code
stringlengths
82
54.1k
code_codestyle
int64
0
699
style_context
stringlengths
111
35.6k
style_context_codestyle
int64
0
699
label
int64
0
1
from typing import Dict, List, Optional, Union import numpy as np from .feature_extraction_utils import BatchFeature, FeatureExtractionMixin from .utils import PaddingStrategy, TensorType, is_tf_tensor, is_torch_tensor, logging, to_numpy _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : float , **UpperCamelCase__ : Union[str, Any] ): A = feature_size A = sampling_rate A = padding_value A = kwargs.pop('padding_side' , 'right' ) A = kwargs.pop('return_attention_mask' , UpperCamelCase__ ) super().__init__(**UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Union[ BatchFeature, List[BatchFeature], Dict[str, BatchFeature], Dict[str, List[BatchFeature]], List[Dict[str, BatchFeature]], ] , UpperCamelCase__ : Union[bool, str, PaddingStrategy] = True , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[Union[str, TensorType]] = None , ): # If we have a list of dicts, let's convert it in a dict of lists # We do this to allow using this method as a collate_fn function in PyTorch Dataloader if isinstance(UpperCamelCase__ , (list, tuple) ) and isinstance(processed_features[0] , (dict, BatchFeature) ): A = { key: [example[key] for example in processed_features] for key in processed_features[0].keys() } # The model's main input name, usually `input_values`, has be passed for padding if self.model_input_names[0] not in processed_features: raise ValueError( 'You should supply an instance of `transformers.BatchFeature` or list of `transformers.BatchFeature`' f''' to this method that includes {self.model_input_names[0]}, but you provided''' f''' {list(processed_features.keys() )}''' ) A = processed_features[self.model_input_names[0]] A = ( return_attention_mask if return_attention_mask is not None else self.return_attention_mask ) if len(UpperCamelCase__ ) == 0: if return_attention_mask: A = [] return processed_features # If we have PyTorch/TF tensors or lists as inputs, we cast them as Numpy arrays # and rebuild them afterwards if no return_tensors is specified # Note that we lose the specific device the tensor may be on for PyTorch A = required_input[0] if isinstance(UpperCamelCase__ , (list, tuple) ): # first_element might be an empty list/tuple in some edge cases so we grab the first non empty element. A = 0 while len(required_input[index] ) == 0: index += 1 if index < len(UpperCamelCase__ ): A = required_input[index][0] if return_tensors is None: if is_tf_tensor(UpperCamelCase__ ): A = 'tf' elif is_torch_tensor(UpperCamelCase__ ): A = 'pt' elif isinstance(UpperCamelCase__ , (int, float, list, tuple, np.ndarray) ): A = 'np' else: raise ValueError( f'''type of {first_element} unknown: {type(UpperCamelCase__ )}. ''' 'Should be one of a python, numpy, pytorch or tensorflow object.' ) for key, value in processed_features.items(): if isinstance(value[0] , (int, float) ): A = to_numpy(UpperCamelCase__ ) else: A = [to_numpy(UpperCamelCase__ ) for v in value] # Convert padding_strategy in PaddingStrategy A = self._get_padding_strategies(padding=UpperCamelCase__ , max_length=UpperCamelCase__ ) A = processed_features[self.model_input_names[0]] A = len(UpperCamelCase__ ) if not all(len(UpperCamelCase__ ) == batch_size for v in processed_features.values() ): raise ValueError('Some items in the output dictionary have a different batch size than others.' ) A = [] for i in range(UpperCamelCase__ ): A = {k: v[i] for k, v in processed_features.items()} # truncation A = self._truncate( UpperCamelCase__ , max_length=UpperCamelCase__ , pad_to_multiple_of=UpperCamelCase__ , truncation=UpperCamelCase__ , ) truncated_inputs.append(UpperCamelCase__ ) if padding_strategy == PaddingStrategy.LONGEST: # make sure that `max_length` cannot be longer than the longest truncated length A = max(len(input_slice[self.model_input_names[0]] ) for input_slice in truncated_inputs ) A = PaddingStrategy.MAX_LENGTH A = {} for i in range(UpperCamelCase__ ): # padding A = self._pad( truncated_inputs[i] , max_length=UpperCamelCase__ , padding_strategy=UpperCamelCase__ , pad_to_multiple_of=UpperCamelCase__ , return_attention_mask=UpperCamelCase__ , ) for key, value in outputs.items(): if key not in batch_outputs: A = [] if value.dtype is np.dtype(np.floataa ): A = value.astype(np.floataa ) batch_outputs[key].append(UpperCamelCase__ ) return BatchFeature(UpperCamelCase__ , tensor_type=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Union[Dict[str, np.ndarray], BatchFeature] , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : PaddingStrategy = PaddingStrategy.DO_NOT_PAD , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , ): A = processed_features[self.model_input_names[0]] if padding_strategy == PaddingStrategy.LONGEST: A = len(UpperCamelCase__ ) if max_length is not None and pad_to_multiple_of is not None and (max_length % pad_to_multiple_of != 0): A = ((max_length // pad_to_multiple_of) + 1) * pad_to_multiple_of A = padding_strategy != PaddingStrategy.DO_NOT_PAD and len(UpperCamelCase__ ) < max_length if return_attention_mask and "attention_mask" not in processed_features: A = np.ones(len(UpperCamelCase__ ) , dtype=np.intaa ) if needs_to_be_padded: A = max_length - len(UpperCamelCase__ ) if self.padding_side == "right": if return_attention_mask: A = np.pad( processed_features['attention_mask'] , (0, difference) ) A = ((0, difference), (0, 0)) if self.feature_size > 1 else (0, difference) A = np.pad( UpperCamelCase__ , UpperCamelCase__ , 'constant' , constant_values=self.padding_value ) elif self.padding_side == "left": if return_attention_mask: A = np.pad( processed_features['attention_mask'] , (difference, 0) ) A = ((difference, 0), (0, 0)) if self.feature_size > 1 else (difference, 0) A = np.pad( UpperCamelCase__ , UpperCamelCase__ , 'constant' , constant_values=self.padding_value ) else: raise ValueError('Invalid padding strategy:' + str(self.padding_side ) ) return processed_features def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[Dict[str, np.ndarray], BatchFeature] , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , ): if not truncation: return processed_features elif truncation and max_length is None: raise ValueError('When setting ``truncation=True``, make sure that ``max_length`` is defined.' ) A = processed_features[self.model_input_names[0]] # find `max_length` that fits `pad_to_multiple_of` if max_length is not None and pad_to_multiple_of is not None and (max_length % pad_to_multiple_of != 0): A = ((max_length // pad_to_multiple_of) + 1) * pad_to_multiple_of A = len(UpperCamelCase__ ) > max_length if needs_to_be_truncated: A = processed_features[self.model_input_names[0]][:max_length] if "attention_mask" in processed_features: A = processed_features['attention_mask'][:max_length] return processed_features def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[Any]=False , UpperCamelCase__ : List[str]=None ): # Get padding strategy if padding is not False: if padding is True: A = PaddingStrategy.LONGEST # Default to pad to the longest sequence in the batch elif not isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = PaddingStrategy(UpperCamelCase__ ) elif isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = padding else: A = PaddingStrategy.DO_NOT_PAD # Set max length if needed if max_length is None: if padding_strategy == PaddingStrategy.MAX_LENGTH: raise ValueError( f'''When setting ``padding={PaddingStrategy.MAX_LENGTH}``, make sure that max_length is defined''' ) # Test if we have a padding value if padding_strategy != PaddingStrategy.DO_NOT_PAD and (self.padding_value is None): raise ValueError( 'Asking to pad but the feature_extractor does not have a padding value. Please select a value to use' ' as `padding_value`. For example: `feature_extractor.padding_value = 0.0`.' ) return padding_strategy
699
from __future__ import annotations import math def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : bool, lowerCAmelCase : list[int], lowerCAmelCase : float ) -> int: if depth < 0: raise ValueError('Depth cannot be less than 0' ) if not scores: raise ValueError('Scores cannot be empty' ) if depth == height: return scores[node_index] return ( max( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) if is_max else min( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) ) def __UpperCamelCase () -> None: A = [90, 23, 6, 33, 21, 65, 123, 34_423] A = math.log(len(lowerCAmelCase ), 2 ) print(f'''Optimal value : {minimax(0, 0, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )}''' ) if __name__ == "__main__": import doctest doctest.testmod() main()
699
1
from copy import deepcopy class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : list[int] | None = None , UpperCamelCase__ : int | None = None ): if arr is None and size is not None: A = size A = [0] * size elif arr is not None: self.init(UpperCamelCase__ ) else: raise ValueError('Either arr or size must be specified' ) def UpperCamelCase ( self : Any , UpperCamelCase__ : list[int] ): A = len(UpperCamelCase__ ) A = deepcopy(UpperCamelCase__ ) for i in range(1 , self.size ): A = self.next_(UpperCamelCase__ ) if j < self.size: self.tree[j] += self.tree[i] def UpperCamelCase ( self : List[str] ): A = self.tree[:] for i in range(self.size - 1 , 0 , -1 ): A = self.next_(UpperCamelCase__ ) if j < self.size: arr[j] -= arr[i] return arr @staticmethod def UpperCamelCase ( UpperCamelCase__ : int ): return index + (index & (-index)) @staticmethod def UpperCamelCase ( UpperCamelCase__ : int ): return index - (index & (-index)) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int ): if index == 0: self.tree[0] += value return while index < self.size: self.tree[index] += value A = self.next_(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): self.add(UpperCamelCase__ , value - self.get(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int ): if right == 0: return 0 A = self.tree[0] right -= 1 # make right inclusive while right > 0: result += self.tree[right] A = self.prev(UpperCamelCase__ ) return result def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int ): return self.prefix(UpperCamelCase__ ) - self.prefix(UpperCamelCase__ ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): return self.query(UpperCamelCase__ , index + 1 ) def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : int ): value -= self.tree[0] if value < 0: return -1 A = 1 # Largest power of 2 <= size while j * 2 < self.size: j *= 2 A = 0 while j > 0: if i + j < self.size and self.tree[i + j] <= value: value -= self.tree[i + j] i += j j //= 2 return i if __name__ == "__main__": import doctest doctest.testmod()
699
import logging import os from dataclasses import dataclass, field from typing import Dict, Optional import datasets import numpy as np import tensorflow as tf from transformers import ( AutoConfig, AutoTokenizer, EvalPrediction, HfArgumentParser, PreTrainedTokenizer, TFAutoModelForSequenceClassification, TFTrainer, TFTrainingArguments, ) from transformers.utils import logging as hf_logging hf_logging.set_verbosity_info() hf_logging.enable_default_handler() hf_logging.enable_explicit_format() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : PreTrainedTokenizer, lowerCAmelCase : int, lowerCAmelCase : Optional[int] = None, ) -> Dict: A = {} if train_file is not None: A = [train_file] if eval_file is not None: A = [eval_file] if test_file is not None: A = [test_file] A = datasets.load_dataset('csv', data_files=lowerCAmelCase ) A = list(ds[list(files.keys() )[0]].features.keys() ) A = features_name.pop(lowerCAmelCase ) A = list(set(ds[list(files.keys() )[0]][label_name] ) ) A = {label: i for i, label in enumerate(lowerCAmelCase )} A = tokenizer.model_input_names A = {} if len(lowerCAmelCase ) == 1: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( example[features_name[0]], truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length' ), batched=lowerCAmelCase, ) elif len(lowerCAmelCase ) == 2: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( (example[features_name[0]], example[features_name[1]]), truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length', ), batched=lowerCAmelCase, ) def gen_train(): for ex in transformed_ds[datasets.Split.TRAIN]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_val(): for ex in transformed_ds[datasets.Split.VALIDATION]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_test(): for ex in transformed_ds[datasets.Split.TEST]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TRAIN in transformed_ds else None ) if train_ds is not None: A = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.VALIDATION in transformed_ds else None ) if val_ds is not None: A = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TEST in transformed_ds else None ) if test_ds is not None: A = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) ) return train_ds, val_ds, test_ds, labelaid _UpperCAmelCase = logging.getLogger(__name__) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : int = field(metadata={'''help''': '''Which column contains the label'''} ) SCREAMING_SNAKE_CASE : str = field(default=__lowercase , metadata={'''help''': '''The path of the training file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the development file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the test file'''} ) SCREAMING_SNAKE_CASE : int = field( default=1_28 , metadata={ '''help''': ( '''The maximum total input sequence length after tokenization. Sequences longer ''' '''than this will be truncated, sequences shorter will be padded.''' ) } , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Overwrite the cached training and evaluation sets'''} ) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : str = field( metadata={'''help''': '''Path to pretrained model or model identifier from huggingface.co/models'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained config name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained tokenizer name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : bool = field(default=__lowercase , metadata={'''help''': '''Set this flag to use fast tokenization.'''} ) # If you want to tweak more attributes on your tokenizer, you should do it in a distinct script, # or just modify its tokenizer_config.json. SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Where do you want to store the pretrained models downloaded from huggingface.co'''} , ) def __UpperCamelCase () -> Any: # See all possible arguments in src/transformers/training_args.py # or by passing the --help flag to this script. # We now keep distinct sets of args, for a cleaner separation of concerns. A = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) ) A , A , A = parser.parse_args_into_dataclasses() if ( os.path.exists(training_args.output_dir ) and os.listdir(training_args.output_dir ) and training_args.do_train and not training_args.overwrite_output_dir ): raise ValueError( f'''Output directory ({training_args.output_dir}) already exists and is not empty. Use''' ' --overwrite_output_dir to overcome.' ) # Setup logging logging.basicConfig( format='%(asctime)s - %(levelname)s - %(name)s - %(message)s', datefmt='%m/%d/%Y %H:%M:%S', level=logging.INFO, ) logger.info( f'''n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, ''' f'''16-bits training: {training_args.fpaa}''' ) logger.info(f'''Training/evaluation parameters {training_args}''' ) # Load pretrained model and tokenizer # # Distributed training: # The .from_pretrained methods guarantee that only one local process can concurrently # download model & vocab. A = AutoTokenizer.from_pretrained( model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path, cache_dir=model_args.cache_dir, ) A , A , A , A = get_tfds( train_file=data_args.train_file, eval_file=data_args.dev_file, test_file=data_args.test_file, tokenizer=lowerCAmelCase, label_column_id=data_args.label_column_id, max_seq_length=data_args.max_seq_length, ) A = AutoConfig.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, num_labels=len(lowerCAmelCase ), labelaid=lowerCAmelCase, idalabel={id: label for label, id in labelaid.items()}, finetuning_task='text-classification', cache_dir=model_args.cache_dir, ) with training_args.strategy.scope(): A = TFAutoModelForSequenceClassification.from_pretrained( model_args.model_name_or_path, from_pt=bool('.bin' in model_args.model_name_or_path ), config=lowerCAmelCase, cache_dir=model_args.cache_dir, ) def compute_metrics(lowerCAmelCase : EvalPrediction ) -> Dict: A = np.argmax(p.predictions, axis=1 ) return {"acc": (preds == p.label_ids).mean()} # Initialize our Trainer A = TFTrainer( model=lowerCAmelCase, args=lowerCAmelCase, train_dataset=lowerCAmelCase, eval_dataset=lowerCAmelCase, compute_metrics=lowerCAmelCase, ) # Training if training_args.do_train: trainer.train() trainer.save_model() tokenizer.save_pretrained(training_args.output_dir ) # Evaluation A = {} if training_args.do_eval: logger.info('*** Evaluate ***' ) A = trainer.evaluate() A = os.path.join(training_args.output_dir, 'eval_results.txt' ) with open(lowerCAmelCase, 'w' ) as writer: logger.info('***** Eval results *****' ) for key, value in result.items(): logger.info(f''' {key} = {value}''' ) writer.write(f'''{key} = {value}\n''' ) results.update(lowerCAmelCase ) return results if __name__ == "__main__": main()
699
1
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, )
699
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = {"configuration_xglm": ["XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "XGLMConfig"]} try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizer"] try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "XGLMForCausalLM", "XGLMModel", "XGLMPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxXGLMForCausalLM", "FlaxXGLMModel", "FlaxXGLMPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "TFXGLMForCausalLM", "TFXGLMModel", "TFXGLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm import XGLMTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm_fast import XGLMTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_xglm import ( TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, TFXGLMForCausalLM, TFXGLMModel, TFXGLMPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure)
699
1
from typing import Dict from .base import GenericTensor, Pipeline class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def UpperCamelCase ( self : Any , UpperCamelCase__ : Tuple=None , UpperCamelCase__ : Dict=None , UpperCamelCase__ : Optional[Any]=None , **UpperCamelCase__ : List[str] ): if tokenize_kwargs is None: A = {} if truncation is not None: if "truncation" in tokenize_kwargs: raise ValueError( 'truncation parameter defined twice (given as keyword argument as well as in tokenize_kwargs)' ) A = truncation A = tokenize_kwargs A = {} if return_tensors is not None: A = return_tensors return preprocess_params, {}, postprocess_params def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, Any] , **UpperCamelCase__ : Optional[int] ): A = self.framework A = self.tokenizer(UpperCamelCase__ , return_tensors=UpperCamelCase__ , **UpperCamelCase__ ) return model_inputs def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Tuple ): A = self.model(**UpperCamelCase__ ) return model_outputs def UpperCamelCase ( self : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : int=False ): # [0] is the first available tensor, logits or last_hidden_state. if return_tensors: return model_outputs[0] if self.framework == "pt": return model_outputs[0].tolist() elif self.framework == "tf": return model_outputs[0].numpy().tolist() def __call__( self : str , *UpperCamelCase__ : Any , **UpperCamelCase__ : Any ): return super().__call__(*UpperCamelCase__ , **UpperCamelCase__ )
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Optional[int]: if b == 0: return 1 if (b % 2) == 0: return actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) else: return a * actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> float: if b < 0: return 1 / actual_power(lowerCAmelCase, lowerCAmelCase ) return actual_power(lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": print(power(-2, -3))
699
1
from __future__ import annotations import pandas as pd def __UpperCamelCase (lowerCAmelCase : list[int], lowerCAmelCase : list[int], lowerCAmelCase : int ) -> list[int]: A = [0] * no_of_processes A = [0] * no_of_processes # Copy the burst time into remaining_time[] for i in range(lowerCAmelCase ): A = burst_time[i] A = 0 A = 0 A = 999_999_999 A = 0 A = False # Process until all processes are completed while complete != no_of_processes: for j in range(lowerCAmelCase ): if arrival_time[j] <= increment_time and remaining_time[j] > 0: if remaining_time[j] < minm: A = remaining_time[j] A = j A = True if not check: increment_time += 1 continue remaining_time[short] -= 1 A = remaining_time[short] if minm == 0: A = 999_999_999 if remaining_time[short] == 0: complete += 1 A = False # Find finish time of current process A = increment_time + 1 # Calculate waiting time A = finish_time - arrival_time[short] A = finar - burst_time[short] if waiting_time[short] < 0: A = 0 # Increment time increment_time += 1 return waiting_time def __UpperCamelCase (lowerCAmelCase : list[int], lowerCAmelCase : int, lowerCAmelCase : list[int] ) -> list[int]: A = [0] * no_of_processes for i in range(lowerCAmelCase ): A = burst_time[i] + waiting_time[i] return turn_around_time def __UpperCamelCase (lowerCAmelCase : list[int], lowerCAmelCase : list[int], lowerCAmelCase : int ) -> None: A = 0 A = 0 for i in range(lowerCAmelCase ): A = total_waiting_time + waiting_time[i] A = total_turn_around_time + turn_around_time[i] print(f'''Average waiting time = {total_waiting_time / no_of_processes:.5f}''' ) print('Average turn around time =', total_turn_around_time / no_of_processes ) if __name__ == "__main__": print("Enter how many process you want to analyze") _UpperCAmelCase = int(input()) _UpperCAmelCase = [0] * no_of_processes _UpperCAmelCase = [0] * no_of_processes _UpperCAmelCase = list(range(1, no_of_processes + 1)) for i in range(no_of_processes): print("Enter the arrival time and burst time for process:--" + str(i + 1)) _UpperCAmelCase , _UpperCAmelCase = map(int, input().split()) _UpperCAmelCase = calculate_waitingtime(arrival_time, burst_time, no_of_processes) _UpperCAmelCase = burst_time _UpperCAmelCase = no_of_processes _UpperCAmelCase = waiting_time _UpperCAmelCase = calculate_turnaroundtime(bt, n, wt) calculate_average_times(waiting_time, turn_around_time, no_of_processes) _UpperCAmelCase = pd.DataFrame( list(zip(processes, burst_time, arrival_time, waiting_time, turn_around_time)), columns=[ "Process", "BurstTime", "ArrivalTime", "WaitingTime", "TurnAroundTime", ], ) # Printing the dataFrame pd.set_option("display.max_rows", fcfs.shape[0] + 1) print(fcfs)
699
def __UpperCamelCase (lowerCAmelCase : list[int] ) -> int: if not numbers: return 0 if not isinstance(lowerCAmelCase, (list, tuple) ) or not all( isinstance(lowerCAmelCase, lowerCAmelCase ) for number in numbers ): raise ValueError('numbers must be an iterable of integers' ) A = A = A = numbers[0] for i in range(1, len(lowerCAmelCase ) ): # update the maximum and minimum subarray products A = numbers[i] if number < 0: A , A = min_till_now, max_till_now A = max(lowerCAmelCase, max_till_now * number ) A = min(lowerCAmelCase, min_till_now * number ) # update the maximum product found till now A = max(lowerCAmelCase, lowerCAmelCase ) return max_prod
699
1
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> bool: return all(number % divisor != 0 for divisor in range(2, isqrt(lowerCAmelCase ) + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int = 10**6 ) -> int: A = 0 A = 1 A = 7 while prime_candidate < max_prime: primes_count += is_prime(lowerCAmelCase ) cube_index += 1 prime_candidate += 6 * cube_index return primes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
from .glue import glue_convert_examples_to_features, glue_output_modes, glue_processors, glue_tasks_num_labels from .squad import SquadExample, SquadFeatures, SquadVaProcessor, SquadVaProcessor, squad_convert_examples_to_features from .utils import DataProcessor, InputExample, InputFeatures, SingleSentenceClassificationProcessor from .xnli import xnli_output_modes, xnli_processors, xnli_tasks_num_labels
699
1
class _UpperCAmelCase : '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : list[int] ): A = len(UpperCamelCase__ ) A = [0] * len_array if len_array > 0: A = array[0] for i in range(1 , UpperCamelCase__ ): A = self.prefix_sum[i - 1] + array[i] def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): if start == 0: return self.prefix_sum[end] return self.prefix_sum[end] - self.prefix_sum[start - 1] def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int ): A = {0} for sum_item in self.prefix_sum: if sum_item - target_sum in sums: return True sums.add(UpperCamelCase__ ) return False if __name__ == "__main__": import doctest doctest.testmod()
699
from ...configuration_utils import PretrainedConfig from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "microsoft/biogpt": "https://huggingface.co/microsoft/biogpt/resolve/main/config.json", # See all BioGPT models at https://huggingface.co/models?filter=biogpt } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = '''biogpt''' def __init__( self : Optional[Any] , UpperCamelCase__ : str=42384 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : Dict=24 , UpperCamelCase__ : Any=16 , UpperCamelCase__ : str=4096 , UpperCamelCase__ : Tuple="gelu" , UpperCamelCase__ : List[str]=0.1 , UpperCamelCase__ : Any=0.1 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : List[Any]=0.02 , UpperCamelCase__ : Dict=1e-1_2 , UpperCamelCase__ : Any=True , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Any=1 , UpperCamelCase__ : List[str]=0 , UpperCamelCase__ : Optional[Any]=2 , **UpperCamelCase__ : List[Any] , ): A = vocab_size A = max_position_embeddings A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = initializer_range A = layer_norm_eps A = scale_embedding A = use_cache A = layerdrop A = activation_dropout super().__init__(pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , **UpperCamelCase__ )
699
1
from __future__ import annotations import unittest from transformers import XGLMConfig, XGLMTokenizer, is_tf_available from transformers.testing_utils import require_tf, slow from ...test_configuration_common import ConfigTester from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask from ...test_pipeline_mixin import PipelineTesterMixin if is_tf_available(): import tensorflow as tf from transformers.models.xglm.modeling_tf_xglm import ( TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, TFXGLMForCausalLM, TFXGLMModel, ) @require_tf class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : int = XGLMConfig SCREAMING_SNAKE_CASE : Any = {} SCREAMING_SNAKE_CASE : Any = '''gelu''' def __init__( self : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : List[str]=14 , UpperCamelCase__ : List[str]=7 , UpperCamelCase__ : Dict=True , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : Optional[int]=99 , UpperCamelCase__ : List[Any]=32 , UpperCamelCase__ : Optional[int]=2 , UpperCamelCase__ : List[str]=4 , UpperCamelCase__ : Optional[Any]=37 , UpperCamelCase__ : List[str]="gelu" , UpperCamelCase__ : Optional[int]=0.1 , UpperCamelCase__ : Optional[int]=0.1 , UpperCamelCase__ : List[str]=512 , UpperCamelCase__ : Any=0.02 , ): A = parent A = batch_size A = seq_length A = is_training A = use_input_mask A = use_labels A = vocab_size A = d_model A = num_hidden_layers A = num_attention_heads A = ffn_dim A = activation_function A = activation_dropout A = attention_dropout A = max_position_embeddings A = initializer_range A = None A = 0 A = 2 A = 1 def UpperCamelCase ( self : Optional[int] ): return XGLMConfig.from_pretrained('facebook/xglm-564M' ) def UpperCamelCase ( self : List[str] ): A = tf.clip_by_value( ids_tensor([self.batch_size, self.seq_length] , self.vocab_size ) , clip_value_min=0 , clip_value_max=3 ) A = None if self.use_input_mask: A = random_attention_mask([self.batch_size, self.seq_length] ) A = self.get_config() A = floats_tensor([self.num_hidden_layers, self.num_attention_heads] , 2 ) return ( config, input_ids, input_mask, head_mask, ) def UpperCamelCase ( self : Optional[int] ): return XGLMConfig( vocab_size=self.vocab_size , d_model=self.hidden_size , num_layers=self.num_hidden_layers , attention_heads=self.num_attention_heads , ffn_dim=self.ffn_dim , activation_function=self.activation_function , activation_dropout=self.activation_dropout , attention_dropout=self.attention_dropout , max_position_embeddings=self.max_position_embeddings , initializer_range=self.initializer_range , use_cache=UpperCamelCase__ , bos_token_id=self.bos_token_id , eos_token_id=self.eos_token_id , pad_token_id=self.pad_token_id , return_dict=UpperCamelCase__ , ) def UpperCamelCase ( self : Any ): A = self.prepare_config_and_inputs() ( ( A ) , ( A ) , ( A ) , ( A ) , ) = config_and_inputs A = { 'input_ids': input_ids, 'head_mask': head_mask, } return config, inputs_dict @require_tf class _UpperCAmelCase ( __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = (TFXGLMModel, TFXGLMForCausalLM) if is_tf_available() else () SCREAMING_SNAKE_CASE : List[Any] = (TFXGLMForCausalLM,) if is_tf_available() else () SCREAMING_SNAKE_CASE : List[Any] = ( {'''feature-extraction''': TFXGLMModel, '''text-generation''': TFXGLMForCausalLM} if is_tf_available() else {} ) SCREAMING_SNAKE_CASE : Union[str, Any] = False SCREAMING_SNAKE_CASE : Union[str, Any] = False SCREAMING_SNAKE_CASE : List[str] = False def UpperCamelCase ( self : Tuple ): A = TFXGLMModelTester(self ) A = ConfigTester(self , config_class=UpperCamelCase__ , n_embd=37 ) def UpperCamelCase ( self : str ): self.config_tester.run_common_tests() @slow def UpperCamelCase ( self : int ): for model_name in TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST[:1]: A = TFXGLMModel.from_pretrained(UpperCamelCase__ ) self.assertIsNotNone(UpperCamelCase__ ) @unittest.skip(reason='Currently, model embeddings are going to undergo a major refactor.' ) def UpperCamelCase ( self : int ): super().test_resize_token_embeddings() @require_tf class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' @slow def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict=True ): A = TFXGLMForCausalLM.from_pretrained('facebook/xglm-564M' ) A = tf.convert_to_tensor([[2, 268, 9865]] , dtype=tf.intaa ) # The dog # </s> The dog is a very friendly dog. He is very affectionate and loves to play with other # fmt: off A = [2, 268, 9865, 67, 11, 1988, 57252, 9865, 5, 984, 67, 1988, 213838, 1658, 53, 70446, 33, 6657, 278, 1581] # fmt: on A = model.generate(UpperCamelCase__ , do_sample=UpperCamelCase__ , num_beams=1 ) if verify_outputs: self.assertListEqual(output_ids[0].numpy().tolist() , UpperCamelCase__ ) @slow def UpperCamelCase ( self : int ): A = XGLMTokenizer.from_pretrained('facebook/xglm-564M' ) A = TFXGLMForCausalLM.from_pretrained('facebook/xglm-564M' ) tf.random.set_seed(0 ) A = tokenizer('Today is a nice day and' , return_tensors='tf' ) A = tokenized.input_ids # forces the generation to happen on CPU, to avoid GPU-related quirks (and assure same output regardless of the available devices) with tf.device(':/CPU:0' ): A = model.generate(UpperCamelCase__ , do_sample=UpperCamelCase__ , seed=[7, 0] ) A = tokenizer.decode(output_ids[0] , skip_special_tokens=UpperCamelCase__ ) A = ( 'Today is a nice day and warm evening here over Southern Alberta!! Today when they closed schools due' ) self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) @slow def UpperCamelCase ( self : Tuple ): A = TFXGLMForCausalLM.from_pretrained('facebook/xglm-564M' ) A = XGLMTokenizer.from_pretrained('facebook/xglm-564M' ) A = 'left' # use different length sentences to test batching A = [ 'This is an extremelly long sentence that only exists to test the ability of the model to cope with ' 'left-padding, such as in batched generation. The output for the sequence below should be the same ' 'regardless of whether left padding is applied or not. When', 'Hello, my dog is a little', ] A = tokenizer(UpperCamelCase__ , return_tensors='tf' , padding=UpperCamelCase__ ) A = inputs['input_ids'] A = model.generate(input_ids=UpperCamelCase__ , attention_mask=inputs['attention_mask'] , max_new_tokens=12 ) A = tokenizer(sentences[0] , return_tensors='tf' ).input_ids A = model.generate(input_ids=UpperCamelCase__ , max_new_tokens=12 ) A = tokenizer(sentences[1] , return_tensors='tf' ).input_ids A = model.generate(input_ids=UpperCamelCase__ , max_new_tokens=12 ) A = tokenizer.batch_decode(UpperCamelCase__ , skip_special_tokens=UpperCamelCase__ ) A = tokenizer.decode(output_non_padded[0] , skip_special_tokens=UpperCamelCase__ ) A = tokenizer.decode(output_padded[0] , skip_special_tokens=UpperCamelCase__ ) A = [ 'This is an extremelly long sentence that only exists to test the ability of the model to cope with ' 'left-padding, such as in batched generation. The output for the sequence below should be the same ' 'regardless of whether left padding is applied or not. When left padding is applied, the sequence will be ' 'a single', 'Hello, my dog is a little bit of a shy one, but he is very friendly', ] self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , [non_padded_sentence, padded_sentence] )
699
import sys def __UpperCamelCase (lowerCAmelCase : Dict ) -> Dict: A = len(lowerCAmelCase ) A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] for chain_length in range(2, lowerCAmelCase ): for a in range(1, n - chain_length + 1 ): A = a + chain_length - 1 A = sys.maxsize for c in range(lowerCAmelCase, lowerCAmelCase ): A = ( matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b] ) if cost < matrix[a][b]: A = cost A = c return matrix, sol def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Union[str, Any], lowerCAmelCase : Union[str, Any] ) -> List[str]: if i == j: print('A' + str(lowerCAmelCase ), end=' ' ) else: print('(', end=' ' ) print_optiomal_solution(lowerCAmelCase, lowerCAmelCase, optimal_solution[i][j] ) print_optiomal_solution(lowerCAmelCase, optimal_solution[i][j] + 1, lowerCAmelCase ) print(')', end=' ' ) def __UpperCamelCase () -> List[str]: A = [30, 35, 15, 5, 10, 20, 25] A = len(lowerCAmelCase ) # Size of matrix created from above array will be # 30*35 35*15 15*5 5*10 10*20 20*25 A , A = matrix_chain_order(lowerCAmelCase ) print('No. of Operation required: ' + str(matrix[1][n - 1] ) ) print_optiomal_solution(lowerCAmelCase, 1, n - 1 ) if __name__ == "__main__": main()
699
1
# This is the module that test_patching.py uses to test patch_submodule() import os # noqa: this is just for tests import os as renamed_os # noqa: this is just for tests from os import path # noqa: this is just for tests from os import path as renamed_path # noqa: this is just for tests from os.path import join # noqa: this is just for tests from os.path import join as renamed_join # noqa: this is just for tests _UpperCAmelCase = open # noqa: we just need to have a builtin inside this module to test it properly
699
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> bool: return all(number % divisor != 0 for divisor in range(2, isqrt(lowerCAmelCase ) + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int = 10**6 ) -> int: A = 0 A = 1 A = 7 while prime_candidate < max_prime: primes_count += is_prime(lowerCAmelCase ) cube_index += 1 prime_candidate += 6 * cube_index return primes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
1
from __future__ import annotations import inspect import unittest import numpy as np from transformers import DeiTConfig from transformers.testing_utils import require_tf, require_vision, slow from transformers.utils import cached_property, is_tf_available, is_vision_available from ...test_configuration_common import ConfigTester from ...test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor from ...test_pipeline_mixin import PipelineTesterMixin if is_tf_available(): import tensorflow as tf from transformers import ( TFDeiTForImageClassification, TFDeiTForImageClassificationWithTeacher, TFDeiTForMaskedImageModeling, TFDeiTModel, ) from transformers.models.deit.modeling_tf_deit import TF_DEIT_PRETRAINED_MODEL_ARCHIVE_LIST if is_vision_available(): from PIL import Image from transformers import DeiTImageProcessor class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[int]=13 , UpperCamelCase__ : Optional[Any]=30 , UpperCamelCase__ : Union[str, Any]=2 , UpperCamelCase__ : Dict=3 , UpperCamelCase__ : Optional[Any]=True , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : List[str]=32 , UpperCamelCase__ : Any=2 , UpperCamelCase__ : Tuple=4 , UpperCamelCase__ : Dict=37 , UpperCamelCase__ : Optional[int]="gelu" , UpperCamelCase__ : Tuple=0.1 , UpperCamelCase__ : Optional[int]=0.1 , UpperCamelCase__ : str=10 , UpperCamelCase__ : List[str]=0.02 , UpperCamelCase__ : List[Any]=3 , UpperCamelCase__ : Dict=None , UpperCamelCase__ : Tuple=2 , ): A = parent A = batch_size A = image_size A = patch_size A = num_channels A = is_training A = use_labels A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = type_sequence_label_size A = initializer_range A = scope A = encoder_stride # in DeiT, the seq length equals the number of patches + 2 (we add 2 for the [CLS] and distilation tokens) A = (image_size // patch_size) ** 2 A = num_patches + 2 def UpperCamelCase ( self : str ): A = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] ) A = None if self.use_labels: A = ids_tensor([self.batch_size] , self.type_sequence_label_size ) A = self.get_config() return config, pixel_values, labels def UpperCamelCase ( self : str ): return DeiTConfig( image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=UpperCamelCase__ , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , ) def UpperCamelCase ( self : str , UpperCamelCase__ : str , UpperCamelCase__ : List[Any] , UpperCamelCase__ : int ): A = TFDeiTModel(config=UpperCamelCase__ ) A = model(UpperCamelCase__ ) self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : List[str] ): A = TFDeiTForMaskedImageModeling(config=UpperCamelCase__ ) A = model(UpperCamelCase__ ) self.parent.assertEqual( result.reconstruction.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) ) # test greyscale images A = 1 A = TFDeiTForMaskedImageModeling(UpperCamelCase__ ) A = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] ) A = model(UpperCamelCase__ ) self.parent.assertEqual(result.reconstruction.shape , (self.batch_size, 1, self.image_size, self.image_size) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : int ): A = self.type_sequence_label_size A = TFDeiTForImageClassification(UpperCamelCase__ ) A = model(UpperCamelCase__ , labels=UpperCamelCase__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) ) # test greyscale images A = 1 A = TFDeiTForImageClassification(UpperCamelCase__ ) A = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] ) A = model(UpperCamelCase__ , labels=UpperCamelCase__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) ) def UpperCamelCase ( self : int ): A = self.prepare_config_and_inputs() A , A , A = config_and_inputs A = {'pixel_values': pixel_values} return config, inputs_dict @require_tf class _UpperCAmelCase ( __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ( ( TFDeiTModel, TFDeiTForImageClassification, TFDeiTForImageClassificationWithTeacher, TFDeiTForMaskedImageModeling, ) if is_tf_available() else () ) SCREAMING_SNAKE_CASE : int = ( { '''feature-extraction''': TFDeiTModel, '''image-classification''': (TFDeiTForImageClassification, TFDeiTForImageClassificationWithTeacher), } if is_tf_available() else {} ) SCREAMING_SNAKE_CASE : List[Any] = False SCREAMING_SNAKE_CASE : Dict = False SCREAMING_SNAKE_CASE : Any = False SCREAMING_SNAKE_CASE : Any = False def UpperCamelCase ( self : int ): A = TFDeiTModelTester(self ) A = ConfigTester(self , config_class=UpperCamelCase__ , has_text_modality=UpperCamelCase__ , hidden_size=37 ) def UpperCamelCase ( self : Union[str, Any] ): self.config_tester.run_common_tests() @unittest.skip(reason='DeiT does not use inputs_embeds' ) def UpperCamelCase ( self : str ): pass def UpperCamelCase ( self : int ): A , A = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: A = model_class(UpperCamelCase__ ) self.assertIsInstance(model.get_input_embeddings() , (tf.keras.layers.Layer) ) A = model.get_output_embeddings() self.assertTrue(x is None or isinstance(UpperCamelCase__ , tf.keras.layers.Dense ) ) def UpperCamelCase ( self : int ): A , A = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: A = model_class(UpperCamelCase__ ) A = inspect.signature(model.call ) # signature.parameters is an OrderedDict => so arg_names order is deterministic A = [*signature.parameters.keys()] A = ['pixel_values'] self.assertListEqual(arg_names[:1] , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_model(*UpperCamelCase__ ) def UpperCamelCase ( self : List[Any] ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_masked_image_modeling(*UpperCamelCase__ ) def UpperCamelCase ( self : List[Any] ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_image_classification(*UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any]=False ): A = super()._prepare_for_class(UpperCamelCase__ , UpperCamelCase__ , return_labels=UpperCamelCase__ ) if return_labels: if "labels" in inputs_dict and "labels" not in inspect.signature(model_class.call ).parameters: del inputs_dict["labels"] return inputs_dict @slow def UpperCamelCase ( self : Dict ): for model_name in TF_DEIT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]: A = TFDeiTModel.from_pretrained(UpperCamelCase__ ) self.assertIsNotNone(UpperCamelCase__ ) def __UpperCamelCase () -> Dict: A = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' ) return image @require_tf @require_vision class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' @cached_property def UpperCamelCase ( self : Optional[int] ): return ( DeiTImageProcessor.from_pretrained('facebook/deit-base-distilled-patch16-224' ) if is_vision_available() else None ) @slow def UpperCamelCase ( self : Union[str, Any] ): A = TFDeiTForImageClassificationWithTeacher.from_pretrained('facebook/deit-base-distilled-patch16-224' ) A = self.default_image_processor A = prepare_img() A = image_processor(images=UpperCamelCase__ , return_tensors='tf' ) # forward pass A = model(**UpperCamelCase__ ) # verify the logits A = tf.TensorShape((1, 1000) ) self.assertEqual(outputs.logits.shape , UpperCamelCase__ ) A = tf.constant([-1.0_266, 0.1_912, -1.2_861] ) self.assertTrue(np.allclose(outputs.logits[0, :3] , UpperCamelCase__ , atol=1e-4 ) )
699
import warnings from ...utils import logging from .image_processing_imagegpt import ImageGPTImageProcessor _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , *UpperCamelCase__ : List[Any] , **UpperCamelCase__ : Tuple ): warnings.warn( 'The class ImageGPTFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use ImageGPTImageProcessor instead.' , UpperCamelCase__ , ) super().__init__(*UpperCamelCase__ , **UpperCamelCase__ )
699
1
def __UpperCamelCase (lowerCAmelCase : int ) -> list[int]: if length <= 0 or not isinstance(lowerCAmelCase, lowerCAmelCase ): raise ValueError('Length must be a positive integer.' ) return [n * (2 * n - 1) for n in range(lowerCAmelCase )] if __name__ == "__main__": print(hexagonal_numbers(length=5)) print(hexagonal_numbers(length=10))
699
