.join(F"""{it:>8}""" for it in item ) + '\n' ) print(' ' * 9 + 'System Resource Table' ) for item in self.__maximum_claim_table: print( F"""P{self.__maximum_claim_table.index(_a ) + 1}""" + ' '.join(F"""{it:>8}""" for it in item ) + '\n' ) print( 'Current Usage by Active Processes: ' + ' '.join(str(_a ) for x in self.__claim_vector ) ) print( 'Initial Available Resources: ' + ' '.join(str(_a ) for x in self.__available_resources() ) ) time.sleep(1 ) if __name__ == "__main__": import doctest doctest.testmod()
271
'''simple docstring''' from typing import List, Optional, Tuple, Union import torch from ...models import UNetaDModel from ...schedulers import KarrasVeScheduler from ...utils import randn_tensor from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : UNetaDModel __UpperCAmelCase : KarrasVeScheduler def __init__( self : Union[str, Any] ,_a : UNetaDModel ,_a : KarrasVeScheduler ): '''simple docstring''' super().__init__() self.register_modules(unet=_a ,scheduler=_a ) @torch.no_grad() def __call__( self : List[Any] ,_a : int = 1 ,_a : int = 50 ,_a : Optional[Union[torch.Generator, List[torch.Generator]]] = None ,_a : Optional[str] = "pil" ,_a : bool = True ,**_a : List[Any] ,): '''simple docstring''' _a : Any = self.unet.config.sample_size _a : Optional[int] = (batch_size, 3, img_size, img_size) _a : Dict = self.unet # sample x_0 ~ N(0, sigma_0^2 * I) _a : Dict = randn_tensor(_a ,generator=_a ,device=self.device ) * self.scheduler.init_noise_sigma self.scheduler.set_timesteps(_a ) for t in self.progress_bar(self.scheduler.timesteps ): # here sigma_t == t_i from the paper _a : Optional[int] = self.scheduler.schedule[t] _a : List[str] = self.scheduler.schedule[t - 1] if t > 0 else 0 # 1. Select temporarily increased noise level sigma_hat # 2. Add new noise to move from sample_i to sample_hat _a, _a : List[Any] = self.scheduler.add_noise_to_input(_a ,_a ,generator=_a ) # 3. Predict the noise residual given the noise magnitude `sigma_hat` # The model inputs and output are adjusted by following eq. (213) in [1]. _a : Optional[int] = (sigma_hat / 2) * model((sample_hat + 1) / 2 ,sigma_hat / 2 ).sample # 4. Evaluate dx/dt at sigma_hat # 5. Take Euler step from sigma to sigma_prev _a : Tuple = self.scheduler.step(_a ,_a ,_a ,_a ) if sigma_prev != 0: # 6. Apply 2nd order correction # The model inputs and output are adjusted by following eq. (213) in [1]. _a : Optional[int] = (sigma_prev / 2) * model((step_output.prev_sample + 1) / 2 ,sigma_prev / 2 ).sample _a : Optional[Any] = self.scheduler.step_correct( _a ,_a ,_a ,_a ,step_output.prev_sample ,step_output['derivative'] ,) _a : Dict = step_output.prev_sample _a : Tuple = (sample / 2 + 0.5).clamp(0 ,1 ) _a : Optional[Any] = sample.cpu().permute(0 ,2 ,3 ,1 ).numpy() if output_type == "pil": _a : List[str] = self.numpy_to_pil(_a ) if not return_dict: return (image,) return ImagePipelineOutput(images=_a )
271
1
'''simple docstring''' from __future__ import annotations from sys import maxsize from typing import Generic, TypeVar __lowerCAmelCase = TypeVar("""T""") def UpperCAmelCase_ (__a : int ): """simple docstring""" return (position - 1) // 2 def UpperCAmelCase_ (__a : int ): """simple docstring""" return (2 * position) + 1 def UpperCAmelCase_ (__a : int ): """simple docstring""" return (2 * position) + 2 class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Union[str, Any] ): '''simple docstring''' _a : list[tuple[T, int]] = [] _a : dict[T, int] = {} _a : int = 0 def __len__( self : Optional[int] ): '''simple docstring''' return self.elements def __repr__( self : Dict ): '''simple docstring''' return str(self.heap ) def __lowercase ( self : Any ): '''simple docstring''' return self.elements == 0 def __lowercase ( self : Union[str, Any] ,_a : T ,_a : int ): '''simple docstring''' self.heap.append((elem, weight) ) _a : List[str] = self.elements self.elements += 1 self._bubble_up(_a ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' if self.elements > 1: self._swap_nodes(0 ,self.elements - 1 ) _a, _a : Any = self.heap.pop() del self.position_map[elem] self.elements -= 1 if self.elements > 0: _a, _a : int = self.heap[0] self._bubble_down(_a ) return elem def __lowercase ( self : Union[str, Any] ,_a : T ,_a : int ): '''simple docstring''' _a : Any = self.position_map[elem] _a : str = (elem, weight) if position > 0: _a : Dict = get_parent_position(_a ) _a, _a : str = self.heap[parent_position] if parent_weight > weight: self._bubble_up(_a ) else: self._bubble_down(_a ) else: self._bubble_down(_a ) def __lowercase ( self : int ,_a : T ): '''simple docstring''' _a : Any = self.position_map[elem] if curr_pos == 0: return None _a : Optional[Any] = get_parent_position(_a ) _a, _a : Any = self.heap[curr_pos] _a, _a : Optional[int] = self.heap[parent_position] if parent_weight > weight: self._swap_nodes(_a ,_a ) return self._bubble_up(_a ) return None def __lowercase ( self : Optional[int] ,_a : T ): '''simple docstring''' _a : int = self.position_map[elem] _a, _a : Optional[int] = self.heap[curr_pos] _a : Dict = get_child_left_position(_a ) _a : List[str] = get_child_right_position(_a ) if child_left_position < self.elements and child_right_position < self.elements: _a, _a : Union[str, Any] = self.heap[child_left_position] _a, _a : Tuple = self.heap[child_right_position] if child_right_weight < child_left_weight and child_right_weight < weight: self._swap_nodes(_a ,_a ) return self._bubble_down(_a ) if child_left_position < self.elements: _a, _a : Optional[int] = self.heap[child_left_position] if child_left_weight < weight: self._swap_nodes(_a ,_a ) return self._bubble_down(_a ) else: return None if child_right_position < self.elements: _a, _a : Tuple = self.heap[child_right_position] if child_right_weight < weight: self._swap_nodes(_a ,_a ) return self._bubble_down(_a ) return None def __lowercase ( self : Optional[int] ,_a : int ,_a : int ): '''simple docstring''' _a : Optional[Any] = self.heap[nodea_pos][0] _a : Optional[Any] = self.heap[nodea_pos][0] _a, _a : Tuple = ( self.heap[nodea_pos], self.heap[nodea_pos], ) _a : Any = nodea_pos _a : Optional[int] = nodea_pos class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Optional[int] ): '''simple docstring''' _a : dict[T, dict[T, int]] = {} _a : int = 0 def __repr__( self : int ): '''simple docstring''' return str(self.connections ) def __len__( self : int ): '''simple docstring''' return self.nodes def __lowercase ( self : Optional[int] ,_a : T ): '''simple docstring''' if node not in self.connections: _a : Optional[int] = {} self.nodes += 1 def __lowercase ( self : str ,_a : T ,_a : T ,_a : int ): '''simple docstring''' self.add_node(_a ) self.add_node(_a ) _a : Optional[Any] = weight _a : Any = weight def UpperCAmelCase_ (__a : GraphUndirectedWeighted[T] , ): """simple docstring""" _a : dict[T, int] = {node: maxsize for node in graph.connections} _a : dict[T, T | None] = {node: None for node in graph.connections} _a : MinPriorityQueue[T] = MinPriorityQueue() for node, weight in dist.items(): priority_queue.push(__a , __a ) if priority_queue.is_empty(): return dist, parent # initialization _a : Optional[Any] = priority_queue.extract_min() _a : List[Any] = 0 for neighbour in graph.connections[node]: if dist[neighbour] > dist[node] + graph.connections[node][neighbour]: _a : Optional[int] = dist[node] + graph.connections[node][neighbour] priority_queue.update_key(__a , dist[neighbour] ) _a : Dict = node # running prim's algorithm while not priority_queue.is_empty(): _a : Union[str, Any] = priority_queue.extract_min() for neighbour in graph.connections[node]: if dist[neighbour] > dist[node] + graph.connections[node][neighbour]: _a : Dict = dist[node] + graph.connections[node][neighbour] priority_queue.update_key(__a , dist[neighbour] ) _a : List[str] = node return dist, parent
271
'''simple docstring''' import importlib import inspect import json import os import re import shutil import sys from pathlib import Path from typing import Dict, Optional, Union from urllib import request from huggingface_hub import HfFolder, cached_download, hf_hub_download, model_info from packaging import version from .. import __version__ from . import DIFFUSERS_DYNAMIC_MODULE_NAME, HF_MODULES_CACHE, logging __lowerCAmelCase = ( """https://raw.githubusercontent.com/huggingface/diffusers/{revision}/examples/community/{pipeline}.py""" ) __lowerCAmelCase = logging.get_logger(__name__) # pylint: disable=invalid-name def UpperCAmelCase_ (): """simple docstring""" _a : Optional[int] = 'https://pypi.org/pypi/diffusers/json' _a : int = json.loads(request.urlopen(__a ).read() )['releases'].keys() return sorted(__a , key=lambda __a : version.Version(__a ) ) def UpperCAmelCase_ (): """simple docstring""" if HF_MODULES_CACHE in sys.path: return sys.path.append(__a ) os.makedirs(__a , exist_ok=__a ) _a : str = Path(__a ) / '__init__.py' if not init_path.exists(): init_path.touch() def UpperCAmelCase_ (__a : Union[str, os.PathLike] ): """simple docstring""" init_hf_modules() _a : Dict = Path(__a ) / name # If the parent module does not exist yet, recursively create it. if not dynamic_module_path.parent.exists(): create_dynamic_module(dynamic_module_path.parent ) os.makedirs(__a , exist_ok=__a ) _a : Optional[int] = dynamic_module_path / '__init__.py' if not init_path.exists(): init_path.touch() def UpperCAmelCase_ (__a : str ): """simple docstring""" with open(__a , 'r' , encoding='utf-8' ) as f: _a : int = f.read() # Imports of the form `import .xxx` _a : Tuple = re.findall('^\s*import\s+\.(\S+)\s*$' , __a , flags=re.MULTILINE ) # Imports of the form `from .xxx import yyy` relative_imports += re.findall('^\s*from\s+\.(\S+)\s+import' , __a , flags=re.MULTILINE ) # Unique-ify return list(set(__a ) ) def UpperCAmelCase_ (__a : Any ): """simple docstring""" _a : Optional[int] = False _a : Optional[int] = [module_file] _a : List[str] = [] # Let's recurse through all relative imports while not no_change: _a : str = [] for f in files_to_check: new_imports.extend(get_relative_imports(__a ) ) _a : Union[str, Any] = Path(__a ).parent _a : str = [str(module_path / m ) for m in new_imports] _a : Tuple = [f for f in new_import_files if f not in all_relative_imports] _a : Dict = [f"""{f}.py""" for f in new_import_files] _a : List[str] = len(__a ) == 0 all_relative_imports.extend(__a ) return all_relative_imports def UpperCAmelCase_ (__a : Tuple ): """simple docstring""" with open(__a , 'r' , encoding='utf-8' ) as f: _a : Dict = f.read() # Imports of the form `import xxx` _a : Optional[int] = re.findall('^\s*import\s+(\S+)\s*$' , __a , flags=re.MULTILINE ) # Imports of the form `from xxx import yyy` imports += re.findall('^\s*from\s+(\S+)\s+import' , __a , flags=re.MULTILINE ) # Only keep the top-level module _a : List[str] = [imp.split('.' )[0] for imp in imports if not imp.startswith('.' )] # Unique-ify and test we got them all _a : Optional[int] = list(set(__a ) ) _a : List[str] = [] for imp in imports: try: importlib.import_module(__a ) except ImportError: missing_packages.append(__a ) if len(__a ) > 0: raise ImportError( 'This modeling file requires the following packages that were not found in your environment: ' f"""{', '.join(__a )}. Run `pip install {' '.join(__a )}`""" ) return get_relative_imports(__a ) def UpperCAmelCase_ (__a : Any , __a : str ): """simple docstring""" _a : Any = module_path.replace(os.path.sep , '.' ) _a : Union[str, Any] = importlib.import_module(__a ) if class_name is None: return find_pipeline_class(__a ) return getattr(__a , __a ) def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" from ..pipelines import DiffusionPipeline _a : List[str] = dict(inspect.getmembers(__a , inspect.isclass ) ) _a : str = None for cls_name, cls in cls_members.items(): if ( cls_name != DiffusionPipeline.__name__ and issubclass(cls , __a ) and cls.__module__.split('.' )[0] != "diffusers" ): if pipeline_class is not None: raise ValueError( f"""Multiple classes that inherit from {DiffusionPipeline.__name__} have been found:""" f""" {pipeline_class.__name__}, and {cls_name}. Please make sure to define only one in""" f""" {loaded_module}.""" ) _a : Any = cls return pipeline_class def UpperCAmelCase_ (__a : Union[str, os.PathLike] , __a : str , __a : Optional[Union[str, os.PathLike]] = None , __a : bool = False , __a : bool = False , __a : Optional[Dict[str, str]] = None , __a : Optional[Union[bool, str]] = None , __a : Optional[str] = None , __a : bool = False , ): """simple docstring""" _a : str = str(__a ) _a : Optional[Any] = os.path.join(__a , __a ) if os.path.isfile(__a ): _a : Tuple = module_file_or_url _a : Optional[Any] = 'local' elif pretrained_model_name_or_path.count('/' ) == 0: _a : int = get_diffusers_versions() # cut ".dev0" _a : Any = 'v' + '.'.join(__version__.split('.' )[:3] ) # retrieve github version that matches if revision is None: _a : Any = latest_version if latest_version[1:] in available_versions else 'main' logger.info(f"""Defaulting to latest_version: {revision}.""" ) elif revision in available_versions: _a : Any = f"""v{revision}""" elif revision == "main": _a : Optional[int] = revision else: raise ValueError( f"""`custom_revision`: {revision} does not exist. Please make sure to choose one of""" f""" {', '.join(available_versions + ['main'] )}.""" ) # community pipeline on GitHub _a : Tuple = COMMUNITY_PIPELINES_URL.format(revision=__a , pipeline=__a ) try: _a : Any = cached_download( __a , cache_dir=__a , force_download=__a , proxies=__a , resume_download=__a , local_files_only=__a , use_auth_token=__a , ) _a : List[Any] = 'git' _a : Any = pretrained_model_name_or_path + '.py' except EnvironmentError: logger.error(f"""Could not locate the {module_file} inside {pretrained_model_name_or_path}.""" ) raise else: try: # Load from URL or cache if already cached _a : Optional[Any] = hf_hub_download( __a , __a , cache_dir=__a , force_download=__a , proxies=__a , resume_download=__a , local_files_only=__a , use_auth_token=__a , ) _a : List[Any] = os.path.join('local' , '--'.join(pretrained_model_name_or_path.split('/' ) ) ) except EnvironmentError: logger.error(f"""Could not locate the {module_file} inside {pretrained_model_name_or_path}.""" ) raise # Check we have all the requirements in our environment _a : Optional[int] = check_imports(__a ) # Now we move the module inside our cached dynamic modules. _a : Optional[Any] = DIFFUSERS_DYNAMIC_MODULE_NAME + os.path.sep + submodule create_dynamic_module(__a ) _a : Any = Path(__a ) / full_submodule if submodule == "local" or submodule == "git": # We always copy local files (we could hash the file to see if there was a change, and give them the name of # that hash, to only copy when there is a modification but it seems overkill for now). # The only reason we do the copy is to avoid putting too many folders in sys.path. shutil.copy(__a , submodule_path / module_file ) for module_needed in modules_needed: _a : Dict = f"""{module_needed}.py""" shutil.copy(os.path.join(__a , __a ) , submodule_path / module_needed ) else: # Get the commit hash # TODO: we will get this info in the etag soon, so retrieve it from there and not here. if isinstance(__a , __a ): _a : Optional[Any] = use_auth_token elif use_auth_token is True: _a : List[Any] = HfFolder.get_token() else: _a : Dict = None _a : int = model_info(__a , revision=__a , token=__a ).sha # The module file will end up being placed in a subfolder with the git hash of the repo. This way we get the # benefit of versioning. _a : Optional[int] = submodule_path / commit_hash _a : str = full_submodule + os.path.sep + commit_hash create_dynamic_module(__a ) if not (submodule_path / module_file).exists(): shutil.copy(__a , submodule_path / module_file ) # Make sure we also have every file with relative for module_needed in modules_needed: if not (submodule_path / module_needed).exists(): get_cached_module_file( __a , f"""{module_needed}.py""" , cache_dir=__a , force_download=__a , resume_download=__a , proxies=__a , use_auth_token=__a , revision=__a , local_files_only=__a , ) return os.path.join(__a , __a ) def UpperCAmelCase_ (__a : Union[str, os.PathLike] , __a : str , __a : Optional[str] = None , __a : Optional[Union[str, os.PathLike]] = None , __a : bool = False , __a : bool = False , __a : Optional[Dict[str, str]] = None , __a : Optional[Union[bool, str]] = None , __a : Optional[str] = None , __a : bool = False , **__a : str , ): """simple docstring""" _a : Dict = get_cached_module_file( __a , __a , cache_dir=__a , force_download=__a , resume_download=__a , proxies=__a , use_auth_token=__a , revision=__a , local_files_only=__a , ) return get_class_in_module(__a , final_module.replace('.py' , '' ) )
271
1
'''simple docstring''' import random import torch from huggingface_hub import HfApi from diffusers import UNetaDModel __lowerCAmelCase = HfApi() __lowerCAmelCase = {} # fmt: off __lowerCAmelCase = torch.tensor([ -0.7_515, -1.6_883, 0.2_420, 0.0_300, 0.6_347, 1.3_433, -1.1_743, -3.7_467, 1.2_342, -2.2_485, 0.4_636, 0.8_076, -0.7_991, 0.3_969, 0.8_498, 0.9_189, -1.8_887, -3.3_522, 0.7_639, 0.2_040, 0.6_271, -2.7_148, -1.6_316, 3.0_839, 0.3_186, 0.2_721, -0.9_759, -1.2_461, 2.6_257, 1.3_557 ]) __lowerCAmelCase = torch.tensor([ -2.3_639, -2.5_344, 0.0_054, -0.6_674, 1.5_990, 1.0_158, 0.3_124, -2.1_436, 1.8_795, -2.5_429, -0.1_566, -0.3_973, 1.2_490, 2.6_447, 1.2_283, -0.5_208, -2.8_154, -3.5_119, 2.3_838, 1.2_033, 1.7_201, -2.1_256, -1.4_576, 2.7_948, 2.4_204, -0.9_752, -1.2_546, 0.8_027, 3.2_758, 3.1_365 ]) __lowerCAmelCase = torch.tensor([ -0.6_531, -0.6_891, -0.3_172, -0.5_375, -0.9_140, -0.5_367, -0.1_175, -0.7_869, -0.3_808, -0.4_513, -0.2_098, -0.0_083, 0.3_183, 0.5_140, 0.2_247, -0.1_304, -0.1_302, -0.2_802, -0.2_084, -0.2_025, -0.4_967, -0.4_873, -0.0_861, 0.6_925, 0.0_250, 0.1_290, -0.1_543, 0.6_316, 1.0_460, 1.4_943 ]) __lowerCAmelCase = torch.tensor([ 0.0_911, 0.1_107, 0.0_182, 0.0_435, -0.0_805, -0.0_608, 0.0_381, 0.2_172, -0.0_280, 0.1_327, -0.0_299, -0.0_255, -0.0_050, -0.1_170, -0.1_046, 0.0_309, 0.1_367, 0.1_728, -0.0_533, -0.0_748, -0.0_534, 0.1_624, 0.0_384, -0.1_805, -0.0_707, 0.0_642, 0.0_220, -0.0_134, -0.1_333, -0.1_505 ]) __lowerCAmelCase = torch.tensor([ 0.1_321, 0.1_337, 0.0_440, 0.0_622, -0.0_591, -0.0_370, 0.0_503, 0.2_133, -0.0_177, 0.1_415, -0.0_116, -0.0_112, 0.0_044, -0.0_980, -0.0_789, 0.0_395, 0.1_502, 0.1_785, -0.0_488, -0.0_514, -0.0_404, 0.1_539, 0.0_454, -0.1_559, -0.0_665, 0.0_659, 0.0_383, -0.0_005, -0.1_266, -0.1_386 ]) __lowerCAmelCase = torch.tensor([ 0.1_154, 0.1_218, 0.0_307, 0.0_526, -0.0_711, -0.0_541, 0.0_366, 0.2_078, -0.0_267, 0.1_317, -0.0_226, -0.0_193, -0.0_014, -0.1_055, -0.0_902, 0.0_330, 0.1_391, 0.1_709, -0.0_562, -0.0_693, -0.0_560, 0.1_482, 0.0_381, -0.1_683, -0.0_681, 0.0_661, 0.0_331, -0.0_046, -0.1_268, -0.1_431 ]) __lowerCAmelCase = torch.tensor([ 0.1_192, 0.1_240, 0.0_414, 0.0_606, -0.0_557, -0.0_412, 0.0_430, 0.2_042, -0.0_200, 0.1_385, -0.0_115, -0.0_132, 0.0_017, -0.0_965, -0.0_802, 0.0_398, 0.1_433, 0.1_747, -0.0_458, -0.0_533, -0.0_407, 0.1_545, 0.0_419, -0.1_574, -0.0_645, 0.0_626, 0.0_341, -0.0_010, -0.1_199, -0.1_390 ]) __lowerCAmelCase = torch.tensor([ 0.1_075, 0.1_074, 0.0_205, 0.0_431, -0.0_774, -0.0_607, 0.0_298, 0.2_042, -0.0_320, 0.1_267, -0.0_281, -0.0_250, -0.0_064, -0.1_091, -0.0_946, 0.0_290, 0.1_328, 0.1_650, -0.0_580, -0.0_738, -0.0_586, 0.1_440, 0.0_337, -0.1_746, -0.0_712, 0.0_605, 0.0_250, -0.0_099, -0.1_316, -0.1_473 ]) __lowerCAmelCase = torch.tensor([ -1.4_572, -2.0_481, -0.0_414, -0.6_005, 1.4_136, 0.5_848, 0.4_028, -2.7_330, 1.2_212, -2.1_228, 0.2_155, 0.4_039, 0.7_662, 2.0_535, 0.7_477, -0.3_243, -2.1_758, -2.7_648, 1.6_947, 0.7_026, 1.2_338, -1.6_078, -0.8_682, 2.2_810, 1.8_574, -0.5_718, -0.5_586, -0.0_186, 2.3_415, 2.1_251]) __lowerCAmelCase = torch.tensor([ -1.3_690, -1.9_720, -0.4_090, -0.6_966, 1.4_660, 0.9_938, -0.1_385, -2.7_324, 0.7_736, -1.8_917, 0.2_923, 0.4_293, 0.1_693, 1.4_112, 1.1_887, -0.3_181, -2.2_160, -2.6_381, 1.3_170, 0.8_163, 0.9_240, -1.6_544, -0.6_099, 2.5_259, 1.6_430, -0.9_090, -0.9_392, -0.0_126, 2.4_268, 2.3_266 ]) __lowerCAmelCase = torch.tensor([ -1.3_525, -1.9_628, -0.3_956, -0.6_860, 1.4_664, 1.0_014, -0.1_259, -2.7_212, 0.7_772, -1.8_811, 0.2_996, 0.4_388, 0.1_704, 1.4_029, 1.1_701, -0.3_027, -2.2_053, -2.6_287, 1.3_350, 0.8_131, 0.9_274, -1.6_292, -0.6_098, 2.5_131, 1.6_505, -0.8_958, -0.9_298, -0.0_151, 2.4_257, 2.3_355 ]) __lowerCAmelCase = torch.tensor([ -2.0_585, -2.7_897, -0.2_850, -0.8_940, 1.9_052, 0.5_702, 0.6_345, -3.8_959, 1.5_932, -3.2_319, 0.1_974, 0.0_287, 1.7_566, 2.6_543, 0.8_387, -0.5_351, -3.2_736, -4.3_375, 2.9_029, 1.6_390, 1.4_640, -2.1_701, -1.9_013, 2.9_341, 3.4_981, -0.6_255, -1.1_644, -0.1_591, 3.7_097, 3.2_066 ]) __lowerCAmelCase = torch.tensor([ -2.3_139, -2.5_594, -0.0_197, -0.6_785, 1.7_001, 1.1_606, 0.3_075, -2.1_740, 1.8_071, -2.5_630, -0.0_926, -0.3_811, 1.2_116, 2.6_246, 1.2_731, -0.5_398, -2.8_153, -3.6_140, 2.3_893, 1.3_262, 1.6_258, -2.1_856, -1.3_267, 2.8_395, 2.3_779, -1.0_623, -1.2_468, 0.8_959, 3.3_367, 3.2_243 ]) __lowerCAmelCase = torch.tensor([ -2.0_628, -2.7_667, -0.2_089, -0.8_263, 2.0_539, 0.5_992, 0.6_495, -3.8_336, 1.6_025, -3.2_817, 0.1_721, -0.0_633, 1.7_516, 2.7_039, 0.8_100, -0.5_908, -3.2_113, -4.4_343, 2.9_257, 1.3_632, 1.5_562, -2.1_489, -1.9_894, 3.0_560, 3.3_396, -0.7_328, -1.0_417, 0.0_383, 3.7_093, 3.2_343 ]) __lowerCAmelCase = torch.tensor([ -1.4_574, -2.0_569, -0.0_473, -0.6_117, 1.4_018, 0.5_769, 0.4_129, -2.7_344, 1.2_241, -2.1_397, 0.2_000, 0.3_937, 0.7_616, 2.0_453, 0.7_324, -0.3_391, -2.1_746, -2.7_744, 1.6_963, 0.6_921, 1.2_187, -1.6_172, -0.8_877, 2.2_439, 1.8_471, -0.5_839, -0.5_605, -0.0_464, 2.3_250, 2.1_219 ]) # fmt: on __lowerCAmelCase = api.list_models(filter="""diffusers""") for mod in models: if "google" in mod.author or mod.modelId == "CompVis/ldm-celebahq-256": __lowerCAmelCase = """/home/patrick/google_checkpoints/""" + mod.modelId.split("""/""")[-1] print(f'''Started running {mod.modelId}!!!''') if mod.modelId.startswith("""CompVis"""): __lowerCAmelCase = UNetaDModel.from_pretrained(local_checkpoint, subfolder="""unet""") else: __lowerCAmelCase = UNetaDModel.from_pretrained(local_checkpoint) torch.manual_seed(0) random.seed(0) __lowerCAmelCase = torch.randn(1, model.config.in_channels, model.config.sample_size, model.config.sample_size) __lowerCAmelCase = torch.tensor([1_0] * noise.shape[0]) with torch.no_grad(): __lowerCAmelCase = model(noise, time_step).sample assert torch.allclose( logits[0, 0, 0, :3_0], results["""_""".join("""_""".join(mod.modelId.split("""/""")).split("""-"""))], atol=1e-3 ) print(f'''{mod.modelId} has passed successfully!!!''')
271
'''simple docstring''' def UpperCAmelCase_ (__a : list , __a : list , __a : int ): """simple docstring""" _a : Optional[Any] = len(__a ) _a : int = [[0] * n for i in range(__a )] for i in range(__a ): _a : Tuple = y_points[i] for i in range(2 , __a ): for j in range(__a , __a ): _a : Tuple = ( (xa - x_points[j - i + 1]) * q[j][i - 1] - (xa - x_points[j]) * q[j - 1][i - 1] ) / (x_points[j] - x_points[j - i + 1]) return [q[n - 1][n - 1], q] if __name__ == "__main__": import doctest doctest.testmod()
271
1
'''simple docstring''' import os from huggingface_hub.constants import HUGGINGFACE_HUB_CACHE, hf_cache_home __lowerCAmelCase = HUGGINGFACE_HUB_CACHE __lowerCAmelCase = """config.json""" __lowerCAmelCase = """diffusion_pytorch_model.bin""" __lowerCAmelCase = """diffusion_flax_model.msgpack""" __lowerCAmelCase = """model.onnx""" __lowerCAmelCase = """diffusion_pytorch_model.safetensors""" __lowerCAmelCase = """weights.pb""" __lowerCAmelCase = """https://huggingface.co""" __lowerCAmelCase = default_cache_path __lowerCAmelCase = """diffusers_modules""" __lowerCAmelCase = os.getenv("""HF_MODULES_CACHE""", os.path.join(hf_cache_home, """modules""")) __lowerCAmelCase = ["""fp16""", """non-ema"""] __lowerCAmelCase = """.self_attn"""
271
'''simple docstring''' import inspect import os import unittest from pathlib import Path import torch import accelerate from accelerate.test_utils import execute_subprocess_async from accelerate.test_utils.testing import run_command class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[Any] = inspect.getfile(accelerate.test_utils ) __UpperCAmelCase : List[str] = os.path.sep.join(mod_file.split(os.path.sep )[:-1] + ['''scripts''', '''test_cli.py'''] ) __UpperCAmelCase : Dict = ['''accelerate''', '''launch'''] __UpperCAmelCase : Dict = Path.home() / '''.cache/huggingface/accelerate''' __UpperCAmelCase : Dict = '''default_config.yaml''' __UpperCAmelCase : Optional[Any] = config_folder / config_file __UpperCAmelCase : Dict = config_folder / '''_default_config.yaml''' __UpperCAmelCase : Any = Path('''tests/test_configs''' ) @classmethod def __lowercase ( cls : int ): '''simple docstring''' if cls.config_path.is_file(): cls.config_path.rename(cls.changed_path ) @classmethod def __lowercase ( cls : List[Any] ): '''simple docstring''' if cls.changed_path.is_file(): cls.changed_path.rename(cls.config_path ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Dict = self.base_cmd if torch.cuda.is_available() and (torch.cuda.device_count() > 1): cmd += ["--multi_gpu"] execute_subprocess_async(cmd + [self.test_file_path] ,env=os.environ.copy() ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' for config in sorted(self.test_config_path.glob('**/*.yaml' ) ): with self.subTest(config_file=_a ): execute_subprocess_async( self.base_cmd + ['--config_file', str(_a ), self.test_file_path] ,env=os.environ.copy() ) def __lowercase ( self : Optional[int] ): '''simple docstring''' execute_subprocess_async(['accelerate', 'test'] ,env=os.environ.copy() ) class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[Any] = '''test-tpu''' __UpperCAmelCase : Any = '''us-central1-a''' __UpperCAmelCase : List[Any] = '''ls''' __UpperCAmelCase : Any = ['''accelerate''', '''tpu-config'''] __UpperCAmelCase : Dict = '''cd /usr/share''' __UpperCAmelCase : Any = '''tests/test_samples/test_command_file.sh''' __UpperCAmelCase : List[Any] = '''Running gcloud compute tpus tpu-vm ssh''' def __lowercase ( self : Dict ): '''simple docstring''' _a : Optional[Any] = run_command( self.cmd + ['--command', self.command, '--tpu_zone', self.tpu_zone, '--tpu_name', self.tpu_name, '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls --worker all""" ,_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Any = run_command( self.cmd + [ '--config_file', 'tests/test_configs/0_12_0.yaml', '--command', self.command, '--tpu_zone', self.tpu_zone, '--tpu_name', self.tpu_name, '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls --worker all""" ,_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--debug'] ,return_stdout=_a ) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : int ): '''simple docstring''' _a : Optional[Any] = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--command', self.command, '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls --worker all""" ,_a ,) def __lowercase ( self : str ): '''simple docstring''' _a : List[str] = run_command( self.cmd + [ '--config_file', 'tests/test_configs/latest.yaml', '--command', self.command, '--command', 'echo "Hello World"', '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls; echo \"Hello World\" --worker all""" ,_a ,) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Any = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--command_file', self.command_file, '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Union[str, Any] = run_command( self.cmd + [ '--config_file', 'tests/test_configs/0_12_0.yaml', '--command_file', self.command_file, '--tpu_zone', self.tpu_zone, '--tpu_name', self.tpu_name, '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : Any ): '''simple docstring''' _a : Optional[int] = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--install_accelerate', '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; pip install accelerate -U; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = run_command( self.cmd + [ '--config_file', 'tests/test_configs/latest.yaml', '--install_accelerate', '--accelerate_version', '12.0.0', '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; pip install accelerate==12.0.0; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,)
271
1
'''simple docstring''' import argparse import json import os from collections import OrderedDict import numpy as np import tensorflow as tf import torch def UpperCAmelCase_ (__a : Dict ): """simple docstring""" _a : List[Any] = os.path.join(args.tf_model_dir , 'parameters.json' ) _a : List[Any] = json.loads(open(__a ).read() ) if not params: raise ValueError( f"""It seems that the json file at {parameter_file} is empty. Make sure you have a correct json file.""" ) if not args.output.endswith('.pt' ): _a : Optional[int] = args.output + '.pt' _a : str = OrderedDict() with tf.device('/CPU:0' ): _a : Dict = tf.train.load_checkpoint(args.tf_model_dir ) _a : Union[str, Any] = reader.get_variable_to_shape_map() for key_name in shapes.keys(): _a : Optional[Any] = reader.get_tensor(__a ).astype(np.floataa ) if key_name.endswith('/adam_m' ) or key_name.endswith('/adam_v' ): continue if key_name.startswith('pasts/' ): if key_name.startswith('pasts/mlp' ): _a : int = int(key_name[9] ) elif key_name.startswith('pasts/out' ): _a : Tuple = 8 _a : Tuple = 'model.sqout.%d.weight' % (player * 2) # enter to nn.Sequencial with Tanh, so 2 at a time _a : Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix _a : Union[str, Any] = torch.tensor(__a ) elif key_name.startswith('model/moe' ): _a : str = int(key_name[9:].split('/' )[0] ) if key_name.endswith('/switch_gating/kernel' ): _a : Union[str, Any] = 'model.blocks.%d.feed_forward.mlp.router.classifier.weight' % player _a : Optional[int] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix _a : List[str] = torch.tensor(__a ) elif key_name.endswith('/softmlp/kernel' ): _a : List[str] = 'model.blocks.%d.feed_forward.soft_bypass_mlp.weight' % player _a : List[str] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix _a : str = torch.tensor(__a ) elif key_name.endswith('/wo/kernel' ) or key_name.endswith('/wi/kernel' ): _a : Union[str, Any] = key_name[-9:-7] for i in range(1_6 ): _a : str = 'model.blocks.%d.feed_forward.mlp.experts.expert_%d.%s.weight' % (player, i, nlayer) _a : List[Any] = ( vnp[i].transpose([1, 0] ).copy() ) # In Mesh-Tensorflow, it is one array, so it is divided _a : str = torch.tensor(__a ) elif key_name.startswith('model/mlp' ): _a : Optional[Any] = int(key_name[9:].split('/' )[0] ) if key_name.endswith('/p1/kernel' ): _a : List[Any] = 'model.blocks.%d.feed_forward.mlp.wi.weight' % player _a : List[str] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix _a : Tuple = torch.tensor(__a ) elif key_name.endswith('/p1/bias' ): _a : List[str] = 'model.blocks.%d.feed_forward.mlp.wi.bias' % player _a : Optional[int] = vnp.copy() # same because it is one dimensional _a : int = torch.tensor(__a ) elif key_name.endswith('/p2/kernel' ): _a : str = 'model.blocks.%d.feed_forward.mlp.wo.weight' % player _a : Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix _a : Union[str, Any] = torch.tensor(__a ) elif key_name.endswith('/p2/bias' ): _a : Union[str, Any] = 'model.blocks.%d.feed_forward.mlp.wo.bias' % player _a : Optional[int] = vnp.copy() # same because it is one dimensional _a : int = torch.tensor(__a ) elif key_name.startswith('model/ln' ): _a : List[Any] = int(key_name[8:].split('/' )[0] ) if key_name.endswith('/b' ): _a : List[Any] = 'model.blocks.%d.feed_forward.norm.bias' % player _a : Tuple = vnp.copy() # same because it is one dimensional _a : Union[str, Any] = torch.tensor(__a ) elif key_name.endswith('/g' ): _a : str = 'model.blocks.%d.feed_forward.norm.weight' % player _a : Dict = vnp.copy() # same because it is one dimensional _a : str = torch.tensor(__a ) elif key_name.startswith('model/att' ): _a : List[Any] = int(key_name[9:].split('/' )[0] ) if key_name.endswith('/qkv/kernel' ): _a : Dict = vnp.copy() # Compute same dimension as Mesh-tensorflow using einsum _a : str = state[:, 0, :, :] _a : Optional[Any] = state[:, 1, :, :] _a : List[str] = state[:, 2, :, :] _a : Dict = ( state_q.reshape([state_q.shape[0], state_q.shape[1] * state_q.shape[2]] ) .transpose([1, 0] ) .copy() ) # Mesh-Tensorflow is a diagonal matrix _a : Dict = ( state_k.reshape([state_k.shape[0], state_k.shape[1] * state_k.shape[2]] ) .transpose([1, 0] ) .copy() ) # Mesh-Tensorflow is a diagonal matrix _a : Union[str, Any] = ( state_v.reshape([state_v.shape[0], state_v.shape[1] * state_v.shape[2]] ) .transpose([1, 0] ) .copy() ) # Mesh-Tensorflow is a diagonal matrix _a : List[Any] = 'model.blocks.%d.self_attn.self_attn.q_proj.weight' % player _a : Any = torch.tensor(__a ) _a : Optional[int] = 'model.blocks.%d.self_attn.self_attn.k_proj.weight' % player _a : str = torch.tensor(__a ) _a : Optional[int] = 'model.blocks.%d.self_attn.self_attn.v_proj.weight' % player _a : List[str] = torch.tensor(__a ) elif key_name.endswith('/o/kernel' ): _a : List[Any] = 'model.blocks.%d.self_attn.self_attn.out_proj.weight' % player _a : Any = ( vnp.reshape([vnp.shape[0] * vnp.shape[1], vnp.shape[2]] ).transpose([1, 0] ).copy() ) # Mesh-Tensorflow is a diagonal matrix _a : List[str] = torch.tensor(__a ) elif key_name.startswith('model/an' ): _a : List[Any] = int(key_name[8:].split('/' )[0] ) if key_name.endswith('/b' ): _a : Optional[int] = 'model.blocks.%d.self_attn.norm.bias' % player _a : Optional[int] = vnp.copy() # same because it is one dimensional _a : Union[str, Any] = torch.tensor(__a ) elif key_name.endswith('/g' ): _a : int = 'model.blocks.%d.self_attn.norm.weight' % player _a : Optional[Any] = vnp.copy() # same because it is one dimensional _a : Optional[int] = torch.tensor(__a ) elif ( key_name.startswith('model/wte' ) or key_name.startswith('model/wpe' ) or key_name.startswith('model/ete' ) ): _a : Optional[Any] = {'wte': 'embed_tokens', 'wpe': 'position_embeddings', 'ete': 'extra_position_embeddings'}[ key_name[-3:] ] _a : List[str] = 'model.%s.weight' % nlayer _a : Any = vnp.copy() # same in embedded _a : int = torch.tensor(__a ) if key_name.startswith('model/wte' ): _a : Tuple = 'lm_head.weight' _a : List[Any] = vnp.copy() # same in embedded _a : int = torch.tensor(__a ) elif key_name.startswith('model/wob' ): _a : Union[str, Any] = 'final_logits_bias' _a : List[Any] = vnp.copy() # same in embedded _a : Any = state.reshape((1, -1) ) _a : List[Any] = torch.tensor(__a ) elif key_name == "model/dense/kernel": _a : Optional[Any] = 'model.last_project.weight' _a : List[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix _a : Any = torch.tensor(__a ) elif key_name == "model/dense_1/bias": _a : Union[str, Any] = 'model.last_project.bias' _a : Tuple = vnp.copy() # same because it is one dimensional _a : str = torch.tensor(__a ) torch.save(__a , args.output ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser( description="""model converter.""", formatter_class=argparse.ArgumentDefaultsHelpFormatter ) parser.add_argument("""--tf_model_dir""", metavar="""PATH""", type=str, required=True, help="""import model""") parser.add_argument("""--output""", metavar="""PATH""", type=str, required=True, help="""output model""") __lowerCAmelCase = parser.parse_args() convert_tf_gptsan_to_pt(args)
271
'''simple docstring''' from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar __lowerCAmelCase = TypeVar("""T""") class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Tuple ,_a : T ): '''simple docstring''' _a : List[str] = data _a : Node[T] | None = None def __str__( self : Dict ): '''simple docstring''' return F"""{self.data}""" class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Optional[int] ): '''simple docstring''' _a : Node[T] | None = None def __iter__( self : str ): '''simple docstring''' _a : Tuple = self.top while node: yield node.data _a : int = node.next def __str__( self : str ): '''simple docstring''' return "->".join([str(_a ) for item in self] ) def __len__( self : Optional[Any] ): '''simple docstring''' return len(tuple(iter(self ) ) ) def __lowercase ( self : str ): '''simple docstring''' return self.top is None def __lowercase ( self : List[Any] ,_a : T ): '''simple docstring''' _a : int = Node(_a ) if not self.is_empty(): _a : Optional[Any] = self.top _a : List[str] = node def __lowercase ( self : Tuple ): '''simple docstring''' if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top ,_a ) _a : List[Any] = self.top _a : int = self.top.next return pop_node.data def __lowercase ( self : List[str] ): '''simple docstring''' if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = None if __name__ == "__main__": from doctest import testmod testmod()
271
1
'''simple docstring''' from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_tf_available, is_torch_available, ) __lowerCAmelCase = {"""configuration_encoder_decoder""": ["""EncoderDecoderConfig"""]} try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""EncoderDecoderModel"""] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""TFEncoderDecoderModel"""] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = ["""FlaxEncoderDecoderModel"""] if TYPE_CHECKING: from .configuration_encoder_decoder import EncoderDecoderConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_encoder_decoder import EncoderDecoderModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_encoder_decoder import TFEncoderDecoderModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_encoder_decoder import FlaxEncoderDecoderModel else: import sys __lowerCAmelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
271
'''simple docstring''' import unittest import numpy as np import torch from diffusers import PNDMPipeline, PNDMScheduler, UNetaDModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : List[str] ): '''simple docstring''' torch.manual_seed(0 ) _a : int = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : str = self.dummy_uncond_unet _a : int = PNDMScheduler() _a : str = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Any = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ,return_dict=_a )[0] _a : List[Any] = image[0, -3:, -3:, -1] _a : Any = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : List[Any] = np.array([1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2 assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1E-2 @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[str] = 'google/ddpm-cifar10-32' _a : str = UNetaDModel.from_pretrained(_a ) _a : Union[str, Any] = PNDMScheduler() _a : Tuple = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : str = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,output_type='numpy' ).images _a : Dict = image[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : Tuple = np.array([0.1564, 0.1_4645, 0.1406, 0.1_4715, 0.1_2425, 0.1_4045, 0.1_3115, 0.1_2175, 0.125] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
271
1
'''simple docstring''' import shutil import tempfile import unittest import numpy as np import pytest from transformers import is_speech_available, is_vision_available from transformers.testing_utils import require_torch if is_vision_available(): from transformers import TvltImageProcessor if is_speech_available(): from transformers import TvltFeatureExtractor from transformers import TvltProcessor @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Tuple ): '''simple docstring''' _a : str = 'ZinengTang/tvlt-base' _a : Optional[Any] = tempfile.mkdtemp() def __lowercase ( self : List[str] ,**_a : int ): '''simple docstring''' return TvltImageProcessor.from_pretrained(self.checkpoint ,**_a ) def __lowercase ( self : List[str] ,**_a : Optional[Any] ): '''simple docstring''' return TvltFeatureExtractor.from_pretrained(self.checkpoint ,**_a ) def __lowercase ( self : List[Any] ): '''simple docstring''' shutil.rmtree(self.tmpdirname ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : int = self.get_image_processor() _a : Optional[Any] = self.get_feature_extractor() _a : Any = TvltProcessor(image_processor=_a ,feature_extractor=_a ) processor.save_pretrained(self.tmpdirname ) _a : Any = TvltProcessor.from_pretrained(self.tmpdirname ) self.assertIsInstance(processor.feature_extractor ,_a ) self.assertIsInstance(processor.image_processor ,_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : str = self.get_image_processor() _a : Any = self.get_feature_extractor() _a : List[str] = TvltProcessor(image_processor=_a ,feature_extractor=_a ) _a : Tuple = np.ones([1_2000] ) _a : Optional[Any] = feature_extractor(_a ,return_tensors='np' ) _a : str = processor(audio=_a ,return_tensors='np' ) for key in audio_dict.keys(): self.assertAlmostEqual(audio_dict[key].sum() ,input_processor[key].sum() ,delta=1E-2 ) def __lowercase ( self : str ): '''simple docstring''' _a : Tuple = self.get_image_processor() _a : Tuple = self.get_feature_extractor() _a : List[str] = TvltProcessor(image_processor=_a ,feature_extractor=_a ) _a : Dict = np.ones([3, 224, 224] ) _a : List[str] = image_processor(_a ,return_tensors='np' ) _a : Tuple = processor(images=_a ,return_tensors='np' ) for key in image_dict.keys(): self.assertAlmostEqual(image_dict[key].sum() ,input_processor[key].sum() ,delta=1E-2 ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : int = self.get_image_processor() _a : List[str] = self.get_feature_extractor() _a : Optional[Any] = TvltProcessor(image_processor=_a ,feature_extractor=_a ) _a : int = np.ones([1_2000] ) _a : int = np.ones([3, 224, 224] ) _a : str = processor(audio=_a ,images=_a ) self.assertListEqual(list(inputs.keys() ) ,['audio_values', 'audio_mask', 'pixel_values', 'pixel_mask'] ) # test if it raises when no input is passed with pytest.raises(_a ): processor() def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : int = self.get_image_processor() _a : Tuple = self.get_feature_extractor() _a : str = TvltProcessor(image_processor=_a ,feature_extractor=_a ) self.assertListEqual( processor.model_input_names ,image_processor.model_input_names + feature_extractor.model_input_names ,msg='`processor` and `image_processor`+`feature_extractor` model input names do not match' ,)
271
'''simple docstring''' import doctest import logging import os import unittest from pathlib import Path from typing import List, Union import transformers from transformers.testing_utils import require_tf, require_torch, slow __lowerCAmelCase = logging.getLogger() @unittest.skip('''Temporarily disable the doc tests.''' ) @require_torch @require_tf @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : str ,_a : Path ,_a : Union[str, None] = None ,_a : Union[List[str], None] = None ,_a : Union[str, List[str], None] = None ,_a : bool = True ,): '''simple docstring''' _a : Optional[int] = [file for file in os.listdir(_a ) if os.path.isfile(os.path.join(_a ,_a ) )] if identifier is not None: _a : List[str] = [file for file in files if identifier in file] if n_identifier is not None: if isinstance(_a ,_a ): for n_ in n_identifier: _a : Tuple = [file for file in files if n_ not in file] else: _a : Optional[Any] = [file for file in files if n_identifier not in file] _a : List[str] = ignore_files or [] ignore_files.append('__init__.py' ) _a : Tuple = [file for file in files if file not in ignore_files] for file in files: # Open all files print('Testing' ,_a ) if only_modules: _a : Any = file.split('.' )[0] try: _a : List[str] = getattr(_a ,_a ) _a : int = doctest.DocTestSuite(_a ) _a : Any = unittest.TextTestRunner().run(_a ) self.assertIs(len(result.failures ) ,0 ) except AttributeError: logger.info(F"""{module_identifier} is not a module.""" ) else: _a : Union[str, Any] = doctest.testfile(str('..' / directory / file ) ,optionflags=doctest.ELLIPSIS ) self.assertIs(result.failed ,0 ) def __lowercase ( self : Any ): '''simple docstring''' _a : int = Path('src/transformers' ) _a : List[Any] = 'modeling' _a : Optional[Any] = [ 'modeling_ctrl.py', 'modeling_tf_ctrl.py', ] self.analyze_directory(_a ,identifier=_a ,ignore_files=_a ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Optional[Any] = Path('src/transformers' ) _a : Optional[Any] = 'tokenization' self.analyze_directory(_a ,identifier=_a ) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Dict = Path('src/transformers' ) _a : str = 'configuration' self.analyze_directory(_a ,identifier=_a ) def __lowercase ( self : Any ): '''simple docstring''' _a : Tuple = Path('src/transformers' ) _a : List[Any] = ['configuration', 'modeling', 'tokenization'] self.analyze_directory(_a ,n_identifier=_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[Any] = Path('docs/source' ) _a : List[str] = ['favicon.ico'] self.analyze_directory(_a ,ignore_files=_a ,only_modules=_a )