from typing import Any, Dict, Optional import torch import torch.nn.functional as F from torch import nn from ..utils import maybe_allow_in_graph from .activations import get_activation from .attention_processor import Attention from .embeddings import CombinedTimestepLabelEmbeddings @maybe_allow_in_graph class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int]=0.0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : str = "layer_norm" , UpperCamelCase__ : bool = False , ): super().__init__() A = only_cross_attention A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm_zero' A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm' if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None: raise ValueError( f'''`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to''' f''' define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}.''' ) # Define 3 blocks. Each block has its own normalization layer. # 1. Self-Attn if self.use_ada_layer_norm: A = AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A = AdaLayerNormZero(UpperCamelCase__ , UpperCamelCase__ ) else: A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = Attention( query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if only_cross_attention else None , upcast_attention=UpperCamelCase__ , ) # 2. Cross-Attn if cross_attention_dim is not None or double_self_attention: # We currently only use AdaLayerNormZero for self attention where there will only be one attention block. # I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during # the second cross attention block. A = ( AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) ) A = Attention( query_dim=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if not double_self_attention else None , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , upcast_attention=UpperCamelCase__ , ) # is self-attn if encoder_hidden_states is none else: A = None A = None # 3. Feed-forward A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = FeedForward(UpperCamelCase__ , dropout=UpperCamelCase__ , activation_fn=UpperCamelCase__ , final_dropout=UpperCamelCase__ ) # let chunk size default to None A = None A = 0 def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int ): # Sets chunk feed-forward A = chunk_size A = dim def UpperCamelCase ( self : Dict , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , UpperCamelCase__ : Dict[str, Any] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , ): # Notice that normalization is always applied before the real computation in the following blocks. # 1. Self-Attention if self.use_ada_layer_norm: A = self.norma(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A , A , A , A , A = self.norma( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=hidden_states.dtype ) else: A = self.norma(UpperCamelCase__ ) A = cross_attention_kwargs if cross_attention_kwargs is not None else {} A = self.attna( UpperCamelCase__ , encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) if self.use_ada_layer_norm_zero: A = gate_msa.unsqueeze(1 ) * attn_output A = attn_output + hidden_states # 2. Cross-Attention if self.attna is not None: A = ( self.norma(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else self.norma(UpperCamelCase__ ) ) A = self.attna( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) A = attn_output + hidden_states # 3. Feed-forward A = self.norma(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None] if self._chunk_size is not None: # "feed_forward_chunk_size" can be used to save memory if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0: raise ValueError( f'''`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`.''' ) A = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size A = torch.cat( [self.ff(UpperCamelCase__ ) for hid_slice in norm_hidden_states.chunk(UpperCamelCase__ , dim=self._chunk_dim )] , dim=self._chunk_dim , ) else: A = self.ff(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = gate_mlp.unsqueeze(1 ) * ff_output A = ff_output + hidden_states return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 4 , UpperCamelCase__ : float = 0.0 , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : bool = False , ): super().__init__() A = int(dim * mult ) A = dim_out if dim_out is not None else dim if activation_fn == "gelu": A = GELU(UpperCamelCase__ , UpperCamelCase__ ) if activation_fn == "gelu-approximate": A = GELU(UpperCamelCase__ , UpperCamelCase__ , approximate='tanh' ) elif activation_fn == "geglu": A = GEGLU(UpperCamelCase__ , UpperCamelCase__ ) elif activation_fn == "geglu-approximate": A = ApproximateGELU(UpperCamelCase__ , UpperCamelCase__ ) A = nn.ModuleList([] ) # project in self.net.append(UpperCamelCase__ ) # project dropout self.net.append(nn.Dropout(UpperCamelCase__ ) ) # project out self.net.append(nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) ) # FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout if final_dropout: self.net.append(nn.Dropout(UpperCamelCase__ ) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int ): for module in self.net: A = module(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : str = "none" ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) A = approximate def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ , approximate=self.approximate ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) , approximate=self.approximate ).to(dtype=gate.dtype ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int ): A = self.proj(UpperCamelCase__ ) A = self.gelu(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , dim_out * 2 ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) ).to(dtype=gate.dtype ) def UpperCamelCase ( self : str , UpperCamelCase__ : str ): A , A = self.proj(UpperCamelCase__ ).chunk(2 , dim=-1 ) return hidden_states * self.gelu(UpperCamelCase__ ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[int] ): A = self.proj(UpperCamelCase__ ) return x * torch.sigmoid(1.702 * x ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Tuple ): super().__init__() A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , embedding_dim * 2 ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[Any] ): A = self.linear(self.silu(self.emb(UpperCamelCase__ ) ) ) A , A = torch.chunk(UpperCamelCase__ , 2 ) A = self.norm(UpperCamelCase__ ) * (1 + scale) + shift return x class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : int , UpperCamelCase__ : List[str] ): super().__init__() A = CombinedTimestepLabelEmbeddings(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , 6 * embedding_dim , bias=UpperCamelCase__ ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ , eps=1e-6 ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : str , UpperCamelCase__ : Tuple=None ): A = self.linear(self.silu(self.emb(UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=UpperCamelCase__ ) ) ) A , A , A , A , A , A = emb.chunk(6 , dim=1 ) A = self.norm(UpperCamelCase__ ) * (1 + scale_msa[:, None]) + shift_msa[:, None] return x, gate_msa, shift_mlp, scale_mlp, gate_mlp class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[str] = None , UpperCamelCase__ : float = 1e-5 ): super().__init__() A = num_groups A = eps if act_fn is None: A = None else: A = get_activation(UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , out_dim * 2 ) def UpperCamelCase ( self : Any , UpperCamelCase__ : str , UpperCamelCase__ : str ): if self.act: A = self.act(UpperCamelCase__ ) A = self.linear(UpperCamelCase__ ) A = emb[:, :, None, None] A , A = emb.chunk(2 , dim=1 ) A = F.group_norm(UpperCamelCase__ , self.num_groups , eps=self.eps ) A = x * (1 + scale) + shift return x
699
1
import argparse import logging import os from datetime import datetime import numpy as np import torch from torch import nn from torch.utils.data import DataLoader, RandomSampler, TensorDataset from tqdm import tqdm from transformers import GPTaLMHeadModel _UpperCAmelCase = logging.getLogger(__name__) def __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : str ) -> Optional[int]: # save results if os.path.exists(lowerCAmelCase ): if os.path.exists(os.path.join(lowerCAmelCase, 'config.json' ) ) and os.path.isfile( os.path.join(lowerCAmelCase, 'config.json' ) ): os.remove(os.path.join(lowerCAmelCase, 'config.json' ) ) if os.path.exists(os.path.join(lowerCAmelCase, 'pytorch_model.bin' ) ) and os.path.isfile( os.path.join(lowerCAmelCase, 'pytorch_model.bin' ) ): os.remove(os.path.join(lowerCAmelCase, 'pytorch_model.bin' ) ) else: os.makedirs(lowerCAmelCase ) model.save_pretrained(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str=False ) -> Any: A = 2 if unlogit: A = torch.pow(lowerCAmelCase, lowerCAmelCase ) A = p * torch.log(lowerCAmelCase ) A = 0 return -plogp.sum(dim=-1 ) def __UpperCamelCase (lowerCAmelCase : Dict ) -> Union[str, Any]: logger.info('lv, h >\t' + '\t'.join(f'''{x + 1}''' for x in range(len(lowerCAmelCase ) ) ) ) for row in range(len(lowerCAmelCase ) ): if tensor.dtype != torch.long: logger.info(f'''layer {row + 1}:\t''' + '\t'.join(f'''{x:.5f}''' for x in tensor[row].cpu().data ) ) else: logger.info(f'''layer {row + 1}:\t''' + '\t'.join(f'''{x:d}''' for x in tensor[row].cpu().data ) ) def __UpperCamelCase (lowerCAmelCase : Any, lowerCAmelCase : int, lowerCAmelCase : Any, lowerCAmelCase : Optional[Any]=True, lowerCAmelCase : Optional[int]=True, lowerCAmelCase : Optional[int]=None, lowerCAmelCase : Union[str, Any]=False ) -> Any: A , A = model.config.num_hidden_layers, model.config.num_attention_heads A = torch.zeros(lowerCAmelCase, lowerCAmelCase ).to(args.device ) A = torch.zeros(lowerCAmelCase, lowerCAmelCase ).to(args.device ) if head_mask is None: A = torch.ones(lowerCAmelCase, lowerCAmelCase ).to(args.device ) head_mask.requires_grad_(requires_grad=lowerCAmelCase ) # If actually pruned attention multi-head, set head mask to None to avoid shape mismatch if actually_pruned: A = None A = 0.0 A = 0.0 for step, inputs in enumerate(tqdm(lowerCAmelCase, desc='Iteration', disable=args.local_rank not in [-1, 0] ) ): A = tuple(t.to(args.device ) for t in inputs ) ((A) , ) = inputs # Do a forward pass (not with torch.no_grad() since we need gradients for importance score - see below) A = model(lowerCAmelCase, labels=lowerCAmelCase, head_mask=lowerCAmelCase ) # (loss), lm_logits, presents, (all hidden_states), (attentions) A , A , A = ( outputs[0], outputs[1], outputs[-1], ) # Loss and logits are the first, attention the last loss.backward() # Backpropagate to populate the gradients in the head mask total_loss += loss.detach().cpu().numpy() if compute_entropy: for layer, attn in enumerate(lowerCAmelCase ): A = entropy(attn.detach(), lowerCAmelCase ) attn_entropy[layer] += masked_entropy.sum(-1 ).sum(0 ).sum(0 ).detach() if compute_importance: head_importance += head_mask.grad.abs().detach() tot_tokens += torch.ones_like(lowerCAmelCase ).float().detach().sum().data # Normalize attn_entropy /= tot_tokens head_importance /= tot_tokens # Layerwise importance normalization if not args.dont_normalize_importance_by_layer: A = 2 A = torch.pow(torch.pow(lowerCAmelCase, lowerCAmelCase ).sum(-1 ), 1 / exponent ) head_importance /= norm_by_layer.unsqueeze(-1 ) + 1E-20 if not args.dont_normalize_global_importance: A = (head_importance - head_importance.min()) / (head_importance.max() - head_importance.min()) # Print matrices if compute_entropy: logger.info('Attention entropies' ) print_ad_tensor(lowerCAmelCase ) if compute_importance: logger.info('Head importance scores' ) print_ad_tensor(lowerCAmelCase ) logger.info('Head ranked by importance scores' ) A = torch.zeros(head_importance.numel(), dtype=torch.long, device=args.device ) A = torch.arange( head_importance.numel(), device=args.device ) A = head_ranks.view_as(lowerCAmelCase ) print_ad_tensor(lowerCAmelCase ) return attn_entropy, head_importance, total_loss def __UpperCamelCase (lowerCAmelCase : List[Any], lowerCAmelCase : Dict, lowerCAmelCase : Optional[Any] ) -> List[Any]: A , A , A = compute_heads_importance(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, compute_entropy=lowerCAmelCase ) A = 1 / loss # instead of downsteam score use the LM loss logger.info('Pruning: original score: %f, threshold: %f', lowerCAmelCase, original_score * args.masking_threshold ) A = torch.ones_like(lowerCAmelCase ) A = max(1, int(new_head_mask.numel() * args.masking_amount ) ) A = original_score while current_score >= original_score * args.masking_threshold: A = new_head_mask.clone().detach() # save current head mask # heads from least important to most - keep only not-masked heads A = float('Inf' ) A = head_importance.view(-1 ).sort()[1] if len(lowerCAmelCase ) <= num_to_mask: print('BREAK BY num_to_mask' ) break # mask heads A = current_heads_to_mask[:num_to_mask] logger.info('Heads to mask: %s', str(current_heads_to_mask.tolist() ) ) A = new_head_mask.view(-1 ) A = 0.0 A = new_head_mask.view_as(lowerCAmelCase ) A = new_head_mask.clone().detach() print_ad_tensor(lowerCAmelCase ) # Compute metric and head importance again A , A , A = compute_heads_importance( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, compute_entropy=lowerCAmelCase, head_mask=lowerCAmelCase ) A = 1 / loss logger.info( 'Masking: current score: %f, remaining heads %d (%.1f percents)', lowerCAmelCase, new_head_mask.sum(), new_head_mask.sum() / new_head_mask.numel() * 100, ) logger.info('Final head mask' ) print_ad_tensor(lowerCAmelCase ) np.save(os.path.join(args.output_dir, 'head_mask.npy' ), head_mask.detach().cpu().numpy() ) return head_mask def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Any, lowerCAmelCase : Union[str, Any], lowerCAmelCase : List[Any] ) -> Dict: A = datetime.now() A , A , A = compute_heads_importance( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, compute_entropy=lowerCAmelCase, compute_importance=lowerCAmelCase, head_mask=lowerCAmelCase ) A = 1 / loss A = datetime.now() - before_time A = sum(p.numel() for p in model.parameters() ) A = { layer: (1 - head_mask[layer].long()).nonzero().squeeze().tolist() for layer in range(len(lowerCAmelCase ) ) } for k, v in heads_to_prune.items(): if isinstance(lowerCAmelCase, lowerCAmelCase ): A = [ v, ] assert sum(len(lowerCAmelCase ) for h in heads_to_prune.values() ) == (1 - head_mask.long()).sum().item() model.prune_heads(lowerCAmelCase ) A = sum(p.numel() for p in model.parameters() ) A = datetime.now() A , A , A = compute_heads_importance( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, compute_entropy=lowerCAmelCase, compute_importance=lowerCAmelCase, head_mask=lowerCAmelCase, actually_pruned=lowerCAmelCase, ) A = 1 / loss A = datetime.now() - before_time logger.info( 'Pruning: original num of params: %.2e, after pruning %.2e (%.1f percents)', lowerCAmelCase, lowerCAmelCase, pruned_num_params / original_num_params * 100, ) logger.info('Pruning: score with masking: %f score with pruning: %f', lowerCAmelCase, lowerCAmelCase ) logger.info('Pruning: speed ratio (original timing / new timing): %f percents', original_time / new_time * 100 ) save_model(lowerCAmelCase, args.output_dir ) def __UpperCamelCase () -> Any: A = argparse.ArgumentParser() # Required parameters parser.add_argument( '--data_dir', default=lowerCAmelCase, type=lowerCAmelCase, required=lowerCAmelCase, help='The input data dir. Should contain the .tsv files (or other data files) for the task.', ) parser.add_argument( '--model_name_or_path', default=lowerCAmelCase, type=lowerCAmelCase, required=lowerCAmelCase, help='Path to pretrained model or model identifier from huggingface.co/models', ) parser.add_argument( '--output_dir', default=lowerCAmelCase, type=lowerCAmelCase, required=lowerCAmelCase, help='The output directory where the model predictions and checkpoints will be written.', ) # Other parameters parser.add_argument( '--config_name', default='', type=lowerCAmelCase, help='Pretrained config name or path if not the same as model_name_or_path', ) parser.add_argument( '--tokenizer_name', default='', type=lowerCAmelCase, help='Pretrained tokenizer name or path if not the same as model_name_or_path', ) parser.add_argument( '--cache_dir', default=lowerCAmelCase, type=lowerCAmelCase, help='Where do you want to store the pre-trained models downloaded from s3', ) parser.add_argument( '--data_subset', type=lowerCAmelCase, default=-1, help='If > 0: limit the data to a subset of data_subset instances.' ) parser.add_argument( '--overwrite_output_dir', action='store_true', help='Whether to overwrite data in output directory' ) parser.add_argument( '--overwrite_cache', action='store_true', help='Overwrite the cached training and evaluation sets' ) parser.add_argument( '--dont_normalize_importance_by_layer', action='store_true', help='Don\'t normalize importance score by layers' ) parser.add_argument( '--dont_normalize_global_importance', action='store_true', help='Don\'t normalize all importance scores between 0 and 1', ) parser.add_argument( '--try_masking', action='store_true', help='Whether to try to mask head until a threshold of accuracy.' ) parser.add_argument( '--masking_threshold', default=0.9, type=lowerCAmelCase, help='masking threshold in term of metrics (stop masking when metric < threshold * original metric value).', ) parser.add_argument( '--masking_amount', default=0.1, type=lowerCAmelCase, help='Amount to heads to masking at each masking step.' ) parser.add_argument('--metric_name', default='acc', type=lowerCAmelCase, help='Metric to use for head masking.' ) parser.add_argument( '--max_seq_length', default=128, type=lowerCAmelCase, help=( 'The maximum total input sequence length after WordPiece tokenization. \n' 'Sequences longer than this will be truncated, sequences shorter padded.' ), ) parser.add_argument('--batch_size', default=1, type=lowerCAmelCase, help='Batch size.' ) parser.add_argument('--seed', type=lowerCAmelCase, default=42 ) parser.add_argument('--local_rank', type=lowerCAmelCase, default=-1, help='local_rank for distributed training on gpus' ) parser.add_argument('--no_cuda', action='store_true', help='Whether not to use CUDA when available' ) parser.add_argument('--server_ip', type=lowerCAmelCase, default='', help='Can be used for distant debugging.' ) parser.add_argument('--server_port', type=lowerCAmelCase, default='', help='Can be used for distant debugging.' ) A = parser.parse_args() if args.server_ip and args.server_port: # Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script import ptvsd print('Waiting for debugger attach' ) ptvsd.enable_attach(address=(args.server_ip, args.server_port), redirect_output=lowerCAmelCase ) ptvsd.wait_for_attach() # Setup devices and distributed training if args.local_rank == -1 or args.no_cuda: A = torch.device('cuda' if torch.cuda.is_available() and not args.no_cuda else 'cpu' ) A = 0 if args.no_cuda else torch.cuda.device_count() else: torch.cuda.set_device(args.local_rank ) A = torch.device('cuda', args.local_rank ) A = 1 torch.distributed.init_process_group(backend='nccl' ) # Initializes the distributed backend # Setup logging logging.basicConfig(level=logging.INFO if args.local_rank in [-1, 0] else logging.WARN ) logger.info('device: {} n_gpu: {}, distributed: {}'.format(args.device, args.n_gpu, bool(args.local_rank != -1 ) ) ) A = GPTaLMHeadModel.from_pretrained(args.model_name_or_path ) # Distributed and parallel training model.to(args.device ) if args.local_rank != -1: A = nn.parallel.DistributedDataParallel( lowerCAmelCase, device_ids=[args.local_rank], output_device=args.local_rank, find_unused_parameters=lowerCAmelCase ) elif args.n_gpu > 1: A = nn.DataParallel(lowerCAmelCase ) # Print/save training arguments os.makedirs(args.output_dir, exist_ok=lowerCAmelCase ) torch.save(lowerCAmelCase, os.path.join(args.output_dir, 'run_args.bin' ) ) logger.info('Training/evaluation parameters %s', lowerCAmelCase ) # Prepare dataset A = np.concatenate( [ np.loadtxt(args.data_dir, dtype=np.intaa ), ] ) A = (torch.from_numpy(lowerCAmelCase ),) A = TensorDataset(*lowerCAmelCase ) A = RandomSampler(lowerCAmelCase ) A = DataLoader(lowerCAmelCase, sampler=lowerCAmelCase, batch_size=args.batch_size ) # Compute head entropy and importance score compute_heads_importance(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) # Try head masking (set heads to zero until the score goes under a threshole) # and head pruning (remove masked heads and see the effect on the network) if args.try_masking and args.masking_threshold > 0.0 and args.masking_threshold < 1.0: A = mask_heads(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) prune_heads(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": main()
699
import json import os from typing import Dict, List, Optional, Tuple from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "vocab_file": "vocab.json", "tokenizer_config_file": "tokenizer_config.json", "merges_file": "merges.txt", } _UpperCAmelCase = { "vocab_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/vocab.json" ), }, "tokenizer_config_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/tokenizer_config.json" ), }, "merges_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/merges.txt" ), }, } _UpperCAmelCase = "</w>" _UpperCAmelCase = "@@ " def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> List[str]: A = set() A = word[0] for char in word[1:]: pairs.add((prev_char, char) ) A = char return pairs # Speech2Text2 has no max input length _UpperCAmelCase = {"facebook/s2t-wav2vec2-large-en-de": 1_024} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] def __init__( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="<pad>" , UpperCamelCase__ : int="</s>" , UpperCamelCase__ : Tuple="<unk>" , UpperCamelCase__ : List[str]=False , UpperCamelCase__ : List[str]=None , **UpperCamelCase__ : Optional[int] , ): super().__init__( unk_token=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , **UpperCamelCase__ , ) A = do_lower_case with open(UpperCamelCase__ , encoding='utf-8' ) as vocab_handle: A = json.load(UpperCamelCase__ ) A = {v: k for k, v in self.encoder.items()} if merges_file is None: logger.info(f'''No merges files provided. {self.__class__.__name__} can only be used for decoding.''' ) A = None A = None else: with open(UpperCamelCase__ , encoding='utf-8' ) as merges_handle: A = merges_handle.read().split('\n' )[:-1] A = [tuple(merge.split()[:2] ) for merge in merges] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = {} @property def UpperCamelCase ( self : Union[str, Any] ): return len(self.decoder ) def UpperCamelCase ( self : Optional[Any] ): return dict(self.encoder , **self.added_tokens_encoder ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): A = tuple(token[:-1] ) + (token[-1] + BPE_TOKEN_MERGES,) if token in self.cache: return self.cache[token] A = get_pairs(UpperCamelCase__ ) if not pairs: return token while True: A = min(UpperCamelCase__ , key=lambda UpperCamelCase__ : self.bpe_ranks.get(UpperCamelCase__ , float('inf' ) ) ) if bigram not in self.bpe_ranks: break A , A = bigram A = [] A = 0 while i < len(UpperCamelCase__ ): try: A = word.index(UpperCamelCase__ , UpperCamelCase__ ) except ValueError: new_word.extend(word[i:] ) break else: new_word.extend(word[i:j] ) A = j if word[i] == first and i < len(UpperCamelCase__ ) - 1 and word[i + 1] == second: new_word.append(first + second ) i += 2 else: new_word.append(word[i] ) i += 1 A = tuple(UpperCamelCase__ ) A = new_word if len(UpperCamelCase__ ) == 1: break else: A = get_pairs(UpperCamelCase__ ) A = ' '.join(UpperCamelCase__ ) if word == "\n " + BPE_TOKEN_MERGES: A = '\n' + BPE_TOKEN_MERGES if word.endswith(UpperCamelCase__ ): A = word.replace(UpperCamelCase__ , '' ) A = word.replace(' ' , UpperCamelCase__ ) A = word return word def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict ): if self.bpe_ranks is None: raise ValueError( 'This tokenizer was instantiated without a `merges.txt` file, so' ' that it can only be used for decoding, not for encoding.' 'Make sure to provide `merges.txt` file at instantiation to enable ' 'encoding.' ) if self.do_lower_case: A = text.lower() A = text.split() A = [] for token in text: if token: split_tokens.extend(list(self.bpe(UpperCamelCase__ ).split(' ' ) ) ) return split_tokens def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : str ): return self.encoder.get(UpperCamelCase__ , self.encoder.get(self.unk_token ) ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): A = self.decoder.get(UpperCamelCase__ , self.unk_token ) return result def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] ): A = ' '.join(UpperCamelCase__ ) # make sure @@ tokens are concatenated A = ''.join(string.split(UpperCamelCase__ ) ) return string def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] ) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as f: f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCamelCase__ , ensure_ascii=UpperCamelCase__ ) + '\n' ) A = 0 if self.bpe_ranks is None: return (vocab_file,) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as writer: for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCamelCase__ : kv[1] ): if index != token_index: logger.warning( f'''Saving vocabulary to {merges_file}: BPE merge indices are not consecutive.''' ' Please check that the tokenizer is not corrupted!' ) A = token_index writer.write(' '.join(UpperCamelCase__ ) + '\n' ) index += 1 return (vocab_file, merges_file)
699
1
import argparse import json import os import evaluate import torch from datasets import load_dataset from torch.optim import AdamW from torch.utils.data import DataLoader from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed from accelerate import Accelerator, DistributedType from accelerate.utils.deepspeed import DummyOptim, DummyScheduler _UpperCAmelCase = 16 _UpperCAmelCase = 32 def __UpperCamelCase (lowerCAmelCase : Accelerator, lowerCAmelCase : int = 16, lowerCAmelCase : str = "bert-base-cased" ) -> Union[str, Any]: A = AutoTokenizer.from_pretrained(lowerCAmelCase ) A = load_dataset('glue', 'mrpc' ) def tokenize_function(lowerCAmelCase : Union[str, Any] ): # max_length=None => use the model max length (it's actually the default) A = 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 A = datasets.map( lowerCAmelCase, batched=lowerCAmelCase, remove_columns=['idx', 'sentence1', 'sentence2'], load_from_cache_file=lowerCAmelCase ) # We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the # transformers library A = tokenized_datasets.rename_column('label', 'labels' ) def collate_fn(lowerCAmelCase : Any ): # On TPU it's best to pad everything to the same length or training will be very slow. if accelerator.distributed_type == DistributedType.TPU: return tokenizer.pad(lowerCAmelCase, padding='max_length', max_length=128, return_tensors='pt' ) return tokenizer.pad(lowerCAmelCase, padding='longest', return_tensors='pt' ) # Instantiate dataloaders. A = DataLoader( tokenized_datasets['train'], shuffle=lowerCAmelCase, collate_fn=lowerCAmelCase, batch_size=lowerCAmelCase ) A = DataLoader( tokenized_datasets['validation'], shuffle=lowerCAmelCase, collate_fn=lowerCAmelCase, batch_size=lowerCAmelCase ) return train_dataloader, eval_dataloader def __UpperCamelCase (lowerCAmelCase : Union[str, Any], lowerCAmelCase : Any ) -> Union[str, Any]: # Initialize accelerator A = Accelerator() # Sample hyper-parameters for learning rate, batch size, seed and a few other HPs A = config['lr'] A = int(config['num_epochs'] ) A = int(config['seed'] ) A = int(config['batch_size'] ) A = args.model_name_or_path set_seed(lowerCAmelCase ) A , A = get_dataloaders(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) # Instantiate the model (we build the model here so that the seed also control new weights initialization) A = AutoModelForSequenceClassification.from_pretrained(lowerCAmelCase, return_dict=lowerCAmelCase ) # Instantiate optimizer A = ( AdamW if accelerator.state.deepspeed_plugin is None or 'optimizer' not in accelerator.state.deepspeed_plugin.deepspeed_config else DummyOptim ) A = optimizer_cls(params=model.parameters(), lr=lowerCAmelCase ) if accelerator.state.deepspeed_plugin is not None: A = accelerator.state.deepspeed_plugin.deepspeed_config[ 'gradient_accumulation_steps' ] else: A = 1 A = (len(lowerCAmelCase ) * num_epochs) // gradient_accumulation_steps # Instantiate scheduler if ( accelerator.state.deepspeed_plugin is None or "scheduler" not in accelerator.state.deepspeed_plugin.deepspeed_config ): A = get_linear_schedule_with_warmup( optimizer=lowerCAmelCase, num_warmup_steps=0, num_training_steps=lowerCAmelCase, ) else: A = DummyScheduler(lowerCAmelCase, total_num_steps=lowerCAmelCase, warmup_num_steps=0 ) # Prepare everything # There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the # prepare method. A , A , A , A , A = accelerator.prepare( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) # We need to keep track of how many total steps we have iterated over A = 0 # We also need to keep track of the stating epoch so files are named properly A = 0 # Now we train the model A = evaluate.load('glue', 'mrpc' ) A = 0 A = {} for epoch in range(lowerCAmelCase, lowerCAmelCase ): model.train() for step, batch in enumerate(lowerCAmelCase ): A = model(**lowerCAmelCase ) A = outputs.loss A = loss / gradient_accumulation_steps accelerator.backward(lowerCAmelCase ) if step % gradient_accumulation_steps == 0: optimizer.step() lr_scheduler.step() optimizer.zero_grad() overall_step += 1 model.eval() A = 0 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(): A = model(**lowerCAmelCase ) A = outputs.logits.argmax(dim=-1 ) # It is slightly faster to call this once, than multiple times A , A = accelerator.gather( (predictions, batch['labels']) ) # If we are in a multiprocess environment, the last batch has duplicates if accelerator.use_distributed: if step == len(lowerCAmelCase ) - 1: A = predictions[: len(eval_dataloader.dataset ) - samples_seen] A = references[: len(eval_dataloader.dataset ) - samples_seen] else: samples_seen += references.shape[0] metric.add_batch( predictions=lowerCAmelCase, references=lowerCAmelCase, ) A = metric.compute() # Use accelerator.print to print only on the main process. accelerator.print(f'''epoch {epoch}:''', lowerCAmelCase ) A = eval_metric['accuracy'] if best_performance < eval_metric["accuracy"]: A = eval_metric['accuracy'] if args.performance_lower_bound is not None: assert ( args.performance_lower_bound <= best_performance ), f'''Best performance metric {best_performance} is lower than the lower bound {args.performance_lower_bound}''' accelerator.wait_for_everyone() if accelerator.is_main_process: with open(os.path.join(args.output_dir, 'all_results.json' ), 'w' ) as f: json.dump(lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase () -> Tuple: A = argparse.ArgumentParser(description='Simple example of training script tracking peak GPU memory usage.' ) parser.add_argument( '--model_name_or_path', type=lowerCAmelCase, default='bert-base-cased', help='Path to pretrained model or model identifier from huggingface.co/models.', required=lowerCAmelCase, ) parser.add_argument( '--output_dir', type=lowerCAmelCase, default='.', help='Optional save directory where all checkpoint folders will be stored. Default is the current working directory.', ) parser.add_argument( '--performance_lower_bound', type=lowerCAmelCase, default=lowerCAmelCase, help='Optional lower bound for the performance metric. If set, the training will throw error when the performance metric drops below this value.', ) parser.add_argument( '--num_epochs', type=lowerCAmelCase, default=3, help='Number of train epochs.', ) A = parser.parse_args() A = {'lr': 2E-5, 'num_epochs': args.num_epochs, 'seed': 42, 'batch_size': 16} training_function(lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": main()
699
# Copyright 2023 The HuggingFace Inc. team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import torch from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer from .base import PipelineTool class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = '''facebook/bart-large-mnli''' SCREAMING_SNAKE_CASE : Union[str, Any] = ( '''This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which ''' '''should be the text to classify, and `labels`, which should be the list of labels to use for classification. ''' '''It returns the most likely label in the list of provided `labels` for the input text.''' ) SCREAMING_SNAKE_CASE : Any = '''text_classifier''' SCREAMING_SNAKE_CASE : Any = AutoTokenizer SCREAMING_SNAKE_CASE : Dict = AutoModelForSequenceClassification SCREAMING_SNAKE_CASE : List[Any] = ['''text''', ['''text''']] SCREAMING_SNAKE_CASE : Dict = ['''text'''] def UpperCamelCase ( self : List[str] ): super().setup() A = self.model.config A = -1 for idx, label in config.idalabel.items(): if label.lower().startswith('entail' ): A = int(UpperCamelCase__ ) if self.entailment_id == -1: raise ValueError('Could not determine the entailment ID from the model config, please pass it at init.' ) def UpperCamelCase ( self : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict ): A = labels return self.pre_processor( [text] * len(UpperCamelCase__ ) , [f'''This example is {label}''' for label in labels] , return_tensors='pt' , padding='max_length' , ) def UpperCamelCase ( self : int , UpperCamelCase__ : List[str] ): A = outputs.logits A = torch.argmax(logits[:, 2] ).item() return self._labels[label_id]
699
1
import io import json import fsspec import pytest from datasets import Dataset, DatasetDict, Features, NamedSplit, Value from datasets.io.json import JsonDatasetReader, JsonDatasetWriter from ..utils import assert_arrow_memory_doesnt_increase, assert_arrow_memory_increases def __UpperCamelCase (lowerCAmelCase : List[Any], lowerCAmelCase : int ) -> Dict: assert isinstance(lowerCAmelCase, lowerCAmelCase ) assert dataset.num_rows == 4 assert dataset.num_columns == 3 assert dataset.column_names == ["col_1", "col_2", "col_3"] for feature, expected_dtype in expected_features.items(): assert dataset.features[feature].dtype == expected_dtype @pytest.mark.parametrize('keep_in_memory', [False, True] ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Union[str, Any], lowerCAmelCase : List[Any] ) -> List[Any]: A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase(): A = JsonDatasetReader(lowerCAmelCase, cache_dir=lowerCAmelCase, keep_in_memory=lowerCAmelCase ).read() _check_json_dataset(lowerCAmelCase, lowerCAmelCase ) @pytest.mark.parametrize( 'features', [ None, {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'}, {'col_1': 'string', 'col_2': 'string', 'col_3': 'string'}, {'col_1': 'int32', 'col_2': 'int32', 'col_3': 'int32'}, {'col_1': 'float32', 'col_2': 'float32', 'col_3': 'float32'}, ], ) def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Tuple, lowerCAmelCase : Optional[int] ) -> Tuple: A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} A = features.copy() if features else default_expected_features A = ( Features({feature: Value(lowerCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None ) A = JsonDatasetReader(lowerCAmelCase, features=lowerCAmelCase, cache_dir=lowerCAmelCase ).read() _check_json_dataset(lowerCAmelCase, lowerCAmelCase ) @pytest.mark.parametrize( 'features', [ None, {'col_3': 'float64', 'col_1': 'string', 'col_2': 'int64'}, ], ) def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Tuple, lowerCAmelCase : Dict ) -> Union[str, Any]: A = tmp_path / 'cache' A = {'col_3': 'float64', 'col_1': 'string', 'col_2': 'int64'} A = features.copy() if features else default_expected_features A = ( Features({feature: Value(lowerCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None ) A = JsonDatasetReader(lowerCAmelCase, features=lowerCAmelCase, cache_dir=lowerCAmelCase ).read() assert isinstance(lowerCAmelCase, lowerCAmelCase ) assert dataset.num_rows == 2 assert dataset.num_columns == 3 assert dataset.column_names == ["col_3", "col_1", "col_2"] for feature, expected_dtype in expected_features.items(): assert dataset.features[feature].dtype == expected_dtype def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Dict ) -> Dict: # jsonl_312_path features are {"col_3": "float64", "col_1": "string", "col_2": "int64"} A = {'col_2': 'int64', 'col_3': 'float64', 'col_1': 'string'} A = features.copy() A = ( Features({feature: Value(lowerCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None ) A = tmp_path / 'cache' A = JsonDatasetReader(lowerCAmelCase, features=lowerCAmelCase, cache_dir=lowerCAmelCase ).read() assert isinstance(lowerCAmelCase, lowerCAmelCase ) assert dataset.num_rows == 2 assert dataset.num_columns == 3 assert dataset.column_names == ["col_2", "col_3", "col_1"] for feature, expected_dtype in expected_features.items(): assert dataset.features[feature].dtype == expected_dtype @pytest.mark.parametrize('split', [None, NamedSplit('train' ), 'train', 'test'] ) def __UpperCamelCase (lowerCAmelCase : Union[str, Any], lowerCAmelCase : Any, lowerCAmelCase : Union[str, Any] ) -> str: A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} A = JsonDatasetReader(lowerCAmelCase, cache_dir=lowerCAmelCase, split=lowerCAmelCase ).read() _check_json_dataset(lowerCAmelCase, lowerCAmelCase ) assert dataset.split == split if split else "train" @pytest.mark.parametrize('path_type', [str, list] ) def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Any, lowerCAmelCase : Tuple ) -> Any: if issubclass(lowerCAmelCase, lowerCAmelCase ): A = jsonl_path elif issubclass(lowerCAmelCase, lowerCAmelCase ): A = [jsonl_path] A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} A = JsonDatasetReader(lowerCAmelCase, cache_dir=lowerCAmelCase ).read() _check_json_dataset(lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : List[Any], lowerCAmelCase : Dict, lowerCAmelCase : List[Any]=("train",) ) -> str: assert isinstance(lowerCAmelCase, lowerCAmelCase ) for split in splits: A = dataset_dict[split] assert dataset.num_rows == 4 assert dataset.num_columns == 3 assert dataset.column_names == ["col_1", "col_2", "col_3"] for feature, expected_dtype in expected_features.items(): assert dataset.features[feature].dtype == expected_dtype @pytest.mark.parametrize('keep_in_memory', [False, True] ) def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Tuple, lowerCAmelCase : Optional[int] ) -> int: A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} with assert_arrow_memory_increases() if keep_in_memory else assert_arrow_memory_doesnt_increase(): A = JsonDatasetReader({'train': jsonl_path}, cache_dir=lowerCAmelCase, keep_in_memory=lowerCAmelCase ).read() _check_json_datasetdict(lowerCAmelCase, lowerCAmelCase ) @pytest.mark.parametrize( 'features', [ None, {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'}, {'col_1': 'string', 'col_2': 'string', 'col_3': 'string'}, {'col_1': 'int32', 'col_2': 'int32', 'col_3': 'int32'}, {'col_1': 'float32', 'col_2': 'float32', 'col_3': 'float32'}, ], ) def __UpperCamelCase (lowerCAmelCase : Dict, lowerCAmelCase : List[Any], lowerCAmelCase : Tuple ) -> Any: A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} A = features.copy() if features else default_expected_features A = ( Features({feature: Value(lowerCAmelCase ) for feature, dtype in features.items()} ) if features is not None else None ) A = JsonDatasetReader({'train': jsonl_path}, features=lowerCAmelCase, cache_dir=lowerCAmelCase ).read() _check_json_datasetdict(lowerCAmelCase, lowerCAmelCase ) @pytest.mark.parametrize('split', [None, NamedSplit('train' ), 'train', 'test'] ) def __UpperCamelCase (lowerCAmelCase : Any, lowerCAmelCase : Optional[int], lowerCAmelCase : Optional[int] ) -> List[Any]: if split: A = {split: jsonl_path} else: A = 'train' A = {'train': jsonl_path, 'test': jsonl_path} A = tmp_path / 'cache' A = {'col_1': 'string', 'col_2': 'int64', 'col_3': 'float64'} A = JsonDatasetReader(lowerCAmelCase, cache_dir=lowerCAmelCase ).read() _check_json_datasetdict(lowerCAmelCase, lowerCAmelCase, splits=list(path.keys() ) ) assert all(dataset[split].split == split for split in path.keys() ) def __UpperCamelCase (lowerCAmelCase : Any ) -> Any: return json.load(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : Union[str, Any] ) -> Union[str, Any]: return [json.loads(lowerCAmelCase ) for line in buffer] class _UpperCAmelCase : '''simple docstring''' @pytest.mark.parametrize('lines, load_json_function' , [(True, load_json_lines), (False, load_json)] ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : Union[str, Any] ): with io.BytesIO() as buffer: JsonDatasetWriter(UpperCamelCase__ , UpperCamelCase__ , lines=UpperCamelCase__ ).write() buffer.seek(0 ) A = load_json_function(UpperCamelCase__ ) assert isinstance(UpperCamelCase__ , UpperCamelCase__ ) assert isinstance(exported_content[0] , UpperCamelCase__ ) assert len(UpperCamelCase__ ) == 10 @pytest.mark.parametrize( 'orient, container, keys, len_at' , [ ('records', list, {'tokens', 'labels', 'answers', 'id'}, None), ('split', dict, {'columns', 'data'}, 'data'), ('index', dict, set('0123456789' ), None), ('columns', dict, {'tokens', 'labels', 'answers', 'id'}, 'tokens'), ('values', list, None, None), ('table', dict, {'schema', 'data'}, 'data'), ] , ) def UpperCamelCase ( self : str , UpperCamelCase__ : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Tuple ): with io.BytesIO() as buffer: JsonDatasetWriter(UpperCamelCase__ , UpperCamelCase__ , lines=UpperCamelCase__ , orient=UpperCamelCase__ ).write() buffer.seek(0 ) A = load_json(UpperCamelCase__ ) assert isinstance(UpperCamelCase__ , UpperCamelCase__ ) if keys: if container is dict: assert exported_content.keys() == keys else: assert exported_content[0].keys() == keys else: assert not hasattr(UpperCamelCase__ , 'keys' ) and not hasattr(exported_content[0] , 'keys' ) if len_at: assert len(exported_content[len_at] ) == 10 else: assert len(UpperCamelCase__ ) == 10 @pytest.mark.parametrize('lines, load_json_function' , [(True, load_json_lines), (False, load_json)] ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : str ): with io.BytesIO() as buffer: JsonDatasetWriter(UpperCamelCase__ , UpperCamelCase__ , lines=UpperCamelCase__ , num_proc=2 ).write() buffer.seek(0 ) A = load_json_function(UpperCamelCase__ ) assert isinstance(UpperCamelCase__ , UpperCamelCase__ ) assert isinstance(exported_content[0] , UpperCamelCase__ ) assert len(UpperCamelCase__ ) == 10 @pytest.mark.parametrize( 'orient, container, keys, len_at' , [ ('records', list, {'tokens', 'labels', 'answers', 'id'}, None), ('split', dict, {'columns', 'data'}, 'data'), ('index', dict, set('0123456789' ), None), ('columns', dict, {'tokens', 'labels', 'answers', 'id'}, 'tokens'), ('values', list, None, None), ('table', dict, {'schema', 'data'}, 'data'), ] , ) def UpperCamelCase ( self : str , UpperCamelCase__ : str , UpperCamelCase__ : Any , UpperCamelCase__ : int , UpperCamelCase__ : Dict , UpperCamelCase__ : Any ): with io.BytesIO() as buffer: JsonDatasetWriter(UpperCamelCase__ , UpperCamelCase__ , lines=UpperCamelCase__ , orient=UpperCamelCase__ , num_proc=2 ).write() buffer.seek(0 ) A = load_json(UpperCamelCase__ ) assert isinstance(UpperCamelCase__ , UpperCamelCase__ ) if keys: if container is dict: assert exported_content.keys() == keys else: assert exported_content[0].keys() == keys else: assert not hasattr(UpperCamelCase__ , 'keys' ) and not hasattr(exported_content[0] , 'keys' ) if len_at: assert len(exported_content[len_at] ) == 10 else: assert len(UpperCamelCase__ ) == 10 def UpperCamelCase ( self : str , UpperCamelCase__ : Any ): with pytest.raises(UpperCamelCase__ ): with io.BytesIO() as buffer: JsonDatasetWriter(UpperCamelCase__ , UpperCamelCase__ , num_proc=0 ) @pytest.mark.parametrize('compression, extension' , [('gzip', 'gz'), ('bz2', 'bz2'), ('xz', 'xz')] ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Dict , UpperCamelCase__ : str , UpperCamelCase__ : int , UpperCamelCase__ : Any , UpperCamelCase__ : List[Any] ): A = tmp_path_factory.mktemp('data' ) / f'''test.json.{extension}''' A = str(shared_datadir / f'''test_file.json.{extension}''' ) JsonDatasetWriter(UpperCamelCase__ , UpperCamelCase__ , compression=UpperCamelCase__ ).write() with fsspec.open(UpperCamelCase__ , 'rb' , compression='infer' ) as f: A = f.read() with fsspec.open(UpperCamelCase__ , 'rb' , compression='infer' ) as f: A = f.read() assert exported_content == original_content
699
import re import jax.numpy as jnp from flax.traverse_util import flatten_dict, unflatten_dict from jax.random import PRNGKey from ..utils import logging _UpperCAmelCase = logging.get_logger(__name__) def __UpperCamelCase (lowerCAmelCase : List[str] ) -> Dict: A = r'\w+[.]\d+' A = re.findall(lowerCAmelCase, lowerCAmelCase ) for pat in pats: A = key.replace(lowerCAmelCase, '_'.join(pat.split('.' ) ) ) return key def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Dict, lowerCAmelCase : Dict ) -> Any: A = pt_tuple_key[:-1] + ('scale',) if ( any('norm' in str_ for str_ in pt_tuple_key ) and (pt_tuple_key[-1] == "bias") and (pt_tuple_key[:-1] + ("bias",) not in random_flax_state_dict) and (pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict) ): A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor elif pt_tuple_key[-1] in ["weight", "gamma"] and pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor # embedding if pt_tuple_key[-1] == "weight" and pt_tuple_key[:-1] + ("embedding",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('embedding',) return renamed_pt_tuple_key, pt_tensor # conv layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight" and pt_tensor.ndim == 4: A = pt_tensor.transpose(2, 3, 1, 0 ) return renamed_pt_tuple_key, pt_tensor # linear layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight": A = pt_tensor.T return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm weight A = pt_tuple_key[:-1] + ('weight',) if pt_tuple_key[-1] == "gamma": return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm bias A = pt_tuple_key[:-1] + ('bias',) if pt_tuple_key[-1] == "beta": return renamed_pt_tuple_key, pt_tensor return pt_tuple_key, pt_tensor def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Any, lowerCAmelCase : str=42 ) -> Any: # Step 1: Convert pytorch tensor to numpy A = {k: v.numpy() for k, v in pt_state_dict.items()} # Step 2: Since the model is stateless, get random Flax params A = flax_model.init_weights(PRNGKey(lowerCAmelCase ) ) A = flatten_dict(lowerCAmelCase ) A = {} # Need to change some parameters name to match Flax names for pt_key, pt_tensor in pt_state_dict.items(): A = rename_key(lowerCAmelCase ) A = tuple(renamed_pt_key.split('.' ) ) # Correctly rename weight parameters A , A = rename_key_and_reshape_tensor(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if flax_key in random_flax_state_dict: if flax_tensor.shape != random_flax_state_dict[flax_key].shape: raise ValueError( f'''PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape ''' f'''{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.''' ) # also add unexpected weight so that warning is thrown A = jnp.asarray(lowerCAmelCase ) return unflatten_dict(lowerCAmelCase )