271
1
'''simple docstring''' # limitations under the License. # NOTE: This file is deprecated and will be removed in a future version. # It only exists so that temporarely `from diffusers.pipelines import DiffusionPipeline` works from .pipelines import DiffusionPipeline, ImagePipelineOutput # noqa: F401 from .utils import deprecate deprecate( """pipelines_utils""", """0.22.0""", """Importing `DiffusionPipeline` or `ImagePipelineOutput` from diffusers.pipeline_utils is deprecated. Please import from diffusers.pipelines.pipeline_utils instead.""", standard_warn=False, stacklevel=3, )
271
'''simple docstring''' import argparse import pickle import numpy as np import torch from torch import nn from transformers import ReformerConfig, ReformerModelWithLMHead from transformers.utils import logging logging.set_verbosity_info() def UpperCAmelCase_ (__a : Optional[Any] , __a : str , __a : Optional[Any]=None ): """simple docstring""" assert torch_layer.weight.shape == weight.shape, f"""{torch_layer} layer.weight does not match""" _a : str = nn.Parameter(__a ) if bias is not None: assert torch_layer.bias.shape == bias.shape, f"""{torch_layer} layer.bias does not match""" _a : Any = nn.Parameter(__a ) def UpperCAmelCase_ (__a : int , __a : Optional[Any] , __a : int ): """simple docstring""" _a : Tuple = np.asarray(weights[0] ) _a : Union[str, Any] = np.asarray(weights[1] ) _a : Dict = np.asarray(weights[2] ) set_param( torch_layer.self_attention.query_key , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.value , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.output.dense , torch.tensor(__a ).view(-1 , __a ).contiguous().transpose(0 , 1 ) , ) def UpperCAmelCase_ (__a : Optional[Any] , __a : Optional[int] , __a : List[str] ): """simple docstring""" _a : Dict = np.asarray(weights[0] ) _a : Union[str, Any] = np.asarray(weights[1] ) _a : str = np.asarray(weights[2] ) _a : int = np.asarray(weights[3] ) set_param( torch_layer.self_attention.query , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.key , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.value , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.output.dense , torch.tensor(__a ).view(-1 , __a ).contiguous().transpose(0 , 1 ) , ) def UpperCAmelCase_ (__a : Any , __a : Any , __a : Optional[Any] ): """simple docstring""" _a : List[str] = weights[0][0][0] _a : List[Any] = np.asarray(layer_norm_a[0] ) _a : List[str] = np.asarray(layer_norm_a[1] ) set_param( torch_block.attention.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # lsh weights + output _a : List[str] = weights[0][1] if len(__a ) < 4: set_layer_weights_in_torch_lsh(__a , torch_block.attention , __a ) else: set_layer_weights_in_torch_local(__a , torch_block.attention , __a ) # intermediate weighs _a : Optional[Any] = weights[2][0][1][2] # Chunked Feed Forward if len(__a ) == 4: _a : Union[str, Any] = intermediate_weights[2] # layernorm 2 _a : Any = np.asarray(intermediate_weights[0][0] ) _a : List[Any] = np.asarray(intermediate_weights[0][1] ) set_param( torch_block.feed_forward.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # intermediate dense _a : Any = np.asarray(intermediate_weights[1][0] ) _a : Any = np.asarray(intermediate_weights[1][1] ) set_param( torch_block.feed_forward.dense.dense , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) # intermediate out _a : Optional[int] = np.asarray(intermediate_weights[4][0] ) _a : int = np.asarray(intermediate_weights[4][1] ) set_param( torch_block.feed_forward.output.dense , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) def UpperCAmelCase_ (__a : Dict , __a : Dict , __a : List[Any] ): """simple docstring""" _a : Optional[int] = torch_model.reformer # word embeds _a : Tuple = np.asarray(weights[1] ) set_param( torch_model_reformer.embeddings.word_embeddings , torch.tensor(__a ) , ) if isinstance(weights[3] , __a ): _a : Any = torch_model_reformer.embeddings.position_embeddings for emb_idx in range(len(position_embeddings.weights ) ): _a : List[Any] = np.asarray(weights[3][emb_idx][0] ) assert ( position_embeddings.weights[emb_idx].shape == emb_weights.shape ), f"""{position_embeddings[emb_idx]} emb does not match""" _a : Any = nn.Parameter(torch.tensor(__a ) ) _a : List[str] = weights[5] assert len(torch_model_reformer.encoder.layers ) * 4 == len( __a ), "HF and trax model do not have the same number of layers" for layer_idx, layer in enumerate(torch_model_reformer.encoder.layers ): _a : Tuple = trax_layer_weights[4 * layer_idx : 4 * (layer_idx + 1)] set_block_weights_in_torch(__a , __a , __a ) # output layer norm _a : Optional[Any] = np.asarray(weights[7][0] ) _a : int = np.asarray(weights[7][1] ) set_param( torch_model_reformer.encoder.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # output embeddings _a : List[str] = np.asarray(weights[9][0] ) _a : int = np.asarray(weights[9][1] ) set_param( torch_model.lm_head.decoder , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) def UpperCAmelCase_ (__a : Tuple , __a : Optional[Any] , __a : Dict ): """simple docstring""" _a : List[Any] = ReformerConfig.from_json_file(__a ) print(f"""Building PyTorch model from configuration: {config}""" ) _a : int = ReformerModelWithLMHead(__a ) with open(__a , 'rb' ) as f: _a : Optional[Any] = pickle.load(__a )['weights'] set_model_weights_in_torch(__a , __a , config.hidden_size ) # Save pytorch-model print(f"""Save PyTorch model to {pytorch_dump_path}""" ) torch.save(model.state_dict() , __a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser() # Required parameters parser.add_argument( """--trax_model_pkl_path""", default=None, type=str, required=True, help="""Path to the TensorFlow checkpoint path.""" ) parser.add_argument( """--config_file""", default=None, type=str, required=True, help=( """The config json file corresponding to the pre-trained Reformer model. \n""" """This specifies the model architecture.""" ), ) parser.add_argument( """--pytorch_dump_path""", default=None, type=str, required=True, help="""Path to the output PyTorch model.""" ) __lowerCAmelCase = parser.parse_args() convert_trax_checkpoint_to_pytorch(args.trax_model_pkl_path, args.config_file, args.pytorch_dump_path)
271
1
'''simple docstring''' import collections import json import math import os import re import time from fnmatch import fnmatch from typing import Dict import requests from slack_sdk import WebClient __lowerCAmelCase = WebClient(token=os.environ["""CI_SLACK_BOT_TOKEN"""]) def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" _a : List[str] = test_results.split(' ' ) _a : Any = 0 _a : Union[str, Any] = 0 # When the output is short enough, the output is surrounded by = signs: "== OUTPUT ==" # When it is too long, those signs are not present. _a : Any = expressions[-2] if '=' in expressions[-1] else expressions[-1] for i, expression in enumerate(__a ): if "failed" in expression: failed += int(expressions[i - 1] ) if "passed" in expression: success += int(expressions[i - 1] ) return failed, success, time_spent def UpperCAmelCase_ (__a : Any ): """simple docstring""" _a : Optional[Any] = {} _a : Tuple = None _a : str = False for line in failures_short_lines.split('\n' ): if re.search(R'_ \[doctest\]' , __a ): _a : Tuple = True _a : str = line.split(' ' )[2] elif in_error and not line.split(' ' )[0].isdigit(): _a : Dict = line _a : List[Any] = False return failures class UpperCAmelCase__ : """simple docstring""" def __init__( self : Tuple ,_a : str ,_a : Dict ): '''simple docstring''' _a : Tuple = title _a : Tuple = doc_test_results['time_spent'].split(',' )[0] _a : List[str] = doc_test_results['success'] _a : List[Any] = doc_test_results['failures'] _a : str = self.n_success + self.n_failures # Failures and success of the modeling tests _a : Union[str, Any] = doc_test_results @property def __lowercase ( self : str ): '''simple docstring''' _a : Any = [self._time_spent] _a : Union[str, Any] = 0 for time in time_spent: _a : Any = time.split(':' ) # Time can be formatted as xx:xx:xx, as .xx, or as x.xx if the time spent was less than a minute. if len(_a ) == 1: _a : Optional[int] = [0, 0, time_parts[0]] _a, _a, _a : Union[str, Any] = int(time_parts[0] ), int(time_parts[1] ), float(time_parts[2] ) total_secs += hours * 3600 + minutes * 60 + seconds _a, _a, _a : List[str] = total_secs // 3600, (total_secs % 3600) // 60, total_secs % 60 return F"""{int(_a )}h{int(_a )}m{int(_a )}s""" @property def __lowercase ( self : Any ): '''simple docstring''' return {"type": "header", "text": {"type": "plain_text", "text": self.title}} @property def __lowercase ( self : List[str] ): '''simple docstring''' return { "type": "section", "text": { "type": "plain_text", "text": F"""🌞 There were no failures: all {self.n_tests} tests passed. The suite ran in {self.time}.""", "emoji": True, }, "accessory": { "type": "button", "text": {"type": "plain_text", "text": "Check Action results", "emoji": True}, "url": F"""https://github.com/huggingface/transformers/actions/runs/{os.environ['GITHUB_RUN_ID']}""", }, } @property def __lowercase ( self : int ): '''simple docstring''' return { "type": "section", "text": { "type": "plain_text", "text": ( F"""There were {self.n_failures} failures, out of {self.n_tests} tests.\nThe suite ran in""" F""" {self.time}.""" ), "emoji": True, }, "accessory": { "type": "button", "text": {"type": "plain_text", "text": "Check Action results", "emoji": True}, "url": F"""https://github.com/huggingface/transformers/actions/runs/{os.environ['GITHUB_RUN_ID']}""", }, } @property def __lowercase ( self : Tuple ): '''simple docstring''' _a : Any = 40 _a : List[Any] = {k: v['failed'] for k, v in doc_test_results.items() if isinstance(_a ,_a )} _a : Any = '' for category, failures in category_failures.items(): if len(_a ) == 0: continue if report != "": report += "\n\n" report += F"""*{category} failures*:""".ljust(line_length // 2 ).rjust(line_length // 2 ) + "\n" report += "`" report += "`\n`".join(_a ) report += "`" return { "type": "section", "text": { "type": "mrkdwn", "text": F"""The following examples had failures:\n\n\n{report}\n""", }, } @property def __lowercase ( self : Any ): '''simple docstring''' _a : Tuple = [self.header] if self.n_failures > 0: blocks.append(self.failures ) if self.n_failures > 0: blocks.extend([self.category_failures] ) if self.n_failures == 0: blocks.append(self.no_failures ) return json.dumps(_a ) @staticmethod def __lowercase ( ): '''simple docstring''' _a : Optional[int] = [ { 'type': 'section', 'text': { 'type': 'plain_text', 'text': 'There was an issue running the tests.', }, 'accessory': { 'type': 'button', 'text': {'type': 'plain_text', 'text': 'Check Action results', 'emoji': True}, 'url': F"""https://github.com/huggingface/transformers/actions/runs/{os.environ['GITHUB_RUN_ID']}""", }, } ] print('Sending the following payload' ) print(json.dumps({'blocks': json.loads(_a )} ) ) client.chat_postMessage( channel=os.environ['CI_SLACK_CHANNEL_ID_DAILY'] ,text='There was an issue running the tests.' ,blocks=_a ,) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' print('Sending the following payload' ) print(json.dumps({'blocks': json.loads(self.payload )} ) ) _a : Optional[int] = F"""{self.n_failures} failures out of {self.n_tests} tests,""" if self.n_failures else 'All tests passed.' _a : Union[str, Any] = client.chat_postMessage( channel=os.environ['CI_SLACK_CHANNEL_ID_DAILY'] ,blocks=self.payload ,text=_a ,) def __lowercase ( self : Dict ,_a : List[Any] ,_a : Any ,_a : List[Any] ,_a : Tuple ): '''simple docstring''' _a : List[Any] = '' for key, value in failures.items(): _a : Dict = value[:200] + ' [Truncated]' if len(_a ) > 250 else value failures_text += F"""*{key}*\n_{value}_\n\n""" _a : Optional[int] = job_name _a : Optional[Any] = {'type': 'section', 'text': {'type': 'mrkdwn', 'text': text}} if job_link is not None: _a : str = { 'type': 'button', 'text': {'type': 'plain_text', 'text': 'GitHub Action job', 'emoji': True}, 'url': job_link, } return [ {"type": "header", "text": {"type": "plain_text", "text": title.upper(), "emoji": True}}, content, {"type": "section", "text": {"type": "mrkdwn", "text": failures_text}}, ] def __lowercase ( self : Any ): '''simple docstring''' if self.thread_ts is None: raise ValueError('Can only post reply if a post has been made.' ) _a : int = self.doc_test_results.pop('job_link' ) self.doc_test_results.pop('failures' ) self.doc_test_results.pop('success' ) self.doc_test_results.pop('time_spent' ) _a : Tuple = sorted(self.doc_test_results.items() ,key=lambda _a : t[0] ) for job, job_result in sorted_dict: if len(job_result['failures'] ): _a : Optional[Any] = F"""*Num failures* :{len(job_result['failed'] )} \n""" _a : str = job_result['failures'] _a : str = self.get_reply_blocks(_a ,_a ,_a ,text=_a ) print('Sending the following reply' ) print(json.dumps({'blocks': blocks} ) ) client.chat_postMessage( channel=os.environ['CI_SLACK_CHANNEL_ID_DAILY'] ,text=F"""Results for {job}""" ,blocks=_a ,thread_ts=self.thread_ts['ts'] ,) time.sleep(1 ) def UpperCAmelCase_ (): """simple docstring""" _a : Dict = os.environ['GITHUB_RUN_ID'] _a : List[Any] = f"""https://api.github.com/repos/huggingface/transformers/actions/runs/{run_id}/jobs?per_page=100""" _a : Union[str, Any] = requests.get(__a ).json() _a : Union[str, Any] = {} try: jobs.update({job['name']: job['html_url'] for job in result['jobs']} ) _a : Tuple = math.ceil((result['total_count'] - 1_0_0) / 1_0_0 ) for i in range(__a ): _a : Optional[Any] = requests.get(url + f"""&page={i + 2}""" ).json() jobs.update({job['name']: job['html_url'] for job in result['jobs']} ) return jobs except Exception as e: print('Unknown error, could not fetch links.' , __a ) return {} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Optional[int] = {} if os.path.exists(__a ): _a : List[Any] = os.listdir(__a ) for file in files: try: with open(os.path.join(__a , __a ) , encoding='utf-8' ) as f: _a : int = f.read() except UnicodeDecodeError as e: raise ValueError(f"""Could not open {os.path.join(__a , __a )}.""" ) from e return _artifact def UpperCAmelCase_ (): """simple docstring""" class UpperCAmelCase__ : """simple docstring""" def __init__( self : Optional[Any] ,_a : str ): '''simple docstring''' _a : Union[str, Any] = name _a : str = [] def __str__( self : Tuple ): '''simple docstring''' return self.name def __lowercase ( self : Dict ,_a : str ): '''simple docstring''' self.paths.append({'name': self.name, 'path': path} ) _a : Dict[str, Artifact] = {} _a : Any = filter(os.path.isdir , os.listdir() ) for directory in directories: _a : Dict = directory if artifact_name not in _available_artifacts: _a : int = Artifact(__a ) _available_artifacts[artifact_name].add_path(__a ) return _available_artifacts if __name__ == "__main__": __lowerCAmelCase = get_job_links() __lowerCAmelCase = retrieve_available_artifacts() __lowerCAmelCase = collections.OrderedDict( [ ("""*.py""", """API Examples"""), ("""*.md""", """MD Examples"""), ] ) # This dict will contain all the information relative to each doc test category: # - failed: list of failed tests # - failures: dict in the format 'test': 'error_message' __lowerCAmelCase = { v: { """failed""": [], """failures""": {}, } for v in docs.values() } # Link to the GitHub Action job __lowerCAmelCase = github_actions_job_links.get("""run_doctests""") __lowerCAmelCase = available_artifacts["""doc_tests_gpu_test_reports"""].paths[0] __lowerCAmelCase = retrieve_artifact(artifact_path["""name"""]) if "stats" in artifact: __lowerCAmelCase , __lowerCAmelCase , __lowerCAmelCase = handle_test_results(artifact["""stats"""]) __lowerCAmelCase = failed __lowerCAmelCase = success __lowerCAmelCase = time_spent[1:-1] + """, """ __lowerCAmelCase = extract_first_line_failure(artifact["""failures_short"""]) for line in artifact["summary_short"].split("""\n"""): if re.search("""FAILED""", line): __lowerCAmelCase = line.replace("""FAILED """, """""") __lowerCAmelCase = line.split()[0].replace("""\n""", """""") if "::" in line: __lowerCAmelCase , __lowerCAmelCase = line.split("""::""") else: __lowerCAmelCase , __lowerCAmelCase = line, line for file_regex in docs.keys(): if fnmatch(file_path, file_regex): __lowerCAmelCase = docs[file_regex] doc_test_results[category]["failed"].append(test) __lowerCAmelCase = all_failures[test] if test in all_failures else """N/A""" __lowerCAmelCase = failure break __lowerCAmelCase = Message("""🤗 Results of the doc tests.""", doc_test_results) message.post() message.post_reply()
271
'''simple docstring''' import unittest import numpy as np import torch from transformers import CLIPTextConfig, CLIPTextModel from diffusers import DDIMScheduler, LDMPipeline, UNetaDModel, VQModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' torch.manual_seed(0 ) _a : Any = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model @property def __lowercase ( self : Optional[int] ): '''simple docstring''' torch.manual_seed(0 ) _a : Union[str, Any] = VQModel( block_out_channels=[32, 64] ,in_channels=3 ,out_channels=3 ,down_block_types=['DownEncoderBlock2D', 'DownEncoderBlock2D'] ,up_block_types=['UpDecoderBlock2D', 'UpDecoderBlock2D'] ,latent_channels=3 ,) return model @property def __lowercase ( self : Optional[int] ): '''simple docstring''' torch.manual_seed(0 ) _a : Any = 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 ,) return CLIPTextModel(_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : Dict = self.dummy_uncond_unet _a : List[Any] = DDIMScheduler() _a : List[Any] = self.dummy_vq_model _a : str = LDMPipeline(unet=_a ,vqvae=_a ,scheduler=_a ) ldm.to(_a ) ldm.set_progress_bar_config(disable=_a ) _a : List[str] = torch.manual_seed(0 ) _a : List[str] = ldm(generator=_a ,num_inference_steps=2 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Union[str, Any] = ldm(generator=_a ,num_inference_steps=2 ,output_type='numpy' ,return_dict=_a )[0] _a : Tuple = image[0, -3:, -3:, -1] _a : Optional[Any] = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 64, 64, 3) _a : int = np.array([0.8512, 0.818, 0.6411, 0.6808, 0.4465, 0.5618, 0.46, 0.6231, 0.5172] ) _a : Any = 1E-2 if torch_device != 'mps' else 3E-2 assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < tolerance @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : List[str] = LDMPipeline.from_pretrained('CompVis/ldm-celebahq-256' ) ldm.to(_a ) ldm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Dict = ldm(generator=_a ,num_inference_steps=5 ,output_type='numpy' ).images _a : str = image[0, -3:, -3:, -1] assert image.shape == (1, 256, 256, 3) _a : Optional[int] = np.array([0.4399, 0.4_4975, 0.4_6825, 0.474, 0.4359, 0.4581, 0.4_5095, 0.4341, 0.4447] ) _a : int = 1E-2 if torch_device != 'mps' else 3E-2 assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
271
1
'''simple docstring''' import copy import os from typing import Union from ...configuration_utils import PretrainedConfig from ...models.auto.modeling_auto import MODEL_FOR_CAUSAL_LM_MAPPING_NAMES from ...utils import logging from ..auto import CONFIG_MAPPING __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """salesforce/blip2-opt-2.7b""": """https://huggingface.co/salesforce/blip2-opt-2.7b/resolve/main/config.json""", } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Dict = '''blip_2_vision_model''' def __init__( self : Tuple ,_a : int=1408 ,_a : str=6144 ,_a : Union[str, Any]=39 ,_a : List[Any]=16 ,_a : int=224 ,_a : str=14 ,_a : Optional[int]="gelu" ,_a : List[Any]=0.0_0001 ,_a : Any=0.0 ,_a : Any=1E-10 ,_a : int=True ,**_a : Union[str, Any] ,): '''simple docstring''' super().__init__(**_a ) _a : int = hidden_size _a : Dict = intermediate_size _a : str = num_hidden_layers _a : str = num_attention_heads _a : Dict = patch_size _a : str = image_size _a : List[str] = initializer_range _a : str = attention_dropout _a : List[Any] = layer_norm_eps _a : str = hidden_act _a : Any = qkv_bias @classmethod def __lowercase ( cls : List[str] ,_a : Union[str, os.PathLike] ,**_a : Optional[Any] ): '''simple docstring''' cls._set_token_in_kwargs(_a ) _a, _a : Optional[Any] = cls.get_config_dict(_a ,**_a ) # get the vision config dict if we are loading from Blip2Config if config_dict.get('model_type' ) == "blip-2": _a : List[Any] = 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(_a ,**_a ) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Any = '''blip_2_qformer''' def __init__( self : int ,_a : Any=3_0522 ,_a : Any=768 ,_a : Optional[Any]=12 ,_a : Union[str, Any]=12 ,_a : Union[str, Any]=3072 ,_a : Dict="gelu" ,_a : Optional[int]=0.1 ,_a : Dict=0.1 ,_a : Optional[Any]=512 ,_a : Union[str, Any]=0.02 ,_a : Optional[int]=1E-12 ,_a : Any=0 ,_a : Dict="absolute" ,_a : Tuple=2 ,_a : Tuple=1408 ,**_a : Dict ,): '''simple docstring''' super().__init__(pad_token_id=_a ,**_a ) _a : Optional[int] = vocab_size _a : Any = hidden_size _a : List[Any] = num_hidden_layers _a : Any = num_attention_heads _a : List[str] = hidden_act _a : Dict = intermediate_size _a : Dict = hidden_dropout_prob _a : Dict = attention_probs_dropout_prob _a : Dict = max_position_embeddings _a : str = initializer_range _a : int = layer_norm_eps _a : List[Any] = position_embedding_type _a : Union[str, Any] = cross_attention_frequency _a : str = encoder_hidden_size @classmethod def __lowercase ( cls : Tuple ,_a : Union[str, os.PathLike] ,**_a : Any ): '''simple docstring''' cls._set_token_in_kwargs(_a ) _a, _a : Any = cls.get_config_dict(_a ,**_a ) # get the qformer config dict if we are loading from Blip2Config if config_dict.get('model_type' ) == "blip-2": _a : Dict = config_dict['qformer_config'] if "model_type" in config_dict and hasattr(cls ,'model_type' ) and config_dict["model_type"] != cls.model_type: logger.warning( F"""You are using a model of type {config_dict['model_type']} to instantiate a model of type """ F"""{cls.model_type}. This is not supported for all configurations of models and can yield errors.""" ) return cls.from_dict(_a ,**_a ) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[Any] = '''blip-2''' __UpperCAmelCase : Dict = True def __init__( self : List[str] ,_a : str=None ,_a : Optional[Any]=None ,_a : Tuple=None ,_a : Optional[int]=32 ,**_a : Union[str, Any] ): '''simple docstring''' super().__init__(**_a ) if vision_config is None: _a : List[str] = {} logger.info('vision_config is None. initializing the Blip2VisionConfig with default values.' ) if qformer_config is None: _a : str = {} logger.info('qformer_config is None. Initializing the Blip2QFormerConfig with default values.' ) if text_config is None: _a : Union[str, Any] = {} logger.info('text_config is None. Initializing the text config with default values (`OPTConfig`).' ) _a : Optional[int] = BlipaVisionConfig(**_a ) _a : Optional[int] = BlipaQFormerConfig(**_a ) _a : Dict = text_config['model_type'] if 'model_type' in text_config else 'opt' _a : Optional[Any] = CONFIG_MAPPING[text_model_type](**_a ) _a : str = self.text_config.tie_word_embeddings _a : Any = self.text_config.is_encoder_decoder _a : Tuple = num_query_tokens _a : Dict = self.vision_config.hidden_size _a : Tuple = self.text_config.model_type in MODEL_FOR_CAUSAL_LM_MAPPING_NAMES _a : int = 1.0 _a : Dict = 0.02 @classmethod def __lowercase ( cls : int ,_a : BlipaVisionConfig ,_a : BlipaQFormerConfig ,_a : PretrainedConfig ,**_a : Tuple ,): '''simple docstring''' return cls( vision_config=vision_config.to_dict() ,qformer_config=qformer_config.to_dict() ,text_config=text_config.to_dict() ,**_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : List[Any] = copy.deepcopy(self.__dict__ ) _a : List[Any] = self.vision_config.to_dict() _a : Dict = self.qformer_config.to_dict() _a : Optional[int] = self.text_config.to_dict() _a : Union[str, Any] = self.__class__.model_type return output
271
'''simple docstring''' import warnings from ...utils import logging from .image_processing_beit import BeitImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : int ,*_a : Optional[int] ,**_a : str ): '''simple docstring''' warnings.warn( 'The class BeitFeatureExtractor is deprecated and will be removed in version 5 of Transformers. Please' ' use BeitImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
1
'''simple docstring''' import copy import tempfile import unittest from huggingface_hub import HfFolder, delete_repo from parameterized import parameterized from requests.exceptions import HTTPError from transformers import AutoConfig, GenerationConfig from transformers.testing_utils import TOKEN, USER, is_staging_test class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @parameterized.expand([(None,), ('foo.json',)] ) def __lowercase ( self : Optional[Any] ,_a : str ): '''simple docstring''' _a : List[str] = GenerationConfig( do_sample=_a ,temperature=0.7 ,length_penalty=1.0 ,bad_words_ids=[[1, 2, 3], [4, 5]] ,) with tempfile.TemporaryDirectory() as tmp_dir: config.save_pretrained(_a ,config_name=_a ) _a : Tuple = GenerationConfig.from_pretrained(_a ,config_name=_a ) # Checks parameters that were specified self.assertEqual(loaded_config.do_sample ,_a ) self.assertEqual(loaded_config.temperature ,0.7 ) self.assertEqual(loaded_config.length_penalty ,1.0 ) self.assertEqual(loaded_config.bad_words_ids ,[[1, 2, 3], [4, 5]] ) # Checks parameters that were not specified (defaults) self.assertEqual(loaded_config.top_k ,50 ) self.assertEqual(loaded_config.max_length ,20 ) self.assertEqual(loaded_config.max_time ,_a ) def __lowercase ( self : int ): '''simple docstring''' _a : Union[str, Any] = AutoConfig.from_pretrained('gpt2' ) _a : Tuple = GenerationConfig.from_model_config(_a ) _a : List[Any] = GenerationConfig() # The generation config has loaded a few non-default parameters from the model config self.assertNotEqual(_a ,_a ) # One of those parameters is eos_token_id -- check if it matches self.assertNotEqual(generation_config_from_model.eos_token_id ,default_generation_config.eos_token_id ) self.assertEqual(generation_config_from_model.eos_token_id ,model_config.eos_token_id ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Optional[Any] = GenerationConfig() _a : Dict = { 'max_new_tokens': 1024, 'foo': 'bar', } _a : Optional[int] = copy.deepcopy(_a ) _a : List[str] = generation_config.update(**_a ) # update_kwargs was not modified (no side effects) self.assertEqual(_a ,_a ) # update_kwargs was used to update the config on valid attributes self.assertEqual(generation_config.max_new_tokens ,1024 ) # `.update()` returns a dictionary of unused kwargs self.assertEqual(_a ,{'foo': 'bar'} ) def __lowercase ( self : Dict ): '''simple docstring''' _a : Optional[Any] = GenerationConfig() _a : Optional[Any] = 'bar' with tempfile.TemporaryDirectory('test-generation-config' ) as tmp_dir: generation_config.save_pretrained(_a ) _a : Optional[int] = GenerationConfig.from_pretrained(_a ) # update_kwargs was used to update the config on valid attributes self.assertEqual(new_config.foo ,'bar' ) _a : List[str] = GenerationConfig.from_model_config(_a ) assert not hasattr(_a ,'foo' ) # no new kwargs should be initialized if from config def __lowercase ( self : List[Any] ): '''simple docstring''' _a : List[str] = GenerationConfig() self.assertEqual(default_config.temperature ,1.0 ) self.assertEqual(default_config.do_sample ,_a ) self.assertEqual(default_config.num_beams ,1 ) _a : List[Any] = GenerationConfig( do_sample=_a ,temperature=0.7 ,length_penalty=1.0 ,bad_words_ids=[[1, 2, 3], [4, 5]] ,) self.assertEqual(config.temperature ,0.7 ) self.assertEqual(config.do_sample ,_a ) self.assertEqual(config.num_beams ,1 ) with tempfile.TemporaryDirectory() as tmp_dir: config.save_pretrained(_a ) _a : Optional[Any] = GenerationConfig.from_pretrained(_a ,temperature=1.0 ) self.assertEqual(loaded_config.temperature ,1.0 ) self.assertEqual(loaded_config.do_sample ,_a ) self.assertEqual(loaded_config.num_beams ,1 ) # default value @is_staging_test class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @classmethod def __lowercase ( cls : str ): '''simple docstring''' _a : List[str] = TOKEN HfFolder.save_token(_a ) @classmethod def __lowercase ( cls : Optional[int] ): '''simple docstring''' try: delete_repo(token=cls._token ,repo_id='test-generation-config' ) except HTTPError: pass try: delete_repo(token=cls._token ,repo_id='valid_org/test-generation-config-org' ) except HTTPError: pass def __lowercase ( self : str ): '''simple docstring''' _a : List[Any] = GenerationConfig( do_sample=_a ,temperature=0.7 ,length_penalty=1.0 ,) config.push_to_hub('test-generation-config' ,use_auth_token=self._token ) _a : Optional[Any] = GenerationConfig.from_pretrained(F"""{USER}/test-generation-config""" ) for k, v in config.to_dict().items(): if k != "transformers_version": self.assertEqual(_a ,getattr(_a ,_a ) ) # Reset repo delete_repo(token=self._token ,repo_id='test-generation-config' ) # Push to hub via save_pretrained with tempfile.TemporaryDirectory() as tmp_dir: config.save_pretrained( _a ,repo_id='test-generation-config' ,push_to_hub=_a ,use_auth_token=self._token ) _a : Optional[int] = GenerationConfig.from_pretrained(F"""{USER}/test-generation-config""" ) for k, v in config.to_dict().items(): if k != "transformers_version": self.assertEqual(_a ,getattr(_a ,_a ) ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : Tuple = GenerationConfig( do_sample=_a ,temperature=0.7 ,length_penalty=1.0 ,) config.push_to_hub('valid_org/test-generation-config-org' ,use_auth_token=self._token ) _a : Union[str, Any] = GenerationConfig.from_pretrained('valid_org/test-generation-config-org' ) for k, v in config.to_dict().items(): if k != "transformers_version": self.assertEqual(_a ,getattr(_a ,_a ) ) # Reset repo delete_repo(token=self._token ,repo_id='valid_org/test-generation-config-org' ) # Push to hub via save_pretrained with tempfile.TemporaryDirectory() as tmp_dir: config.save_pretrained( _a ,repo_id='valid_org/test-generation-config-org' ,push_to_hub=_a ,use_auth_token=self._token ) _a : Tuple = GenerationConfig.from_pretrained('valid_org/test-generation-config-org' ) for k, v in config.to_dict().items(): if k != "transformers_version": self.assertEqual(_a ,getattr(_a ,_a ) )
271
'''simple docstring''' from typing import Any, Dict, List, Optional, Tuple, Union import torch from torch import nn from torch.utils.data import DistributedSampler, RandomSampler from transformers import PreTrainedModel, Trainer, logging from transformers.integrations import is_fairscale_available from transformers.models.fsmt.configuration_fsmt import FSMTConfig from transformers.optimization import ( Adafactor, AdamW, get_constant_schedule, get_constant_schedule_with_warmup, get_cosine_schedule_with_warmup, get_cosine_with_hard_restarts_schedule_with_warmup, get_linear_schedule_with_warmup, get_polynomial_decay_schedule_with_warmup, ) from transformers.trainer_pt_utils import get_tpu_sampler from transformers.training_args import ParallelMode from transformers.utils import is_torch_tpu_available if is_fairscale_available(): from fairscale.optim import OSS __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """linear""": get_linear_schedule_with_warmup, """cosine""": get_cosine_schedule_with_warmup, """cosine_w_restarts""": get_cosine_with_hard_restarts_schedule_with_warmup, """polynomial""": get_polynomial_decay_schedule_with_warmup, """constant""": get_constant_schedule, """constant_w_warmup""": get_constant_schedule_with_warmup, } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Optional[int] ,_a : Optional[Any]=None ,_a : Dict=None ,*_a : int ,**_a : str ): '''simple docstring''' super().__init__(*_a ,**_a ) if config is None: assert isinstance(self.model ,_a ), ( "If no `config` is passed the model to be trained has to be of type `PreTrainedModel`, but is" F""" {self.model.__class__}""" ) _a : List[Any] = self.model.config else: _a : Optional[int] = config _a : List[str] = data_args _a : List[Any] = self.config.tgt_vocab_size if isinstance(self.config ,_a ) else self.config.vocab_size if self.args.label_smoothing != 0 or (self.data_args is not None and self.data_args.ignore_pad_token_for_loss): assert self.config.pad_token_id is not None, ( "Make sure that `config.pad_token_id` is correcly defined when ignoring `pad_token` for loss" " calculation or doing label smoothing." ) if self.config.pad_token_id is None and self.config.eos_token_id is not None: logger.warning( F"""The `config.pad_token_id` is `None`. Using `config.eos_token_id` = {self.config.eos_token_id} for""" ' padding..' ) if self.args.label_smoothing == 0: _a : List[str] = torch.nn.CrossEntropyLoss(ignore_index=self.config.pad_token_id ) else: # dynamically import label_smoothed_nll_loss from utils import label_smoothed_nll_loss _a : Tuple = label_smoothed_nll_loss def __lowercase ( self : List[str] ,_a : int ): '''simple docstring''' if self.optimizer is None: _a : Union[str, Any] = ['bias', 'LayerNorm.weight'] _a : Tuple = [ { 'params': [p for n, p in self.model.named_parameters() if not any(nd in n for nd in no_decay )], 'weight_decay': self.args.weight_decay, }, { 'params': [p for n, p in self.model.named_parameters() if any(nd in n for nd in no_decay )], 'weight_decay': 0.0, }, ] _a : Optional[int] = Adafactor if self.args.adafactor else AdamW if self.args.adafactor: _a : Any = Adafactor _a : Dict = {'scale_parameter': False, 'relative_step': False} else: _a : Union[str, Any] = AdamW _a : str = { 'betas': (self.args.adam_betaa, self.args.adam_betaa), 'eps': self.args.adam_epsilon, } _a : Union[str, Any] = self.args.learning_rate if self.sharded_ddp: _a : str = OSS( params=_a ,optim=_a ,**_a ,) else: _a : Tuple = optimizer_cls(_a ,**_a ) if self.lr_scheduler is None: _a : List[Any] = self._get_lr_scheduler(_a ) else: # ignoring --lr_scheduler logger.warning('scheduler is passed to `Seq2SeqTrainer`, `--lr_scheduler` arg is ignored.' ) def __lowercase ( self : List[Any] ,_a : List[Any] ): '''simple docstring''' _a : str = arg_to_scheduler[self.args.lr_scheduler] if self.args.lr_scheduler == "constant": _a : int = schedule_func(self.optimizer ) elif self.args.lr_scheduler == "constant_w_warmup": _a : List[str] = schedule_func(self.optimizer ,num_warmup_steps=self.args.warmup_steps ) else: _a : Optional[int] = schedule_func( self.optimizer ,num_warmup_steps=self.args.warmup_steps ,num_training_steps=_a ) return scheduler def __lowercase ( self : Tuple ): '''simple docstring''' if isinstance(self.train_dataset ,torch.utils.data.IterableDataset ): return None elif is_torch_tpu_available(): return get_tpu_sampler(self.train_dataset ) else: if self.args.sortish_sampler: self.train_dataset.make_sortish_sampler( self.args.per_device_train_batch_size ,distributed=(self.args.parallel_mode == ParallelMode.DISTRIBUTED) ,) return ( RandomSampler(self.train_dataset ) if self.args.local_rank == -1 else DistributedSampler(self.train_dataset ) ) def __lowercase ( self : Dict ,_a : Dict ,_a : Any ,_a : Dict ): '''simple docstring''' if self.args.label_smoothing == 0: if self.data_args is not None and self.data_args.ignore_pad_token_for_loss: # force training to ignore pad token _a : List[Any] = model(**_a ,use_cache=_a )[0] _a : Union[str, Any] = self.loss_fn(logits.view(-1 ,logits.shape[-1] ) ,labels.view(-1 ) ) else: # compute usual loss via models _a, _a : Union[str, Any] = model(**_a ,labels=_a ,use_cache=_a )[:2] else: # compute label smoothed loss _a : List[Any] = model(**_a ,use_cache=_a )[0] _a : Any = torch.nn.functional.log_softmax(_a ,dim=-1 ) _a, _a : List[str] = self.loss_fn(_a ,_a ,self.args.label_smoothing ,ignore_index=self.config.pad_token_id ) return loss, logits def __lowercase ( self : Optional[int] ,_a : Union[str, Any] ,_a : List[Any] ): '''simple docstring''' _a : Optional[int] = inputs.pop('labels' ) _a, _a : int = self._compute_loss(_a ,_a ,_a ) return loss def __lowercase ( self : Optional[Any] ,_a : nn.Module ,_a : Dict[str, Union[torch.Tensor, Any]] ,_a : bool ,_a : Optional[List[str]] = None ,): '''simple docstring''' _a : int = self._prepare_inputs(_a ) _a : Any = { 'max_length': self.data_args.val_max_target_length if self.data_args is not None else self.config.max_length, 'num_beams': self.data_args.eval_beams if self.data_args is not None else self.config.num_beams, } if self.args.predict_with_generate and not self.args.prediction_loss_only: _a : int = self.model.generate( inputs['input_ids'] ,attention_mask=inputs['attention_mask'] ,**_a ,) # in case the batch is shorter than max length, the output should be padded if generated_tokens.shape[-1] < gen_kwargs["max_length"]: _a : int = self._pad_tensors_to_max_len(_a ,gen_kwargs['max_length'] ) _a : Union[str, Any] = inputs.pop('labels' ) with torch.no_grad(): # compute loss on predict data _a, _a : Optional[int] = self._compute_loss(_a ,_a ,_a ) _a : Optional[Any] = loss.mean().detach() if self.args.prediction_loss_only: return (loss, None, None) _a : Optional[Any] = generated_tokens if self.args.predict_with_generate else logits if labels.shape[-1] < gen_kwargs["max_length"]: _a : Dict = self._pad_tensors_to_max_len(_a ,gen_kwargs['max_length'] ) return (loss, logits, labels) def __lowercase ( self : str ,_a : Tuple ,_a : Tuple ): '''simple docstring''' _a : List[Any] = self.config.pad_token_id if self.config.pad_token_id is not None else self.config.eos_token_id if pad_token_id is None: raise ValueError( 'Make sure that either `config.pad_token_id` or `config.eos_token_id` is defined if tensor has to be' F""" padded to `max_length`={max_length}""" ) _a : int = pad_token_id * torch.ones( (tensor.shape[0], max_length) ,dtype=tensor.dtype ,device=tensor.device ) _a : Union[str, Any] = tensor return padded_tensor
271
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, BatchEncoding, PreTrainedTokenizer from ...utils import logging __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = """▁""" __lowerCAmelCase = {"""vocab_file""": """sentencepiece.bpe.model"""} __lowerCAmelCase = { """vocab_file""": { """facebook/mbart-large-50-one-to-many-mmt""": ( """https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model""" ), } } __lowerCAmelCase = { """facebook/mbart-large-50-one-to-many-mmt""": 1_0_2_4, } # fmt: off __lowerCAmelCase = ["""ar_AR""", """cs_CZ""", """de_DE""", """en_XX""", """es_XX""", """et_EE""", """fi_FI""", """fr_XX""", """gu_IN""", """hi_IN""", """it_IT""", """ja_XX""", """kk_KZ""", """ko_KR""", """lt_LT""", """lv_LV""", """my_MM""", """ne_NP""", """nl_XX""", """ro_RO""", """ru_RU""", """si_LK""", """tr_TR""", """vi_VN""", """zh_CN""", """af_ZA""", """az_AZ""", """bn_IN""", """fa_IR""", """he_IL""", """hr_HR""", """id_ID""", """ka_GE""", """km_KH""", """mk_MK""", """ml_IN""", """mn_MN""", """mr_IN""", """pl_PL""", """ps_AF""", """pt_XX""", """sv_SE""", """sw_KE""", """ta_IN""", """te_IN""", """th_TH""", """tl_XX""", """uk_UA""", """ur_PK""", """xh_ZA""", """gl_ES""", """sl_SI"""] class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Tuple = VOCAB_FILES_NAMES __UpperCAmelCase : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES __UpperCAmelCase : str = PRETRAINED_VOCAB_FILES_MAP __UpperCAmelCase : int = ['''input_ids''', '''attention_mask'''] __UpperCAmelCase : List[int] = [] __UpperCAmelCase : List[int] = [] def __init__( self : Any ,_a : int ,_a : Union[str, Any]=None ,_a : Optional[int]=None ,_a : int="</s>" ,_a : Optional[Any]="</s>" ,_a : Dict="<s>" ,_a : List[Any]="<unk>" ,_a : Tuple="<pad>" ,_a : List[Any]="<mask>" ,_a : Optional[Dict[str, Any]] = None ,**_a : List[str] ,): '''simple docstring''' _a : Optional[Any] = AddedToken(_a ,lstrip=_a ,rstrip=_a ) if isinstance(_a ,_a ) else mask_token _a : List[str] = {} if sp_model_kwargs is None else sp_model_kwargs _a : Tuple = kwargs.get('additional_special_tokens' ,[] ) kwargs["additional_special_tokens"] += [ code for code in FAIRSEQ_LANGUAGE_CODES if code not in kwargs["additional_special_tokens"] ] super().__init__( src_lang=_a ,tgt_lang=_a ,eos_token=_a ,unk_token=_a ,sep_token=_a ,cls_token=_a ,pad_token=_a ,mask_token=_a ,sp_model_kwargs=self.sp_model_kwargs ,**_a ,) _a : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.Load(str(_a ) ) _a : List[Any] = vocab_file # Original fairseq vocab and spm vocab must be "aligned": # Vocab | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 # -------- | ------- | ------- | ------ | ------- | --- | --- | --- | ----- | ----- | ---- # fairseq | '<s>' | '<pad>' | '</s>' | '<unk>' | ',' | '.' | '▁' | 's' | '▁de' | '-' # spm | '<unk>' | '<s>' | '</s>' | ',' | '.' | '▁' | 's' | '▁de' | '-' | '▁a' # Mimic fairseq token-to-id alignment for the first 4 token _a : Optional[int] = {'<s>': 0, '<pad>': 1, '</s>': 2, '<unk>': 3} # The first "real" token "," has position 4 in the original fairseq vocab and position 3 in the spm vocab _a : Union[str, Any] = 1 _a : Optional[Any] = len(self.sp_model ) _a : Tuple = { code: self.sp_model_size + i + self.fairseq_offset for i, code in enumerate(_a ) } _a : Any = {v: k for k, v in self.lang_code_to_id.items()} _a : Dict = len(self.sp_model ) + len(self.lang_code_to_id ) + self.fairseq_offset self.fairseq_tokens_to_ids.update(self.lang_code_to_id ) _a : int = {v: k for k, v in self.fairseq_tokens_to_ids.items()} _a : Any = src_lang if src_lang is not None else 'en_XX' _a : List[Any] = self.lang_code_to_id[self._src_lang] _a : Union[str, Any] = tgt_lang self.set_src_lang_special_tokens(self._src_lang ) @property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' return len(self.sp_model ) + len(self.lang_code_to_id ) + self.fairseq_offset + 1 # Plus 1 for the mask token @property def __lowercase ( self : int ): '''simple docstring''' return self._src_lang @src_lang.setter def __lowercase ( self : Any ,_a : str ): '''simple docstring''' _a : Optional[Any] = new_src_lang self.set_src_lang_special_tokens(self._src_lang ) def __getstate__( self : Dict ): '''simple docstring''' _a : Union[str, Any] = self.__dict__.copy() _a : int = None return state def __setstate__( self : int ,_a : Dict ): '''simple docstring''' _a : int = d # for backward compatibility if not hasattr(self ,'sp_model_kwargs' ): _a : Dict = {} _a : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.Load(self.vocab_file ) def __lowercase ( self : Any ): '''simple docstring''' _a : Dict = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )} vocab.update(self.added_tokens_encoder ) return vocab def __lowercase ( self : str ,_a : str ): '''simple docstring''' return self.sp_model.encode(_a ,out_type=_a ) def __lowercase ( self : str ,_a : str ): '''simple docstring''' if token in self.fairseq_tokens_to_ids: return self.fairseq_tokens_to_ids[token] _a : Tuple = self.sp_model.PieceToId(_a ) # Need to return unknown token if the SP model returned 0 return spm_id + self.fairseq_offset if spm_id else self.unk_token_id def __lowercase ( self : Optional[int] ,_a : int ): '''simple docstring''' if index in self.fairseq_ids_to_tokens: return self.fairseq_ids_to_tokens[index] return self.sp_model.IdToPiece(index - self.fairseq_offset ) def __lowercase ( self : Dict ,_a : str ): '''simple docstring''' _a : Union[str, Any] = [] _a : Dict = '' _a : Any = False for token in tokens: # make sure that special tokens are not decoded using sentencepiece model if token in self.all_special_tokens: if not prev_is_special: out_string += " " out_string += self.sp_model.decode(_a ) + token _a : Any = True _a : int = [] else: current_sub_tokens.append(_a ) _a : int = False out_string += self.sp_model.decode(_a ) return out_string.strip() def __lowercase ( self : Optional[int] ,_a : str ,_a : Optional[str] = None ): '''simple docstring''' if not os.path.isdir(_a ): logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" ) return _a : List[Any] = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ): copyfile(self.vocab_file ,_a ) elif not os.path.isfile(self.vocab_file ): with open(_a ,'wb' ) as fi: _a : Optional[int] = self.sp_model.serialized_model_proto() fi.write(_a ) return (out_vocab_file,) def __lowercase ( self : List[str] ,_a : List[int] ,_a : Optional[List[int]] = None ,_a : bool = False ): '''simple docstring''' if already_has_special_tokens: return super().get_special_tokens_mask( token_ids_a=_a ,token_ids_a=_a ,already_has_special_tokens=_a ) _a : Optional[Any] = [1] * len(self.prefix_tokens ) _a : List[Any] = [1] * len(self.suffix_tokens ) if token_ids_a is None: return prefix_ones + ([0] * len(_a )) + suffix_ones return prefix_ones + ([0] * len(_a )) + ([0] * len(_a )) + suffix_ones def __lowercase ( self : Dict ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' if token_ids_a is None: return self.prefix_tokens + token_ids_a + self.suffix_tokens # We don't expect to process pairs, but leave the pair logic for API consistency return self.prefix_tokens + token_ids_a + token_ids_a + self.suffix_tokens def __lowercase ( self : Union[str, Any] ,_a : Optional[Any] ,_a : str ,_a : Optional[str] ,_a : Optional[str] ,**_a : str ): '''simple docstring''' if src_lang is None or tgt_lang is None: raise ValueError('Translation requires a `src_lang` and a `tgt_lang` for this model' ) _a : str = src_lang _a : Union[str, Any] = self(_a ,add_special_tokens=_a ,return_tensors=_a ,**_a ) _a : Any = self.convert_tokens_to_ids(_a ) _a : int = tgt_lang_id return inputs def __lowercase ( self : Optional[Any] ,_a : List[str] ,_a : str = "en_XX" ,_a : Optional[List[str]] = None ,_a : str = "ro_RO" ,**_a : str ,): '''simple docstring''' _a : Dict = src_lang _a : List[Any] = tgt_lang return super().prepare_seqaseq_batch(_a ,_a ,**_a ) def __lowercase ( self : Tuple ): '''simple docstring''' return self.set_src_lang_special_tokens(self.src_lang ) def __lowercase ( self : Dict ): '''simple docstring''' return self.set_tgt_lang_special_tokens(self.tgt_lang ) def __lowercase ( self : Optional[int] ,_a : str ): '''simple docstring''' _a : str = self.lang_code_to_id[src_lang] _a : List[Any] = [self.cur_lang_code_id] _a : Optional[int] = [self.eos_token_id] def __lowercase ( self : Optional[int] ,_a : str ): '''simple docstring''' _a : int = self.lang_code_to_id[tgt_lang] _a : Optional[Any] = [self.cur_lang_code_id] _a : str = [self.eos_token_id]