699
1
import gc import unittest import numpy as np import torch import torch.nn.functional as F from transformers import ( ClapTextConfig, ClapTextModelWithProjection, RobertaTokenizer, SpeechTaHifiGan, SpeechTaHifiGanConfig, ) from diffusers import ( AudioLDMPipeline, AutoencoderKL, DDIMScheduler, LMSDiscreteScheduler, PNDMScheduler, UNetaDConditionModel, ) from diffusers.utils import is_xformers_available, slow, torch_device from diffusers.utils.testing_utils import enable_full_determinism from ..pipeline_params import TEXT_TO_AUDIO_BATCH_PARAMS, TEXT_TO_AUDIO_PARAMS from ..test_pipelines_common import PipelineTesterMixin enable_full_determinism() class _UpperCAmelCase ( __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = AudioLDMPipeline SCREAMING_SNAKE_CASE : Union[str, Any] = TEXT_TO_AUDIO_PARAMS SCREAMING_SNAKE_CASE : Union[str, Any] = TEXT_TO_AUDIO_BATCH_PARAMS SCREAMING_SNAKE_CASE : Optional[int] = frozenset( [ '''num_inference_steps''', '''num_waveforms_per_prompt''', '''generator''', '''latents''', '''output_type''', '''return_dict''', '''callback''', '''callback_steps''', ] ) def UpperCamelCase ( self : Dict ): torch.manual_seed(0 ) A = UNetaDConditionModel( block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=('DownBlock2D', 'CrossAttnDownBlock2D') , up_block_types=('CrossAttnUpBlock2D', 'UpBlock2D') , cross_attention_dim=(32, 64) , class_embed_type='simple_projection' , projection_class_embeddings_input_dim=32 , class_embeddings_concat=UpperCamelCase__ , ) A = DDIMScheduler( beta_start=0.00_085 , beta_end=0.012 , beta_schedule='scaled_linear' , clip_sample=UpperCamelCase__ , set_alpha_to_one=UpperCamelCase__ , ) torch.manual_seed(0 ) A = AutoencoderKL( block_out_channels=[32, 64] , in_channels=1 , out_channels=1 , down_block_types=['DownEncoderBlock2D', 'DownEncoderBlock2D'] , up_block_types=['UpDecoderBlock2D', 'UpDecoderBlock2D'] , latent_channels=4 , ) torch.manual_seed(0 ) A = ClapTextConfig( bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-0_5 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1000 , projection_dim=32 , ) A = ClapTextModelWithProjection(UpperCamelCase__ ) A = RobertaTokenizer.from_pretrained('hf-internal-testing/tiny-random-roberta' , model_max_length=77 ) A = SpeechTaHifiGanConfig( model_in_dim=8 , sampling_rate=16000 , upsample_initial_channel=16 , upsample_rates=[2, 2] , upsample_kernel_sizes=[4, 4] , resblock_kernel_sizes=[3, 7] , resblock_dilation_sizes=[[1, 3, 5], [1, 3, 5]] , normalize_before=UpperCamelCase__ , ) A = SpeechTaHifiGan(UpperCamelCase__ ) A = { 'unet': unet, 'scheduler': scheduler, 'vae': vae, 'text_encoder': text_encoder, 'tokenizer': tokenizer, 'vocoder': vocoder, } return components def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : Tuple=0 ): if str(UpperCamelCase__ ).startswith('mps' ): A = torch.manual_seed(UpperCamelCase__ ) else: A = torch.Generator(device=UpperCamelCase__ ).manual_seed(UpperCamelCase__ ) A = { 'prompt': 'A hammer hitting a wooden surface', 'generator': generator, 'num_inference_steps': 2, 'guidance_scale': 6.0, } return inputs def UpperCamelCase ( self : Tuple ): A = 'cpu' # ensure determinism for the device-dependent torch.Generator A = self.get_dummy_components() A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = self.get_dummy_inputs(UpperCamelCase__ ) A = audioldm_pipe(**UpperCamelCase__ ) A = output.audios[0] assert audio.ndim == 1 assert len(UpperCamelCase__ ) == 256 A = audio[:10] A = np.array( [-0.0_050, 0.0_050, -0.0_060, 0.0_033, -0.0_026, 0.0_033, -0.0_027, 0.0_033, -0.0_028, 0.0_033] ) assert np.abs(audio_slice - expected_slice ).max() < 1e-2 def UpperCamelCase ( self : Tuple ): A = self.get_dummy_components() A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = self.get_dummy_inputs(UpperCamelCase__ ) A = 3 * [inputs['prompt']] # forward A = audioldm_pipe(**UpperCamelCase__ ) A = output.audios[0] A = self.get_dummy_inputs(UpperCamelCase__ ) A = 3 * [inputs.pop('prompt' )] A = audioldm_pipe.tokenizer( UpperCamelCase__ , padding='max_length' , max_length=audioldm_pipe.tokenizer.model_max_length , truncation=UpperCamelCase__ , return_tensors='pt' , ) A = text_inputs['input_ids'].to(UpperCamelCase__ ) A = audioldm_pipe.text_encoder( UpperCamelCase__ , ) A = prompt_embeds.text_embeds # additional L_2 normalization over each hidden-state A = F.normalize(UpperCamelCase__ , dim=-1 ) A = prompt_embeds # forward A = audioldm_pipe(**UpperCamelCase__ ) A = output.audios[0] assert np.abs(audio_a - audio_a ).max() < 1e-2 def UpperCamelCase ( self : Optional[Any] ): A = self.get_dummy_components() A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = self.get_dummy_inputs(UpperCamelCase__ ) A = 3 * ['this is a negative prompt'] A = negative_prompt A = 3 * [inputs['prompt']] # forward A = audioldm_pipe(**UpperCamelCase__ ) A = output.audios[0] A = self.get_dummy_inputs(UpperCamelCase__ ) A = 3 * [inputs.pop('prompt' )] A = [] for p in [prompt, negative_prompt]: A = audioldm_pipe.tokenizer( UpperCamelCase__ , padding='max_length' , max_length=audioldm_pipe.tokenizer.model_max_length , truncation=UpperCamelCase__ , return_tensors='pt' , ) A = text_inputs['input_ids'].to(UpperCamelCase__ ) A = audioldm_pipe.text_encoder( UpperCamelCase__ , ) A = text_embeds.text_embeds # additional L_2 normalization over each hidden-state A = F.normalize(UpperCamelCase__ , dim=-1 ) embeds.append(UpperCamelCase__ ) A , A = embeds # forward A = audioldm_pipe(**UpperCamelCase__ ) A = output.audios[0] assert np.abs(audio_a - audio_a ).max() < 1e-2 def UpperCamelCase ( self : Tuple ): A = 'cpu' # ensure determinism for the device-dependent torch.Generator A = self.get_dummy_components() A = PNDMScheduler(skip_prk_steps=UpperCamelCase__ ) A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = self.get_dummy_inputs(UpperCamelCase__ ) A = 'egg cracking' A = audioldm_pipe(**UpperCamelCase__ , negative_prompt=UpperCamelCase__ ) A = output.audios[0] assert audio.ndim == 1 assert len(UpperCamelCase__ ) == 256 A = audio[:10] A = np.array( [-0.0_051, 0.0_050, -0.0_060, 0.0_034, -0.0_026, 0.0_033, -0.0_027, 0.0_033, -0.0_028, 0.0_032] ) assert np.abs(audio_slice - expected_slice ).max() < 1e-2 def UpperCamelCase ( self : Optional[Any] ): A = 'cpu' # ensure determinism for the device-dependent torch.Generator A = self.get_dummy_components() A = PNDMScheduler(skip_prk_steps=UpperCamelCase__ ) A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = 'A hammer hitting a wooden surface' # test num_waveforms_per_prompt=1 (default) A = audioldm_pipe(UpperCamelCase__ , num_inference_steps=2 ).audios assert audios.shape == (1, 256) # test num_waveforms_per_prompt=1 (default) for batch of prompts A = 2 A = audioldm_pipe([prompt] * batch_size , num_inference_steps=2 ).audios assert audios.shape == (batch_size, 256) # test num_waveforms_per_prompt for single prompt A = 2 A = audioldm_pipe(UpperCamelCase__ , num_inference_steps=2 , num_waveforms_per_prompt=UpperCamelCase__ ).audios assert audios.shape == (num_waveforms_per_prompt, 256) # test num_waveforms_per_prompt for batch of prompts A = 2 A = audioldm_pipe( [prompt] * batch_size , num_inference_steps=2 , num_waveforms_per_prompt=UpperCamelCase__ ).audios assert audios.shape == (batch_size * num_waveforms_per_prompt, 256) def UpperCamelCase ( self : Any ): A = 'cpu' # ensure determinism for the device-dependent torch.Generator A = self.get_dummy_components() A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = audioldm_pipe.vocoder.config.sampling_rate A = self.get_dummy_inputs(UpperCamelCase__ ) A = audioldm_pipe(audio_length_in_s=0.016 , **UpperCamelCase__ ) A = output.audios[0] assert audio.ndim == 1 assert len(UpperCamelCase__ ) / vocoder_sampling_rate == 0.016 A = audioldm_pipe(audio_length_in_s=0.032 , **UpperCamelCase__ ) A = output.audios[0] assert audio.ndim == 1 assert len(UpperCamelCase__ ) / vocoder_sampling_rate == 0.032 def UpperCamelCase ( self : int ): A = self.get_dummy_components() A = AudioLDMPipeline(**UpperCamelCase__ ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = ['hey'] A = audioldm_pipe(UpperCamelCase__ , num_inference_steps=1 ) A = output.audios.shape assert audio_shape == (1, 256) A = audioldm_pipe.vocoder.config config.model_in_dim *= 2 A = SpeechTaHifiGan(UpperCamelCase__ ).to(UpperCamelCase__ ) A = audioldm_pipe(UpperCamelCase__ , num_inference_steps=1 ) A = output.audios.shape # waveform shape is unchanged, we just have 2x the number of mel channels in the spectrogram assert audio_shape == (1, 256) def UpperCamelCase ( self : List[str] ): self._test_attention_slicing_forward_pass(test_mean_pixel_difference=UpperCamelCase__ ) def UpperCamelCase ( self : Dict ): self._test_inference_batch_single_identical(test_mean_pixel_difference=UpperCamelCase__ ) @unittest.skipIf( torch_device != 'cuda' or not is_xformers_available() , reason='XFormers attention is only available with CUDA and `xformers` installed' , ) def UpperCamelCase ( self : Tuple ): self._test_xformers_attention_forwardGenerator_pass(test_mean_pixel_difference=UpperCamelCase__ ) @slow class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : Tuple ): super().tearDown() gc.collect() torch.cuda.empty_cache() def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : List[Any]="cpu" , UpperCamelCase__ : Optional[int]=torch.floataa , UpperCamelCase__ : Union[str, Any]=0 ): A = torch.Generator(device=UpperCamelCase__ ).manual_seed(UpperCamelCase__ ) A = np.random.RandomState(UpperCamelCase__ ).standard_normal((1, 8, 128, 16) ) A = torch.from_numpy(UpperCamelCase__ ).to(device=UpperCamelCase__ , dtype=UpperCamelCase__ ) A = { 'prompt': 'A hammer hitting a wooden surface', 'latents': latents, 'generator': generator, 'num_inference_steps': 3, 'guidance_scale': 2.5, } return inputs def UpperCamelCase ( self : Optional[int] ): A = AudioLDMPipeline.from_pretrained('cvssp/audioldm' ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = self.get_inputs(UpperCamelCase__ ) A = 25 A = audioldm_pipe(**UpperCamelCase__ ).audios[0] assert audio.ndim == 1 assert len(UpperCamelCase__ ) == 81920 A = audio[77230:77240] A = np.array( [-0.4_884, -0.4_607, 0.0_023, 0.5_007, 0.5_896, 0.5_151, 0.3_813, -0.0_208, -0.3_687, -0.4_315] ) A = np.abs(expected_slice - audio_slice ).max() assert max_diff < 1e-2 def UpperCamelCase ( self : Any ): A = AudioLDMPipeline.from_pretrained('cvssp/audioldm' ) A = LMSDiscreteScheduler.from_config(audioldm_pipe.scheduler.config ) A = audioldm_pipe.to(UpperCamelCase__ ) audioldm_pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = self.get_inputs(UpperCamelCase__ ) A = audioldm_pipe(**UpperCamelCase__ ).audios[0] assert audio.ndim == 1 assert len(UpperCamelCase__ ) == 81920 A = audio[27780:27790] A = np.array([-0.2_131, -0.0_873, -0.0_124, -0.0_189, 0.0_569, 0.1_373, 0.1_883, 0.2_886, 0.3_297, 0.2_212] ) A = np.abs(expected_slice - audio_slice ).max() assert max_diff < 3e-2
699
from __future__ import annotations import math import random from collections.abc import Collection from typing import overload class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Collection[float] | None = None ): if components is None: A = [] A = list(UpperCamelCase__ ) def __len__( self : List[Any] ): return len(self.__components ) def __str__( self : str ): return "(" + ",".join(map(UpperCamelCase__ , self.__components ) ) + ")" def __add__( self : str , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] + other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: raise Exception('must have the same size' ) def __sub__( self : Dict , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] - other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: # error case raise Exception('must have the same size' ) @overload def __mul__( self : Tuple , UpperCamelCase__ : float ): ... @overload def __mul__( self : Dict , UpperCamelCase__ : Vector ): ... def __mul__( self : Union[str, Any] , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , (float, int) ): A = [c * other for c in self.__components] return Vector(UpperCamelCase__ ) elif isinstance(UpperCamelCase__ , UpperCamelCase__ ) and len(self ) == len(UpperCamelCase__ ): A = len(self ) A = [self.__components[i] * other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return sum(UpperCamelCase__ ) else: # error case raise Exception('invalid operand!' ) def UpperCamelCase ( self : Union[str, Any] ): return Vector(self.__components ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ) and -len(self.__components ) <= i < len(self.__components ): return self.__components[i] else: raise Exception('index out of range' ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : float ): assert -len(self.__components ) <= pos < len(self.__components ) A = value def UpperCamelCase ( self : str ): if len(self.__components ) == 0: raise Exception('Vector is empty' ) A = [c**2 for c in self.__components] return math.sqrt(sum(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Vector , UpperCamelCase__ : bool = False ): A = self * other A = self.euclidean_length() * other.euclidean_length() if deg: return math.degrees(math.acos(num / den ) ) else: return math.acos(num / den ) def __UpperCamelCase (lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) return Vector([0] * dimension ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, lowerCAmelCase )) A = [0] * dimension A = 1 return Vector(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : float, lowerCAmelCase : Vector, lowerCAmelCase : Vector ) -> Vector: assert ( isinstance(lowerCAmelCase, lowerCAmelCase ) and isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, (int, float) )) ) return x * scalar + y def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: random.seed(lowerCAmelCase ) A = [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] return Vector(lowerCAmelCase ) class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : list[list[float]] , UpperCamelCase__ : int , UpperCamelCase__ : int ): A = matrix A = w A = h def __str__( self : int ): A = '' for i in range(self.__height ): ans += "|" for j in range(self.__width ): if j < self.__width - 1: ans += str(self.__matrix[i][j] ) + "," else: ans += str(self.__matrix[i][j] ) + "|\n" return ans def __add__( self : Optional[Any] , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] + other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrix must have the same dimension!' ) def __sub__( self : Dict , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] - other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrices must have the same dimension!' ) @overload def __mul__( self : int , UpperCamelCase__ : float ): ... @overload def __mul__( self : Union[str, Any] , UpperCamelCase__ : Vector ): ... def __mul__( self : Tuple , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): # matrix-vector if len(UpperCamelCase__ ) == self.__width: A = zero_vector(self.__height ) for i in range(self.__height ): A = [ self.__matrix[i][j] * other.component(UpperCamelCase__ ) for j in range(self.__width ) ] ans.change_component(UpperCamelCase__ , sum(UpperCamelCase__ ) ) return ans else: raise Exception( 'vector must have the same size as the ' 'number of columns of the matrix!' ) elif isinstance(UpperCamelCase__ , (int, float) ): # matrix-scalar A = [ [self.__matrix[i][j] * other for j in range(self.__width )] for i in range(self.__height ) ] return Matrix(UpperCamelCase__ , self.__width , self.__height ) return None def UpperCamelCase ( self : Optional[int] ): return self.__height def UpperCamelCase ( self : List[Any] ): return self.__width def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int ): if 0 <= x < self.__height and 0 <= y < self.__width: return self.__matrix[x][y] else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : float ): if 0 <= x < self.__height and 0 <= y < self.__width: A = value else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) A = self.__matrix[:x] + self.__matrix[x + 1 :] for i in range(len(UpperCamelCase__ ) ): A = minor[i][:y] + minor[i][y + 1 :] return Matrix(UpperCamelCase__ , self.__width - 1 , self.__height - 1 ).determinant() def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if 0 <= x < self.__height and 0 <= y < self.__width: return (-1) ** (x + y) * self.minor(UpperCamelCase__ , UpperCamelCase__ ) else: raise Exception('Indices out of bounds' ) def UpperCamelCase ( self : Tuple ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if self.__height < 1: raise Exception('Matrix has no element' ) elif self.__height == 1: return self.__matrix[0][0] elif self.__height == 2: return ( self.__matrix[0][0] * self.__matrix[1][1] - self.__matrix[0][1] * self.__matrix[1][0] ) else: A = [ self.__matrix[0][y] * self.cofactor(0 , UpperCamelCase__ ) for y in range(self.__width ) ] return sum(UpperCamelCase__ ) def __UpperCamelCase (lowerCAmelCase : int ) -> Matrix: A = [[0] * n for _ in range(lowerCAmelCase )] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Matrix: random.seed(lowerCAmelCase ) A = [ [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] for _ in range(lowerCAmelCase ) ] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )
699
1
import argparse import re import torch from CLAP import create_model from transformers import AutoFeatureExtractor, ClapConfig, ClapModel _UpperCAmelCase = { "text_branch": "text_model", "audio_branch": "audio_model.audio_encoder", "attn": "attention.self", "self.proj": "output.dense", "attention.self_mask": "attn_mask", "mlp.fc1": "intermediate.dense", "mlp.fc2": "output.dense", "norm1": "layernorm_before", "norm2": "layernorm_after", "bn0": "batch_norm", } _UpperCAmelCase = AutoFeatureExtractor.from_pretrained("laion/clap-htsat-unfused", truncation="rand_trunc") def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Union[str, Any]=False ) -> List[Any]: A , A = create_model( 'HTSAT-tiny', 'roberta', lowerCAmelCase, precision='fp32', device='cuda:0' if torch.cuda.is_available() else 'cpu', enable_fusion=lowerCAmelCase, fusion_type='aff_2d' if enable_fusion else None, ) return model, model_cfg def __UpperCamelCase (lowerCAmelCase : int ) -> Dict: A = {} A = r'.*sequential.(\d+).*' A = r'.*_projection.(\d+).*' for key, value in state_dict.items(): # check if any key needs to be modified for key_to_modify, new_key in KEYS_TO_MODIFY_MAPPING.items(): if key_to_modify in key: A = key.replace(lowerCAmelCase, lowerCAmelCase ) if re.match(lowerCAmelCase, lowerCAmelCase ): # replace sequential layers with list A = re.match(lowerCAmelCase, lowerCAmelCase ).group(1 ) A = key.replace(f'''sequential.{sequential_layer}.''', f'''layers.{int(lowerCAmelCase )//3}.linear.''' ) elif re.match(lowerCAmelCase, lowerCAmelCase ): A = int(re.match(lowerCAmelCase, lowerCAmelCase ).group(1 ) ) # Because in CLAP they use `nn.Sequential`... A = 1 if projecton_layer == 0 else 2 A = key.replace(f'''_projection.{projecton_layer}.''', f'''_projection.linear{transformers_projection_layer}.''' ) if "audio" and "qkv" in key: # split qkv into query key and value A = value A = mixed_qkv.size(0 ) // 3 A = mixed_qkv[:qkv_dim] A = mixed_qkv[qkv_dim : qkv_dim * 2] A = mixed_qkv[qkv_dim * 2 :] A = query_layer A = key_layer A = value_layer else: A = value return model_state_dict def __UpperCamelCase (lowerCAmelCase : Any, lowerCAmelCase : int, lowerCAmelCase : str, lowerCAmelCase : List[Any]=False ) -> int: A , A = init_clap(lowerCAmelCase, enable_fusion=lowerCAmelCase ) clap_model.eval() A = clap_model.state_dict() A = rename_state_dict(lowerCAmelCase ) A = ClapConfig() A = enable_fusion A = ClapModel(lowerCAmelCase ) # ignore the spectrogram embedding layer model.load_state_dict(lowerCAmelCase, strict=lowerCAmelCase ) model.save_pretrained(lowerCAmelCase ) transformers_config.save_pretrained(lowerCAmelCase ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") parser.add_argument("--enable_fusion", action="store_true", help="Whether to enable fusion or not") _UpperCAmelCase = parser.parse_args() convert_clap_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.enable_fusion)
699
from collections import OrderedDict from typing import Any, Mapping, Optional from ... import PreTrainedTokenizer from ...configuration_utils import PretrainedConfig from ...file_utils import TensorType, is_torch_available from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast from ...onnx.utils import compute_effective_axis_dimension from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "facebook/blenderbot_small-90M": "https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/config.json", # See all BlenderbotSmall models at https://huggingface.co/models?filter=blenderbot_small } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = '''blenderbot-small''' SCREAMING_SNAKE_CASE : Any = ['''past_key_values'''] SCREAMING_SNAKE_CASE : List[str] = {'''num_attention_heads''': '''encoder_attention_heads''', '''hidden_size''': '''d_model'''} def __init__( self : List[str] , UpperCamelCase__ : Optional[Any]=50265 , UpperCamelCase__ : Optional[int]=512 , UpperCamelCase__ : int=8 , UpperCamelCase__ : Optional[int]=2048 , UpperCamelCase__ : Optional[Any]=16 , UpperCamelCase__ : Optional[Any]=8 , UpperCamelCase__ : List[Any]=2048 , UpperCamelCase__ : int=16 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : int=True , UpperCamelCase__ : Optional[Any]="gelu" , UpperCamelCase__ : Any=512 , UpperCamelCase__ : Union[str, Any]=0.1 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Union[str, Any]=0.0 , UpperCamelCase__ : Dict=0.02 , UpperCamelCase__ : Optional[Any]=1 , UpperCamelCase__ : Any=False , UpperCamelCase__ : Dict=0 , UpperCamelCase__ : Optional[int]=1 , UpperCamelCase__ : List[str]=2 , UpperCamelCase__ : Dict=2 , **UpperCamelCase__ : List[str] , ): A = vocab_size A = max_position_embeddings A = d_model A = encoder_ffn_dim A = encoder_layers A = encoder_attention_heads A = decoder_ffn_dim A = decoder_layers A = decoder_attention_heads A = dropout A = attention_dropout A = activation_dropout A = activation_function A = init_std A = encoder_layerdrop A = decoder_layerdrop A = use_cache A = encoder_layers A = scale_embedding # scale factor will be sqrt(d_model) if True super().__init__( pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , is_encoder_decoder=UpperCamelCase__ , decoder_start_token_id=UpperCamelCase__ , forced_eos_token_id=UpperCamelCase__ , **UpperCamelCase__ , ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' @property def UpperCamelCase ( self : List[Any] ): if self.task in ["default", "seq2seq-lm"]: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A = {0: 'batch'} A = {0: 'batch', 1: 'past_decoder_sequence + sequence'} else: A = {0: 'batch', 1: 'decoder_sequence'} A = {0: 'batch', 1: 'decoder_sequence'} if self.use_past: self.fill_with_past_key_values_(UpperCamelCase__ , direction='inputs' ) elif self.task == "causal-lm": # TODO: figure this case out. A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} else: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ('decoder_input_ids', {0: 'batch', 1: 'decoder_sequence'}), ('decoder_attention_mask', {0: 'batch', 1: 'decoder_sequence'}), ] ) return common_inputs @property def UpperCamelCase ( self : int ): if self.task in ["default", "seq2seq-lm"]: A = super().outputs else: A = super(UpperCamelCase__ , self ).outputs if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} return common_outputs def UpperCamelCase ( self : int , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) # Generate decoder inputs A = seq_length if not self.use_past else 1 A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) A = {f'''decoder_{name}''': tensor for name, tensor in decoder_inputs.items()} A = dict(**UpperCamelCase__ , **UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape A = common_inputs['decoder_input_ids'].shape[1] A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, encoder_seq_length, self._config.hidden_size // num_encoder_attention_heads, ) A = decoder_seq_length + 3 A = ( batch, num_decoder_attention_heads, decoder_past_length, self._config.hidden_size // num_decoder_attention_heads, ) A = torch.cat( [common_inputs['decoder_attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ )] , dim=1 ) A = [] # If the number of encoder and decoder layers are present in the model configuration, both are considered A , A = self.num_layers A = min(UpperCamelCase__ , UpperCamelCase__ ) A = max(UpperCamelCase__ , UpperCamelCase__ ) - min_num_layers A = 'encoder' if num_encoder_layers > num_decoder_layers else 'decoder' for _ in range(UpperCamelCase__ ): common_inputs["past_key_values"].append( ( torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), ) ) # TODO: test this. A = encoder_shape if remaining_side_name == 'encoder' else decoder_shape for _ in range(UpperCamelCase__ , UpperCamelCase__ ): common_inputs["past_key_values"].append((torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) ) return common_inputs def UpperCamelCase ( self : Tuple , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape # Not using the same length for past_key_values A = seqlen + 2 A , A = self.num_layers A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, past_key_values_length, self._config.hidden_size // num_encoder_attention_heads, ) A = common_inputs['attention_mask'].dtype A = torch.cat( [common_inputs['attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ , dtype=UpperCamelCase__ )] , dim=1 ) A = [ (torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) for _ in range(UpperCamelCase__ ) ] return common_inputs def UpperCamelCase ( self : List[str] , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): # Copied from OnnxConfig.generate_dummy_inputs # Did not use super(OnnxConfigWithPast, self).generate_dummy_inputs for code clarity. # If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 ) # If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX A = tokenizer.num_special_tokens_to_add(UpperCamelCase__ ) A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=UpperCamelCase__ ) # Generate dummy inputs according to compute batch and sequence A = [' '.join([tokenizer.unk_token] ) * seq_length] * batch_size A = dict(tokenizer(UpperCamelCase__ , return_tensors=UpperCamelCase__ ) ) return common_inputs def UpperCamelCase ( self : Any , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): if self.task in ["default", "seq2seq-lm"]: A = self._generate_dummy_inputs_for_default_and_seqaseq_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) elif self.task == "causal-lm": A = self._generate_dummy_inputs_for_causal_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) else: A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) return common_inputs def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : List[str] , UpperCamelCase__ : str , UpperCamelCase__ : Tuple ): if self.task in ["default", "seq2seq-lm"]: A = super()._flatten_past_key_values_(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) else: A = super(UpperCamelCase__ , self )._flatten_past_key_values_( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
699
1
from math import factorial _UpperCAmelCase = {str(digit): factorial(digit) for digit in range(10)} def __UpperCamelCase (lowerCAmelCase : int ) -> int: if not isinstance(lowerCAmelCase, lowerCAmelCase ): raise TypeError('Parameter number must be int' ) if number < 0: raise ValueError('Parameter number must be greater than or equal to 0' ) # Converts number in string to iterate on its digits and adds its factorial. return sum(DIGIT_FACTORIAL[digit] for digit in str(lowerCAmelCase ) ) def __UpperCamelCase (lowerCAmelCase : int = 60, lowerCAmelCase : int = 1_000_000 ) -> int: if not isinstance(lowerCAmelCase, lowerCAmelCase ) or not isinstance(lowerCAmelCase, lowerCAmelCase ): raise TypeError('Parameters chain_length and number_limit must be int' ) if chain_length <= 0 or number_limit <= 0: raise ValueError( 'Parameters chain_length and number_limit must be greater than 0' ) # the counter for the chains with the exact desired length A = 0 # the cached sizes of the previous chains A = {} for start_chain_element in range(1, lowerCAmelCase ): # The temporary set will contain the elements of the chain A = set() A = 0 # Stop computing the chain when you find a cached size, a repeating item or the # length is greater then the desired one. A = start_chain_element while ( chain_element not in chain_sets_lengths and chain_element not in chain_set and chain_set_length <= chain_length ): chain_set.add(lowerCAmelCase ) chain_set_length += 1 A = digit_factorial_sum(lowerCAmelCase ) if chain_element in chain_sets_lengths: chain_set_length += chain_sets_lengths[chain_element] A = chain_set_length # If chain contains the exact amount of elements increase the counter if chain_set_length == chain_length: chains_counter += 1 return chains_counter if __name__ == "__main__": import doctest doctest.testmod() print(F'''{solution()}''')
699
from typing import List, Optional, Union from ...processing_utils import ProcessorMixin from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy from ...utils import TensorType class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = ['''image_processor''', '''tokenizer'''] SCREAMING_SNAKE_CASE : List[str] = '''BridgeTowerImageProcessor''' SCREAMING_SNAKE_CASE : Tuple = ('''RobertaTokenizer''', '''RobertaTokenizerFast''') def __init__( self : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int] ): super().__init__(UpperCamelCase__ , UpperCamelCase__ ) def __call__( self : Optional[int] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None , UpperCamelCase__ : bool = True , UpperCamelCase__ : Union[bool, str, PaddingStrategy] = False , UpperCamelCase__ : Union[bool, str, TruncationStrategy] = None , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : Optional[Union[str, TensorType]] = None , **UpperCamelCase__ : List[Any] , ): A = self.tokenizer( text=UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , padding=UpperCamelCase__ , truncation=UpperCamelCase__ , max_length=UpperCamelCase__ , stride=UpperCamelCase__ , pad_to_multiple_of=UpperCamelCase__ , return_token_type_ids=UpperCamelCase__ , return_attention_mask=UpperCamelCase__ , return_overflowing_tokens=UpperCamelCase__ , return_special_tokens_mask=UpperCamelCase__ , return_offsets_mapping=UpperCamelCase__ , return_length=UpperCamelCase__ , verbose=UpperCamelCase__ , return_tensors=UpperCamelCase__ , **UpperCamelCase__ , ) # add pixel_values + pixel_mask A = self.image_processor( UpperCamelCase__ , return_tensors=UpperCamelCase__ , do_normalize=UpperCamelCase__ , do_center_crop=UpperCamelCase__ , **UpperCamelCase__ ) encoding.update(UpperCamelCase__ ) return encoding def UpperCamelCase ( self : Dict , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : Any ): return self.tokenizer.batch_decode(*UpperCamelCase__ , **UpperCamelCase__ ) def UpperCamelCase ( self : int , *UpperCamelCase__ : int , **UpperCamelCase__ : List[str] ): return self.tokenizer.decode(*UpperCamelCase__ , **UpperCamelCase__ ) @property def UpperCamelCase ( self : Any ): A = self.tokenizer.model_input_names A = self.image_processor.model_input_names return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
699
1
import json import os import pickle import shutil import tempfile from unittest import TestCase from unittest.mock import patch import numpy as np from datasets import Dataset from transformers import is_faiss_available from transformers.models.bart.configuration_bart import BartConfig from transformers.models.bart.tokenization_bart import BartTokenizer from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES as DPR_VOCAB_FILES_NAMES from transformers.models.dpr.configuration_dpr import DPRConfig from transformers.models.dpr.tokenization_dpr import DPRContextEncoderTokenizer, DPRQuestionEncoderTokenizer from transformers.models.rag.configuration_rag import RagConfig from transformers.models.rag.retrieval_rag import CustomHFIndex, RagRetriever from transformers.models.roberta.tokenization_roberta import VOCAB_FILES_NAMES as BART_VOCAB_FILES_NAMES from transformers.testing_utils import require_faiss, require_sentencepiece, require_tokenizers, require_torch if is_faiss_available(): import faiss @require_faiss class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def UpperCamelCase ( self : str ): A = tempfile.mkdtemp() A = 8 # DPR tok A = [ '[UNK]', '[CLS]', '[SEP]', '[PAD]', '[MASK]', 'want', '##want', '##ed', 'wa', 'un', 'runn', '##ing', ',', 'low', 'lowest', ] A = os.path.join(self.tmpdirname , 'dpr_tokenizer' ) os.makedirs(UpperCamelCase__ , exist_ok=UpperCamelCase__ ) A = os.path.join(UpperCamelCase__ , DPR_VOCAB_FILES_NAMES['vocab_file'] ) with open(self.vocab_file , 'w' , encoding='utf-8' ) as vocab_writer: vocab_writer.write(''.join([x + '\n' for x in vocab_tokens] ) ) # BART tok A = [ 'l', 'o', 'w', 'e', 'r', 's', 't', 'i', 'd', 'n', '\u0120', '\u0120l', '\u0120n', '\u0120lo', '\u0120low', 'er', '\u0120lowest', '\u0120newer', '\u0120wider', '<unk>', ] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = ['#version: 0.2', '\u0120 l', '\u0120l o', '\u0120lo w', 'e r', ''] A = {'unk_token': '<unk>'} A = os.path.join(self.tmpdirname , 'bart_tokenizer' ) os.makedirs(UpperCamelCase__ , exist_ok=UpperCamelCase__ ) A = os.path.join(UpperCamelCase__ , BART_VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join(UpperCamelCase__ , BART_VOCAB_FILES_NAMES['merges_file'] ) with open(self.vocab_file , 'w' , encoding='utf-8' ) as fp: fp.write(json.dumps(UpperCamelCase__ ) + '\n' ) with open(self.merges_file , 'w' , encoding='utf-8' ) as fp: fp.write('\n'.join(UpperCamelCase__ ) ) def UpperCamelCase ( self : List[str] ): return DPRQuestionEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , 'dpr_tokenizer' ) ) def UpperCamelCase ( self : int ): return DPRContextEncoderTokenizer.from_pretrained(os.path.join(self.tmpdirname , 'dpr_tokenizer' ) ) def UpperCamelCase ( self : Union[str, Any] ): return BartTokenizer.from_pretrained(os.path.join(self.tmpdirname , 'bart_tokenizer' ) ) def UpperCamelCase ( self : Any ): shutil.rmtree(self.tmpdirname ) def UpperCamelCase ( self : List[Any] ): A = Dataset.from_dict( { 'id': ['0', '1'], 'text': ['foo', 'bar'], 'title': ['Foo', 'Bar'], 'embeddings': [np.ones(self.retrieval_vector_size ), 2 * np.ones(self.retrieval_vector_size )], } ) dataset.add_faiss_index('embeddings' , string_factory='Flat' , metric_type=faiss.METRIC_INNER_PRODUCT ) return dataset def UpperCamelCase ( self : Optional[Any] ): A = self.get_dummy_dataset() A = RagConfig( retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , ) with patch('transformers.models.rag.retrieval_rag.load_dataset' ) as mock_load_dataset: A = dataset A = RagRetriever( UpperCamelCase__ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , ) return retriever def UpperCamelCase ( self : Dict , UpperCamelCase__ : bool ): A = self.get_dummy_dataset() A = RagConfig( retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='custom' , ) if from_disk: A = os.path.join(self.tmpdirname , 'dataset' ) A = os.path.join(self.tmpdirname , 'index.faiss' ) dataset.get_index('embeddings' ).save(os.path.join(self.tmpdirname , 'index.faiss' ) ) dataset.drop_index('embeddings' ) dataset.save_to_disk(os.path.join(self.tmpdirname , 'dataset' ) ) del dataset A = RagRetriever( UpperCamelCase__ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , ) else: A = RagRetriever( UpperCamelCase__ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() , index=CustomHFIndex(config.retrieval_vector_size , UpperCamelCase__ ) , ) return retriever def UpperCamelCase ( self : Optional[Any] ): A = Dataset.from_dict( { 'id': ['0', '1'], 'text': ['foo', 'bar'], 'title': ['Foo', 'Bar'], 'embeddings': [np.ones(self.retrieval_vector_size + 1 ), 2 * np.ones(self.retrieval_vector_size + 1 )], } ) dataset.add_faiss_index('embeddings' , string_factory='Flat' , metric_type=faiss.METRIC_INNER_PRODUCT ) A = os.path.join(self.tmpdirname , 'hf_bert_base.hnswSQ8_correct_phi_128.c_index' ) dataset.save_faiss_index('embeddings' , index_file_name + '.index.dpr' ) pickle.dump(dataset['id'] , open(index_file_name + '.index_meta.dpr' , 'wb' ) ) A = os.path.join(self.tmpdirname , 'psgs_w100.tsv.pkl' ) A = {sample['id']: [sample['text'], sample['title']] for sample in dataset} pickle.dump(UpperCamelCase__ , open(UpperCamelCase__ , 'wb' ) ) A = RagConfig( retrieval_vector_size=self.retrieval_vector_size , question_encoder=DPRConfig().to_dict() , generator=BartConfig().to_dict() , index_name='legacy' , index_path=self.tmpdirname , ) A = RagRetriever( UpperCamelCase__ , question_encoder_tokenizer=self.get_dpr_tokenizer() , generator_tokenizer=self.get_bart_tokenizer() ) return retriever def UpperCamelCase ( self : Optional[int] ): A = 1 A = self.get_dummy_canonical_hf_index_retriever() A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A , A , A = retriever.retrieve(UpperCamelCase__ , n_docs=UpperCamelCase__ ) self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) ) self.assertEqual(len(UpperCamelCase__ ) , 2 ) self.assertEqual(sorted(doc_dicts[0] ) , ['embeddings', 'id', 'text', 'title'] ) self.assertEqual(len(doc_dicts[0]['id'] ) , UpperCamelCase__ ) self.assertEqual(doc_dicts[0]['id'][0] , '1' ) # max inner product is reached with second doc self.assertEqual(doc_dicts[1]['id'][0] , '0' ) # max inner product is reached with first doc self.assertListEqual(doc_ids.tolist() , [[1], [0]] ) def UpperCamelCase ( self : List[Any] ): A = self.get_dummy_canonical_hf_index_retriever() with tempfile.TemporaryDirectory() as tmp_dirname: with patch('transformers.models.rag.retrieval_rag.load_dataset' ) as mock_load_dataset: A = self.get_dummy_dataset() retriever.save_pretrained(UpperCamelCase__ ) A = RagRetriever.from_pretrained(UpperCamelCase__ ) self.assertIsInstance(UpperCamelCase__ , UpperCamelCase__ ) A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A = retriever.retrieve(UpperCamelCase__ , n_docs=1 ) self.assertTrue(out is not None ) def UpperCamelCase ( self : Optional[Any] ): A = 1 A = self.get_dummy_custom_hf_index_retriever(from_disk=UpperCamelCase__ ) A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A , A , A = retriever.retrieve(UpperCamelCase__ , n_docs=UpperCamelCase__ ) self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) ) self.assertEqual(len(UpperCamelCase__ ) , 2 ) self.assertEqual(sorted(doc_dicts[0] ) , ['embeddings', 'id', 'text', 'title'] ) self.assertEqual(len(doc_dicts[0]['id'] ) , UpperCamelCase__ ) self.assertEqual(doc_dicts[0]['id'][0] , '1' ) # max inner product is reached with second doc self.assertEqual(doc_dicts[1]['id'][0] , '0' ) # max inner product is reached with first doc self.assertListEqual(doc_ids.tolist() , [[1], [0]] ) def UpperCamelCase ( self : Tuple ): A = self.get_dummy_custom_hf_index_retriever(from_disk=UpperCamelCase__ ) with tempfile.TemporaryDirectory() as tmp_dirname: retriever.save_pretrained(UpperCamelCase__ ) A = RagRetriever.from_pretrained(UpperCamelCase__ ) self.assertIsInstance(UpperCamelCase__ , UpperCamelCase__ ) A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A = retriever.retrieve(UpperCamelCase__ , n_docs=1 ) self.assertTrue(out is not None ) def UpperCamelCase ( self : Tuple ): A = 1 A = self.get_dummy_custom_hf_index_retriever(from_disk=UpperCamelCase__ ) A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A , A , A = retriever.retrieve(UpperCamelCase__ , n_docs=UpperCamelCase__ ) self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) ) self.assertEqual(len(UpperCamelCase__ ) , 2 ) self.assertEqual(sorted(doc_dicts[0] ) , ['embeddings', 'id', 'text', 'title'] ) self.assertEqual(len(doc_dicts[0]['id'] ) , UpperCamelCase__ ) self.assertEqual(doc_dicts[0]['id'][0] , '1' ) # max inner product is reached with second doc self.assertEqual(doc_dicts[1]['id'][0] , '0' ) # max inner product is reached with first doc self.assertListEqual(doc_ids.tolist() , [[1], [0]] ) def UpperCamelCase ( self : List[Any] ): A = self.get_dummy_custom_hf_index_retriever(from_disk=UpperCamelCase__ ) with tempfile.TemporaryDirectory() as tmp_dirname: retriever.save_pretrained(UpperCamelCase__ ) A = RagRetriever.from_pretrained(UpperCamelCase__ ) self.assertIsInstance(UpperCamelCase__ , UpperCamelCase__ ) A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A = retriever.retrieve(UpperCamelCase__ , n_docs=1 ) self.assertTrue(out is not None ) def UpperCamelCase ( self : Dict ): A = 1 A = self.get_dummy_legacy_index_retriever() A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A , A , A = retriever.retrieve(UpperCamelCase__ , n_docs=UpperCamelCase__ ) self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) ) self.assertEqual(len(UpperCamelCase__ ) , 2 ) self.assertEqual(sorted(doc_dicts[0] ) , ['text', 'title'] ) self.assertEqual(len(doc_dicts[0]['text'] ) , UpperCamelCase__ ) self.assertEqual(doc_dicts[0]['text'][0] , 'bar' ) # max inner product is reached with second doc self.assertEqual(doc_dicts[1]['text'][0] , 'foo' ) # max inner product is reached with first doc self.assertListEqual(doc_ids.tolist() , [[1], [0]] ) def UpperCamelCase ( self : str ): A = self.get_dummy_legacy_index_retriever() with tempfile.TemporaryDirectory() as tmp_dirname: retriever.save_pretrained(UpperCamelCase__ ) A = RagRetriever.from_pretrained(UpperCamelCase__ ) self.assertIsInstance(UpperCamelCase__ , UpperCamelCase__ ) A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A = retriever.retrieve(UpperCamelCase__ , n_docs=1 ) self.assertTrue(out is not None ) @require_torch @require_tokenizers @require_sentencepiece def UpperCamelCase ( self : Optional[Any] ): import torch A = 1 A = self.get_dummy_canonical_hf_index_retriever() A = [[5, 7], [10, 11]] A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A = retriever(UpperCamelCase__ , UpperCamelCase__ , prefix=retriever.config.generator.prefix , n_docs=UpperCamelCase__ ) A , A , A = ( out['context_input_ids'], out['context_attention_mask'], out['retrieved_doc_embeds'], ) self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) ) self.assertIsInstance(UpperCamelCase__ , UpperCamelCase__ ) self.assertIsInstance(UpperCamelCase__ , UpperCamelCase__ ) self.assertIsInstance(UpperCamelCase__ , np.ndarray ) A = retriever( UpperCamelCase__ , UpperCamelCase__ , prefix=retriever.config.generator.prefix , n_docs=UpperCamelCase__ , return_tensors='pt' , ) A , A , A , A = ( # noqa: F841 out['context_input_ids'], out['context_attention_mask'], out['retrieved_doc_embeds'], out['doc_ids'], ) self.assertEqual(retrieved_doc_embeds.shape , (2, n_docs, self.retrieval_vector_size) ) self.assertIsInstance(UpperCamelCase__ , torch.Tensor ) self.assertIsInstance(UpperCamelCase__ , torch.Tensor ) self.assertIsInstance(UpperCamelCase__ , torch.Tensor ) @require_torch @require_tokenizers @require_sentencepiece def UpperCamelCase ( self : str ): A = self.get_dpr_ctx_encoder_tokenizer() A = 1 A = self.get_dummy_custom_hf_index_retriever(from_disk=UpperCamelCase__ ) retriever.set_ctx_encoder_tokenizer(UpperCamelCase__ ) A = [[5, 7], [10, 11]] A = np.array( [np.ones(self.retrieval_vector_size ), -np.ones(self.retrieval_vector_size )] , dtype=np.floataa ) A = retriever(UpperCamelCase__ , UpperCamelCase__ , prefix=retriever.config.generator.prefix , n_docs=UpperCamelCase__ ) self.assertEqual( len(UpperCamelCase__ ) , 6 ) # check whether the retriever output consist of 6 attributes including tokenized docs self.assertEqual( all(k in out for k in ('tokenized_doc_ids', 'tokenized_doc_attention_mask') ) , UpperCamelCase__ ) # check for doc token related keys in dictionary.