271
'''simple docstring''' import gzip import hashlib import json import multiprocessing import os import re import shutil import time from pathlib import Path import numpy as np from arguments import PreprocessingArguments from datasets import load_dataset from minhash_deduplication import deduplicate_dataset from transformers import AutoTokenizer, HfArgumentParser __lowerCAmelCase = re.compile(r"""\s+""") def UpperCAmelCase_ (__a : Any ): """simple docstring""" return {"hash": hashlib.mda(re.sub(__a , '' , example['content'] ).encode('utf-8' ) ).hexdigest()} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : List[str] = [len(__a ) for line in example['content'].splitlines()] return {"line_mean": np.mean(__a ), "line_max": max(__a )} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Union[str, Any] = np.mean([c.isalnum() for c in example['content']] ) return {"alpha_frac": alpha_frac} def UpperCAmelCase_ (__a : Optional[int] , __a : Any ): """simple docstring""" if example["hash"] in uniques: uniques.remove(example['hash'] ) return True else: return False def UpperCAmelCase_ (__a : int , __a : Union[str, Any]=5 ): """simple docstring""" _a : Optional[int] = ['auto-generated', 'autogenerated', 'automatically generated'] _a : List[str] = example['content'].splitlines() for _, line in zip(range(__a ) , __a ): for keyword in keywords: if keyword in line.lower(): return {"autogenerated": True} else: return {"autogenerated": False} def UpperCAmelCase_ (__a : List[str] , __a : Dict=5 , __a : Tuple=0.05 ): """simple docstring""" _a : Optional[int] = ['unit tests', 'test file', 'configuration file'] _a : int = example['content'].splitlines() _a : int = 0 _a : Dict = 0 # first test for _, line in zip(range(__a ) , __a ): for keyword in keywords: if keyword in line.lower(): return {"config_or_test": True} # second test _a : int = example['content'].count('\n' ) _a : int = int(coeff * nlines ) for line in lines: count_config += line.lower().count('config' ) count_test += line.lower().count('test' ) if count_config > threshold or count_test > threshold: return {"config_or_test": True} return {"config_or_test": False} def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" _a : List[str] = ['def ', 'class ', 'for ', 'while '] _a : str = example['content'].splitlines() for line in lines: for keyword in keywords: if keyword in line.lower(): return {"has_no_keywords": False} return {"has_no_keywords": True} def UpperCAmelCase_ (__a : int , __a : Any=4 ): """simple docstring""" _a : List[str] = example['content'].splitlines() _a : Dict = 0 for line in lines: counter += line.lower().count('=' ) if counter > minimum: return {"has_few_assignments": False} return {"has_few_assignments": True} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Optional[Any] = tokenizer(example['content'] , truncation=__a )['input_ids'] _a : Optional[int] = len(example['content'] ) / len(__a ) return {"ratio": ratio} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Dict = {} results.update(get_hash(__a ) ) results.update(line_stats(__a ) ) results.update(alpha_stats(__a ) ) results.update(char_token_ratio(__a ) ) results.update(is_autogenerated(__a ) ) results.update(is_config_or_test(__a ) ) results.update(has_no_keywords(__a ) ) results.update(has_few_assignments(__a ) ) return results def UpperCAmelCase_ (__a : Any , __a : Any , __a : str ): """simple docstring""" if not check_uniques(__a , __a ): return False elif example["autogenerated"]: return False elif example["line_max"] > args.line_max: return False elif example["line_mean"] > args.line_mean: return False elif example["alpha_frac"] < args.alpha_frac: return False elif example["ratio"] < args.min_token_ratio: return False elif example["config_or_test"] and np.random.rand() <= args.filter_proba: return False elif example["has_no_keywords"] and np.random.rand() <= args.filter_proba: return False elif example["has_few_assignments"]: return False else: return True def UpperCAmelCase_ (__a : Union[str, Any] ): """simple docstring""" with open(__a , 'rb' ) as f_in: with gzip.open(str(__a ) + '.gz' , 'wb' , compresslevel=6 ) as f_out: shutil.copyfileobj(__a , __a ) os.unlink(__a ) # Settings __lowerCAmelCase = HfArgumentParser(PreprocessingArguments) __lowerCAmelCase = parser.parse_args() if args.num_workers is None: __lowerCAmelCase = multiprocessing.cpu_count() __lowerCAmelCase = AutoTokenizer.from_pretrained(args.tokenizer_dir) # Load dataset __lowerCAmelCase = time.time() __lowerCAmelCase = load_dataset(args.dataset_name, split="""train""") print(f'''Time to load dataset: {time.time()-t_start:.2f}''') # Run preprocessing __lowerCAmelCase = time.time() __lowerCAmelCase = ds.map(preprocess, num_proc=args.num_workers) print(f'''Time to preprocess dataset: {time.time()-t_start:.2f}''') # Deduplicate hashes __lowerCAmelCase = set(ds.unique("""hash""")) __lowerCAmelCase = len(uniques) / len(ds) print(f'''Fraction of duplicates: {1-frac:.2%}''') # Deduplicate data and apply heuristics __lowerCAmelCase = time.time() __lowerCAmelCase = ds.filter(filter, fn_kwargs={"""uniques""": uniques, """args""": args}) print(f'''Time to filter dataset: {time.time()-t_start:.2f}''') print(f'''Size of filtered dataset: {len(ds_filter)}''') # Deduplicate with minhash and jaccard similarity if args.near_deduplication: __lowerCAmelCase = time.time() __lowerCAmelCase , __lowerCAmelCase = deduplicate_dataset(ds_filter, args.jaccard_threshold) print(f'''Time to deduplicate dataset: {time.time()-t_start:.2f}''') print(f'''Size of deduplicate dataset: {len(ds_filter)}''') # Save data in batches of samples_per_file __lowerCAmelCase = Path(args.output_dir) output_dir.mkdir(exist_ok=True) # save duplicate_clusters in the output_dir as artifacts # not sure it is the right place the save it if args.near_deduplication: with open(output_dir / """duplicate_clusters.json""", """w""") as f: json.dump(duplicate_clusters, f) __lowerCAmelCase = output_dir / """data""" data_dir.mkdir(exist_ok=True) __lowerCAmelCase = time.time() for file_number, index in enumerate(range(0, len(ds_filter), args.samples_per_file)): __lowerCAmelCase = str(data_dir / f'''file-{file_number+1:012}.json''') __lowerCAmelCase = min(len(ds_filter), index + args.samples_per_file) ds_filter.select(list(range(index, end_index))).to_json(file_path) compress_file(file_path) print(f'''Time to save dataset: {time.time()-t_start:.2f}''')
271
1
'''simple docstring''' from unittest import TestCase from datasets import Dataset from minhash_deduplication import deduplicate_dataset, make_duplicate_clusters def UpperCAmelCase_ (): """simple docstring""" _a : Optional[int] = { 'repo_name': ['test_repo1', 'test_repo2', 'test_repo3'], 'path': ['test_1.py', 'test_2.py', 'unit_test.py'], 'content': ['a ' * 2_0, 'a ' * 3_0, 'b ' * 7], } _a : Tuple = Dataset.from_dict(__a ) return dataset class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __lowercase ( self : int ): '''simple docstring''' _a : List[str] = get_dataset() _a : str = make_duplicate_clusters(_a ,0.85 ) self.assertEqual(len(duplicate_clusters[0] ) ,2 ) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Union[str, Any] = get_dataset() _a, _a : Union[str, Any] = deduplicate_dataset(_a ) self.assertEqual(len(_a ) ,2 ) print(_a ) self.assertEqual(duplicate_clusters[0][0]['copies'] ,2 ) self.assertEqual(duplicate_clusters[0][0]['is_extreme'] ,_a )
271
'''simple docstring''' import argparse from typing import List import evaluate import numpy as np import torch from datasets import DatasetDict, load_dataset # New Code # # We'll be using StratifiedKFold for this example from sklearn.model_selection import StratifiedKFold from torch.optim import AdamW from torch.utils.data import DataLoader from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed from accelerate import Accelerator, DistributedType ######################################################################## # This is a fully working simple example to use Accelerate, # specifically showcasing how to perform Cross Validation, # and builds off the `nlp_example.py` script. # # This example trains a Bert base model on GLUE MRPC # in any of the following settings (with the same script): # - single CPU or single GPU # - multi GPUS (using PyTorch distributed mode) # - (multi) TPUs # - fp16 (mixed-precision) or fp32 (normal precision) # # To help focus on the differences in the code, building `DataLoaders` # was refactored into its own function. # New additions from the base script can be found quickly by # looking for the # New Code # tags # # To run it in each of these various modes, follow the instructions # in the readme for examples: # https://github.com/huggingface/accelerate/tree/main/examples # ######################################################################## __lowerCAmelCase = 1_6 __lowerCAmelCase = 3_2 def UpperCAmelCase_ (__a : Accelerator , __a : DatasetDict , __a : List[int] , __a : List[int] , __a : int = 1_6 ): """simple docstring""" _a : Union[str, Any] = AutoTokenizer.from_pretrained('bert-base-cased' ) _a : str = DatasetDict( { 'train': dataset['train'].select(__a ), 'validation': dataset['train'].select(__a ), 'test': dataset['validation'], } ) def tokenize_function(__a : List[Any] ): # max_length=None => use the model max length (it's actually the default) _a : Optional[int] = tokenizer(examples['sentence1'] , examples['sentence2'] , truncation=__a , max_length=__a ) return outputs # Apply the method we just defined to all the examples in all the splits of the dataset # starting with the main process first: with accelerator.main_process_first(): _a : List[str] = datasets.map( __a , batched=__a , remove_columns=['idx', 'sentence1', 'sentence2'] , ) # We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the # transformers library _a : List[Any] = tokenized_datasets.rename_column('label' , 'labels' ) def collate_fn(__a : int ): # On TPU it's best to pad everything to the same length or training will be very slow. _a : Dict = 1_2_8 if accelerator.distributed_type == DistributedType.TPU else None # When using mixed precision we want round multiples of 8/16 if accelerator.mixed_precision == "fp8": _a : Tuple = 1_6 elif accelerator.mixed_precision != "no": _a : List[Any] = 8 else: _a : List[Any] = None return tokenizer.pad( __a , padding='longest' , max_length=__a , pad_to_multiple_of=__a , return_tensors='pt' , ) # Instantiate dataloaders. _a : Any = DataLoader( tokenized_datasets['train'] , shuffle=__a , collate_fn=__a , batch_size=__a ) _a : Optional[int] = DataLoader( tokenized_datasets['validation'] , shuffle=__a , collate_fn=__a , batch_size=__a ) _a : Optional[Any] = DataLoader( tokenized_datasets['test'] , shuffle=__a , collate_fn=__a , batch_size=__a ) return train_dataloader, eval_dataloader, test_dataloader def UpperCAmelCase_ (__a : Any , __a : Union[str, Any] ): """simple docstring""" _a : Dict = [] # Download the dataset _a : Tuple = load_dataset('glue' , 'mrpc' ) # Create our splits _a : Union[str, Any] = StratifiedKFold(n_splits=int(args.num_folds ) ) # Initialize accelerator _a : Any = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision ) # Sample hyper-parameters for learning rate, batch size, seed and a few other HPs _a : Optional[Any] = config['lr'] _a : Optional[int] = int(config['num_epochs'] ) _a : Dict = int(config['seed'] ) _a : Dict = int(config['batch_size'] ) _a : Optional[int] = evaluate.load('glue' , 'mrpc' ) # If the batch size is too big we use gradient accumulation _a : List[Any] = 1 if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU: _a : Any = batch_size // MAX_GPU_BATCH_SIZE _a : List[str] = MAX_GPU_BATCH_SIZE set_seed(__a ) # New Code # # Create our folds: _a : int = kfold.split(np.zeros(datasets['train'].num_rows ) , datasets['train']['label'] ) _a : Any = [] # Iterate over them for i, (train_idxs, valid_idxs) in enumerate(__a ): _a, _a, _a : Optional[Any] = get_fold_dataloaders( __a , __a , __a , __a , ) # Instantiate the model (we build the model here so that the seed also control new weights initialization) _a : Dict = AutoModelForSequenceClassification.from_pretrained('bert-base-cased' , return_dict=__a ) # We could avoid this line since the accelerator is set with `device_placement=True` (default value). # Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer # creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that). _a : List[Any] = model.to(accelerator.device ) # Instantiate optimizer _a : List[str] = AdamW(params=model.parameters() , lr=__a ) # Instantiate scheduler _a : List[Any] = get_linear_schedule_with_warmup( optimizer=__a , num_warmup_steps=1_0_0 , num_training_steps=(len(__a ) * num_epochs) // gradient_accumulation_steps , ) # Prepare everything # There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the # prepare method. _a, _a, _a, _a, _a : Union[str, Any] = accelerator.prepare( __a , __a , __a , __a , __a ) # Now we train the model for epoch in range(__a ): model.train() for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) _a : Dict = model(**__a ) _a : int = outputs.loss _a : Any = loss / gradient_accumulation_steps accelerator.backward(__a ) if step % gradient_accumulation_steps == 0: optimizer.step() lr_scheduler.step() optimizer.zero_grad() model.eval() for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): _a : Union[str, Any] = model(**__a ) _a : Tuple = outputs.logits.argmax(dim=-1 ) _a, _a : Any = accelerator.gather_for_metrics((predictions, batch['labels']) ) metric.add_batch( predictions=__a , references=__a , ) _a : List[Any] = metric.compute() # Use accelerator.print to print only on the main process. accelerator.print(f"""epoch {epoch}:""" , __a ) # New Code # # We also run predictions on the test set at the very end _a : Any = [] for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): _a : Tuple = model(**__a ) _a : Dict = outputs.logits _a, _a : Optional[int] = accelerator.gather_for_metrics((predictions, batch['labels']) ) fold_predictions.append(predictions.cpu() ) if i == 0: # We need all of the test predictions test_references.append(references.cpu() ) # Use accelerator.print to print only on the main process. test_predictions.append(torch.cat(__a , dim=0 ) ) # We now need to release all our memory and get rid of the current model, optimizer, etc accelerator.free_memory() # New Code # # Finally we check the accuracy of our folded results: _a : Dict = torch.cat(__a , dim=0 ) _a : Any = torch.stack(__a , dim=0 ).sum(dim=0 ).div(int(args.num_folds ) ).argmax(dim=-1 ) _a : str = metric.compute(predictions=__a , references=__a ) accelerator.print('Average test metrics from all folds:' , __a ) def UpperCAmelCase_ (): """simple docstring""" _a : Any = argparse.ArgumentParser(description='Simple example of training script.' ) parser.add_argument( '--mixed_precision' , type=__a , default=__a , choices=['no', 'fp16', 'bf16', 'fp8'] , help='Whether to use mixed precision. Choose' 'between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.' 'and an Nvidia Ampere GPU.' , ) parser.add_argument('--cpu' , action='store_true' , help='If passed, will train on the CPU.' ) # New Code # parser.add_argument('--num_folds' , type=__a , default=3 , help='The number of splits to perform across the dataset' ) _a : Any = parser.parse_args() _a : int = {'lr': 2e-5, 'num_epochs': 3, 'seed': 4_2, 'batch_size': 1_6} training_function(__a , __a ) if __name__ == "__main__": main()
271
1
'''simple docstring''' import unittest import numpy as np def UpperCAmelCase_ (__a : np.ndarray , __a : np.ndarray , __a : np.ndarray , __a : np.ndarray | None = None , ): """simple docstring""" _a : Optional[int] = np.shape(__a ) _a : Dict = np.shape(__a ) _a : List[Any] = np.shape(__a ) if shape_a[0] != shape_b[0]: _a : int = ( 'Expected the same number of rows for A and B. ' f"""Instead found A of size {shape_a} and B of size {shape_b}""" ) raise ValueError(__a ) if shape_b[1] != shape_c[1]: _a : Tuple = ( 'Expected the same number of columns for B and C. ' f"""Instead found B of size {shape_b} and C of size {shape_c}""" ) raise ValueError(__a ) _a : Optional[Any] = pseudo_inv if a_inv is None: try: _a : Union[str, Any] = np.linalg.inv(__a ) except np.linalg.LinAlgError: raise ValueError( 'Input matrix A is not invertible. Cannot compute Schur complement.' ) return mat_c - mat_b.T @ a_inv @ mat_b class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : List[str] ): '''simple docstring''' _a : int = np.array([[1, 2, 1], [2, 1, 2], [3, 2, 4]] ) _a : Dict = np.array([[0, 3], [3, 0], [2, 3]] ) _a : List[str] = np.array([[2, 1], [6, 3]] ) _a : Any = schur_complement(_a ,_a ,_a ) _a : Any = np.block([[a, b], [b.T, c]] ) _a : str = np.linalg.det(_a ) _a : Dict = np.linalg.det(_a ) _a : Union[str, Any] = np.linalg.det(_a ) self.assertAlmostEqual(_a ,det_a * det_s ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : Optional[Any] = np.array([[1, 2, 1], [2, 1, 2], [3, 2, 4]] ) _a : Optional[Any] = np.array([[0, 3], [3, 0], [2, 3]] ) _a : str = np.array([[2, 1], [6, 3]] ) with self.assertRaises(_a ): schur_complement(_a ,_a ,_a ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Optional[Any] = np.array([[1, 2, 1], [2, 1, 2], [3, 2, 4]] ) _a : Optional[Any] = np.array([[0, 3], [3, 0], [2, 3]] ) _a : Optional[Any] = np.array([[2, 1, 3], [6, 3, 5]] ) with self.assertRaises(_a ): schur_complement(_a ,_a ,_a ) if __name__ == "__main__": import doctest doctest.testmod() unittest.main()
271
'''simple docstring''' from __future__ import annotations __lowerCAmelCase = [-1_0, -5, 0, 5, 5.1, 1_1, 1_3, 2_1, 3, 4, -2_1, -1_0, -5, -1, 0] __lowerCAmelCase = [-5, 0, 5, 5.1, 1_1, 1_3, 2_1, -1, 4, -1, -1_0, -5, -1, 0, -1] def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : Optional[int] = [] _a : int = len(__a ) for i in range(__a ): _a : float = -1 for j in range(i + 1 , __a ): if arr[i] < arr[j]: _a : Any = arr[j] break result.append(__a ) return result def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : Tuple = [] for i, outer in enumerate(__a ): _a : float = -1 for inner in arr[i + 1 :]: if outer < inner: _a : Dict = inner break result.append(__a ) return result def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : int = len(__a ) _a : list[float] = [] _a : list[float] = [-1] * arr_size for index in reversed(range(__a ) ): if stack: while stack[-1] <= arr[index]: stack.pop() if not stack: break if stack: _a : Dict = stack[-1] stack.append(arr[index] ) return result if __name__ == "__main__": from doctest import testmod from timeit import timeit testmod() print(next_greatest_element_slow(arr)) print(next_greatest_element_fast(arr)) print(next_greatest_element(arr)) __lowerCAmelCase = ( """from __main__ import arr, next_greatest_element_slow, """ """next_greatest_element_fast, next_greatest_element""" ) print( """next_greatest_element_slow():""", timeit("""next_greatest_element_slow(arr)""", setup=setup), ) print( """next_greatest_element_fast():""", timeit("""next_greatest_element_fast(arr)""", setup=setup), ) print( """ next_greatest_element():""", timeit("""next_greatest_element(arr)""", setup=setup), )
271
1
'''simple docstring''' import os def UpperCAmelCase_ (__a : List[str] ): """simple docstring""" _a : Dict = len(grid[0] ) _a : Dict = len(__a ) _a : Dict = 0 _a : Optional[int] = 0 _a : Any = 0 # Check vertically, horizontally, diagonally at the same time (only works # for nxn grid) for i in range(__a ): for j in range(n_rows - 3 ): _a : int = grid[j][i] * grid[j + 1][i] * grid[j + 2][i] * grid[j + 3][i] _a : Any = grid[i][j] * grid[i][j + 1] * grid[i][j + 2] * grid[i][j + 3] # Left-to-right diagonal (\) product if i < n_columns - 3: _a : Union[str, Any] = ( grid[i][j] * grid[i + 1][j + 1] * grid[i + 2][j + 2] * grid[i + 3][j + 3] ) # Right-to-left diagonal(/) product if i > 2: _a : Tuple = ( grid[i][j] * grid[i - 1][j + 1] * grid[i - 2][j + 2] * grid[i - 3][j + 3] ) _a : Union[str, Any] = max( __a , __a , __a , __a ) if max_product > largest: _a : Any = max_product return largest def UpperCAmelCase_ (): """simple docstring""" _a : Any = [] with open(os.path.dirname(__a ) + '/grid.txt' ) as file: for line in file: grid.append(line.strip('\n' ).split(' ' ) ) _a : Any = [[int(__a ) for i in grid[j]] for j in range(len(__a ) )] return largest_product(__a ) if __name__ == "__main__": print(solution())
271
'''simple docstring''' import os from huggingface_hub.constants import HUGGINGFACE_HUB_CACHE, hf_cache_home __lowerCAmelCase = HUGGINGFACE_HUB_CACHE __lowerCAmelCase = """config.json""" __lowerCAmelCase = """diffusion_pytorch_model.bin""" __lowerCAmelCase = """diffusion_flax_model.msgpack""" __lowerCAmelCase = """model.onnx""" __lowerCAmelCase = """diffusion_pytorch_model.safetensors""" __lowerCAmelCase = """weights.pb""" __lowerCAmelCase = """https://huggingface.co""" __lowerCAmelCase = default_cache_path __lowerCAmelCase = """diffusers_modules""" __lowerCAmelCase = os.getenv("""HF_MODULES_CACHE""", os.path.join(hf_cache_home, """modules""")) __lowerCAmelCase = ["""fp16""", """non-ema"""] __lowerCAmelCase = """.self_attn"""
271
1
'''simple docstring''' import argparse import json import os import pickle import shutil import numpy as np import torch from distiller import Distiller from lm_seqs_dataset import LmSeqsDataset from transformers import ( BertConfig, BertForMaskedLM, BertTokenizer, DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer, GPTaConfig, GPTaLMHeadModel, GPTaTokenizer, RobertaConfig, RobertaForMaskedLM, RobertaTokenizer, ) from utils import git_log, init_gpu_params, logger, set_seed __lowerCAmelCase = { """distilbert""": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer), """roberta""": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer), """bert""": (BertConfig, BertForMaskedLM, BertTokenizer), """gpt2""": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer), } def UpperCAmelCase_ (__a : Optional[Any] ): """simple docstring""" assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0) assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0) if args.mlm: assert os.path.isfile(args.token_counts ) assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"]) else: assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"]) assert args.teacher_type == args.student_type or ( args.student_type == "distilbert" and args.teacher_type == "bert" ) assert os.path.isfile(args.student_config ) if args.student_pretrained_weights is not None: assert os.path.isfile(args.student_pretrained_weights ) if args.freeze_token_type_embds: assert args.student_type in ["roberta"] assert args.alpha_ce >= 0.0 assert args.alpha_mlm >= 0.0 assert args.alpha_clm >= 0.0 assert args.alpha_mse >= 0.0 assert args.alpha_cos >= 0.0 assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0 def UpperCAmelCase_ (__a : Optional[int] , __a : Union[str, Any] ): """simple docstring""" if args.student_type == "roberta": _a : Optional[Any] = False elif args.student_type == "gpt2": _a : int = False def UpperCAmelCase_ (__a : Any , __a : Dict ): """simple docstring""" if args.student_type == "roberta": _a : Dict = False def UpperCAmelCase_ (): """simple docstring""" _a : Optional[int] = argparse.ArgumentParser(description='Training' ) parser.add_argument('--force' , action='store_true' , help='Overwrite dump_path if it already exists.' ) parser.add_argument( '--dump_path' , type=__a , required=__a , help='The output directory (log, checkpoints, parameters, etc.)' ) parser.add_argument( '--data_file' , type=__a , required=__a , help='The binarized file (tokenized + tokens_to_ids) and grouped by sequence.' , ) parser.add_argument( '--student_type' , type=__a , choices=['distilbert', 'roberta', 'gpt2'] , required=__a , help='The student type (DistilBERT, RoBERTa).' , ) parser.add_argument('--student_config' , type=__a , required=__a , help='Path to the student configuration.' ) parser.add_argument( '--student_pretrained_weights' , default=__a , type=__a , help='Load student initialization checkpoint.' ) parser.add_argument( '--teacher_type' , choices=['bert', 'roberta', 'gpt2'] , required=__a , help='Teacher type (BERT, RoBERTa).' ) parser.add_argument('--teacher_name' , type=__a , required=__a , help='The teacher model.' ) parser.add_argument('--temperature' , default=2.0 , type=__a , help='Temperature for the softmax temperature.' ) parser.add_argument( '--alpha_ce' , default=0.5 , type=__a , help='Linear weight for the distillation loss. Must be >=0.' ) parser.add_argument( '--alpha_mlm' , default=0.0 , type=__a , help='Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.' , ) parser.add_argument('--alpha_clm' , default=0.5 , type=__a , help='Linear weight for the CLM loss. Must be >=0.' ) parser.add_argument('--alpha_mse' , default=0.0 , type=__a , help='Linear weight of the MSE loss. Must be >=0.' ) parser.add_argument( '--alpha_cos' , default=0.0 , type=__a , help='Linear weight of the cosine embedding loss. Must be >=0.' ) parser.add_argument( '--mlm' , action='store_true' , help='The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.' ) parser.add_argument( '--mlm_mask_prop' , default=0.15 , type=__a , help='Proportion of tokens for which we need to make a prediction.' , ) parser.add_argument('--word_mask' , default=0.8 , type=__a , help='Proportion of tokens to mask out.' ) parser.add_argument('--word_keep' , default=0.1 , type=__a , help='Proportion of tokens to keep.' ) parser.add_argument('--word_rand' , default=0.1 , type=__a , help='Proportion of tokens to randomly replace.' ) parser.add_argument( '--mlm_smoothing' , default=0.7 , type=__a , help='Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).' , ) parser.add_argument('--token_counts' , type=__a , help='The token counts in the data_file for MLM.' ) parser.add_argument( '--restrict_ce_to_mask' , action='store_true' , help='If true, compute the distillation loss only the [MLM] prediction distribution.' , ) parser.add_argument( '--freeze_pos_embs' , action='store_true' , help='Freeze positional embeddings during distillation. For student_type in [\'roberta\', \'gpt2\'] only.' , ) parser.add_argument( '--freeze_token_type_embds' , action='store_true' , help='Freeze token type embeddings during distillation if existent. For student_type in [\'roberta\'] only.' , ) parser.add_argument('--n_epoch' , type=__a , default=3 , help='Number of pass on the whole dataset.' ) parser.add_argument('--batch_size' , type=__a , default=5 , help='Batch size (for each process).' ) parser.add_argument( '--group_by_size' , action='store_false' , help='If true, group sequences that have similar length into the same batch. Default is true.' , ) parser.add_argument( '--gradient_accumulation_steps' , type=__a , default=5_0 , help='Gradient accumulation for larger training batches.' , ) parser.add_argument('--warmup_prop' , default=0.05 , type=__a , help='Linear warmup proportion.' ) parser.add_argument('--weight_decay' , default=0.0 , type=__a , help='Weight decay if we apply some.' ) parser.add_argument('--learning_rate' , default=5e-4 , type=__a , help='The initial learning rate for Adam.' ) parser.add_argument('--adam_epsilon' , default=1e-6 , type=__a , help='Epsilon for Adam optimizer.' ) parser.add_argument('--max_grad_norm' , default=5.0 , type=__a , help='Max gradient norm.' ) parser.add_argument('--initializer_range' , default=0.02 , type=__a , help='Random initialization range.' ) parser.add_argument( '--fp16' , action='store_true' , help='Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit' , ) parser.add_argument( '--fp16_opt_level' , type=__a , default='O1' , help=( 'For fp16: Apex AMP optimization level selected in [\'O0\', \'O1\', \'O2\', and \'O3\'].' 'See details at https://nvidia.github.io/apex/amp.html' ) , ) parser.add_argument('--n_gpu' , type=__a , default=1 , help='Number of GPUs in the node.' ) parser.add_argument('--local_rank' , type=__a , default=-1 , help='Distributed training - Local rank' ) parser.add_argument('--seed' , type=__a , default=5_6 , help='Random seed' ) parser.add_argument('--log_interval' , type=__a , default=5_0_0 , help='Tensorboard logging interval.' ) parser.add_argument('--checkpoint_interval' , type=__a , default=4_0_0_0 , help='Checkpoint interval.' ) _a : Any = parser.parse_args() sanity_checks(__a ) # ARGS # init_gpu_params(__a ) set_seed(__a ) if args.is_master: if os.path.exists(args.dump_path ): if not args.force: raise ValueError( f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite""" ' itUse `--force` if you want to overwrite it' ) else: shutil.rmtree(args.dump_path ) if not os.path.exists(args.dump_path ): os.makedirs(args.dump_path ) logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" ) # SAVE PARAMS # logger.info(f"""Param: {args}""" ) with open(os.path.join(args.dump_path , 'parameters.json' ) , 'w' ) as f: json.dump(vars(__a ) , __a , indent=4 ) git_log(args.dump_path ) _a, _a, _a : Union[str, Any] = MODEL_CLASSES[args.student_type] _a, _a, _a : Tuple = MODEL_CLASSES[args.teacher_type] # TOKENIZER # _a : List[str] = teacher_tokenizer_class.from_pretrained(args.teacher_name ) _a : Optional[int] = {} for tok_name, tok_symbol in tokenizer.special_tokens_map.items(): _a : Any = tokenizer.all_special_tokens.index(__a ) _a : int = tokenizer.all_special_ids[idx] logger.info(f"""Special tokens {special_tok_ids}""" ) _a : List[str] = special_tok_ids _a : Any = tokenizer.max_model_input_sizes[args.teacher_name] # DATA LOADER # logger.info(f"""Loading data from {args.data_file}""" ) with open(args.data_file , 'rb' ) as fp: _a : Optional[int] = pickle.load(__a ) if args.mlm: logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" ) with open(args.token_counts , 'rb' ) as fp: _a : Tuple = pickle.load(__a ) _a : Union[str, Any] = np.maximum(__a , 1 ) ** -args.mlm_smoothing for idx in special_tok_ids.values(): _a : str = 0.0 # do not predict special tokens _a : Optional[int] = torch.from_numpy(__a ) else: _a : Optional[Any] = None _a : Union[str, Any] = LmSeqsDataset(params=__a , data=__a ) logger.info('Data loader created.' ) # STUDENT # logger.info(f"""Loading student config from {args.student_config}""" ) _a : Dict = student_config_class.from_pretrained(args.student_config ) _a : int = True if args.student_pretrained_weights is not None: logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" ) _a : List[str] = student_model_class.from_pretrained(args.student_pretrained_weights , config=__a ) else: _a : List[str] = student_model_class(__a ) if args.n_gpu > 0: student.to(f"""cuda:{args.local_rank}""" ) logger.info('Student loaded.' ) # TEACHER # _a : int = teacher_model_class.from_pretrained(args.teacher_name , output_hidden_states=__a ) if args.n_gpu > 0: teacher.to(f"""cuda:{args.local_rank}""" ) logger.info(f"""Teacher loaded from {args.teacher_name}.""" ) # FREEZING # if args.freeze_pos_embs: freeze_pos_embeddings(__a , __a ) if args.freeze_token_type_embds: freeze_token_type_embeddings(__a , __a ) # SANITY CHECKS # assert student.config.vocab_size == teacher.config.vocab_size assert student.config.hidden_size == teacher.config.hidden_size assert student.config.max_position_embeddings == teacher.config.max_position_embeddings if args.mlm: assert token_probs.size(0 ) == stu_architecture_config.vocab_size # DISTILLER # torch.cuda.empty_cache() _a : Dict = Distiller( params=__a , dataset=__a , token_probs=__a , student=__a , teacher=__a ) distiller.train() logger.info('Let\'s go get some drinks.' ) if __name__ == "__main__": main()
271
'''simple docstring''' import inspect import unittest import numpy as np from tests.test_modeling_common import floats_tensor from transformers import MaskaFormerConfig, is_torch_available, is_vision_available from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device from transformers.utils import cached_property from ...test_configuration_common import ConfigTester from ...test_modeling_common import ModelTesterMixin from ...test_pipeline_mixin import PipelineTesterMixin if is_torch_available(): import torch from transformers import MaskaFormerForUniversalSegmentation, MaskaFormerModel if is_vision_available(): from transformers import MaskaFormerImageProcessor if is_vision_available(): from PIL import Image class UpperCAmelCase__ : """simple docstring""" def __init__( self : int ,_a : Any ,_a : Optional[int]=2 ,_a : Optional[Any]=True ,_a : Dict=False ,_a : Dict=10 ,_a : Any=3 ,_a : str=32 * 8 ,_a : Optional[int]=32 * 8 ,_a : int=4 ,_a : str=64 ,): '''simple docstring''' _a : Dict = parent _a : Union[str, Any] = batch_size _a : Tuple = is_training _a : List[str] = use_auxiliary_loss _a : Optional[Any] = num_queries _a : str = num_channels _a : List[str] = min_size _a : int = max_size _a : Optional[int] = num_labels _a : List[str] = hidden_dim _a : int = hidden_dim def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Tuple = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to( _a ) _a : Optional[Any] = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=_a ) _a : Union[str, Any] = ( torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=_a ) > 0.5 ).float() _a : Tuple = (torch.rand((self.batch_size, self.num_labels) ,device=_a ) > 0.5).long() _a : Dict = self.get_config() return config, pixel_values, pixel_mask, mask_labels, class_labels def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : int = MaskaFormerConfig( hidden_size=self.hidden_dim ,) _a : str = self.num_queries _a : Union[str, Any] = self.num_labels _a : Tuple = [1, 1, 1, 1] _a : Dict = self.num_channels _a : str = 64 _a : Tuple = 128 _a : Optional[Any] = self.hidden_dim _a : Union[str, Any] = self.hidden_dim _a : List[Any] = self.hidden_dim return config def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a, _a, _a, _a, _a : Optional[Any] = self.prepare_config_and_inputs() _a : str = {'pixel_values': pixel_values, 'pixel_mask': pixel_mask} return config, inputs_dict def __lowercase ( self : List[str] ,_a : Optional[Any] ,_a : str ): '''simple docstring''' _a : str = output.encoder_hidden_states _a : Any = output.pixel_decoder_hidden_states _a : Optional[Any] = output.transformer_decoder_hidden_states self.parent.assertTrue(len(_a ) ,len(config.backbone_config.depths ) ) self.parent.assertTrue(len(_a ) ,len(config.backbone_config.depths ) ) self.parent.assertTrue(len(_a ) ,config.decoder_layers ) def __lowercase ( self : List[str] ,_a : str ,_a : List[Any] ,_a : Any ,_a : Union[str, Any]=False ): '''simple docstring''' with torch.no_grad(): _a : str = MaskaFormerModel(config=_a ) model.to(_a ) model.eval() _a : Any = model(pixel_values=_a ,pixel_mask=_a ) _a : Optional[Any] = model(_a ,output_hidden_states=_a ) self.parent.assertEqual( output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.hidden_dim) ,) # let's ensure the other two hidden state exists self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None ) self.parent.assertTrue(output.encoder_last_hidden_state is not None ) if output_hidden_states: self.check_output_hidden_state(_a ,_a ) def __lowercase ( self : Tuple ,_a : List[Any] ,_a : Union[str, Any] ,_a : Tuple ,_a : List[str] ,_a : Any ): '''simple docstring''' _a : int = MaskaFormerForUniversalSegmentation(config=_a ) model.to(_a ) model.eval() def comm_check_on_output(_a : Any ): # let's still check that all the required stuff is there self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None ) self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None ) self.parent.assertTrue(result.encoder_last_hidden_state is not None ) # okay, now we need to check the logits shape # due to the encoder compression, masks have a //4 spatial size self.parent.assertEqual( result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,) # + 1 for null class self.parent.assertEqual( result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) ) with torch.no_grad(): _a : Any = model(pixel_values=_a ,pixel_mask=_a ) _a : Optional[int] = model(_a ) comm_check_on_output(_a ) _a : List[str] = model( pixel_values=_a ,pixel_mask=_a ,mask_labels=_a ,class_labels=_a ) comm_check_on_output(_a ) self.parent.assertTrue(result.loss is not None ) self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) ) @require_torch class UpperCAmelCase__ ( lowercase__ , lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[int] = (MaskaFormerModel, MaskaFormerForUniversalSegmentation) if is_torch_available() else () __UpperCAmelCase : Dict = {'''feature-extraction''': MaskaFormerModel} if is_torch_available() else {} __UpperCAmelCase : Dict = False __UpperCAmelCase : Tuple = False __UpperCAmelCase : Dict = False __UpperCAmelCase : List[Any] = False def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Union[str, Any] = MaskaFormerModelTester(self ) _a : Dict = ConfigTester(self ,config_class=_a ,has_text_modality=_a ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' self.config_tester.run_common_tests() def __lowercase ( self : Optional[int] ): '''simple docstring''' _a, _a : List[str] = self.model_tester.prepare_config_and_inputs_for_common() self.model_tester.create_and_check_maskaformer_model(_a ,**_a ,output_hidden_states=_a ) def __lowercase ( self : str ): '''simple docstring''' _a : Dict = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_maskaformer_instance_segmentation_head_model(*_a ) @unittest.skip(reason='Mask2Former does not use inputs_embeds' ) def __lowercase ( self : Any ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former does not have a get_input_embeddings method' ) def __lowercase ( self : str ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former is not a generative model' ) def __lowercase ( self : List[Any] ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former does not use token embeddings' ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' pass @require_torch_multi_gpu @unittest.skip( reason='Mask2Former has some layers using `add_module` which doesn\'t work well with `nn.DataParallel`' ) def __lowercase ( self : Dict ): '''simple docstring''' pass @unittest.skip('Will be fixed soon by reducing the size of the model used for common tests.' ) def __lowercase ( self : List[Any] ): '''simple docstring''' pass def __lowercase ( self : int ): '''simple docstring''' _a, _a : Any = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a : Union[str, Any] = model_class(_a ) _a : List[str] = inspect.signature(model.forward ) # signature.parameters is an OrderedDict => so arg_names order is deterministic _a : Optional[Any] = [*signature.parameters.keys()] _a : List[Any] = ['pixel_values'] self.assertListEqual(arg_names[:1] ,_a ) @slow def __lowercase ( self : List[str] ): '''simple docstring''' for model_name in ["facebook/mask2former-swin-small-coco-instance"]: _a : Dict = MaskaFormerModel.from_pretrained(_a ) self.assertIsNotNone(_a ) def __lowercase ( self : List[Any] ): '''simple docstring''' _a : int = (self.model_tester.min_size,) * 2 _a : Any = { 'pixel_values': torch.randn((2, 3, *size) ,device=_a ), 'mask_labels': torch.randn((2, 10, *size) ,device=_a ), 'class_labels': torch.zeros(2 ,10 ,device=_a ).long(), } _a : List[Any] = self.model_tester.get_config() _a : int = MaskaFormerForUniversalSegmentation(_a ).to(_a ) _a : str = model(**_a ) self.assertTrue(outputs.loss is not None ) def __lowercase ( self : List[str] ): '''simple docstring''' _a, _a : List[Any] = self.model_tester.prepare_config_and_inputs_for_common() self.model_tester.create_and_check_maskaformer_model(_a ,**_a ,output_hidden_states=_a ) def __lowercase ( self : int ): '''simple docstring''' _a, _a : Dict = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a : Any = model_class(_a ).to(_a ) _a : Optional[int] = model(**_a ,output_attentions=_a ) self.assertTrue(outputs.attentions is not None ) def __lowercase ( self : Tuple ): '''simple docstring''' if not self.model_tester.is_training: return _a : List[str] = self.all_model_classes[1] _a, _a, _a, _a, _a : List[str] = self.model_tester.prepare_config_and_inputs() _a : Any = model_class(_a ) model.to(_a ) model.train() _a : Union[str, Any] = model(_a ,mask_labels=_a ,class_labels=_a ).loss loss.backward() def __lowercase ( self : int ): '''simple docstring''' _a : int = self.all_model_classes[1] _a, _a, _a, _a, _a : List[Any] = self.model_tester.prepare_config_and_inputs() _a : str = True _a : str = True _a : List[str] = model_class(_a ).to(_a ) model.train() _a : Optional[int] = model(_a ,mask_labels=_a ,class_labels=_a ) _a : Tuple = outputs.encoder_hidden_states[0] encoder_hidden_states.retain_grad() _a : str = outputs.pixel_decoder_hidden_states[0] pixel_decoder_hidden_states.retain_grad() _a : Dict = outputs.transformer_decoder_hidden_states[0] transformer_decoder_hidden_states.retain_grad() _a : List[str] = outputs.attentions[0] attentions.retain_grad() outputs.loss.backward(retain_graph=_a ) self.assertIsNotNone(encoder_hidden_states.grad ) self.assertIsNotNone(pixel_decoder_hidden_states.grad ) self.assertIsNotNone(transformer_decoder_hidden_states.grad ) self.assertIsNotNone(attentions.grad ) __lowerCAmelCase = 1e-4 def UpperCAmelCase_ (): """simple docstring""" _a : int = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' ) return image @require_vision @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @cached_property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' return "facebook/mask2former-swin-small-coco-instance" @cached_property def __lowercase ( self : Any ): '''simple docstring''' return MaskaFormerImageProcessor.from_pretrained(self.model_checkpoints ) if is_vision_available() else None def __lowercase ( self : Any ): '''simple docstring''' _a : List[str] = MaskaFormerModel.from_pretrained(self.model_checkpoints ).to(_a ) _a : int = self.default_image_processor _a : Tuple = prepare_img() _a : Any = image_processor(_a ,return_tensors='pt' ).to(_a ) _a : Union[str, Any] = inputs['pixel_values'].shape # check size is divisible by 32 self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 ) # check size self.assertEqual(_a ,(1, 3, 384, 384) ) with torch.no_grad(): _a : Optional[Any] = model(**_a ) _a : List[Any] = torch.tensor( [[-0.2790, -1.0717, -1.1668], [-0.5128, -0.3128, -0.4987], [-0.5832, 0.1971, -0.0197]] ).to(_a ) self.assertTrue( torch.allclose( outputs.encoder_last_hidden_state[0, 0, :3, :3] ,_a ,atol=_a ) ) _a : str = torch.tensor( [[0.8973, 1.1847, 1.1776], [1.1934, 1.5040, 1.5128], [1.1153, 1.4486, 1.4951]] ).to(_a ) self.assertTrue( torch.allclose( outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,_a ,atol=_a ) ) _a : Any = torch.tensor( [[2.1152, 1.7000, -0.8603], [1.5808, 1.8004, -0.9353], [1.6043, 1.7495, -0.5999]] ).to(_a ) self.assertTrue( torch.allclose( outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,_a ,atol=_a ) ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[Any] = MaskaFormerForUniversalSegmentation.from_pretrained(self.model_checkpoints ).to(_a ).eval() _a : Optional[Any] = self.default_image_processor _a : List[Any] = prepare_img() _a : str = image_processor(_a ,return_tensors='pt' ).to(_a ) _a : Any = inputs['pixel_values'].shape # check size is divisible by 32 self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 ) # check size self.assertEqual(_a ,(1, 3, 384, 384) ) with torch.no_grad(): _a : Optional[int] = model(**_a ) # masks_queries_logits _a : Dict = outputs.masks_queries_logits self.assertEqual( masks_queries_logits.shape ,(1, model.config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ) _a : Dict = [ [-8.7839, -9.0056, -8.8121], [-7.4104, -7.0313, -6.5401], [-6.6105, -6.3427, -6.4675], ] _a : Optional[Any] = torch.tensor(_a ).to(_a ) self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,_a ,atol=_a ) ) # class_queries_logits _a : str = outputs.class_queries_logits self.assertEqual(class_queries_logits.shape ,(1, model.config.num_queries, model.config.num_labels + 1) ) _a : str = torch.tensor( [ [1.8324, -8.0835, -4.1922], [0.8450, -9.0050, -3.6053], [0.3045, -7.7293, -3.0275], ] ).to(_a ) self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,_a ,atol=_a ) ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : Any = MaskaFormerForUniversalSegmentation.from_pretrained(self.model_checkpoints ).to(_a ).eval() _a : Tuple = self.default_image_processor _a : Tuple = image_processor( [np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors='pt' ,) _a : str = inputs['pixel_values'].to(_a ) _a : str = [el.to(_a ) for el in inputs['mask_labels']] _a : Dict = [el.to(_a ) for el in inputs['class_labels']] with torch.no_grad(): _a : List[str] = model(**_a ) self.assertTrue(outputs.loss is not None )
271
1
'''simple docstring''' import os from argparse import ArgumentParser, Namespace from ..data import SingleSentenceClassificationProcessor as Processor from ..pipelines import TextClassificationPipeline from ..utils import is_tf_available, is_torch_available, logging from . import BaseTransformersCLICommand if not is_tf_available() and not is_torch_available(): raise RuntimeError("""At least one of PyTorch or TensorFlow 2.0+ should be installed to use CLI training""") # TF training parameters __lowerCAmelCase = False __lowerCAmelCase = False def UpperCAmelCase_ (__a : Namespace ): """simple docstring""" return TrainCommand(__a ) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" @staticmethod def __lowercase ( _a : ArgumentParser ): '''simple docstring''' _a : Optional[Any] = parser.add_parser('train' ,help='CLI tool to train a model on a task.' ) train_parser.add_argument( '--train_data' ,type=_a ,required=_a ,help='path to train (and optionally evaluation) dataset as a csv with tab separated labels and sentences.' ,) train_parser.add_argument( '--column_label' ,type=_a ,default=0 ,help='Column of the dataset csv file with example labels.' ) train_parser.add_argument( '--column_text' ,type=_a ,default=1 ,help='Column of the dataset csv file with example texts.' ) train_parser.add_argument( '--column_id' ,type=_a ,default=2 ,help='Column of the dataset csv file with example ids.' ) train_parser.add_argument( '--skip_first_row' ,action='store_true' ,help='Skip the first row of the csv file (headers).' ) train_parser.add_argument('--validation_data' ,type=_a ,default='' ,help='path to validation dataset.' ) train_parser.add_argument( '--validation_split' ,type=_a ,default=0.1 ,help='if validation dataset is not provided, fraction of train dataset to use as validation dataset.' ,) train_parser.add_argument('--output' ,type=_a ,default='./' ,help='path to saved the trained model.' ) train_parser.add_argument( '--task' ,type=_a ,default='text_classification' ,help='Task to train the model on.' ) train_parser.add_argument( '--model' ,type=_a ,default='bert-base-uncased' ,help='Model\'s name or path to stored model.' ) train_parser.add_argument('--train_batch_size' ,type=_a ,default=32 ,help='Batch size for training.' ) train_parser.add_argument('--valid_batch_size' ,type=_a ,default=64 ,help='Batch size for validation.' ) train_parser.add_argument('--learning_rate' ,type=_a ,default=3E-5 ,help='Learning rate.' ) train_parser.add_argument('--adam_epsilon' ,type=_a ,default=1E-08 ,help='Epsilon for Adam optimizer.' ) train_parser.set_defaults(func=_a ) def __init__( self : int ,_a : Namespace ): '''simple docstring''' _a : Union[str, Any] = logging.get_logger('transformers-cli/training' ) _a : int = 'tf' if is_tf_available() else 'torch' os.makedirs(args.output ,exist_ok=_a ) _a : List[str] = args.output _a : Optional[Any] = args.column_label _a : Any = args.column_text _a : Union[str, Any] = args.column_id self.logger.info(F"""Loading {args.task} pipeline for {args.model}""" ) if args.task == "text_classification": _a : Dict = TextClassificationPipeline.from_pretrained(args.model ) elif args.task == "token_classification": raise NotImplementedError elif args.task == "question_answering": raise NotImplementedError self.logger.info(F"""Loading dataset from {args.train_data}""" ) _a : Any = Processor.create_from_csv( args.train_data ,column_label=args.column_label ,column_text=args.column_text ,column_id=args.column_id ,skip_first_row=args.skip_first_row ,) _a : Tuple = None if args.validation_data: self.logger.info(F"""Loading validation dataset from {args.validation_data}""" ) _a : Optional[int] = Processor.create_from_csv( args.validation_data ,column_label=args.column_label ,column_text=args.column_text ,column_id=args.column_id ,skip_first_row=args.skip_first_row ,) _a : Any = args.validation_split _a : Optional[Any] = args.train_batch_size _a : Dict = args.valid_batch_size _a : Optional[int] = args.learning_rate _a : int = args.adam_epsilon def __lowercase ( self : Any ): '''simple docstring''' if self.framework == "tf": return self.run_tf() return self.run_torch() def __lowercase ( self : Tuple ): '''simple docstring''' raise NotImplementedError def __lowercase ( self : List[str] ): '''simple docstring''' self.pipeline.fit( self.train_dataset ,validation_data=self.valid_dataset ,validation_split=self.validation_split ,learning_rate=self.learning_rate ,adam_epsilon=self.adam_epsilon ,train_batch_size=self.train_batch_size ,valid_batch_size=self.valid_batch_size ,) # Save trained pipeline self.pipeline.save_pretrained(self.output )
271
'''simple docstring''' import argparse import json from typing import List from ltp import LTP from transformers import BertTokenizer def UpperCAmelCase_ (__a : List[Any] ): """simple docstring""" if ( (cp >= 0x4E_00 and cp <= 0x9F_FF) or (cp >= 0x34_00 and cp <= 0x4D_BF) # or (cp >= 0x2_00_00 and cp <= 0x2_A6_DF) # or (cp >= 0x2_A7_00 and cp <= 0x2_B7_3F) # or (cp >= 0x2_B7_40 and cp <= 0x2_B8_1F) # or (cp >= 0x2_B8_20 and cp <= 0x2_CE_AF) # or (cp >= 0xF9_00 and cp <= 0xFA_FF) or (cp >= 0x2_F8_00 and cp <= 0x2_FA_1F) # ): # return True return False def UpperCAmelCase_ (__a : str ): """simple docstring""" for char in word: _a : Union[str, Any] = ord(__a ) if not _is_chinese_char(__a ): return 0 return 1 def UpperCAmelCase_ (__a : List[str] ): """simple docstring""" _a : Dict = set() for token in tokens: _a : str = len(__a ) > 1 and is_chinese(__a ) if chinese_word: word_set.add(__a ) _a : Optional[Any] = list(__a ) return word_list def UpperCAmelCase_ (__a : List[str] , __a : set() ): """simple docstring""" if not chinese_word_set: return bert_tokens _a : Optional[Any] = max([len(__a ) for w in chinese_word_set] ) _a : Optional[int] = bert_tokens _a, _a : Any = 0, len(__a ) while start < end: _a : Tuple = True if is_chinese(bert_word[start] ): _a : Union[str, Any] = min(end - start , __a ) for i in range(__a , 1 , -1 ): _a : Optional[Any] = ''.join(bert_word[start : start + i] ) if whole_word in chinese_word_set: for j in range(start + 1 , start + i ): _a : Any = '##' + bert_word[j] _a : Union[str, Any] = start + i _a : int = False break if single_word: start += 1 return bert_word def UpperCAmelCase_ (__a : List[str] , __a : LTP , __a : BertTokenizer ): """simple docstring""" _a : int = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : Union[str, Any] = ltp_tokenizer.seg(lines[i : i + 1_0_0] )[0] _a : Optional[Any] = [get_chinese_word(__a ) for r in res] ltp_res.extend(__a ) assert len(__a ) == len(__a ) _a : str = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : List[str] = bert_tokenizer(lines[i : i + 1_0_0] , add_special_tokens=__a , truncation=__a , max_length=5_1_2 ) bert_res.extend(res['input_ids'] ) assert len(__a ) == len(__a ) _a : List[str] = [] for input_ids, chinese_word in zip(__a , __a ): _a : int = [] for id in input_ids: _a : Optional[int] = bert_tokenizer._convert_id_to_token(__a ) input_tokens.append(__a ) _a : List[str] = add_sub_symbol(__a , __a ) _a : Tuple = [] # We only save pos of chinese subwords start with ##, which mean is part of a whole word. for i, token in enumerate(__a ): if token[:2] == "##": _a : str = token[2:] # save chinese tokens' pos if len(__a ) == 1 and _is_chinese_char(ord(__a ) ): ref_id.append(__a ) ref_ids.append(__a ) assert len(__a ) == len(__a ) return ref_ids def UpperCAmelCase_ (__a : Optional[Any] ): """simple docstring""" with open(args.file_name , 'r' , encoding='utf-8' ) as f: _a : Dict = f.readlines() _a : int = [line.strip() for line in data if len(__a ) > 0 and not line.isspace()] # avoid delimiter like '\u2029' _a : int = LTP(args.ltp ) # faster in GPU device _a : Tuple = BertTokenizer.from_pretrained(args.bert ) _a : int = prepare_ref(__a , __a , __a ) with open(args.save_path , 'w' , encoding='utf-8' ) as f: _a : Optional[Any] = [json.dumps(__a ) + '\n' for ref in ref_ids] f.writelines(__a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser(description="""prepare_chinese_ref""") parser.add_argument( """--file_name""", type=str, default="""./resources/chinese-demo.txt""", help="""file need process, same as training data in lm""", ) parser.add_argument( """--ltp""", type=str, default="""./resources/ltp""", help="""resources for LTP tokenizer, usually a path""" ) parser.add_argument("""--bert""", type=str, default="""./resources/robert""", help="""resources for Bert tokenizer""") parser.add_argument("""--save_path""", type=str, default="""./resources/ref.txt""", help="""path to save res""") __lowerCAmelCase = parser.parse_args() main(args)
271
1
'''simple docstring''' from collections.abc import Generator from math import sin def UpperCAmelCase_ (__a : bytes ): """simple docstring""" if len(__a ) != 3_2: raise ValueError('Input must be of length 32' ) _a : List[Any] = b'' for i in [3, 2, 1, 0]: little_endian += string_aa[8 * i : 8 * i + 8] return little_endian def UpperCAmelCase_ (__a : int ): """simple docstring""" if i < 0: raise ValueError('Input must be non-negative' ) _a : List[str] = format(__a , '08x' )[-8:] _a : Dict = b'' for i in [3, 2, 1, 0]: little_endian_hex += hex_rep[2 * i : 2 * i + 2].encode('utf-8' ) return little_endian_hex def UpperCAmelCase_ (__a : bytes ): """simple docstring""" _a : List[Any] = b'' for char in message: bit_string += format(__a , '08b' ).encode('utf-8' ) _a : List[str] = format(len(__a ) , '064b' ).encode('utf-8' ) # Pad bit_string to a multiple of 512 chars bit_string += b"1" while len(__a ) % 5_1_2 != 4_4_8: bit_string += b"0" bit_string += to_little_endian(start_len[3_2:] ) + to_little_endian(start_len[:3_2] ) return bit_string def UpperCAmelCase_ (__a : bytes ): """simple docstring""" if len(__a ) % 5_1_2 != 0: raise ValueError('Input must have length that\'s a multiple of 512' ) for pos in range(0 , len(__a ) , 5_1_2 ): _a : Optional[Any] = bit_string[pos : pos + 5_1_2] _a : Optional[Any] = [] for i in range(0 , 5_1_2 , 3_2 ): block_words.append(int(to_little_endian(block[i : i + 3_2] ) , 2 ) ) yield block_words def UpperCAmelCase_ (__a : int ): """simple docstring""" if i < 0: raise ValueError('Input must be non-negative' ) _a : Union[str, Any] = format(__a , '032b' ) _a : Dict = '' for c in i_str: new_str += "1" if c == "0" else "0" return int(__a , 2 ) def UpperCAmelCase_ (__a : int , __a : int ): """simple docstring""" return (a + b) % 2**3_2 def UpperCAmelCase_ (__a : int , __a : int ): """simple docstring""" if i < 0: raise ValueError('Input must be non-negative' ) if shift < 0: raise ValueError('Shift must be non-negative' ) return ((i << shift) ^ (i >> (3_2 - shift))) % 2**3_2 def UpperCAmelCase_ (__a : bytes ): """simple docstring""" _a : Optional[Any] = preprocess(__a ) _a : List[str] = [int(2**3_2 * abs(sin(i + 1 ) ) ) for i in range(6_4 )] # Starting states _a : str = 0x67_45_23_01 _a : Dict = 0xEF_CD_AB_89 _a : Union[str, Any] = 0x98_BA_DC_FE _a : List[str] = 0x10_32_54_76 _a : str = [ 7, 1_2, 1_7, 2_2, 7, 1_2, 1_7, 2_2, 7, 1_2, 1_7, 2_2, 7, 1_2, 1_7, 2_2, 5, 9, 1_4, 2_0, 5, 9, 1_4, 2_0, 5, 9, 1_4, 2_0, 5, 9, 1_4, 2_0, 4, 1_1, 1_6, 2_3, 4, 1_1, 1_6, 2_3, 4, 1_1, 1_6, 2_3, 4, 1_1, 1_6, 2_3, 6, 1_0, 1_5, 2_1, 6, 1_0, 1_5, 2_1, 6, 1_0, 1_5, 2_1, 6, 1_0, 1_5, 2_1, ] # Process bit string in chunks, each with 16 32-char words for block_words in get_block_words(__a ): _a : Union[str, Any] = aa _a : List[str] = ba _a : List[Any] = ca _a : Tuple = da # Hash current chunk for i in range(6_4 ): if i <= 1_5: # f = (b & c) | (not_32(b) & d) # Alternate definition for f _a : Dict = d ^ (b & (c ^ d)) _a : Any = i elif i <= 3_1: # f = (d & b) | (not_32(d) & c) # Alternate definition for f _a : Dict = c ^ (d & (b ^ c)) _a : Union[str, Any] = (5 * i + 1) % 1_6 elif i <= 4_7: _a : List[str] = b ^ c ^ d _a : str = (3 * i + 5) % 1_6 else: _a : Dict = c ^ (b | not_aa(__a )) _a : Dict = (7 * i) % 1_6 _a : Optional[int] = (f + a + added_consts[i] + block_words[g]) % 2**3_2 _a : Union[str, Any] = d _a : str = c _a : List[Any] = b _a : Any = sum_aa(__a , left_rotate_aa(__a , shift_amounts[i] ) ) # Add hashed chunk to running total _a : Optional[int] = sum_aa(__a , __a ) _a : Any = sum_aa(__a , __a ) _a : Tuple = sum_aa(__a , __a ) _a : Dict = sum_aa(__a , __a ) _a : str = reformat_hex(__a ) + reformat_hex(__a ) + reformat_hex(__a ) + reformat_hex(__a ) return digest if __name__ == "__main__": import doctest doctest.testmod()
271
'''simple docstring''' import warnings from ...utils import logging from .image_processing_videomae import VideoMAEImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Tuple ,*_a : List[str] ,**_a : Any ): '''simple docstring''' warnings.warn( 'The class VideoMAEFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use VideoMAEImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
1
'''simple docstring''' def UpperCAmelCase_ (__a : int ): """simple docstring""" if not isinstance(__a , __a ): raise TypeError('only integers accepted as input' ) else: _a : int = str(abs(__a ) ) _a : Any = [list(__a ) for char in range(len(__a ) )] for index in range(len(__a ) ): num_transpositions[index].pop(__a ) return max( int(''.join(list(__a ) ) ) for transposition in num_transpositions ) if __name__ == "__main__": __import__("""doctest""").testmod()
271
'''simple docstring''' from __future__ import annotations from random import choice def UpperCAmelCase_ (__a : str ): """simple docstring""" return choice(__a ) def UpperCAmelCase_ (__a : list[int] , __a : int ): """simple docstring""" _a : Dict = random_pivot(__a ) # partition based on pivot # linear time _a : Optional[int] = [e for e in lst if e < pivot] _a : List[str] = [e for e in lst if e > pivot] # if we get lucky, pivot might be the element we want. # we can easily see this: # small (elements smaller than k) # + pivot (kth element) # + big (elements larger than k) if len(__a ) == k - 1: return pivot # pivot is in elements bigger than k elif len(__a ) < k - 1: return kth_number(__a , k - len(__a ) - 1 ) # pivot is in elements smaller than k else: return kth_number(__a , __a ) if __name__ == "__main__": import doctest doctest.testmod()
271
1
'''simple docstring''' from itertools import count def UpperCAmelCase_ (__a : int = 5_0 ): """simple docstring""" _a : int = [1] * min_block_length for n in count(__a ): fill_count_functions.append(1 ) for block_length in range(__a , n + 1 ): for block_start in range(n - block_length ): fill_count_functions[n] += fill_count_functions[ n - block_start - block_length - 1 ] fill_count_functions[n] += 1 if fill_count_functions[n] > 1_0_0_0_0_0_0: break return n if __name__ == "__main__": print(f'''{solution() = }''')
271
'''simple docstring''' class UpperCAmelCase__ : """simple docstring""" def __init__( self : Dict ): '''simple docstring''' _a : Dict = {} def __lowercase ( self : Union[str, Any] ): '''simple docstring''' print(self.vertex ) for i in self.vertex: print(_a ,' -> ' ,' -> '.join([str(_a ) for j in self.vertex[i]] ) ) def __lowercase ( self : Dict ,_a : int ,_a : int ): '''simple docstring''' if from_vertex in self.vertex: self.vertex[from_vertex].append(_a ) else: # else make a new vertex _a : int = [to_vertex] def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Tuple = [False] * len(self.vertex ) # call the recursive helper function for i in range(len(self.vertex ) ): if not visited[i]: self.dfs_recursive(_a ,_a ) def __lowercase ( self : Union[str, Any] ,_a : int ,_a : list ): '''simple docstring''' _a : List[Any] = True print(_a ,end=' ' ) # Recur for all the vertices that are adjacent to this node for i in self.vertex: if not visited[i]: self.dfs_recursive(_a ,_a ) if __name__ == "__main__": __lowerCAmelCase = Graph() g.add_edge(0, 1) g.add_edge(0, 2) g.add_edge(1, 2) g.add_edge(2, 0) g.add_edge(2, 3) g.add_edge(3, 3) g.print_graph() print("""DFS:""") g.dfs() # OUTPUT: # 0 -> 1 -> 2 # 1 -> 2 # 2 -> 0 -> 3 # 3 -> 3 # DFS: # 0 1 2 3
271
1
'''simple docstring''' import argparse from torch import nn # transformers_old should correspond to branch `save_old_prophetnet_model_structure` here # original prophetnet_checkpoints are saved under `patrickvonplaten/..._old` respectively from transformers_old.modeling_prophetnet import ( ProphetNetForConditionalGeneration as ProphetNetForConditionalGenerationOld, ) from transformers_old.modeling_xlm_prophetnet import ( XLMProphetNetForConditionalGeneration as XLMProphetNetForConditionalGenerationOld, ) from transformers import ProphetNetForConditionalGeneration, XLMProphetNetForConditionalGeneration, logging __lowerCAmelCase = logging.get_logger(__name__) logging.set_verbosity_info() def UpperCAmelCase_ (__a : str , __a : str ): """simple docstring""" if "xprophetnet" in prophetnet_checkpoint_path: _a : int = XLMProphetNetForConditionalGenerationOld.from_pretrained(__a ) _a, _a : Any = XLMProphetNetForConditionalGeneration.from_pretrained( __a , output_loading_info=__a ) else: _a : Optional[int] = ProphetNetForConditionalGenerationOld.from_pretrained(__a ) _a, _a : int = ProphetNetForConditionalGeneration.from_pretrained( __a , output_loading_info=__a ) _a : Optional[int] = ['key_proj', 'value_proj', 'query_proj'] _a : Optional[int] = { 'self_attn': 'ngram_self_attn', 'cross_attn': 'encoder_attn', 'cross_attn_layer_norm': 'encoder_attn_layer_norm', 'feed_forward_layer_norm': 'final_layer_norm', 'feed_forward': '', 'intermediate': 'fc1', 'output': 'fc2', 'key_proj': 'k_proj', 'query_proj': 'q_proj', 'value_proj': 'v_proj', 'word_embeddings': 'embed_tokens', 'embeddings_layer_norm': 'emb_layer_norm', 'relative_pos_embeddings': 'relative_linear', 'ngram_embeddings': 'ngram_input_embed', 'position_embeddings': 'embed_positions', } for key in loading_info["missing_keys"]: _a : Union[str, Any] = key.split('.' ) if attributes[0] == "lm_head": _a : Union[str, Any] = prophet _a : str = prophet_old else: _a : Dict = prophet.prophetnet _a : Union[str, Any] = prophet_old.model _a : List[str] = False for attribute in attributes: if attribute in mapping: _a : Optional[int] = mapping[attribute] if not hasattr(__a , __a ) and len(__a ) > 0: _a : str = attribute elif hasattr(__a , __a ): _a : List[str] = attribute if attribute == "weight": assert old_model.weight.shape == model.weight.shape, "Shapes have to match!" _a : Dict = old_model.weight logger.info(f"""{attribute} is initialized.""" ) _a : List[Any] = True break elif attribute == "bias": assert old_model.bias.shape == model.bias.shape, "Shapes have to match!" _a : List[Any] = old_model.bias logger.info(f"""{attribute} is initialized""" ) _a : int = True break elif attribute in special_keys and hasattr(__a , 'in_proj_weight' ): _a : Any = old_model.in_proj_weight.shape[0] // 3 _a : Union[str, Any] = getattr(__a , __a ) param.weight.shape == old_model.in_proj_weight[:embed_dim, :].shape, "Shapes have to match" param.bias.shape == old_model.in_proj_bias[:embed_dim].shape, "Shapes have to match" if attribute == "query_proj": _a : Optional[Any] = nn.Parameter(old_model.in_proj_weight[:embed_dim, :] ) _a : Optional[int] = nn.Parameter(old_model.in_proj_bias[:embed_dim] ) elif attribute == "key_proj": _a : Tuple = nn.Parameter(old_model.in_proj_weight[embed_dim : 2 * embed_dim, :] ) _a : List[Any] = nn.Parameter(old_model.in_proj_bias[embed_dim : 2 * embed_dim] ) elif attribute == "value_proj": _a : Optional[int] = nn.Parameter(old_model.in_proj_weight[2 * embed_dim :, :] ) _a : Optional[int] = nn.Parameter(old_model.in_proj_bias[2 * embed_dim :] ) _a : Any = True break elif attribute == "position_embeddings": assert ( model.position_embeddings.weight.shape[-1] == old_model.embed_positions.weight.shape[-1] ), "Hidden size has to match" assert model.position_embeddings.weight.shape[0] == 5_1_2, "We want 512 position_embeddings." _a : List[str] = nn.Parameter(old_model.embed_positions.weight[:5_1_2, :] ) _a : List[Any] = True break if attribute.isdigit(): _a : str = model[int(__a )] _a : List[str] = old_model[int(__a )] else: _a : Optional[int] = getattr(__a , __a ) if old_attribute == "": _a : Optional[int] = old_model else: if not hasattr(__a , __a ): raise ValueError(f"""{old_model} does not have {old_attribute}""" ) _a : Union[str, Any] = getattr(__a , __a ) if not is_key_init: raise ValueError(f"""{key} was not correctly initialized!""" ) print(f"""Saving model to {pytorch_dump_folder_path}""" ) prophet.save_pretrained(__a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser() # Required parameters parser.add_argument( """--prophetnet_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.""" ) __lowerCAmelCase = parser.parse_args() convert_prophetnet_checkpoint_to_pytorch(args.prophetnet_checkpoint_path, args.pytorch_dump_folder_path)
271
'''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 __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = """▁""" __lowerCAmelCase = {"""vocab_file""": """sentencepiece.bpe.model""", """monolingual_vocab_file""": """dict.txt"""} __lowerCAmelCase = { """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""", }, } __lowerCAmelCase = {"""vinai/bartpho-syllable""": 1_0_2_4} class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[Any] = VOCAB_FILES_NAMES __UpperCAmelCase : Dict = PRETRAINED_VOCAB_FILES_MAP __UpperCAmelCase : Optional[int] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES __UpperCAmelCase : Dict = ['''input_ids''', '''attention_mask'''] def __init__( self : str ,_a : str ,_a : Any ,_a : Any="<s>" ,_a : Dict="</s>" ,_a : int="</s>" ,_a : Union[str, Any]="<s>" ,_a : List[Any]="<unk>" ,_a : Optional[Any]="<pad>" ,_a : List[str]="<mask>" ,_a : Optional[Dict[str, Any]] = None ,**_a : int ,): '''simple docstring''' _a : Any = AddedToken(_a ,lstrip=_a ,rstrip=_a ) if isinstance(_a ,_a ) else mask_token _a : Optional[Any] = {} if sp_model_kwargs is None else sp_model_kwargs super().__init__( bos_token=_a ,eos_token=_a ,unk_token=_a ,sep_token=_a ,cls_token=_a ,pad_token=_a ,mask_token=_a ,sp_model_kwargs=self.sp_model_kwargs ,**_a ,) _a : Optional[int] = vocab_file _a : Union[str, Any] = monolingual_vocab_file _a : List[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.Load(str(_a ) ) # Load the reduced vocab # Keep order of special tokens for backward compatibility _a : Union[str, Any] = {} _a : int = 0 for token in [bos_token, pad_token, eos_token, unk_token, sep_token, cls_token]: if str(_a ) not in self.fairseq_tokens_to_ids: _a : int = cnt cnt += 1 with open(_a ,'r' ,encoding='utf-8' ) as f: for line in f.readlines(): _a : str = line.strip().split()[0] _a : Tuple = len(self.fairseq_tokens_to_ids ) if str(_a ) not in self.fairseq_tokens_to_ids: _a : List[str] = len(self.fairseq_tokens_to_ids ) _a : Tuple = {v: k for k, v in self.fairseq_tokens_to_ids.items()} def __getstate__( self : Union[str, Any] ): '''simple docstring''' _a : int = self.__dict__.copy() _a : str = None _a : Optional[Any] = self.sp_model.serialized_model_proto() return state def __setstate__( self : Tuple ,_a : Tuple ): '''simple docstring''' _a : Tuple = d # for backward compatibility if not hasattr(self ,'sp_model_kwargs' ): _a : List[str] = {} _a : List[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.LoadFromSerializedProto(self.sp_model_proto ) def __lowercase ( self : Dict ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' if token_ids_a is None: return [self.cls_token_id] + token_ids_a + [self.sep_token_id] _a : Dict = [self.cls_token_id] _a : int = [self.sep_token_id] return cls + token_ids_a + sep + sep + token_ids_a + sep def __lowercase ( self : int ,_a : List[int] ,_a : Optional[List[int]] = None ,_a : bool = False ): '''simple docstring''' if already_has_special_tokens: return super().get_special_tokens_mask( token_ids_a=_a ,token_ids_a=_a ,already_has_special_tokens=_a ) if token_ids_a is None: return [1] + ([0] * len(_a )) + [1] return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1] def __lowercase ( self : Tuple ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' _a : List[str] = [self.sep_token_id] _a : List[str] = [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 __lowercase ( self : Dict ): '''simple docstring''' return len(self.fairseq_ids_to_tokens ) def __lowercase ( self : Dict ): '''simple docstring''' _a : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )} vocab.update(self.added_tokens_encoder ) return vocab def __lowercase ( self : Tuple ,_a : str ): '''simple docstring''' return self.sp_model.encode(_a ,out_type=_a ) def __lowercase ( self : Union[str, Any] ,_a : Union[str, Any] ): '''simple docstring''' if token in self.fairseq_tokens_to_ids: return self.fairseq_tokens_to_ids[token] else: return self.unk_token_id def __lowercase ( self : Any ,_a : int ): '''simple docstring''' return self.fairseq_ids_to_tokens[index] def __lowercase ( self : Tuple ,_a : Union[str, Any] ): '''simple docstring''' _a : str = ''.join(_a ).replace(_a ,' ' ).strip() return out_string def __lowercase ( self : Union[str, Any] ,_a : str ,_a : Optional[str] = None ): '''simple docstring''' if not os.path.isdir(_a ): logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" ) return _a : int = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) _a : int = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['monolingual_vocab_file'] ,) if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ): copyfile(self.vocab_file ,_a ) elif not os.path.isfile(self.vocab_file ): with open(_a ,'wb' ) as fi: _a : List[Any] = self.sp_model.serialized_model_proto() fi.write(_a ) if os.path.abspath(self.monolingual_vocab_file ) != os.path.abspath( _a ) and os.path.isfile(self.monolingual_vocab_file ): copyfile(self.monolingual_vocab_file ,_a ) elif not os.path.isfile(self.monolingual_vocab_file ): with open(_a ,'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(_a )} \n""" ) return out_vocab_file, out_monolingual_vocab_file
271
1
'''simple docstring''' from statistics import mean, stdev def UpperCAmelCase_ (__a : list , __a : int = 3 ): """simple docstring""" _a : Optional[Any] = min(__a ) _a : int = max(__a ) # normalize data return [round((x - x_min) / (x_max - x_min) , __a ) for x in data] def UpperCAmelCase_ (__a : list , __a : int = 3 ): """simple docstring""" _a : Dict = mean(__a ) _a : Union[str, Any] = stdev(__a ) # standardize data return [round((x - mu) / (sigma) , __a ) for x in data]
271
'''simple docstring''' import numpy as np from transformers import BatchFeature from transformers.testing_utils import require_tf, require_torch from .test_feature_extraction_common import FeatureExtractionSavingTestMixin class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Union[str, Any] = None __UpperCAmelCase : List[Any] = None @property def __lowercase ( self : Dict ): '''simple docstring''' return self.feat_extract_tester.prepare_feat_extract_dict() def __lowercase ( self : str ): '''simple docstring''' _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) self.assertTrue(hasattr(_a ,'feature_size' ) ) self.assertTrue(hasattr(_a ,'sampling_rate' ) ) self.assertTrue(hasattr(_a ,'padding_value' ) ) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.feat_extract_tester.prepare_inputs_for_common() _a : str = self.feature_extraction_class(**self.feat_extract_dict ) _a : int = feat_extract.model_input_names[0] _a : List[Any] = BatchFeature({input_name: speech_inputs} ) self.assertTrue(all(len(_a ) == len(_a ) for x, y in zip(_a ,processed_features[input_name] ) ) ) _a : Any = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Union[str, Any] = BatchFeature({input_name: speech_inputs} ,tensor_type='np' ) _a : Union[str, Any] = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : Optional[int] = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) @require_torch def __lowercase ( self : Any ): '''simple docstring''' _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : int = feat_extract.model_input_names[0] _a : str = BatchFeature({input_name: speech_inputs} ,tensor_type='pt' ) _a : str = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : str = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) @require_tf def __lowercase ( self : int ): '''simple docstring''' _a : int = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : Tuple = feat_extract.model_input_names[0] _a : int = BatchFeature({input_name: speech_inputs} ,tensor_type='tf' ) _a : Optional[int] = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : Optional[Any] = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) def __lowercase ( self : Dict ,_a : Any=False ): '''simple docstring''' def _inputs_have_equal_length(_a : Tuple ): _a : Tuple = len(input[0] ) for input_slice in input[1:]: if len(_a ) != length: return False return True def _inputs_are_equal(_a : Optional[Any] ,_a : Union[str, Any] ): if len(_a ) != len(_a ): return False for input_slice_a, input_slice_a in zip(_a ,_a ): if not np.allclose(np.asarray(_a ) ,np.asarray(_a ) ,atol=1E-3 ): return False return True _a : int = self.feature_extraction_class(**self.feat_extract_dict ) _a : Tuple = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a ) _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Tuple = BatchFeature({input_name: speech_inputs} ) _a : str = self.feat_extract_tester.seq_length_diff _a : Dict = self.feat_extract_tester.max_seq_length + pad_diff _a : Dict = self.feat_extract_tester.min_seq_length _a : Optional[Any] = self.feat_extract_tester.batch_size _a : Tuple = self.feat_extract_tester.feature_size # test padding for List[int] + numpy _a : int = feat_extract.pad(_a ,padding=_a ) _a : List[Any] = input_a[input_name] _a : Tuple = feat_extract.pad(_a ,padding='longest' ) _a : Any = input_a[input_name] _a : Optional[Any] = feat_extract.pad(_a ,padding='max_length' ,max_length=len(speech_inputs[-1] ) ) _a : List[str] = input_a[input_name] _a : List[str] = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' ) _a : str = input_a[input_name] # max_length parameter has to be provided when setting `padding="max_length"` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='max_length' )[input_name] _a : int = feat_extract.pad( _a ,padding='max_length' ,max_length=_a ,return_tensors='np' ) _a : Optional[int] = input_a[input_name] self.assertFalse(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) self.assertTrue(len(input_a[0] ) == pad_min_length ) self.assertTrue(len(input_a[1] ) == pad_min_length + pad_diff ) self.assertTrue(input_a.shape[:2] == (batch_size, len(input_a[0] )) ) self.assertTrue(input_a.shape[:2] == (batch_size, pad_max_length) ) if feature_size > 1: self.assertTrue(input_a.shape[2] == input_a.shape[2] == feature_size ) # test padding for `pad_to_multiple_of` for List[int] + numpy _a : Tuple = feat_extract.pad(_a ,pad_to_multiple_of=10 ) _a : List[str] = input_a[input_name] _a : str = feat_extract.pad(_a ,padding='longest' ,pad_to_multiple_of=10 ) _a : Tuple = input_a[input_name] _a : Optional[int] = feat_extract.pad( _a ,padding='max_length' ,pad_to_multiple_of=10 ,max_length=_a ) _a : Any = input_a[input_name] _a : Optional[int] = feat_extract.pad( _a ,padding='max_length' ,pad_to_multiple_of=10 ,max_length=_a ,return_tensors='np' ,) _a : Dict = input_a[input_name] self.assertTrue(all(len(_a ) % 10 == 0 for x in input_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) _a : List[str] = pad_max_length if pad_max_length % 10 == 0 else (pad_max_length // 10 + 1) * 10 self.assertTrue(all(len(_a ) == expected_mult_pad_length for x in input_a ) ) self.assertEqual(input_a.shape[:2] ,(batch_size, expected_mult_pad_length) ) if feature_size > 1: self.assertTrue(input_a.shape[2] == feature_size ) # Check padding value is correct _a : Any = (np.ones(self.feat_extract_tester.feature_size ) * feat_extract.padding_value).sum() self.assertTrue( abs(np.asarray(input_a[0] )[pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1E-3 ) self.assertTrue( abs( np.asarray(input_a[1] )[pad_min_length + pad_diff :].sum() - padding_vector_sum * (pad_max_length - pad_min_length - pad_diff) ) < 1E-3 ) self.assertTrue( abs( np.asarray(input_a[2] )[pad_min_length + 2 * pad_diff :].sum() - padding_vector_sum * (pad_max_length - pad_min_length - 2 * pad_diff) ) < 1E-3 ) self.assertTrue( abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1E-3 ) self.assertTrue( abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (expected_mult_pad_length - pad_min_length) ) < 1E-3 ) def __lowercase ( self : List[Any] ,_a : Optional[int]=False ): '''simple docstring''' def _inputs_have_equal_length(_a : List[str] ): _a : Union[str, Any] = len(input[0] ) for input_slice in input[1:]: if len(_a ) != length: return False return True def _inputs_are_equal(_a : List[str] ,_a : List[str] ): if len(_a ) != len(_a ): return False for input_slice_a, input_slice_a in zip(_a ,_a ): if not np.allclose(np.asarray(_a ) ,np.asarray(_a ) ,atol=1E-3 ): return False return True _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : str = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a ) _a : Any = feat_extract.model_input_names[0] _a : List[Any] = BatchFeature({input_name: speech_inputs} ) # truncate to smallest _a : Union[str, Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,truncation=_a ) _a : str = input_a[input_name] _a : List[str] = feat_extract.pad(_a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ) _a : Tuple = input_a[input_name] self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertFalse(_inputs_have_equal_length(_a ) ) # truncate to smallest with np _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,return_tensors='np' ,truncation=_a ,) _a : Any = input_a[input_name] _a : List[Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,return_tensors='np' ) _a : int = input_a[input_name] self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(input_a.shape[1] == len(speech_inputs[0] ) ) # since truncation forces padding to be smaller than longest input # function can't return `np.ndarray`, but has to return list self.assertFalse(_inputs_have_equal_length(_a ) ) # truncate to middle _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,truncation=_a ,return_tensors='np' ,) _a : List[Any] = input_a[input_name] _a : Tuple = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,truncation=_a ) _a : Tuple = input_a[input_name] _a : Tuple = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,return_tensors='np' ) _a : Dict = input_a[input_name] self.assertTrue(input_a.shape[1] == len(speech_inputs[1] ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) # since truncation forces padding to be smaller than longest input # function can't return `np.ndarray`, but has to return list self.assertFalse(_inputs_have_equal_length(_a ) ) self.assertTrue(len(input_a[-1] ) == len(speech_inputs[-1] ) ) # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,truncation=_a )[input_name] # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='longest' ,truncation=_a )[input_name] # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='longest' ,truncation=_a )[input_name] # max_length parameter has to be provided when setting `truncation=True` and padding="max_length" with self.assertRaises(_a ): feat_extract.pad(_a ,padding='max_length' ,truncation=_a )[input_name] # test truncation for `pad_to_multiple_of` for List[int] + numpy _a : Optional[Any] = 12 _a : List[Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,pad_to_multiple_of=_a ,truncation=_a ,) _a : Tuple = input_a[input_name] _a : str = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,pad_to_multiple_of=_a ,) _a : List[Any] = input_a[input_name] # retrieve expected_length as multiple of pad_to_multiple_of _a : List[Any] = len(speech_inputs[0] ) if expected_length % pad_to_multiple_of != 0: _a : Union[str, Any] = ((len(speech_inputs[0] ) // pad_to_multiple_of) + 1) * pad_to_multiple_of self.assertTrue(len(input_a[0] ) == expected_length ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertFalse(_inputs_have_equal_length(_a ) ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' self._check_padding(numpify=_a ) def __lowercase ( self : Tuple ): '''simple docstring''' self._check_padding(numpify=_a ) def __lowercase ( self : Dict ): '''simple docstring''' self._check_truncation(numpify=_a ) def __lowercase ( self : str ): '''simple docstring''' self._check_truncation(numpify=_a ) @require_torch def __lowercase ( self : Dict ): '''simple docstring''' _a : Any = self.feature_extraction_class(**self.feat_extract_dict ) _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Optional[int] = BatchFeature({input_name: speech_inputs} ) _a : List[Any] = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' )[input_name] _a : List[str] = feat_extract.pad(_a ,padding='longest' ,return_tensors='pt' )[input_name] self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_pt.numpy().astype(np.floataa ).sum() ) < 1E-2 ) @require_tf def __lowercase ( self : int ): '''simple docstring''' _a : List[str] = self.feature_extraction_class(**self.feat_extract_dict ) _a : Optional[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Dict = feat_extract.model_input_names[0] _a : Optional[Any] = BatchFeature({input_name: speech_inputs} ) _a : Dict = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' )[input_name] _a : Any = feat_extract.pad(_a ,padding='longest' ,return_tensors='tf' )[input_name] self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_tf.numpy().astype(np.floataa ).sum() ) < 1E-2 ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : str = self.feat_extract_dict _a : List[Any] = True _a : Optional[int] = self.feature_extraction_class(**_a ) _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Tuple = [len(_a ) for x in speech_inputs] _a : int = feat_extract.model_input_names[0] _a : Optional[Any] = BatchFeature({input_name: speech_inputs} ) _a : str = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' ) self.assertIn('attention_mask' ,_a ) self.assertListEqual(list(processed.attention_mask.shape ) ,list(processed[input_name].shape[:2] ) ) self.assertListEqual(processed.attention_mask.sum(-1 ).tolist() ,_a ) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.feat_extract_dict _a : Tuple = True _a : Optional[int] = self.feature_extraction_class(**_a ) _a : Dict = self.feat_extract_tester.prepare_inputs_for_common() _a : Dict = [len(_a ) for x in speech_inputs] _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Any = BatchFeature({input_name: speech_inputs} ) _a : List[Any] = min(_a ) _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=_a ,truncation=_a ,return_tensors='np' ) self.assertIn('attention_mask' ,_a ) self.assertListEqual( list(processed_pad.attention_mask.shape ) ,[processed_pad[input_name].shape[0], max_length] ) self.assertListEqual( processed_pad.attention_mask[:, :max_length].sum(-1 ).tolist() ,[max_length for x in speech_inputs] )
271
1