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> str: 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=10))
699
1
import os def __UpperCamelCase () -> Optional[int]: A = 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())
699
import math import torch from torch import nn from ..configuration_utils import ConfigMixin, register_to_config from .attention_processor import Attention from .embeddings import get_timestep_embedding from .modeling_utils import ModelMixin class _UpperCAmelCase ( __lowercase , __lowercase ): '''simple docstring''' @register_to_config def __init__( self : Any , UpperCamelCase__ : int = 128 , UpperCamelCase__ : int = 256 , UpperCamelCase__ : float = 2_000.0 , UpperCamelCase__ : int = 768 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 64 , UpperCamelCase__ : int = 2048 , UpperCamelCase__ : float = 0.1 , ): super().__init__() A = nn.Sequential( nn.Linear(UpperCamelCase__ , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , nn.Linear(d_model * 4 , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , ) A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = False A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.ModuleList() for lyr_num in range(UpperCamelCase__ ): # FiLM conditional T5 decoder A = DecoderLayer(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) self.decoders.append(UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Any , UpperCamelCase__ : int ): A = torch.mul(query_input.unsqueeze(-1 ) , key_input.unsqueeze(-2 ) ) return mask.unsqueeze(-3 ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[int] ): A , A , A = decoder_input_tokens.shape assert decoder_noise_time.shape == (batch,) # decoder_noise_time is in [0, 1), so rescale to expected timing range. A = get_timestep_embedding( decoder_noise_time * self.config.max_decoder_noise_time , embedding_dim=self.config.d_model , max_period=self.config.max_decoder_noise_time , ).to(dtype=self.dtype ) A = self.conditioning_emb(UpperCamelCase__ ).unsqueeze(1 ) assert conditioning_emb.shape == (batch, 1, self.config.d_model * 4) A = decoder_input_tokens.shape[1] # If we want to use relative positions for audio context, we can just offset # this sequence by the length of encodings_and_masks. A = torch.broadcast_to( torch.arange(UpperCamelCase__ , device=decoder_input_tokens.device ) , (batch, seq_length) , ) A = self.position_encoding(UpperCamelCase__ ) A = self.continuous_inputs_projection(UpperCamelCase__ ) inputs += position_encodings A = self.dropout(UpperCamelCase__ ) # decoder: No padding present. A = torch.ones( decoder_input_tokens.shape[:2] , device=decoder_input_tokens.device , dtype=inputs.dtype ) # Translate encoding masks to encoder-decoder masks. A = [(x, self.encoder_decoder_mask(UpperCamelCase__ , UpperCamelCase__ )) for x, y in encodings_and_masks] # cross attend style: concat encodings A = torch.cat([x[0] for x in encodings_and_encdec_masks] , dim=1 ) A = torch.cat([x[1] for x in encodings_and_encdec_masks] , dim=-1 ) for lyr in self.decoders: A = lyr( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , )[0] A = self.decoder_norm(UpperCamelCase__ ) A = self.post_dropout(UpperCamelCase__ ) A = self.spec_out(UpperCamelCase__ ) return spec_out class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Union[str, Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Union[str, Any]=1e-6 ): super().__init__() A = nn.ModuleList() # cond self attention: layer 0 self.layer.append( TaLayerSelfAttentionCond(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) ) # cross attention: layer 1 self.layer.append( TaLayerCrossAttention( d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ , ) ) # Film Cond MLP + dropout: last layer self.layer.append( TaLayerFFCond(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ ) ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : int=None , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[Any]=None , ): A = self.layer[0]( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) if encoder_hidden_states is not None: A = torch.where(encoder_attention_mask > 0 , 0 , -1e1_0 ).to( encoder_hidden_states.dtype ) A = self.layer[1]( UpperCamelCase__ , key_value_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) # Apply Film Conditional Feed Forward layer A = self.layer[-1](UpperCamelCase__ , UpperCamelCase__ ) return (hidden_states,) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Dict ): super().__init__() A = TaLayerNorm(UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : str , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Tuple=None , ): # pre_self_attention_layer_norm A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.FiLMLayer(UpperCamelCase__ , UpperCamelCase__ ) # Self-attention block A = self.attention(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[Any] ): super().__init__() A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[str]=None , ): A = self.layer_norm(UpperCamelCase__ ) A = self.attention( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=attention_mask.squeeze(1 ) , ) A = hidden_states + self.dropout(UpperCamelCase__ ) return layer_output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any ): super().__init__() A = TaDenseGatedActDense(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any=None ): A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.film(UpperCamelCase__ , UpperCamelCase__ ) A = self.DenseReluDense(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) A = NewGELUActivation() def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : List[Any] ): A = self.act(self.wi_a(UpperCamelCase__ ) ) A = self.wi_a(UpperCamelCase__ ) A = hidden_gelu * hidden_linear A = self.dropout(UpperCamelCase__ ) A = self.wo(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Tuple=1e-6 ): super().__init__() A = nn.Parameter(torch.ones(UpperCamelCase__ ) ) A = eps def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int ): # T5 uses a layer_norm which only scales and doesn't shift, which is also known as Root Mean # Square Layer Normalization https://arxiv.org/abs/1910.07467 thus variance is calculated # w/o mean and there is no bias. Additionally we want to make sure that the accumulation for # half-precision inputs is done in fp32 A = hidden_states.to(torch.floataa ).pow(2 ).mean(-1 , keepdim=UpperCamelCase__ ) A = hidden_states * torch.rsqrt(variance + self.variance_epsilon ) # convert into half-precision if necessary if self.weight.dtype in [torch.floataa, torch.bfloataa]: A = hidden_states.to(self.weight.dtype ) return self.weight * hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def UpperCamelCase ( self : Any , UpperCamelCase__ : torch.Tensor ): return 0.5 * input * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi ) * (input + 0.044_715 * torch.pow(UpperCamelCase__ , 3.0 )) )) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , out_features * 2 , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : List[Any] ): A = self.scale_bias(UpperCamelCase__ ) A , A = torch.chunk(UpperCamelCase__ , 2 , -1 ) A = x * (1 + scale) + shift return x
699
1
import os import jsonlines import numpy as np from tqdm import tqdm _UpperCAmelCase = 2_048 _UpperCAmelCase = 4_096 _UpperCAmelCase = 42 _UpperCAmelCase = os.environ.pop("PROCESS_TRAIN", "false") _UpperCAmelCase = {"null": 0, "short": 1, "long": 2, "yes": 3, "no": 4} def __UpperCamelCase (lowerCAmelCase : Optional[Any] ) -> Union[str, Any]: def choose_first(lowerCAmelCase : Any, lowerCAmelCase : str=False ): assert isinstance(lowerCAmelCase, lowerCAmelCase ) if len(lowerCAmelCase ) == 1: A = answer[0] return {k: [answer[k]] for k in answer} if is_long_answer else answer for a in answer: if is_long_answer: A = {k: [a[k]] for k in a} if len(a['start_token'] ) > 0: break return a A = {'id': example['id']} A = example['annotations'] A = annotation['yes_no_answer'] if 0 in yes_no_answer or 1 in yes_no_answer: A = ['yes'] if 1 in yes_no_answer else ['no'] A = A = [] A = A = [] A = ['<cls>'] else: A = ['short'] A = choose_first(annotation['short_answers'] ) if len(out['start_token'] ) == 0: # answer will be long if short is not available A = ['long'] A = choose_first(annotation['long_answer'], is_long_answer=lowerCAmelCase ) A = [] answer.update(lowerCAmelCase ) # disregard some samples if len(answer['start_token'] ) > 1 or answer["start_token"] == answer["end_token"]: A = True else: A = False A = ['start_token', 'end_token', 'start_byte', 'end_byte', 'text'] if not all(isinstance(answer[k], lowerCAmelCase ) for k in cols ): raise ValueError('Issue in ID', example['id'] ) return answer def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Optional[Any]=False ) -> List[Any]: A = _get_single_answer(lowerCAmelCase ) # bytes are of no use del answer["start_byte"] del answer["end_byte"] # handle yes_no answers explicitly if answer["category"][0] in ["yes", "no"]: # category is list with one element A = example['document']['tokens'] A = [] for i in range(len(doc['token'] ) ): if not doc["is_html"][i]: context.append(doc['token'][i] ) return { "context": " ".join(lowerCAmelCase ), "answer": { "start_token": -100, # ignore index in cross-entropy "end_token": -100, # ignore index in cross-entropy "category": answer["category"], "span": answer["category"], # extra }, } # later, help in removing all no answers if answer["start_token"] == [-1]: return { "context": "None", "answer": { "start_token": -1, "end_token": -1, "category": "null", "span": "None", # extra }, } # handling normal samples A = ['start_token', 'end_token'] answer.update({k: answer[k][0] if len(answer[k] ) > 0 else answer[k] for k in cols} ) # e.g. [10] == 10 A = example['document']['tokens'] A = answer['start_token'] A = answer['end_token'] A = [] for i in range(len(doc['token'] ) ): if not doc["is_html"][i]: context.append(doc['token'][i] ) else: if answer["start_token"] > i: start_token -= 1 if answer["end_token"] > i: end_token -= 1 A = ' '.join(context[start_token:end_token] ) # checking above code if assertion: A = doc['is_html'][answer['start_token'] : answer['end_token']] A = doc['token'][answer['start_token'] : answer['end_token']] A = ' '.join([old[i] for i in range(len(lowerCAmelCase ) ) if not is_html[i]] ) if new != old: print('ID:', example['id'] ) print('New:', lowerCAmelCase, end='\n' ) print('Old:', lowerCAmelCase, end='\n\n' ) return { "context": " ".join(lowerCAmelCase ), "answer": { "start_token": start_token, "end_token": end_token - 1, # this makes it inclusive "category": answer["category"], # either long or short "span": new, # extra }, } def __UpperCamelCase (lowerCAmelCase : Dict, lowerCAmelCase : Tuple, lowerCAmelCase : int=2_048, lowerCAmelCase : Optional[int]=4_096, lowerCAmelCase : List[Any]=True ) -> Dict: # overlap will be of doc_stride - q_len A = get_context_and_ans(lowerCAmelCase, assertion=lowerCAmelCase ) A = out['answer'] # later, removing these samples if answer["start_token"] == -1: return { "example_id": example["id"], "input_ids": [[-1]], "labels": { "start_token": [-1], "end_token": [-1], "category": ["null"], }, } A = tokenizer(example['question']['text'], out['context'] ).input_ids A = input_ids.index(tokenizer.sep_token_id ) + 1 # return yes/no if answer["category"][0] in ["yes", "no"]: # category is list with one element A = [] A = [] A = input_ids[:q_len] A = range(lowerCAmelCase, len(lowerCAmelCase ), max_length - doc_stride ) for i in doc_start_indices: A = i + max_length - q_len A = input_ids[i:end_index] inputs.append(q_indices + slice ) category.append(answer['category'][0] ) if slice[-1] == tokenizer.sep_token_id: break return { "example_id": example["id"], "input_ids": inputs, "labels": { "start_token": [-100] * len(lowerCAmelCase ), "end_token": [-100] * len(lowerCAmelCase ), "category": category, }, } A = out['context'].split() A = splitted_context[answer['end_token']] A = len( tokenizer( ' '.join(splitted_context[: answer['start_token']] ), add_special_tokens=lowerCAmelCase, ).input_ids ) A = len( tokenizer(' '.join(splitted_context[: answer['end_token']] ), add_special_tokens=lowerCAmelCase ).input_ids ) answer["start_token"] += q_len answer["end_token"] += q_len # fixing end token A = len(tokenizer(lowerCAmelCase, add_special_tokens=lowerCAmelCase ).input_ids ) if num_sub_tokens > 1: answer["end_token"] += num_sub_tokens - 1 A = input_ids[answer['start_token'] : answer['end_token'] + 1] # right & left are inclusive A = answer['start_token'] A = answer['end_token'] if assertion: A = tokenizer.decode(lowerCAmelCase ) if answer["span"] != new: print('ISSUE IN TOKENIZATION' ) print('OLD:', answer['span'] ) print('NEW:', lowerCAmelCase, end='\n\n' ) if len(lowerCAmelCase ) <= max_length: return { "example_id": example["id"], "input_ids": [input_ids], "labels": { "start_token": [answer["start_token"]], "end_token": [answer["end_token"]], "category": answer["category"], }, } A = input_ids[:q_len] A = range(lowerCAmelCase, len(lowerCAmelCase ), max_length - doc_stride ) A = [] A = [] A = [] A = [] # null, yes, no, long, short for i in doc_start_indices: A = i + max_length - q_len A = input_ids[i:end_index] inputs.append(q_indices + slice ) assert len(inputs[-1] ) <= max_length, "Issue in truncating length" if start_token >= i and end_token <= end_index - 1: A = start_token - i + q_len A = end_token - i + q_len answers_category.append(answer['category'][0] ) # ["short"] -> "short" else: A = -100 A = -100 answers_category.append('null' ) A = inputs[-1][start_token : end_token + 1] answers_start_token.append(lowerCAmelCase ) answers_end_token.append(lowerCAmelCase ) if assertion: if new != old and new != [tokenizer.cls_token_id]: print('ISSUE in strided for ID:', example['id'] ) print('New:', tokenizer.decode(lowerCAmelCase ) ) print('Old:', tokenizer.decode(lowerCAmelCase ), end='\n\n' ) if slice[-1] == tokenizer.sep_token_id: break return { "example_id": example["id"], "input_ids": inputs, "labels": { "start_token": answers_start_token, "end_token": answers_end_token, "category": answers_category, }, } def __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Tuple, lowerCAmelCase : Union[str, Any]=2_048, lowerCAmelCase : List[Any]=4_096, lowerCAmelCase : str=False ) -> List[str]: A = get_strided_contexts_and_ans( lowerCAmelCase, lowerCAmelCase, doc_stride=lowerCAmelCase, max_length=lowerCAmelCase, assertion=lowerCAmelCase, ) return example def __UpperCamelCase (lowerCAmelCase : List[Any], lowerCAmelCase : Union[str, Any] ) -> Any: with jsonlines.open(lowerCAmelCase, 'a' ) as writer: for example in tqdm(lowerCAmelCase, total=len(lowerCAmelCase ), desc='Saving samples ... ' ): A = example['labels'] for ids, start, end, cat in zip( example['input_ids'], labels['start_token'], labels['end_token'], labels['category'], ): if start == -1 and end == -1: continue # leave waste samples with no answer if cat == "null" and np.random.rand() < 0.6: continue # removing 50 % samples writer.write( { 'input_ids': ids, 'start_token': start, 'end_token': end, 'category': CATEGORY_MAPPING[cat], } ) if __name__ == "__main__": from datasets import load_dataset from transformers import BigBirdTokenizer _UpperCAmelCase = load_dataset("natural_questions") _UpperCAmelCase = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base") _UpperCAmelCase = data["train" if PROCESS_TRAIN == "true" else "validation"] _UpperCAmelCase = { "tokenizer": tokenizer, "doc_stride": DOC_STRIDE, "max_length": MAX_LENGTH, "assertion": False, } _UpperCAmelCase = data.map(prepare_inputs, fn_kwargs=fn_kwargs) _UpperCAmelCase = data.remove_columns(["annotations", "document", "id", "question"]) print(data) np.random.seed(SEED) _UpperCAmelCase = "nq-training.jsonl" if PROCESS_TRAIN == "true" else "nq-validation.jsonl" save_to_disk(data, file_name=cache_file_name)
699
import numpy as np import torch from imwatermark import WatermarkEncoder # Copied from https://github.com/Stability-AI/generative-models/blob/613af104c6b85184091d42d374fef420eddb356d/scripts/demo/streamlit_helpers.py#L66 _UpperCAmelCase = 0b10_11_00_11_11_10_11_00_10_01_00_00_01_11_10_11_10_11_00_01_10_01_11_10 # bin(x)[2:] gives bits of x as str, use int to convert them to 0/1 _UpperCAmelCase = [int(bit) for bit in bin(WATERMARK_MESSAGE)[2:]] class _UpperCAmelCase : '''simple docstring''' def __init__( self : Union[str, Any] ): A = WATERMARK_BITS A = WatermarkEncoder() self.encoder.set_watermark('bits' , self.watermark ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : torch.FloatTensor ): # can't encode images that are smaller than 256 if images.shape[-1] < 256: return images A = (255 * (images / 2 + 0.5)).cpu().permute(0 , 2 , 3 , 1 ).float().numpy() A = [self.encoder.encode(UpperCamelCase__ , 'dwtDct' ) for image in images] A = torch.from_numpy(np.array(UpperCamelCase__ ) ).permute(0 , 3 , 1 , 2 ) A = torch.clamp(2 * (images / 255 - 0.5) , min=-1.0 , max=1.0 ) return images
699
1
# 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 ...utils import deprecate from ..controlnet.pipeline_flax_controlnet import FlaxStableDiffusionControlNetPipeline # noqa: F401 deprecate( "stable diffusion controlnet", "0.22.0", "Importing `FlaxStableDiffusionControlNetPipeline` from diffusers.pipelines.stable_diffusion.flax_pipeline_stable_diffusion_controlnet is deprecated. Please import `from diffusers import FlaxStableDiffusionControlNetPipeline` instead.", standard_warn=False, stacklevel=3, )
699
import argparse import fairseq import torch from transformers import UniSpeechSatConfig, UniSpeechSatForCTC, UniSpeechSatForPreTraining, logging logging.set_verbosity_info() _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "post_extract_proj": "feature_projection.projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.layer_norm": "encoder.layer_norm", "encoder.layer_norm_for_extract": "layer_norm_for_extract", "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": "lm_head", "label_embs_concat": "label_embeddings_concat", "mask_emb": "masked_spec_embed", "spk_proj": "speaker_proj", } _UpperCAmelCase = [ "lm_head", "quantizer.weight_proj", "quantizer.codevectors", "project_q", "project_hid", "label_embeddings_concat", "speaker_proj", "layer_norm_for_extract", ] def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Dict, lowerCAmelCase : Optional[int], lowerCAmelCase : List[Any], lowerCAmelCase : str ) -> int: 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 if hf_shape != value.shape: raise ValueError( 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 __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Optional[int] ) -> Dict: A = [] A = fairseq_model.state_dict() A = hf_model.unispeech_sat.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 = 'unispeech_sat.' + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key if key in name or key.split('w2v_model.' )[-1] == name.split('.' )[0]: if "layer_norm_for_extract" in name and (".".join(name.split('.' )[:-1] ) != key): # special case since naming is very similar continue 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 "bias" in name: A = 'bias' elif "weight" in name: # TODO: don't match quantizer.weight_proj A = 'weight' 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 __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : Tuple, lowerCAmelCase : List[Any], lowerCAmelCase : int ) -> Dict: 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.bias.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.weight.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape: raise ValueError( f'''{full_name} has size {value.shape}, but''' f''' {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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape: raise ValueError( 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 ) @torch.no_grad() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : Dict, lowerCAmelCase : Union[str, Any]=None, lowerCAmelCase : str=None, lowerCAmelCase : List[Any]=True ) -> Union[str, Any]: if config_path is not None: A = UniSpeechSatConfig.from_pretrained(lowerCAmelCase ) else: A = UniSpeechSatConfig() A = '' if is_finetuned: A = UniSpeechSatForCTC(lowerCAmelCase ) else: A = UniSpeechSatForPreTraining(lowerCAmelCase ) A , A , A = fairseq.checkpoint_utils.load_model_ensemble_and_task( [checkpoint_path], arg_overrides={'data': '/'.join(dict_path.split('/' )[:-1] )} ) A = model[0].eval() recursively_load_weights(lowerCAmelCase, lowerCAmelCase ) hf_wavavec.save_pretrained(lowerCAmelCase ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--dict_path", default=None, type=str, help="Path to dict of fine-tuned model") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") parser.add_argument( "--not_finetuned", action="store_true", help="Whether the model to convert is a fine-tuned model or not" ) _UpperCAmelCase = parser.parse_args() convert_unispeech_sat_checkpoint( args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned )
699
1
from typing import Any, Dict, List, Union from ..utils import add_end_docstrings, is_torch_available, is_vision_available, logging, requires_backends from .base import PIPELINE_INIT_ARGS, ChunkPipeline if is_vision_available(): from PIL import Image from ..image_utils import load_image if is_torch_available(): import torch from transformers.modeling_outputs import BaseModelOutput from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING _UpperCAmelCase = logging.get_logger(__name__) @add_end_docstrings(__lowercase ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : Optional[int] , **UpperCamelCase__ : int ): super().__init__(**UpperCamelCase__ ) if self.framework == "tf": raise ValueError(f'''The {self.__class__} is only available in PyTorch.''' ) requires_backends(self , 'vision' ) self.check_model_type(UpperCamelCase__ ) def __call__( self : List[Any] , UpperCamelCase__ : Union[str, "Image.Image", List[Dict[str, Any]]] , UpperCamelCase__ : Union[str, List[str]] = None , **UpperCamelCase__ : Dict , ): if "text_queries" in kwargs: A = kwargs.pop('text_queries' ) if isinstance(UpperCamelCase__ , (str, Image.Image) ): A = {'image': image, 'candidate_labels': candidate_labels} else: A = image A = super().__call__(UpperCamelCase__ , **UpperCamelCase__ ) return results def UpperCamelCase ( self : List[str] , **UpperCamelCase__ : List[Any] ): A = {} if "threshold" in kwargs: A = kwargs['threshold'] if "top_k" in kwargs: A = kwargs['top_k'] return {}, {}, postprocess_params def UpperCamelCase ( self : List[str] , UpperCamelCase__ : List[Any] ): A = load_image(inputs['image'] ) A = inputs['candidate_labels'] if isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = candidate_labels.split(',' ) A = torch.tensor([[image.height, image.width]] , dtype=torch.intaa ) for i, candidate_label in enumerate(UpperCamelCase__ ): A = self.tokenizer(UpperCamelCase__ , return_tensors=self.framework ) A = self.image_processor(UpperCamelCase__ , return_tensors=self.framework ) yield { "is_last": i == len(UpperCamelCase__ ) - 1, "target_size": target_size, "candidate_label": candidate_label, **text_inputs, **image_features, } def UpperCamelCase ( self : Dict , UpperCamelCase__ : Any ): A = model_inputs.pop('target_size' ) A = model_inputs.pop('candidate_label' ) A = model_inputs.pop('is_last' ) A = self.model(**UpperCamelCase__ ) A = {'target_size': target_size, 'candidate_label': candidate_label, 'is_last': is_last, **outputs} return model_outputs def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Any=0.1 , UpperCamelCase__ : List[Any]=None ): A = [] for model_output in model_outputs: A = model_output['candidate_label'] A = BaseModelOutput(UpperCamelCase__ ) A = self.image_processor.post_process_object_detection( outputs=UpperCamelCase__ , threshold=UpperCamelCase__ , target_sizes=model_output['target_size'] )[0] for index in outputs["scores"].nonzero(): A = outputs['scores'][index].item() A = self._get_bounding_box(outputs['boxes'][index][0] ) A = {'score': score, 'label': label, 'box': box} results.append(UpperCamelCase__ ) A = sorted(UpperCamelCase__ , key=lambda UpperCamelCase__ : x["score"] , reverse=UpperCamelCase__ ) if top_k: A = results[:top_k] return results def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : "torch.Tensor" ): if self.framework != "pt": raise ValueError('The ZeroShotObjectDetectionPipeline is only available in PyTorch.' ) A , A , A , A = box.int().tolist() A = { 'xmin': xmin, 'ymin': ymin, 'xmax': xmax, 'ymax': ymax, } return bbox
699
from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar _UpperCAmelCase = TypeVar("T") class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : T ): A = data A = None def __str__( self : Optional[int] ): return f'''{self.data}''' class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple ): A = None def __iter__( self : int ): A = self.top while node: yield node.data A = node.next def __str__( self : Any ): return "->".join([str(UpperCamelCase__ ) for item in self] ) def __len__( self : Dict ): return len(tuple(iter(self ) ) ) def UpperCamelCase ( self : List[str] ): return self.top is None def UpperCamelCase ( self : Dict , UpperCamelCase__ : T ): A = Node(UpperCamelCase__ ) if not self.is_empty(): A = self.top A = node def UpperCamelCase ( self : Dict ): if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top , UpperCamelCase__ ) A = self.top A = self.top.next return pop_node.data def UpperCamelCase ( self : List[str] ): if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def UpperCamelCase ( self : List[str] ): A = None if __name__ == "__main__": from doctest import testmod testmod()
699
1
class _UpperCAmelCase : '''simple docstring''' def __init__( self : Tuple ): A = {} def UpperCamelCase ( self : List[Any] ): print(self.vertex ) for i in self.vertex: print(UpperCamelCase__ , ' -> ' , ' -> '.join([str(UpperCamelCase__ ) for j in self.vertex[i]] ) ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int ): # check if vertex is already present, if from_vertex in self.vertex: self.vertex[from_vertex].append(UpperCamelCase__ ) else: # else make a new vertex A = [to_vertex] def UpperCamelCase ( self : List[Any] ): # visited array for storing already visited nodes A = [False] * len(self.vertex ) # call the recursive helper function for i in range(len(self.vertex ) ): if not visited[i]: self.dfs_recursive(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : list ): # mark start vertex as visited A = True print(UpperCamelCase__ , end=' ' ) # Recur for all the vertices that are adjacent to this node for i in self.vertex: if not visited[i]: self.dfs_recursive(UpperCamelCase__ , UpperCamelCase__ ) if __name__ == "__main__": _UpperCAmelCase = 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
699
from __future__ import annotations import math def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : bool, lowerCAmelCase : list[int], lowerCAmelCase : float ) -> int: if depth < 0: raise ValueError('Depth cannot be less than 0' ) if not scores: raise ValueError('Scores cannot be empty' ) if depth == height: return scores[node_index] return ( max( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) if is_max else min( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) ) def __UpperCamelCase () -> None: A = [90, 23, 6, 33, 21, 65, 123, 34_423] A = math.log(len(lowerCAmelCase ), 2 ) print(f'''Optimal value : {minimax(0, 0, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )}''' ) if __name__ == "__main__": import doctest doctest.testmod() main()
699
1
from __future__ import annotations def __UpperCamelCase (lowerCAmelCase : list, lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> list: A = [] A , A = input_list[low:mid], input_list[mid : high + 1] while left and right: result.append((left if left[0] <= right[0] else right).pop(0 ) ) A = result + left + right return input_list def __UpperCamelCase (lowerCAmelCase : list ) -> list: if len(lowerCAmelCase ) <= 1: return input_list A = list(lowerCAmelCase ) # iteration for two-way merging A = 2 while p <= len(lowerCAmelCase ): # getting low, high and middle value for merge-sort of single list for i in range(0, len(lowerCAmelCase ), lowerCAmelCase ): A = i A = i + p - 1 A = (low + high + 1) // 2 A = merge(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) # final merge of last two parts if p * 2 >= len(lowerCAmelCase ): A = i A = merge(lowerCAmelCase, 0, lowerCAmelCase, len(lowerCAmelCase ) - 1 ) break p *= 2 return input_list if __name__ == "__main__": _UpperCAmelCase = input("Enter numbers separated by a comma:\n").strip() if user_input == "": _UpperCAmelCase = [] else: _UpperCAmelCase = [int(item.strip()) for item in user_input.split(",")] print(iter_merge_sort(unsorted))
699
import logging import os from dataclasses import dataclass, field from typing import Dict, Optional import datasets import numpy as np import tensorflow as tf from transformers import ( AutoConfig, AutoTokenizer, EvalPrediction, HfArgumentParser, PreTrainedTokenizer, TFAutoModelForSequenceClassification, TFTrainer, TFTrainingArguments, ) from transformers.utils import logging as hf_logging hf_logging.set_verbosity_info() hf_logging.enable_default_handler() hf_logging.enable_explicit_format() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : PreTrainedTokenizer, lowerCAmelCase : int, lowerCAmelCase : Optional[int] = None, ) -> Dict: A = {} if train_file is not None: A = [train_file] if eval_file is not None: A = [eval_file] if test_file is not None: A = [test_file] A = datasets.load_dataset('csv', data_files=lowerCAmelCase ) A = list(ds[list(files.keys() )[0]].features.keys() ) A = features_name.pop(lowerCAmelCase ) A = list(set(ds[list(files.keys() )[0]][label_name] ) ) A = {label: i for i, label in enumerate(lowerCAmelCase )} A = tokenizer.model_input_names A = {} if len(lowerCAmelCase ) == 1: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( example[features_name[0]], truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length' ), batched=lowerCAmelCase, ) elif len(lowerCAmelCase ) == 2: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( (example[features_name[0]], example[features_name[1]]), truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length', ), batched=lowerCAmelCase, ) def gen_train(): for ex in transformed_ds[datasets.Split.TRAIN]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_val(): for ex in transformed_ds[datasets.Split.VALIDATION]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_test(): for ex in transformed_ds[datasets.Split.TEST]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TRAIN in transformed_ds else None ) if train_ds is not None: A = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.VALIDATION in transformed_ds else None ) if val_ds is not None: A = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TEST in transformed_ds else None ) if test_ds is not None: A = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) ) return train_ds, val_ds, test_ds, labelaid _UpperCAmelCase = logging.getLogger(__name__) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : int = field(metadata={'''help''': '''Which column contains the label'''} ) SCREAMING_SNAKE_CASE : str = field(default=__lowercase , metadata={'''help''': '''The path of the training file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the development file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the test file'''} ) SCREAMING_SNAKE_CASE : int = field( default=1_28 , metadata={ '''help''': ( '''The maximum total input sequence length after tokenization. Sequences longer ''' '''than this will be truncated, sequences shorter will be padded.''' ) } , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Overwrite the cached training and evaluation sets'''} ) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : str = field( metadata={'''help''': '''Path to pretrained model or model identifier from huggingface.co/models'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained config name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained tokenizer name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : bool = field(default=__lowercase , metadata={'''help''': '''Set this flag to use fast tokenization.'''} ) # If you want to tweak more attributes on your tokenizer, you should do it in a distinct script, # or just modify its tokenizer_config.json. SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Where do you want to store the pretrained models downloaded from huggingface.co'''} , ) def __UpperCamelCase () -> Any: # See all possible arguments in src/transformers/training_args.py # or by passing the --help flag to this script. # We now keep distinct sets of args, for a cleaner separation of concerns. A = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) ) A , A , A = parser.parse_args_into_dataclasses() if ( os.path.exists(training_args.output_dir ) and os.listdir(training_args.output_dir ) and training_args.do_train and not training_args.overwrite_output_dir ): raise ValueError( f'''Output directory ({training_args.output_dir}) already exists and is not empty. Use''' ' --overwrite_output_dir to overcome.' ) # Setup logging logging.basicConfig( format='%(asctime)s - %(levelname)s - %(name)s - %(message)s', datefmt='%m/%d/%Y %H:%M:%S', level=logging.INFO, ) logger.info( f'''n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, ''' f'''16-bits training: {training_args.fpaa}''' ) logger.info(f'''Training/evaluation parameters {training_args}''' ) # Load pretrained model and tokenizer # # Distributed training: # The .from_pretrained methods guarantee that only one local process can concurrently # download model & vocab. A = AutoTokenizer.from_pretrained( model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path, cache_dir=model_args.cache_dir, ) A , A , A , A = get_tfds( train_file=data_args.train_file, eval_file=data_args.dev_file, test_file=data_args.test_file, tokenizer=lowerCAmelCase, label_column_id=data_args.label_column_id, max_seq_length=data_args.max_seq_length, ) A = AutoConfig.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, num_labels=len(lowerCAmelCase ), labelaid=lowerCAmelCase, idalabel={id: label for label, id in labelaid.items()}, finetuning_task='text-classification', cache_dir=model_args.cache_dir, ) with training_args.strategy.scope(): A = TFAutoModelForSequenceClassification.from_pretrained( model_args.model_name_or_path, from_pt=bool('.bin' in model_args.model_name_or_path ), config=lowerCAmelCase, cache_dir=model_args.cache_dir, ) def compute_metrics(lowerCAmelCase : EvalPrediction ) -> Dict: A = np.argmax(p.predictions, axis=1 ) return {"acc": (preds == p.label_ids).mean()} # Initialize our Trainer A = TFTrainer( model=lowerCAmelCase, args=lowerCAmelCase, train_dataset=lowerCAmelCase, eval_dataset=lowerCAmelCase, compute_metrics=lowerCAmelCase, ) # Training if training_args.do_train: trainer.train() trainer.save_model() tokenizer.save_pretrained(training_args.output_dir ) # Evaluation A = {} if training_args.do_eval: logger.info('*** Evaluate ***' ) A = trainer.evaluate() A = os.path.join(training_args.output_dir, 'eval_results.txt' ) with open(lowerCAmelCase, 'w' ) as writer: logger.info('***** Eval results *****' ) for key, value in result.items(): logger.info(f''' {key} = {value}''' ) writer.write(f'''{key} = {value}\n''' ) results.update(lowerCAmelCase ) return results if __name__ == "__main__": main()
699
1
from typing import List import datasets from datasets.tasks import AudioClassification from ..folder_based_builder import folder_based_builder _UpperCAmelCase = datasets.utils.logging.get_logger(__name__) class _UpperCAmelCase ( folder_based_builder.FolderBasedBuilderConfig ): '''simple docstring''' SCREAMING_SNAKE_CASE : bool = None SCREAMING_SNAKE_CASE : bool = None class _UpperCAmelCase ( folder_based_builder.FolderBasedBuilder ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = datasets.Audio() SCREAMING_SNAKE_CASE : Union[str, Any] = '''audio''' SCREAMING_SNAKE_CASE : List[Any] = AudioFolderConfig SCREAMING_SNAKE_CASE : List[str] # definition at the bottom of the script SCREAMING_SNAKE_CASE : Any = AudioClassification(audio_column='''audio''' , label_column='''label''' ) _UpperCAmelCase = [ ".aiff", ".au", ".avr", ".caf", ".flac", ".htk", ".svx", ".mat4", ".mat5", ".mpc2k", ".ogg", ".paf", ".pvf", ".raw", ".rf64", ".sd2", ".sds", ".ircam", ".voc", ".w64", ".wav", ".nist", ".wavex", ".wve", ".xi", ".mp3", ".opus", ] _UpperCAmelCase = AUDIO_EXTENSIONS
699
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = {"configuration_xglm": ["XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "XGLMConfig"]} try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizer"] try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "XGLMForCausalLM", "XGLMModel", "XGLMPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxXGLMForCausalLM", "FlaxXGLMModel", "FlaxXGLMPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "TFXGLMForCausalLM", "TFXGLMModel", "TFXGLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm import XGLMTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm_fast import XGLMTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_xglm import ( TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, TFXGLMForCausalLM, TFXGLMModel, TFXGLMPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure)
699
1
import unittest from transformers import AutoTokenizer, FalconConfig, is_torch_available from transformers.testing_utils import require_torch, slow, torch_device from ...generation.test_utils import GenerationTesterMixin from ...test_configuration_common import ConfigTester from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask from ...test_pipeline_mixin import PipelineTesterMixin if is_torch_available(): import torch from transformers import ( FalconForCausalLM, FalconForQuestionAnswering, FalconForSequenceClassification, FalconForTokenClassification, FalconModel, ) class _UpperCAmelCase : '''simple docstring''' def __init__( self : str , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any]=3 , UpperCamelCase__ : Optional[Any]=7 , UpperCamelCase__ : Any=True , UpperCamelCase__ : List[Any]=True , UpperCamelCase__ : Union[str, Any]=False , UpperCamelCase__ : str=True , UpperCamelCase__ : Union[str, Any]=99 , UpperCamelCase__ : Optional[int]=32 , UpperCamelCase__ : Optional[int]=5 , UpperCamelCase__ : Tuple=4 , UpperCamelCase__ : Any=37 , UpperCamelCase__ : Optional[Any]="gelu" , UpperCamelCase__ : str=0.1 , UpperCamelCase__ : Tuple=0.1 , UpperCamelCase__ : Dict=512 , UpperCamelCase__ : int=16 , UpperCamelCase__ : Optional[int]=2 , UpperCamelCase__ : Optional[int]=0.02 , UpperCamelCase__ : Tuple=3 , UpperCamelCase__ : Tuple=4 , UpperCamelCase__ : List[str]=None , ): A = parent A = batch_size A = seq_length A = is_training A = use_input_mask A = use_token_type_ids A = use_labels A = vocab_size A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = max_position_embeddings A = type_vocab_size A = type_sequence_label_size A = initializer_range A = num_labels A = num_choices A = scope def UpperCamelCase ( self : Any ): A = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size ) A = None if self.use_input_mask: A = random_attention_mask([self.batch_size, self.seq_length] ) A = None A = None A = None A = None if self.use_labels: A = ids_tensor([self.batch_size] , self.type_sequence_label_size ) A = ids_tensor([self.batch_size, self.seq_length] , self.num_labels ) A = ids_tensor([self.batch_size] , self.num_choices ) A = self.get_config() return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels def UpperCamelCase ( self : Tuple ): return FalconConfig( vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=UpperCamelCase__ , initializer_range=self.initializer_range , pad_token_id=1 , new_decoder_architecture=UpperCamelCase__ , ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Dict , UpperCamelCase__ : int , UpperCamelCase__ : str , UpperCamelCase__ : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : int ): A = FalconModel(config=UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , attention_mask=UpperCamelCase__ ) A = model(UpperCamelCase__ ) self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Any , UpperCamelCase__ : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Optional[Any] , ): A = True A = FalconModel(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model( UpperCamelCase__ , attention_mask=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , ) A = model( UpperCamelCase__ , attention_mask=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , ) A = model(UpperCamelCase__ , attention_mask=UpperCamelCase__ ) self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) ) def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Any , UpperCamelCase__ : Dict , ): A = FalconForCausalLM(config=UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , attention_mask=UpperCamelCase__ , labels=UpperCamelCase__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Dict , UpperCamelCase__ : Any , UpperCamelCase__ : List[str] , UpperCamelCase__ : Dict , UpperCamelCase__ : int , UpperCamelCase__ : List[Any] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : Union[str, Any] , ): A = True A = True A = FalconForCausalLM(config=UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() # first forward pass A = model( UpperCamelCase__ , attention_mask=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , use_cache=UpperCamelCase__ , ) A = outputs.past_key_values # create hypothetical multiple next token and extent to next_input_ids A = ids_tensor((self.batch_size, 3) , config.vocab_size ) A = ids_tensor((self.batch_size, 3) , vocab_size=2 ) # append to next input_ids and A = torch.cat([input_ids, next_tokens] , dim=-1 ) A = torch.cat([input_mask, next_mask] , dim=-1 ) A = model( UpperCamelCase__ , attention_mask=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , output_hidden_states=UpperCamelCase__ , )['hidden_states'][0] A = model( UpperCamelCase__ , attention_mask=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , past_key_values=UpperCamelCase__ , output_hidden_states=UpperCamelCase__ , )['hidden_states'][0] # select random slice A = ids_tensor((1,) , output_from_past.shape[-1] ).item() A = output_from_no_past[:, -3:, random_slice_idx].detach() A = output_from_past[:, :, random_slice_idx].detach() self.parent.assertTrue(output_from_past_slice.shape[1] == next_tokens.shape[1] ) # test that outputs are equal for slice self.parent.assertTrue(torch.allclose(UpperCamelCase__ , UpperCamelCase__ , atol=1e-3 ) ) def UpperCamelCase ( self : Union[str, Any] ): A = self.prepare_config_and_inputs() ( ( A ) , ( A ) , ( A ) , ( A ) , ( A ) , ( A ) , ( A ) , ) = config_and_inputs A = {'input_ids': input_ids, 'attention_mask': input_mask} return config, inputs_dict @require_torch class _UpperCAmelCase ( __lowercase , __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ( ( FalconModel, FalconForCausalLM, FalconForSequenceClassification, FalconForTokenClassification, FalconForQuestionAnswering, ) if is_torch_available() else () ) SCREAMING_SNAKE_CASE : Tuple = (FalconForCausalLM,) if is_torch_available() else () SCREAMING_SNAKE_CASE : Dict = ( { '''feature-extraction''': FalconModel, '''text-classification''': FalconForSequenceClassification, '''text-generation''': FalconForCausalLM, '''question-answering''': FalconForQuestionAnswering, '''token-classification''': FalconForTokenClassification, '''zero-shot''': FalconForSequenceClassification, } if is_torch_available() else {} ) SCREAMING_SNAKE_CASE : int = False SCREAMING_SNAKE_CASE : Union[str, Any] = False def UpperCamelCase ( self : List[str] ): A = FalconModelTester(self ) A = ConfigTester(self , config_class=UpperCamelCase__ , hidden_size=37 ) def UpperCamelCase ( self : str ): self.config_tester.run_common_tests() def UpperCamelCase ( self : List[str] ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_model(*UpperCamelCase__ ) def UpperCamelCase ( self : Tuple ): A , *A = self.model_tester.prepare_config_and_inputs() for alibi in [True, False]: A = alibi self.model_tester.create_and_check_model(UpperCamelCase__ , *UpperCamelCase__ ) def UpperCamelCase ( self : Dict ): A , A = self.model_tester.prepare_config_and_inputs_for_common() A = 3 A = input_dict['input_ids'] A = input_ids.ne(1 ).to(UpperCamelCase__ ) A = ids_tensor([self.model_tester.batch_size] , self.model_tester.type_sequence_label_size ) A = FalconForSequenceClassification(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , attention_mask=UpperCamelCase__ , labels=UpperCamelCase__ ) self.assertEqual(result.logits.shape , (self.model_tester.batch_size, self.model_tester.num_labels) ) def UpperCamelCase ( self : Dict ): A , A = self.model_tester.prepare_config_and_inputs_for_common() A = 3 A = 'single_label_classification' A = input_dict['input_ids'] A = input_ids.ne(1 ).to(UpperCamelCase__ ) A = ids_tensor([self.model_tester.batch_size] , self.model_tester.type_sequence_label_size ) A = FalconForSequenceClassification(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , attention_mask=UpperCamelCase__ , labels=UpperCamelCase__ ) self.assertEqual(result.logits.shape , (self.model_tester.batch_size, self.model_tester.num_labels) ) def UpperCamelCase ( self : Union[str, Any] ): A , A = self.model_tester.prepare_config_and_inputs_for_common() A = input_dict['input_ids'] A = FalconForCausalLM(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , use_cache=UpperCamelCase__ ) A = input_ids.shape[0] A = model._convert_to_rw_cache(result.past_key_values ) A = model._convert_cache_to_standard_format(UpperCamelCase__ , UpperCamelCase__ ) for layer in range(len(UpperCamelCase__ ) ): for tensor_idx in range(2 ): self.assertTrue(rw_cache[layer][tensor_idx].ndim == 3 ) self.assertTrue(result.past_key_values[layer][tensor_idx].ndim == 4 ) self.assertTrue( torch.all(result.past_key_values[layer][tensor_idx] == standard_cache[layer][tensor_idx] ) ) def UpperCamelCase ( self : Optional[Any] ): A , A = self.model_tester.prepare_config_and_inputs_for_common() A = 3 A = 'multi_label_classification' A = input_dict['input_ids'] A = input_ids.ne(1 ).to(UpperCamelCase__ ) A = ids_tensor( [self.model_tester.batch_size, config.num_labels] , self.model_tester.type_sequence_label_size ).to(torch.float ) A = FalconForSequenceClassification(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , attention_mask=UpperCamelCase__ , labels=UpperCamelCase__ ) self.assertEqual(result.logits.shape , (self.model_tester.batch_size, self.model_tester.num_labels) ) def UpperCamelCase ( self : Optional[Any] ): # Falcon can have different numbers of KV-heads than the number of query heads, so we need # to override this test to use the right head counts. for model_class in self.all_generative_model_classes: A , A = self.model_tester.prepare_config_and_inputs_for_common() # If it doesn't support cache, pass the test if not hasattr(UpperCamelCase__ , 'use_cache' ): return A = model_class(UpperCamelCase__ ).to(UpperCamelCase__ ) if "use_cache" not in inputs: A = True A = model(**UpperCamelCase__ ) # If "past_key_values" is not returned, pass the test (e.g. RWKV uses a different cache name and format) if "past_key_values" not in outputs: return A = ( getattr(UpperCamelCase__ , 'decoder_layers' , UpperCamelCase__ ) or getattr(UpperCamelCase__ , 'num_decoder_layers' , UpperCamelCase__ ) or config.num_hidden_layers ) A = getattr(UpperCamelCase__ , 'num_kv_heads' , config.num_attention_heads ) A = getattr(UpperCamelCase__ , 'd_model' , config.hidden_size ) A = embed_dim // num_attention_heads A = outputs['past_key_values'] self.assertEqual(len(UpperCamelCase__ ) , UpperCamelCase__ ) A , A = inputs['input_ids'].shape for i in range(UpperCamelCase__ ): if config.new_decoder_architecture: A = config.num_attention_heads elif config.multi_query: A = 1 self.assertEqual(len(past_kv[0] ) , 2 ) # K V for the decoder = 2 self.assertEqual( past_kv[i][0].shape , (batch_size, num_attention_heads, seq_length, per_head_embed_dim) ) self.assertEqual( past_kv[i][1].shape , (batch_size, num_attention_heads, seq_length, per_head_embed_dim) ) @require_torch class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' @slow def UpperCamelCase ( self : Union[str, Any] ): A = AutoTokenizer.from_pretrained('Rocketknight1/falcon-rw-1b' ) A = FalconForCausalLM.from_pretrained('Rocketknight1/falcon-rw-1b' ) model.eval() model.to(UpperCamelCase__ ) A = tokenizer('My favorite food is' , return_tensors='pt' ).to(UpperCamelCase__ ) A = ( 'My favorite food is pizza. I love it so much that I have a pizza party every year for my birthday.' ) A = model.generate(**UpperCamelCase__ , do_sample=UpperCamelCase__ , max_new_tokens=19 ) A = tokenizer.batch_decode(UpperCamelCase__ )[0] self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) @slow def UpperCamelCase ( self : Any ): # The big models are way too big for the CI, so we use tiny random models that resemble their # architectures but with much smaller and fewer layers for repo in ["Rocketknight1/tiny-random-falcon-7b", "Rocketknight1/tiny-random-falcon-40b"]: A = AutoTokenizer.from_pretrained(UpperCamelCase__ ) A = FalconForCausalLM.from_pretrained(UpperCamelCase__ ) model.eval() model.to(UpperCamelCase__ ) A = tokenizer('My favorite food is' , return_tensors='pt' ).to(UpperCamelCase__ ) # We just test that these run without errors - the models are randomly initialized # and so the actual text outputs will be garbage model.generate(**UpperCamelCase__ , do_sample=UpperCamelCase__ , max_new_tokens=4 ) model.generate(**UpperCamelCase__ , do_sample=UpperCamelCase__ , max_new_tokens=4 ) model.generate(**UpperCamelCase__ , num_beams=2 , max_new_tokens=4 ) @slow def UpperCamelCase ( self : Optional[Any] ): # The big models are way too big for the CI, so we use tiny random models that resemble their # architectures but with much smaller and fewer layers with torch.no_grad(): for repo in [ "Rocketknight1/falcon-rw-1b", "Rocketknight1/tiny-random-falcon-7b", "Rocketknight1/tiny-random-falcon-40b", ]: A = AutoTokenizer.from_pretrained(UpperCamelCase__ ) A = FalconForCausalLM.from_pretrained(UpperCamelCase__ ) model.eval() model.to(device=UpperCamelCase__ ) A = tokenizer('My favorite food is' , return_tensors='pt' ).to(UpperCamelCase__ ) # Test results are the same with and without cache A = model.generate(**UpperCamelCase__ , do_sample=UpperCamelCase__ , max_new_tokens=20 , use_cache=UpperCamelCase__ ) A = model.generate(**UpperCamelCase__ , do_sample=UpperCamelCase__ , max_new_tokens=20 , use_cache=UpperCamelCase__ ) self.assertTrue((outputs_cache - outputs_no_cache).sum().item() == 0 )
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Optional[int]: if b == 0: return 1 if (b % 2) == 0: return actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) else: return a * actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> float: if b < 0: return 1 / actual_power(lowerCAmelCase, lowerCAmelCase ) return actual_power(lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": print(power(-2, -3))
699
1