'''simple docstring''' import importlib import inspect import json import os import re import shutil import sys from pathlib import Path from typing import Dict, Optional, Union from urllib import request from huggingface_hub import HfFolder, cached_download, hf_hub_download, model_info from packaging import version from .. import __version__ from . import DIFFUSERS_DYNAMIC_MODULE_NAME, HF_MODULES_CACHE, logging __lowerCAmelCase = ( """https://raw.githubusercontent.com/huggingface/diffusers/{revision}/examples/community/{pipeline}.py""" ) __lowerCAmelCase = logging.get_logger(__name__) # pylint: disable=invalid-name def UpperCAmelCase_ (): """simple docstring""" _a : Optional[int] = 'https://pypi.org/pypi/diffusers/json' _a : int = json.loads(request.urlopen(__a ).read() )['releases'].keys() return sorted(__a , key=lambda __a : version.Version(__a ) ) def UpperCAmelCase_ (): """simple docstring""" if HF_MODULES_CACHE in sys.path: return sys.path.append(__a ) os.makedirs(__a , exist_ok=__a ) _a : str = Path(__a ) / '__init__.py' if not init_path.exists(): init_path.touch() def UpperCAmelCase_ (__a : Union[str, os.PathLike] ): """simple docstring""" init_hf_modules() _a : Dict = Path(__a ) / name # If the parent module does not exist yet, recursively create it. if not dynamic_module_path.parent.exists(): create_dynamic_module(dynamic_module_path.parent ) os.makedirs(__a , exist_ok=__a ) _a : Optional[int] = dynamic_module_path / '__init__.py' if not init_path.exists(): init_path.touch() def UpperCAmelCase_ (__a : str ): """simple docstring""" with open(__a , 'r' , encoding='utf-8' ) as f: _a : int = f.read() # Imports of the form `import .xxx` _a : Tuple = re.findall('^\s*import\s+\.(\S+)\s*$' , __a , flags=re.MULTILINE ) # Imports of the form `from .xxx import yyy` relative_imports += re.findall('^\s*from\s+\.(\S+)\s+import' , __a , flags=re.MULTILINE ) # Unique-ify return list(set(__a ) ) def UpperCAmelCase_ (__a : Any ): """simple docstring""" _a : Optional[int] = False _a : Optional[int] = [module_file] _a : List[str] = [] # Let's recurse through all relative imports while not no_change: _a : str = [] for f in files_to_check: new_imports.extend(get_relative_imports(__a ) ) _a : Union[str, Any] = Path(__a ).parent _a : str = [str(module_path / m ) for m in new_imports] _a : Tuple = [f for f in new_import_files if f not in all_relative_imports] _a : Dict = [f"""{f}.py""" for f in new_import_files] _a : List[str] = len(__a ) == 0 all_relative_imports.extend(__a ) return all_relative_imports def UpperCAmelCase_ (__a : Tuple ): """simple docstring""" with open(__a , 'r' , encoding='utf-8' ) as f: _a : Dict = f.read() # Imports of the form `import xxx` _a : Optional[int] = re.findall('^\s*import\s+(\S+)\s*$' , __a , flags=re.MULTILINE ) # Imports of the form `from xxx import yyy` imports += re.findall('^\s*from\s+(\S+)\s+import' , __a , flags=re.MULTILINE ) # Only keep the top-level module _a : List[str] = [imp.split('.' )[0] for imp in imports if not imp.startswith('.' )] # Unique-ify and test we got them all _a : Optional[int] = list(set(__a ) ) _a : List[str] = [] for imp in imports: try: importlib.import_module(__a ) except ImportError: missing_packages.append(__a ) if len(__a ) > 0: raise ImportError( 'This modeling file requires the following packages that were not found in your environment: ' f"""{', '.join(__a )}. Run `pip install {' '.join(__a )}`""" ) return get_relative_imports(__a ) def UpperCAmelCase_ (__a : Any , __a : str ): """simple docstring""" _a : Any = module_path.replace(os.path.sep , '.' ) _a : Union[str, Any] = importlib.import_module(__a ) if class_name is None: return find_pipeline_class(__a ) return getattr(__a , __a ) def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" from ..pipelines import DiffusionPipeline _a : List[str] = dict(inspect.getmembers(__a , inspect.isclass ) ) _a : str = None for cls_name, cls in cls_members.items(): if ( cls_name != DiffusionPipeline.__name__ and issubclass(cls , __a ) and cls.__module__.split('.' )[0] != "diffusers" ): if pipeline_class is not None: raise ValueError( f"""Multiple classes that inherit from {DiffusionPipeline.__name__} have been found:""" f""" {pipeline_class.__name__}, and {cls_name}. Please make sure to define only one in""" f""" {loaded_module}.""" ) _a : Any = cls return pipeline_class def UpperCAmelCase_ (__a : Union[str, os.PathLike] , __a : str , __a : Optional[Union[str, os.PathLike]] = None , __a : bool = False , __a : bool = False , __a : Optional[Dict[str, str]] = None , __a : Optional[Union[bool, str]] = None , __a : Optional[str] = None , __a : bool = False , ): """simple docstring""" _a : str = str(__a ) _a : Optional[Any] = os.path.join(__a , __a ) if os.path.isfile(__a ): _a : Tuple = module_file_or_url _a : Optional[Any] = 'local' elif pretrained_model_name_or_path.count('/' ) == 0: _a : int = get_diffusers_versions() # cut ".dev0" _a : Any = 'v' + '.'.join(__version__.split('.' )[:3] ) # retrieve github version that matches if revision is None: _a : Any = latest_version if latest_version[1:] in available_versions else 'main' logger.info(f"""Defaulting to latest_version: {revision}.""" ) elif revision in available_versions: _a : Any = f"""v{revision}""" elif revision == "main": _a : Optional[int] = revision else: raise ValueError( f"""`custom_revision`: {revision} does not exist. Please make sure to choose one of""" f""" {', '.join(available_versions + ['main'] )}.""" ) # community pipeline on GitHub _a : Tuple = COMMUNITY_PIPELINES_URL.format(revision=__a , pipeline=__a ) try: _a : Any = cached_download( __a , cache_dir=__a , force_download=__a , proxies=__a , resume_download=__a , local_files_only=__a , use_auth_token=__a , ) _a : List[Any] = 'git' _a : Any = pretrained_model_name_or_path + '.py' except EnvironmentError: logger.error(f"""Could not locate the {module_file} inside {pretrained_model_name_or_path}.""" ) raise else: try: # Load from URL or cache if already cached _a : Optional[Any] = hf_hub_download( __a , __a , cache_dir=__a , force_download=__a , proxies=__a , resume_download=__a , local_files_only=__a , use_auth_token=__a , ) _a : List[Any] = os.path.join('local' , '--'.join(pretrained_model_name_or_path.split('/' ) ) ) except EnvironmentError: logger.error(f"""Could not locate the {module_file} inside {pretrained_model_name_or_path}.""" ) raise # Check we have all the requirements in our environment _a : Optional[int] = check_imports(__a ) # Now we move the module inside our cached dynamic modules. _a : Optional[Any] = DIFFUSERS_DYNAMIC_MODULE_NAME + os.path.sep + submodule create_dynamic_module(__a ) _a : Any = Path(__a ) / full_submodule if submodule == "local" or submodule == "git": # We always copy local files (we could hash the file to see if there was a change, and give them the name of # that hash, to only copy when there is a modification but it seems overkill for now). # The only reason we do the copy is to avoid putting too many folders in sys.path. shutil.copy(__a , submodule_path / module_file ) for module_needed in modules_needed: _a : Dict = f"""{module_needed}.py""" shutil.copy(os.path.join(__a , __a ) , submodule_path / module_needed ) else: # Get the commit hash # TODO: we will get this info in the etag soon, so retrieve it from there and not here. if isinstance(__a , __a ): _a : Optional[Any] = use_auth_token elif use_auth_token is True: _a : List[Any] = HfFolder.get_token() else: _a : Dict = None _a : int = model_info(__a , revision=__a , token=__a ).sha # The module file will end up being placed in a subfolder with the git hash of the repo. This way we get the # benefit of versioning. _a : Optional[int] = submodule_path / commit_hash _a : str = full_submodule + os.path.sep + commit_hash create_dynamic_module(__a ) if not (submodule_path / module_file).exists(): shutil.copy(__a , submodule_path / module_file ) # Make sure we also have every file with relative for module_needed in modules_needed: if not (submodule_path / module_needed).exists(): get_cached_module_file( __a , f"""{module_needed}.py""" , cache_dir=__a , force_download=__a , resume_download=__a , proxies=__a , use_auth_token=__a , revision=__a , local_files_only=__a , ) return os.path.join(__a , __a ) def UpperCAmelCase_ (__a : Union[str, os.PathLike] , __a : str , __a : Optional[str] = None , __a : Optional[Union[str, os.PathLike]] = None , __a : bool = False , __a : bool = False , __a : Optional[Dict[str, str]] = None , __a : Optional[Union[bool, str]] = None , __a : Optional[str] = None , __a : bool = False , **__a : str , ): """simple docstring""" _a : Dict = get_cached_module_file( __a , __a , cache_dir=__a , force_download=__a , resume_download=__a , proxies=__a , use_auth_token=__a , revision=__a , local_files_only=__a , ) return get_class_in_module(__a , final_module.replace('.py' , '' ) )
271
'''simple docstring''' from typing import List, Optional, Tuple, Union import torch from ...models import UNetaDModel from ...schedulers import KarrasVeScheduler from ...utils import randn_tensor from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : UNetaDModel __UpperCAmelCase : KarrasVeScheduler def __init__( self : Union[str, Any] ,_a : UNetaDModel ,_a : KarrasVeScheduler ): '''simple docstring''' super().__init__() self.register_modules(unet=_a ,scheduler=_a ) @torch.no_grad() def __call__( self : List[Any] ,_a : int = 1 ,_a : int = 50 ,_a : Optional[Union[torch.Generator, List[torch.Generator]]] = None ,_a : Optional[str] = "pil" ,_a : bool = True ,**_a : List[Any] ,): '''simple docstring''' _a : Any = self.unet.config.sample_size _a : Optional[int] = (batch_size, 3, img_size, img_size) _a : Dict = self.unet # sample x_0 ~ N(0, sigma_0^2 * I) _a : Dict = randn_tensor(_a ,generator=_a ,device=self.device ) * self.scheduler.init_noise_sigma self.scheduler.set_timesteps(_a ) for t in self.progress_bar(self.scheduler.timesteps ): # here sigma_t == t_i from the paper _a : Optional[int] = self.scheduler.schedule[t] _a : List[str] = self.scheduler.schedule[t - 1] if t > 0 else 0 # 1. Select temporarily increased noise level sigma_hat # 2. Add new noise to move from sample_i to sample_hat _a, _a : List[Any] = self.scheduler.add_noise_to_input(_a ,_a ,generator=_a ) # 3. Predict the noise residual given the noise magnitude `sigma_hat` # The model inputs and output are adjusted by following eq. (213) in [1]. _a : Optional[int] = (sigma_hat / 2) * model((sample_hat + 1) / 2 ,sigma_hat / 2 ).sample # 4. Evaluate dx/dt at sigma_hat # 5. Take Euler step from sigma to sigma_prev _a : Tuple = self.scheduler.step(_a ,_a ,_a ,_a ) if sigma_prev != 0: # 6. Apply 2nd order correction # The model inputs and output are adjusted by following eq. (213) in [1]. _a : Optional[int] = (sigma_prev / 2) * model((step_output.prev_sample + 1) / 2 ,sigma_prev / 2 ).sample _a : Optional[Any] = self.scheduler.step_correct( _a ,_a ,_a ,_a ,step_output.prev_sample ,step_output['derivative'] ,) _a : Dict = step_output.prev_sample _a : Tuple = (sample / 2 + 0.5).clamp(0 ,1 ) _a : Optional[Any] = sample.cpu().permute(0 ,2 ,3 ,1 ).numpy() if output_type == "pil": _a : List[str] = self.numpy_to_pil(_a ) if not return_dict: return (image,) return ImagePipelineOutput(images=_a )
271
1
'''simple docstring''' import json from typing import List, Optional, Tuple from tokenizers import normalizers from ...tokenization_utils_fast import PreTrainedTokenizerFast from .tokenization_electra import ElectraTokenizer __lowerCAmelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""} __lowerCAmelCase = { """vocab_file""": { """google/electra-small-generator""": ( """https://huggingface.co/google/electra-small-generator/resolve/main/vocab.txt""" ), """google/electra-base-generator""": """https://huggingface.co/google/electra-base-generator/resolve/main/vocab.txt""", """google/electra-large-generator""": ( """https://huggingface.co/google/electra-large-generator/resolve/main/vocab.txt""" ), """google/electra-small-discriminator""": ( """https://huggingface.co/google/electra-small-discriminator/resolve/main/vocab.txt""" ), """google/electra-base-discriminator""": ( """https://huggingface.co/google/electra-base-discriminator/resolve/main/vocab.txt""" ), """google/electra-large-discriminator""": ( """https://huggingface.co/google/electra-large-discriminator/resolve/main/vocab.txt""" ), }, """tokenizer_file""": { """google/electra-small-generator""": ( """https://huggingface.co/google/electra-small-generator/resolve/main/tokenizer.json""" ), """google/electra-base-generator""": ( """https://huggingface.co/google/electra-base-generator/resolve/main/tokenizer.json""" ), """google/electra-large-generator""": ( """https://huggingface.co/google/electra-large-generator/resolve/main/tokenizer.json""" ), """google/electra-small-discriminator""": ( """https://huggingface.co/google/electra-small-discriminator/resolve/main/tokenizer.json""" ), """google/electra-base-discriminator""": ( """https://huggingface.co/google/electra-base-discriminator/resolve/main/tokenizer.json""" ), """google/electra-large-discriminator""": ( """https://huggingface.co/google/electra-large-discriminator/resolve/main/tokenizer.json""" ), }, } __lowerCAmelCase = { """google/electra-small-generator""": 5_1_2, """google/electra-base-generator""": 5_1_2, """google/electra-large-generator""": 5_1_2, """google/electra-small-discriminator""": 5_1_2, """google/electra-base-discriminator""": 5_1_2, """google/electra-large-discriminator""": 5_1_2, } __lowerCAmelCase = { """google/electra-small-generator""": {"""do_lower_case""": True}, """google/electra-base-generator""": {"""do_lower_case""": True}, """google/electra-large-generator""": {"""do_lower_case""": True}, """google/electra-small-discriminator""": {"""do_lower_case""": True}, """google/electra-base-discriminator""": {"""do_lower_case""": True}, """google/electra-large-discriminator""": {"""do_lower_case""": True}, } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[int] = VOCAB_FILES_NAMES __UpperCAmelCase : Optional[Any] = PRETRAINED_VOCAB_FILES_MAP __UpperCAmelCase : List[str] = PRETRAINED_INIT_CONFIGURATION __UpperCAmelCase : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES __UpperCAmelCase : Optional[int] = ElectraTokenizer def __init__( self : Optional[Any] ,_a : str=None ,_a : List[Any]=None ,_a : List[str]=True ,_a : Tuple="[UNK]" ,_a : List[Any]="[SEP]" ,_a : List[Any]="[PAD]" ,_a : int="[CLS]" ,_a : List[str]="[MASK]" ,_a : List[Any]=True ,_a : List[str]=None ,**_a : int ,): '''simple docstring''' super().__init__( _a ,tokenizer_file=_a ,do_lower_case=_a ,unk_token=_a ,sep_token=_a ,pad_token=_a ,cls_token=_a ,mask_token=_a ,tokenize_chinese_chars=_a ,strip_accents=_a ,**_a ,) _a : Union[str, Any] = json.loads(self.backend_tokenizer.normalizer.__getstate__() ) if ( normalizer_state.get('lowercase' ,_a ) != do_lower_case or normalizer_state.get('strip_accents' ,_a ) != strip_accents or normalizer_state.get('handle_chinese_chars' ,_a ) != tokenize_chinese_chars ): _a : Tuple = getattr(_a ,normalizer_state.pop('type' ) ) _a : Tuple = do_lower_case _a : int = strip_accents _a : Dict = tokenize_chinese_chars _a : List[Any] = normalizer_class(**_a ) _a : int = do_lower_case def __lowercase ( self : Optional[int] ,_a : str ,_a : str=None ): '''simple docstring''' _a : str = [self.cls_token_id] + token_ids_a + [self.sep_token_id] if token_ids_a: output += token_ids_a + [self.sep_token_id] return output def __lowercase ( self : List[str] ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' _a : List[str] = [self.sep_token_id] _a : Union[str, Any] = [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 ) * [0] + len(token_ids_a + sep ) * [1] def __lowercase ( self : str ,_a : str ,_a : Optional[str] = None ): '''simple docstring''' _a : str = self._tokenizer.model.save(_a ,name=_a ) return tuple(_a )
271
'''simple docstring''' import importlib import inspect import json import os import re import shutil import sys from pathlib import Path from typing import Dict, Optional, Union from urllib import request from huggingface_hub import HfFolder, cached_download, hf_hub_download, model_info from packaging import version from .. import __version__ from . import DIFFUSERS_DYNAMIC_MODULE_NAME, HF_MODULES_CACHE, logging __lowerCAmelCase = ( """https://raw.githubusercontent.com/huggingface/diffusers/{revision}/examples/community/{pipeline}.py""" ) __lowerCAmelCase = logging.get_logger(__name__) # pylint: disable=invalid-name def UpperCAmelCase_ (): """simple docstring""" _a : Optional[int] = 'https://pypi.org/pypi/diffusers/json' _a : int = json.loads(request.urlopen(__a ).read() )['releases'].keys() return sorted(__a , key=lambda __a : version.Version(__a ) ) def UpperCAmelCase_ (): """simple docstring""" if HF_MODULES_CACHE in sys.path: return sys.path.append(__a ) os.makedirs(__a , exist_ok=__a ) _a : str = Path(__a ) / '__init__.py' if not init_path.exists(): init_path.touch() def UpperCAmelCase_ (__a : Union[str, os.PathLike] ): """simple docstring""" init_hf_modules() _a : Dict = Path(__a ) / name # If the parent module does not exist yet, recursively create it. if not dynamic_module_path.parent.exists(): create_dynamic_module(dynamic_module_path.parent ) os.makedirs(__a , exist_ok=__a ) _a : Optional[int] = dynamic_module_path / '__init__.py' if not init_path.exists(): init_path.touch() def UpperCAmelCase_ (__a : str ): """simple docstring""" with open(__a , 'r' , encoding='utf-8' ) as f: _a : int = f.read() # Imports of the form `import .xxx` _a : Tuple = re.findall('^\s*import\s+\.(\S+)\s*$' , __a , flags=re.MULTILINE ) # Imports of the form `from .xxx import yyy` relative_imports += re.findall('^\s*from\s+\.(\S+)\s+import' , __a , flags=re.MULTILINE ) # Unique-ify return list(set(__a ) ) def UpperCAmelCase_ (__a : Any ): """simple docstring""" _a : Optional[int] = False _a : Optional[int] = [module_file] _a : List[str] = [] # Let's recurse through all relative imports while not no_change: _a : str = [] for f in files_to_check: new_imports.extend(get_relative_imports(__a ) ) _a : Union[str, Any] = Path(__a ).parent _a : str = [str(module_path / m ) for m in new_imports] _a : Tuple = [f for f in new_import_files if f not in all_relative_imports] _a : Dict = [f"""{f}.py""" for f in new_import_files] _a : List[str] = len(__a ) == 0 all_relative_imports.extend(__a ) return all_relative_imports def UpperCAmelCase_ (__a : Tuple ): """simple docstring""" with open(__a , 'r' , encoding='utf-8' ) as f: _a : Dict = f.read() # Imports of the form `import xxx` _a : Optional[int] = re.findall('^\s*import\s+(\S+)\s*$' , __a , flags=re.MULTILINE ) # Imports of the form `from xxx import yyy` imports += re.findall('^\s*from\s+(\S+)\s+import' , __a , flags=re.MULTILINE ) # Only keep the top-level module _a : List[str] = [imp.split('.' )[0] for imp in imports if not imp.startswith('.' )] # Unique-ify and test we got them all _a : Optional[int] = list(set(__a ) ) _a : List[str] = [] for imp in imports: try: importlib.import_module(__a ) except ImportError: missing_packages.append(__a ) if len(__a ) > 0: raise ImportError( 'This modeling file requires the following packages that were not found in your environment: ' f"""{', '.join(__a )}. Run `pip install {' '.join(__a )}`""" ) return get_relative_imports(__a ) def UpperCAmelCase_ (__a : Any , __a : str ): """simple docstring""" _a : Any = module_path.replace(os.path.sep , '.' ) _a : Union[str, Any] = importlib.import_module(__a ) if class_name is None: return find_pipeline_class(__a ) return getattr(__a , __a ) def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" from ..pipelines import DiffusionPipeline _a : List[str] = dict(inspect.getmembers(__a , inspect.isclass ) ) _a : str = None for cls_name, cls in cls_members.items(): if ( cls_name != DiffusionPipeline.__name__ and issubclass(cls , __a ) and cls.__module__.split('.' )[0] != "diffusers" ): if pipeline_class is not None: raise ValueError( f"""Multiple classes that inherit from {DiffusionPipeline.__name__} have been found:""" f""" {pipeline_class.__name__}, and {cls_name}. Please make sure to define only one in""" f""" {loaded_module}.""" ) _a : Any = cls return pipeline_class def UpperCAmelCase_ (__a : Union[str, os.PathLike] , __a : str , __a : Optional[Union[str, os.PathLike]] = None , __a : bool = False , __a : bool = False , __a : Optional[Dict[str, str]] = None , __a : Optional[Union[bool, str]] = None , __a : Optional[str] = None , __a : bool = False , ): """simple docstring""" _a : str = str(__a ) _a : Optional[Any] = os.path.join(__a , __a ) if os.path.isfile(__a ): _a : Tuple = module_file_or_url _a : Optional[Any] = 'local' elif pretrained_model_name_or_path.count('/' ) == 0: _a : int = get_diffusers_versions() # cut ".dev0" _a : Any = 'v' + '.'.join(__version__.split('.' )[:3] ) # retrieve github version that matches if revision is None: _a : Any = latest_version if latest_version[1:] in available_versions else 'main' logger.info(f"""Defaulting to latest_version: {revision}.""" ) elif revision in available_versions: _a : Any = f"""v{revision}""" elif revision == "main": _a : Optional[int] = revision else: raise ValueError( f"""`custom_revision`: {revision} does not exist. Please make sure to choose one of""" f""" {', '.join(available_versions + ['main'] )}.""" ) # community pipeline on GitHub _a : Tuple = COMMUNITY_PIPELINES_URL.format(revision=__a , pipeline=__a ) try: _a : Any = cached_download( __a , cache_dir=__a , force_download=__a , proxies=__a , resume_download=__a , local_files_only=__a , use_auth_token=__a , ) _a : List[Any] = 'git' _a : Any = pretrained_model_name_or_path + '.py' except EnvironmentError: logger.error(f"""Could not locate the {module_file} inside {pretrained_model_name_or_path}.""" ) raise else: try: # Load from URL or cache if already cached _a : Optional[Any] = hf_hub_download( __a , __a , cache_dir=__a , force_download=__a , proxies=__a , resume_download=__a , local_files_only=__a , use_auth_token=__a , ) _a : List[Any] = os.path.join('local' , '--'.join(pretrained_model_name_or_path.split('/' ) ) ) except EnvironmentError: logger.error(f"""Could not locate the {module_file} inside {pretrained_model_name_or_path}.""" ) raise # Check we have all the requirements in our environment _a : Optional[int] = check_imports(__a ) # Now we move the module inside our cached dynamic modules. _a : Optional[Any] = DIFFUSERS_DYNAMIC_MODULE_NAME + os.path.sep + submodule create_dynamic_module(__a ) _a : Any = Path(__a ) / full_submodule if submodule == "local" or submodule == "git": # We always copy local files (we could hash the file to see if there was a change, and give them the name of # that hash, to only copy when there is a modification but it seems overkill for now). # The only reason we do the copy is to avoid putting too many folders in sys.path. shutil.copy(__a , submodule_path / module_file ) for module_needed in modules_needed: _a : Dict = f"""{module_needed}.py""" shutil.copy(os.path.join(__a , __a ) , submodule_path / module_needed ) else: # Get the commit hash # TODO: we will get this info in the etag soon, so retrieve it from there and not here. if isinstance(__a , __a ): _a : Optional[Any] = use_auth_token elif use_auth_token is True: _a : List[Any] = HfFolder.get_token() else: _a : Dict = None _a : int = model_info(__a , revision=__a , token=__a ).sha # The module file will end up being placed in a subfolder with the git hash of the repo. This way we get the # benefit of versioning. _a : Optional[int] = submodule_path / commit_hash _a : str = full_submodule + os.path.sep + commit_hash create_dynamic_module(__a ) if not (submodule_path / module_file).exists(): shutil.copy(__a , submodule_path / module_file ) # Make sure we also have every file with relative for module_needed in modules_needed: if not (submodule_path / module_needed).exists(): get_cached_module_file( __a , f"""{module_needed}.py""" , cache_dir=__a , force_download=__a , resume_download=__a , proxies=__a , use_auth_token=__a , revision=__a , local_files_only=__a , ) return os.path.join(__a , __a ) def UpperCAmelCase_ (__a : Union[str, os.PathLike] , __a : str , __a : Optional[str] = None , __a : Optional[Union[str, os.PathLike]] = None , __a : bool = False , __a : bool = False , __a : Optional[Dict[str, str]] = None , __a : Optional[Union[bool, str]] = None , __a : Optional[str] = None , __a : bool = False , **__a : str , ): """simple docstring""" _a : Dict = get_cached_module_file( __a , __a , cache_dir=__a , force_download=__a , resume_download=__a , proxies=__a , use_auth_token=__a , revision=__a , local_files_only=__a , ) return get_class_in_module(__a , final_module.replace('.py' , '' ) )
271
1
'''simple docstring''' __lowerCAmelCase = 8.3_144_598 def UpperCAmelCase_ (__a : float , __a : float ): """simple docstring""" if temperature < 0: raise Exception('Temperature cannot be less than 0 K' ) if molar_mass <= 0: raise Exception('Molar mass cannot be less than or equal to 0 kg/mol' ) else: return (3 * UNIVERSAL_GAS_CONSTANT * temperature / molar_mass) ** 0.5 if __name__ == "__main__": import doctest # run doctest doctest.testmod() # example __lowerCAmelCase = 3_0_0 __lowerCAmelCase = 2_8 __lowerCAmelCase = rms_speed_of_molecule(temperature, molar_mass) print(f'''Vrms of Nitrogen gas at 300 K is {vrms} m/s''')
271
'''simple docstring''' def UpperCAmelCase_ (__a : list , __a : list , __a : int ): """simple docstring""" _a : Optional[Any] = len(__a ) _a : int = [[0] * n for i in range(__a )] for i in range(__a ): _a : Tuple = y_points[i] for i in range(2 , __a ): for j in range(__a , __a ): _a : Tuple = ( (xa - x_points[j - i + 1]) * q[j][i - 1] - (xa - x_points[j]) * q[j - 1][i - 1] ) / (x_points[j] - x_points[j - i + 1]) return [q[n - 1][n - 1], q] if __name__ == "__main__": import doctest doctest.testmod()
271
1
'''simple docstring''' from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar __lowerCAmelCase = TypeVar("""T""") class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Tuple ,_a : T ): '''simple docstring''' _a : List[str] = data _a : Node[T] | None = None def __str__( self : Dict ): '''simple docstring''' return F"""{self.data}""" class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Optional[int] ): '''simple docstring''' _a : Node[T] | None = None def __iter__( self : str ): '''simple docstring''' _a : Tuple = self.top while node: yield node.data _a : int = node.next def __str__( self : str ): '''simple docstring''' return "->".join([str(_a ) for item in self] ) def __len__( self : Optional[Any] ): '''simple docstring''' return len(tuple(iter(self ) ) ) def __lowercase ( self : str ): '''simple docstring''' return self.top is None def __lowercase ( self : List[Any] ,_a : T ): '''simple docstring''' _a : int = Node(_a ) if not self.is_empty(): _a : Optional[Any] = self.top _a : List[str] = node def __lowercase ( self : Tuple ): '''simple docstring''' if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top ,_a ) _a : List[Any] = self.top _a : int = self.top.next return pop_node.data def __lowercase ( self : List[str] ): '''simple docstring''' if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = None if __name__ == "__main__": from doctest import testmod testmod()
271
'''simple docstring''' import inspect import os import unittest from pathlib import Path import torch import accelerate from accelerate.test_utils import execute_subprocess_async from accelerate.test_utils.testing import run_command class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[Any] = inspect.getfile(accelerate.test_utils ) __UpperCAmelCase : List[str] = os.path.sep.join(mod_file.split(os.path.sep )[:-1] + ['''scripts''', '''test_cli.py'''] ) __UpperCAmelCase : Dict = ['''accelerate''', '''launch'''] __UpperCAmelCase : Dict = Path.home() / '''.cache/huggingface/accelerate''' __UpperCAmelCase : Dict = '''default_config.yaml''' __UpperCAmelCase : Optional[Any] = config_folder / config_file __UpperCAmelCase : Dict = config_folder / '''_default_config.yaml''' __UpperCAmelCase : Any = Path('''tests/test_configs''' ) @classmethod def __lowercase ( cls : int ): '''simple docstring''' if cls.config_path.is_file(): cls.config_path.rename(cls.changed_path ) @classmethod def __lowercase ( cls : List[Any] ): '''simple docstring''' if cls.changed_path.is_file(): cls.changed_path.rename(cls.config_path ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Dict = self.base_cmd if torch.cuda.is_available() and (torch.cuda.device_count() > 1): cmd += ["--multi_gpu"] execute_subprocess_async(cmd + [self.test_file_path] ,env=os.environ.copy() ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' for config in sorted(self.test_config_path.glob('**/*.yaml' ) ): with self.subTest(config_file=_a ): execute_subprocess_async( self.base_cmd + ['--config_file', str(_a ), self.test_file_path] ,env=os.environ.copy() ) def __lowercase ( self : Optional[int] ): '''simple docstring''' execute_subprocess_async(['accelerate', 'test'] ,env=os.environ.copy() ) class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[Any] = '''test-tpu''' __UpperCAmelCase : Any = '''us-central1-a''' __UpperCAmelCase : List[Any] = '''ls''' __UpperCAmelCase : Any = ['''accelerate''', '''tpu-config'''] __UpperCAmelCase : Dict = '''cd /usr/share''' __UpperCAmelCase : Any = '''tests/test_samples/test_command_file.sh''' __UpperCAmelCase : List[Any] = '''Running gcloud compute tpus tpu-vm ssh''' def __lowercase ( self : Dict ): '''simple docstring''' _a : Optional[Any] = run_command( self.cmd + ['--command', self.command, '--tpu_zone', self.tpu_zone, '--tpu_name', self.tpu_name, '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls --worker all""" ,_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Any = run_command( self.cmd + [ '--config_file', 'tests/test_configs/0_12_0.yaml', '--command', self.command, '--tpu_zone', self.tpu_zone, '--tpu_name', self.tpu_name, '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls --worker all""" ,_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--debug'] ,return_stdout=_a ) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : int ): '''simple docstring''' _a : Optional[Any] = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--command', self.command, '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls --worker all""" ,_a ,) def __lowercase ( self : str ): '''simple docstring''' _a : List[str] = run_command( self.cmd + [ '--config_file', 'tests/test_configs/latest.yaml', '--command', self.command, '--command', 'echo "Hello World"', '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; ls; echo \"Hello World\" --worker all""" ,_a ,) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Any = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--command_file', self.command_file, '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Union[str, Any] = run_command( self.cmd + [ '--config_file', 'tests/test_configs/0_12_0.yaml', '--command_file', self.command_file, '--tpu_zone', self.tpu_zone, '--tpu_name', self.tpu_name, '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : Any ): '''simple docstring''' _a : Optional[int] = run_command( self.cmd + ['--config_file', 'tests/test_configs/latest.yaml', '--install_accelerate', '--debug'] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; pip install accelerate -U; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = run_command( self.cmd + [ '--config_file', 'tests/test_configs/latest.yaml', '--install_accelerate', '--accelerate_version', '12.0.0', '--debug', ] ,return_stdout=_a ,) self.assertIn( F"""{self.gcloud} test-tpu --zone us-central1-a --command {self.base_output}; pip install accelerate==12.0.0; echo \"hello world\"; echo \"this is a second command\" --worker all""" ,_a ,)
271
1
'''simple docstring''' import warnings from contextlib import contextmanager from ...processing_utils import ProcessorMixin class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[Any] = '''Speech2TextFeatureExtractor''' __UpperCAmelCase : Any = '''Speech2TextTokenizer''' def __init__( self : Tuple ,_a : List[str] ,_a : Any ): '''simple docstring''' super().__init__(_a ,_a ) _a : str = self.feature_extractor _a : List[str] = False def __call__( self : List[str] ,*_a : str ,**_a : List[Any] ): '''simple docstring''' if self._in_target_context_manager: return self.current_processor(*_a ,**_a ) if "raw_speech" in kwargs: warnings.warn('Using `raw_speech` as a keyword argument is deprecated. Use `audio` instead.' ) _a : Optional[int] = kwargs.pop('raw_speech' ) else: _a : Optional[Any] = kwargs.pop('audio' ,_a ) _a : Any = kwargs.pop('sampling_rate' ,_a ) _a : str = kwargs.pop('text' ,_a ) if len(_a ) > 0: _a : str = args[0] _a : Dict = args[1:] if audio is None and text is None: raise ValueError('You need to specify either an `audio` or `text` input to process.' ) if audio is not None: _a : Optional[Any] = self.feature_extractor(_a ,*_a ,sampling_rate=_a ,**_a ) if text is not None: _a : Any = self.tokenizer(_a ,**_a ) if text is None: return inputs elif audio is None: return encodings else: _a : int = encodings['input_ids'] return inputs def __lowercase ( self : List[str] ,*_a : Optional[Any] ,**_a : Union[str, Any] ): '''simple docstring''' return self.tokenizer.batch_decode(*_a ,**_a ) def __lowercase ( self : List[Any] ,*_a : Optional[Any] ,**_a : Optional[Any] ): '''simple docstring''' return self.tokenizer.decode(*_a ,**_a ) @contextmanager def __lowercase ( self : List[str] ): '''simple docstring''' warnings.warn( '`as_target_processor` is deprecated and will be removed in v5 of Transformers. You can process your ' 'labels by using the argument `text` of the regular `__call__` method (either in the same call as ' 'your audio inputs, or in a separate call.' ) _a : int = True _a : Dict = self.tokenizer yield _a : Any = self.feature_extractor _a : Any = False
271
'''simple docstring''' from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar __lowerCAmelCase = TypeVar("""T""") class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Tuple ,_a : T ): '''simple docstring''' _a : List[str] = data _a : Node[T] | None = None def __str__( self : Dict ): '''simple docstring''' return F"""{self.data}""" class UpperCAmelCase__ ( Generic[T] ): """simple docstring""" def __init__( self : Optional[int] ): '''simple docstring''' _a : Node[T] | None = None def __iter__( self : str ): '''simple docstring''' _a : Tuple = self.top while node: yield node.data _a : int = node.next def __str__( self : str ): '''simple docstring''' return "->".join([str(_a ) for item in self] ) def __len__( self : Optional[Any] ): '''simple docstring''' return len(tuple(iter(self ) ) ) def __lowercase ( self : str ): '''simple docstring''' return self.top is None def __lowercase ( self : List[Any] ,_a : T ): '''simple docstring''' _a : int = Node(_a ) if not self.is_empty(): _a : Optional[Any] = self.top _a : List[str] = node def __lowercase ( self : Tuple ): '''simple docstring''' if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top ,_a ) _a : List[Any] = self.top _a : int = self.top.next return pop_node.data def __lowercase ( self : List[str] ): '''simple docstring''' if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = None if __name__ == "__main__": from doctest import testmod testmod()
271
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 UpperCAmelCase_ (__a : Dict ): """simple docstring""" _a : Union[str, Any] = [tensor.shape for tensor in tensor_list] return all(shape == shapes[0] for shape in shapes[1:] ) class UpperCAmelCase__ ( lowercase__ , lowercase__ , lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : List[Any] = StableDiffusionLatentUpscalePipeline __UpperCAmelCase : Any = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - { '''height''', '''width''', '''cross_attention_kwargs''', '''negative_prompt_embeds''', '''prompt_embeds''', } __UpperCAmelCase : Optional[int] = PipelineTesterMixin.required_optional_params - {'''num_images_per_prompt'''} __UpperCAmelCase : Union[str, Any] = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS __UpperCAmelCase : Any = frozenset( [] ) # TO-DO: update image_params once pipeline is refactored with VaeImageProcessor.preprocess __UpperCAmelCase : Dict = frozenset([] ) __UpperCAmelCase : Tuple = True @property def __lowercase ( self : List[Any] ): '''simple docstring''' _a : Dict = 1 _a : int = 4 _a : Union[str, Any] = (16, 16) _a : List[str] = floats_tensor((batch_size, num_channels) + sizes ,rng=random.Random(0 ) ).to(_a ) return image def __lowercase ( self : List[str] ): '''simple docstring''' torch.manual_seed(0 ) _a : Optional[int] = UNetaDConditionModel( act_fn='gelu' ,attention_head_dim=8 ,norm_num_groups=_a ,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=_a ,only_cross_attention=_a ,out_channels=5 ,resnet_time_scale_shift='scale_shift' ,time_embedding_type='fourier' ,timestep_post_act='gelu' ,up_block_types=('KCrossAttnUpBlock2D', 'KCrossAttnUpBlock2D', 'KCrossAttnUpBlock2D', 'KUpBlock2D') ,) _a : Union[str, Any] = 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 ,) _a : Optional[Any] = EulerDiscreteScheduler(prediction_type='sample' ) _a : List[Any] = 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 ,) _a : Tuple = CLIPTextModel(_a ) _a : Optional[Any] = CLIPTokenizer.from_pretrained('hf-internal-testing/tiny-random-clip' ) _a : List[Any] = { 'unet': model.eval(), 'vae': vae.eval(), 'scheduler': scheduler, 'text_encoder': text_encoder, 'tokenizer': tokenizer, } return components def __lowercase ( self : List[str] ,_a : int ,_a : List[str]=0 ): '''simple docstring''' if str(_a ).startswith('mps' ): _a : List[Any] = torch.manual_seed(_a ) else: _a : Dict = torch.Generator(device=_a ).manual_seed(_a ) _a : int = { '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 __lowercase ( self : int ): '''simple docstring''' _a : str = 'cpu' _a : List[Any] = self.get_dummy_components() _a : Any = self.pipeline_class(**_a ) pipe.to(_a ) pipe.set_progress_bar_config(disable=_a ) _a : str = self.get_dummy_inputs(_a ) _a : Optional[int] = pipe(**_a ).images _a : Optional[int] = image[0, -3:, -3:, -1] self.assertEqual(image.shape ,(1, 256, 256, 3) ) _a : Union[str, Any] = 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] ) _a : int = np.abs(image_slice.flatten() - expected_slice ).max() self.assertLessEqual(_a ,1E-3 ) def __lowercase ( self : Tuple ): '''simple docstring''' super().test_attention_slicing_forward_pass(expected_max_diff=7E-3 ) def __lowercase ( self : List[Any] ): '''simple docstring''' super().test_cpu_offload_forward_pass(expected_max_diff=3E-3 ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' super().test_dict_tuple_outputs_equivalent(expected_max_difference=3E-3 ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' super().test_inference_batch_single_identical(expected_max_diff=7E-3 ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' super().test_pt_np_pil_outputs_equivalent(expected_max_diff=3E-3 ) def __lowercase ( self : int ): '''simple docstring''' super().test_save_load_local(expected_max_difference=3E-3 ) def __lowercase ( self : str ): '''simple docstring''' super().test_save_load_optional_components(expected_max_difference=3E-3 ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : List[Any] = [ 'DDIMScheduler', 'DDPMScheduler', 'PNDMScheduler', 'HeunDiscreteScheduler', 'EulerAncestralDiscreteScheduler', 'KDPM2DiscreteScheduler', 'KDPM2AncestralDiscreteScheduler', 'DPMSolverSDEScheduler', ] _a : Optional[int] = self.get_dummy_components() _a : Optional[int] = self.pipeline_class(**_a ) # make sure that PNDM does not need warm-up pipe.scheduler.register_to_config(skip_prk_steps=_a ) pipe.to(_a ) pipe.set_progress_bar_config(disable=_a ) _a : Union[str, Any] = self.get_dummy_inputs(_a ) _a : Tuple = 2 _a : List[str] = [] for scheduler_enum in KarrasDiffusionSchedulers: if scheduler_enum.name in skip_schedulers: # no sigma schedulers are not supported # no schedulers continue _a : int = getattr(_a ,scheduler_enum.name ) _a : Union[str, Any] = scheduler_cls.from_config(pipe.scheduler.config ) _a : List[str] = pipe(**_a )[0] outputs.append(_a ) assert check_same_shape(_a ) @require_torch_gpu @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : str ): '''simple docstring''' super().tearDown() gc.collect() torch.cuda.empty_cache() def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Any = torch.manual_seed(33 ) _a : List[Any] = StableDiffusionPipeline.from_pretrained('CompVis/stable-diffusion-v1-4' ,torch_dtype=torch.floataa ) pipe.to('cuda' ) _a : Optional[Any] = StableDiffusionLatentUpscalePipeline.from_pretrained( 'stabilityai/sd-x2-latent-upscaler' ,torch_dtype=torch.floataa ) upscaler.to('cuda' ) _a : Tuple = 'a photo of an astronaut high resolution, unreal engine, ultra realistic' _a : Union[str, Any] = pipe(_a ,generator=_a ,output_type='latent' ).images _a : List[Any] = upscaler( prompt=_a ,image=_a ,num_inference_steps=20 ,guidance_scale=0 ,generator=_a ,output_type='np' ,).images[0] _a : Tuple = 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 __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Optional[int] = torch.manual_seed(33 ) _a : Union[str, Any] = StableDiffusionLatentUpscalePipeline.from_pretrained( 'stabilityai/sd-x2-latent-upscaler' ,torch_dtype=torch.floataa ) upscaler.to('cuda' ) _a : Tuple = 'the temple of fire by Ross Tran and Gerardo Dottori, oil on canvas' _a : str = load_image( 'https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/latent-upscaler/fire_temple_512.png' ) _a : List[str] = upscaler( prompt=_a ,image=_a ,num_inference_steps=20 ,guidance_scale=0 ,generator=_a ,output_type='np' ,).images[0] _a : Tuple = 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
271
'''simple docstring''' import unittest import numpy as np import torch from diffusers import PNDMPipeline, PNDMScheduler, UNetaDModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : List[str] ): '''simple docstring''' torch.manual_seed(0 ) _a : int = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : str = self.dummy_uncond_unet _a : int = PNDMScheduler() _a : str = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Any = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ,return_dict=_a )[0] _a : List[Any] = image[0, -3:, -3:, -1] _a : Any = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : List[Any] = np.array([1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2 assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1E-2 @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[str] = 'google/ddpm-cifar10-32' _a : str = UNetaDModel.from_pretrained(_a ) _a : Union[str, Any] = PNDMScheduler() _a : Tuple = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : str = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,output_type='numpy' ).images _a : Dict = image[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : Tuple = np.array([0.1564, 0.1_4645, 0.1406, 0.1_4715, 0.1_2425, 0.1_4045, 0.1_3115, 0.1_2175, 0.125] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
271
1
'''simple docstring''' import inspect import os import torch from transformers import AutoModel from transformers.testing_utils import mockenv_context from transformers.trainer_utils import set_seed import accelerate from accelerate.accelerator import Accelerator from accelerate.state import AcceleratorState from accelerate.test_utils.testing import ( AccelerateTestCase, TempDirTestCase, execute_subprocess_async, require_cuda, require_fsdp, require_multi_gpu, slow, ) from accelerate.utils.constants import ( FSDP_AUTO_WRAP_POLICY, FSDP_BACKWARD_PREFETCH, FSDP_SHARDING_STRATEGY, FSDP_STATE_DICT_TYPE, ) from accelerate.utils.dataclasses import FullyShardedDataParallelPlugin from accelerate.utils.other import patch_environment set_seed(4_2) __lowerCAmelCase = """bert-base-cased""" __lowerCAmelCase = """fp16""" __lowerCAmelCase = """bf16""" __lowerCAmelCase = [FPaa, BFaa] @require_fsdp @require_cuda class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __lowercase ( self : Dict ): '''simple docstring''' super().setUp() _a : int = dict( ACCELERATE_USE_FSDP='true' ,MASTER_ADDR='localhost' ,MASTER_PORT='10999' ,RANK='0' ,LOCAL_RANK='0' ,WORLD_SIZE='1' ,) def __lowercase ( self : Any ): '''simple docstring''' from torch.distributed.fsdp.fully_sharded_data_parallel import ShardingStrategy for i, strategy in enumerate(_a ): _a : Union[str, Any] = self.dist_env.copy() _a : Optional[Any] = F"""{i + 1}""" _a : str = strategy with mockenv_context(**_a ): _a : Optional[int] = FullyShardedDataParallelPlugin() self.assertEqual(fsdp_plugin.sharding_strategy ,ShardingStrategy(i + 1 ) ) def __lowercase ( self : str ): '''simple docstring''' from torch.distributed.fsdp.fully_sharded_data_parallel import BackwardPrefetch for i, prefetch_policy in enumerate(_a ): _a : Tuple = self.dist_env.copy() _a : List[str] = prefetch_policy with mockenv_context(**_a ): _a : Optional[int] = FullyShardedDataParallelPlugin() if prefetch_policy == "NO_PREFETCH": self.assertIsNone(fsdp_plugin.backward_prefetch ) else: self.assertEqual(fsdp_plugin.backward_prefetch ,BackwardPrefetch(i + 1 ) ) def __lowercase ( self : Tuple ): '''simple docstring''' from torch.distributed.fsdp.fully_sharded_data_parallel import StateDictType for i, state_dict_type in enumerate(_a ): _a : List[Any] = self.dist_env.copy() _a : str = state_dict_type with mockenv_context(**_a ): _a : Optional[Any] = FullyShardedDataParallelPlugin() self.assertEqual(fsdp_plugin.state_dict_type ,StateDictType(i + 1 ) ) if state_dict_type == "FULL_STATE_DICT": self.assertTrue(fsdp_plugin.state_dict_config.offload_to_cpu ) self.assertTrue(fsdp_plugin.state_dict_config.ranka_only ) def __lowercase ( self : List[Any] ): '''simple docstring''' _a : Dict = AutoModel.from_pretrained(_a ) for policy in FSDP_AUTO_WRAP_POLICY: _a : Optional[int] = self.dist_env.copy() _a : Union[str, Any] = policy if policy == "TRANSFORMER_BASED_WRAP": _a : int = 'BertLayer' elif policy == "SIZE_BASED_WRAP": _a : Optional[int] = '2000' with mockenv_context(**_a ): _a : Union[str, Any] = FullyShardedDataParallelPlugin() fsdp_plugin.set_auto_wrap_policy(_a ) if policy == "NO_WRAP": self.assertIsNone(fsdp_plugin.auto_wrap_policy ) else: self.assertIsNotNone(fsdp_plugin.auto_wrap_policy ) _a : Tuple = self.dist_env.copy() _a : List[str] = 'TRANSFORMER_BASED_WRAP' _a : Optional[Any] = 'T5Layer' with mockenv_context(**_a ): _a : int = FullyShardedDataParallelPlugin() with self.assertRaises(_a ) as cm: fsdp_plugin.set_auto_wrap_policy(_a ) self.assertTrue('Could not find the transformer layer class to wrap in the model.' in str(cm.exception ) ) _a : Tuple = self.dist_env.copy() _a : Optional[int] = 'SIZE_BASED_WRAP' _a : str = '0' with mockenv_context(**_a ): _a : Optional[int] = FullyShardedDataParallelPlugin() fsdp_plugin.set_auto_wrap_policy(_a ) self.assertIsNone(fsdp_plugin.auto_wrap_policy ) def __lowercase ( self : int ): '''simple docstring''' from torch.distributed.fsdp.fully_sharded_data_parallel import MixedPrecision from torch.distributed.fsdp.sharded_grad_scaler import ShardedGradScaler for mp_dtype in dtypes: _a : Any = self.dist_env.copy() _a : List[str] = mp_dtype with mockenv_context(**_a ): _a : Dict = Accelerator() if mp_dtype == "fp16": _a : List[str] = torch.floataa elif mp_dtype == "bf16": _a : Any = torch.bfloataa _a : Any = MixedPrecision(param_dtype=_a ,reduce_dtype=_a ,buffer_dtype=_a ) self.assertEqual(accelerator.state.fsdp_plugin.mixed_precision_policy ,_a ) if mp_dtype == FPaa: self.assertTrue(isinstance(accelerator.scaler ,_a ) ) elif mp_dtype == BFaa: self.assertIsNone(accelerator.scaler ) AcceleratorState._reset_state(_a ) def __lowercase ( self : str ): '''simple docstring''' from torch.distributed.fsdp.fully_sharded_data_parallel import CPUOffload for flag in [True, False]: _a : List[str] = self.dist_env.copy() _a : Tuple = str(_a ).lower() with mockenv_context(**_a ): _a : List[str] = FullyShardedDataParallelPlugin() self.assertEqual(fsdp_plugin.cpu_offload ,CPUOffload(offload_params=_a ) ) @require_fsdp @require_multi_gpu @slow class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __lowercase ( self : List[str] ): '''simple docstring''' super().setUp() _a : int = 0.82 _a : Union[str, Any] = [ 'fsdp_shard_grad_op_transformer_based_wrap', 'fsdp_full_shard_transformer_based_wrap', ] _a : Dict = { 'multi_gpu_fp16': 3200, 'fsdp_shard_grad_op_transformer_based_wrap_fp16': 2000, 'fsdp_full_shard_transformer_based_wrap_fp16': 1900, # Disabling below test as it overwhelms the RAM memory usage # on CI self-hosted runner leading to tests getting killed. # "fsdp_full_shard_cpu_offload_transformer_based_wrap_fp32": 1500, # fp16 was leading to indefinite hang } _a : Tuple = 160 _a : Dict = 160 _a : List[str] = inspect.getfile(accelerate.test_utils ) _a : Optional[Any] = os.path.sep.join(mod_file.split(os.path.sep )[:-1] + ['scripts', 'external_deps'] ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : List[str] = os.path.join(self.test_scripts_folder ,'test_performance.py' ) _a : Any = ['accelerate', 'launch', '--num_processes=2', '--num_machines=1', '--machine_rank=0', '--use_fsdp'] for config in self.performance_configs: _a : int = cmd.copy() for i, strategy in enumerate(_a ): if strategy.lower() in config: cmd_config.append(F"""--fsdp_sharding_strategy={i+1}""" ) break if "fp32" in config: cmd_config.append('--mixed_precision=no' ) else: cmd_config.append('--mixed_precision=fp16' ) if "cpu_offload" in config: cmd_config.append('--fsdp_offload_params=True' ) for policy in FSDP_AUTO_WRAP_POLICY: if policy.lower() in config: cmd_config.append(F"""--fsdp_auto_wrap_policy={policy}""" ) break if policy == "TRANSFORMER_BASED_WRAP": cmd_config.append('--fsdp_transformer_layer_cls_to_wrap=BertLayer' ) elif policy == "SIZE_BASED_WRAP": cmd_config.append('--fsdp_min_num_params=2000' ) cmd_config.extend( [ self.test_file_path, F"""--output_dir={self.tmpdir}""", F"""--performance_lower_bound={self.performance_lower_bound}""", ] ) with patch_environment(omp_num_threads=1 ): execute_subprocess_async(_a ,env=os.environ.copy() ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : Optional[Any] = os.path.join(self.test_scripts_folder ,'test_checkpointing.py' ) _a : Union[str, Any] = [ 'accelerate', 'launch', '--num_processes=2', '--num_machines=1', '--machine_rank=0', '--use_fsdp', '--mixed_precision=fp16', '--fsdp_transformer_layer_cls_to_wrap=BertLayer', ] for i, strategy in enumerate(_a ): _a : int = cmd.copy() cmd_config.append(F"""--fsdp_sharding_strategy={i+1}""" ) if strategy != "FULL_SHARD": continue _a : int = len(_a ) for state_dict_type in FSDP_STATE_DICT_TYPE: _a : str = cmd_config[:state_dict_config_index] cmd_config.append(F"""--fsdp_state_dict_type={state_dict_type}""" ) cmd_config.extend( [ self.test_file_path, F"""--output_dir={self.tmpdir}""", '--partial_train_epoch=1', ] ) with patch_environment(omp_num_threads=1 ): execute_subprocess_async(_a ,env=os.environ.copy() ) _a : Optional[int] = cmd_config[:-1] _a : Tuple = os.path.join(self.tmpdir ,'epoch_0' ) cmd_config.extend( [ F"""--resume_from_checkpoint={resume_from_checkpoint}""", ] ) with patch_environment(omp_num_threads=1 ): execute_subprocess_async(_a ,env=os.environ.copy() ) def __lowercase ( self : List[str] ): '''simple docstring''' _a : int = os.path.join(self.test_scripts_folder ,'test_peak_memory_usage.py' ) _a : Optional[int] = [ 'accelerate', 'launch', '--num_processes=2', '--num_machines=1', '--machine_rank=0', ] for spec, peak_mem_upper_bound in self.peak_memory_usage_upper_bound.items(): _a : Any = cmd.copy() if "fp16" in spec: cmd_config.extend(['--mixed_precision=fp16'] ) else: cmd_config.extend(['--mixed_precision=no'] ) if "multi_gpu" in spec: continue else: cmd_config.extend(['--use_fsdp'] ) for i, strategy in enumerate(_a ): if strategy.lower() in spec: cmd_config.append(F"""--fsdp_sharding_strategy={i+1}""" ) break if "cpu_offload" in spec: cmd_config.append('--fsdp_offload_params=True' ) for policy in FSDP_AUTO_WRAP_POLICY: if policy.lower() in spec: cmd_config.append(F"""--fsdp_auto_wrap_policy={policy}""" ) break if policy == "TRANSFORMER_BASED_WRAP": cmd_config.append('--fsdp_transformer_layer_cls_to_wrap=BertLayer' ) elif policy == "SIZE_BASED_WRAP": cmd_config.append('--fsdp_min_num_params=2000' ) cmd_config.extend( [ self.test_file_path, F"""--output_dir={self.tmpdir}""", F"""--peak_memory_upper_bound={peak_mem_upper_bound}""", F"""--n_train={self.n_train}""", F"""--n_val={self.n_val}""", ] ) with patch_environment(omp_num_threads=1 ): execute_subprocess_async(_a ,env=os.environ.copy() )
271
'''simple docstring''' import doctest import logging import os import unittest from pathlib import Path from typing import List, Union import transformers from transformers.testing_utils import require_tf, require_torch, slow __lowerCAmelCase = logging.getLogger() @unittest.skip('''Temporarily disable the doc tests.''' ) @require_torch @require_tf @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : str ,_a : Path ,_a : Union[str, None] = None ,_a : Union[List[str], None] = None ,_a : Union[str, List[str], None] = None ,_a : bool = True ,): '''simple docstring''' _a : Optional[int] = [file for file in os.listdir(_a ) if os.path.isfile(os.path.join(_a ,_a ) )] if identifier is not None: _a : List[str] = [file for file in files if identifier in file] if n_identifier is not None: if isinstance(_a ,_a ): for n_ in n_identifier: _a : Tuple = [file for file in files if n_ not in file] else: _a : Optional[Any] = [file for file in files if n_identifier not in file] _a : List[str] = ignore_files or [] ignore_files.append('__init__.py' ) _a : Tuple = [file for file in files if file not in ignore_files] for file in files: # Open all files print('Testing' ,_a ) if only_modules: _a : Any = file.split('.' )[0] try: _a : List[str] = getattr(_a ,_a ) _a : int = doctest.DocTestSuite(_a ) _a : Any = unittest.TextTestRunner().run(_a ) self.assertIs(len(result.failures ) ,0 ) except AttributeError: logger.info(F"""{module_identifier} is not a module.""" ) else: _a : Union[str, Any] = doctest.testfile(str('..' / directory / file ) ,optionflags=doctest.ELLIPSIS ) self.assertIs(result.failed ,0 ) def __lowercase ( self : Any ): '''simple docstring''' _a : int = Path('src/transformers' ) _a : List[Any] = 'modeling' _a : Optional[Any] = [ 'modeling_ctrl.py', 'modeling_tf_ctrl.py', ] self.analyze_directory(_a ,identifier=_a ,ignore_files=_a ) def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Optional[Any] = Path('src/transformers' ) _a : Optional[Any] = 'tokenization' self.analyze_directory(_a ,identifier=_a ) def __lowercase ( self : List[str] ): '''simple docstring''' _a : Dict = Path('src/transformers' ) _a : str = 'configuration' self.analyze_directory(_a ,identifier=_a ) def __lowercase ( self : Any ): '''simple docstring''' _a : Tuple = Path('src/transformers' ) _a : List[Any] = ['configuration', 'modeling', 'tokenization'] self.analyze_directory(_a ,n_identifier=_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[Any] = Path('docs/source' ) _a : List[str] = ['favicon.ico'] self.analyze_directory(_a ,ignore_files=_a ,only_modules=_a )
271
1
'''simple docstring''' import unittest 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 if is_torch_available(): import torch if is_vision_available(): from PIL import Image from transformers import ChineseCLIPImageProcessor class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __init__( self : Union[str, Any] ,_a : int ,_a : List[Any]=7 ,_a : Any=3 ,_a : Optional[Any]=18 ,_a : Any=30 ,_a : List[str]=400 ,_a : Dict=True ,_a : int=None ,_a : Optional[Any]=True ,_a : str=None ,_a : Optional[int]=True ,_a : Any=[0.4814_5466, 0.457_8275, 0.4082_1073] ,_a : str=[0.2686_2954, 0.2613_0258, 0.2757_7711] ,_a : Optional[int]=True ,): '''simple docstring''' _a : Tuple = size if size is not None else {'height': 224, 'width': 224} _a : Any = crop_size if crop_size is not None else {'height': 18, 'width': 18} _a : Optional[Any] = parent _a : List[str] = batch_size _a : Union[str, Any] = num_channels _a : str = image_size _a : Dict = min_resolution _a : Union[str, Any] = max_resolution _a : List[Any] = do_resize _a : Optional[int] = size _a : int = do_center_crop _a : Any = crop_size _a : Optional[int] = do_normalize _a : Union[str, Any] = image_mean _a : Union[str, Any] = image_std _a : Any = do_convert_rgb def __lowercase ( self : Optional[int] ): '''simple docstring''' return { "do_resize": self.do_resize, "size": self.size, "do_center_crop": self.do_center_crop, "crop_size": self.crop_size, "do_normalize": self.do_normalize, "image_mean": self.image_mean, "image_std": self.image_std, "do_convert_rgb": self.do_convert_rgb, } def __lowercase ( self : Optional[int] ,_a : Optional[Any]=False ,_a : Union[str, Any]=False ,_a : int=False ): '''simple docstring''' assert not (numpify and torchify), "You cannot specify both numpy and PyTorch tensors at the same time" if equal_resolution: _a : Dict = [] for i in range(self.batch_size ): image_inputs.append( np.random.randint( 255 ,size=(self.num_channels, self.max_resolution, self.max_resolution) ,dtype=np.uinta ) ) else: _a : int = [] for i in range(self.batch_size ): _a, _a : Union[str, Any] = np.random.choice(np.arange(self.min_resolution ,self.max_resolution ) ,2 ) image_inputs.append(np.random.randint(255 ,size=(self.num_channels, width, height) ,dtype=np.uinta ) ) if not numpify and not torchify: # PIL expects the channel dimension as last dimension _a : List[Any] = [Image.fromarray(np.moveaxis(_a ,0 ,-1 ) ) for x in image_inputs] if torchify: _a : Any = [torch.from_numpy(_a ) for x in image_inputs] return image_inputs @require_torch @require_vision class UpperCAmelCase__ ( lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[int] = ChineseCLIPImageProcessor if is_vision_available() else None def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : int = ChineseCLIPImageProcessingTester(self ,do_center_crop=_a ) @property def __lowercase ( self : Tuple ): '''simple docstring''' return self.image_processor_tester.prepare_image_processor_dict() def __lowercase ( self : List[Any] ): '''simple docstring''' _a : List[Any] = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(_a ,'do_resize' ) ) self.assertTrue(hasattr(_a ,'size' ) ) self.assertTrue(hasattr(_a ,'do_center_crop' ) ) self.assertTrue(hasattr(_a ,'center_crop' ) ) self.assertTrue(hasattr(_a ,'do_normalize' ) ) self.assertTrue(hasattr(_a ,'image_mean' ) ) self.assertTrue(hasattr(_a ,'image_std' ) ) self.assertTrue(hasattr(_a ,'do_convert_rgb' ) ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[str] = self.image_processing_class.from_dict(self.image_processor_dict ) self.assertEqual(image_processor.size ,{'height': 224, 'width': 224} ) self.assertEqual(image_processor.crop_size ,{'height': 18, 'width': 18} ) _a : List[Any] = self.image_processing_class.from_dict(self.image_processor_dict ,size=42 ,crop_size=84 ) self.assertEqual(image_processor.size ,{'shortest_edge': 42} ) self.assertEqual(image_processor.crop_size ,{'height': 84, 'width': 84} ) def __lowercase ( self : List[str] ): '''simple docstring''' pass def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : List[str] = self.image_processing_class(**self.image_processor_dict ) # create random PIL images _a : Dict = self.image_processor_tester.prepare_inputs(equal_resolution=_a ) for image in image_inputs: self.assertIsInstance(_a ,Image.Image ) # Test not batched input _a : Optional[int] = image_processing(image_inputs[0] ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( 1, self.image_processor_tester.num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) # Test batched _a : str = image_processing(_a ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( self.image_processor_tester.batch_size, self.image_processor_tester.num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) def __lowercase ( self : List[Any] ): '''simple docstring''' _a : Optional[int] = self.image_processing_class(**self.image_processor_dict ) # create random numpy tensors _a : Optional[int] = self.image_processor_tester.prepare_inputs(equal_resolution=_a ,numpify=_a ) for image in image_inputs: self.assertIsInstance(_a ,np.ndarray ) # Test not batched input _a : List[Any] = image_processing(image_inputs[0] ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( 1, self.image_processor_tester.num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) # Test batched _a : Any = image_processing(_a ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( self.image_processor_tester.batch_size, self.image_processor_tester.num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) def __lowercase ( self : Dict ): '''simple docstring''' _a : Optional[int] = self.image_processing_class(**self.image_processor_dict ) # create random PyTorch tensors _a : int = self.image_processor_tester.prepare_inputs(equal_resolution=_a ,torchify=_a ) for image in image_inputs: self.assertIsInstance(_a ,torch.Tensor ) # Test not batched input _a : List[Any] = image_processing(image_inputs[0] ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( 1, self.image_processor_tester.num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) # Test batched _a : Dict = image_processing(_a ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( self.image_processor_tester.batch_size, self.image_processor_tester.num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) @require_torch @require_vision class UpperCAmelCase__ ( lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : List[Any] = ChineseCLIPImageProcessor if is_vision_available() else None def __lowercase ( self : str ): '''simple docstring''' _a : Union[str, Any] = ChineseCLIPImageProcessingTester(self ,num_channels=4 ,do_center_crop=_a ) _a : Union[str, Any] = 3 @property def __lowercase ( self : str ): '''simple docstring''' return self.image_processor_tester.prepare_image_processor_dict() def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : str = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(_a ,'do_resize' ) ) self.assertTrue(hasattr(_a ,'size' ) ) self.assertTrue(hasattr(_a ,'do_center_crop' ) ) self.assertTrue(hasattr(_a ,'center_crop' ) ) self.assertTrue(hasattr(_a ,'do_normalize' ) ) self.assertTrue(hasattr(_a ,'image_mean' ) ) self.assertTrue(hasattr(_a ,'image_std' ) ) self.assertTrue(hasattr(_a ,'do_convert_rgb' ) ) def __lowercase ( self : Tuple ): '''simple docstring''' pass def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[Any] = self.image_processing_class(**self.image_processor_dict ) # create random PIL images _a : List[str] = self.image_processor_tester.prepare_inputs(equal_resolution=_a ) for image in image_inputs: self.assertIsInstance(_a ,Image.Image ) # Test not batched input _a : Union[str, Any] = image_processing(image_inputs[0] ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( 1, self.expected_encoded_image_num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,) # Test batched _a : Union[str, Any] = image_processing(_a ,return_tensors='pt' ).pixel_values self.assertEqual( encoded_images.shape ,( self.image_processor_tester.batch_size, self.expected_encoded_image_num_channels, self.image_processor_tester.crop_size['height'], self.image_processor_tester.crop_size['width'], ) ,)
271
'''simple docstring''' import argparse import pickle import numpy as np import torch from torch import nn from transformers import ReformerConfig, ReformerModelWithLMHead from transformers.utils import logging logging.set_verbosity_info() def UpperCAmelCase_ (__a : Optional[Any] , __a : str , __a : Optional[Any]=None ): """simple docstring""" assert torch_layer.weight.shape == weight.shape, f"""{torch_layer} layer.weight does not match""" _a : str = nn.Parameter(__a ) if bias is not None: assert torch_layer.bias.shape == bias.shape, f"""{torch_layer} layer.bias does not match""" _a : Any = nn.Parameter(__a ) def UpperCAmelCase_ (__a : int , __a : Optional[Any] , __a : int ): """simple docstring""" _a : Tuple = np.asarray(weights[0] ) _a : Union[str, Any] = np.asarray(weights[1] ) _a : Dict = np.asarray(weights[2] ) set_param( torch_layer.self_attention.query_key , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.value , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.output.dense , torch.tensor(__a ).view(-1 , __a ).contiguous().transpose(0 , 1 ) , ) def UpperCAmelCase_ (__a : Optional[Any] , __a : Optional[int] , __a : List[str] ): """simple docstring""" _a : Dict = np.asarray(weights[0] ) _a : Union[str, Any] = np.asarray(weights[1] ) _a : str = np.asarray(weights[2] ) _a : int = np.asarray(weights[3] ) set_param( torch_layer.self_attention.query , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.key , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.self_attention.value , torch.tensor(__a ).transpose(1 , 2 ).contiguous().view(-1 , __a ) , ) set_param( torch_layer.output.dense , torch.tensor(__a ).view(-1 , __a ).contiguous().transpose(0 , 1 ) , ) def UpperCAmelCase_ (__a : Any , __a : Any , __a : Optional[Any] ): """simple docstring""" _a : List[str] = weights[0][0][0] _a : List[Any] = np.asarray(layer_norm_a[0] ) _a : List[str] = np.asarray(layer_norm_a[1] ) set_param( torch_block.attention.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # lsh weights + output _a : List[str] = weights[0][1] if len(__a ) < 4: set_layer_weights_in_torch_lsh(__a , torch_block.attention , __a ) else: set_layer_weights_in_torch_local(__a , torch_block.attention , __a ) # intermediate weighs _a : Optional[Any] = weights[2][0][1][2] # Chunked Feed Forward if len(__a ) == 4: _a : Union[str, Any] = intermediate_weights[2] # layernorm 2 _a : Any = np.asarray(intermediate_weights[0][0] ) _a : List[Any] = np.asarray(intermediate_weights[0][1] ) set_param( torch_block.feed_forward.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # intermediate dense _a : Any = np.asarray(intermediate_weights[1][0] ) _a : Any = np.asarray(intermediate_weights[1][1] ) set_param( torch_block.feed_forward.dense.dense , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) # intermediate out _a : Optional[int] = np.asarray(intermediate_weights[4][0] ) _a : int = np.asarray(intermediate_weights[4][1] ) set_param( torch_block.feed_forward.output.dense , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) def UpperCAmelCase_ (__a : Dict , __a : Dict , __a : List[Any] ): """simple docstring""" _a : Optional[int] = torch_model.reformer # word embeds _a : Tuple = np.asarray(weights[1] ) set_param( torch_model_reformer.embeddings.word_embeddings , torch.tensor(__a ) , ) if isinstance(weights[3] , __a ): _a : Any = torch_model_reformer.embeddings.position_embeddings for emb_idx in range(len(position_embeddings.weights ) ): _a : List[Any] = np.asarray(weights[3][emb_idx][0] ) assert ( position_embeddings.weights[emb_idx].shape == emb_weights.shape ), f"""{position_embeddings[emb_idx]} emb does not match""" _a : Any = nn.Parameter(torch.tensor(__a ) ) _a : List[str] = weights[5] assert len(torch_model_reformer.encoder.layers ) * 4 == len( __a ), "HF and trax model do not have the same number of layers" for layer_idx, layer in enumerate(torch_model_reformer.encoder.layers ): _a : Tuple = trax_layer_weights[4 * layer_idx : 4 * (layer_idx + 1)] set_block_weights_in_torch(__a , __a , __a ) # output layer norm _a : Optional[Any] = np.asarray(weights[7][0] ) _a : int = np.asarray(weights[7][1] ) set_param( torch_model_reformer.encoder.layer_norm , torch.tensor(__a ) , torch.tensor(__a ) , ) # output embeddings _a : List[str] = np.asarray(weights[9][0] ) _a : int = np.asarray(weights[9][1] ) set_param( torch_model.lm_head.decoder , torch.tensor(__a ).transpose(0 , 1 ).contiguous() , torch.tensor(__a ) , ) def UpperCAmelCase_ (__a : Tuple , __a : Optional[Any] , __a : Dict ): """simple docstring""" _a : List[Any] = ReformerConfig.from_json_file(__a ) print(f"""Building PyTorch model from configuration: {config}""" ) _a : int = ReformerModelWithLMHead(__a ) with open(__a , 'rb' ) as f: _a : Optional[Any] = pickle.load(__a )['weights'] set_model_weights_in_torch(__a , __a , config.hidden_size ) # Save pytorch-model print(f"""Save PyTorch model to {pytorch_dump_path}""" ) torch.save(model.state_dict() , __a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser() # Required parameters parser.add_argument( """--trax_model_pkl_path""", default=None, type=str, required=True, help="""Path to the TensorFlow checkpoint path.""" ) parser.add_argument( """--config_file""", default=None, type=str, required=True, help=( """The config json file corresponding to the pre-trained Reformer model. \n""" """This specifies the model architecture.""" ), ) parser.add_argument( """--pytorch_dump_path""", default=None, type=str, required=True, help="""Path to the output PyTorch model.""" ) __lowerCAmelCase = parser.parse_args() convert_trax_checkpoint_to_pytorch(args.trax_model_pkl_path, args.config_file, args.pytorch_dump_path)
271
1
'''simple docstring''' from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available __lowerCAmelCase = { """configuration_clap""": [ """CLAP_PRETRAINED_MODEL_ARCHIVE_LIST""", """ClapAudioConfig""", """ClapConfig""", """ClapTextConfig""", ], """processing_clap""": ["""ClapProcessor"""], } try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: __lowerCAmelCase = [ """CLAP_PRETRAINED_MODEL_ARCHIVE_LIST""", """ClapModel""", """ClapPreTrainedModel""", """ClapTextModel""", """ClapTextModelWithProjection""", """ClapAudioModel""", """ClapAudioModelWithProjection""", ] __lowerCAmelCase = ["""ClapFeatureExtractor"""] if TYPE_CHECKING: from .configuration_clap import ( CLAP_PRETRAINED_MODEL_ARCHIVE_LIST, ClapAudioConfig, ClapConfig, ClapTextConfig, ) from .processing_clap import ClapProcessor try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .feature_extraction_clap import ClapFeatureExtractor from .modeling_clap import ( CLAP_PRETRAINED_MODEL_ARCHIVE_LIST, ClapAudioModel, ClapAudioModelWithProjection, ClapModel, ClapPreTrainedModel, ClapTextModel, ClapTextModelWithProjection, ) else: import sys __lowerCAmelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
271
'''simple docstring''' import unittest import numpy as np import torch from transformers import CLIPTextConfig, CLIPTextModel from diffusers import DDIMScheduler, LDMPipeline, UNetaDModel, VQModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' torch.manual_seed(0 ) _a : Any = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model @property def __lowercase ( self : Optional[int] ): '''simple docstring''' torch.manual_seed(0 ) _a : Union[str, Any] = VQModel( block_out_channels=[32, 64] ,in_channels=3 ,out_channels=3 ,down_block_types=['DownEncoderBlock2D', 'DownEncoderBlock2D'] ,up_block_types=['UpDecoderBlock2D', 'UpDecoderBlock2D'] ,latent_channels=3 ,) return model @property def __lowercase ( self : Optional[int] ): '''simple docstring''' torch.manual_seed(0 ) _a : Any = 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 ,) return CLIPTextModel(_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : Dict = self.dummy_uncond_unet _a : List[Any] = DDIMScheduler() _a : List[Any] = self.dummy_vq_model _a : str = LDMPipeline(unet=_a ,vqvae=_a ,scheduler=_a ) ldm.to(_a ) ldm.set_progress_bar_config(disable=_a ) _a : List[str] = torch.manual_seed(0 ) _a : List[str] = ldm(generator=_a ,num_inference_steps=2 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Union[str, Any] = ldm(generator=_a ,num_inference_steps=2 ,output_type='numpy' ,return_dict=_a )[0] _a : Tuple = image[0, -3:, -3:, -1] _a : Optional[Any] = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 64, 64, 3) _a : int = np.array([0.8512, 0.818, 0.6411, 0.6808, 0.4465, 0.5618, 0.46, 0.6231, 0.5172] ) _a : Any = 1E-2 if torch_device != 'mps' else 3E-2 assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < tolerance @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : List[str] = LDMPipeline.from_pretrained('CompVis/ldm-celebahq-256' ) ldm.to(_a ) ldm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Dict = ldm(generator=_a ,num_inference_steps=5 ,output_type='numpy' ).images _a : str = image[0, -3:, -3:, -1] assert image.shape == (1, 256, 256, 3) _a : Optional[int] = np.array([0.4399, 0.4_4975, 0.4_6825, 0.474, 0.4359, 0.4581, 0.4_5095, 0.4341, 0.4447] ) _a : int = 1E-2 if torch_device != 'mps' else 3E-2 assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
271
1
'''simple docstring''' # DISCLAIMER: This code is strongly influenced by https://github.com/pesser/pytorch_diffusion # and https://github.com/hojonathanho/diffusion import math from dataclasses import dataclass from typing import List, Optional, Tuple, Union import numpy as np import torch from diffusers.configuration_utils import ConfigMixin, register_to_config from diffusers.schedulers.scheduling_utils import SchedulerMixin from diffusers.utils import BaseOutput, deprecate @dataclass # Copied from diffusers.schedulers.scheduling_ddpm.DDPMSchedulerOutput with DDPM->DDIM class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : torch.FloatTensor __UpperCAmelCase : Optional[torch.FloatTensor] = None def UpperCAmelCase_ (__a : Tuple , __a : Dict=0.999 , __a : List[Any]="cosine" , ): """simple docstring""" if alpha_transform_type == "cosine": def alpha_bar_fn(__a : List[str] ): return math.cos((t + 0.008) / 1.008 * math.pi / 2 ) ** 2 elif alpha_transform_type == "exp": def alpha_bar_fn(__a : Any ): return math.exp(t * -12.0 ) else: raise ValueError(f"""Unsupported alpha_tranform_type: {alpha_transform_type}""" ) _a : Union[str, Any] = [] for i in range(__a ): _a : Any = i / num_diffusion_timesteps _a : List[str] = (i + 1) / num_diffusion_timesteps betas.append(min(1 - alpha_bar_fn(__a ) / alpha_bar_fn(__a ) , __a ) ) return torch.tensor(__a , dtype=torch.floataa ) class UpperCAmelCase__ ( lowercase__ , lowercase__ ): """simple docstring""" __UpperCAmelCase : Dict = 1 @register_to_config def __init__( self : int ,_a : int = 1000 ,_a : float = 0.0001 ,_a : float = 0.02 ,_a : str = "linear" ,_a : Optional[Union[np.ndarray, List[float]]] = None ,_a : bool = True ,_a : bool = True ,_a : int = 0 ,_a : str = "epsilon" ,_a : float = 1.0 ,**_a : Union[str, Any] ,): '''simple docstring''' if kwargs.get('set_alpha_to_one' ,_a ) is not None: _a : str = ( 'The `set_alpha_to_one` argument is deprecated. Please use `set_alpha_to_zero` instead.' ) deprecate('set_alpha_to_one' ,'1.0.0' ,_a ,standard_warn=_a ) _a : Union[str, Any] = kwargs['set_alpha_to_one'] if trained_betas is not None: _a : int = torch.tensor(_a ,dtype=torch.floataa ) elif beta_schedule == "linear": _a : List[str] = torch.linspace(_a ,_a ,_a ,dtype=torch.floataa ) elif beta_schedule == "scaled_linear": # this schedule is very specific to the latent diffusion model. _a : Union[str, Any] = ( torch.linspace(beta_start**0.5 ,beta_end**0.5 ,_a ,dtype=torch.floataa ) ** 2 ) elif beta_schedule == "squaredcos_cap_v2": # Glide cosine schedule _a : Any = betas_for_alpha_bar(_a ) else: raise NotImplementedError(F"""{beta_schedule} does is not implemented for {self.__class__}""" ) _a : List[Any] = 1.0 - self.betas _a : Dict = torch.cumprod(self.alphas ,dim=0 ) # At every step in inverted ddim, we are looking into the next alphas_cumprod # For the final step, there is no next alphas_cumprod, and the index is out of bounds # `set_alpha_to_zero` decides whether we set this parameter simply to zero # in this case, self.step() just output the predicted noise # or whether we use the final alpha of the "non-previous" one. _a : int = torch.tensor(0.0 ) if set_alpha_to_zero else self.alphas_cumprod[-1] # standard deviation of the initial noise distribution _a : Optional[int] = 1.0 # setable values _a : int = None _a : Any = torch.from_numpy(np.arange(0 ,_a ).copy().astype(np.intaa ) ) def __lowercase ( self : List[Any] ,_a : torch.FloatTensor ,_a : Optional[int] = None ): '''simple docstring''' return sample def __lowercase ( self : List[str] ,_a : int ,_a : Union[str, torch.device] = None ): '''simple docstring''' if num_inference_steps > self.config.num_train_timesteps: raise ValueError( F"""`num_inference_steps`: {num_inference_steps} cannot be larger than `self.config.train_timesteps`:""" F""" {self.config.num_train_timesteps} as the unet model trained with this scheduler can only handle""" F""" maximal {self.config.num_train_timesteps} timesteps.""" ) _a : str = num_inference_steps _a : Optional[Any] = self.config.num_train_timesteps // self.num_inference_steps # creates integer timesteps by multiplying by ratio # casting to int to avoid issues when num_inference_step is power of 3 _a : Any = (np.arange(0 ,_a ) * step_ratio).round().copy().astype(np.intaa ) _a : List[Any] = torch.from_numpy(_a ).to(_a ) self.timesteps += self.config.steps_offset def __lowercase ( self : List[str] ,_a : torch.FloatTensor ,_a : int ,_a : torch.FloatTensor ,_a : float = 0.0 ,_a : bool = False ,_a : Optional[torch.FloatTensor] = None ,_a : bool = True ,): '''simple docstring''' _a : Union[str, Any] = timestep + self.config.num_train_timesteps // self.num_inference_steps # 2. compute alphas, betas # change original implementation to exactly match noise levels for analogous forward process _a : List[str] = self.alphas_cumprod[timestep] _a : List[str] = ( self.alphas_cumprod[prev_timestep] if prev_timestep < self.config.num_train_timesteps else self.final_alpha_cumprod ) _a : Union[str, Any] = 1 - alpha_prod_t # 3. compute predicted original sample from predicted noise also called # "predicted x_0" of formula (12) from https://arxiv.org/pdf/2010.02502.pdf if self.config.prediction_type == "epsilon": _a : str = (sample - beta_prod_t ** 0.5 * model_output) / alpha_prod_t ** 0.5 _a : List[str] = model_output elif self.config.prediction_type == "sample": _a : int = model_output _a : List[Any] = (sample - alpha_prod_t ** 0.5 * pred_original_sample) / beta_prod_t ** 0.5 elif self.config.prediction_type == "v_prediction": _a : Any = (alpha_prod_t**0.5) * sample - (beta_prod_t**0.5) * model_output _a : Optional[Any] = (alpha_prod_t**0.5) * model_output + (beta_prod_t**0.5) * sample else: raise ValueError( F"""prediction_type given as {self.config.prediction_type} must be one of `epsilon`, `sample`, or""" ' `v_prediction`' ) # 4. Clip or threshold "predicted x_0" if self.config.clip_sample: _a : Optional[int] = pred_original_sample.clamp( -self.config.clip_sample_range ,self.config.clip_sample_range ) # 5. compute "direction pointing to x_t" of formula (12) from https://arxiv.org/pdf/2010.02502.pdf _a : Any = (1 - alpha_prod_t_prev) ** 0.5 * pred_epsilon # 6. compute x_t without "random noise" of formula (12) from https://arxiv.org/pdf/2010.02502.pdf _a : Dict = alpha_prod_t_prev ** 0.5 * pred_original_sample + pred_sample_direction if not return_dict: return (prev_sample, pred_original_sample) return DDIMSchedulerOutput(prev_sample=_a ,pred_original_sample=_a ) def __len__( self : List[str] ): '''simple docstring''' return self.config.num_train_timesteps
271
'''simple docstring''' import warnings from ...utils import logging from .image_processing_beit import BeitImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : int ,*_a : Optional[int] ,**_a : str ): '''simple docstring''' warnings.warn( 'The class BeitFeatureExtractor is deprecated and will be removed in version 5 of Transformers. Please' ' use BeitImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
1
'''simple docstring''' import json import os import re import unicodedata from json.encoder import INFINITY from typing import Any, Dict, List, Optional, Tuple, Union import numpy as np import regex from ...tokenization_utils import AddedToken, PreTrainedTokenizer from ...tokenization_utils_base import BatchEncoding from ...utils import TensorType, is_flax_available, is_tf_available, is_torch_available, logging from ...utils.generic import _is_jax, _is_numpy __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """artists_file""": """artists.json""", """lyrics_file""": """lyrics.json""", """genres_file""": """genres.json""", } __lowerCAmelCase = { """artists_file""": { """jukebox""": """https://huggingface.co/ArthurZ/jukebox/blob/main/artists.json""", }, """genres_file""": { """jukebox""": """https://huggingface.co/ArthurZ/jukebox/blob/main/genres.json""", }, """lyrics_file""": { """jukebox""": """https://huggingface.co/ArthurZ/jukebox/blob/main/lyrics.json""", }, } __lowerCAmelCase = { """jukebox""": 5_1_2, } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : str = VOCAB_FILES_NAMES __UpperCAmelCase : Optional[int] = PRETRAINED_VOCAB_FILES_MAP __UpperCAmelCase : Union[str, Any] = PRETRAINED_LYRIC_TOKENS_SIZES __UpperCAmelCase : List[str] = ['''input_ids''', '''attention_mask'''] def __init__( self : str ,_a : Any ,_a : Dict ,_a : str ,_a : Any=["v3", "v2", "v2"] ,_a : Tuple=512 ,_a : int=5 ,_a : Any="<|endoftext|>" ,**_a : List[Any] ,): '''simple docstring''' _a : Tuple = AddedToken(_a ,lstrip=_a ,rstrip=_a ) if isinstance(_a ,_a ) else unk_token super().__init__( unk_token=_a ,n_genres=_a ,version=_a ,max_n_lyric_tokens=_a ,**_a ,) _a : int = version _a : Dict = max_n_lyric_tokens _a : Dict = n_genres with open(_a ,encoding='utf-8' ) as vocab_handle: _a : List[Any] = json.load(_a ) with open(_a ,encoding='utf-8' ) as vocab_handle: _a : str = json.load(_a ) with open(_a ,encoding='utf-8' ) as vocab_handle: _a : Union[str, Any] = json.load(_a ) _a : Optional[int] = R'[^A-Za-z0-9.,:;!?\-\'\"()\[\] \t\n]+' # In v2, we had a n_vocab=80 and in v3 we missed + and so n_vocab=79 of characters. if len(self.lyrics_encoder ) == 79: _a : Union[str, Any] = oov.replace(R'\-\'' ,R'\-+\'' ) _a : str = regex.compile(_a ) _a : Union[str, Any] = {v: k for k, v in self.artists_encoder.items()} _a : str = {v: k for k, v in self.genres_encoder.items()} _a : Dict = {v: k for k, v in self.lyrics_encoder.items()} @property def __lowercase ( self : Optional[int] ): '''simple docstring''' return len(self.artists_encoder ) + len(self.genres_encoder ) + len(self.lyrics_encoder ) def __lowercase ( self : Dict ): '''simple docstring''' return dict(self.artists_encoder ,self.genres_encoder ,self.lyrics_encoder ) def __lowercase ( self : Optional[int] ,_a : Union[str, Any] ,_a : Dict ,_a : Union[str, Any] ): '''simple docstring''' _a : Tuple = [self.artists_encoder.get(_a ,0 ) for artist in list_artists] for genres in range(len(_a ) ): _a : Dict = [self.genres_encoder.get(_a ,0 ) for genre in list_genres[genres]] _a : Any = list_genres[genres] + [-1] * (self.n_genres - len(list_genres[genres] )) _a : Optional[Any] = [[self.lyrics_encoder.get(_a ,0 ) for character in list_lyrics[0]], [], []] return artists_id, list_genres, lyric_ids def __lowercase ( self : Optional[Any] ,_a : Union[str, Any] ): '''simple docstring''' return list(_a ) def __lowercase ( self : str ,_a : int ,_a : Tuple ,_a : int ,**_a : List[str] ): '''simple docstring''' _a, _a, _a : str = self.prepare_for_tokenization(_a ,_a ,_a ) _a : List[Any] = self._tokenize(_a ) return artist, genre, lyrics def __lowercase ( self : Any ,_a : str ,_a : str ,_a : str ,_a : bool = False ): '''simple docstring''' for idx in range(len(self.version ) ): if self.version[idx] == "v3": _a : Tuple = artists[idx].lower() _a : List[str] = [genres[idx].lower()] else: _a : List[Any] = self._normalize(artists[idx] ) + '.v2' _a : Union[str, Any] = [ self._normalize(_a ) + '.v2' for genre in genres[idx].split('_' ) ] # split is for the full dictionary with combined genres if self.version[0] == "v2": _a : int = regex.compile(R'[^A-Za-z0-9.,:;!?\-\'\"()\[\] \t\n]+' ) _a : List[Any] = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789.,:;!?-+\'\"()[] \t\n' _a : List[str] = {vocab[index]: index + 1 for index in range(len(_a ) )} _a : str = 0 _a : Dict = len(_a ) + 1 _a : Any = self.vocab _a : str = {v: k for k, v in self.vocab.items()} _a : Dict = '' else: _a : str = regex.compile(R'[^A-Za-z0-9.,:;!?\-+\'\"()\[\] \t\n]+' ) _a : str = self._run_strip_accents(_a ) _a : List[Any] = lyrics.replace('\\' ,'\n' ) _a : List[str] = self.out_of_vocab.sub('' ,_a ), [], [] return artists, genres, lyrics def __lowercase ( self : Any ,_a : List[str] ): '''simple docstring''' _a : Dict = unicodedata.normalize('NFD' ,_a ) _a : Union[str, Any] = [] for char in text: _a : Optional[Any] = unicodedata.category(_a ) if cat == "Mn": continue output.append(_a ) return "".join(_a ) def __lowercase ( self : Any ,_a : str ): '''simple docstring''' _a : Optional[Any] = ( [chr(_a ) for i in range(ord('a' ) ,ord('z' ) + 1 )] + [chr(_a ) for i in range(ord('A' ) ,ord('Z' ) + 1 )] + [chr(_a ) for i in range(ord('0' ) ,ord('9' ) + 1 )] + ['.'] ) _a : Any = frozenset(_a ) _a : Optional[int] = re.compile(R'_+' ) _a : Tuple = ''.join([c if c in accepted else '_' for c in text.lower()] ) _a : Any = pattern.sub('_' ,_a ).strip('_' ) return text def __lowercase ( self : Tuple ,_a : List[str] ): '''simple docstring''' return " ".join(_a ) def __lowercase ( self : str ,_a : Optional[int] ,_a : Optional[Union[str, TensorType]] = None ,_a : bool = False ): '''simple docstring''' if not isinstance(_a ,_a ): _a : List[str] = TensorType(_a ) # Get a function reference for the correct framework if tensor_type == TensorType.TENSORFLOW: if not is_tf_available(): raise ImportError( 'Unable to convert output to TensorFlow tensors format, TensorFlow is not installed.' ) import tensorflow as tf _a : Dict = tf.constant _a : str = tf.is_tensor elif tensor_type == TensorType.PYTORCH: if not is_torch_available(): raise ImportError('Unable to convert output to PyTorch tensors format, PyTorch is not installed.' ) import torch _a : Tuple = torch.tensor _a : List[Any] = torch.is_tensor elif tensor_type == TensorType.JAX: if not is_flax_available(): raise ImportError('Unable to convert output to JAX tensors format, JAX is not installed.' ) import jax.numpy as jnp # noqa: F811 _a : Optional[Any] = jnp.array _a : Optional[Any] = _is_jax else: _a : Dict = np.asarray _a : Optional[int] = _is_numpy # Do the tensor conversion in batch try: if prepend_batch_axis: _a : Union[str, Any] = [inputs] if not is_tensor(_a ): _a : str = as_tensor(_a ) except: # noqa E722 raise ValueError( 'Unable to create tensor, you should probably activate truncation and/or padding ' 'with \'padding=True\' \'truncation=True\' to have batched tensors with the same length.' ) return inputs def __call__( self : int ,_a : int ,_a : Optional[int] ,_a : Dict="" ,_a : Any="pt" ): '''simple docstring''' _a : Optional[Any] = [0, 0, 0] _a : Dict = [artist] * len(self.version ) _a : List[Any] = [genres] * len(self.version ) _a, _a, _a : int = self.tokenize(_a ,_a ,_a ) _a, _a, _a : Dict = self._convert_token_to_id(_a ,_a ,_a ) _a : Optional[Any] = [-INFINITY] * len(full_tokens[-1] ) _a : List[Any] = [ self.convert_to_tensors( [input_ids + [artists_id[i]] + genres_ids[i] + full_tokens[i]] ,tensor_type=_a ) for i in range(len(self.version ) ) ] return BatchEncoding({'input_ids': input_ids, 'attention_masks': attention_masks} ) def __lowercase ( self : int ,_a : str ,_a : Optional[str] = None ): '''simple docstring''' if not os.path.isdir(_a ): logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" ) return _a : Union[str, Any] = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['artists_file'] ) with open(_a ,'w' ,encoding='utf-8' ) as f: f.write(json.dumps(self.artists_encoder ,ensure_ascii=_a ) ) _a : List[str] = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['genres_file'] ) with open(_a ,'w' ,encoding='utf-8' ) as f: f.write(json.dumps(self.genres_encoder ,ensure_ascii=_a ) ) _a : List[str] = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['lyrics_file'] ) with open(_a ,'w' ,encoding='utf-8' ) as f: f.write(json.dumps(self.lyrics_encoder ,ensure_ascii=_a ) ) return (artists_file, genres_file, lyrics_file) def __lowercase ( self : str ,_a : Dict ,_a : Dict ,_a : Optional[int] ): '''simple docstring''' _a : List[Any] = self.artists_decoder.get(_a ) _a : List[str] = [self.genres_decoder.get(_a ) for genre in genres_index] _a : Union[str, Any] = [self.lyrics_decoder.get(_a ) for character in lyric_index] return artist, genres, lyrics
271