import argparse import gdown import numpy as np import torch from huggingface_hub import hf_hub_download from transformers import ( CLIPTokenizer, CLIPTokenizerFast, VideoMAEImageProcessor, XCLIPConfig, XCLIPModel, XCLIPProcessor, XCLIPTextConfig, XCLIPVisionConfig, ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Optional[int] ) -> Tuple: A = XCLIPTextConfig() # derive patch size from model name A = model_name.find('patch' ) A = int(model_name[start_idx + len('patch' ) : start_idx + len('patch' ) + 2] ) A = XCLIPVisionConfig(patch_size=lowerCAmelCase, num_frames=lowerCAmelCase ) if "large" in model_name: A = 768 A = 3_072 A = 12 A = 1_024 A = 4_096 A = 16 A = 24 A = 768 A = 3_072 if model_name == "xclip-large-patch14-16-frames": A = 336 A = XCLIPConfig.from_text_vision_configs(lowerCAmelCase, lowerCAmelCase ) if "large" in model_name: A = 768 return config def __UpperCamelCase (lowerCAmelCase : Dict ) -> List[Any]: # text encoder if name == "token_embedding.weight": A = name.replace('token_embedding.weight', 'text_model.embeddings.token_embedding.weight' ) if name == "positional_embedding": A = name.replace('positional_embedding', 'text_model.embeddings.position_embedding.weight' ) if "ln_1" in name: A = name.replace('ln_1', 'layer_norm1' ) if "ln_2" in name: A = name.replace('ln_2', 'layer_norm2' ) if "c_fc" in name: A = name.replace('c_fc', 'fc1' ) if "c_proj" in name: A = name.replace('c_proj', 'fc2' ) if name.startswith('transformer.resblocks' ): A = name.replace('transformer.resblocks', 'text_model.encoder.layers' ) if "attn.out_proj" in name and "message" not in name: A = name.replace('attn.out_proj', 'self_attn.out_proj' ) if "ln_final" in name: A = name.replace('ln_final', 'text_model.final_layer_norm' ) # visual encoder if name == "visual.class_embedding": A = name.replace('visual.class_embedding', 'vision_model.embeddings.class_embedding' ) if name == "visual.positional_embedding": A = name.replace('visual.positional_embedding', 'vision_model.embeddings.position_embedding.weight' ) if name.startswith('visual.transformer.resblocks' ): A = name.replace('visual.transformer.resblocks', 'vision_model.encoder.layers' ) if "visual.conv1" in name: A = name.replace('visual.conv1', 'vision_model.embeddings.patch_embedding' ) if "visual.ln_pre" in name: A = name.replace('visual.ln_pre', 'vision_model.pre_layernorm' ) if "visual.ln_post" in name: A = name.replace('visual.ln_post', 'vision_model.post_layernorm' ) if "visual.proj" in name: A = name.replace('visual.proj', 'visual_projection.weight' ) if "text_projection" in name: A = name.replace('text_projection', 'text_projection.weight' ) # things on top if "prompts_visual_proj" in name: A = name.replace('prompts_visual_proj', 'prompts_visual_projection' ) if "prompts_visual_ln" in name: A = name.replace('prompts_visual_ln', 'prompts_visual_layernorm' ) # mit if name == "mit.positional_embedding": A = name.replace('positional', 'position' ) if name.startswith('mit.resblocks' ): A = name.replace('mit.resblocks', 'mit.encoder.layers' ) # prompts generator if name.startswith('prompts_generator.norm' ): A = name.replace('prompts_generator.norm', 'prompts_generator.layernorm' ) return name def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : List[str] ) -> int: for key in orig_state_dict.copy().keys(): A = orig_state_dict.pop(lowerCAmelCase ) if "attn.in_proj" in key: A = key.split('.' ) if key.startswith('visual' ): A = key_split[3] A = config.vision_config.hidden_size if "message_attn" in key: if "weight" in key: A = val[ :dim, : ] A = val[ dim : dim * 2, : ] A = val[ -dim:, : ] else: A = val[ :dim ] A = val[ dim : dim * 2 ] A = val[ -dim: ] else: if "weight" in key: A = val[ :dim, : ] A = val[ dim : dim * 2, : ] A = val[ -dim:, : ] else: A = val[:dim] A = val[ dim : dim * 2 ] A = val[-dim:] elif key.startswith('mit' ): A = key_split[2] A = config.vision_config.mit_hidden_size if "weight" in key: A = val[:dim, :] A = val[dim : dim * 2, :] A = val[-dim:, :] else: A = val[:dim] A = val[dim : dim * 2] A = val[-dim:] else: A = key_split[2] A = config.text_config.hidden_size if "weight" in key: A = val[:dim, :] A = val[ dim : dim * 2, : ] A = val[-dim:, :] else: A = val[:dim] A = val[ dim : dim * 2 ] A = val[-dim:] else: A = rename_key(lowerCAmelCase ) if new_key_name in ["visual_projection.weight", "text_projection.weight"]: A = val.T A = val return orig_state_dict def __UpperCamelCase (lowerCAmelCase : Tuple ) -> Optional[int]: if num_frames == 8: A = 'eating_spaghetti_8_frames.npy' elif num_frames == 16: A = 'eating_spaghetti.npy' elif num_frames == 32: A = 'eating_spaghetti_32_frames.npy' A = hf_hub_download( repo_id='hf-internal-testing/spaghetti-video', filename=lowerCAmelCase, repo_type='dataset', ) A = np.load(lowerCAmelCase ) return list(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : List[str]=None, lowerCAmelCase : Optional[Any]=False ) -> Union[str, Any]: A = { # fully supervised kinetics-400 checkpoints 'xclip-base-patch32': 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth', 'xclip-base-patch32-16-frames': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth' ), 'xclip-base-patch16': 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth', 'xclip-base-patch16-16-frames': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth' ), 'xclip-large-patch14': 'https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&amp;export=download&amp;confirm=t&amp;uuid=b26caedc-88e2-473e-830a-9d158b653cdb', 'xclip-large-patch14-16-frames': 'https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&amp;export=download&amp;confirm=t&amp;uuid=538fa810-e671-4050-b385-9a623f89804f', # fully supervised kinetics-600 checkpoints 'xclip-base-patch16-kinetics-600': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth' ), 'xclip-base-patch16-kinetics-600-16-frames': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth' ), 'xclip-large-patch14-kinetics-600': 'https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&amp;export=download&amp;confirm=t&amp;uuid=141d4977-4a65-44ae-864f-4b0c19f838be', # few shot 'xclip-base-patch16-hmdb-2-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth' ), 'xclip-base-patch16-hmdb-4-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth' ), 'xclip-base-patch16-hmdb-8-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth' ), 'xclip-base-patch16-hmdb-16-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth' ), 'xclip-base-patch16-ucf-2-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth' ), 'xclip-base-patch16-ucf-4-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth' ), 'xclip-base-patch16-ucf-8-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth' ), 'xclip-base-patch16-ucf-16-shot': ( 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth' ), # zero shot 'xclip-base-patch16-zero-shot': 'https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth', } A = model_to_url[model_name] A = 8 if "16-frames" in model_name: A = 16 elif "shot" in model_name: A = 32 A = get_xclip_config(lowerCAmelCase, lowerCAmelCase ) A = XCLIPModel(lowerCAmelCase ) model.eval() if "drive" in checkpoint_url: A = 'pytorch_model.bin' gdown.cached_download(lowerCAmelCase, lowerCAmelCase, quiet=lowerCAmelCase ) A = torch.load(lowerCAmelCase, map_location='cpu' )['model'] else: A = torch.hub.load_state_dict_from_url(lowerCAmelCase )['model'] A = convert_state_dict(lowerCAmelCase, lowerCAmelCase ) A = XCLIPModel(lowerCAmelCase ) A , A = model.load_state_dict(lowerCAmelCase, strict=lowerCAmelCase ) assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"] model.eval() A = 336 if model_name == 'xclip-large-patch14-16-frames' else 224 A = VideoMAEImageProcessor(size=lowerCAmelCase ) A = CLIPTokenizer.from_pretrained('openai/clip-vit-base-patch32' ) A = CLIPTokenizerFast.from_pretrained('openai/clip-vit-base-patch32' ) A = XCLIPProcessor(image_processor=lowerCAmelCase, tokenizer=lowerCAmelCase ) A = prepare_video(lowerCAmelCase ) A = processor( text=['playing sports', 'eating spaghetti', 'go shopping'], videos=lowerCAmelCase, return_tensors='pt', padding=lowerCAmelCase ) print('Shape of pixel values:', inputs.pixel_values.shape ) with torch.no_grad(): A = model(**lowerCAmelCase ) # Verify outputs A = outputs.logits_per_video A = logits_per_video.softmax(dim=1 ) print('Probs:', lowerCAmelCase ) # kinetics-400 if model_name == "xclip-base-patch32": A = torch.tensor([[0.0019, 0.9951, 0.0030]] ) elif model_name == "xclip-base-patch32-16-frames": A = torch.tensor([[7.0999E-04, 9.9883E-01, 4.5580E-04]] ) elif model_name == "xclip-base-patch16": A = torch.tensor([[0.0083, 0.9681, 0.0236]] ) elif model_name == "xclip-base-patch16-16-frames": A = torch.tensor([[7.6937E-04, 9.9728E-01, 1.9473E-03]] ) elif model_name == "xclip-large-patch14": A = torch.tensor([[0.0062, 0.9864, 0.0075]] ) elif model_name == "xclip-large-patch14-16-frames": A = torch.tensor([[3.3877E-04, 9.9937E-01, 2.8888E-04]] ) # kinetics-600 elif model_name == "xclip-base-patch16-kinetics-600": A = torch.tensor([[0.0555, 0.8914, 0.0531]] ) elif model_name == "xclip-base-patch16-kinetics-600-16-frames": A = torch.tensor([[3.8554E-04, 9.9929E-01, 3.2754E-04]] ) elif model_name == "xclip-large-patch14-kinetics-600": A = torch.tensor([[0.0036, 0.9920, 0.0045]] ) # few shot elif model_name == "xclip-base-patch16-hmdb-2-shot": A = torch.tensor([[7.1890E-06, 9.9994E-01, 5.6559E-05]] ) elif model_name == "xclip-base-patch16-hmdb-4-shot": A = torch.tensor([[1.0320E-05, 9.9993E-01, 6.2435E-05]] ) elif model_name == "xclip-base-patch16-hmdb-8-shot": A = torch.tensor([[4.1377E-06, 9.9990E-01, 9.8386E-05]] ) elif model_name == "xclip-base-patch16-hmdb-16-shot": A = torch.tensor([[4.1347E-05, 9.9962E-01, 3.3411E-04]] ) elif model_name == "xclip-base-patch16-ucf-2-shot": A = torch.tensor([[8.5857E-05, 9.9928E-01, 6.3291E-04]] ) elif model_name == "xclip-base-patch16-ucf-4-shot": A = torch.tensor([[8.5857E-05, 9.9928E-01, 6.3291E-04]] ) elif model_name == "xclip-base-patch16-ucf-8-shot": A = torch.tensor([[0.0027, 0.9904, 0.0070]] ) elif model_name == "xclip-base-patch16-ucf-16-shot": A = torch.tensor([[9.8219E-04, 9.9593E-01, 3.0863E-03]] ) # zero shot elif model_name == "xclip-base-patch16-zero-shot": A = torch.tensor([[3.5082E-04, 9.9785E-01, 1.7966E-03]] ) else: raise ValueError(f'''Model name {model_name} not supported''' ) assert torch.allclose(lowerCAmelCase, lowerCAmelCase, atol=1E-3 ) print('Looks ok!' ) if pytorch_dump_folder_path is not None: print(f'''Saving model {model_name} to {pytorch_dump_folder_path}''' ) model.save_pretrained(lowerCAmelCase ) if push_to_hub: print('Pushing model, processor and slow tokenizer files to the hub...' ) model.push_to_hub(lowerCAmelCase, organization='nielsr' ) processor.push_to_hub(lowerCAmelCase, organization='nielsr' ) slow_tokenizer.push_to_hub(lowerCAmelCase, organization='nielsr' ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() # Required parameters parser.add_argument( "--model_name", default="xclip-base-patch32", type=str, help="Name of the model.", ) parser.add_argument( "--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory." ) parser.add_argument( "--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub." ) _UpperCAmelCase = parser.parse_args() convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
699
def __UpperCamelCase (lowerCAmelCase : list[int] ) -> int: if not numbers: return 0 if not isinstance(lowerCAmelCase, (list, tuple) ) or not all( isinstance(lowerCAmelCase, lowerCAmelCase ) for number in numbers ): raise ValueError('numbers must be an iterable of integers' ) A = A = A = numbers[0] for i in range(1, len(lowerCAmelCase ) ): # update the maximum and minimum subarray products A = numbers[i] if number < 0: A , A = min_till_now, max_till_now A = max(lowerCAmelCase, max_till_now * number ) A = min(lowerCAmelCase, min_till_now * number ) # update the maximum product found till now A = max(lowerCAmelCase, lowerCAmelCase ) return max_prod
699
1
from typing import List, Union import numpy as np from ..tokenization_utils import TruncationStrategy from ..utils import add_end_docstrings, logging from .base import PIPELINE_INIT_ARGS, ArgumentHandler, ChunkPipeline _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def UpperCamelCase ( self : int , UpperCamelCase__ : List[str] ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = [label.strip() for label in labels.split(',' ) if label.strip()] return labels def __call__( self : Dict , UpperCamelCase__ : List[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Tuple ): if len(UpperCamelCase__ ) == 0 or len(UpperCamelCase__ ) == 0: raise ValueError('You must include at least one label and at least one sequence.' ) if hypothesis_template.format(labels[0] ) == hypothesis_template: raise ValueError( ( 'The provided hypothesis_template "{}" was not able to be formatted with the target labels. ' 'Make sure the passed template includes formatting syntax such as {{}} where the label should go.' ).format(UpperCamelCase__ ) ) if isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = [sequences] A = [] for sequence in sequences: sequence_pairs.extend([[sequence, hypothesis_template.format(UpperCamelCase__ )] for label in labels] ) return sequence_pairs, sequences @add_end_docstrings(__lowercase ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : Any=ZeroShotClassificationArgumentHandler() , *UpperCamelCase__ : Any , **UpperCamelCase__ : Dict ): A = args_parser super().__init__(*UpperCamelCase__ , **UpperCamelCase__ ) if self.entailment_id == -1: logger.warning( 'Failed to determine \'entailment\' label id from the label2id mapping in the model config. Setting to ' '-1. Define a descriptive label2id mapping in the model config to ensure correct outputs.' ) @property def UpperCamelCase ( self : str ): for label, ind in self.model.config.labelaid.items(): if label.lower().startswith('entail' ): return ind return -1 def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Any=True , UpperCamelCase__ : Dict=True , UpperCamelCase__ : Optional[Any]=TruncationStrategy.ONLY_FIRST , **UpperCamelCase__ : str ): A = self.framework if self.tokenizer.pad_token is None: # Override for tokenizers not supporting padding logger.error( 'Tokenizer was not supporting padding necessary for zero-shot, attempting to use ' ' `pad_token=eos_token`' ) A = self.tokenizer.eos_token try: A = self.tokenizer( UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , return_tensors=UpperCamelCase__ , padding=UpperCamelCase__ , truncation=UpperCamelCase__ , ) except Exception as e: if "too short" in str(UpperCamelCase__ ): # tokenizers might yell that we want to truncate # to a value that is not even reached by the input. # In that case we don't want to truncate. # It seems there's not a really better way to catch that # exception. A = self.tokenizer( UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , return_tensors=UpperCamelCase__ , padding=UpperCamelCase__ , truncation=TruncationStrategy.DO_NOT_TRUNCATE , ) else: raise e return inputs def UpperCamelCase ( self : Any , **UpperCamelCase__ : List[str] ): if kwargs.get('multi_class' , UpperCamelCase__ ) is not None: A = kwargs['multi_class'] logger.warning( 'The `multi_class` argument has been deprecated and renamed to `multi_label`. ' '`multi_class` will be removed in a future version of Transformers.' ) A = {} if "candidate_labels" in kwargs: A = self._args_parser._parse_labels(kwargs['candidate_labels'] ) if "hypothesis_template" in kwargs: A = kwargs['hypothesis_template'] A = {} if "multi_label" in kwargs: A = kwargs['multi_label'] return preprocess_params, {}, postprocess_params def __call__( self : Optional[int] , UpperCamelCase__ : Union[str, List[str]] , *UpperCamelCase__ : Tuple , **UpperCamelCase__ : List[str] , ): if len(UpperCamelCase__ ) == 0: pass elif len(UpperCamelCase__ ) == 1 and "candidate_labels" not in kwargs: A = args[0] else: raise ValueError(f'''Unable to understand extra arguments {args}''' ) return super().__call__(UpperCamelCase__ , **UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Dict=None , UpperCamelCase__ : str="This example is {}." ): A , A = self._args_parser(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) for i, (candidate_label, sequence_pair) in enumerate(zip(UpperCamelCase__ , UpperCamelCase__ ) ): A = self._parse_and_tokenize([sequence_pair] ) yield { "candidate_label": candidate_label, "sequence": sequences[0], "is_last": i == len(UpperCamelCase__ ) - 1, **model_input, } def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : int ): A = inputs['candidate_label'] A = inputs['sequence'] A = {k: inputs[k] for k in self.tokenizer.model_input_names} A = self.model(**UpperCamelCase__ ) A = { 'candidate_label': candidate_label, 'sequence': sequence, 'is_last': inputs['is_last'], **outputs, } return model_outputs def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Union[str, Any]=False ): A = [outputs['candidate_label'] for outputs in model_outputs] A = [outputs['sequence'] for outputs in model_outputs] A = np.concatenate([output['logits'].numpy() for output in model_outputs] ) A = logits.shape[0] A = len(UpperCamelCase__ ) A = N // n A = logits.reshape((num_sequences, n, -1) ) if multi_label or len(UpperCamelCase__ ) == 1: # softmax over the entailment vs. contradiction dim for each label independently A = self.entailment_id A = -1 if entailment_id == 0 else 0 A = reshaped_outputs[..., [contradiction_id, entailment_id]] A = np.exp(UpperCamelCase__ ) / np.exp(UpperCamelCase__ ).sum(-1 , keepdims=UpperCamelCase__ ) A = scores[..., 1] else: # softmax the "entailment" logits over all candidate labels A = reshaped_outputs[..., self.entailment_id] A = np.exp(UpperCamelCase__ ) / np.exp(UpperCamelCase__ ).sum(-1 , keepdims=UpperCamelCase__ ) A = list(reversed(scores[0].argsort() ) ) return { "sequence": sequences[0], "labels": [candidate_labels[i] for i in top_inds], "scores": scores[0, top_inds].tolist(), }
699
from .glue import glue_convert_examples_to_features, glue_output_modes, glue_processors, glue_tasks_num_labels from .squad import SquadExample, SquadFeatures, SquadVaProcessor, SquadVaProcessor, squad_convert_examples_to_features from .utils import DataProcessor, InputExample, InputFeatures, SingleSentenceClassificationProcessor from .xnli import xnli_output_modes, xnli_processors, xnli_tasks_num_labels
699
1
import numpy as np import qiskit def __UpperCamelCase (lowerCAmelCase : int = 8, lowerCAmelCase : int | None = None ) -> str: A = np.random.default_rng(seed=lowerCAmelCase ) # Roughly 25% of the qubits will contribute to the key. # So we take more than we need. A = 6 * key_len # Measurement basis for Alice's qubits. A = rng.integers(2, size=lowerCAmelCase ) # The set of states Alice will prepare. A = rng.integers(2, size=lowerCAmelCase ) # Measurement basis for Bob's qubits. A = rng.integers(2, size=lowerCAmelCase ) # Quantum Circuit to simulate BB84 A = qiskit.QuantumCircuit(lowerCAmelCase, name='BB84' ) # Alice prepares her qubits according to rules above. for index, _ in enumerate(lowerCAmelCase ): if alice_state[index] == 1: bbaa_circ.x(lowerCAmelCase ) if alice_basis[index] == 1: bbaa_circ.h(lowerCAmelCase ) bbaa_circ.barrier() # Bob measures the received qubits according to rules above. for index, _ in enumerate(lowerCAmelCase ): if bob_basis[index] == 1: bbaa_circ.h(lowerCAmelCase ) bbaa_circ.barrier() bbaa_circ.measure_all() # Simulate the quantum circuit. A = qiskit.Aer.get_backend('aer_simulator' ) # We only need to run one shot because the key is unique. # Multiple shots will produce the same key. A = qiskit.execute(lowerCAmelCase, lowerCAmelCase, shots=1, seed_simulator=lowerCAmelCase ) # Returns the result of measurement. A = job.result().get_counts(lowerCAmelCase ).most_frequent() # Extracting the generated key from the simulation results. # Only keep measurement results where Alice and Bob chose the same basis. A = ''.join( [ result_bit for alice_basis_bit, bob_basis_bit, result_bit in zip( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if alice_basis_bit == bob_basis_bit ] ) # Get final key. Pad with 0 if too short, otherwise truncate. A = gen_key[:key_len] if len(lowerCAmelCase ) >= key_len else gen_key.ljust(lowerCAmelCase, '0' ) return key if __name__ == "__main__": print(F'''The generated key is : {bbaa(8, seed=0)}''') from doctest import testmod testmod()
699
from ...configuration_utils import PretrainedConfig from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "microsoft/biogpt": "https://huggingface.co/microsoft/biogpt/resolve/main/config.json", # See all BioGPT models at https://huggingface.co/models?filter=biogpt } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = '''biogpt''' def __init__( self : Optional[Any] , UpperCamelCase__ : str=42384 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : Dict=24 , UpperCamelCase__ : Any=16 , UpperCamelCase__ : str=4096 , UpperCamelCase__ : Tuple="gelu" , UpperCamelCase__ : List[str]=0.1 , UpperCamelCase__ : Any=0.1 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : List[Any]=0.02 , UpperCamelCase__ : Dict=1e-1_2 , UpperCamelCase__ : Any=True , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Any=1 , UpperCamelCase__ : List[str]=0 , UpperCamelCase__ : Optional[Any]=2 , **UpperCamelCase__ : List[Any] , ): A = vocab_size A = max_position_embeddings A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = initializer_range A = layer_norm_eps A = scale_embedding A = use_cache A = layerdrop A = activation_dropout super().__init__(pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , **UpperCamelCase__ )
699
1
import os import tempfile import unittest import numpy as np from diffusers.utils import is_flax_available from diffusers.utils.testing_utils import require_flax, slow if is_flax_available(): import jax import jax.numpy as jnp from flax.jax_utils import replicate from flax.training.common_utils import shard from diffusers import FlaxDDIMScheduler, FlaxDiffusionPipeline, FlaxStableDiffusionPipeline @require_flax class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : List[str] ): with tempfile.TemporaryDirectory() as tmpdirname: # pipeline has Flax weights A = FlaxDiffusionPipeline.from_pretrained( 'hf-internal-testing/tiny-stable-diffusion-pipe' , safety_checker=UpperCamelCase__ , cache_dir=UpperCamelCase__ ) A = [t[-1] for t in os.walk(os.path.join(UpperCamelCase__ , os.listdir(UpperCamelCase__ )[0] , 'snapshots' ) )] A = [item for sublist in all_root_files for item in sublist] # None of the downloaded files should be a PyTorch file even if we have some here: # https://huggingface.co/hf-internal-testing/tiny-stable-diffusion-pipe/blob/main/unet/diffusion_pytorch_model.bin assert not any(f.endswith('.bin' ) for f in files ) @slow @require_flax class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : Dict ): A , A = FlaxStableDiffusionPipeline.from_pretrained( 'hf-internal-testing/tiny-stable-diffusion-pipe' , safety_checker=UpperCamelCase__ ) A = ( 'A cinematic film still of Morgan Freeman starring as Jimi Hendrix, portrait, 40mm lens, shallow depth of' ' field, close up, split lighting, cinematic' ) A = jax.random.PRNGKey(0 ) A = 4 A = jax.device_count() A = num_samples * [prompt] A = pipeline.prepare_inputs(UpperCamelCase__ ) # shard inputs and rng A = replicate(UpperCamelCase__ ) A = jax.random.split(UpperCamelCase__ , UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images.shape == (num_samples, 1, 64, 64, 3) if jax.device_count() == 8: assert np.abs(np.abs(images[0, 0, :2, :2, -2:] , dtype=np.floataa ).sum() - 4.1_514_745 ) < 1e-3 assert np.abs(np.abs(UpperCamelCase__ , dtype=np.floataa ).sum() - 49_947.875 ) < 5e-1 A = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:] ) ) ) assert len(UpperCamelCase__ ) == num_samples def UpperCamelCase ( self : Union[str, Any] ): A , A = FlaxStableDiffusionPipeline.from_pretrained( 'CompVis/stable-diffusion-v1-4' , revision='flax' , safety_checker=UpperCamelCase__ ) A = ( 'A cinematic film still of Morgan Freeman starring as Jimi Hendrix, portrait, 40mm lens, shallow depth of' ' field, close up, split lighting, cinematic' ) A = jax.random.PRNGKey(0 ) A = 50 A = jax.device_count() A = num_samples * [prompt] A = pipeline.prepare_inputs(UpperCamelCase__ ) # shard inputs and rng A = replicate(UpperCamelCase__ ) A = jax.random.split(UpperCamelCase__ , UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images.shape == (num_samples, 1, 512, 512, 3) if jax.device_count() == 8: assert np.abs((np.abs(images[0, 0, :2, :2, -2:] , dtype=np.floataa ).sum() - 0.05_652_401) ) < 1e-3 assert np.abs((np.abs(UpperCamelCase__ , dtype=np.floataa ).sum() - 2_383_808.2) ) < 5e-1 def UpperCamelCase ( self : Tuple ): A , A = FlaxStableDiffusionPipeline.from_pretrained( 'CompVis/stable-diffusion-v1-4' , revision='bf16' , dtype=jnp.bfloataa , safety_checker=UpperCamelCase__ ) A = ( 'A cinematic film still of Morgan Freeman starring as Jimi Hendrix, portrait, 40mm lens, shallow depth of' ' field, close up, split lighting, cinematic' ) A = jax.random.PRNGKey(0 ) A = 50 A = jax.device_count() A = num_samples * [prompt] A = pipeline.prepare_inputs(UpperCamelCase__ ) # shard inputs and rng A = replicate(UpperCamelCase__ ) A = jax.random.split(UpperCamelCase__ , UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images.shape == (num_samples, 1, 512, 512, 3) if jax.device_count() == 8: assert np.abs((np.abs(images[0, 0, :2, :2, -2:] , dtype=np.floataa ).sum() - 0.04_003_906) ) < 1e-3 assert np.abs((np.abs(UpperCamelCase__ , dtype=np.floataa ).sum() - 2_373_516.75) ) < 5e-1 def UpperCamelCase ( self : Union[str, Any] ): A , A = FlaxStableDiffusionPipeline.from_pretrained( 'CompVis/stable-diffusion-v1-4' , revision='bf16' , dtype=jnp.bfloataa ) A = ( 'A cinematic film still of Morgan Freeman starring as Jimi Hendrix, portrait, 40mm lens, shallow depth of' ' field, close up, split lighting, cinematic' ) A = jax.random.PRNGKey(0 ) A = 50 A = jax.device_count() A = num_samples * [prompt] A = pipeline.prepare_inputs(UpperCamelCase__ ) # shard inputs and rng A = replicate(UpperCamelCase__ ) A = jax.random.split(UpperCamelCase__ , UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images.shape == (num_samples, 1, 512, 512, 3) if jax.device_count() == 8: assert np.abs((np.abs(images[0, 0, :2, :2, -2:] , dtype=np.floataa ).sum() - 0.04_003_906) ) < 1e-3 assert np.abs((np.abs(UpperCamelCase__ , dtype=np.floataa ).sum() - 2_373_516.75) ) < 5e-1 def UpperCamelCase ( self : Union[str, Any] ): A = FlaxDDIMScheduler( beta_start=0.00_085 , beta_end=0.012 , beta_schedule='scaled_linear' , set_alpha_to_one=UpperCamelCase__ , steps_offset=1 , ) A , A = FlaxStableDiffusionPipeline.from_pretrained( 'CompVis/stable-diffusion-v1-4' , revision='bf16' , dtype=jnp.bfloataa , scheduler=UpperCamelCase__ , safety_checker=UpperCamelCase__ , ) A = scheduler.create_state() A = scheduler_state A = ( 'A cinematic film still of Morgan Freeman starring as Jimi Hendrix, portrait, 40mm lens, shallow depth of' ' field, close up, split lighting, cinematic' ) A = jax.random.PRNGKey(0 ) A = 50 A = jax.device_count() A = num_samples * [prompt] A = pipeline.prepare_inputs(UpperCamelCase__ ) # shard inputs and rng A = replicate(UpperCamelCase__ ) A = jax.random.split(UpperCamelCase__ , UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images.shape == (num_samples, 1, 512, 512, 3) if jax.device_count() == 8: assert np.abs((np.abs(images[0, 0, :2, :2, -2:] , dtype=np.floataa ).sum() - 0.045_043_945) ) < 1e-3 assert np.abs((np.abs(UpperCamelCase__ , dtype=np.floataa ).sum() - 2_347_693.5) ) < 5e-1 def UpperCamelCase ( self : Any ): A = ( 'A cinematic film still of Morgan Freeman starring as Jimi Hendrix, portrait, 40mm lens, shallow depth of' ' field, close up, split lighting, cinematic' ) A = jax.device_count() A = num_samples * [prompt] A = jax.random.split(jax.random.PRNGKey(0 ) , UpperCamelCase__ ) A , A = FlaxStableDiffusionPipeline.from_pretrained( 'CompVis/stable-diffusion-v1-4' , revision='bf16' , dtype=jnp.bfloataa , safety_checker=UpperCamelCase__ , ) A = replicate(UpperCamelCase__ ) A = pipeline.prepare_inputs(UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images.shape == (num_samples, 1, 512, 512, 3) A = images[2, 0, 256, 10:17, 1] # With memory efficient attention A , A = FlaxStableDiffusionPipeline.from_pretrained( 'CompVis/stable-diffusion-v1-4' , revision='bf16' , dtype=jnp.bfloataa , safety_checker=UpperCamelCase__ , use_memory_efficient_attention=UpperCamelCase__ , ) A = replicate(UpperCamelCase__ ) A = pipeline.prepare_inputs(UpperCamelCase__ ) A = shard(UpperCamelCase__ ) A = pipeline(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , jit=UpperCamelCase__ ).images assert images_eff.shape == (num_samples, 1, 512, 512, 3) A = images[2, 0, 256, 10:17, 1] # I checked the results visually and they are very similar. However, I saw that the max diff is `1` and the `sum` # over the 8 images is exactly `256`, which is very suspicious. Testing a random slice for now. assert abs(slice_eff - slice ).max() < 1e-2
699
import sys def __UpperCamelCase (lowerCAmelCase : Dict ) -> Dict: A = len(lowerCAmelCase ) A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] for chain_length in range(2, lowerCAmelCase ): for a in range(1, n - chain_length + 1 ): A = a + chain_length - 1 A = sys.maxsize for c in range(lowerCAmelCase, lowerCAmelCase ): A = ( matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b] ) if cost < matrix[a][b]: A = cost A = c return matrix, sol def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Union[str, Any], lowerCAmelCase : Union[str, Any] ) -> List[str]: if i == j: print('A' + str(lowerCAmelCase ), end=' ' ) else: print('(', end=' ' ) print_optiomal_solution(lowerCAmelCase, lowerCAmelCase, optimal_solution[i][j] ) print_optiomal_solution(lowerCAmelCase, optimal_solution[i][j] + 1, lowerCAmelCase ) print(')', end=' ' ) def __UpperCamelCase () -> List[str]: A = [30, 35, 15, 5, 10, 20, 25] A = len(lowerCAmelCase ) # Size of matrix created from above array will be # 30*35 35*15 15*5 5*10 10*20 20*25 A , A = matrix_chain_order(lowerCAmelCase ) print('No. of Operation required: ' + str(matrix[1][n - 1] ) ) print_optiomal_solution(lowerCAmelCase, 1, n - 1 ) if __name__ == "__main__": main()
699
1
from collections import OrderedDict from typing import Any, List, Mapping, Optional from ... import PreTrainedTokenizer, TensorType, is_torch_available from ...configuration_utils import PretrainedConfig from ...onnx import OnnxConfigWithPast, PatchingSpec from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "Salesforce/codegen-350M-nl": "https://huggingface.co/Salesforce/codegen-350M-nl/resolve/main/config.json", "Salesforce/codegen-350M-multi": "https://huggingface.co/Salesforce/codegen-350M-multi/resolve/main/config.json", "Salesforce/codegen-350M-mono": "https://huggingface.co/Salesforce/codegen-350M-mono/resolve/main/config.json", "Salesforce/codegen-2B-nl": "https://huggingface.co/Salesforce/codegen-2B-nl/resolve/main/config.json", "Salesforce/codegen-2B-multi": "https://huggingface.co/Salesforce/codegen-2B-multi/resolve/main/config.json", "Salesforce/codegen-2B-mono": "https://huggingface.co/Salesforce/codegen-2B-mono/resolve/main/config.json", "Salesforce/codegen-6B-nl": "https://huggingface.co/Salesforce/codegen-6B-nl/resolve/main/config.json", "Salesforce/codegen-6B-multi": "https://huggingface.co/Salesforce/codegen-6B-multi/resolve/main/config.json", "Salesforce/codegen-6B-mono": "https://huggingface.co/Salesforce/codegen-6B-mono/resolve/main/config.json", "Salesforce/codegen-16B-nl": "https://huggingface.co/Salesforce/codegen-16B-nl/resolve/main/config.json", "Salesforce/codegen-16B-multi": "https://huggingface.co/Salesforce/codegen-16B-multi/resolve/main/config.json", "Salesforce/codegen-16B-mono": "https://huggingface.co/Salesforce/codegen-16B-mono/resolve/main/config.json", } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = '''codegen''' SCREAMING_SNAKE_CASE : Optional[int] = { '''max_position_embeddings''': '''n_positions''', '''hidden_size''': '''n_embd''', '''num_attention_heads''': '''n_head''', '''num_hidden_layers''': '''n_layer''', } def __init__( self : int , UpperCamelCase__ : List[Any]=50400 , UpperCamelCase__ : Optional[int]=2048 , UpperCamelCase__ : Any=2048 , UpperCamelCase__ : List[str]=4096 , UpperCamelCase__ : Tuple=28 , UpperCamelCase__ : Optional[int]=16 , UpperCamelCase__ : Dict=64 , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[str]="gelu_new" , UpperCamelCase__ : Optional[int]=0.0 , UpperCamelCase__ : Dict=0.0 , UpperCamelCase__ : Any=0.0 , UpperCamelCase__ : Tuple=1e-5 , UpperCamelCase__ : int=0.02 , UpperCamelCase__ : int=True , UpperCamelCase__ : Any=50256 , UpperCamelCase__ : str=50256 , UpperCamelCase__ : Optional[Any]=False , **UpperCamelCase__ : Any , ): A = vocab_size A = n_ctx A = n_positions A = n_embd A = n_layer A = n_head A = n_inner A = rotary_dim A = activation_function A = resid_pdrop A = embd_pdrop A = attn_pdrop A = layer_norm_epsilon A = initializer_range A = use_cache A = bos_token_id A = eos_token_id super().__init__( bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , tie_word_embeddings=UpperCamelCase__ , **UpperCamelCase__ ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : PretrainedConfig , UpperCamelCase__ : str = "default" , UpperCamelCase__ : List[PatchingSpec] = None , UpperCamelCase__ : bool = False , ): super().__init__(UpperCamelCase__ , task=UpperCamelCase__ , patching_specs=UpperCamelCase__ , use_past=UpperCamelCase__ ) if not getattr(self._config , 'pad_token_id' , UpperCamelCase__ ): # TODO: how to do that better? A = 0 @property def UpperCamelCase ( self : List[Any] ): A = OrderedDict({'input_ids': {0: 'batch', 1: 'sequence'}} ) if self.use_past: self.fill_with_past_key_values_(UpperCamelCase__ , direction='inputs' ) A = {0: 'batch', 1: 'past_sequence + sequence'} else: A = {0: 'batch', 1: 'sequence'} return common_inputs @property def UpperCamelCase ( self : Optional[int] ): return self._config.n_layer @property def UpperCamelCase ( self : Any ): return self._config.n_head def UpperCamelCase ( self : Tuple , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = super(UpperCamelCase__ , self ).generate_dummy_inputs( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) # We need to order the input in the way they appears in the forward() A = OrderedDict({'input_ids': common_inputs['input_ids']} ) # Need to add the past_keys if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape # Not using the same length for past_key_values A = seqlen + 2 A = ( batch, self.num_attention_heads, past_key_values_length, self._config.hidden_size // self.num_attention_heads, ) A = [ (torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) for _ in range(self.num_layers ) ] A = common_inputs['attention_mask'] if self.use_past: A = ordered_inputs['attention_mask'].dtype A = torch.cat( [ordered_inputs['attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ , dtype=UpperCamelCase__ )] , dim=1 ) return ordered_inputs @property def UpperCamelCase ( self : List[Any] ): return 13
699
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> bool: return all(number % divisor != 0 for divisor in range(2, isqrt(lowerCAmelCase ) + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int = 10**6 ) -> int: A = 0 A = 1 A = 7 while prime_candidate < max_prime: primes_count += is_prime(lowerCAmelCase ) cube_index += 1 prime_candidate += 6 * cube_index return primes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
1
import math from typing import List, Optional, Tuple, Union import numpy as np import torch from ..configuration_utils import ConfigMixin, register_to_config from .scheduling_utils import SchedulerMixin, SchedulerOutput class _UpperCAmelCase ( __lowercase , __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = 1 @register_to_config def __init__( self : Optional[Any] , UpperCamelCase__ : int = 1000 , UpperCamelCase__ : Optional[Union[np.ndarray, List[float]]] = None ): # set `betas`, `alphas`, `timesteps` self.set_timesteps(UpperCamelCase__ ) # standard deviation of the initial noise distribution A = 1.0 # For now we only support F-PNDM, i.e. the runge-kutta method # For more information on the algorithm please take a look at the paper: https://arxiv.org/pdf/2202.09778.pdf # mainly at formula (9), (12), (13) and the Algorithm 2. A = 4 # running values A = [] def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : Union[str, torch.device] = None ): A = num_inference_steps A = torch.linspace(1 , 0 , num_inference_steps + 1 )[:-1] A = torch.cat([steps, torch.tensor([0.0] )] ) if self.config.trained_betas is not None: A = torch.tensor(self.config.trained_betas , dtype=torch.floataa ) else: A = torch.sin(steps * math.pi / 2 ) ** 2 A = (1.0 - self.betas**2) ** 0.5 A = (torch.atana(self.betas , self.alphas ) / math.pi * 2)[:-1] A = timesteps.to(UpperCamelCase__ ) A = [] def UpperCamelCase ( self : Tuple , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : int , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : bool = True , ): if self.num_inference_steps is None: raise ValueError( 'Number of inference steps is \'None\', you need to run \'set_timesteps\' after creating the scheduler' ) A = (self.timesteps == timestep).nonzero().item() A = timestep_index + 1 A = sample * self.betas[timestep_index] + model_output * self.alphas[timestep_index] self.ets.append(UpperCamelCase__ ) if len(self.ets ) == 1: A = self.ets[-1] elif len(self.ets ) == 2: A = (3 * self.ets[-1] - self.ets[-2]) / 2 elif len(self.ets ) == 3: A = (23 * self.ets[-1] - 16 * self.ets[-2] + 5 * self.ets[-3]) / 12 else: A = (1 / 24) * (55 * self.ets[-1] - 59 * self.ets[-2] + 37 * self.ets[-3] - 9 * self.ets[-4]) A = self._get_prev_sample(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) if not return_dict: return (prev_sample,) return SchedulerOutput(prev_sample=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : torch.FloatTensor , *UpperCamelCase__ : Union[str, Any] , **UpperCamelCase__ : Any ): return sample def UpperCamelCase ( self : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : List[Any] ): A = self.alphas[timestep_index] A = self.betas[timestep_index] A = self.alphas[prev_timestep_index] A = self.betas[prev_timestep_index] A = (sample - sigma * ets) / max(UpperCamelCase__ , 1e-8 ) A = next_alpha * pred + ets * next_sigma return prev_sample def __len__( self : List[str] ): return self.config.num_train_timesteps
699
import warnings from ...utils import logging from .image_processing_imagegpt import ImageGPTImageProcessor _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , *UpperCamelCase__ : List[Any] , **UpperCamelCase__ : Tuple ): warnings.warn( 'The class ImageGPTFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use ImageGPTImageProcessor instead.' , UpperCamelCase__ , ) super().__init__(*UpperCamelCase__ , **UpperCamelCase__ )
699
1
# Copyright 2023 The HuggingFace Team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import torch from accelerate import PartialState from accelerate.utils.operations import broadcast, gather, gather_object, pad_across_processes, reduce def __UpperCamelCase (lowerCAmelCase : Any ) -> List[str]: return (torch.arange(state.num_processes ) + 1.0 + (state.num_processes * state.process_index)).to(state.device ) def __UpperCamelCase (lowerCAmelCase : Optional[Any] ) -> Any: A = create_tensor(lowerCAmelCase ) A = gather(lowerCAmelCase ) assert gathered_tensor.tolist() == list(range(1, state.num_processes**2 + 1 ) ) def __UpperCamelCase (lowerCAmelCase : Any ) -> Tuple: A = [state.process_index] A = gather_object(lowerCAmelCase ) assert len(lowerCAmelCase ) == state.num_processes, f'''{gathered_obj}, {len(lowerCAmelCase )} != {state.num_processes}''' assert gathered_obj == list(range(state.num_processes ) ), f'''{gathered_obj} != {list(range(state.num_processes ) )}''' def __UpperCamelCase (lowerCAmelCase : Union[str, Any] ) -> int: A = create_tensor(lowerCAmelCase ) A = broadcast(lowerCAmelCase ) assert broadcasted_tensor.shape == torch.Size([state.num_processes] ) assert broadcasted_tensor.tolist() == list(range(1, state.num_processes + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int ) -> List[Any]: # We need to pad the tensor with one more element if we are the main process # to ensure that we can pad if state.is_main_process: A = torch.arange(state.num_processes + 1 ).to(state.device ) else: A = torch.arange(state.num_processes ).to(state.device ) A = pad_across_processes(lowerCAmelCase ) assert padded_tensor.shape == torch.Size([state.num_processes + 1] ) if not state.is_main_process: assert padded_tensor.tolist() == list(range(0, state.num_processes ) ) + [0] def __UpperCamelCase (lowerCAmelCase : Union[str, Any] ) -> List[str]: # For now runs on only two processes if state.num_processes != 2: return A = create_tensor(lowerCAmelCase ) A = reduce(lowerCAmelCase, 'sum' ) A = torch.tensor([4.0, 6] ).to(state.device ) assert torch.allclose(lowerCAmelCase, lowerCAmelCase ), f'''{reduced_tensor} != {truth_tensor}''' def __UpperCamelCase (lowerCAmelCase : str ) -> Union[str, Any]: # For now runs on only two processes if state.num_processes != 2: return A = create_tensor(lowerCAmelCase ) A = reduce(lowerCAmelCase, 'mean' ) A = torch.tensor([2.0, 3] ).to(state.device ) assert torch.allclose(lowerCAmelCase, lowerCAmelCase ), f'''{reduced_tensor} != {truth_tensor}''' def __UpperCamelCase (lowerCAmelCase : List[Any] ) -> Dict: # For xla_spawn (TPUs) main() def __UpperCamelCase () -> Union[str, Any]: A = PartialState() state.print(f'''State: {state}''' ) state.print('testing gather' ) test_gather(lowerCAmelCase ) state.print('testing gather_object' ) test_gather_object(lowerCAmelCase ) state.print('testing broadcast' ) test_broadcast(lowerCAmelCase ) state.print('testing pad_across_processes' ) test_pad_across_processes(lowerCAmelCase ) state.print('testing reduce_sum' ) test_reduce_sum(lowerCAmelCase ) state.print('testing reduce_mean' ) test_reduce_mean(lowerCAmelCase ) if __name__ == "__main__": main()
699
from typing import Any, Dict, Optional import torch import torch.nn.functional as F from torch import nn from ..utils import maybe_allow_in_graph from .activations import get_activation from .attention_processor import Attention from .embeddings import CombinedTimestepLabelEmbeddings @maybe_allow_in_graph class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int]=0.0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : str = "layer_norm" , UpperCamelCase__ : bool = False , ): super().__init__() A = only_cross_attention A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm_zero' A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm' if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None: raise ValueError( f'''`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to''' f''' define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}.''' ) # Define 3 blocks. Each block has its own normalization layer. # 1. Self-Attn if self.use_ada_layer_norm: A = AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A = AdaLayerNormZero(UpperCamelCase__ , UpperCamelCase__ ) else: A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = Attention( query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if only_cross_attention else None , upcast_attention=UpperCamelCase__ , ) # 2. Cross-Attn if cross_attention_dim is not None or double_self_attention: # We currently only use AdaLayerNormZero for self attention where there will only be one attention block. # I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during # the second cross attention block. A = ( AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) ) A = Attention( query_dim=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if not double_self_attention else None , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , upcast_attention=UpperCamelCase__ , ) # is self-attn if encoder_hidden_states is none else: A = None A = None # 3. Feed-forward A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = FeedForward(UpperCamelCase__ , dropout=UpperCamelCase__ , activation_fn=UpperCamelCase__ , final_dropout=UpperCamelCase__ ) # let chunk size default to None A = None A = 0 def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int ): # Sets chunk feed-forward A = chunk_size A = dim def UpperCamelCase ( self : Dict , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , UpperCamelCase__ : Dict[str, Any] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , ): # Notice that normalization is always applied before the real computation in the following blocks. # 1. Self-Attention if self.use_ada_layer_norm: A = self.norma(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A , A , A , A , A = self.norma( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=hidden_states.dtype ) else: A = self.norma(UpperCamelCase__ ) A = cross_attention_kwargs if cross_attention_kwargs is not None else {} A = self.attna( UpperCamelCase__ , encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) if self.use_ada_layer_norm_zero: A = gate_msa.unsqueeze(1 ) * attn_output A = attn_output + hidden_states # 2. Cross-Attention if self.attna is not None: A = ( self.norma(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else self.norma(UpperCamelCase__ ) ) A = self.attna( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) A = attn_output + hidden_states # 3. Feed-forward A = self.norma(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None] if self._chunk_size is not None: # "feed_forward_chunk_size" can be used to save memory if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0: raise ValueError( f'''`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`.''' ) A = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size A = torch.cat( [self.ff(UpperCamelCase__ ) for hid_slice in norm_hidden_states.chunk(UpperCamelCase__ , dim=self._chunk_dim )] , dim=self._chunk_dim , ) else: A = self.ff(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = gate_mlp.unsqueeze(1 ) * ff_output A = ff_output + hidden_states return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 4 , UpperCamelCase__ : float = 0.0 , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : bool = False , ): super().__init__() A = int(dim * mult ) A = dim_out if dim_out is not None else dim if activation_fn == "gelu": A = GELU(UpperCamelCase__ , UpperCamelCase__ ) if activation_fn == "gelu-approximate": A = GELU(UpperCamelCase__ , UpperCamelCase__ , approximate='tanh' ) elif activation_fn == "geglu": A = GEGLU(UpperCamelCase__ , UpperCamelCase__ ) elif activation_fn == "geglu-approximate": A = ApproximateGELU(UpperCamelCase__ , UpperCamelCase__ ) A = nn.ModuleList([] ) # project in self.net.append(UpperCamelCase__ ) # project dropout self.net.append(nn.Dropout(UpperCamelCase__ ) ) # project out self.net.append(nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) ) # FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout if final_dropout: self.net.append(nn.Dropout(UpperCamelCase__ ) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int ): for module in self.net: A = module(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : str = "none" ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) A = approximate def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ , approximate=self.approximate ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) , approximate=self.approximate ).to(dtype=gate.dtype ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int ): A = self.proj(UpperCamelCase__ ) A = self.gelu(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , dim_out * 2 ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) ).to(dtype=gate.dtype ) def UpperCamelCase ( self : str , UpperCamelCase__ : str ): A , A = self.proj(UpperCamelCase__ ).chunk(2 , dim=-1 ) return hidden_states * self.gelu(UpperCamelCase__ ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[int] ): A = self.proj(UpperCamelCase__ ) return x * torch.sigmoid(1.702 * x ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Tuple ): super().__init__() A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , embedding_dim * 2 ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[Any] ): A = self.linear(self.silu(self.emb(UpperCamelCase__ ) ) ) A , A = torch.chunk(UpperCamelCase__ , 2 ) A = self.norm(UpperCamelCase__ ) * (1 + scale) + shift return x class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : int , UpperCamelCase__ : List[str] ): super().__init__() A = CombinedTimestepLabelEmbeddings(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , 6 * embedding_dim , bias=UpperCamelCase__ ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ , eps=1e-6 ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : str , UpperCamelCase__ : Tuple=None ): A = self.linear(self.silu(self.emb(UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=UpperCamelCase__ ) ) ) A , A , A , A , A , A = emb.chunk(6 , dim=1 ) A = self.norm(UpperCamelCase__ ) * (1 + scale_msa[:, None]) + shift_msa[:, None] return x, gate_msa, shift_mlp, scale_mlp, gate_mlp class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[str] = None , UpperCamelCase__ : float = 1e-5 ): super().__init__() A = num_groups A = eps if act_fn is None: A = None else: A = get_activation(UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , out_dim * 2 ) def UpperCamelCase ( self : Any , UpperCamelCase__ : str , UpperCamelCase__ : str ): if self.act: A = self.act(UpperCamelCase__ ) A = self.linear(UpperCamelCase__ ) A = emb[:, :, None, None] A , A = emb.chunk(2 , dim=1 ) A = F.group_norm(UpperCamelCase__ , self.num_groups , eps=self.eps ) A = x * (1 + scale) + shift return x
699
1
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_camembert import CamembertTokenizer else: _UpperCAmelCase = None _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = {"vocab_file": "sentencepiece.bpe.model", "tokenizer_file": "tokenizer.json"} _UpperCAmelCase = { "vocab_file": { "camembert-base": "https://huggingface.co/camembert-base/resolve/main/sentencepiece.bpe.model", }, "tokenizer_file": { "camembert-base": "https://huggingface.co/camembert-base/resolve/main/tokenizer.json", }, } _UpperCAmelCase = { "camembert-base": 512, } _UpperCAmelCase = "▁" class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Tuple = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[Any] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] SCREAMING_SNAKE_CASE : Dict = CamembertTokenizer def __init__( self : Any , UpperCamelCase__ : List[str]=None , UpperCamelCase__ : int=None , UpperCamelCase__ : str="<s>" , UpperCamelCase__ : Union[str, Any]="</s>" , UpperCamelCase__ : str="</s>" , UpperCamelCase__ : int="<s>" , UpperCamelCase__ : List[Any]="<unk>" , UpperCamelCase__ : List[str]="<pad>" , UpperCamelCase__ : Optional[Any]="<mask>" , UpperCamelCase__ : Tuple=["<s>NOTUSED", "</s>NOTUSED"] , **UpperCamelCase__ : int , ): # Mask token behave like a normal word, i.e. include the space before it A = AddedToken(UpperCamelCase__ , lstrip=UpperCamelCase__ , rstrip=UpperCamelCase__ ) if isinstance(UpperCamelCase__ , UpperCamelCase__ ) else mask_token super().__init__( UpperCamelCase__ , tokenizer_file=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , sep_token=UpperCamelCase__ , cls_token=UpperCamelCase__ , unk_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , mask_token=UpperCamelCase__ , additional_special_tokens=UpperCamelCase__ , **UpperCamelCase__ , ) A = vocab_file A = False if not self.vocab_file else True def UpperCamelCase ( self : List[str] , UpperCamelCase__ : List[int] , UpperCamelCase__ : Optional[List[int]] = None ): if token_ids_a is None: return [self.cls_token_id] + token_ids_a + [self.sep_token_id] A = [self.cls_token_id] A = [self.sep_token_id] return cls + token_ids_a + sep + sep + token_ids_a + sep def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : List[int] , UpperCamelCase__ : Optional[List[int]] = None ): A = [self.sep_token_id] A = [self.cls_token_id] if token_ids_a is None: return len(cls + token_ids_a + sep ) * [0] return len(cls + token_ids_a + sep + sep + token_ids_a + sep ) * [0] def UpperCamelCase ( self : Any , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not self.can_save_slow_tokenizer: raise ValueError( 'Your fast tokenizer does not have the necessary information to save the vocabulary for a slow ' 'tokenizer.' ) if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCamelCase__ ): copyfile(self.vocab_file , UpperCamelCase__ ) return (out_vocab_file,)
699
import json import os from typing import Dict, List, Optional, Tuple from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "vocab_file": "vocab.json", "tokenizer_config_file": "tokenizer_config.json", "merges_file": "merges.txt", } _UpperCAmelCase = { "vocab_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/vocab.json" ), }, "tokenizer_config_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/tokenizer_config.json" ), }, "merges_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/merges.txt" ), }, } _UpperCAmelCase = "</w>" _UpperCAmelCase = "@@ " def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> List[str]: A = set() A = word[0] for char in word[1:]: pairs.add((prev_char, char) ) A = char return pairs # Speech2Text2 has no max input length _UpperCAmelCase = {"facebook/s2t-wav2vec2-large-en-de": 1_024} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] def __init__( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="<pad>" , UpperCamelCase__ : int="</s>" , UpperCamelCase__ : Tuple="<unk>" , UpperCamelCase__ : List[str]=False , UpperCamelCase__ : List[str]=None , **UpperCamelCase__ : Optional[int] , ): super().__init__( unk_token=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , **UpperCamelCase__ , ) A = do_lower_case with open(UpperCamelCase__ , encoding='utf-8' ) as vocab_handle: A = json.load(UpperCamelCase__ ) A = {v: k for k, v in self.encoder.items()} if merges_file is None: logger.info(f'''No merges files provided. {self.__class__.__name__} can only be used for decoding.''' ) A = None A = None else: with open(UpperCamelCase__ , encoding='utf-8' ) as merges_handle: A = merges_handle.read().split('\n' )[:-1] A = [tuple(merge.split()[:2] ) for merge in merges] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = {} @property def UpperCamelCase ( self : Union[str, Any] ): return len(self.decoder ) def UpperCamelCase ( self : Optional[Any] ): return dict(self.encoder , **self.added_tokens_encoder ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): A = tuple(token[:-1] ) + (token[-1] + BPE_TOKEN_MERGES,) if token in self.cache: return self.cache[token] A = get_pairs(UpperCamelCase__ ) if not pairs: return token while True: A = min(UpperCamelCase__ , key=lambda UpperCamelCase__ : self.bpe_ranks.get(UpperCamelCase__ , float('inf' ) ) ) if bigram not in self.bpe_ranks: break A , A = bigram A = [] A = 0 while i < len(UpperCamelCase__ ): try: A = word.index(UpperCamelCase__ , UpperCamelCase__ ) except ValueError: new_word.extend(word[i:] ) break else: new_word.extend(word[i:j] ) A = j if word[i] == first and i < len(UpperCamelCase__ ) - 1 and word[i + 1] == second: new_word.append(first + second ) i += 2 else: new_word.append(word[i] ) i += 1 A = tuple(UpperCamelCase__ ) A = new_word if len(UpperCamelCase__ ) == 1: break else: A = get_pairs(UpperCamelCase__ ) A = ' '.join(UpperCamelCase__ ) if word == "\n " + BPE_TOKEN_MERGES: A = '\n' + BPE_TOKEN_MERGES if word.endswith(UpperCamelCase__ ): A = word.replace(UpperCamelCase__ , '' ) A = word.replace(' ' , UpperCamelCase__ ) A = word return word def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict ): if self.bpe_ranks is None: raise ValueError( 'This tokenizer was instantiated without a `merges.txt` file, so' ' that it can only be used for decoding, not for encoding.' 'Make sure to provide `merges.txt` file at instantiation to enable ' 'encoding.' ) if self.do_lower_case: A = text.lower() A = text.split() A = [] for token in text: if token: split_tokens.extend(list(self.bpe(UpperCamelCase__ ).split(' ' ) ) ) return split_tokens def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : str ): return self.encoder.get(UpperCamelCase__ , self.encoder.get(self.unk_token ) ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): A = self.decoder.get(UpperCamelCase__ , self.unk_token ) return result def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] ): A = ' '.join(UpperCamelCase__ ) # make sure @@ tokens are concatenated A = ''.join(string.split(UpperCamelCase__ ) ) return string def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] ) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as f: f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCamelCase__ , ensure_ascii=UpperCamelCase__ ) + '\n' ) A = 0 if self.bpe_ranks is None: return (vocab_file,) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as writer: for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCamelCase__ : kv[1] ): if index != token_index: logger.warning( f'''Saving vocabulary to {merges_file}: BPE merge indices are not consecutive.''' ' Please check that the tokenizer is not corrupted!' ) A = token_index writer.write(' '.join(UpperCamelCase__ ) + '\n' ) index += 1 return (vocab_file, merges_file)
699
1
import random import unittest import torch from diffusers import IFInpaintingPipeline from diffusers.utils import floats_tensor from diffusers.utils.import_utils import is_xformers_available from diffusers.utils.testing_utils import skip_mps, torch_device from ..pipeline_params import ( TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS, TEXT_GUIDED_IMAGE_INPAINTING_PARAMS, ) from ..test_pipelines_common import PipelineTesterMixin from . import IFPipelineTesterMixin @skip_mps class _UpperCAmelCase ( __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[Any] = IFInpaintingPipeline SCREAMING_SNAKE_CASE : str = TEXT_GUIDED_IMAGE_INPAINTING_PARAMS - {'''width''', '''height'''} SCREAMING_SNAKE_CASE : Any = TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS SCREAMING_SNAKE_CASE : Union[str, Any] = PipelineTesterMixin.required_optional_params - {'''latents'''} def UpperCamelCase ( self : List[str] ): return self._get_dummy_components() def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict , UpperCamelCase__ : Tuple=0 ): if str(UpperCamelCase__ ).startswith('mps' ): A = torch.manual_seed(UpperCamelCase__ ) else: A = torch.Generator(device=UpperCamelCase__ ).manual_seed(UpperCamelCase__ ) A = floats_tensor((1, 3, 32, 32) , rng=random.Random(UpperCamelCase__ ) ).to(UpperCamelCase__ ) A = floats_tensor((1, 3, 32, 32) , rng=random.Random(UpperCamelCase__ ) ).to(UpperCamelCase__ ) A = { 'prompt': 'A painting of a squirrel eating a burger', 'image': image, 'mask_image': mask_image, 'generator': generator, 'num_inference_steps': 2, 'output_type': 'numpy', } return inputs @unittest.skipIf( torch_device != 'cuda' or not is_xformers_available() , reason='XFormers attention is only available with CUDA and `xformers` installed' , ) def UpperCamelCase ( self : str ): self._test_xformers_attention_forwardGenerator_pass(expected_max_diff=1e-3 ) def UpperCamelCase ( self : Tuple ): self._test_save_load_optional_components() @unittest.skipIf(torch_device != 'cuda' , reason='float16 requires CUDA' ) def UpperCamelCase ( self : int ): # Due to non-determinism in save load of the hf-internal-testing/tiny-random-t5 text encoder super().test_save_load_floataa(expected_max_diff=1e-1 ) def UpperCamelCase ( self : Optional[Any] ): self._test_attention_slicing_forward_pass(expected_max_diff=1e-2 ) def UpperCamelCase ( self : str ): self._test_save_load_local() def UpperCamelCase ( self : Dict ): self._test_inference_batch_single_identical( expected_max_diff=1e-2 , )
699
# Copyright 2023 The HuggingFace Inc. team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import torch from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer from .base import PipelineTool class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = '''facebook/bart-large-mnli''' SCREAMING_SNAKE_CASE : Union[str, Any] = ( '''This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which ''' '''should be the text to classify, and `labels`, which should be the list of labels to use for classification. ''' '''It returns the most likely label in the list of provided `labels` for the input text.''' ) SCREAMING_SNAKE_CASE : Any = '''text_classifier''' SCREAMING_SNAKE_CASE : Any = AutoTokenizer SCREAMING_SNAKE_CASE : Dict = AutoModelForSequenceClassification SCREAMING_SNAKE_CASE : List[Any] = ['''text''', ['''text''']] SCREAMING_SNAKE_CASE : Dict = ['''text'''] def UpperCamelCase ( self : List[str] ): super().setup() A = self.model.config A = -1 for idx, label in config.idalabel.items(): if label.lower().startswith('entail' ): A = int(UpperCamelCase__ ) if self.entailment_id == -1: raise ValueError('Could not determine the entailment ID from the model config, please pass it at init.' ) def UpperCamelCase ( self : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict ): A = labels return self.pre_processor( [text] * len(UpperCamelCase__ ) , [f'''This example is {label}''' for label in labels] , return_tensors='pt' , padding='max_length' , ) def UpperCamelCase ( self : int , UpperCamelCase__ : List[str] ): A = outputs.logits A = torch.argmax(logits[:, 2] ).item() return self._labels[label_id]
699
1
from __future__ import annotations import math import random from collections.abc import Collection from typing import overload class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Collection[float] | None = None ): if components is None: A = [] A = list(UpperCamelCase__ ) def __len__( self : List[Any] ): return len(self.__components ) def __str__( self : str ): return "(" + ",".join(map(UpperCamelCase__ , self.__components ) ) + ")" def __add__( self : str , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] + other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: raise Exception('must have the same size' ) def __sub__( self : Dict , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] - other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: # error case raise Exception('must have the same size' ) @overload def __mul__( self : Tuple , UpperCamelCase__ : float ): ... @overload def __mul__( self : Dict , UpperCamelCase__ : Vector ): ... def __mul__( self : Union[str, Any] , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , (float, int) ): A = [c * other for c in self.__components] return Vector(UpperCamelCase__ ) elif isinstance(UpperCamelCase__ , UpperCamelCase__ ) and len(self ) == len(UpperCamelCase__ ): A = len(self ) A = [self.__components[i] * other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return sum(UpperCamelCase__ ) else: # error case raise Exception('invalid operand!' ) def UpperCamelCase ( self : Union[str, Any] ): return Vector(self.__components ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ) and -len(self.__components ) <= i < len(self.__components ): return self.__components[i] else: raise Exception('index out of range' ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : float ): assert -len(self.__components ) <= pos < len(self.__components ) A = value def UpperCamelCase ( self : str ): if len(self.__components ) == 0: raise Exception('Vector is empty' ) A = [c**2 for c in self.__components] return math.sqrt(sum(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Vector , UpperCamelCase__ : bool = False ): A = self * other A = self.euclidean_length() * other.euclidean_length() if deg: return math.degrees(math.acos(num / den ) ) else: return math.acos(num / den ) def __UpperCamelCase (lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) return Vector([0] * dimension ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, lowerCAmelCase )) A = [0] * dimension A = 1 return Vector(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : float, lowerCAmelCase : Vector, lowerCAmelCase : Vector ) -> Vector: assert ( isinstance(lowerCAmelCase, lowerCAmelCase ) and isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, (int, float) )) ) return x * scalar + y def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: random.seed(lowerCAmelCase ) A = [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] return Vector(lowerCAmelCase ) class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : list[list[float]] , UpperCamelCase__ : int , UpperCamelCase__ : int ): A = matrix A = w A = h def __str__( self : int ): A = '' for i in range(self.__height ): ans += "|" for j in range(self.__width ): if j < self.__width - 1: ans += str(self.__matrix[i][j] ) + "," else: ans += str(self.__matrix[i][j] ) + "|\n" return ans def __add__( self : Optional[Any] , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] + other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrix must have the same dimension!' ) def __sub__( self : Dict , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] - other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrices must have the same dimension!' ) @overload def __mul__( self : int , UpperCamelCase__ : float ): ... @overload def __mul__( self : Union[str, Any] , UpperCamelCase__ : Vector ): ... def __mul__( self : Tuple , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): # matrix-vector if len(UpperCamelCase__ ) == self.__width: A = zero_vector(self.__height ) for i in range(self.__height ): A = [ self.__matrix[i][j] * other.component(UpperCamelCase__ ) for j in range(self.__width ) ] ans.change_component(UpperCamelCase__ , sum(UpperCamelCase__ ) ) return ans else: raise Exception( 'vector must have the same size as the ' 'number of columns of the matrix!' ) elif isinstance(UpperCamelCase__ , (int, float) ): # matrix-scalar A = [ [self.__matrix[i][j] * other for j in range(self.__width )] for i in range(self.__height ) ] return Matrix(UpperCamelCase__ , self.__width , self.__height ) return None def UpperCamelCase ( self : Optional[int] ): return self.__height def UpperCamelCase ( self : List[Any] ): return self.__width def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int ): if 0 <= x < self.__height and 0 <= y < self.__width: return self.__matrix[x][y] else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : float ): if 0 <= x < self.__height and 0 <= y < self.__width: A = value else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) A = self.__matrix[:x] + self.__matrix[x + 1 :] for i in range(len(UpperCamelCase__ ) ): A = minor[i][:y] + minor[i][y + 1 :] return Matrix(UpperCamelCase__ , self.__width - 1 , self.__height - 1 ).determinant() def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if 0 <= x < self.__height and 0 <= y < self.__width: return (-1) ** (x + y) * self.minor(UpperCamelCase__ , UpperCamelCase__ ) else: raise Exception('Indices out of bounds' ) def UpperCamelCase ( self : Tuple ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if self.__height < 1: raise Exception('Matrix has no element' ) elif self.__height == 1: return self.__matrix[0][0] elif self.__height == 2: return ( self.__matrix[0][0] * self.__matrix[1][1] - self.__matrix[0][1] * self.__matrix[1][0] ) else: A = [ self.__matrix[0][y] * self.cofactor(0 , UpperCamelCase__ ) for y in range(self.__width ) ] return sum(UpperCamelCase__ ) def __UpperCamelCase (lowerCAmelCase : int ) -> Matrix: A = [[0] * n for _ in range(lowerCAmelCase )] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Matrix: random.seed(lowerCAmelCase ) A = [ [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] for _ in range(lowerCAmelCase ) ] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )
699
import re import jax.numpy as jnp from flax.traverse_util import flatten_dict, unflatten_dict from jax.random import PRNGKey from ..utils import logging _UpperCAmelCase = logging.get_logger(__name__) def __UpperCamelCase (lowerCAmelCase : List[str] ) -> Dict: A = r'\w+[.]\d+' A = re.findall(lowerCAmelCase, lowerCAmelCase ) for pat in pats: A = key.replace(lowerCAmelCase, '_'.join(pat.split('.' ) ) ) return key def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Dict, lowerCAmelCase : Dict ) -> Any: A = pt_tuple_key[:-1] + ('scale',) if ( any('norm' in str_ for str_ in pt_tuple_key ) and (pt_tuple_key[-1] == "bias") and (pt_tuple_key[:-1] + ("bias",) not in random_flax_state_dict) and (pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict) ): A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor elif pt_tuple_key[-1] in ["weight", "gamma"] and pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor # embedding if pt_tuple_key[-1] == "weight" and pt_tuple_key[:-1] + ("embedding",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('embedding',) return renamed_pt_tuple_key, pt_tensor # conv layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight" and pt_tensor.ndim == 4: A = pt_tensor.transpose(2, 3, 1, 0 ) return renamed_pt_tuple_key, pt_tensor # linear layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight": A = pt_tensor.T return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm weight A = pt_tuple_key[:-1] + ('weight',) if pt_tuple_key[-1] == "gamma": return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm bias A = pt_tuple_key[:-1] + ('bias',) if pt_tuple_key[-1] == "beta": return renamed_pt_tuple_key, pt_tensor return pt_tuple_key, pt_tensor def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Any, lowerCAmelCase : str=42 ) -> Any: # Step 1: Convert pytorch tensor to numpy A = {k: v.numpy() for k, v in pt_state_dict.items()} # Step 2: Since the model is stateless, get random Flax params A = flax_model.init_weights(PRNGKey(lowerCAmelCase ) ) A = flatten_dict(lowerCAmelCase ) A = {} # Need to change some parameters name to match Flax names for pt_key, pt_tensor in pt_state_dict.items(): A = rename_key(lowerCAmelCase ) A = tuple(renamed_pt_key.split('.' ) ) # Correctly rename weight parameters A , A = rename_key_and_reshape_tensor(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if flax_key in random_flax_state_dict: if flax_tensor.shape != random_flax_state_dict[flax_key].shape: raise ValueError( f'''PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape ''' f'''{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.''' ) # also add unexpected weight so that warning is thrown A = jnp.asarray(lowerCAmelCase ) return unflatten_dict(lowerCAmelCase )
699
1
import argparse import math import os from copy import deepcopy import torch from audio_diffusion.models import DiffusionAttnUnetaD from diffusion import sampling from torch import nn from diffusers import DanceDiffusionPipeline, IPNDMScheduler, UNetaDModel _UpperCAmelCase = { "gwf-440k": { "url": "https://model-server.zqevans2.workers.dev/gwf-440k.ckpt", "sample_rate": 48_000, "sample_size": 65_536, }, "jmann-small-190k": { "url": "https://model-server.zqevans2.workers.dev/jmann-small-190k.ckpt", "sample_rate": 48_000, "sample_size": 65_536, }, "jmann-large-580k": { "url": "https://model-server.zqevans2.workers.dev/jmann-large-580k.ckpt", "sample_rate": 48_000, "sample_size": 131_072, }, "maestro-uncond-150k": { "url": "https://model-server.zqevans2.workers.dev/maestro-uncond-150k.ckpt", "sample_rate": 16_000, "sample_size": 65_536, }, "unlocked-uncond-250k": { "url": "https://model-server.zqevans2.workers.dev/unlocked-uncond-250k.ckpt", "sample_rate": 16_000, "sample_size": 65_536, }, "honk-140k": { "url": "https://model-server.zqevans2.workers.dev/honk-140k.ckpt", "sample_rate": 16_000, "sample_size": 65_536, }, } def __UpperCamelCase (lowerCAmelCase : Any, lowerCAmelCase : Union[str, Any] ) -> List[str]: return torch.atana(lowerCAmelCase, lowerCAmelCase ) / math.pi * 2 def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> Optional[Any]: A = torch.sin(t * math.pi / 2 ) ** 2 A = (1 - sigma**2) ** 0.5 return alpha_sigma_to_t(lowerCAmelCase, lowerCAmelCase ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' pass class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Optional[Any] ): super().__init__() A = DiffusionAttnUnetaD(UpperCamelCase__ , n_attn_layers=4 ) A = deepcopy(self.diffusion ) A = torch.quasirandom.SobolEngine(1 , scramble=UpperCamelCase__ ) def __UpperCamelCase (lowerCAmelCase : int ) -> Union[str, Any]: A = MODELS_MAP[model_name]['url'] os.system(f'''wget {url} ./''' ) return f'''./{model_name}.ckpt''' _UpperCAmelCase = { "1": "resnets.0", "2": "attentions.0", "3": "resnets.1", "4": "attentions.1", "5": "resnets.2", "6": "attentions.2", } _UpperCAmelCase = { "8": "resnets.0", "9": "attentions.0", "10": "resnets.1", "11": "attentions.1", "12": "resnets.2", "13": "attentions.2", } _UpperCAmelCase = { "1": "resnets.0", "2": "attentions.0", "3": "resnets.1", "4": "attentions.1", "5": "resnets.2", "6": "attentions.2", "8": "resnets.3", "9": "attentions.3", "10": "resnets.4", "11": "attentions.4", "12": "resnets.5", "13": "attentions.5", } _UpperCAmelCase = { "0": "resnets.0", "1": "resnets.1", "2": "resnets.2", "4": "resnets.0", "5": "resnets.1", "6": "resnets.2", } _UpperCAmelCase = { "skip": "conv_skip", "main.0": "conv_1", "main.1": "group_norm_1", "main.3": "conv_2", "main.4": "group_norm_2", } _UpperCAmelCase = { "norm": "group_norm", "qkv_proj": ["query", "key", "value"], "out_proj": ["proj_attn"], } def __UpperCamelCase (lowerCAmelCase : Dict ) -> str: if name.startswith('skip' ): return name.replace('skip', RES_CONV_MAP['skip'] ) # name has to be of format main.{digit} if not name.startswith('main.' ): raise ValueError(f'''ResConvBlock error with {name}''' ) return name.replace(name[:6], RES_CONV_MAP[name[:6]] ) def __UpperCamelCase (lowerCAmelCase : Dict ) -> Union[str, Any]: for key, value in ATTN_MAP.items(): if name.startswith(lowerCAmelCase ) and not isinstance(lowerCAmelCase, lowerCAmelCase ): return name.replace(lowerCAmelCase, lowerCAmelCase ) elif name.startswith(lowerCAmelCase ): return [name.replace(lowerCAmelCase, lowerCAmelCase ) for v in value] raise ValueError(f'''Attn error with {name}''' ) def __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Union[str, Any]=13 ) -> List[Any]: A = input_string if string.split('.' )[0] == "timestep_embed": return string.replace('timestep_embed', 'time_proj' ) A = 0 if string.startswith('net.3.' ): depth += 1 A = string[6:] elif string.startswith('net.' ): A = string[4:] while string.startswith('main.7.' ): depth += 1 A = string[7:] if string.startswith('main.' ): A = string[5:] # mid block if string[:2].isdigit(): A = string[:2] A = string[2:] else: A = string[0] A = string[1:] if depth == max_depth: A = MID_NUM_TO_LAYER[layer_num] A = 'mid_block' elif depth > 0 and int(lowerCAmelCase ) < 7: A = DOWN_NUM_TO_LAYER[layer_num] A = f'''down_blocks.{depth}''' elif depth > 0 and int(lowerCAmelCase ) > 7: A = UP_NUM_TO_LAYER[layer_num] A = f'''up_blocks.{max_depth - depth - 1}''' elif depth == 0: A = DEPTH_0_TO_LAYER[layer_num] A = f'''up_blocks.{max_depth - 1}''' if int(lowerCAmelCase ) > 3 else 'down_blocks.0' if not string_left.startswith('.' ): raise ValueError(f'''Naming error with {input_string} and string_left: {string_left}.''' ) A = string_left[1:] if "resnets" in new_layer: A = convert_resconv_naming(lowerCAmelCase ) elif "attentions" in new_layer: A = convert_attn_naming(lowerCAmelCase ) A = new_string_left if not isinstance(lowerCAmelCase, lowerCAmelCase ): A = prefix + '.' + new_layer + '.' + string_left else: A = [prefix + '.' + new_layer + '.' + s for s in string_left] return new_string def __UpperCamelCase (lowerCAmelCase : List[Any] ) -> Dict: A = {} for k, v in state_dict.items(): if k.endswith('kernel' ): # up- and downsample layers, don't have trainable weights continue A = rename(lowerCAmelCase ) # check if we need to transform from Conv => Linear for attention if isinstance(lowerCAmelCase, lowerCAmelCase ): A = transform_conv_attns(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) else: A = v return new_state_dict def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Dict, lowerCAmelCase : List[str] ) -> List[Any]: if len(lowerCAmelCase ) == 1: if len(v.shape ) == 3: # weight A = v[:, :, 0] else: # bias A = v else: # qkv matrices A = v.shape[0] A = trippled_shape // 3 for i in range(3 ): if len(v.shape ) == 3: A = v[i * single_shape : (i + 1) * single_shape, :, 0] else: A = v[i * single_shape : (i + 1) * single_shape] return new_state_dict def __UpperCamelCase (lowerCAmelCase : Any ) -> str: A = torch.device('cuda' if torch.cuda.is_available() else 'cpu' ) A = args.model_path.split('/' )[-1].split('.' )[0] if not os.path.isfile(args.model_path ): assert ( model_name == args.model_path ), f'''Make sure to provide one of the official model names {MODELS_MAP.keys()}''' A = download(lowerCAmelCase ) A = MODELS_MAP[model_name]['sample_rate'] A = MODELS_MAP[model_name]['sample_size'] A = Object() A = sample_size A = sample_rate A = 0 A = UNetaDModel(sample_size=lowerCAmelCase, sample_rate=lowerCAmelCase ) A = diffusers_model.state_dict() A = DiffusionUncond(lowerCAmelCase ) orig_model.load_state_dict(torch.load(args.model_path, map_location=lowerCAmelCase )['state_dict'] ) A = orig_model.diffusion_ema.eval() A = orig_model.state_dict() A = rename_orig_weights(lowerCAmelCase ) A = set(renamed_state_dict.keys() ) - set(diffusers_state_dict.keys() ) A = set(diffusers_state_dict.keys() ) - set(renamed_state_dict.keys() ) assert len(lowerCAmelCase ) == 0, f'''Problem with {renamed_minus_diffusers}''' assert all(k.endswith('kernel' ) for k in list(lowerCAmelCase ) ), f'''Problem with {diffusers_minus_renamed}''' for key, value in renamed_state_dict.items(): assert ( diffusers_state_dict[key].squeeze().shape == value.squeeze().shape ), f'''Shape for {key} doesn\'t match. Diffusers: {diffusers_state_dict[key].shape} vs. {value.shape}''' if key == "time_proj.weight": A = value.squeeze() A = value diffusers_model.load_state_dict(lowerCAmelCase ) A = 100 A = 33 A = IPNDMScheduler(num_train_timesteps=lowerCAmelCase ) A = torch.manual_seed(lowerCAmelCase ) A = torch.randn([1, 2, config.sample_size], generator=lowerCAmelCase ).to(lowerCAmelCase ) A = torch.linspace(1, 0, steps + 1, device=lowerCAmelCase )[:-1] A = get_crash_schedule(lowerCAmelCase ) A = DanceDiffusionPipeline(unet=lowerCAmelCase, scheduler=lowerCAmelCase ) A = torch.manual_seed(33 ) A = pipe(num_inference_steps=lowerCAmelCase, generator=lowerCAmelCase ).audios A = sampling.iplms_sample(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, {} ) A = generated.clamp(-1, 1 ) A = (generated - audio).abs().sum() A = (generated - audio).abs().max() if args.save: pipe.save_pretrained(args.checkpoint_path ) print('Diff sum', lowerCAmelCase ) print('Diff max', lowerCAmelCase ) assert diff_max < 1E-3, f'''Diff max: {diff_max} is too much :-/''' print(f'''Conversion for {model_name} successful!''' ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--model_path", default=None, type=str, required=True, help="Path to the model to convert.") parser.add_argument( "--save", default=True, type=bool, required=False, help="Whether to save the converted model or not." ) parser.add_argument("--checkpoint_path", default=None, type=str, required=True, help="Path to the output model.") _UpperCAmelCase = parser.parse_args() main(args)
699
from __future__ import annotations import math import random from collections.abc import Collection from typing import overload class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Collection[float] | None = None ): if components is None: A = [] A = list(UpperCamelCase__ ) def __len__( self : List[Any] ): return len(self.__components ) def __str__( self : str ): return "(" + ",".join(map(UpperCamelCase__ , self.__components ) ) + ")" def __add__( self : str , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] + other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: raise Exception('must have the same size' ) def __sub__( self : Dict , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] - other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: # error case raise Exception('must have the same size' ) @overload def __mul__( self : Tuple , UpperCamelCase__ : float ): ... @overload def __mul__( self : Dict , UpperCamelCase__ : Vector ): ... def __mul__( self : Union[str, Any] , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , (float, int) ): A = [c * other for c in self.__components] return Vector(UpperCamelCase__ ) elif isinstance(UpperCamelCase__ , UpperCamelCase__ ) and len(self ) == len(UpperCamelCase__ ): A = len(self ) A = [self.__components[i] * other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return sum(UpperCamelCase__ ) else: # error case raise Exception('invalid operand!' ) def UpperCamelCase ( self : Union[str, Any] ): return Vector(self.__components ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ) and -len(self.__components ) <= i < len(self.__components ): return self.__components[i] else: raise Exception('index out of range' ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : float ): assert -len(self.__components ) <= pos < len(self.__components ) A = value def UpperCamelCase ( self : str ): if len(self.__components ) == 0: raise Exception('Vector is empty' ) A = [c**2 for c in self.__components] return math.sqrt(sum(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Vector , UpperCamelCase__ : bool = False ): A = self * other A = self.euclidean_length() * other.euclidean_length() if deg: return math.degrees(math.acos(num / den ) ) else: return math.acos(num / den ) def __UpperCamelCase (lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) return Vector([0] * dimension ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, lowerCAmelCase )) A = [0] * dimension A = 1 return Vector(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : float, lowerCAmelCase : Vector, lowerCAmelCase : Vector ) -> Vector: assert ( isinstance(lowerCAmelCase, lowerCAmelCase ) and isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, (int, float) )) ) return x * scalar + y def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: random.seed(lowerCAmelCase ) A = [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] return Vector(lowerCAmelCase ) class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : list[list[float]] , UpperCamelCase__ : int , UpperCamelCase__ : int ): A = matrix A = w A = h def __str__( self : int ): A = '' for i in range(self.__height ): ans += "|" for j in range(self.__width ): if j < self.__width - 1: ans += str(self.__matrix[i][j] ) + "," else: ans += str(self.__matrix[i][j] ) + "|\n" return ans def __add__( self : Optional[Any] , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] + other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrix must have the same dimension!' ) def __sub__( self : Dict , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] - other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrices must have the same dimension!' ) @overload def __mul__( self : int , UpperCamelCase__ : float ): ... @overload def __mul__( self : Union[str, Any] , UpperCamelCase__ : Vector ): ... def __mul__( self : Tuple , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): # matrix-vector if len(UpperCamelCase__ ) == self.__width: A = zero_vector(self.__height ) for i in range(self.__height ): A = [ self.__matrix[i][j] * other.component(UpperCamelCase__ ) for j in range(self.__width ) ] ans.change_component(UpperCamelCase__ , sum(UpperCamelCase__ ) ) return ans else: raise Exception( 'vector must have the same size as the ' 'number of columns of the matrix!' ) elif isinstance(UpperCamelCase__ , (int, float) ): # matrix-scalar A = [ [self.__matrix[i][j] * other for j in range(self.__width )] for i in range(self.__height ) ] return Matrix(UpperCamelCase__ , self.__width , self.__height ) return None def UpperCamelCase ( self : Optional[int] ): return self.__height def UpperCamelCase ( self : List[Any] ): return self.__width def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int ): if 0 <= x < self.__height and 0 <= y < self.__width: return self.__matrix[x][y] else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : float ): if 0 <= x < self.__height and 0 <= y < self.__width: A = value else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) A = self.__matrix[:x] + self.__matrix[x + 1 :] for i in range(len(UpperCamelCase__ ) ): A = minor[i][:y] + minor[i][y + 1 :] return Matrix(UpperCamelCase__ , self.__width - 1 , self.__height - 1 ).determinant() def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if 0 <= x < self.__height and 0 <= y < self.__width: return (-1) ** (x + y) * self.minor(UpperCamelCase__ , UpperCamelCase__ ) else: raise Exception('Indices out of bounds' ) def UpperCamelCase ( self : Tuple ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if self.__height < 1: raise Exception('Matrix has no element' ) elif self.__height == 1: return self.__matrix[0][0] elif self.__height == 2: return ( self.__matrix[0][0] * self.__matrix[1][1] - self.__matrix[0][1] * self.__matrix[1][0] ) else: A = [ self.__matrix[0][y] * self.cofactor(0 , UpperCamelCase__ ) for y in range(self.__width ) ] return sum(UpperCamelCase__ ) def __UpperCamelCase (lowerCAmelCase : int ) -> Matrix: A = [[0] * n for _ in range(lowerCAmelCase )] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Matrix: random.seed(lowerCAmelCase ) A = [ [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] for _ in range(lowerCAmelCase ) ] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )
699
1
from __future__ import annotations from random import random class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : int | None = None ): A = value A = random() A = None A = None def __repr__( self : List[Any] ): from pprint import pformat if self.left is None and self.right is None: return f'''\'{self.value}: {self.prior:.5}\'''' else: return pformat( {f'''{self.value}: {self.prior:.5}''': (self.left, self.right)} , indent=1 ) def __str__( self : List[str] ): A = str(self.value ) + ' ' A = str(self.left or '' ) A = str(self.right or '' ) return value + left + right def __UpperCamelCase (lowerCAmelCase : Node | None, lowerCAmelCase : int ) -> tuple[Node | None, Node | None]: if root is None: # None tree is split into 2 Nones return None, None elif root.value is None: return None, None else: if value < root.value: A , A = split(root.left, lowerCAmelCase ) return left, root else: A , A = split(root.right, lowerCAmelCase ) return root, right def __UpperCamelCase (lowerCAmelCase : Node | None, lowerCAmelCase : Node | None ) -> Node | None: if (not left) or (not right): # If one node is None, return the other return left or right elif left.prior < right.prior: A = merge(left.right, lowerCAmelCase ) return left else: A = merge(lowerCAmelCase, right.left ) return right def __UpperCamelCase (lowerCAmelCase : Node | None, lowerCAmelCase : int ) -> Node | None: A = Node(lowerCAmelCase ) A , A = split(lowerCAmelCase, lowerCAmelCase ) return merge(merge(lowerCAmelCase, lowerCAmelCase ), lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : Node | None, lowerCAmelCase : int ) -> Node | None: A , A = split(lowerCAmelCase, value - 1 ) A , A = split(lowerCAmelCase, lowerCAmelCase ) return merge(lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : Node | None ) -> None: if not root: # None return else: inorder(root.left ) print(root.value, end=',' ) inorder(root.right ) def __UpperCamelCase (lowerCAmelCase : Node | None, lowerCAmelCase : str ) -> Node | None: for arg in args.split(): if arg[0] == "+": A = insert(lowerCAmelCase, int(arg[1:] ) ) elif arg[0] == "-": A = erase(lowerCAmelCase, int(arg[1:] ) ) else: print('Unknown command' ) return root def __UpperCamelCase () -> None: A = None print( 'enter numbers to create a tree, + value to add value into treap, ' '- value to erase all nodes with value. \'q\' to quit. ' ) A = input() while args != "q": A = interact_treap(lowerCAmelCase, lowerCAmelCase ) print(lowerCAmelCase ) A = input() print('good by!' ) if __name__ == "__main__": import doctest doctest.testmod() main()
699
from collections import OrderedDict from typing import Any, Mapping, Optional from ... import PreTrainedTokenizer from ...configuration_utils import PretrainedConfig from ...file_utils import TensorType, is_torch_available from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast from ...onnx.utils import compute_effective_axis_dimension from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "facebook/blenderbot_small-90M": "https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/config.json", # See all BlenderbotSmall models at https://huggingface.co/models?filter=blenderbot_small } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = '''blenderbot-small''' SCREAMING_SNAKE_CASE : Any = ['''past_key_values'''] SCREAMING_SNAKE_CASE : List[str] = {'''num_attention_heads''': '''encoder_attention_heads''', '''hidden_size''': '''d_model'''} def __init__( self : List[str] , UpperCamelCase__ : Optional[Any]=50265 , UpperCamelCase__ : Optional[int]=512 , UpperCamelCase__ : int=8 , UpperCamelCase__ : Optional[int]=2048 , UpperCamelCase__ : Optional[Any]=16 , UpperCamelCase__ : Optional[Any]=8 , UpperCamelCase__ : List[Any]=2048 , UpperCamelCase__ : int=16 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : int=True , UpperCamelCase__ : Optional[Any]="gelu" , UpperCamelCase__ : Any=512 , UpperCamelCase__ : Union[str, Any]=0.1 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Union[str, Any]=0.0 , UpperCamelCase__ : Dict=0.02 , UpperCamelCase__ : Optional[Any]=1 , UpperCamelCase__ : Any=False , UpperCamelCase__ : Dict=0 , UpperCamelCase__ : Optional[int]=1 , UpperCamelCase__ : List[str]=2 , UpperCamelCase__ : Dict=2 , **UpperCamelCase__ : List[str] , ): A = vocab_size A = max_position_embeddings A = d_model A = encoder_ffn_dim A = encoder_layers A = encoder_attention_heads A = decoder_ffn_dim A = decoder_layers A = decoder_attention_heads A = dropout A = attention_dropout A = activation_dropout A = activation_function A = init_std A = encoder_layerdrop A = decoder_layerdrop A = use_cache A = encoder_layers A = scale_embedding # scale factor will be sqrt(d_model) if True super().__init__( pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , is_encoder_decoder=UpperCamelCase__ , decoder_start_token_id=UpperCamelCase__ , forced_eos_token_id=UpperCamelCase__ , **UpperCamelCase__ , ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' @property def UpperCamelCase ( self : List[Any] ): if self.task in ["default", "seq2seq-lm"]: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A = {0: 'batch'} A = {0: 'batch', 1: 'past_decoder_sequence + sequence'} else: A = {0: 'batch', 1: 'decoder_sequence'} A = {0: 'batch', 1: 'decoder_sequence'} if self.use_past: self.fill_with_past_key_values_(UpperCamelCase__ , direction='inputs' ) elif self.task == "causal-lm": # TODO: figure this case out. A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} else: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ('decoder_input_ids', {0: 'batch', 1: 'decoder_sequence'}), ('decoder_attention_mask', {0: 'batch', 1: 'decoder_sequence'}), ] ) return common_inputs @property def UpperCamelCase ( self : int ): if self.task in ["default", "seq2seq-lm"]: A = super().outputs else: A = super(UpperCamelCase__ , self ).outputs if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} return common_outputs def UpperCamelCase ( self : int , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) # Generate decoder inputs A = seq_length if not self.use_past else 1 A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) A = {f'''decoder_{name}''': tensor for name, tensor in decoder_inputs.items()} A = dict(**UpperCamelCase__ , **UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape A = common_inputs['decoder_input_ids'].shape[1] A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, encoder_seq_length, self._config.hidden_size // num_encoder_attention_heads, ) A = decoder_seq_length + 3 A = ( batch, num_decoder_attention_heads, decoder_past_length, self._config.hidden_size // num_decoder_attention_heads, ) A = torch.cat( [common_inputs['decoder_attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ )] , dim=1 ) A = [] # If the number of encoder and decoder layers are present in the model configuration, both are considered A , A = self.num_layers A = min(UpperCamelCase__ , UpperCamelCase__ ) A = max(UpperCamelCase__ , UpperCamelCase__ ) - min_num_layers A = 'encoder' if num_encoder_layers > num_decoder_layers else 'decoder' for _ in range(UpperCamelCase__ ): common_inputs["past_key_values"].append( ( torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), ) ) # TODO: test this. A = encoder_shape if remaining_side_name == 'encoder' else decoder_shape for _ in range(UpperCamelCase__ , UpperCamelCase__ ): common_inputs["past_key_values"].append((torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) ) return common_inputs def UpperCamelCase ( self : Tuple , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape # Not using the same length for past_key_values A = seqlen + 2 A , A = self.num_layers A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, past_key_values_length, self._config.hidden_size // num_encoder_attention_heads, ) A = common_inputs['attention_mask'].dtype A = torch.cat( [common_inputs['attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ , dtype=UpperCamelCase__ )] , dim=1 ) A = [ (torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) for _ in range(UpperCamelCase__ ) ] return common_inputs def UpperCamelCase ( self : List[str] , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): # Copied from OnnxConfig.generate_dummy_inputs # Did not use super(OnnxConfigWithPast, self).generate_dummy_inputs for code clarity. # If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 ) # If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX A = tokenizer.num_special_tokens_to_add(UpperCamelCase__ ) A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=UpperCamelCase__ ) # Generate dummy inputs according to compute batch and sequence A = [' '.join([tokenizer.unk_token] ) * seq_length] * batch_size A = dict(tokenizer(UpperCamelCase__ , return_tensors=UpperCamelCase__ ) ) return common_inputs def UpperCamelCase ( self : Any , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): if self.task in ["default", "seq2seq-lm"]: A = self._generate_dummy_inputs_for_default_and_seqaseq_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) elif self.task == "causal-lm": A = self._generate_dummy_inputs_for_causal_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) else: A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) return common_inputs def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : List[str] , UpperCamelCase__ : str , UpperCamelCase__ : Tuple ): if self.task in ["default", "seq2seq-lm"]: A = super()._flatten_past_key_values_(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) else: A = super(UpperCamelCase__ , self )._flatten_past_key_values_( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
699
1
import numpy as np import torch import torch.nn as nn from transformers import CLIPConfig, CLIPVisionModelWithProjection, PreTrainedModel from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[Any] = CLIPConfig SCREAMING_SNAKE_CASE : Any = ['''CLIPEncoderLayer'''] def __init__( self : str , UpperCamelCase__ : CLIPConfig ): super().__init__(UpperCamelCase__ ) A = CLIPVisionModelWithProjection(config.vision_config ) A = nn.Linear(config.vision_config.projection_dim , 1 ) A = nn.Linear(config.vision_config.projection_dim , 1 ) @torch.no_grad() def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[Any] , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int]=0.5 , UpperCamelCase__ : Union[str, Any]=0.5 ): A = self.vision_model(UpperCamelCase__ )[0] A = self.p_head(UpperCamelCase__ ) A = nsfw_detected.flatten() A = nsfw_detected > p_threshold A = nsfw_detected.tolist() if any(UpperCamelCase__ ): logger.warning( 'Potential NSFW content was detected in one or more images. A black image will be returned instead.' ' Try again with a different prompt and/or seed.' ) for idx, nsfw_detected_ in enumerate(UpperCamelCase__ ): if nsfw_detected_: A = np.zeros(images[idx].shape ) A = self.w_head(UpperCamelCase__ ) A = watermark_detected.flatten() A = watermark_detected > w_threshold A = watermark_detected.tolist() if any(UpperCamelCase__ ): logger.warning( 'Potential watermarked content was detected in one or more images. A black image will be returned instead.' ' Try again with a different prompt and/or seed.' ) for idx, watermark_detected_ in enumerate(UpperCamelCase__ ): if watermark_detected_: A = np.zeros(images[idx].shape ) return images, nsfw_detected, watermark_detected
699
from typing import List, Optional, Union from ...processing_utils import ProcessorMixin from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy from ...utils import TensorType class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = ['''image_processor''', '''tokenizer'''] SCREAMING_SNAKE_CASE : List[str] = '''BridgeTowerImageProcessor''' SCREAMING_SNAKE_CASE : Tuple = ('''RobertaTokenizer''', '''RobertaTokenizerFast''') def __init__( self : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int] ): super().__init__(UpperCamelCase__ , UpperCamelCase__ ) def __call__( self : Optional[int] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None , UpperCamelCase__ : bool = True , UpperCamelCase__ : Union[bool, str, PaddingStrategy] = False , UpperCamelCase__ : Union[bool, str, TruncationStrategy] = None , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : Optional[Union[str, TensorType]] = None , **UpperCamelCase__ : List[Any] , ): A = self.tokenizer( text=UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , padding=UpperCamelCase__ , truncation=UpperCamelCase__ , max_length=UpperCamelCase__ , stride=UpperCamelCase__ , pad_to_multiple_of=UpperCamelCase__ , return_token_type_ids=UpperCamelCase__ , return_attention_mask=UpperCamelCase__ , return_overflowing_tokens=UpperCamelCase__ , return_special_tokens_mask=UpperCamelCase__ , return_offsets_mapping=UpperCamelCase__ , return_length=UpperCamelCase__ , verbose=UpperCamelCase__ , return_tensors=UpperCamelCase__ , **UpperCamelCase__ , ) # add pixel_values + pixel_mask A = self.image_processor( UpperCamelCase__ , return_tensors=UpperCamelCase__ , do_normalize=UpperCamelCase__ , do_center_crop=UpperCamelCase__ , **UpperCamelCase__ ) encoding.update(UpperCamelCase__ ) return encoding def UpperCamelCase ( self : Dict , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : Any ): return self.tokenizer.batch_decode(*UpperCamelCase__ , **UpperCamelCase__ ) def UpperCamelCase ( self : int , *UpperCamelCase__ : int , **UpperCamelCase__ : List[str] ): return self.tokenizer.decode(*UpperCamelCase__ , **UpperCamelCase__ ) @property def UpperCamelCase ( self : Any ): A = self.tokenizer.model_input_names A = self.image_processor.model_input_names return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
699
1
# Copyright 2023 The HuggingFace Team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available _UpperCAmelCase = {"configuration_timm_backbone": ["TimmBackboneConfig"]} try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["TimmBackbone"] if TYPE_CHECKING: from .configuration_timm_backbone import TimmBackboneConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_timm_backbone import TimmBackbone else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> str: 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=10))
699
1
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, prepare_image_inputs if is_torch_available(): import torch if is_vision_available(): from PIL import Image from transformers import MobileViTImageProcessor class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[int]=7 , UpperCamelCase__ : Dict=3 , UpperCamelCase__ : Union[str, Any]=18 , UpperCamelCase__ : Optional[Any]=30 , UpperCamelCase__ : Optional[int]=400 , UpperCamelCase__ : Optional[Any]=True , UpperCamelCase__ : Union[str, Any]=None , UpperCamelCase__ : str=True , UpperCamelCase__ : Any=None , UpperCamelCase__ : int=True , ): A = size if size is not None else {'shortest_edge': 20} A = crop_size if crop_size is not None else {'height': 18, 'width': 18} A = parent A = batch_size A = num_channels A = image_size A = min_resolution A = max_resolution A = do_resize A = size A = do_center_crop A = crop_size A = do_flip_channel_order def UpperCamelCase ( self : Optional[int] ): return { "do_resize": self.do_resize, "size": self.size, "do_center_crop": self.do_center_crop, "crop_size": self.crop_size, "do_flip_channel_order": self.do_flip_channel_order, } @require_torch @require_vision class _UpperCAmelCase ( __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = MobileViTImageProcessor if is_vision_available() else None def UpperCamelCase ( self : Any ): A = MobileViTImageProcessingTester(self ) @property def UpperCamelCase ( self : Tuple ): return self.image_processor_tester.prepare_image_processor_dict() def UpperCamelCase ( self : str ): A = self.image_processing_class(**self.image_processor_dict ) self.assertTrue(hasattr(UpperCamelCase__ , 'do_resize' ) ) self.assertTrue(hasattr(UpperCamelCase__ , 'size' ) ) self.assertTrue(hasattr(UpperCamelCase__ , 'do_center_crop' ) ) self.assertTrue(hasattr(UpperCamelCase__ , 'center_crop' ) ) self.assertTrue(hasattr(UpperCamelCase__ , 'do_flip_channel_order' ) ) def UpperCamelCase ( self : Optional[int] ): A = self.image_processing_class.from_dict(self.image_processor_dict ) self.assertEqual(image_processor.size , {'shortest_edge': 20} ) self.assertEqual(image_processor.crop_size , {'height': 18, 'width': 18} ) A = 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 UpperCamelCase ( self : Tuple ): pass def UpperCamelCase ( self : Any ): # Initialize image_processing A = self.image_processing_class(**self.image_processor_dict ) # create random PIL images A = prepare_image_inputs(self.image_processor_tester , equal_resolution=UpperCamelCase__ ) for image in image_inputs: self.assertIsInstance(UpperCamelCase__ , Image.Image ) # Test not batched input A = 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 = image_processing(UpperCamelCase__ , 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 UpperCamelCase ( self : List[str] ): # Initialize image_processing A = self.image_processing_class(**self.image_processor_dict ) # create random numpy tensors A = prepare_image_inputs(self.image_processor_tester , equal_resolution=UpperCamelCase__ , numpify=UpperCamelCase__ ) for image in image_inputs: self.assertIsInstance(UpperCamelCase__ , np.ndarray ) # Test not batched input A = 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 = image_processing(UpperCamelCase__ , 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 UpperCamelCase ( self : List[Any] ): # Initialize image_processing A = self.image_processing_class(**self.image_processor_dict ) # create random PyTorch tensors A = prepare_image_inputs(self.image_processor_tester , equal_resolution=UpperCamelCase__ , torchify=UpperCamelCase__ ) for image in image_inputs: self.assertIsInstance(UpperCamelCase__ , torch.Tensor ) # Test not batched input A = 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 = image_processing(UpperCamelCase__ , 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'], ) , )
699