'''simple docstring''' from typing import Any, Dict, List, Optional, Tuple, Union import torch from torch import nn from torch.utils.data import DistributedSampler, RandomSampler from transformers import PreTrainedModel, Trainer, logging from transformers.integrations import is_fairscale_available from transformers.models.fsmt.configuration_fsmt import FSMTConfig from transformers.optimization import ( Adafactor, AdamW, get_constant_schedule, get_constant_schedule_with_warmup, get_cosine_schedule_with_warmup, get_cosine_with_hard_restarts_schedule_with_warmup, get_linear_schedule_with_warmup, get_polynomial_decay_schedule_with_warmup, ) from transformers.trainer_pt_utils import get_tpu_sampler from transformers.training_args import ParallelMode from transformers.utils import is_torch_tpu_available if is_fairscale_available(): from fairscale.optim import OSS __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = { """linear""": get_linear_schedule_with_warmup, """cosine""": get_cosine_schedule_with_warmup, """cosine_w_restarts""": get_cosine_with_hard_restarts_schedule_with_warmup, """polynomial""": get_polynomial_decay_schedule_with_warmup, """constant""": get_constant_schedule, """constant_w_warmup""": get_constant_schedule_with_warmup, } class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Optional[int] ,_a : Optional[Any]=None ,_a : Dict=None ,*_a : int ,**_a : str ): '''simple docstring''' super().__init__(*_a ,**_a ) if config is None: assert isinstance(self.model ,_a ), ( "If no `config` is passed the model to be trained has to be of type `PreTrainedModel`, but is" F""" {self.model.__class__}""" ) _a : List[Any] = self.model.config else: _a : Optional[int] = config _a : List[str] = data_args _a : List[Any] = self.config.tgt_vocab_size if isinstance(self.config ,_a ) else self.config.vocab_size if self.args.label_smoothing != 0 or (self.data_args is not None and self.data_args.ignore_pad_token_for_loss): assert self.config.pad_token_id is not None, ( "Make sure that `config.pad_token_id` is correcly defined when ignoring `pad_token` for loss" " calculation or doing label smoothing." ) if self.config.pad_token_id is None and self.config.eos_token_id is not None: logger.warning( F"""The `config.pad_token_id` is `None`. Using `config.eos_token_id` = {self.config.eos_token_id} for""" ' padding..' ) if self.args.label_smoothing == 0: _a : List[str] = torch.nn.CrossEntropyLoss(ignore_index=self.config.pad_token_id ) else: # dynamically import label_smoothed_nll_loss from utils import label_smoothed_nll_loss _a : Tuple = label_smoothed_nll_loss def __lowercase ( self : List[str] ,_a : int ): '''simple docstring''' if self.optimizer is None: _a : Union[str, Any] = ['bias', 'LayerNorm.weight'] _a : Tuple = [ { 'params': [p for n, p in self.model.named_parameters() if not any(nd in n for nd in no_decay )], 'weight_decay': self.args.weight_decay, }, { 'params': [p for n, p in self.model.named_parameters() if any(nd in n for nd in no_decay )], 'weight_decay': 0.0, }, ] _a : Optional[int] = Adafactor if self.args.adafactor else AdamW if self.args.adafactor: _a : Any = Adafactor _a : Dict = {'scale_parameter': False, 'relative_step': False} else: _a : Union[str, Any] = AdamW _a : str = { 'betas': (self.args.adam_betaa, self.args.adam_betaa), 'eps': self.args.adam_epsilon, } _a : Union[str, Any] = self.args.learning_rate if self.sharded_ddp: _a : str = OSS( params=_a ,optim=_a ,**_a ,) else: _a : Tuple = optimizer_cls(_a ,**_a ) if self.lr_scheduler is None: _a : List[Any] = self._get_lr_scheduler(_a ) else: # ignoring --lr_scheduler logger.warning('scheduler is passed to `Seq2SeqTrainer`, `--lr_scheduler` arg is ignored.' ) def __lowercase ( self : List[Any] ,_a : List[Any] ): '''simple docstring''' _a : str = arg_to_scheduler[self.args.lr_scheduler] if self.args.lr_scheduler == "constant": _a : int = schedule_func(self.optimizer ) elif self.args.lr_scheduler == "constant_w_warmup": _a : List[str] = schedule_func(self.optimizer ,num_warmup_steps=self.args.warmup_steps ) else: _a : Optional[int] = schedule_func( self.optimizer ,num_warmup_steps=self.args.warmup_steps ,num_training_steps=_a ) return scheduler def __lowercase ( self : Tuple ): '''simple docstring''' if isinstance(self.train_dataset ,torch.utils.data.IterableDataset ): return None elif is_torch_tpu_available(): return get_tpu_sampler(self.train_dataset ) else: if self.args.sortish_sampler: self.train_dataset.make_sortish_sampler( self.args.per_device_train_batch_size ,distributed=(self.args.parallel_mode == ParallelMode.DISTRIBUTED) ,) return ( RandomSampler(self.train_dataset ) if self.args.local_rank == -1 else DistributedSampler(self.train_dataset ) ) def __lowercase ( self : Dict ,_a : Dict ,_a : Any ,_a : Dict ): '''simple docstring''' if self.args.label_smoothing == 0: if self.data_args is not None and self.data_args.ignore_pad_token_for_loss: # force training to ignore pad token _a : List[Any] = model(**_a ,use_cache=_a )[0] _a : Union[str, Any] = self.loss_fn(logits.view(-1 ,logits.shape[-1] ) ,labels.view(-1 ) ) else: # compute usual loss via models _a, _a : Union[str, Any] = model(**_a ,labels=_a ,use_cache=_a )[:2] else: # compute label smoothed loss _a : List[Any] = model(**_a ,use_cache=_a )[0] _a : Any = torch.nn.functional.log_softmax(_a ,dim=-1 ) _a, _a : List[str] = self.loss_fn(_a ,_a ,self.args.label_smoothing ,ignore_index=self.config.pad_token_id ) return loss, logits def __lowercase ( self : Optional[int] ,_a : Union[str, Any] ,_a : List[Any] ): '''simple docstring''' _a : Optional[int] = inputs.pop('labels' ) _a, _a : int = self._compute_loss(_a ,_a ,_a ) return loss def __lowercase ( self : Optional[Any] ,_a : nn.Module ,_a : Dict[str, Union[torch.Tensor, Any]] ,_a : bool ,_a : Optional[List[str]] = None ,): '''simple docstring''' _a : int = self._prepare_inputs(_a ) _a : Any = { 'max_length': self.data_args.val_max_target_length if self.data_args is not None else self.config.max_length, 'num_beams': self.data_args.eval_beams if self.data_args is not None else self.config.num_beams, } if self.args.predict_with_generate and not self.args.prediction_loss_only: _a : int = self.model.generate( inputs['input_ids'] ,attention_mask=inputs['attention_mask'] ,**_a ,) # in case the batch is shorter than max length, the output should be padded if generated_tokens.shape[-1] < gen_kwargs["max_length"]: _a : int = self._pad_tensors_to_max_len(_a ,gen_kwargs['max_length'] ) _a : Union[str, Any] = inputs.pop('labels' ) with torch.no_grad(): # compute loss on predict data _a, _a : Optional[int] = self._compute_loss(_a ,_a ,_a ) _a : Optional[Any] = loss.mean().detach() if self.args.prediction_loss_only: return (loss, None, None) _a : Optional[Any] = generated_tokens if self.args.predict_with_generate else logits if labels.shape[-1] < gen_kwargs["max_length"]: _a : Dict = self._pad_tensors_to_max_len(_a ,gen_kwargs['max_length'] ) return (loss, logits, labels) def __lowercase ( self : str ,_a : Tuple ,_a : Tuple ): '''simple docstring''' _a : List[Any] = self.config.pad_token_id if self.config.pad_token_id is not None else self.config.eos_token_id if pad_token_id is None: raise ValueError( 'Make sure that either `config.pad_token_id` or `config.eos_token_id` is defined if tensor has to be' F""" padded to `max_length`={max_length}""" ) _a : int = pad_token_id * torch.ones( (tensor.shape[0], max_length) ,dtype=tensor.dtype ,device=tensor.device ) _a : Union[str, Any] = tensor return padded_tensor
271
1
'''simple docstring''' def UpperCAmelCase_ (__a : int = 1_0_0 ): """simple docstring""" _a : List[Any] = n * (n + 1) * (2 * n + 1) / 6 _a : int = (n * (n + 1) / 2) ** 2 return int(square_of_sum - sum_of_squares ) if __name__ == "__main__": print(f'''{solution() = }''')
271
'''simple docstring''' import gzip import hashlib import json import multiprocessing import os import re import shutil import time from pathlib import Path import numpy as np from arguments import PreprocessingArguments from datasets import load_dataset from minhash_deduplication import deduplicate_dataset from transformers import AutoTokenizer, HfArgumentParser __lowerCAmelCase = re.compile(r"""\s+""") def UpperCAmelCase_ (__a : Any ): """simple docstring""" return {"hash": hashlib.mda(re.sub(__a , '' , example['content'] ).encode('utf-8' ) ).hexdigest()} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : List[str] = [len(__a ) for line in example['content'].splitlines()] return {"line_mean": np.mean(__a ), "line_max": max(__a )} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Union[str, Any] = np.mean([c.isalnum() for c in example['content']] ) return {"alpha_frac": alpha_frac} def UpperCAmelCase_ (__a : Optional[int] , __a : Any ): """simple docstring""" if example["hash"] in uniques: uniques.remove(example['hash'] ) return True else: return False def UpperCAmelCase_ (__a : int , __a : Union[str, Any]=5 ): """simple docstring""" _a : Optional[int] = ['auto-generated', 'autogenerated', 'automatically generated'] _a : List[str] = example['content'].splitlines() for _, line in zip(range(__a ) , __a ): for keyword in keywords: if keyword in line.lower(): return {"autogenerated": True} else: return {"autogenerated": False} def UpperCAmelCase_ (__a : List[str] , __a : Dict=5 , __a : Tuple=0.05 ): """simple docstring""" _a : Optional[int] = ['unit tests', 'test file', 'configuration file'] _a : int = example['content'].splitlines() _a : int = 0 _a : Dict = 0 # first test for _, line in zip(range(__a ) , __a ): for keyword in keywords: if keyword in line.lower(): return {"config_or_test": True} # second test _a : int = example['content'].count('\n' ) _a : int = int(coeff * nlines ) for line in lines: count_config += line.lower().count('config' ) count_test += line.lower().count('test' ) if count_config > threshold or count_test > threshold: return {"config_or_test": True} return {"config_or_test": False} def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" _a : List[str] = ['def ', 'class ', 'for ', 'while '] _a : str = example['content'].splitlines() for line in lines: for keyword in keywords: if keyword in line.lower(): return {"has_no_keywords": False} return {"has_no_keywords": True} def UpperCAmelCase_ (__a : int , __a : Any=4 ): """simple docstring""" _a : List[str] = example['content'].splitlines() _a : Dict = 0 for line in lines: counter += line.lower().count('=' ) if counter > minimum: return {"has_few_assignments": False} return {"has_few_assignments": True} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Optional[Any] = tokenizer(example['content'] , truncation=__a )['input_ids'] _a : Optional[int] = len(example['content'] ) / len(__a ) return {"ratio": ratio} def UpperCAmelCase_ (__a : str ): """simple docstring""" _a : Dict = {} results.update(get_hash(__a ) ) results.update(line_stats(__a ) ) results.update(alpha_stats(__a ) ) results.update(char_token_ratio(__a ) ) results.update(is_autogenerated(__a ) ) results.update(is_config_or_test(__a ) ) results.update(has_no_keywords(__a ) ) results.update(has_few_assignments(__a ) ) return results def UpperCAmelCase_ (__a : Any , __a : Any , __a : str ): """simple docstring""" if not check_uniques(__a , __a ): return False elif example["autogenerated"]: return False elif example["line_max"] > args.line_max: return False elif example["line_mean"] > args.line_mean: return False elif example["alpha_frac"] < args.alpha_frac: return False elif example["ratio"] < args.min_token_ratio: return False elif example["config_or_test"] and np.random.rand() <= args.filter_proba: return False elif example["has_no_keywords"] and np.random.rand() <= args.filter_proba: return False elif example["has_few_assignments"]: return False else: return True def UpperCAmelCase_ (__a : Union[str, Any] ): """simple docstring""" with open(__a , 'rb' ) as f_in: with gzip.open(str(__a ) + '.gz' , 'wb' , compresslevel=6 ) as f_out: shutil.copyfileobj(__a , __a ) os.unlink(__a ) # Settings __lowerCAmelCase = HfArgumentParser(PreprocessingArguments) __lowerCAmelCase = parser.parse_args() if args.num_workers is None: __lowerCAmelCase = multiprocessing.cpu_count() __lowerCAmelCase = AutoTokenizer.from_pretrained(args.tokenizer_dir) # Load dataset __lowerCAmelCase = time.time() __lowerCAmelCase = load_dataset(args.dataset_name, split="""train""") print(f'''Time to load dataset: {time.time()-t_start:.2f}''') # Run preprocessing __lowerCAmelCase = time.time() __lowerCAmelCase = ds.map(preprocess, num_proc=args.num_workers) print(f'''Time to preprocess dataset: {time.time()-t_start:.2f}''') # Deduplicate hashes __lowerCAmelCase = set(ds.unique("""hash""")) __lowerCAmelCase = len(uniques) / len(ds) print(f'''Fraction of duplicates: {1-frac:.2%}''') # Deduplicate data and apply heuristics __lowerCAmelCase = time.time() __lowerCAmelCase = ds.filter(filter, fn_kwargs={"""uniques""": uniques, """args""": args}) print(f'''Time to filter dataset: {time.time()-t_start:.2f}''') print(f'''Size of filtered dataset: {len(ds_filter)}''') # Deduplicate with minhash and jaccard similarity if args.near_deduplication: __lowerCAmelCase = time.time() __lowerCAmelCase , __lowerCAmelCase = deduplicate_dataset(ds_filter, args.jaccard_threshold) print(f'''Time to deduplicate dataset: {time.time()-t_start:.2f}''') print(f'''Size of deduplicate dataset: {len(ds_filter)}''') # Save data in batches of samples_per_file __lowerCAmelCase = Path(args.output_dir) output_dir.mkdir(exist_ok=True) # save duplicate_clusters in the output_dir as artifacts # not sure it is the right place the save it if args.near_deduplication: with open(output_dir / """duplicate_clusters.json""", """w""") as f: json.dump(duplicate_clusters, f) __lowerCAmelCase = output_dir / """data""" data_dir.mkdir(exist_ok=True) __lowerCAmelCase = time.time() for file_number, index in enumerate(range(0, len(ds_filter), args.samples_per_file)): __lowerCAmelCase = str(data_dir / f'''file-{file_number+1:012}.json''') __lowerCAmelCase = min(len(ds_filter), index + args.samples_per_file) ds_filter.select(list(range(index, end_index))).to_json(file_path) compress_file(file_path) print(f'''Time to save dataset: {time.time()-t_start:.2f}''')
271
1
'''simple docstring''' import io import json import unittest from parameterized import parameterized from transformers import FSMTForConditionalGeneration, FSMTTokenizer from transformers.testing_utils import get_tests_dir, require_torch, slow, torch_device from utils import calculate_bleu __lowerCAmelCase = get_tests_dir() + """/test_data/fsmt/fsmt_val_data.json""" with io.open(filename, """r""", encoding="""utf-8""") as f: __lowerCAmelCase = json.load(f) @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : List[str] ,_a : Dict ): '''simple docstring''' return FSMTTokenizer.from_pretrained(_a ) def __lowercase ( self : int ,_a : List[str] ): '''simple docstring''' _a : List[str] = FSMTForConditionalGeneration.from_pretrained(_a ).to(_a ) if torch_device == "cuda": model.half() return model @parameterized.expand( [ ['en-ru', 26.0], ['ru-en', 22.0], ['en-de', 22.0], ['de-en', 29.0], ] ) @slow def __lowercase ( self : int ,_a : Union[str, Any] ,_a : Union[str, Any] ): '''simple docstring''' _a : int = F"""facebook/wmt19-{pair}""" _a : Any = self.get_tokenizer(_a ) _a : Any = self.get_model(_a ) _a : Union[str, Any] = bleu_data[pair]['src'] _a : Optional[int] = bleu_data[pair]['tgt'] _a : Any = tokenizer(_a ,return_tensors='pt' ,truncation=_a ,padding='longest' ).to(_a ) _a : int = model.generate( input_ids=batch.input_ids ,num_beams=8 ,) _a : Optional[int] = tokenizer.batch_decode( _a ,skip_special_tokens=_a ,clean_up_tokenization_spaces=_a ) _a : Optional[Any] = calculate_bleu(_a ,_a ) print(_a ) self.assertGreaterEqual(scores['bleu'] ,_a )
271
'''simple docstring''' import argparse from typing import List import evaluate import numpy as np import torch from datasets import DatasetDict, load_dataset # New Code # # We'll be using StratifiedKFold for this example from sklearn.model_selection import StratifiedKFold from torch.optim import AdamW from torch.utils.data import DataLoader from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed from accelerate import Accelerator, DistributedType ######################################################################## # This is a fully working simple example to use Accelerate, # specifically showcasing how to perform Cross Validation, # and builds off the `nlp_example.py` script. # # This example trains a Bert base model on GLUE MRPC # in any of the following settings (with the same script): # - single CPU or single GPU # - multi GPUS (using PyTorch distributed mode) # - (multi) TPUs # - fp16 (mixed-precision) or fp32 (normal precision) # # To help focus on the differences in the code, building `DataLoaders` # was refactored into its own function. # New additions from the base script can be found quickly by # looking for the # New Code # tags # # To run it in each of these various modes, follow the instructions # in the readme for examples: # https://github.com/huggingface/accelerate/tree/main/examples # ######################################################################## __lowerCAmelCase = 1_6 __lowerCAmelCase = 3_2 def UpperCAmelCase_ (__a : Accelerator , __a : DatasetDict , __a : List[int] , __a : List[int] , __a : int = 1_6 ): """simple docstring""" _a : Union[str, Any] = AutoTokenizer.from_pretrained('bert-base-cased' ) _a : str = DatasetDict( { 'train': dataset['train'].select(__a ), 'validation': dataset['train'].select(__a ), 'test': dataset['validation'], } ) def tokenize_function(__a : List[Any] ): # max_length=None => use the model max length (it's actually the default) _a : Optional[int] = tokenizer(examples['sentence1'] , examples['sentence2'] , truncation=__a , max_length=__a ) return outputs # Apply the method we just defined to all the examples in all the splits of the dataset # starting with the main process first: with accelerator.main_process_first(): _a : List[str] = datasets.map( __a , batched=__a , remove_columns=['idx', 'sentence1', 'sentence2'] , ) # We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the # transformers library _a : List[Any] = tokenized_datasets.rename_column('label' , 'labels' ) def collate_fn(__a : int ): # On TPU it's best to pad everything to the same length or training will be very slow. _a : Dict = 1_2_8 if accelerator.distributed_type == DistributedType.TPU else None # When using mixed precision we want round multiples of 8/16 if accelerator.mixed_precision == "fp8": _a : Tuple = 1_6 elif accelerator.mixed_precision != "no": _a : List[Any] = 8 else: _a : List[Any] = None return tokenizer.pad( __a , padding='longest' , max_length=__a , pad_to_multiple_of=__a , return_tensors='pt' , ) # Instantiate dataloaders. _a : Any = DataLoader( tokenized_datasets['train'] , shuffle=__a , collate_fn=__a , batch_size=__a ) _a : Optional[int] = DataLoader( tokenized_datasets['validation'] , shuffle=__a , collate_fn=__a , batch_size=__a ) _a : Optional[Any] = DataLoader( tokenized_datasets['test'] , shuffle=__a , collate_fn=__a , batch_size=__a ) return train_dataloader, eval_dataloader, test_dataloader def UpperCAmelCase_ (__a : Any , __a : Union[str, Any] ): """simple docstring""" _a : Dict = [] # Download the dataset _a : Tuple = load_dataset('glue' , 'mrpc' ) # Create our splits _a : Union[str, Any] = StratifiedKFold(n_splits=int(args.num_folds ) ) # Initialize accelerator _a : Any = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision ) # Sample hyper-parameters for learning rate, batch size, seed and a few other HPs _a : Optional[Any] = config['lr'] _a : Optional[int] = int(config['num_epochs'] ) _a : Dict = int(config['seed'] ) _a : Dict = int(config['batch_size'] ) _a : Optional[int] = evaluate.load('glue' , 'mrpc' ) # If the batch size is too big we use gradient accumulation _a : List[Any] = 1 if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU: _a : Any = batch_size // MAX_GPU_BATCH_SIZE _a : List[str] = MAX_GPU_BATCH_SIZE set_seed(__a ) # New Code # # Create our folds: _a : int = kfold.split(np.zeros(datasets['train'].num_rows ) , datasets['train']['label'] ) _a : Any = [] # Iterate over them for i, (train_idxs, valid_idxs) in enumerate(__a ): _a, _a, _a : Optional[Any] = get_fold_dataloaders( __a , __a , __a , __a , ) # Instantiate the model (we build the model here so that the seed also control new weights initialization) _a : Dict = AutoModelForSequenceClassification.from_pretrained('bert-base-cased' , return_dict=__a ) # We could avoid this line since the accelerator is set with `device_placement=True` (default value). # Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer # creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that). _a : List[Any] = model.to(accelerator.device ) # Instantiate optimizer _a : List[str] = AdamW(params=model.parameters() , lr=__a ) # Instantiate scheduler _a : List[Any] = get_linear_schedule_with_warmup( optimizer=__a , num_warmup_steps=1_0_0 , num_training_steps=(len(__a ) * num_epochs) // gradient_accumulation_steps , ) # Prepare everything # There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the # prepare method. _a, _a, _a, _a, _a : Union[str, Any] = accelerator.prepare( __a , __a , __a , __a , __a ) # Now we train the model for epoch in range(__a ): model.train() for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) _a : Dict = model(**__a ) _a : int = outputs.loss _a : Any = loss / gradient_accumulation_steps accelerator.backward(__a ) if step % gradient_accumulation_steps == 0: optimizer.step() lr_scheduler.step() optimizer.zero_grad() model.eval() for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): _a : Union[str, Any] = model(**__a ) _a : Tuple = outputs.logits.argmax(dim=-1 ) _a, _a : Any = accelerator.gather_for_metrics((predictions, batch['labels']) ) metric.add_batch( predictions=__a , references=__a , ) _a : List[Any] = metric.compute() # Use accelerator.print to print only on the main process. accelerator.print(f"""epoch {epoch}:""" , __a ) # New Code # # We also run predictions on the test set at the very end _a : Any = [] for step, batch in enumerate(__a ): # We could avoid this line since we set the accelerator with `device_placement=True`. batch.to(accelerator.device ) with torch.no_grad(): _a : Tuple = model(**__a ) _a : Dict = outputs.logits _a, _a : Optional[int] = accelerator.gather_for_metrics((predictions, batch['labels']) ) fold_predictions.append(predictions.cpu() ) if i == 0: # We need all of the test predictions test_references.append(references.cpu() ) # Use accelerator.print to print only on the main process. test_predictions.append(torch.cat(__a , dim=0 ) ) # We now need to release all our memory and get rid of the current model, optimizer, etc accelerator.free_memory() # New Code # # Finally we check the accuracy of our folded results: _a : Dict = torch.cat(__a , dim=0 ) _a : Any = torch.stack(__a , dim=0 ).sum(dim=0 ).div(int(args.num_folds ) ).argmax(dim=-1 ) _a : str = metric.compute(predictions=__a , references=__a ) accelerator.print('Average test metrics from all folds:' , __a ) def UpperCAmelCase_ (): """simple docstring""" _a : Any = argparse.ArgumentParser(description='Simple example of training script.' ) parser.add_argument( '--mixed_precision' , type=__a , default=__a , choices=['no', 'fp16', 'bf16', 'fp8'] , help='Whether to use mixed precision. Choose' 'between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10.' 'and an Nvidia Ampere GPU.' , ) parser.add_argument('--cpu' , action='store_true' , help='If passed, will train on the CPU.' ) # New Code # parser.add_argument('--num_folds' , type=__a , default=3 , help='The number of splits to perform across the dataset' ) _a : Any = parser.parse_args() _a : int = {'lr': 2e-5, 'num_epochs': 3, 'seed': 4_2, 'batch_size': 1_6} training_function(__a , __a ) if __name__ == "__main__": main()
271
1
'''simple docstring''' from __future__ import annotations from random import choice def UpperCAmelCase_ (__a : str ): """simple docstring""" return choice(__a ) def UpperCAmelCase_ (__a : list[int] , __a : int ): """simple docstring""" _a : Dict = random_pivot(__a ) # partition based on pivot # linear time _a : Optional[int] = [e for e in lst if e < pivot] _a : List[str] = [e for e in lst if e > pivot] # if we get lucky, pivot might be the element we want. # we can easily see this: # small (elements smaller than k) # + pivot (kth element) # + big (elements larger than k) if len(__a ) == k - 1: return pivot # pivot is in elements bigger than k elif len(__a ) < k - 1: return kth_number(__a , k - len(__a ) - 1 ) # pivot is in elements smaller than k else: return kth_number(__a , __a ) if __name__ == "__main__": import doctest doctest.testmod()
271
'''simple docstring''' from __future__ import annotations __lowerCAmelCase = [-1_0, -5, 0, 5, 5.1, 1_1, 1_3, 2_1, 3, 4, -2_1, -1_0, -5, -1, 0] __lowerCAmelCase = [-5, 0, 5, 5.1, 1_1, 1_3, 2_1, -1, 4, -1, -1_0, -5, -1, 0, -1] def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : Optional[int] = [] _a : int = len(__a ) for i in range(__a ): _a : float = -1 for j in range(i + 1 , __a ): if arr[i] < arr[j]: _a : Any = arr[j] break result.append(__a ) return result def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : Tuple = [] for i, outer in enumerate(__a ): _a : float = -1 for inner in arr[i + 1 :]: if outer < inner: _a : Dict = inner break result.append(__a ) return result def UpperCAmelCase_ (__a : list[float] ): """simple docstring""" _a : int = len(__a ) _a : list[float] = [] _a : list[float] = [-1] * arr_size for index in reversed(range(__a ) ): if stack: while stack[-1] <= arr[index]: stack.pop() if not stack: break if stack: _a : Dict = stack[-1] stack.append(arr[index] ) return result if __name__ == "__main__": from doctest import testmod from timeit import timeit testmod() print(next_greatest_element_slow(arr)) print(next_greatest_element_fast(arr)) print(next_greatest_element(arr)) __lowerCAmelCase = ( """from __main__ import arr, next_greatest_element_slow, """ """next_greatest_element_fast, next_greatest_element""" ) print( """next_greatest_element_slow():""", timeit("""next_greatest_element_slow(arr)""", setup=setup), ) print( """next_greatest_element_fast():""", timeit("""next_greatest_element_fast(arr)""", setup=setup), ) print( """ next_greatest_element():""", timeit("""next_greatest_element(arr)""", setup=setup), )
271
1
'''simple docstring''' import unittest from transformers import ( MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING, TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING, TextaTextGenerationPipeline, pipeline, ) from transformers.testing_utils import is_pipeline_test, require_tf, require_torch from transformers.utils import is_torch_available from .test_pipelines_common import ANY if is_torch_available(): import torch @is_pipeline_test class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" __UpperCAmelCase : int = MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING __UpperCAmelCase : int = TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING def __lowercase ( self : Any ,_a : Any ,_a : List[str] ,_a : str ): '''simple docstring''' _a : Optional[Any] = TextaTextGenerationPipeline(model=_a ,tokenizer=_a ) return generator, ["Something to write", "Something else"] def __lowercase ( self : int ,_a : int ,_a : Optional[Any] ): '''simple docstring''' _a : Dict = generator('Something there' ) self.assertEqual(_a ,[{'generated_text': ANY(_a )}] ) # These are encoder decoder, they don't just append to incoming string self.assertFalse(outputs[0]['generated_text'].startswith('Something there' ) ) _a : List[str] = generator(['This is great !', 'Something else'] ,num_return_sequences=2 ,do_sample=_a ) self.assertEqual( _a ,[ [{'generated_text': ANY(_a )}, {'generated_text': ANY(_a )}], [{'generated_text': ANY(_a )}, {'generated_text': ANY(_a )}], ] ,) _a : Dict = generator( ['This is great !', 'Something else'] ,num_return_sequences=2 ,batch_size=2 ,do_sample=_a ) self.assertEqual( _a ,[ [{'generated_text': ANY(_a )}, {'generated_text': ANY(_a )}], [{'generated_text': ANY(_a )}, {'generated_text': ANY(_a )}], ] ,) with self.assertRaises(_a ): generator(4 ) @require_torch def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Dict = pipeline('text2text-generation' ,model='patrickvonplaten/t5-tiny-random' ,framework='pt' ) # do_sample=False necessary for reproducibility _a : int = generator('Something there' ,do_sample=_a ) self.assertEqual(_a ,[{'generated_text': ''}] ) _a : List[str] = 3 _a : int = generator( 'Something there' ,num_return_sequences=_a ,num_beams=_a ,) _a : Optional[int] = [ {'generated_text': 'Beide Beide Beide Beide Beide Beide Beide Beide Beide'}, {'generated_text': 'Beide Beide Beide Beide Beide Beide Beide Beide'}, {'generated_text': ''}, ] self.assertEqual(_a ,_a ) _a : Any = generator('This is a test' ,do_sample=_a ,num_return_sequences=2 ,return_tensors=_a ) self.assertEqual( _a ,[ {'generated_token_ids': ANY(torch.Tensor )}, {'generated_token_ids': ANY(torch.Tensor )}, ] ,) _a : List[Any] = generator.model.config.eos_token_id _a : Optional[Any] = '<pad>' _a : List[Any] = generator( ['This is a test', 'This is a second test'] ,do_sample=_a ,num_return_sequences=2 ,batch_size=2 ,return_tensors=_a ,) self.assertEqual( _a ,[ [ {'generated_token_ids': ANY(torch.Tensor )}, {'generated_token_ids': ANY(torch.Tensor )}, ], [ {'generated_token_ids': ANY(torch.Tensor )}, {'generated_token_ids': ANY(torch.Tensor )}, ], ] ,) @require_tf def __lowercase ( self : List[str] ): '''simple docstring''' _a : Optional[int] = pipeline('text2text-generation' ,model='patrickvonplaten/t5-tiny-random' ,framework='tf' ) # do_sample=False necessary for reproducibility _a : str = generator('Something there' ,do_sample=_a ) self.assertEqual(_a ,[{'generated_text': ''}] )
271
'''simple docstring''' import os from huggingface_hub.constants import HUGGINGFACE_HUB_CACHE, hf_cache_home __lowerCAmelCase = HUGGINGFACE_HUB_CACHE __lowerCAmelCase = """config.json""" __lowerCAmelCase = """diffusion_pytorch_model.bin""" __lowerCAmelCase = """diffusion_flax_model.msgpack""" __lowerCAmelCase = """model.onnx""" __lowerCAmelCase = """diffusion_pytorch_model.safetensors""" __lowerCAmelCase = """weights.pb""" __lowerCAmelCase = """https://huggingface.co""" __lowerCAmelCase = default_cache_path __lowerCAmelCase = """diffusers_modules""" __lowerCAmelCase = os.getenv("""HF_MODULES_CACHE""", os.path.join(hf_cache_home, """modules""")) __lowerCAmelCase = ["""fp16""", """non-ema"""] __lowerCAmelCase = """.self_attn"""
271
1
'''simple docstring''' import json import os import subprocess import unittest from ast import literal_eval import pytest from parameterized import parameterized, parameterized_class from . import is_sagemaker_available if is_sagemaker_available(): from sagemaker import Session, TrainingJobAnalytics from sagemaker.huggingface import HuggingFace @pytest.mark.skipif( literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , ) @pytest.mark.usefixtures('''sm_env''' ) @parameterized_class( [ { '''framework''': '''pytorch''', '''script''': '''run_glue_model_parallelism.py''', '''model_name_or_path''': '''roberta-large''', '''instance_type''': '''ml.p3dn.24xlarge''', '''results''': {'''train_runtime''': 1600, '''eval_accuracy''': 0.3, '''eval_loss''': 1.2}, }, { '''framework''': '''pytorch''', '''script''': '''run_glue.py''', '''model_name_or_path''': '''roberta-large''', '''instance_type''': '''ml.p3dn.24xlarge''', '''results''': {'''train_runtime''': 1600, '''eval_accuracy''': 0.3, '''eval_loss''': 1.2}, }, ] ) class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Optional[int] ): '''simple docstring''' if self.framework == "pytorch": subprocess.run( F"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() ,encoding='utf-8' ,check=_a ,) assert hasattr(self ,'env' ) def __lowercase ( self : int ,_a : int ): '''simple docstring''' _a : Dict = { 'enabled': True, 'processes_per_host': 8, } _a : Union[str, Any] = { 'enabled': True, 'parameters': { 'microbatches': 4, 'placement_strategy': 'spread', 'pipeline': 'interleaved', 'optimize': 'speed', 'partitions': 4, 'ddp': True, }, } _a : List[str] = {'smdistributed': {'modelparallel': smp_options}, 'mpi': mpi_options} _a : List[str] = 'trainer' if self.script == 'run_glue.py' else 'smtrainer' # creates estimator return HuggingFace( entry_point=self.script ,source_dir=self.env.test_path ,role=self.env.role ,image_uri=self.env.image_uri ,base_job_name=F"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" ,instance_count=_a ,instance_type=self.instance_type ,debugger_hook_config=_a ,hyperparameters={ **self.env.hyperparameters, 'model_name_or_path': self.model_name_or_path, 'max_steps': 500, } ,metric_definitions=self.env.metric_definitions ,distribution=_a ,py_version='py36' ,) def __lowercase ( self : int ,_a : List[str] ): '''simple docstring''' TrainingJobAnalytics(_a ).export_csv(F"""{self.env.test_path}/{job_name}_metrics.csv""" ) @parameterized.expand([(1,)] ) def __lowercase ( self : Union[str, Any] ,_a : List[str] ): '''simple docstring''' _a : str = self.create_estimator(_a ) # run training estimator.fit() # result dataframe _a : Union[str, Any] = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe() # extract kpis _a : Dict = list(result_metrics_df[result_metrics_df.metric_name == 'eval_accuracy']['value'] ) _a : List[str] = list(result_metrics_df[result_metrics_df.metric_name == 'eval_loss']['value'] ) # get train time from SageMaker job, this includes starting, preprocessing, stopping _a : List[str] = ( Session().describe_training_job(estimator.latest_training_job.name ).get('TrainingTimeInSeconds' ,99_9999 ) ) # assert kpis assert train_runtime <= self.results["train_runtime"] assert all(t >= self.results['eval_accuracy'] for t in eval_accuracy ) assert all(t <= self.results['eval_loss'] for t in eval_loss ) # dump tests result into json file to share in PR with open(F"""{estimator.latest_training_job.name}.json""" ,'w' ) as outfile: json.dump({'train_time': train_runtime, 'eval_accuracy': eval_accuracy, 'eval_loss': eval_loss} ,_a )
271
'''simple docstring''' import inspect import unittest import numpy as np from tests.test_modeling_common import floats_tensor from transformers import MaskaFormerConfig, is_torch_available, is_vision_available from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device from transformers.utils import cached_property from ...test_configuration_common import ConfigTester from ...test_modeling_common import ModelTesterMixin from ...test_pipeline_mixin import PipelineTesterMixin if is_torch_available(): import torch from transformers import MaskaFormerForUniversalSegmentation, MaskaFormerModel if is_vision_available(): from transformers import MaskaFormerImageProcessor if is_vision_available(): from PIL import Image class UpperCAmelCase__ : """simple docstring""" def __init__( self : int ,_a : Any ,_a : Optional[int]=2 ,_a : Optional[Any]=True ,_a : Dict=False ,_a : Dict=10 ,_a : Any=3 ,_a : str=32 * 8 ,_a : Optional[int]=32 * 8 ,_a : int=4 ,_a : str=64 ,): '''simple docstring''' _a : Dict = parent _a : Union[str, Any] = batch_size _a : Tuple = is_training _a : List[str] = use_auxiliary_loss _a : Optional[Any] = num_queries _a : str = num_channels _a : List[str] = min_size _a : int = max_size _a : Optional[int] = num_labels _a : List[str] = hidden_dim _a : int = hidden_dim def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : Tuple = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to( _a ) _a : Optional[Any] = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=_a ) _a : Union[str, Any] = ( torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=_a ) > 0.5 ).float() _a : Tuple = (torch.rand((self.batch_size, self.num_labels) ,device=_a ) > 0.5).long() _a : Dict = self.get_config() return config, pixel_values, pixel_mask, mask_labels, class_labels def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : int = MaskaFormerConfig( hidden_size=self.hidden_dim ,) _a : str = self.num_queries _a : Union[str, Any] = self.num_labels _a : Tuple = [1, 1, 1, 1] _a : Dict = self.num_channels _a : str = 64 _a : Tuple = 128 _a : Optional[Any] = self.hidden_dim _a : Union[str, Any] = self.hidden_dim _a : List[Any] = self.hidden_dim return config def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a, _a, _a, _a, _a : Optional[Any] = self.prepare_config_and_inputs() _a : str = {'pixel_values': pixel_values, 'pixel_mask': pixel_mask} return config, inputs_dict def __lowercase ( self : List[str] ,_a : Optional[Any] ,_a : str ): '''simple docstring''' _a : str = output.encoder_hidden_states _a : Any = output.pixel_decoder_hidden_states _a : Optional[Any] = output.transformer_decoder_hidden_states self.parent.assertTrue(len(_a ) ,len(config.backbone_config.depths ) ) self.parent.assertTrue(len(_a ) ,len(config.backbone_config.depths ) ) self.parent.assertTrue(len(_a ) ,config.decoder_layers ) def __lowercase ( self : List[str] ,_a : str ,_a : List[Any] ,_a : Any ,_a : Union[str, Any]=False ): '''simple docstring''' with torch.no_grad(): _a : str = MaskaFormerModel(config=_a ) model.to(_a ) model.eval() _a : Any = model(pixel_values=_a ,pixel_mask=_a ) _a : Optional[Any] = model(_a ,output_hidden_states=_a ) self.parent.assertEqual( output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.hidden_dim) ,) # let's ensure the other two hidden state exists self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None ) self.parent.assertTrue(output.encoder_last_hidden_state is not None ) if output_hidden_states: self.check_output_hidden_state(_a ,_a ) def __lowercase ( self : Tuple ,_a : List[Any] ,_a : Union[str, Any] ,_a : Tuple ,_a : List[str] ,_a : Any ): '''simple docstring''' _a : int = MaskaFormerForUniversalSegmentation(config=_a ) model.to(_a ) model.eval() def comm_check_on_output(_a : Any ): # let's still check that all the required stuff is there self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None ) self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None ) self.parent.assertTrue(result.encoder_last_hidden_state is not None ) # okay, now we need to check the logits shape # due to the encoder compression, masks have a //4 spatial size self.parent.assertEqual( result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,) # + 1 for null class self.parent.assertEqual( result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) ) with torch.no_grad(): _a : Any = model(pixel_values=_a ,pixel_mask=_a ) _a : Optional[int] = model(_a ) comm_check_on_output(_a ) _a : List[str] = model( pixel_values=_a ,pixel_mask=_a ,mask_labels=_a ,class_labels=_a ) comm_check_on_output(_a ) self.parent.assertTrue(result.loss is not None ) self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) ) @require_torch class UpperCAmelCase__ ( lowercase__ , lowercase__ , unittest.TestCase ): """simple docstring""" __UpperCAmelCase : Optional[int] = (MaskaFormerModel, MaskaFormerForUniversalSegmentation) if is_torch_available() else () __UpperCAmelCase : Dict = {'''feature-extraction''': MaskaFormerModel} if is_torch_available() else {} __UpperCAmelCase : Dict = False __UpperCAmelCase : Tuple = False __UpperCAmelCase : Dict = False __UpperCAmelCase : List[Any] = False def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Union[str, Any] = MaskaFormerModelTester(self ) _a : Dict = ConfigTester(self ,config_class=_a ,has_text_modality=_a ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' self.config_tester.run_common_tests() def __lowercase ( self : Optional[int] ): '''simple docstring''' _a, _a : List[str] = self.model_tester.prepare_config_and_inputs_for_common() self.model_tester.create_and_check_maskaformer_model(_a ,**_a ,output_hidden_states=_a ) def __lowercase ( self : str ): '''simple docstring''' _a : Dict = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_maskaformer_instance_segmentation_head_model(*_a ) @unittest.skip(reason='Mask2Former does not use inputs_embeds' ) def __lowercase ( self : Any ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former does not have a get_input_embeddings method' ) def __lowercase ( self : str ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former is not a generative model' ) def __lowercase ( self : List[Any] ): '''simple docstring''' pass @unittest.skip(reason='Mask2Former does not use token embeddings' ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' pass @require_torch_multi_gpu @unittest.skip( reason='Mask2Former has some layers using `add_module` which doesn\'t work well with `nn.DataParallel`' ) def __lowercase ( self : Dict ): '''simple docstring''' pass @unittest.skip('Will be fixed soon by reducing the size of the model used for common tests.' ) def __lowercase ( self : List[Any] ): '''simple docstring''' pass def __lowercase ( self : int ): '''simple docstring''' _a, _a : Any = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a : Union[str, Any] = model_class(_a ) _a : List[str] = inspect.signature(model.forward ) # signature.parameters is an OrderedDict => so arg_names order is deterministic _a : Optional[Any] = [*signature.parameters.keys()] _a : List[Any] = ['pixel_values'] self.assertListEqual(arg_names[:1] ,_a ) @slow def __lowercase ( self : List[str] ): '''simple docstring''' for model_name in ["facebook/mask2former-swin-small-coco-instance"]: _a : Dict = MaskaFormerModel.from_pretrained(_a ) self.assertIsNotNone(_a ) def __lowercase ( self : List[Any] ): '''simple docstring''' _a : int = (self.model_tester.min_size,) * 2 _a : Any = { 'pixel_values': torch.randn((2, 3, *size) ,device=_a ), 'mask_labels': torch.randn((2, 10, *size) ,device=_a ), 'class_labels': torch.zeros(2 ,10 ,device=_a ).long(), } _a : List[Any] = self.model_tester.get_config() _a : int = MaskaFormerForUniversalSegmentation(_a ).to(_a ) _a : str = model(**_a ) self.assertTrue(outputs.loss is not None ) def __lowercase ( self : List[str] ): '''simple docstring''' _a, _a : List[Any] = self.model_tester.prepare_config_and_inputs_for_common() self.model_tester.create_and_check_maskaformer_model(_a ,**_a ,output_hidden_states=_a ) def __lowercase ( self : int ): '''simple docstring''' _a, _a : Dict = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: _a : Any = model_class(_a ).to(_a ) _a : Optional[int] = model(**_a ,output_attentions=_a ) self.assertTrue(outputs.attentions is not None ) def __lowercase ( self : Tuple ): '''simple docstring''' if not self.model_tester.is_training: return _a : List[str] = self.all_model_classes[1] _a, _a, _a, _a, _a : List[str] = self.model_tester.prepare_config_and_inputs() _a : Any = model_class(_a ) model.to(_a ) model.train() _a : Union[str, Any] = model(_a ,mask_labels=_a ,class_labels=_a ).loss loss.backward() def __lowercase ( self : int ): '''simple docstring''' _a : int = self.all_model_classes[1] _a, _a, _a, _a, _a : List[Any] = self.model_tester.prepare_config_and_inputs() _a : str = True _a : str = True _a : List[str] = model_class(_a ).to(_a ) model.train() _a : Optional[int] = model(_a ,mask_labels=_a ,class_labels=_a ) _a : Tuple = outputs.encoder_hidden_states[0] encoder_hidden_states.retain_grad() _a : str = outputs.pixel_decoder_hidden_states[0] pixel_decoder_hidden_states.retain_grad() _a : Dict = outputs.transformer_decoder_hidden_states[0] transformer_decoder_hidden_states.retain_grad() _a : List[str] = outputs.attentions[0] attentions.retain_grad() outputs.loss.backward(retain_graph=_a ) self.assertIsNotNone(encoder_hidden_states.grad ) self.assertIsNotNone(pixel_decoder_hidden_states.grad ) self.assertIsNotNone(transformer_decoder_hidden_states.grad ) self.assertIsNotNone(attentions.grad ) __lowerCAmelCase = 1e-4 def UpperCAmelCase_ (): """simple docstring""" _a : int = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' ) return image @require_vision @slow class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @cached_property def __lowercase ( self : Union[str, Any] ): '''simple docstring''' return "facebook/mask2former-swin-small-coco-instance" @cached_property def __lowercase ( self : Any ): '''simple docstring''' return MaskaFormerImageProcessor.from_pretrained(self.model_checkpoints ) if is_vision_available() else None def __lowercase ( self : Any ): '''simple docstring''' _a : List[str] = MaskaFormerModel.from_pretrained(self.model_checkpoints ).to(_a ) _a : int = self.default_image_processor _a : Tuple = prepare_img() _a : Any = image_processor(_a ,return_tensors='pt' ).to(_a ) _a : Union[str, Any] = inputs['pixel_values'].shape # check size is divisible by 32 self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 ) # check size self.assertEqual(_a ,(1, 3, 384, 384) ) with torch.no_grad(): _a : Optional[Any] = model(**_a ) _a : List[Any] = torch.tensor( [[-0.2790, -1.0717, -1.1668], [-0.5128, -0.3128, -0.4987], [-0.5832, 0.1971, -0.0197]] ).to(_a ) self.assertTrue( torch.allclose( outputs.encoder_last_hidden_state[0, 0, :3, :3] ,_a ,atol=_a ) ) _a : str = torch.tensor( [[0.8973, 1.1847, 1.1776], [1.1934, 1.5040, 1.5128], [1.1153, 1.4486, 1.4951]] ).to(_a ) self.assertTrue( torch.allclose( outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,_a ,atol=_a ) ) _a : Any = torch.tensor( [[2.1152, 1.7000, -0.8603], [1.5808, 1.8004, -0.9353], [1.6043, 1.7495, -0.5999]] ).to(_a ) self.assertTrue( torch.allclose( outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,_a ,atol=_a ) ) def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[Any] = MaskaFormerForUniversalSegmentation.from_pretrained(self.model_checkpoints ).to(_a ).eval() _a : Optional[Any] = self.default_image_processor _a : List[Any] = prepare_img() _a : str = image_processor(_a ,return_tensors='pt' ).to(_a ) _a : Any = inputs['pixel_values'].shape # check size is divisible by 32 self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 ) # check size self.assertEqual(_a ,(1, 3, 384, 384) ) with torch.no_grad(): _a : Optional[int] = model(**_a ) # masks_queries_logits _a : Dict = outputs.masks_queries_logits self.assertEqual( masks_queries_logits.shape ,(1, model.config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ) _a : Dict = [ [-8.7839, -9.0056, -8.8121], [-7.4104, -7.0313, -6.5401], [-6.6105, -6.3427, -6.4675], ] _a : Optional[Any] = torch.tensor(_a ).to(_a ) self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,_a ,atol=_a ) ) # class_queries_logits _a : str = outputs.class_queries_logits self.assertEqual(class_queries_logits.shape ,(1, model.config.num_queries, model.config.num_labels + 1) ) _a : str = torch.tensor( [ [1.8324, -8.0835, -4.1922], [0.8450, -9.0050, -3.6053], [0.3045, -7.7293, -3.0275], ] ).to(_a ) self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,_a ,atol=_a ) ) def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : Any = MaskaFormerForUniversalSegmentation.from_pretrained(self.model_checkpoints ).to(_a ).eval() _a : Tuple = self.default_image_processor _a : Tuple = image_processor( [np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors='pt' ,) _a : str = inputs['pixel_values'].to(_a ) _a : str = [el.to(_a ) for el in inputs['mask_labels']] _a : Dict = [el.to(_a ) for el in inputs['class_labels']] with torch.no_grad(): _a : List[str] = model(**_a ) self.assertTrue(outputs.loss is not None )