import math import torch from torch import nn from ..configuration_utils import ConfigMixin, register_to_config from .attention_processor import Attention from .embeddings import get_timestep_embedding from .modeling_utils import ModelMixin class _UpperCAmelCase ( __lowercase , __lowercase ): '''simple docstring''' @register_to_config def __init__( self : Any , UpperCamelCase__ : int = 128 , UpperCamelCase__ : int = 256 , UpperCamelCase__ : float = 2_000.0 , UpperCamelCase__ : int = 768 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 64 , UpperCamelCase__ : int = 2048 , UpperCamelCase__ : float = 0.1 , ): super().__init__() A = nn.Sequential( nn.Linear(UpperCamelCase__ , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , nn.Linear(d_model * 4 , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , ) A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = False A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.ModuleList() for lyr_num in range(UpperCamelCase__ ): # FiLM conditional T5 decoder A = DecoderLayer(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) self.decoders.append(UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Any , UpperCamelCase__ : int ): A = torch.mul(query_input.unsqueeze(-1 ) , key_input.unsqueeze(-2 ) ) return mask.unsqueeze(-3 ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[int] ): A , A , A = decoder_input_tokens.shape assert decoder_noise_time.shape == (batch,) # decoder_noise_time is in [0, 1), so rescale to expected timing range. A = get_timestep_embedding( decoder_noise_time * self.config.max_decoder_noise_time , embedding_dim=self.config.d_model , max_period=self.config.max_decoder_noise_time , ).to(dtype=self.dtype ) A = self.conditioning_emb(UpperCamelCase__ ).unsqueeze(1 ) assert conditioning_emb.shape == (batch, 1, self.config.d_model * 4) A = decoder_input_tokens.shape[1] # If we want to use relative positions for audio context, we can just offset # this sequence by the length of encodings_and_masks. A = torch.broadcast_to( torch.arange(UpperCamelCase__ , device=decoder_input_tokens.device ) , (batch, seq_length) , ) A = self.position_encoding(UpperCamelCase__ ) A = self.continuous_inputs_projection(UpperCamelCase__ ) inputs += position_encodings A = self.dropout(UpperCamelCase__ ) # decoder: No padding present. A = torch.ones( decoder_input_tokens.shape[:2] , device=decoder_input_tokens.device , dtype=inputs.dtype ) # Translate encoding masks to encoder-decoder masks. A = [(x, self.encoder_decoder_mask(UpperCamelCase__ , UpperCamelCase__ )) for x, y in encodings_and_masks] # cross attend style: concat encodings A = torch.cat([x[0] for x in encodings_and_encdec_masks] , dim=1 ) A = torch.cat([x[1] for x in encodings_and_encdec_masks] , dim=-1 ) for lyr in self.decoders: A = lyr( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , )[0] A = self.decoder_norm(UpperCamelCase__ ) A = self.post_dropout(UpperCamelCase__ ) A = self.spec_out(UpperCamelCase__ ) return spec_out class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Union[str, Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Union[str, Any]=1e-6 ): super().__init__() A = nn.ModuleList() # cond self attention: layer 0 self.layer.append( TaLayerSelfAttentionCond(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) ) # cross attention: layer 1 self.layer.append( TaLayerCrossAttention( d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ , ) ) # Film Cond MLP + dropout: last layer self.layer.append( TaLayerFFCond(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ ) ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : int=None , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[Any]=None , ): A = self.layer[0]( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) if encoder_hidden_states is not None: A = torch.where(encoder_attention_mask > 0 , 0 , -1e1_0 ).to( encoder_hidden_states.dtype ) A = self.layer[1]( UpperCamelCase__ , key_value_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) # Apply Film Conditional Feed Forward layer A = self.layer[-1](UpperCamelCase__ , UpperCamelCase__ ) return (hidden_states,) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Dict ): super().__init__() A = TaLayerNorm(UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : str , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Tuple=None , ): # pre_self_attention_layer_norm A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.FiLMLayer(UpperCamelCase__ , UpperCamelCase__ ) # Self-attention block A = self.attention(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[Any] ): super().__init__() A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[str]=None , ): A = self.layer_norm(UpperCamelCase__ ) A = self.attention( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=attention_mask.squeeze(1 ) , ) A = hidden_states + self.dropout(UpperCamelCase__ ) return layer_output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any ): super().__init__() A = TaDenseGatedActDense(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any=None ): A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.film(UpperCamelCase__ , UpperCamelCase__ ) A = self.DenseReluDense(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) A = NewGELUActivation() def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : List[Any] ): A = self.act(self.wi_a(UpperCamelCase__ ) ) A = self.wi_a(UpperCamelCase__ ) A = hidden_gelu * hidden_linear A = self.dropout(UpperCamelCase__ ) A = self.wo(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Tuple=1e-6 ): super().__init__() A = nn.Parameter(torch.ones(UpperCamelCase__ ) ) A = eps def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int ): # T5 uses a layer_norm which only scales and doesn't shift, which is also known as Root Mean # Square Layer Normalization https://arxiv.org/abs/1910.07467 thus variance is calculated # w/o mean and there is no bias. Additionally we want to make sure that the accumulation for # half-precision inputs is done in fp32 A = hidden_states.to(torch.floataa ).pow(2 ).mean(-1 , keepdim=UpperCamelCase__ ) A = hidden_states * torch.rsqrt(variance + self.variance_epsilon ) # convert into half-precision if necessary if self.weight.dtype in [torch.floataa, torch.bfloataa]: A = hidden_states.to(self.weight.dtype ) return self.weight * hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def UpperCamelCase ( self : Any , UpperCamelCase__ : torch.Tensor ): return 0.5 * input * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi ) * (input + 0.044_715 * torch.pow(UpperCamelCase__ , 3.0 )) )) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , out_features * 2 , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : List[Any] ): A = self.scale_bias(UpperCamelCase__ ) A , A = torch.chunk(UpperCamelCase__ , 2 , -1 ) A = x * (1 + scale) + shift return x
699
1
from typing import List, Optional, Tuple, Union import torch from torch import nn from torch.nn import CrossEntropyLoss from ... import AutoBackbone from ...modeling_outputs import SemanticSegmenterOutput from ...modeling_utils import PreTrainedModel from ...utils import add_start_docstrings, add_start_docstrings_to_model_forward, replace_return_docstrings from ...utils.backbone_utils import BackboneMixin from .configuration_upernet import UperNetConfig _UpperCAmelCase = [ "openmmlab/upernet-convnext-tiny", # See all UperNet models at https://huggingface.co/models?filter=upernet ] # General docstring _UpperCAmelCase = "UperNetConfig" class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Union[int, Tuple[int, int]] , UpperCamelCase__ : Union[int, Tuple[int, int], str] = 0 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Union[int, Tuple[int, int]] = 1 , ): super().__init__() A = nn.Convad( in_channels=UpperCamelCase__ , out_channels=UpperCamelCase__ , kernel_size=UpperCamelCase__ , padding=UpperCamelCase__ , bias=UpperCamelCase__ , dilation=UpperCamelCase__ , ) A = nn.BatchNormad(UpperCamelCase__ ) A = nn.ReLU() def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : torch.Tensor ): A = self.conv(UpperCamelCase__ ) A = self.batch_norm(UpperCamelCase__ ) A = self.activation(UpperCamelCase__ ) return output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = [ nn.AdaptiveAvgPoolad(UpperCamelCase__ ), UperNetConvModule(UpperCamelCase__ , UpperCamelCase__ , kernel_size=1 ), ] for i, layer in enumerate(self.layers ): self.add_module(str(UpperCamelCase__ ) , UpperCamelCase__ ) def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : torch.Tensor ): A = input for layer in self.layers: A = layer(UpperCamelCase__ ) return hidden_state class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Tuple[int, ...] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : bool ): super().__init__() A = pool_scales A = align_corners A = in_channels A = channels A = [] for i, pool_scale in enumerate(UpperCamelCase__ ): A = UperNetPyramidPoolingBlock(pool_scale=UpperCamelCase__ , in_channels=UpperCamelCase__ , channels=UpperCamelCase__ ) self.blocks.append(UpperCamelCase__ ) self.add_module(str(UpperCamelCase__ ) , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : torch.Tensor ): A = [] for ppm in self.blocks: A = ppm(UpperCamelCase__ ) A = nn.functional.interpolate( UpperCamelCase__ , size=x.size()[2:] , mode='bilinear' , align_corners=self.align_corners ) ppm_outs.append(UpperCamelCase__ ) return ppm_outs class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[Any] ): super().__init__() A = config A = config.pool_scales # e.g. (1, 2, 3, 6) A = in_channels A = config.hidden_size A = False A = nn.Convad(self.channels , config.num_labels , kernel_size=1 ) # PSP Module A = UperNetPyramidPoolingModule( self.pool_scales , self.in_channels[-1] , self.channels , align_corners=self.align_corners , ) A = UperNetConvModule( self.in_channels[-1] + len(self.pool_scales ) * self.channels , self.channels , kernel_size=3 , padding=1 , ) # FPN Module A = nn.ModuleList() A = nn.ModuleList() for in_channels in self.in_channels[:-1]: # skip the top layer A = UperNetConvModule(UpperCamelCase__ , self.channels , kernel_size=1 ) A = UperNetConvModule(self.channels , self.channels , kernel_size=3 , padding=1 ) self.lateral_convs.append(UpperCamelCase__ ) self.fpn_convs.append(UpperCamelCase__ ) A = UperNetConvModule( len(self.in_channels ) * self.channels , self.channels , kernel_size=3 , padding=1 , ) def UpperCamelCase ( self : Optional[int] ): self.apply(self._init_weights ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple ): if isinstance(UpperCamelCase__ , nn.Convad ): module.weight.data.normal_(mean=0.0 , std=self.config.initializer_range ) if module.bias is not None: module.bias.data.zero_() def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : Dict ): A = inputs[-1] A = [x] psp_outs.extend(self.psp_modules(UpperCamelCase__ ) ) A = torch.cat(UpperCamelCase__ , dim=1 ) A = self.bottleneck(UpperCamelCase__ ) return output def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : torch.Tensor ): # build laterals A = [lateral_conv(encoder_hidden_states[i] ) for i, lateral_conv in enumerate(self.lateral_convs )] laterals.append(self.psp_forward(UpperCamelCase__ ) ) # build top-down path A = len(UpperCamelCase__ ) for i in range(used_backbone_levels - 1 , 0 , -1 ): A = laterals[i - 1].shape[2:] A = laterals[i - 1] + nn.functional.interpolate( laterals[i] , size=UpperCamelCase__ , mode='bilinear' , align_corners=self.align_corners ) # build outputs A = [self.fpn_convs[i](laterals[i] ) for i in range(used_backbone_levels - 1 )] # append psp feature fpn_outs.append(laterals[-1] ) for i in range(used_backbone_levels - 1 , 0 , -1 ): A = nn.functional.interpolate( fpn_outs[i] , size=fpn_outs[0].shape[2:] , mode='bilinear' , align_corners=self.align_corners ) A = torch.cat(UpperCamelCase__ , dim=1 ) A = self.fpn_bottleneck(UpperCamelCase__ ) A = self.classifier(UpperCamelCase__ ) return output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : List[Any] , UpperCamelCase__ : int = 2 , UpperCamelCase__ : int = 3 , UpperCamelCase__ : Union[int, Tuple[int, int]] = 1 ): super().__init__() A = config A = config.auxiliary_in_channels A = config.auxiliary_channels A = config.auxiliary_num_convs A = config.auxiliary_concat_input A = in_index A = (kernel_size // 2) * dilation A = [] convs.append( UperNetConvModule( self.in_channels , self.channels , kernel_size=UpperCamelCase__ , padding=UpperCamelCase__ , dilation=UpperCamelCase__ ) ) for i in range(self.num_convs - 1 ): convs.append( UperNetConvModule( self.channels , self.channels , kernel_size=UpperCamelCase__ , padding=UpperCamelCase__ , dilation=UpperCamelCase__ ) ) if self.num_convs == 0: A = nn.Identity() else: A = nn.Sequential(*UpperCamelCase__ ) if self.concat_input: A = UperNetConvModule( self.in_channels + self.channels , self.channels , kernel_size=UpperCamelCase__ , padding=kernel_size // 2 ) A = nn.Convad(self.channels , config.num_labels , kernel_size=1 ) def UpperCamelCase ( self : Optional[Any] ): self.apply(self._init_weights ) def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : List[str] ): if isinstance(UpperCamelCase__ , nn.Convad ): module.weight.data.normal_(mean=0.0 , std=self.config.initializer_range ) if module.bias is not None: module.bias.data.zero_() def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : torch.Tensor ): # just take the relevant feature maps A = encoder_hidden_states[self.in_index] A = self.convs(UpperCamelCase__ ) if self.concat_input: A = self.conv_cat(torch.cat([hidden_states, output] , dim=1 ) ) A = self.classifier(UpperCamelCase__ ) return output class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = UperNetConfig SCREAMING_SNAKE_CASE : List[str] = '''pixel_values''' SCREAMING_SNAKE_CASE : int = True def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): module.backbone.init_weights() module.decode_head.init_weights() module.auxiliary_head.init_weights() def UpperCamelCase ( self : str ): self.backbone.init_weights() self.decode_head.init_weights() self.auxiliary_head.init_weights() def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Optional[int]=False ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = value _UpperCAmelCase = R"\n Parameters:\n This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) sub-class. Use\n it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and\n behavior.\n config ([`UperNetConfig`]): Model configuration class with all the parameters of the model.\n Initializing with a config file does not load the weights associated with the model, only the\n configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.\n" _UpperCAmelCase = R"\n Args:\n pixel_values (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`):\n Pixel values. Padding will be ignored by default should you provide it. Pixel values can be obtained using\n [`AutoImageProcessor`]. See [`SegformerImageProcessor.__call__`] for details.\n output_attentions (`bool`, *optional*):\n Whether or not to return the attentions tensors of all attention layers in case the backbone has them. See\n `attentions` under returned tensors for more detail.\n output_hidden_states (`bool`, *optional*):\n Whether or not to return the hidden states of all layers of the backbone. See `hidden_states` under\n returned tensors for more detail.\n return_dict (`bool`, *optional*):\n Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.\n" @add_start_docstrings( '''UperNet framework leveraging any vision backbone e.g. for ADE20k, CityScapes.''' , __lowercase , ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : Union[str, Any] ): super().__init__(UpperCamelCase__ ) A = AutoBackbone.from_config(config.backbone_config ) # Semantic segmentation head(s) A = UperNetHead(UpperCamelCase__ , in_channels=self.backbone.channels ) A = UperNetFCNHead(UpperCamelCase__ ) if config.use_auxiliary_head else None # Initialize weights and apply final processing self.post_init() @add_start_docstrings_to_model_forward(UPERNET_INPUTS_DOCSTRING.format('batch_size, sequence_length' ) ) @replace_return_docstrings(output_type=UpperCamelCase__ , config_class=_CONFIG_FOR_DOC ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Optional[torch.Tensor] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[torch.Tensor] = None , UpperCamelCase__ : Optional[bool] = None , ): A = return_dict if return_dict is not None else self.config.use_return_dict A = ( output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states ) A = output_attentions if output_attentions is not None else self.config.output_attentions A = self.backbone.forward_with_filtered_kwargs( UpperCamelCase__ , output_hidden_states=UpperCamelCase__ , output_attentions=UpperCamelCase__ ) A = outputs.feature_maps A = self.decode_head(UpperCamelCase__ ) A = nn.functional.interpolate(UpperCamelCase__ , size=pixel_values.shape[2:] , mode='bilinear' , align_corners=UpperCamelCase__ ) A = None if self.auxiliary_head is not None: A = self.auxiliary_head(UpperCamelCase__ ) A = nn.functional.interpolate( UpperCamelCase__ , size=pixel_values.shape[2:] , mode='bilinear' , align_corners=UpperCamelCase__ ) A = None if labels is not None: if self.config.num_labels == 1: raise ValueError('The number of labels should be greater than one' ) else: # compute weighted loss A = CrossEntropyLoss(ignore_index=self.config.loss_ignore_index ) A = loss_fct(UpperCamelCase__ , UpperCamelCase__ ) A = loss_fct(UpperCamelCase__ , UpperCamelCase__ ) A = main_loss + self.config.auxiliary_loss_weight * auxiliary_loss if not return_dict: if output_hidden_states: A = (logits,) + outputs[1:] else: A = (logits,) + outputs[2:] return ((loss,) + output) if loss is not None else output return SemanticSegmenterOutput( loss=UpperCamelCase__ , logits=UpperCamelCase__ , hidden_states=outputs.hidden_states , attentions=outputs.attentions , )
699
import numpy as np import torch from imwatermark import WatermarkEncoder # Copied from https://github.com/Stability-AI/generative-models/blob/613af104c6b85184091d42d374fef420eddb356d/scripts/demo/streamlit_helpers.py#L66 _UpperCAmelCase = 0b10_11_00_11_11_10_11_00_10_01_00_00_01_11_10_11_10_11_00_01_10_01_11_10 # bin(x)[2:] gives bits of x as str, use int to convert them to 0/1 _UpperCAmelCase = [int(bit) for bit in bin(WATERMARK_MESSAGE)[2:]] class _UpperCAmelCase : '''simple docstring''' def __init__( self : Union[str, Any] ): A = WATERMARK_BITS A = WatermarkEncoder() self.encoder.set_watermark('bits' , self.watermark ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : torch.FloatTensor ): # can't encode images that are smaller than 256 if images.shape[-1] < 256: return images A = (255 * (images / 2 + 0.5)).cpu().permute(0 , 2 , 3 , 1 ).float().numpy() A = [self.encoder.encode(UpperCamelCase__ , 'dwtDct' ) for image in images] A = torch.from_numpy(np.array(UpperCamelCase__ ) ).permute(0 , 3 , 1 , 2 ) A = torch.clamp(2 * (images / 255 - 0.5) , min=-1.0 , max=1.0 ) return images
699
1
import os from shutil import copyfile from typing import Any, Dict, List, Optional, Tuple import sentencepiece as spm from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = "▁" _UpperCAmelCase = {"vocab_file": "sentencepiece.bpe.model"} _UpperCAmelCase = { "vocab_file": { "facebook/xglm-564M": "https://huggingface.co/facebook/xglm-564M/resolve/main/sentencepiece.bpe.model", } } _UpperCAmelCase = { "facebook/xglm-564M": 2_048, } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : str = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : Optional[int] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Optional[Any] = ['''input_ids''', '''attention_mask'''] def __init__( self : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="</s>" , UpperCamelCase__ : Tuple="</s>" , UpperCamelCase__ : Optional[Any]="<s>" , UpperCamelCase__ : int="<unk>" , UpperCamelCase__ : Dict="<pad>" , UpperCamelCase__ : Optional[Dict[str, Any]] = None , **UpperCamelCase__ : Tuple , ): A = {} if sp_model_kwargs is None else sp_model_kwargs # Compatibility with the original tokenizer A = 7 A = [f'''<madeupword{i}>''' for i in range(self.num_madeup_words )] A = kwargs.get('additional_special_tokens' , [] ) kwargs["additional_special_tokens"] += [ word for word in madeup_words if word not in kwargs["additional_special_tokens"] ] super().__init__( bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , unk_token=UpperCamelCase__ , sep_token=UpperCamelCase__ , cls_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , sp_model_kwargs=self.sp_model_kwargs , **UpperCamelCase__ , ) A = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.Load(str(UpperCamelCase__ ) ) A = 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' # The first "real" token "," has position 4 in the original fairseq vocab and position 3 in the spm vocab A = 1 # Mimic fairseq token-to-id alignment for the first 4 token A = {'<s>': 0, '<pad>': 1, '</s>': 2, '<unk>': 3} A = len(self.sp_model ) A = {f'''<madeupword{i}>''': sp_size + i + self.fairseq_offset for i in range(self.num_madeup_words )} self.fairseq_tokens_to_ids.update(UpperCamelCase__ ) A = {v: k for k, v in self.fairseq_tokens_to_ids.items()} def __getstate__( self : List[Any] ): A = self.__dict__.copy() A = None A = self.sp_model.serialized_model_proto() return state def __setstate__( self : Union[str, Any] , UpperCamelCase__ : Any ): A = d # for backward compatibility if not hasattr(self , 'sp_model_kwargs' ): A = {} A = spm.SentencePieceProcessor(**self.sp_model_kwargs ) self.sp_model.LoadFromSerializedProto(self.sp_model_proto ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : List[int] , UpperCamelCase__ : Optional[List[int]] = None ): if token_ids_a is None: return [self.sep_token_id] + token_ids_a A = [self.sep_token_id] return sep + token_ids_a + sep + sep + token_ids_a def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : List[int] , UpperCamelCase__ : Optional[List[int]] = None , UpperCamelCase__ : bool = False ): if already_has_special_tokens: return super().get_special_tokens_mask( token_ids_a=UpperCamelCase__ , token_ids_a=UpperCamelCase__ , already_has_special_tokens=UpperCamelCase__ ) if token_ids_a is None: return [1] + ([0] * len(UpperCamelCase__ )) return [1] + ([0] * len(UpperCamelCase__ )) + [1, 1] + ([0] * len(UpperCamelCase__ )) def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : List[int] , UpperCamelCase__ : Optional[List[int]] = None ): A = [self.sep_token_id] if token_ids_a is None: return len(sep + token_ids_a ) * [0] return len(sep + token_ids_a + sep + sep + token_ids_a ) * [0] @property def UpperCamelCase ( self : str ): return len(self.sp_model ) + self.fairseq_offset + self.num_madeup_words def UpperCamelCase ( self : int ): A = {self.convert_ids_to_tokens(UpperCamelCase__ ): i for i in range(self.vocab_size )} vocab.update(self.added_tokens_encoder ) return vocab def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str ): return self.sp_model.encode(UpperCamelCase__ , out_type=UpperCamelCase__ ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Optional[int] ): if token in self.fairseq_tokens_to_ids: return self.fairseq_tokens_to_ids[token] A = self.sp_model.PieceToId(UpperCamelCase__ ) # 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 UpperCamelCase ( self : Dict , UpperCamelCase__ : str ): 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 UpperCamelCase ( self : Any , UpperCamelCase__ : str ): A = ''.join(UpperCamelCase__ ).replace(UpperCamelCase__ , ' ' ).strip() return out_string def UpperCamelCase ( self : str , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) if os.path.abspath(self.vocab_file ) != os.path.abspath(UpperCamelCase__ ) and os.path.isfile(self.vocab_file ): copyfile(self.vocab_file , UpperCamelCase__ ) elif not os.path.isfile(self.vocab_file ): with open(UpperCamelCase__ , 'wb' ) as fi: A = self.sp_model.serialized_model_proto() fi.write(UpperCamelCase__ ) return (out_vocab_file,)
699
import argparse import fairseq import torch from transformers import UniSpeechSatConfig, UniSpeechSatForCTC, UniSpeechSatForPreTraining, logging logging.set_verbosity_info() _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "post_extract_proj": "feature_projection.projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.layer_norm": "encoder.layer_norm", "encoder.layer_norm_for_extract": "layer_norm_for_extract", "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": "lm_head", "label_embs_concat": "label_embeddings_concat", "mask_emb": "masked_spec_embed", "spk_proj": "speaker_proj", } _UpperCAmelCase = [ "lm_head", "quantizer.weight_proj", "quantizer.codevectors", "project_q", "project_hid", "label_embeddings_concat", "speaker_proj", "layer_norm_for_extract", ] def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Dict, lowerCAmelCase : Optional[int], lowerCAmelCase : List[Any], lowerCAmelCase : str ) -> int: 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 if hf_shape != value.shape: raise ValueError( 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 __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Optional[int] ) -> Dict: A = [] A = fairseq_model.state_dict() A = hf_model.unispeech_sat.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 = 'unispeech_sat.' + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key if key in name or key.split('w2v_model.' )[-1] == name.split('.' )[0]: if "layer_norm_for_extract" in name and (".".join(name.split('.' )[:-1] ) != key): # special case since naming is very similar continue 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 "bias" in name: A = 'bias' elif "weight" in name: # TODO: don't match quantizer.weight_proj A = 'weight' 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 __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : Tuple, lowerCAmelCase : List[Any], lowerCAmelCase : int ) -> Dict: 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.bias.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.weight.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape: raise ValueError( f'''{full_name} has size {value.shape}, but''' f''' {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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape: raise ValueError( 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 ) @torch.no_grad() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : Dict, lowerCAmelCase : Union[str, Any]=None, lowerCAmelCase : str=None, lowerCAmelCase : List[Any]=True ) -> Union[str, Any]: if config_path is not None: A = UniSpeechSatConfig.from_pretrained(lowerCAmelCase ) else: A = UniSpeechSatConfig() A = '' if is_finetuned: A = UniSpeechSatForCTC(lowerCAmelCase ) else: A = UniSpeechSatForPreTraining(lowerCAmelCase ) A , A , A = fairseq.checkpoint_utils.load_model_ensemble_and_task( [checkpoint_path], arg_overrides={'data': '/'.join(dict_path.split('/' )[:-1] )} ) A = model[0].eval() recursively_load_weights(lowerCAmelCase, lowerCAmelCase ) hf_wavavec.save_pretrained(lowerCAmelCase ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--dict_path", default=None, type=str, help="Path to dict of fine-tuned model") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") parser.add_argument( "--not_finetuned", action="store_true", help="Whether the model to convert is a fine-tuned model or not" ) _UpperCAmelCase = parser.parse_args() convert_unispeech_sat_checkpoint( args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned )
699
1
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> str: 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=10))
699
from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar _UpperCAmelCase = TypeVar("T") class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : T ): A = data A = None def __str__( self : Optional[int] ): return f'''{self.data}''' class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple ): A = None def __iter__( self : int ): A = self.top while node: yield node.data A = node.next def __str__( self : Any ): return "->".join([str(UpperCamelCase__ ) for item in self] ) def __len__( self : Dict ): return len(tuple(iter(self ) ) ) def UpperCamelCase ( self : List[str] ): return self.top is None def UpperCamelCase ( self : Dict , UpperCamelCase__ : T ): A = Node(UpperCamelCase__ ) if not self.is_empty(): A = self.top A = node def UpperCamelCase ( self : Dict ): if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top , UpperCamelCase__ ) A = self.top A = self.top.next return pop_node.data def UpperCamelCase ( self : List[str] ): if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def UpperCamelCase ( self : List[str] ): A = None if __name__ == "__main__": from doctest import testmod testmod()
699
1
import unittest from transformers import DebertaVaTokenizer, DebertaVaTokenizerFast from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, slow from ...test_tokenization_common import TokenizerTesterMixin _UpperCAmelCase = get_tests_dir("fixtures/spiece.model") @require_sentencepiece @require_tokenizers class _UpperCAmelCase ( __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = DebertaVaTokenizer SCREAMING_SNAKE_CASE : Optional[Any] = DebertaVaTokenizerFast SCREAMING_SNAKE_CASE : Any = True SCREAMING_SNAKE_CASE : str = True def UpperCamelCase ( self : Optional[int] ): super().setUp() # We have a SentencePiece fixture for testing A = DebertaVaTokenizer(UpperCamelCase__ , unk_token='<unk>' ) tokenizer.save_pretrained(self.tmpdirname ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Dict ): A = 'this is a test' A = 'this is a test' return input_text, output_text def UpperCamelCase ( self : Dict ): A = '<pad>' A = 0 self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCamelCase__ ) , UpperCamelCase__ ) self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCamelCase__ ) , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] ): A = list(self.get_tokenizer().get_vocab().keys() ) self.assertEqual(vocab_keys[0] , '<pad>' ) self.assertEqual(vocab_keys[1] , '<unk>' ) self.assertEqual(vocab_keys[-1] , '[PAD]' ) self.assertEqual(len(UpperCamelCase__ ) , 30001 ) def UpperCamelCase ( self : List[Any] ): self.assertEqual(self.get_tokenizer().vocab_size , 30000 ) def UpperCamelCase ( self : Dict ): # fmt: off A = ' \tHeLLo!how \n Are yoU? ' A = ['▁hello', '!', 'how', '▁are', '▁you', '?'] # fmt: on A = DebertaVaTokenizer(UpperCamelCase__ , do_lower_case=UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , do_lower_case=UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) @unittest.skip('There is an inconsistency between slow and fast tokenizer due to a bug in the fast one.' ) def UpperCamelCase ( self : Optional[int] ): pass @unittest.skip('There is an inconsistency between slow and fast tokenizer due to a bug in the fast one.' ) def UpperCamelCase ( self : Tuple ): pass def UpperCamelCase ( self : Union[str, Any] ): # fmt: off A = 'I was born in 92000, and this is falsé.' A = ['▁', '<unk>', '▁was', '▁born', '▁in', '▁9', '2000', '▁', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '▁', '.', ] # fmt: on A = DebertaVaTokenizer(UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : str ): # fmt: off A = 'I was born in 92000, and this is falsé.' A = ['▁i', '▁was', '▁born', '▁in', '▁9', '2000', '▁', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '▁', '.', ] # fmt: on A = DebertaVaTokenizer(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : List[Any] ): # fmt: off A = 'I was born in 92000, and this is falsé.' A = ['▁i', '▁was', '▁born', '▁in', '▁9', '2000', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '.', ] # fmt: on A = DebertaVaTokenizer(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any ): # fmt: off A = 'I was born in 92000, and this is falsé.' A = ['▁', '<unk>', '▁was', '▁born', '▁in', '▁9', '2000', '▁', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '▁', '.', ] # fmt: on A = DebertaVaTokenizer(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] ): # fmt: off A = ' \tHeLLo!how \n Are yoU? ' A = ['▁', '<unk>', 'e', '<unk>', 'o', '!', 'how', '▁', '<unk>', 're', '▁yo', '<unk>', '?'] # fmt: on A = DebertaVaTokenizer(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , do_lower_case=UpperCamelCase__ , split_by_punct=UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[int] ): A = self.get_tokenizer() A = self.get_rust_tokenizer() A = 'I was born in 92000, and this is falsé.' A = tokenizer.convert_ids_to_tokens(tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) A = rust_tokenizer.convert_ids_to_tokens(rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) A = rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = self.get_rust_tokenizer() A = tokenizer.encode(UpperCamelCase__ ) A = rust_tokenizer.encode(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any ): A = 'This is a test' A = [13, 1, 4398, 25, 21, 1289] A = ['▁', 'T', 'his', '▁is', '▁a', '▁test'] A = ['▁', '<unk>', 'his', '▁is', '▁a', '▁test'] A = DebertaVaTokenizer(UpperCamelCase__ , keep_accents=UpperCamelCase__ ) A = DebertaVaTokenizerFast(UpperCamelCase__ , keep_accents=UpperCamelCase__ ) A = tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokenizer.tokenize(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = rust_tokenizer.tokenize(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) # fmt: off A = 'I was born in 92000, and this is falsé.' A = [13, 1, 23, 386, 19, 561, 3050, 15, 17, 48, 25, 8256, 18, 1, 9] A = ['▁', 'I', '▁was', '▁born', '▁in', '▁9', '2000', ',', '▁and', '▁this', '▁is', '▁fal', 's', 'é', '.', ] A = ['▁', '<unk>', '▁was', '▁born', '▁in', '▁9', '2000', ',', '▁and', '▁this', '▁is', '▁fal', 's', '<unk>', '.', ] # fmt: on A = tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokenizer.tokenize(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokenizer.convert_ids_to_tokens(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = rust_tokenizer.tokenize(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = rust_tokenizer.convert_ids_to_tokens(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : List[str] ): A = DebertaVaTokenizer(UpperCamelCase__ ) A = tokenizer.encode('sequence builders' ) A = tokenizer.encode('multi-sequence build' ) A = tokenizer.build_inputs_with_special_tokens(UpperCamelCase__ ) A = tokenizer.build_inputs_with_special_tokens(UpperCamelCase__ , UpperCamelCase__ ) self.assertEqual([tokenizer.cls_token_id] + text + [tokenizer.sep_token_id] , UpperCamelCase__ ) self.assertEqual( [tokenizer.cls_token_id] + text + [tokenizer.sep_token_id] + text_a + [tokenizer.sep_token_id] , UpperCamelCase__ , ) @slow def UpperCamelCase ( self : Tuple ): # fmt: off A = {'input_ids': [[1, 39867, 36, 19390, 486, 27, 35052, 81436, 18, 60685, 1225, 7, 35052, 81436, 18, 9367, 16899, 18, 15937, 53, 594, 773, 18, 16287, 30465, 36, 15937, 6, 41139, 38, 36979, 60763, 191, 6, 34132, 99, 6, 50538, 390, 43230, 6, 34132, 2779, 20850, 14, 699, 1072, 1194, 36, 382, 10901, 53, 7, 699, 1072, 2084, 36, 20422, 630, 53, 19, 105, 3049, 1896, 1053, 16899, 1506, 11, 37978, 4243, 7, 1237, 31869, 200, 16566, 654, 6, 35052, 81436, 7, 55630, 13593, 4, 2], [1, 26, 15011, 13, 667, 8, 1053, 18, 23611, 1237, 72356, 12820, 34, 104134, 1209, 35, 13313, 6627, 21, 202, 347, 7, 164, 2399, 11, 46, 4485, 4, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 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, 5, 1232, 2864, 15785, 14951, 105, 5, 8581, 1250, 4, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], 'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], 'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501 # fmt: on self.tokenizer_integration_test_util( expected_encoding=UpperCamelCase__ , model_name='microsoft/deberta-v2-xlarge' , revision='ad6e42c1532ddf3a15c39246b63f5559d558b670' , )
699
from __future__ import annotations import math def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : bool, lowerCAmelCase : list[int], lowerCAmelCase : float ) -> int: if depth < 0: raise ValueError('Depth cannot be less than 0' ) if not scores: raise ValueError('Scores cannot be empty' ) if depth == height: return scores[node_index] return ( max( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) if is_max else min( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) ) def __UpperCamelCase () -> None: A = [90, 23, 6, 33, 21, 65, 123, 34_423] A = math.log(len(lowerCAmelCase ), 2 ) print(f'''Optimal value : {minimax(0, 0, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )}''' ) if __name__ == "__main__": import doctest doctest.testmod() main()
699
1
import json import os import unittest from transformers.models.xlm.tokenization_xlm import VOCAB_FILES_NAMES, XLMTokenizer from transformers.testing_utils import slow from ...test_tokenization_common import TokenizerTesterMixin class _UpperCAmelCase ( __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : str = XLMTokenizer SCREAMING_SNAKE_CASE : Union[str, Any] = False def UpperCamelCase ( self : Dict ): 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(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = ['l o 123', 'lo w 1456', 'e r</w> 1789', ''] 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(UpperCamelCase__ ) ) with open(self.merges_file , 'w' ) as fp: fp.write('\n'.join(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[int] ): A = 'lower newer' A = 'lower newer' return input_text, output_text def UpperCamelCase ( self : Union[str, Any] ): A = XLMTokenizer(self.vocab_file , self.merges_file ) A = 'lower' A = ['low', 'er</w>'] A = tokenizer.tokenize(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokens + ['<unk>'] A = [14, 15, 20] self.assertListEqual(tokenizer.convert_tokens_to_ids(UpperCamelCase__ ) , UpperCamelCase__ ) @slow def UpperCamelCase ( self : Optional[Any] ): A = XLMTokenizer.from_pretrained('xlm-mlm-en-2048' ) A = tokenizer.encode('sequence builders' , add_special_tokens=UpperCamelCase__ ) A = tokenizer.encode('multi-sequence build' , add_special_tokens=UpperCamelCase__ ) A = tokenizer.build_inputs_with_special_tokens(UpperCamelCase__ ) A = tokenizer.build_inputs_with_special_tokens(UpperCamelCase__ , UpperCamelCase__ ) assert encoded_sentence == [0] + text + [1] assert encoded_pair == [0] + text + [1] + text_a + [1]
699
import logging import os from dataclasses import dataclass, field from typing import Dict, Optional import datasets import numpy as np import tensorflow as tf from transformers import ( AutoConfig, AutoTokenizer, EvalPrediction, HfArgumentParser, PreTrainedTokenizer, TFAutoModelForSequenceClassification, TFTrainer, TFTrainingArguments, ) from transformers.utils import logging as hf_logging hf_logging.set_verbosity_info() hf_logging.enable_default_handler() hf_logging.enable_explicit_format() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : PreTrainedTokenizer, lowerCAmelCase : int, lowerCAmelCase : Optional[int] = None, ) -> Dict: A = {} if train_file is not None: A = [train_file] if eval_file is not None: A = [eval_file] if test_file is not None: A = [test_file] A = datasets.load_dataset('csv', data_files=lowerCAmelCase ) A = list(ds[list(files.keys() )[0]].features.keys() ) A = features_name.pop(lowerCAmelCase ) A = list(set(ds[list(files.keys() )[0]][label_name] ) ) A = {label: i for i, label in enumerate(lowerCAmelCase )} A = tokenizer.model_input_names A = {} if len(lowerCAmelCase ) == 1: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( example[features_name[0]], truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length' ), batched=lowerCAmelCase, ) elif len(lowerCAmelCase ) == 2: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( (example[features_name[0]], example[features_name[1]]), truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length', ), batched=lowerCAmelCase, ) def gen_train(): for ex in transformed_ds[datasets.Split.TRAIN]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_val(): for ex in transformed_ds[datasets.Split.VALIDATION]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_test(): for ex in transformed_ds[datasets.Split.TEST]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TRAIN in transformed_ds else None ) if train_ds is not None: A = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.VALIDATION in transformed_ds else None ) if val_ds is not None: A = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TEST in transformed_ds else None ) if test_ds is not None: A = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) ) return train_ds, val_ds, test_ds, labelaid _UpperCAmelCase = logging.getLogger(__name__) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : int = field(metadata={'''help''': '''Which column contains the label'''} ) SCREAMING_SNAKE_CASE : str = field(default=__lowercase , metadata={'''help''': '''The path of the training file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the development file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the test file'''} ) SCREAMING_SNAKE_CASE : int = field( default=1_28 , metadata={ '''help''': ( '''The maximum total input sequence length after tokenization. Sequences longer ''' '''than this will be truncated, sequences shorter will be padded.''' ) } , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Overwrite the cached training and evaluation sets'''} ) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : str = field( metadata={'''help''': '''Path to pretrained model or model identifier from huggingface.co/models'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained config name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained tokenizer name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : bool = field(default=__lowercase , metadata={'''help''': '''Set this flag to use fast tokenization.'''} ) # If you want to tweak more attributes on your tokenizer, you should do it in a distinct script, # or just modify its tokenizer_config.json. SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Where do you want to store the pretrained models downloaded from huggingface.co'''} , ) def __UpperCamelCase () -> Any: # See all possible arguments in src/transformers/training_args.py # or by passing the --help flag to this script. # We now keep distinct sets of args, for a cleaner separation of concerns. A = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) ) A , A , A = parser.parse_args_into_dataclasses() if ( os.path.exists(training_args.output_dir ) and os.listdir(training_args.output_dir ) and training_args.do_train and not training_args.overwrite_output_dir ): raise ValueError( f'''Output directory ({training_args.output_dir}) already exists and is not empty. Use''' ' --overwrite_output_dir to overcome.' ) # Setup logging logging.basicConfig( format='%(asctime)s - %(levelname)s - %(name)s - %(message)s', datefmt='%m/%d/%Y %H:%M:%S', level=logging.INFO, ) logger.info( f'''n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, ''' f'''16-bits training: {training_args.fpaa}''' ) logger.info(f'''Training/evaluation parameters {training_args}''' ) # Load pretrained model and tokenizer # # Distributed training: # The .from_pretrained methods guarantee that only one local process can concurrently # download model & vocab. A = AutoTokenizer.from_pretrained( model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path, cache_dir=model_args.cache_dir, ) A , A , A , A = get_tfds( train_file=data_args.train_file, eval_file=data_args.dev_file, test_file=data_args.test_file, tokenizer=lowerCAmelCase, label_column_id=data_args.label_column_id, max_seq_length=data_args.max_seq_length, ) A = AutoConfig.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, num_labels=len(lowerCAmelCase ), labelaid=lowerCAmelCase, idalabel={id: label for label, id in labelaid.items()}, finetuning_task='text-classification', cache_dir=model_args.cache_dir, ) with training_args.strategy.scope(): A = TFAutoModelForSequenceClassification.from_pretrained( model_args.model_name_or_path, from_pt=bool('.bin' in model_args.model_name_or_path ), config=lowerCAmelCase, cache_dir=model_args.cache_dir, ) def compute_metrics(lowerCAmelCase : EvalPrediction ) -> Dict: A = np.argmax(p.predictions, axis=1 ) return {"acc": (preds == p.label_ids).mean()} # Initialize our Trainer A = TFTrainer( model=lowerCAmelCase, args=lowerCAmelCase, train_dataset=lowerCAmelCase, eval_dataset=lowerCAmelCase, compute_metrics=lowerCAmelCase, ) # Training if training_args.do_train: trainer.train() trainer.save_model() tokenizer.save_pretrained(training_args.output_dir ) # Evaluation A = {} if training_args.do_eval: logger.info('*** Evaluate ***' ) A = trainer.evaluate() A = os.path.join(training_args.output_dir, 'eval_results.txt' ) with open(lowerCAmelCase, 'w' ) as writer: logger.info('***** Eval results *****' ) for key, value in result.items(): logger.info(f''' {key} = {value}''' ) writer.write(f'''{key} = {value}\n''' ) results.update(lowerCAmelCase ) return results if __name__ == "__main__": main()
699
1
class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] ): A = 0 A = 0 A = {} def UpperCamelCase ( self : Any , UpperCamelCase__ : List[Any] ): if vertex not in self.adjacency: A = {} self.num_vertices += 1 def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : str , UpperCamelCase__ : int ): self.add_vertex(UpperCamelCase__ ) self.add_vertex(UpperCamelCase__ ) if head == tail: return A = weight A = weight def UpperCamelCase ( self : Optional[Any] ): A = self.get_edges() for edge in edges: A , A , A = edge edges.remove((tail, head, weight) ) for i in range(len(UpperCamelCase__ ) ): A = list(edges[i] ) edges.sort(key=lambda UpperCamelCase__ : e[2] ) for i in range(len(UpperCamelCase__ ) - 1 ): if edges[i][2] >= edges[i + 1][2]: A = edges[i][2] + 1 for edge in edges: A , A , A = edge A = weight A = weight def __str__( self : Dict ): A = '' for tail in self.adjacency: for head in self.adjacency[tail]: A = self.adjacency[head][tail] string += f'''{head} -> {tail} == {weight}\n''' return string.rstrip('\n' ) def UpperCamelCase ( self : List[Any] ): A = [] for tail in self.adjacency: for head in self.adjacency[tail]: output.append((tail, head, self.adjacency[head][tail]) ) return output def UpperCamelCase ( self : Union[str, Any] ): return self.adjacency.keys() @staticmethod def UpperCamelCase ( UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : List[str]=None ): A = Graph() if vertices is None: A = [] if edges is None: A = [] for vertex in vertices: g.add_vertex(UpperCamelCase__ ) for edge in edges: g.add_edge(*UpperCamelCase__ ) return g class _UpperCAmelCase : '''simple docstring''' def __init__( self : Optional[int] ): A = {} A = {} def __len__( self : int ): return len(self.parent ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[Any] ): if item in self.parent: return self.find(UpperCamelCase__ ) A = item A = 0 return item def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[Any] ): if item not in self.parent: return self.make_set(UpperCamelCase__ ) if item != self.parent[item]: A = self.find(self.parent[item] ) return self.parent[item] def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Optional[int] ): A = self.find(UpperCamelCase__ ) A = self.find(UpperCamelCase__ ) if roota == roota: return roota if self.rank[roota] > self.rank[roota]: A = roota return roota if self.rank[roota] < self.rank[roota]: A = roota return roota if self.rank[roota] == self.rank[roota]: self.rank[roota] += 1 A = roota return roota return None @staticmethod def UpperCamelCase ( UpperCamelCase__ : str ): A = graph.num_vertices A = Graph.UnionFind() A = [] while num_components > 1: A = {} for vertex in graph.get_vertices(): A = -1 A = graph.get_edges() for edge in edges: A , A , A = edge edges.remove((tail, head, weight) ) for edge in edges: A , A , A = edge A = union_find.find(UpperCamelCase__ ) A = union_find.find(UpperCamelCase__ ) if seta != seta: if cheap_edge[seta] == -1 or cheap_edge[seta][2] > weight: A = [head, tail, weight] if cheap_edge[seta] == -1 or cheap_edge[seta][2] > weight: A = [head, tail, weight] for vertex in cheap_edge: if cheap_edge[vertex] != -1: A , A , A = cheap_edge[vertex] if union_find.find(UpperCamelCase__ ) != union_find.find(UpperCamelCase__ ): union_find.union(UpperCamelCase__ , UpperCamelCase__ ) mst_edges.append(cheap_edge[vertex] ) A = num_components - 1 A = Graph.build(edges=UpperCamelCase__ ) return mst
699
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = {"configuration_xglm": ["XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "XGLMConfig"]} try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizer"] try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "XGLMForCausalLM", "XGLMModel", "XGLMPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxXGLMForCausalLM", "FlaxXGLMModel", "FlaxXGLMPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "TFXGLMForCausalLM", "TFXGLMModel", "TFXGLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm import XGLMTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm_fast import XGLMTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_xglm import ( TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, TFXGLMForCausalLM, TFXGLMModel, TFXGLMPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure)
699
1
import json import os from typing import Dict, List, Optional, Tuple from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "vocab_file": "vocab.json", "tokenizer_config_file": "tokenizer_config.json", "merges_file": "merges.txt", } _UpperCAmelCase = { "vocab_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/vocab.json" ), }, "tokenizer_config_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/tokenizer_config.json" ), }, "merges_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/merges.txt" ), }, } _UpperCAmelCase = "</w>" _UpperCAmelCase = "@@ " def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> List[str]: A = set() A = word[0] for char in word[1:]: pairs.add((prev_char, char) ) A = char return pairs # Speech2Text2 has no max input length _UpperCAmelCase = {"facebook/s2t-wav2vec2-large-en-de": 1_024} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] def __init__( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="<pad>" , UpperCamelCase__ : int="</s>" , UpperCamelCase__ : Tuple="<unk>" , UpperCamelCase__ : List[str]=False , UpperCamelCase__ : List[str]=None , **UpperCamelCase__ : Optional[int] , ): super().__init__( unk_token=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , **UpperCamelCase__ , ) A = do_lower_case with open(UpperCamelCase__ , encoding='utf-8' ) as vocab_handle: A = json.load(UpperCamelCase__ ) A = {v: k for k, v in self.encoder.items()} if merges_file is None: logger.info(f'''No merges files provided. {self.__class__.__name__} can only be used for decoding.''' ) A = None A = None else: with open(UpperCamelCase__ , encoding='utf-8' ) as merges_handle: A = merges_handle.read().split('\n' )[:-1] A = [tuple(merge.split()[:2] ) for merge in merges] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = {} @property def UpperCamelCase ( self : Union[str, Any] ): return len(self.decoder ) def UpperCamelCase ( self : Optional[Any] ): return dict(self.encoder , **self.added_tokens_encoder ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): A = tuple(token[:-1] ) + (token[-1] + BPE_TOKEN_MERGES,) if token in self.cache: return self.cache[token] A = get_pairs(UpperCamelCase__ ) if not pairs: return token while True: A = min(UpperCamelCase__ , key=lambda UpperCamelCase__ : self.bpe_ranks.get(UpperCamelCase__ , float('inf' ) ) ) if bigram not in self.bpe_ranks: break A , A = bigram A = [] A = 0 while i < len(UpperCamelCase__ ): try: A = word.index(UpperCamelCase__ , UpperCamelCase__ ) except ValueError: new_word.extend(word[i:] ) break else: new_word.extend(word[i:j] ) A = j if word[i] == first and i < len(UpperCamelCase__ ) - 1 and word[i + 1] == second: new_word.append(first + second ) i += 2 else: new_word.append(word[i] ) i += 1 A = tuple(UpperCamelCase__ ) A = new_word if len(UpperCamelCase__ ) == 1: break else: A = get_pairs(UpperCamelCase__ ) A = ' '.join(UpperCamelCase__ ) if word == "\n " + BPE_TOKEN_MERGES: A = '\n' + BPE_TOKEN_MERGES if word.endswith(UpperCamelCase__ ): A = word.replace(UpperCamelCase__ , '' ) A = word.replace(' ' , UpperCamelCase__ ) A = word return word def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict ): if self.bpe_ranks is None: raise ValueError( 'This tokenizer was instantiated without a `merges.txt` file, so' ' that it can only be used for decoding, not for encoding.' 'Make sure to provide `merges.txt` file at instantiation to enable ' 'encoding.' ) if self.do_lower_case: A = text.lower() A = text.split() A = [] for token in text: if token: split_tokens.extend(list(self.bpe(UpperCamelCase__ ).split(' ' ) ) ) return split_tokens def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : str ): return self.encoder.get(UpperCamelCase__ , self.encoder.get(self.unk_token ) ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): A = self.decoder.get(UpperCamelCase__ , self.unk_token ) return result def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] ): A = ' '.join(UpperCamelCase__ ) # make sure @@ tokens are concatenated A = ''.join(string.split(UpperCamelCase__ ) ) return string def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] ) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as f: f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCamelCase__ , ensure_ascii=UpperCamelCase__ ) + '\n' ) A = 0 if self.bpe_ranks is None: return (vocab_file,) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as writer: for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCamelCase__ : kv[1] ): if index != token_index: logger.warning( f'''Saving vocabulary to {merges_file}: BPE merge indices are not consecutive.''' ' Please check that the tokenizer is not corrupted!' ) A = token_index writer.write(' '.join(UpperCamelCase__ ) + '\n' ) index += 1 return (vocab_file, merges_file)
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Optional[int]: if b == 0: return 1 if (b % 2) == 0: return actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) else: return a * actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> float: if b < 0: return 1 / actual_power(lowerCAmelCase, lowerCAmelCase ) return actual_power(lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": print(power(-2, -3))
699
1