271
1
'''simple docstring''' from __future__ import annotations import math def UpperCAmelCase_ (__a : int ): """simple docstring""" if 1 < number < 4: # 2 and 3 are primes return True elif number < 2 or number % 2 == 0 or number % 3 == 0: # Negatives, 0, 1, all even numbers, all multiples of 3 are not primes return False # All primes number are in format of 6k +/- 1 for i in range(5 , int(math.sqrt(__a ) + 1 ) , 6 ): if number % i == 0 or number % (i + 2) == 0: return False return True __lowerCAmelCase = [num for num in range(3, 1_0_0_0_0_1, 2) if not is_prime(num)] def UpperCAmelCase_ (__a : int ): """simple docstring""" if not isinstance(__a , __a ): raise ValueError('n must be an integer' ) if n <= 0: raise ValueError('n must be >= 0' ) _a : Dict = [] for num in range(len(__a ) ): _a : List[Any] = 0 while 2 * i * i <= odd_composites[num]: _a : Any = odd_composites[num] - 2 * i * i if is_prime(__a ): break i += 1 else: list_nums.append(odd_composites[num] ) if len(__a ) == n: return list_nums return [] def UpperCAmelCase_ (): """simple docstring""" return compute_nums(1 )[0] if __name__ == "__main__": print(f'''{solution() = }''')
271
'''simple docstring''' import argparse import json from typing import List from ltp import LTP from transformers import BertTokenizer def UpperCAmelCase_ (__a : List[Any] ): """simple docstring""" if ( (cp >= 0x4E_00 and cp <= 0x9F_FF) or (cp >= 0x34_00 and cp <= 0x4D_BF) # or (cp >= 0x2_00_00 and cp <= 0x2_A6_DF) # or (cp >= 0x2_A7_00 and cp <= 0x2_B7_3F) # or (cp >= 0x2_B7_40 and cp <= 0x2_B8_1F) # or (cp >= 0x2_B8_20 and cp <= 0x2_CE_AF) # or (cp >= 0xF9_00 and cp <= 0xFA_FF) or (cp >= 0x2_F8_00 and cp <= 0x2_FA_1F) # ): # return True return False def UpperCAmelCase_ (__a : str ): """simple docstring""" for char in word: _a : Union[str, Any] = ord(__a ) if not _is_chinese_char(__a ): return 0 return 1 def UpperCAmelCase_ (__a : List[str] ): """simple docstring""" _a : Dict = set() for token in tokens: _a : str = len(__a ) > 1 and is_chinese(__a ) if chinese_word: word_set.add(__a ) _a : Optional[Any] = list(__a ) return word_list def UpperCAmelCase_ (__a : List[str] , __a : set() ): """simple docstring""" if not chinese_word_set: return bert_tokens _a : Optional[Any] = max([len(__a ) for w in chinese_word_set] ) _a : Optional[int] = bert_tokens _a, _a : Any = 0, len(__a ) while start < end: _a : Tuple = True if is_chinese(bert_word[start] ): _a : Union[str, Any] = min(end - start , __a ) for i in range(__a , 1 , -1 ): _a : Optional[Any] = ''.join(bert_word[start : start + i] ) if whole_word in chinese_word_set: for j in range(start + 1 , start + i ): _a : Any = '##' + bert_word[j] _a : Union[str, Any] = start + i _a : int = False break if single_word: start += 1 return bert_word def UpperCAmelCase_ (__a : List[str] , __a : LTP , __a : BertTokenizer ): """simple docstring""" _a : int = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : Union[str, Any] = ltp_tokenizer.seg(lines[i : i + 1_0_0] )[0] _a : Optional[Any] = [get_chinese_word(__a ) for r in res] ltp_res.extend(__a ) assert len(__a ) == len(__a ) _a : str = [] for i in range(0 , len(__a ) , 1_0_0 ): _a : List[str] = bert_tokenizer(lines[i : i + 1_0_0] , add_special_tokens=__a , truncation=__a , max_length=5_1_2 ) bert_res.extend(res['input_ids'] ) assert len(__a ) == len(__a ) _a : List[str] = [] for input_ids, chinese_word in zip(__a , __a ): _a : int = [] for id in input_ids: _a : Optional[int] = bert_tokenizer._convert_id_to_token(__a ) input_tokens.append(__a ) _a : List[str] = add_sub_symbol(__a , __a ) _a : Tuple = [] # We only save pos of chinese subwords start with ##, which mean is part of a whole word. for i, token in enumerate(__a ): if token[:2] == "##": _a : str = token[2:] # save chinese tokens' pos if len(__a ) == 1 and _is_chinese_char(ord(__a ) ): ref_id.append(__a ) ref_ids.append(__a ) assert len(__a ) == len(__a ) return ref_ids def UpperCAmelCase_ (__a : Optional[Any] ): """simple docstring""" with open(args.file_name , 'r' , encoding='utf-8' ) as f: _a : Dict = f.readlines() _a : int = [line.strip() for line in data if len(__a ) > 0 and not line.isspace()] # avoid delimiter like '\u2029' _a : int = LTP(args.ltp ) # faster in GPU device _a : Tuple = BertTokenizer.from_pretrained(args.bert ) _a : int = prepare_ref(__a , __a , __a ) with open(args.save_path , 'w' , encoding='utf-8' ) as f: _a : Optional[Any] = [json.dumps(__a ) + '\n' for ref in ref_ids] f.writelines(__a ) if __name__ == "__main__": __lowerCAmelCase = argparse.ArgumentParser(description="""prepare_chinese_ref""") parser.add_argument( """--file_name""", type=str, default="""./resources/chinese-demo.txt""", help="""file need process, same as training data in lm""", ) parser.add_argument( """--ltp""", type=str, default="""./resources/ltp""", help="""resources for LTP tokenizer, usually a path""" ) parser.add_argument("""--bert""", type=str, default="""./resources/robert""", help="""resources for Bert tokenizer""") parser.add_argument("""--save_path""", type=str, default="""./resources/ref.txt""", help="""path to save res""") __lowerCAmelCase = parser.parse_args() main(args)
271
1
'''simple docstring''' class UpperCAmelCase__ : """simple docstring""" def __init__( self : Dict ): '''simple docstring''' _a : Dict = {} def __lowercase ( self : Union[str, Any] ): '''simple docstring''' print(self.vertex ) for i in self.vertex: print(_a ,' -> ' ,' -> '.join([str(_a ) for j in self.vertex[i]] ) ) def __lowercase ( self : Dict ,_a : int ,_a : int ): '''simple docstring''' if from_vertex in self.vertex: self.vertex[from_vertex].append(_a ) else: # else make a new vertex _a : int = [to_vertex] def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Tuple = [False] * len(self.vertex ) # call the recursive helper function for i in range(len(self.vertex ) ): if not visited[i]: self.dfs_recursive(_a ,_a ) def __lowercase ( self : Union[str, Any] ,_a : int ,_a : list ): '''simple docstring''' _a : List[Any] = True print(_a ,end=' ' ) # Recur for all the vertices that are adjacent to this node for i in self.vertex: if not visited[i]: self.dfs_recursive(_a ,_a ) if __name__ == "__main__": __lowerCAmelCase = Graph() g.add_edge(0, 1) g.add_edge(0, 2) g.add_edge(1, 2) g.add_edge(2, 0) g.add_edge(2, 3) g.add_edge(3, 3) g.print_graph() print("""DFS:""") g.dfs() # OUTPUT: # 0 -> 1 -> 2 # 1 -> 2 # 2 -> 0 -> 3 # 3 -> 3 # DFS: # 0 1 2 3
271
'''simple docstring''' import warnings from ...utils import logging from .image_processing_videomae import VideoMAEImageProcessor __lowerCAmelCase = logging.get_logger(__name__) class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : Tuple ,*_a : List[str] ,**_a : Any ): '''simple docstring''' warnings.warn( 'The class VideoMAEFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use VideoMAEImageProcessor instead.' ,_a ,) super().__init__(*_a ,**_a )
271
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 __lowerCAmelCase = """sshleifer/bart-tiny-random""" __lowerCAmelCase = """patrickvonplaten/t5-tiny-random""" @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @cached_property def __lowercase ( self : Optional[int] ): '''simple docstring''' return AutoConfig.from_pretrained(_a ) def __lowercase ( self : Tuple ): '''simple docstring''' _a, *_a : int = create_student_by_copying_alternating_layers(_a ,tempfile.mkdtemp() ,e=1 ,d=1 ) self.assertEqual(student.config.num_hidden_layers ,1 ) def __lowercase ( self : Dict ): '''simple docstring''' _a, *_a : Optional[int] = create_student_by_copying_alternating_layers(_a ,tempfile.mkdtemp() ,e=1 ,d=_a ) def __lowercase ( self : int ): '''simple docstring''' _a, *_a : Tuple = create_student_by_copying_alternating_layers(_a ,tempfile.mkdtemp() ,e=1 ,d=_a ) self.assertEqual(student.config.encoder_layers ,1 ) self.assertEqual(student.config.decoder_layers ,self.teacher_config.encoder_layers ) def __lowercase ( self : str ): '''simple docstring''' _a, *_a : List[Any] = create_student_by_copying_alternating_layers(_a ,tempfile.mkdtemp() ,e=1 ,d=1 ) self.assertEqual(student.config.encoder_layers ,1 ) self.assertEqual(student.config.decoder_layers ,1 ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' with self.assertRaises(_a ): create_student_by_copying_alternating_layers(_a ,tempfile.mkdtemp() ,e=_a ,d=_a )
271
'''simple docstring''' from __future__ import annotations from random import choice def UpperCAmelCase_ (__a : str ): """simple docstring""" return choice(__a ) def UpperCAmelCase_ (__a : list[int] , __a : int ): """simple docstring""" _a : Dict = random_pivot(__a ) # partition based on pivot # linear time _a : Optional[int] = [e for e in lst if e < pivot] _a : List[str] = [e for e in lst if e > pivot] # if we get lucky, pivot might be the element we want. # we can easily see this: # small (elements smaller than k) # + pivot (kth element) # + big (elements larger than k) if len(__a ) == k - 1: return pivot # pivot is in elements bigger than k elif len(__a ) < k - 1: return kth_number(__a , k - len(__a ) - 1 ) # pivot is in elements smaller than k else: return kth_number(__a , __a ) if __name__ == "__main__": import doctest doctest.testmod()
271
1
'''simple docstring''' from dataclasses import dataclass from typing import Optional, Tuple import torch from torch import nn from transformers import RobertaPreTrainedModel, XLMRobertaConfig, XLMRobertaModel from transformers.utils import ModelOutput @dataclass class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[torch.FloatTensor] = None __UpperCAmelCase : torch.FloatTensor = None __UpperCAmelCase : Optional[Tuple[torch.FloatTensor]] = None __UpperCAmelCase : Optional[Tuple[torch.FloatTensor]] = None class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" def __init__( self : List[str] ,_a : Union[str, Any]=1 ,_a : List[Any]=0 ,_a : List[Any]=2 ,_a : str=512 ,_a : Optional[int]="cls" ,_a : Union[str, Any]=False ,_a : Optional[int]=True ,**_a : List[Any] ,): '''simple docstring''' super().__init__(pad_token_id=_a ,bos_token_id=_a ,eos_token_id=_a ,**_a ) _a : Dict = project_dim _a : int = pooler_fn _a : Tuple = learn_encoder _a : Optional[int] = use_attention_mask class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : List[Any] = [R'''pooler''', R'''logit_scale'''] __UpperCAmelCase : int = [R'''position_ids''', R'''predictions.decoder.bias'''] __UpperCAmelCase : Dict = '''roberta''' __UpperCAmelCase : str = RobertaSeriesConfig def __init__( self : Tuple ,_a : List[Any] ): '''simple docstring''' super().__init__(_a ) _a : Union[str, Any] = XLMRobertaModel(_a ) _a : Tuple = nn.Linear(config.hidden_size ,config.project_dim ) _a : str = getattr(_a ,'has_pre_transformation' ,_a ) if self.has_pre_transformation: _a : int = nn.Linear(config.hidden_size ,config.project_dim ) _a : int = nn.LayerNorm(config.hidden_size ,eps=config.layer_norm_eps ) self.post_init() def __lowercase ( self : Union[str, Any] ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[torch.Tensor] = None ,_a : Optional[bool] = None ,_a : Optional[bool] = None ,_a : Optional[bool] = None ,): '''simple docstring''' _a : Optional[int] = return_dict if return_dict is not None else self.config.use_return_dict _a : str = self.base_model( input_ids=_a ,attention_mask=_a ,token_type_ids=_a ,position_ids=_a ,head_mask=_a ,inputs_embeds=_a ,encoder_hidden_states=_a ,encoder_attention_mask=_a ,output_attentions=_a ,output_hidden_states=True if self.has_pre_transformation else output_hidden_states ,return_dict=_a ,) if self.has_pre_transformation: _a : Optional[Any] = outputs['hidden_states'][-2] _a : str = self.pre_LN(_a ) _a : Dict = self.transformation_pre(_a ) return TransformationModelOutput( projection_state=_a ,last_hidden_state=outputs.last_hidden_state ,hidden_states=outputs.hidden_states ,attentions=outputs.attentions ,) else: _a : List[str] = self.transformation(outputs.last_hidden_state ) return TransformationModelOutput( projection_state=_a ,last_hidden_state=outputs.last_hidden_state ,hidden_states=outputs.hidden_states ,attentions=outputs.attentions ,)
271
'''simple docstring''' class UpperCAmelCase__ : """simple docstring""" def __init__( self : Dict ): '''simple docstring''' _a : Dict = {} def __lowercase ( self : Union[str, Any] ): '''simple docstring''' print(self.vertex ) for i in self.vertex: print(_a ,' -> ' ,' -> '.join([str(_a ) for j in self.vertex[i]] ) ) def __lowercase ( self : Dict ,_a : int ,_a : int ): '''simple docstring''' if from_vertex in self.vertex: self.vertex[from_vertex].append(_a ) else: # else make a new vertex _a : int = [to_vertex] def __lowercase ( self : Optional[int] ): '''simple docstring''' _a : Tuple = [False] * len(self.vertex ) # call the recursive helper function for i in range(len(self.vertex ) ): if not visited[i]: self.dfs_recursive(_a ,_a ) def __lowercase ( self : Union[str, Any] ,_a : int ,_a : list ): '''simple docstring''' _a : List[Any] = True print(_a ,end=' ' ) # Recur for all the vertices that are adjacent to this node for i in self.vertex: if not visited[i]: self.dfs_recursive(_a ,_a ) if __name__ == "__main__": __lowerCAmelCase = Graph() g.add_edge(0, 1) g.add_edge(0, 2) g.add_edge(1, 2) g.add_edge(2, 0) g.add_edge(2, 3) g.add_edge(3, 3) g.print_graph() print("""DFS:""") g.dfs() # OUTPUT: # 0 -> 1 -> 2 # 1 -> 2 # 2 -> 0 -> 3 # 3 -> 3 # DFS: # 0 1 2 3
271
1
'''simple docstring''' import logging import math import os from dataclasses import dataclass, field from glob import glob from typing import Optional from torch.utils.data import ConcatDataset import transformers from transformers import ( CONFIG_MAPPING, MODEL_WITH_LM_HEAD_MAPPING, AutoConfig, AutoModelWithLMHead, AutoTokenizer, DataCollatorForLanguageModeling, DataCollatorForPermutationLanguageModeling, DataCollatorForWholeWordMask, HfArgumentParser, LineByLineTextDataset, LineByLineWithRefDataset, PreTrainedTokenizer, TextDataset, Trainer, TrainingArguments, set_seed, ) from transformers.trainer_utils import is_main_process __lowerCAmelCase = logging.getLogger(__name__) __lowerCAmelCase = list(MODEL_WITH_LM_HEAD_MAPPING.keys()) __lowerCAmelCase = tuple(conf.model_type for conf in MODEL_CONFIG_CLASSES) @dataclass class UpperCAmelCase__ : """simple docstring""" __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={ '''help''': ( '''The model checkpoint for weights initialization. Leave None if you want to train a model from''' ''' scratch.''' ) } , ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''If training from scratch, pass a model type from the list: ''' + ''', '''.join(lowercase__ )} , ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''Pretrained config name or path if not the same as model_name'''} ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''Pretrained tokenizer name or path if not the same as model_name'''} ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''Where do you want to store the pretrained models downloaded from huggingface.co'''} , ) @dataclass class UpperCAmelCase__ : """simple docstring""" __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''The input training data file (a text file).'''} ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={ '''help''': ( '''The input training data files (multiple files in glob format). ''' '''Very often splitting large files to smaller files can prevent tokenizer going out of memory''' ) } , ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''An optional input evaluation data file to evaluate the perplexity on (a text file).'''} , ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''An optional input train ref data file for whole word mask in Chinese.'''} , ) __UpperCAmelCase : Optional[str] = field( default=lowercase__ , metadata={'''help''': '''An optional input eval ref data file for whole word mask in Chinese.'''} , ) __UpperCAmelCase : bool = field( default=lowercase__ , metadata={'''help''': '''Whether distinct lines of text in the dataset are to be handled as distinct sequences.'''} , ) __UpperCAmelCase : bool = field( default=lowercase__ , metadata={'''help''': '''Train with masked-language modeling loss instead of language modeling.'''} ) __UpperCAmelCase : bool = field(default=lowercase__ , metadata={'''help''': '''Whether ot not to use whole word mask.'''} ) __UpperCAmelCase : float = field( default=0.1_5 , metadata={'''help''': '''Ratio of tokens to mask for masked language modeling loss'''} ) __UpperCAmelCase : float = field( default=1 / 6 , metadata={ '''help''': ( '''Ratio of length of a span of masked tokens to surrounding context length for permutation language''' ''' modeling.''' ) } , ) __UpperCAmelCase : int = field( default=5 , metadata={'''help''': '''Maximum length of a span of masked tokens for permutation language modeling.'''} ) __UpperCAmelCase : int = field( default=-1 , metadata={ '''help''': ( '''Optional input sequence length after tokenization.''' '''The training dataset will be truncated in block of this size for training.''' '''Default to the model max input length for single sentence inputs (take into account special tokens).''' ) } , ) __UpperCAmelCase : bool = field( default=lowercase__ , metadata={'''help''': '''Overwrite the cached training and evaluation sets'''} ) def UpperCAmelCase_ (__a : DataTrainingArguments , __a : PreTrainedTokenizer , __a : bool = False , __a : Optional[str] = None , ): """simple docstring""" def _dataset(__a : Any , __a : int=None ): if args.line_by_line: if ref_path is not None: if not args.whole_word_mask or not args.mlm: raise ValueError('You need to set world whole masking and mlm to True for Chinese Whole Word Mask' ) return LineByLineWithRefDataset( tokenizer=__a , file_path=__a , block_size=args.block_size , ref_path=__a , ) return LineByLineTextDataset(tokenizer=__a , file_path=__a , block_size=args.block_size ) else: return TextDataset( tokenizer=__a , file_path=__a , block_size=args.block_size , overwrite_cache=args.overwrite_cache , cache_dir=__a , ) if evaluate: return _dataset(args.eval_data_file , args.eval_ref_file ) elif args.train_data_files: return ConcatDataset([_dataset(__a ) for f in glob(args.train_data_files )] ) else: return _dataset(args.train_data_file , args.train_ref_file ) def UpperCAmelCase_ (): """simple docstring""" _a : Any = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments) ) _a, _a, _a : int = parser.parse_args_into_dataclasses() if data_args.eval_data_file is None and training_args.do_eval: raise ValueError( 'Cannot do evaluation without an evaluation data file. Either supply a file to --eval_data_file ' 'or remove the --do_eval argument.' ) 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 if training_args.local_rank in [-1, 0] else logging.WARN , ) logger.warning( 'Process rank: %s, device: %s, n_gpu: %s, distributed training: %s, 16-bits training: %s' , training_args.local_rank , training_args.device , training_args.n_gpu , bool(training_args.local_rank != -1 ) , training_args.fpaa , ) # Set the verbosity to info of the Transformers logger (on main process only): if is_main_process(training_args.local_rank ): transformers.utils.logging.set_verbosity_info() transformers.utils.logging.enable_default_handler() transformers.utils.logging.enable_explicit_format() logger.info('Training/evaluation parameters %s' , __a ) # Set seed set_seed(training_args.seed ) # Load pretrained model and tokenizer # # Distributed training: # The .from_pretrained methods guarantee that only one local process can concurrently # download model & vocab. if model_args.config_name: _a : List[Any] = AutoConfig.from_pretrained(model_args.config_name , cache_dir=model_args.cache_dir ) elif model_args.model_name_or_path: _a : List[str] = AutoConfig.from_pretrained(model_args.model_name_or_path , cache_dir=model_args.cache_dir ) else: _a : Tuple = CONFIG_MAPPING[model_args.model_type]() logger.warning('You are instantiating a new config instance from scratch.' ) if model_args.tokenizer_name: _a : int = AutoTokenizer.from_pretrained(model_args.tokenizer_name , cache_dir=model_args.cache_dir ) elif model_args.model_name_or_path: _a : Any = AutoTokenizer.from_pretrained(model_args.model_name_or_path , cache_dir=model_args.cache_dir ) else: raise ValueError( 'You are instantiating a new tokenizer from scratch. This is not supported, but you can do it from another' ' script, save it,and load it from here, using --tokenizer_name' ) if model_args.model_name_or_path: _a : Any = AutoModelWithLMHead.from_pretrained( model_args.model_name_or_path , from_tf=bool('.ckpt' in model_args.model_name_or_path ) , config=__a , cache_dir=model_args.cache_dir , ) else: logger.info('Training new model from scratch' ) _a : Tuple = AutoModelWithLMHead.from_config(__a ) model.resize_token_embeddings(len(__a ) ) if config.model_type in ["bert", "roberta", "distilbert", "camembert"] and not data_args.mlm: raise ValueError( 'BERT and RoBERTa-like models do not have LM heads but masked LM heads. They must be run using the' '--mlm flag (masked language modeling).' ) if data_args.block_size <= 0: _a : List[Any] = tokenizer.max_len # Our input block size will be the max possible for the model else: _a : Any = min(data_args.block_size , tokenizer.max_len ) # Get datasets _a : List[str] = ( get_dataset(__a , tokenizer=__a , cache_dir=model_args.cache_dir ) if training_args.do_train else None ) _a : int = ( get_dataset(__a , tokenizer=__a , evaluate=__a , cache_dir=model_args.cache_dir ) if training_args.do_eval else None ) if config.model_type == "xlnet": _a : Any = DataCollatorForPermutationLanguageModeling( tokenizer=__a , plm_probability=data_args.plm_probability , max_span_length=data_args.max_span_length , ) else: if data_args.mlm and data_args.whole_word_mask: _a : Optional[int] = DataCollatorForWholeWordMask( tokenizer=__a , mlm_probability=data_args.mlm_probability ) else: _a : int = DataCollatorForLanguageModeling( tokenizer=__a , mlm=data_args.mlm , mlm_probability=data_args.mlm_probability ) # Initialize our Trainer _a : Dict = Trainer( model=__a , args=__a , data_collator=__a , train_dataset=__a , eval_dataset=__a , prediction_loss_only=__a , ) # Training if training_args.do_train: _a : str = ( model_args.model_name_or_path if model_args.model_name_or_path is not None and os.path.isdir(model_args.model_name_or_path ) else None ) trainer.train(model_path=__a ) trainer.save_model() # For convenience, we also re-save the tokenizer to the same directory, # so that you can share your model easily on huggingface.co/models =) if trainer.is_world_master(): tokenizer.save_pretrained(training_args.output_dir ) # Evaluation _a : Optional[int] = {} if training_args.do_eval: logger.info('*** Evaluate ***' ) _a : int = trainer.evaluate() _a : Optional[Any] = math.exp(eval_output['eval_loss'] ) _a : List[Any] = {'perplexity': perplexity} _a : List[str] = os.path.join(training_args.output_dir , 'eval_results_lm.txt' ) if trainer.is_world_master(): with open(__a , 'w' ) as writer: logger.info('***** Eval results *****' ) for key in sorted(result.keys() ): logger.info(' %s = %s' , __a , str(result[key] ) ) writer.write('%s = %s\n' % (key, str(result[key] )) ) results.update(__a ) return results def UpperCAmelCase_ (__a : Optional[int] ): """simple docstring""" main() if __name__ == "__main__": main()
271
'''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 __lowerCAmelCase = logging.get_logger(__name__) __lowerCAmelCase = """▁""" __lowerCAmelCase = {"""vocab_file""": """sentencepiece.bpe.model""", """monolingual_vocab_file""": """dict.txt"""} __lowerCAmelCase = { """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""", }, } __lowerCAmelCase = {"""vinai/bartpho-syllable""": 1_0_2_4} class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Optional[Any] = VOCAB_FILES_NAMES __UpperCAmelCase : Dict = PRETRAINED_VOCAB_FILES_MAP __UpperCAmelCase : Optional[int] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES __UpperCAmelCase : Dict = ['''input_ids''', '''attention_mask'''] def __init__( self : str ,_a : str ,_a : Any ,_a : Any="<s>" ,_a : Dict="</s>" ,_a : int="</s>" ,_a : Union[str, Any]="<s>" ,_a : List[Any]="<unk>" ,_a : Optional[Any]="<pad>" ,_a : List[str]="<mask>" ,_a : Optional[Dict[str, Any]] = None ,**_a : int ,): '''simple docstring''' _a : Any = AddedToken(_a ,lstrip=_a ,rstrip=_a ) if isinstance(_a ,_a ) else mask_token _a : Optional[Any] = {} if sp_model_kwargs is None else sp_model_kwargs super().__init__( bos_token=_a ,eos_token=_a ,unk_token=_a ,sep_token=_a ,cls_token=_a ,pad_token=_a ,mask_token=_a ,sp_model_kwargs=self.sp_model_kwargs ,**_a ,) _a : Optional[int] = vocab_file _a : Union[str, Any] = monolingual_vocab_file _a : List[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.Load(str(_a ) ) # Load the reduced vocab # Keep order of special tokens for backward compatibility _a : Union[str, Any] = {} _a : int = 0 for token in [bos_token, pad_token, eos_token, unk_token, sep_token, cls_token]: if str(_a ) not in self.fairseq_tokens_to_ids: _a : int = cnt cnt += 1 with open(_a ,'r' ,encoding='utf-8' ) as f: for line in f.readlines(): _a : str = line.strip().split()[0] _a : Tuple = len(self.fairseq_tokens_to_ids ) if str(_a ) not in self.fairseq_tokens_to_ids: _a : List[str] = len(self.fairseq_tokens_to_ids ) _a : Tuple = {v: k for k, v in self.fairseq_tokens_to_ids.items()} def __getstate__( self : Union[str, Any] ): '''simple docstring''' _a : int = self.__dict__.copy() _a : str = None _a : Optional[Any] = self.sp_model.serialized_model_proto() return state def __setstate__( self : Tuple ,_a : Tuple ): '''simple docstring''' _a : Tuple = d # for backward compatibility if not hasattr(self ,'sp_model_kwargs' ): _a : List[str] = {} _a : List[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.LoadFromSerializedProto(self.sp_model_proto ) def __lowercase ( self : Dict ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' if token_ids_a is None: return [self.cls_token_id] + token_ids_a + [self.sep_token_id] _a : Dict = [self.cls_token_id] _a : int = [self.sep_token_id] return cls + token_ids_a + sep + sep + token_ids_a + sep def __lowercase ( self : int ,_a : List[int] ,_a : Optional[List[int]] = None ,_a : bool = False ): '''simple docstring''' if already_has_special_tokens: return super().get_special_tokens_mask( token_ids_a=_a ,token_ids_a=_a ,already_has_special_tokens=_a ) if token_ids_a is None: return [1] + ([0] * len(_a )) + [1] return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1] def __lowercase ( self : Tuple ,_a : List[int] ,_a : Optional[List[int]] = None ): '''simple docstring''' _a : List[str] = [self.sep_token_id] _a : List[str] = [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 __lowercase ( self : Dict ): '''simple docstring''' return len(self.fairseq_ids_to_tokens ) def __lowercase ( self : Dict ): '''simple docstring''' _a : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )} vocab.update(self.added_tokens_encoder ) return vocab def __lowercase ( self : Tuple ,_a : str ): '''simple docstring''' return self.sp_model.encode(_a ,out_type=_a ) def __lowercase ( self : Union[str, Any] ,_a : Union[str, Any] ): '''simple docstring''' if token in self.fairseq_tokens_to_ids: return self.fairseq_tokens_to_ids[token] else: return self.unk_token_id def __lowercase ( self : Any ,_a : int ): '''simple docstring''' return self.fairseq_ids_to_tokens[index] def __lowercase ( self : Tuple ,_a : Union[str, Any] ): '''simple docstring''' _a : str = ''.join(_a ).replace(_a ,' ' ).strip() return out_string def __lowercase ( self : Union[str, Any] ,_a : str ,_a : Optional[str] = None ): '''simple docstring''' if not os.path.isdir(_a ): logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" ) return _a : int = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) _a : int = os.path.join( _a ,(filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['monolingual_vocab_file'] ,) if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ): copyfile(self.vocab_file ,_a ) elif not os.path.isfile(self.vocab_file ): with open(_a ,'wb' ) as fi: _a : List[Any] = self.sp_model.serialized_model_proto() fi.write(_a ) if os.path.abspath(self.monolingual_vocab_file ) != os.path.abspath( _a ) and os.path.isfile(self.monolingual_vocab_file ): copyfile(self.monolingual_vocab_file ,_a ) elif not os.path.isfile(self.monolingual_vocab_file ): with open(_a ,'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(_a )} \n""" ) return out_vocab_file, out_monolingual_vocab_file
271
1
'''simple docstring''' import unittest import numpy as np import torch from diffusers import PNDMPipeline, PNDMScheduler, UNetaDModel from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device enable_full_determinism() class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" @property def __lowercase ( self : List[str] ): '''simple docstring''' torch.manual_seed(0 ) _a : int = UNetaDModel( block_out_channels=(32, 64) ,layers_per_block=2 ,sample_size=32 ,in_channels=3 ,out_channels=3 ,down_block_types=('DownBlock2D', 'AttnDownBlock2D') ,up_block_types=('AttnUpBlock2D', 'UpBlock2D') ,) return model def __lowercase ( self : Optional[Any] ): '''simple docstring''' _a : str = self.dummy_uncond_unet _a : int = PNDMScheduler() _a : str = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : Optional[int] = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ).images _a : List[str] = torch.manual_seed(0 ) _a : Any = pndm(generator=_a ,num_inference_steps=20 ,output_type='numpy' ,return_dict=_a )[0] _a : List[Any] = image[0, -3:, -3:, -1] _a : Any = image_from_tuple[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : List[Any] = np.array([1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2 assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1E-2 @slow @require_torch class UpperCAmelCase__ ( unittest.TestCase ): """simple docstring""" def __lowercase ( self : Tuple ): '''simple docstring''' _a : List[str] = 'google/ddpm-cifar10-32' _a : str = UNetaDModel.from_pretrained(_a ) _a : Union[str, Any] = PNDMScheduler() _a : Tuple = PNDMPipeline(unet=_a ,scheduler=_a ) pndm.to(_a ) pndm.set_progress_bar_config(disable=_a ) _a : str = torch.manual_seed(0 ) _a : Optional[Any] = pndm(generator=_a ,output_type='numpy' ).images _a : Dict = image[0, -3:, -3:, -1] assert image.shape == (1, 32, 32, 3) _a : Tuple = np.array([0.1564, 0.1_4645, 0.1406, 0.1_4715, 0.1_2425, 0.1_4045, 0.1_3115, 0.1_2175, 0.125] ) assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
271
'''simple docstring''' import numpy as np from transformers import BatchFeature from transformers.testing_utils import require_tf, require_torch from .test_feature_extraction_common import FeatureExtractionSavingTestMixin class UpperCAmelCase__ ( lowercase__ ): """simple docstring""" __UpperCAmelCase : Union[str, Any] = None __UpperCAmelCase : List[Any] = None @property def __lowercase ( self : Dict ): '''simple docstring''' return self.feat_extract_tester.prepare_feat_extract_dict() def __lowercase ( self : str ): '''simple docstring''' _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) self.assertTrue(hasattr(_a ,'feature_size' ) ) self.assertTrue(hasattr(_a ,'sampling_rate' ) ) self.assertTrue(hasattr(_a ,'padding_value' ) ) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.feat_extract_tester.prepare_inputs_for_common() _a : str = self.feature_extraction_class(**self.feat_extract_dict ) _a : int = feat_extract.model_input_names[0] _a : List[Any] = BatchFeature({input_name: speech_inputs} ) self.assertTrue(all(len(_a ) == len(_a ) for x, y in zip(_a ,processed_features[input_name] ) ) ) _a : Any = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Union[str, Any] = BatchFeature({input_name: speech_inputs} ,tensor_type='np' ) _a : Union[str, Any] = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : Optional[int] = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) @require_torch def __lowercase ( self : Any ): '''simple docstring''' _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : int = feat_extract.model_input_names[0] _a : str = BatchFeature({input_name: speech_inputs} ,tensor_type='pt' ) _a : str = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : str = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) @require_tf def __lowercase ( self : int ): '''simple docstring''' _a : int = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a ) _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : Tuple = feat_extract.model_input_names[0] _a : int = BatchFeature({input_name: speech_inputs} ,tensor_type='tf' ) _a : Optional[int] = processed_features[input_name] if len(batch_features_input.shape ) < 3: _a : Optional[Any] = batch_features_input[:, :, None] self.assertTrue( batch_features_input.shape == (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) ) def __lowercase ( self : Dict ,_a : Any=False ): '''simple docstring''' def _inputs_have_equal_length(_a : Tuple ): _a : Tuple = len(input[0] ) for input_slice in input[1:]: if len(_a ) != length: return False return True def _inputs_are_equal(_a : Optional[Any] ,_a : Union[str, Any] ): if len(_a ) != len(_a ): return False for input_slice_a, input_slice_a in zip(_a ,_a ): if not np.allclose(np.asarray(_a ) ,np.asarray(_a ) ,atol=1E-3 ): return False return True _a : int = self.feature_extraction_class(**self.feat_extract_dict ) _a : Tuple = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a ) _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Tuple = BatchFeature({input_name: speech_inputs} ) _a : str = self.feat_extract_tester.seq_length_diff _a : Dict = self.feat_extract_tester.max_seq_length + pad_diff _a : Dict = self.feat_extract_tester.min_seq_length _a : Optional[Any] = self.feat_extract_tester.batch_size _a : Tuple = self.feat_extract_tester.feature_size # test padding for List[int] + numpy _a : int = feat_extract.pad(_a ,padding=_a ) _a : List[Any] = input_a[input_name] _a : Tuple = feat_extract.pad(_a ,padding='longest' ) _a : Any = input_a[input_name] _a : Optional[Any] = feat_extract.pad(_a ,padding='max_length' ,max_length=len(speech_inputs[-1] ) ) _a : List[str] = input_a[input_name] _a : List[str] = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' ) _a : str = input_a[input_name] # max_length parameter has to be provided when setting `padding="max_length"` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='max_length' )[input_name] _a : int = feat_extract.pad( _a ,padding='max_length' ,max_length=_a ,return_tensors='np' ) _a : Optional[int] = input_a[input_name] self.assertFalse(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) self.assertTrue(len(input_a[0] ) == pad_min_length ) self.assertTrue(len(input_a[1] ) == pad_min_length + pad_diff ) self.assertTrue(input_a.shape[:2] == (batch_size, len(input_a[0] )) ) self.assertTrue(input_a.shape[:2] == (batch_size, pad_max_length) ) if feature_size > 1: self.assertTrue(input_a.shape[2] == input_a.shape[2] == feature_size ) # test padding for `pad_to_multiple_of` for List[int] + numpy _a : Tuple = feat_extract.pad(_a ,pad_to_multiple_of=10 ) _a : List[str] = input_a[input_name] _a : str = feat_extract.pad(_a ,padding='longest' ,pad_to_multiple_of=10 ) _a : Tuple = input_a[input_name] _a : Optional[int] = feat_extract.pad( _a ,padding='max_length' ,pad_to_multiple_of=10 ,max_length=_a ) _a : Any = input_a[input_name] _a : Optional[int] = feat_extract.pad( _a ,padding='max_length' ,pad_to_multiple_of=10 ,max_length=_a ,return_tensors='np' ,) _a : Dict = input_a[input_name] self.assertTrue(all(len(_a ) % 10 == 0 for x in input_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) _a : List[str] = pad_max_length if pad_max_length % 10 == 0 else (pad_max_length // 10 + 1) * 10 self.assertTrue(all(len(_a ) == expected_mult_pad_length for x in input_a ) ) self.assertEqual(input_a.shape[:2] ,(batch_size, expected_mult_pad_length) ) if feature_size > 1: self.assertTrue(input_a.shape[2] == feature_size ) # Check padding value is correct _a : Any = (np.ones(self.feat_extract_tester.feature_size ) * feat_extract.padding_value).sum() self.assertTrue( abs(np.asarray(input_a[0] )[pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1E-3 ) self.assertTrue( abs( np.asarray(input_a[1] )[pad_min_length + pad_diff :].sum() - padding_vector_sum * (pad_max_length - pad_min_length - pad_diff) ) < 1E-3 ) self.assertTrue( abs( np.asarray(input_a[2] )[pad_min_length + 2 * pad_diff :].sum() - padding_vector_sum * (pad_max_length - pad_min_length - 2 * pad_diff) ) < 1E-3 ) self.assertTrue( abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1E-3 ) self.assertTrue( abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (expected_mult_pad_length - pad_min_length) ) < 1E-3 ) def __lowercase ( self : List[Any] ,_a : Optional[int]=False ): '''simple docstring''' def _inputs_have_equal_length(_a : List[str] ): _a : Union[str, Any] = len(input[0] ) for input_slice in input[1:]: if len(_a ) != length: return False return True def _inputs_are_equal(_a : List[str] ,_a : List[str] ): if len(_a ) != len(_a ): return False for input_slice_a, input_slice_a in zip(_a ,_a ): if not np.allclose(np.asarray(_a ) ,np.asarray(_a ) ,atol=1E-3 ): return False return True _a : Dict = self.feature_extraction_class(**self.feat_extract_dict ) _a : str = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a ) _a : Any = feat_extract.model_input_names[0] _a : List[Any] = BatchFeature({input_name: speech_inputs} ) # truncate to smallest _a : Union[str, Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,truncation=_a ) _a : str = input_a[input_name] _a : List[str] = feat_extract.pad(_a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ) _a : Tuple = input_a[input_name] self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertFalse(_inputs_have_equal_length(_a ) ) # truncate to smallest with np _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,return_tensors='np' ,truncation=_a ,) _a : Any = input_a[input_name] _a : List[Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,return_tensors='np' ) _a : int = input_a[input_name] self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(input_a.shape[1] == len(speech_inputs[0] ) ) # since truncation forces padding to be smaller than longest input # function can't return `np.ndarray`, but has to return list self.assertFalse(_inputs_have_equal_length(_a ) ) # truncate to middle _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,truncation=_a ,return_tensors='np' ,) _a : List[Any] = input_a[input_name] _a : Tuple = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,truncation=_a ) _a : Tuple = input_a[input_name] _a : Tuple = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[1] ) ,return_tensors='np' ) _a : Dict = input_a[input_name] self.assertTrue(input_a.shape[1] == len(speech_inputs[1] ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertTrue(_inputs_are_equal(_a ,_a ) ) # since truncation forces padding to be smaller than longest input # function can't return `np.ndarray`, but has to return list self.assertFalse(_inputs_have_equal_length(_a ) ) self.assertTrue(len(input_a[-1] ) == len(speech_inputs[-1] ) ) # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,truncation=_a )[input_name] # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='longest' ,truncation=_a )[input_name] # padding has to be max_length when setting `truncation=True` with self.assertRaises(_a ): feat_extract.pad(_a ,padding='longest' ,truncation=_a )[input_name] # max_length parameter has to be provided when setting `truncation=True` and padding="max_length" with self.assertRaises(_a ): feat_extract.pad(_a ,padding='max_length' ,truncation=_a )[input_name] # test truncation for `pad_to_multiple_of` for List[int] + numpy _a : Optional[Any] = 12 _a : List[Any] = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,pad_to_multiple_of=_a ,truncation=_a ,) _a : Tuple = input_a[input_name] _a : str = feat_extract.pad( _a ,padding='max_length' ,max_length=len(speech_inputs[0] ) ,pad_to_multiple_of=_a ,) _a : List[Any] = input_a[input_name] # retrieve expected_length as multiple of pad_to_multiple_of _a : List[Any] = len(speech_inputs[0] ) if expected_length % pad_to_multiple_of != 0: _a : Union[str, Any] = ((len(speech_inputs[0] ) // pad_to_multiple_of) + 1) * pad_to_multiple_of self.assertTrue(len(input_a[0] ) == expected_length ) self.assertTrue(_inputs_have_equal_length(_a ) ) self.assertFalse(_inputs_have_equal_length(_a ) ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' self._check_padding(numpify=_a ) def __lowercase ( self : Tuple ): '''simple docstring''' self._check_padding(numpify=_a ) def __lowercase ( self : Dict ): '''simple docstring''' self._check_truncation(numpify=_a ) def __lowercase ( self : str ): '''simple docstring''' self._check_truncation(numpify=_a ) @require_torch def __lowercase ( self : Dict ): '''simple docstring''' _a : Any = self.feature_extraction_class(**self.feat_extract_dict ) _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Optional[int] = BatchFeature({input_name: speech_inputs} ) _a : List[Any] = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' )[input_name] _a : List[str] = feat_extract.pad(_a ,padding='longest' ,return_tensors='pt' )[input_name] self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_pt.numpy().astype(np.floataa ).sum() ) < 1E-2 ) @require_tf def __lowercase ( self : int ): '''simple docstring''' _a : List[str] = self.feature_extraction_class(**self.feat_extract_dict ) _a : Optional[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Dict = feat_extract.model_input_names[0] _a : Optional[Any] = BatchFeature({input_name: speech_inputs} ) _a : Dict = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' )[input_name] _a : Any = feat_extract.pad(_a ,padding='longest' ,return_tensors='tf' )[input_name] self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_tf.numpy().astype(np.floataa ).sum() ) < 1E-2 ) def __lowercase ( self : Union[str, Any] ): '''simple docstring''' _a : str = self.feat_extract_dict _a : List[Any] = True _a : Optional[int] = self.feature_extraction_class(**_a ) _a : List[Any] = self.feat_extract_tester.prepare_inputs_for_common() _a : Tuple = [len(_a ) for x in speech_inputs] _a : int = feat_extract.model_input_names[0] _a : Optional[Any] = BatchFeature({input_name: speech_inputs} ) _a : str = feat_extract.pad(_a ,padding='longest' ,return_tensors='np' ) self.assertIn('attention_mask' ,_a ) self.assertListEqual(list(processed.attention_mask.shape ) ,list(processed[input_name].shape[:2] ) ) self.assertListEqual(processed.attention_mask.sum(-1 ).tolist() ,_a ) def __lowercase ( self : int ): '''simple docstring''' _a : Any = self.feat_extract_dict _a : Tuple = True _a : Optional[int] = self.feature_extraction_class(**_a ) _a : Dict = self.feat_extract_tester.prepare_inputs_for_common() _a : Dict = [len(_a ) for x in speech_inputs] _a : Union[str, Any] = feat_extract.model_input_names[0] _a : Any = BatchFeature({input_name: speech_inputs} ) _a : List[Any] = min(_a ) _a : Dict = feat_extract.pad( _a ,padding='max_length' ,max_length=_a ,truncation=_a ,return_tensors='np' ) self.assertIn('attention_mask' ,_a ) self.assertListEqual( list(processed_pad.attention_mask.shape ) ,[processed_pad[input_name].shape[0], max_length] ) self.assertListEqual( processed_pad.attention_mask[:, :max_length].sum(-1 ).tolist() ,[max_length for x in speech_inputs] )
271
1