import json import os import unittest from transformers import AutoTokenizer, GPTaTokenizer, GPTaTokenizerFast from transformers.models.gpta.tokenization_gpta 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''' SCREAMING_SNAKE_CASE : Tuple = GPTaTokenizer SCREAMING_SNAKE_CASE : int = GPTaTokenizerFast SCREAMING_SNAKE_CASE : Optional[int] = True SCREAMING_SNAKE_CASE : List[str] = {'''add_prefix_space''': True} SCREAMING_SNAKE_CASE : Union[str, Any] = False def UpperCamelCase ( self : Optional[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', '\u0120', '\u0120l', '\u0120n', '\u0120lo', '\u0120low', 'er', '\u0120lowest', '\u0120newer', '\u0120wider', '<unk>', '<|endoftext|>', ] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = ['#version: 0.2', '\u0120 l', '\u0120l o', '\u0120lo w', 'e r', ''] A = {'unk_token': '<unk>'} 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' , encoding='utf-8' ) as fp: fp.write(json.dumps(UpperCamelCase__ ) + '\n' ) with open(self.merges_file , 'w' , encoding='utf-8' ) as fp: fp.write('\n'.join(UpperCamelCase__ ) ) def UpperCamelCase ( self : int , **UpperCamelCase__ : Optional[int] ): kwargs.update(self.special_tokens_map ) return GPTaTokenizer.from_pretrained(self.tmpdirname , **UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , **UpperCamelCase__ : List[Any] ): kwargs.update(self.special_tokens_map ) return GPTaTokenizerFast.from_pretrained(self.tmpdirname , **UpperCamelCase__ ) def UpperCamelCase ( self : str , UpperCamelCase__ : str ): A = 'lower newer' A = 'lower newer' return input_text, output_text def UpperCamelCase ( self : Optional[int] ): A = GPTaTokenizer(self.vocab_file , self.merges_file , **self.special_tokens_map ) A = 'lower newer' A = ['\u0120low', 'er', '\u0120', 'n', 'e', 'w', 'er'] A = tokenizer.tokenize(UpperCamelCase__ , add_prefix_space=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) A = tokens + [tokenizer.unk_token] A = [14, 15, 10, 9, 3, 2, 15, 19] self.assertListEqual(tokenizer.convert_tokens_to_ids(UpperCamelCase__ ) , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] ): if not self.test_rust_tokenizer: return A = self.get_tokenizer() A = self.get_rust_tokenizer(add_prefix_space=UpperCamelCase__ ) A = 'lower newer' # Testing tokenization A = tokenizer.tokenize(UpperCamelCase__ , add_prefix_space=UpperCamelCase__ ) A = rust_tokenizer.tokenize(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) # Testing conversion to ids without special tokens A = tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , add_prefix_space=UpperCamelCase__ ) A = rust_tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) # Testing conversion to ids with special tokens A = self.get_rust_tokenizer(add_prefix_space=UpperCamelCase__ ) A = tokenizer.encode(UpperCamelCase__ , add_prefix_space=UpperCamelCase__ ) A = rust_tokenizer.encode(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ ) # Testing the unknown token A = tokens + [rust_tokenizer.unk_token] A = [14, 15, 10, 9, 3, 2, 15, 19] self.assertListEqual(rust_tokenizer.convert_tokens_to_ids(UpperCamelCase__ ) , UpperCamelCase__ ) def UpperCamelCase ( self : Any , *UpperCamelCase__ : List[str] , **UpperCamelCase__ : List[str] ): # It's very difficult to mix/test pretokenization with byte-level # And get both GPT2 and Roberta to work at the same time (mostly an issue of adding a space before the string) pass def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Optional[Any]=15 ): 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(UpperCamelCase__ , **UpperCamelCase__ ) # 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(UpperCamelCase__ , tokenizer_r.encode , UpperCamelCase__ , max_length=UpperCamelCase__ , padding='max_length' ) # Simple input self.assertRaises(UpperCamelCase__ , tokenizer_r.encode_plus , UpperCamelCase__ , max_length=UpperCamelCase__ , padding='max_length' ) # Simple input self.assertRaises( UpperCamelCase__ , tokenizer_r.batch_encode_plus , UpperCamelCase__ , max_length=UpperCamelCase__ , padding='max_length' , ) # Pair input self.assertRaises(UpperCamelCase__ , tokenizer_r.encode , UpperCamelCase__ , max_length=UpperCamelCase__ , padding='max_length' ) # Pair input self.assertRaises(UpperCamelCase__ , tokenizer_r.encode_plus , UpperCamelCase__ , max_length=UpperCamelCase__ , padding='max_length' ) # Pair input self.assertRaises( UpperCamelCase__ , tokenizer_r.batch_encode_plus , UpperCamelCase__ , max_length=UpperCamelCase__ , padding='max_length' , ) def UpperCamelCase ( self : List[Any] ): A = GPTaTokenizer.from_pretrained(self.tmpdirname , pad_token='<pad>' ) # Simple input A = 'This is a simple input' A = ['This is a simple input looooooooong', 'This is a simple input'] A = ('This is a simple input', 'This is a pair') A = [ ('This is a simple input loooooong', 'This is a simple input'), ('This is a simple pair loooooong', 'This is a simple pair'), ] A = tokenizer.pad_token_id A = tokenizer(UpperCamelCase__ , padding='max_length' , max_length=30 , return_tensors='np' ) A = tokenizer(UpperCamelCase__ , padding=UpperCamelCase__ , truncate=UpperCamelCase__ , return_tensors='np' ) A = tokenizer(*UpperCamelCase__ , padding='max_length' , max_length=60 , return_tensors='np' ) A = tokenizer(UpperCamelCase__ , padding=UpperCamelCase__ , truncate=UpperCamelCase__ , return_tensors='np' ) # s # test single string max_length padding self.assertEqual(out_s['input_ids'].shape[-1] , 30 ) self.assertTrue(pad_token_id in out_s['input_ids'] ) self.assertTrue(0 in out_s['attention_mask'] ) # s2 # test automatic padding self.assertEqual(out_sa['input_ids'].shape[-1] , 33 ) # long slice doesn't have padding self.assertFalse(pad_token_id in out_sa['input_ids'][0] ) self.assertFalse(0 in out_sa['attention_mask'][0] ) # short slice does have padding self.assertTrue(pad_token_id in out_sa['input_ids'][1] ) self.assertTrue(0 in out_sa['attention_mask'][1] ) # p # test single pair max_length padding self.assertEqual(out_p['input_ids'].shape[-1] , 60 ) self.assertTrue(pad_token_id in out_p['input_ids'] ) self.assertTrue(0 in out_p['attention_mask'] ) # p2 # test automatic padding pair self.assertEqual(out_pa['input_ids'].shape[-1] , 52 ) # long slice pair doesn't have padding self.assertFalse(pad_token_id in out_pa['input_ids'][0] ) self.assertFalse(0 in out_pa['attention_mask'][0] ) # short slice pair does have padding self.assertTrue(pad_token_id in out_pa['input_ids'][1] ) self.assertTrue(0 in out_pa['attention_mask'][1] ) def UpperCamelCase ( self : int ): A = '$$$' A = GPTaTokenizer.from_pretrained(self.tmpdirname , bos_token=UpperCamelCase__ , add_bos_token=UpperCamelCase__ ) A = 'This is a simple input' A = ['This is a simple input 1', 'This is a simple input 2'] A = tokenizer.bos_token_id A = tokenizer(UpperCamelCase__ ) A = tokenizer(UpperCamelCase__ ) self.assertEqual(out_s.input_ids[0] , UpperCamelCase__ ) self.assertTrue(all(o[0] == bos_token_id for o in out_sa.input_ids ) ) A = tokenizer.decode(out_s.input_ids ) A = tokenizer.batch_decode(out_sa.input_ids ) self.assertEqual(decode_s.split()[0] , UpperCamelCase__ ) self.assertTrue(all(d.split()[0] == bos_token for d in decode_sa ) ) def UpperCamelCase ( self : List[Any] ): pass def UpperCamelCase ( self : Dict ): # TODO: change to self.get_tokenizers() when the fast version is implemented A = [self.get_tokenizer(do_lower_case=UpperCamelCase__ , add_bos_token=UpperCamelCase__ )] for tokenizer in tokenizers: with self.subTest(f'''{tokenizer.__class__.__name__}''' ): A = 'Encode this.' A = 'This one too please.' A = tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) encoded_sequence += tokenizer.encode(UpperCamelCase__ , add_special_tokens=UpperCamelCase__ ) A = tokenizer.encode_plus( UpperCamelCase__ , UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , return_special_tokens_mask=UpperCamelCase__ , ) A = encoded_sequence_dict['input_ids'] A = encoded_sequence_dict['special_tokens_mask'] self.assertEqual(len(UpperCamelCase__ ) , len(UpperCamelCase__ ) ) A = [ (x if not special_tokens_mask[i] else None) for i, x in enumerate(UpperCamelCase__ ) ] A = [x for x in filtered_sequence if x is not None] self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) @require_tokenizers class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : Optional[Any] ): # More context: # https://huggingface.co/wjmcat/opt-350m-paddle/discussions/1 # https://huggingface.slack.com/archives/C01N44FJDHT/p1653511495183519 # https://github.com/huggingface/transformers/pull/17088#discussion_r871246439 A = AutoTokenizer.from_pretrained('facebook/opt-350m' , from_slow=UpperCamelCase__ ) A = 'A photo of a cat' A = tokenizer.encode( UpperCamelCase__ , ) self.assertEqual(UpperCamelCase__ , [2, 250, 1345, 9, 10, 4758] ) tokenizer.save_pretrained('test_opt' ) A = AutoTokenizer.from_pretrained('./test_opt' ) A = tokenizer.encode( UpperCamelCase__ , ) self.assertEqual(UpperCamelCase__ , [2, 250, 1345, 9, 10, 4758] ) def UpperCamelCase ( self : str ): A = AutoTokenizer.from_pretrained('facebook/opt-350m' , use_slow=UpperCamelCase__ ) A = 'A photo of a cat' A = tokenizer.encode( UpperCamelCase__ , ) # Same as above self.assertEqual(UpperCamelCase__ , [2, 250, 1345, 9, 10, 4758] ) @unittest.skip('This test is failing because of a bug in the fast tokenizer' ) def UpperCamelCase ( self : List[str] ): A = AutoTokenizer.from_pretrained('facebook/opt-350m' , from_slow=UpperCamelCase__ ) A = 'bos' A = tokenizer.get_vocab()['bos'] A = 'A photo of a cat' A = tokenizer.encode( UpperCamelCase__ , ) # We changed the bos token self.assertEqual(UpperCamelCase__ , [31957, 250, 1345, 9, 10, 4758] ) tokenizer.save_pretrained('./tok' ) A = AutoTokenizer.from_pretrained('./tok' ) self.assertTrue(tokenizer.is_fast ) A = tokenizer.encode( UpperCamelCase__ , ) self.assertEqual(UpperCamelCase__ , [31957, 250, 1345, 9, 10, 4758] )
699
def __UpperCamelCase (lowerCAmelCase : list[int] ) -> int: if not numbers: return 0 if not isinstance(lowerCAmelCase, (list, tuple) ) or not all( isinstance(lowerCAmelCase, lowerCAmelCase ) for number in numbers ): raise ValueError('numbers must be an iterable of integers' ) A = A = A = numbers[0] for i in range(1, len(lowerCAmelCase ) ): # update the maximum and minimum subarray products A = numbers[i] if number < 0: A , A = min_till_now, max_till_now A = max(lowerCAmelCase, max_till_now * number ) A = min(lowerCAmelCase, min_till_now * number ) # update the maximum product found till now A = max(lowerCAmelCase, lowerCAmelCase ) return max_prod
699
1
from .testing import ( are_the_same_tensors, execute_subprocess_async, require_bnb, require_cpu, require_cuda, require_huggingface_suite, require_mps, require_multi_gpu, require_multi_xpu, require_safetensors, require_single_gpu, require_single_xpu, require_torch_min_version, require_tpu, require_xpu, skip, slow, ) from .training import RegressionDataset, RegressionModel, RegressionModelaXPU from .scripts import test_script, test_sync, test_ops # isort: skip
699
from .glue import glue_convert_examples_to_features, glue_output_modes, glue_processors, glue_tasks_num_labels from .squad import SquadExample, SquadFeatures, SquadVaProcessor, SquadVaProcessor, squad_convert_examples_to_features from .utils import DataProcessor, InputExample, InputFeatures, SingleSentenceClassificationProcessor from .xnli import xnli_output_modes, xnli_processors, xnli_tasks_num_labels
699
1
def __UpperCamelCase (lowerCAmelCase : int = 50 ) -> int: A = [[0] * 3 for _ in range(length + 1 )] for row_length in range(length + 1 ): for tile_length in range(2, 5 ): for tile_start in range(row_length - tile_length + 1 ): different_colour_ways_number[row_length][tile_length - 2] += ( different_colour_ways_number[row_length - tile_start - tile_length][ tile_length - 2 ] + 1 ) return sum(different_colour_ways_number[length] ) if __name__ == "__main__": print(F'''{solution() = }''')
699
from ...configuration_utils import PretrainedConfig from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "microsoft/biogpt": "https://huggingface.co/microsoft/biogpt/resolve/main/config.json", # See all BioGPT models at https://huggingface.co/models?filter=biogpt } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = '''biogpt''' def __init__( self : Optional[Any] , UpperCamelCase__ : str=42384 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : Dict=24 , UpperCamelCase__ : Any=16 , UpperCamelCase__ : str=4096 , UpperCamelCase__ : Tuple="gelu" , UpperCamelCase__ : List[str]=0.1 , UpperCamelCase__ : Any=0.1 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : List[Any]=0.02 , UpperCamelCase__ : Dict=1e-1_2 , UpperCamelCase__ : Any=True , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Any=1 , UpperCamelCase__ : List[str]=0 , UpperCamelCase__ : Optional[Any]=2 , **UpperCamelCase__ : List[Any] , ): A = vocab_size A = max_position_embeddings A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = initializer_range A = layer_norm_eps A = scale_embedding A = use_cache A = layerdrop A = activation_dropout super().__init__(pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , **UpperCamelCase__ )
699
1
from __future__ import annotations def __UpperCamelCase (lowerCAmelCase : dict, lowerCAmelCase : str ) -> set[str]: A , A = set(lowerCAmelCase ), [start] while stack: A = stack.pop() explored.add(lowerCAmelCase ) # Differences from BFS: # 1) pop last element instead of first one # 2) add adjacent elements to stack without exploring them for adj in reversed(graph[v] ): if adj not in explored: stack.append(lowerCAmelCase ) return explored _UpperCAmelCase = { "A": ["B", "C", "D"], "B": ["A", "D", "E"], "C": ["A", "F"], "D": ["B", "D"], "E": ["B", "F"], "F": ["C", "E", "G"], "G": ["F"], } if __name__ == "__main__": import doctest doctest.testmod() print(depth_first_search(G, "A"))
699
import sys def __UpperCamelCase (lowerCAmelCase : Dict ) -> Dict: A = len(lowerCAmelCase ) A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] for chain_length in range(2, lowerCAmelCase ): for a in range(1, n - chain_length + 1 ): A = a + chain_length - 1 A = sys.maxsize for c in range(lowerCAmelCase, lowerCAmelCase ): A = ( matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b] ) if cost < matrix[a][b]: A = cost A = c return matrix, sol def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Union[str, Any], lowerCAmelCase : Union[str, Any] ) -> List[str]: if i == j: print('A' + str(lowerCAmelCase ), end=' ' ) else: print('(', end=' ' ) print_optiomal_solution(lowerCAmelCase, lowerCAmelCase, optimal_solution[i][j] ) print_optiomal_solution(lowerCAmelCase, optimal_solution[i][j] + 1, lowerCAmelCase ) print(')', end=' ' ) def __UpperCamelCase () -> List[str]: A = [30, 35, 15, 5, 10, 20, 25] A = len(lowerCAmelCase ) # Size of matrix created from above array will be # 30*35 35*15 15*5 5*10 10*20 20*25 A , A = matrix_chain_order(lowerCAmelCase ) print('No. of Operation required: ' + str(matrix[1][n - 1] ) ) print_optiomal_solution(lowerCAmelCase, 1, n - 1 ) if __name__ == "__main__": main()
699
1
# Copyright 2023 The HuggingFace Team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. from typing import TYPE_CHECKING from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available _UpperCAmelCase = { "configuration_mgp_str": ["MGP_STR_PRETRAINED_CONFIG_ARCHIVE_MAP", "MgpstrConfig"], "processing_mgp_str": ["MgpstrProcessor"], "tokenization_mgp_str": ["MgpstrTokenizer"], } try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "MGP_STR_PRETRAINED_MODEL_ARCHIVE_LIST", "MgpstrModel", "MgpstrPreTrainedModel", "MgpstrForSceneTextRecognition", ] if TYPE_CHECKING: from .configuration_mgp_str import MGP_STR_PRETRAINED_CONFIG_ARCHIVE_MAP, MgpstrConfig from .processing_mgp_str import MgpstrProcessor from .tokenization_mgp_str import MgpstrTokenizer try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_mgp_str import ( MGP_STR_PRETRAINED_MODEL_ARCHIVE_LIST, MgpstrForSceneTextRecognition, MgpstrModel, MgpstrPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
699
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> bool: return all(number % divisor != 0 for divisor in range(2, isqrt(lowerCAmelCase ) + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int = 10**6 ) -> int: A = 0 A = 1 A = 7 while prime_candidate < max_prime: primes_count += is_prime(lowerCAmelCase ) cube_index += 1 prime_candidate += 6 * cube_index return primes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
1
import random from typing import Any def __UpperCamelCase (lowerCAmelCase : list ) -> list[Any]: for _ in range(len(lowerCAmelCase ) ): A = random.randint(0, len(lowerCAmelCase ) - 1 ) A = random.randint(0, len(lowerCAmelCase ) - 1 ) A , A = data[b], data[a] return data if __name__ == "__main__": _UpperCAmelCase = [0, 1, 2, 3, 4, 5, 6, 7] _UpperCAmelCase = ["python", "says", "hello", "!"] print("Fisher-Yates Shuffle:") print("List", integers, strings) print("FY Shuffle", fisher_yates_shuffle(integers), fisher_yates_shuffle(strings))
699
import warnings from ...utils import logging from .image_processing_imagegpt import ImageGPTImageProcessor _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , *UpperCamelCase__ : List[Any] , **UpperCamelCase__ : Tuple ): warnings.warn( 'The class ImageGPTFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use ImageGPTImageProcessor instead.' , UpperCamelCase__ , ) super().__init__(*UpperCamelCase__ , **UpperCamelCase__ )
699
1
import argparse import fairseq import torch from transformers import UniSpeechSatConfig, UniSpeechSatForCTC, UniSpeechSatForPreTraining, logging logging.set_verbosity_info() _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "post_extract_proj": "feature_projection.projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.layer_norm": "encoder.layer_norm", "encoder.layer_norm_for_extract": "layer_norm_for_extract", "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": "lm_head", "label_embs_concat": "label_embeddings_concat", "mask_emb": "masked_spec_embed", "spk_proj": "speaker_proj", } _UpperCAmelCase = [ "lm_head", "quantizer.weight_proj", "quantizer.codevectors", "project_q", "project_hid", "label_embeddings_concat", "speaker_proj", "layer_norm_for_extract", ] def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Dict, lowerCAmelCase : Optional[int], lowerCAmelCase : List[Any], lowerCAmelCase : str ) -> int: 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 if hf_shape != value.shape: raise ValueError( 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 __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Optional[int] ) -> Dict: A = [] A = fairseq_model.state_dict() A = hf_model.unispeech_sat.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 = 'unispeech_sat.' + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key if key in name or key.split('w2v_model.' )[-1] == name.split('.' )[0]: if "layer_norm_for_extract" in name and (".".join(name.split('.' )[:-1] ) != key): # special case since naming is very similar continue 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 "bias" in name: A = 'bias' elif "weight" in name: # TODO: don't match quantizer.weight_proj A = 'weight' 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 __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : Tuple, lowerCAmelCase : List[Any], lowerCAmelCase : int ) -> Dict: 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.bias.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.weight.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape: raise ValueError( f'''{full_name} has size {value.shape}, but''' f''' {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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape: raise ValueError( 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 ) @torch.no_grad() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : Dict, lowerCAmelCase : Union[str, Any]=None, lowerCAmelCase : str=None, lowerCAmelCase : List[Any]=True ) -> Union[str, Any]: if config_path is not None: A = UniSpeechSatConfig.from_pretrained(lowerCAmelCase ) else: A = UniSpeechSatConfig() A = '' if is_finetuned: A = UniSpeechSatForCTC(lowerCAmelCase ) else: A = UniSpeechSatForPreTraining(lowerCAmelCase ) A , A , A = fairseq.checkpoint_utils.load_model_ensemble_and_task( [checkpoint_path], arg_overrides={'data': '/'.join(dict_path.split('/' )[:-1] )} ) A = model[0].eval() recursively_load_weights(lowerCAmelCase, lowerCAmelCase ) hf_wavavec.save_pretrained(lowerCAmelCase ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--dict_path", default=None, type=str, help="Path to dict of fine-tuned model") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") parser.add_argument( "--not_finetuned", action="store_true", help="Whether the model to convert is a fine-tuned model or not" ) _UpperCAmelCase = parser.parse_args() convert_unispeech_sat_checkpoint( args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned )
699
from typing import Any, Dict, Optional import torch import torch.nn.functional as F from torch import nn from ..utils import maybe_allow_in_graph from .activations import get_activation from .attention_processor import Attention from .embeddings import CombinedTimestepLabelEmbeddings @maybe_allow_in_graph class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int]=0.0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : str = "layer_norm" , UpperCamelCase__ : bool = False , ): super().__init__() A = only_cross_attention A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm_zero' A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm' if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None: raise ValueError( f'''`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to''' f''' define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}.''' ) # Define 3 blocks. Each block has its own normalization layer. # 1. Self-Attn if self.use_ada_layer_norm: A = AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A = AdaLayerNormZero(UpperCamelCase__ , UpperCamelCase__ ) else: A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = Attention( query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if only_cross_attention else None , upcast_attention=UpperCamelCase__ , ) # 2. Cross-Attn if cross_attention_dim is not None or double_self_attention: # We currently only use AdaLayerNormZero for self attention where there will only be one attention block. # I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during # the second cross attention block. A = ( AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) ) A = Attention( query_dim=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if not double_self_attention else None , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , upcast_attention=UpperCamelCase__ , ) # is self-attn if encoder_hidden_states is none else: A = None A = None # 3. Feed-forward A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = FeedForward(UpperCamelCase__ , dropout=UpperCamelCase__ , activation_fn=UpperCamelCase__ , final_dropout=UpperCamelCase__ ) # let chunk size default to None A = None A = 0 def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int ): # Sets chunk feed-forward A = chunk_size A = dim def UpperCamelCase ( self : Dict , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , UpperCamelCase__ : Dict[str, Any] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , ): # Notice that normalization is always applied before the real computation in the following blocks. # 1. Self-Attention if self.use_ada_layer_norm: A = self.norma(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A , A , A , A , A = self.norma( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=hidden_states.dtype ) else: A = self.norma(UpperCamelCase__ ) A = cross_attention_kwargs if cross_attention_kwargs is not None else {} A = self.attna( UpperCamelCase__ , encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) if self.use_ada_layer_norm_zero: A = gate_msa.unsqueeze(1 ) * attn_output A = attn_output + hidden_states # 2. Cross-Attention if self.attna is not None: A = ( self.norma(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else self.norma(UpperCamelCase__ ) ) A = self.attna( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) A = attn_output + hidden_states # 3. Feed-forward A = self.norma(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None] if self._chunk_size is not None: # "feed_forward_chunk_size" can be used to save memory if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0: raise ValueError( f'''`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`.''' ) A = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size A = torch.cat( [self.ff(UpperCamelCase__ ) for hid_slice in norm_hidden_states.chunk(UpperCamelCase__ , dim=self._chunk_dim )] , dim=self._chunk_dim , ) else: A = self.ff(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = gate_mlp.unsqueeze(1 ) * ff_output A = ff_output + hidden_states return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 4 , UpperCamelCase__ : float = 0.0 , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : bool = False , ): super().__init__() A = int(dim * mult ) A = dim_out if dim_out is not None else dim if activation_fn == "gelu": A = GELU(UpperCamelCase__ , UpperCamelCase__ ) if activation_fn == "gelu-approximate": A = GELU(UpperCamelCase__ , UpperCamelCase__ , approximate='tanh' ) elif activation_fn == "geglu": A = GEGLU(UpperCamelCase__ , UpperCamelCase__ ) elif activation_fn == "geglu-approximate": A = ApproximateGELU(UpperCamelCase__ , UpperCamelCase__ ) A = nn.ModuleList([] ) # project in self.net.append(UpperCamelCase__ ) # project dropout self.net.append(nn.Dropout(UpperCamelCase__ ) ) # project out self.net.append(nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) ) # FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout if final_dropout: self.net.append(nn.Dropout(UpperCamelCase__ ) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int ): for module in self.net: A = module(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : str = "none" ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) A = approximate def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ , approximate=self.approximate ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) , approximate=self.approximate ).to(dtype=gate.dtype ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int ): A = self.proj(UpperCamelCase__ ) A = self.gelu(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , dim_out * 2 ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) ).to(dtype=gate.dtype ) def UpperCamelCase ( self : str , UpperCamelCase__ : str ): A , A = self.proj(UpperCamelCase__ ).chunk(2 , dim=-1 ) return hidden_states * self.gelu(UpperCamelCase__ ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[int] ): A = self.proj(UpperCamelCase__ ) return x * torch.sigmoid(1.702 * x ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Tuple ): super().__init__() A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , embedding_dim * 2 ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[Any] ): A = self.linear(self.silu(self.emb(UpperCamelCase__ ) ) ) A , A = torch.chunk(UpperCamelCase__ , 2 ) A = self.norm(UpperCamelCase__ ) * (1 + scale) + shift return x class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : int , UpperCamelCase__ : List[str] ): super().__init__() A = CombinedTimestepLabelEmbeddings(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , 6 * embedding_dim , bias=UpperCamelCase__ ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ , eps=1e-6 ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : str , UpperCamelCase__ : Tuple=None ): A = self.linear(self.silu(self.emb(UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=UpperCamelCase__ ) ) ) A , A , A , A , A , A = emb.chunk(6 , dim=1 ) A = self.norm(UpperCamelCase__ ) * (1 + scale_msa[:, None]) + shift_msa[:, None] return x, gate_msa, shift_mlp, scale_mlp, gate_mlp class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[str] = None , UpperCamelCase__ : float = 1e-5 ): super().__init__() A = num_groups A = eps if act_fn is None: A = None else: A = get_activation(UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , out_dim * 2 ) def UpperCamelCase ( self : Any , UpperCamelCase__ : str , UpperCamelCase__ : str ): if self.act: A = self.act(UpperCamelCase__ ) A = self.linear(UpperCamelCase__ ) A = emb[:, :, None, None] A , A = emb.chunk(2 , dim=1 ) A = F.group_norm(UpperCamelCase__ , self.num_groups , eps=self.eps ) A = x * (1 + scale) + shift return x
699
1
from argparse import ArgumentParser from .add_new_model import AddNewModelCommand from .add_new_model_like import AddNewModelLikeCommand from .convert import ConvertCommand from .download import DownloadCommand from .env import EnvironmentCommand from .lfs import LfsCommands from .pt_to_tf import PTtoTFCommand from .run import RunCommand from .serving import ServeCommand from .user import UserCommands def __UpperCamelCase () -> str: A = ArgumentParser('Transformers CLI tool', usage='transformers-cli <command> [<args>]' ) A = parser.add_subparsers(help='transformers-cli command helpers' ) # Register commands ConvertCommand.register_subcommand(lowerCAmelCase ) DownloadCommand.register_subcommand(lowerCAmelCase ) EnvironmentCommand.register_subcommand(lowerCAmelCase ) RunCommand.register_subcommand(lowerCAmelCase ) ServeCommand.register_subcommand(lowerCAmelCase ) UserCommands.register_subcommand(lowerCAmelCase ) AddNewModelCommand.register_subcommand(lowerCAmelCase ) AddNewModelLikeCommand.register_subcommand(lowerCAmelCase ) LfsCommands.register_subcommand(lowerCAmelCase ) PTtoTFCommand.register_subcommand(lowerCAmelCase ) # Let's go A = parser.parse_args() if not hasattr(lowerCAmelCase, 'func' ): parser.print_help() exit(1 ) # Run A = args.func(lowerCAmelCase ) service.run() if __name__ == "__main__": main()
699
import json import os from typing import Dict, List, Optional, Tuple from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "vocab_file": "vocab.json", "tokenizer_config_file": "tokenizer_config.json", "merges_file": "merges.txt", } _UpperCAmelCase = { "vocab_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/vocab.json" ), }, "tokenizer_config_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/tokenizer_config.json" ), }, "merges_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/merges.txt" ), }, } _UpperCAmelCase = "</w>" _UpperCAmelCase = "@@ " def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> List[str]: A = set() A = word[0] for char in word[1:]: pairs.add((prev_char, char) ) A = char return pairs # Speech2Text2 has no max input length _UpperCAmelCase = {"facebook/s2t-wav2vec2-large-en-de": 1_024} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] def __init__( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="<pad>" , UpperCamelCase__ : int="</s>" , UpperCamelCase__ : Tuple="<unk>" , UpperCamelCase__ : List[str]=False , UpperCamelCase__ : List[str]=None , **UpperCamelCase__ : Optional[int] , ): super().__init__( unk_token=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , **UpperCamelCase__ , ) A = do_lower_case with open(UpperCamelCase__ , encoding='utf-8' ) as vocab_handle: A = json.load(UpperCamelCase__ ) A = {v: k for k, v in self.encoder.items()} if merges_file is None: logger.info(f'''No merges files provided. {self.__class__.__name__} can only be used for decoding.''' ) A = None A = None else: with open(UpperCamelCase__ , encoding='utf-8' ) as merges_handle: A = merges_handle.read().split('\n' )[:-1] A = [tuple(merge.split()[:2] ) for merge in merges] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = {} @property def UpperCamelCase ( self : Union[str, Any] ): return len(self.decoder ) def UpperCamelCase ( self : Optional[Any] ): return dict(self.encoder , **self.added_tokens_encoder ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): A = tuple(token[:-1] ) + (token[-1] + BPE_TOKEN_MERGES,) if token in self.cache: return self.cache[token] A = get_pairs(UpperCamelCase__ ) if not pairs: return token while True: A = min(UpperCamelCase__ , key=lambda UpperCamelCase__ : self.bpe_ranks.get(UpperCamelCase__ , float('inf' ) ) ) if bigram not in self.bpe_ranks: break A , A = bigram A = [] A = 0 while i < len(UpperCamelCase__ ): try: A = word.index(UpperCamelCase__ , UpperCamelCase__ ) except ValueError: new_word.extend(word[i:] ) break else: new_word.extend(word[i:j] ) A = j if word[i] == first and i < len(UpperCamelCase__ ) - 1 and word[i + 1] == second: new_word.append(first + second ) i += 2 else: new_word.append(word[i] ) i += 1 A = tuple(UpperCamelCase__ ) A = new_word if len(UpperCamelCase__ ) == 1: break else: A = get_pairs(UpperCamelCase__ ) A = ' '.join(UpperCamelCase__ ) if word == "\n " + BPE_TOKEN_MERGES: A = '\n' + BPE_TOKEN_MERGES if word.endswith(UpperCamelCase__ ): A = word.replace(UpperCamelCase__ , '' ) A = word.replace(' ' , UpperCamelCase__ ) A = word return word def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict ): if self.bpe_ranks is None: raise ValueError( 'This tokenizer was instantiated without a `merges.txt` file, so' ' that it can only be used for decoding, not for encoding.' 'Make sure to provide `merges.txt` file at instantiation to enable ' 'encoding.' ) if self.do_lower_case: A = text.lower() A = text.split() A = [] for token in text: if token: split_tokens.extend(list(self.bpe(UpperCamelCase__ ).split(' ' ) ) ) return split_tokens def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : str ): return self.encoder.get(UpperCamelCase__ , self.encoder.get(self.unk_token ) ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): A = self.decoder.get(UpperCamelCase__ , self.unk_token ) return result def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] ): A = ' '.join(UpperCamelCase__ ) # make sure @@ tokens are concatenated A = ''.join(string.split(UpperCamelCase__ ) ) return string def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] ) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as f: f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCamelCase__ , ensure_ascii=UpperCamelCase__ ) + '\n' ) A = 0 if self.bpe_ranks is None: return (vocab_file,) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as writer: for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCamelCase__ : kv[1] ): if index != token_index: logger.warning( f'''Saving vocabulary to {merges_file}: BPE merge indices are not consecutive.''' ' Please check that the tokenizer is not corrupted!' ) A = token_index writer.write(' '.join(UpperCamelCase__ ) + '\n' ) index += 1 return (vocab_file, merges_file)
699
1
import time import warnings from abc import ABC from copy import deepcopy from typing import Optional import torch from ..utils import add_start_docstrings, logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = R"\n Args:\n input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):\n Indices of input sequence tokens in the vocabulary.\n\n Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and\n [`PreTrainedTokenizer.__call__`] for details.\n\n [What are input IDs?](../glossary#input-ids)\n scores (`torch.FloatTensor` of shape `(batch_size, config.vocab_size)`):\n Prediction scores of a language modeling head. These can be scores for each vocabulary token before SoftMax\n or scores for each vocabulary token after SoftMax.\n kwargs (`Dict[str, Any]`, *optional*):\n Additional stopping criteria specific kwargs.\n\n Return:\n `bool`. `False` indicates we should continue, `True` indicates we should stop.\n\n" class _UpperCAmelCase ( __lowercase ): '''simple docstring''' @add_start_docstrings(UpperCamelCase__ ) def __call__( self : Dict , UpperCamelCase__ : torch.LongTensor , UpperCamelCase__ : torch.FloatTensor , **UpperCamelCase__ : Any ): raise NotImplementedError('StoppingCriteria needs to be subclassed' ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] = None ): A = max_length A = max_position_embeddings @add_start_docstrings(UpperCamelCase__ ) def __call__( self : List[str] , UpperCamelCase__ : torch.LongTensor , UpperCamelCase__ : torch.FloatTensor , **UpperCamelCase__ : Optional[int] ): A = input_ids.shape[-1] A = cur_len >= self.max_length if self.max_position_embeddings is not None and not is_done and cur_len >= self.max_position_embeddings: logger.warning_once( 'This is a friendly reminder - the current text generation call will exceed the model\'s predefined ' f'''maximum length ({self.max_position_embeddings}). Depending on the model, you may observe ''' 'exceptions, performance degradation, or nothing at all.' ) return is_done class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): warnings.warn( 'The class `MaxNewTokensCriteria` is deprecated. ' f'''Please use `MaxLengthCriteria(max_length={start_length + max_new_tokens})` ''' 'with `max_length = start_length + max_new_tokens` instead.' , UpperCamelCase__ , ) A = start_length A = max_new_tokens A = start_length + max_new_tokens @add_start_docstrings(UpperCamelCase__ ) def __call__( self : Tuple , UpperCamelCase__ : torch.LongTensor , UpperCamelCase__ : torch.FloatTensor , **UpperCamelCase__ : int ): return input_ids.shape[-1] >= self.max_length class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : Union[str, Any] , UpperCamelCase__ : float , UpperCamelCase__ : Optional[float] = None ): A = max_time A = time.time() if initial_timestamp is None else initial_timestamp @add_start_docstrings(UpperCamelCase__ ) def __call__( self : Dict , UpperCamelCase__ : torch.LongTensor , UpperCamelCase__ : torch.FloatTensor , **UpperCamelCase__ : Tuple ): return time.time() - self.initial_timestamp > self.max_time class _UpperCAmelCase ( __lowercase ): '''simple docstring''' @add_start_docstrings(UpperCamelCase__ ) def __call__( self : int , UpperCamelCase__ : torch.LongTensor , UpperCamelCase__ : torch.FloatTensor , **UpperCamelCase__ : List[Any] ): return any(criteria(UpperCamelCase__ , UpperCamelCase__ ) for criteria in self ) @property def UpperCamelCase ( self : List[str] ): for stopping_criterium in self: if isinstance(UpperCamelCase__ , UpperCamelCase__ ): return stopping_criterium.max_length elif isinstance(UpperCamelCase__ , UpperCamelCase__ ): return stopping_criterium.max_length return None def __UpperCamelCase (lowerCAmelCase : StoppingCriteriaList, lowerCAmelCase : int ) -> StoppingCriteriaList: A = stopping_criteria.max_length A = deepcopy(lowerCAmelCase ) if stopping_max_length is not None and stopping_max_length != max_length: warnings.warn('You set different `max_length` for stopping criteria and `max_length` parameter', lowerCAmelCase ) elif stopping_max_length is None: new_stopping_criteria.append(MaxLengthCriteria(max_length=lowerCAmelCase ) ) return new_stopping_criteria
699
# Copyright 2023 The HuggingFace Inc. team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import torch from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer from .base import PipelineTool class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = '''facebook/bart-large-mnli''' SCREAMING_SNAKE_CASE : Union[str, Any] = ( '''This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which ''' '''should be the text to classify, and `labels`, which should be the list of labels to use for classification. ''' '''It returns the most likely label in the list of provided `labels` for the input text.''' ) SCREAMING_SNAKE_CASE : Any = '''text_classifier''' SCREAMING_SNAKE_CASE : Any = AutoTokenizer SCREAMING_SNAKE_CASE : Dict = AutoModelForSequenceClassification SCREAMING_SNAKE_CASE : List[Any] = ['''text''', ['''text''']] SCREAMING_SNAKE_CASE : Dict = ['''text'''] def UpperCamelCase ( self : List[str] ): super().setup() A = self.model.config A = -1 for idx, label in config.idalabel.items(): if label.lower().startswith('entail' ): A = int(UpperCamelCase__ ) if self.entailment_id == -1: raise ValueError('Could not determine the entailment ID from the model config, please pass it at init.' ) def UpperCamelCase ( self : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict ): A = labels return self.pre_processor( [text] * len(UpperCamelCase__ ) , [f'''This example is {label}''' for label in labels] , return_tensors='pt' , padding='max_length' , ) def UpperCamelCase ( self : int , UpperCamelCase__ : List[str] ): A = outputs.logits A = torch.argmax(logits[:, 2] ).item() return self._labels[label_id]
699
1
# tests directory-specific settings - this file is run automatically # by pytest before any tests are run import sys import warnings from os.path import abspath, dirname, join # allow having multiple repository checkouts and not needing to remember to rerun # 'pip install -e .[dev]' when switching between checkouts and running tests. _UpperCAmelCase = abspath(join(dirname(dirname(__file__)), "src")) sys.path.insert(1, git_repo_path) # silence FutureWarning warnings in tests since often we can't act on them until # they become normal warnings - i.e. the tests still need to test the current functionality warnings.simplefilter(action="ignore", category=FutureWarning) def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> Optional[int]: from diffusers.utils.testing_utils import pytest_addoption_shared pytest_addoption_shared(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : Any ) -> str: from diffusers.utils.testing_utils import pytest_terminal_summary_main A = terminalreporter.config.getoption('--make-reports' ) if make_reports: pytest_terminal_summary_main(lowerCAmelCase, id=lowerCAmelCase )
699
import re import jax.numpy as jnp from flax.traverse_util import flatten_dict, unflatten_dict from jax.random import PRNGKey from ..utils import logging _UpperCAmelCase = logging.get_logger(__name__) def __UpperCamelCase (lowerCAmelCase : List[str] ) -> Dict: A = r'\w+[.]\d+' A = re.findall(lowerCAmelCase, lowerCAmelCase ) for pat in pats: A = key.replace(lowerCAmelCase, '_'.join(pat.split('.' ) ) ) return key def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Dict, lowerCAmelCase : Dict ) -> Any: A = pt_tuple_key[:-1] + ('scale',) if ( any('norm' in str_ for str_ in pt_tuple_key ) and (pt_tuple_key[-1] == "bias") and (pt_tuple_key[:-1] + ("bias",) not in random_flax_state_dict) and (pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict) ): A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor elif pt_tuple_key[-1] in ["weight", "gamma"] and pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor # embedding if pt_tuple_key[-1] == "weight" and pt_tuple_key[:-1] + ("embedding",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('embedding',) return renamed_pt_tuple_key, pt_tensor # conv layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight" and pt_tensor.ndim == 4: A = pt_tensor.transpose(2, 3, 1, 0 ) return renamed_pt_tuple_key, pt_tensor # linear layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight": A = pt_tensor.T return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm weight A = pt_tuple_key[:-1] + ('weight',) if pt_tuple_key[-1] == "gamma": return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm bias A = pt_tuple_key[:-1] + ('bias',) if pt_tuple_key[-1] == "beta": return renamed_pt_tuple_key, pt_tensor return pt_tuple_key, pt_tensor def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Any, lowerCAmelCase : str=42 ) -> Any: # Step 1: Convert pytorch tensor to numpy A = {k: v.numpy() for k, v in pt_state_dict.items()} # Step 2: Since the model is stateless, get random Flax params A = flax_model.init_weights(PRNGKey(lowerCAmelCase ) ) A = flatten_dict(lowerCAmelCase ) A = {} # Need to change some parameters name to match Flax names for pt_key, pt_tensor in pt_state_dict.items(): A = rename_key(lowerCAmelCase ) A = tuple(renamed_pt_key.split('.' ) ) # Correctly rename weight parameters A , A = rename_key_and_reshape_tensor(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if flax_key in random_flax_state_dict: if flax_tensor.shape != random_flax_state_dict[flax_key].shape: raise ValueError( f'''PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape ''' f'''{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.''' ) # also add unexpected weight so that warning is thrown A = jnp.asarray(lowerCAmelCase ) return unflatten_dict(lowerCAmelCase )
699
1
import math import torch from torch import nn from ..configuration_utils import ConfigMixin, register_to_config from .attention_processor import Attention from .embeddings import get_timestep_embedding from .modeling_utils import ModelMixin class _UpperCAmelCase ( __lowercase , __lowercase ): '''simple docstring''' @register_to_config def __init__( self : Any , UpperCamelCase__ : int = 128 , UpperCamelCase__ : int = 256 , UpperCamelCase__ : float = 2_000.0 , UpperCamelCase__ : int = 768 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 64 , UpperCamelCase__ : int = 2048 , UpperCamelCase__ : float = 0.1 , ): super().__init__() A = nn.Sequential( nn.Linear(UpperCamelCase__ , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , nn.Linear(d_model * 4 , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , ) A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = False A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.ModuleList() for lyr_num in range(UpperCamelCase__ ): # FiLM conditional T5 decoder A = DecoderLayer(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) self.decoders.append(UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Any , UpperCamelCase__ : int ): A = torch.mul(query_input.unsqueeze(-1 ) , key_input.unsqueeze(-2 ) ) return mask.unsqueeze(-3 ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[int] ): A , A , A = decoder_input_tokens.shape assert decoder_noise_time.shape == (batch,) # decoder_noise_time is in [0, 1), so rescale to expected timing range. A = get_timestep_embedding( decoder_noise_time * self.config.max_decoder_noise_time , embedding_dim=self.config.d_model , max_period=self.config.max_decoder_noise_time , ).to(dtype=self.dtype ) A = self.conditioning_emb(UpperCamelCase__ ).unsqueeze(1 ) assert conditioning_emb.shape == (batch, 1, self.config.d_model * 4) A = decoder_input_tokens.shape[1] # If we want to use relative positions for audio context, we can just offset # this sequence by the length of encodings_and_masks. A = torch.broadcast_to( torch.arange(UpperCamelCase__ , device=decoder_input_tokens.device ) , (batch, seq_length) , ) A = self.position_encoding(UpperCamelCase__ ) A = self.continuous_inputs_projection(UpperCamelCase__ ) inputs += position_encodings A = self.dropout(UpperCamelCase__ ) # decoder: No padding present. A = torch.ones( decoder_input_tokens.shape[:2] , device=decoder_input_tokens.device , dtype=inputs.dtype ) # Translate encoding masks to encoder-decoder masks. A = [(x, self.encoder_decoder_mask(UpperCamelCase__ , UpperCamelCase__ )) for x, y in encodings_and_masks] # cross attend style: concat encodings A = torch.cat([x[0] for x in encodings_and_encdec_masks] , dim=1 ) A = torch.cat([x[1] for x in encodings_and_encdec_masks] , dim=-1 ) for lyr in self.decoders: A = lyr( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , )[0] A = self.decoder_norm(UpperCamelCase__ ) A = self.post_dropout(UpperCamelCase__ ) A = self.spec_out(UpperCamelCase__ ) return spec_out class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Union[str, Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Union[str, Any]=1e-6 ): super().__init__() A = nn.ModuleList() # cond self attention: layer 0 self.layer.append( TaLayerSelfAttentionCond(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) ) # cross attention: layer 1 self.layer.append( TaLayerCrossAttention( d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ , ) ) # Film Cond MLP + dropout: last layer self.layer.append( TaLayerFFCond(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ ) ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : int=None , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[Any]=None , ): A = self.layer[0]( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) if encoder_hidden_states is not None: A = torch.where(encoder_attention_mask > 0 , 0 , -1e1_0 ).to( encoder_hidden_states.dtype ) A = self.layer[1]( UpperCamelCase__ , key_value_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) # Apply Film Conditional Feed Forward layer A = self.layer[-1](UpperCamelCase__ , UpperCamelCase__ ) return (hidden_states,) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Dict ): super().__init__() A = TaLayerNorm(UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : str , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Tuple=None , ): # pre_self_attention_layer_norm A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.FiLMLayer(UpperCamelCase__ , UpperCamelCase__ ) # Self-attention block A = self.attention(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[Any] ): super().__init__() A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[str]=None , ): A = self.layer_norm(UpperCamelCase__ ) A = self.attention( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=attention_mask.squeeze(1 ) , ) A = hidden_states + self.dropout(UpperCamelCase__ ) return layer_output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any ): super().__init__() A = TaDenseGatedActDense(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any=None ): A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.film(UpperCamelCase__ , UpperCamelCase__ ) A = self.DenseReluDense(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) A = NewGELUActivation() def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : List[Any] ): A = self.act(self.wi_a(UpperCamelCase__ ) ) A = self.wi_a(UpperCamelCase__ ) A = hidden_gelu * hidden_linear A = self.dropout(UpperCamelCase__ ) A = self.wo(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Tuple=1e-6 ): super().__init__() A = nn.Parameter(torch.ones(UpperCamelCase__ ) ) A = eps def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int ): # T5 uses a layer_norm which only scales and doesn't shift, which is also known as Root Mean # Square Layer Normalization https://arxiv.org/abs/1910.07467 thus variance is calculated # w/o mean and there is no bias. Additionally we want to make sure that the accumulation for # half-precision inputs is done in fp32 A = hidden_states.to(torch.floataa ).pow(2 ).mean(-1 , keepdim=UpperCamelCase__ ) A = hidden_states * torch.rsqrt(variance + self.variance_epsilon ) # convert into half-precision if necessary if self.weight.dtype in [torch.floataa, torch.bfloataa]: A = hidden_states.to(self.weight.dtype ) return self.weight * hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def UpperCamelCase ( self : Any , UpperCamelCase__ : torch.Tensor ): return 0.5 * input * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi ) * (input + 0.044_715 * torch.pow(UpperCamelCase__ , 3.0 )) )) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , out_features * 2 , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : List[Any] ): A = self.scale_bias(UpperCamelCase__ ) A , A = torch.chunk(UpperCamelCase__ , 2 , -1 ) A = x * (1 + scale) + shift return x
699
from __future__ import annotations import math import random from collections.abc import Collection from typing import overload class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Collection[float] | None = None ): if components is None: A = [] A = list(UpperCamelCase__ ) def __len__( self : List[Any] ): return len(self.__components ) def __str__( self : str ): return "(" + ",".join(map(UpperCamelCase__ , self.__components ) ) + ")" def __add__( self : str , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] + other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: raise Exception('must have the same size' ) def __sub__( self : Dict , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] - other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: # error case raise Exception('must have the same size' ) @overload def __mul__( self : Tuple , UpperCamelCase__ : float ): ... @overload def __mul__( self : Dict , UpperCamelCase__ : Vector ): ... def __mul__( self : Union[str, Any] , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , (float, int) ): A = [c * other for c in self.__components] return Vector(UpperCamelCase__ ) elif isinstance(UpperCamelCase__ , UpperCamelCase__ ) and len(self ) == len(UpperCamelCase__ ): A = len(self ) A = [self.__components[i] * other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return sum(UpperCamelCase__ ) else: # error case raise Exception('invalid operand!' ) def UpperCamelCase ( self : Union[str, Any] ): return Vector(self.__components ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ) and -len(self.__components ) <= i < len(self.__components ): return self.__components[i] else: raise Exception('index out of range' ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : float ): assert -len(self.__components ) <= pos < len(self.__components ) A = value def UpperCamelCase ( self : str ): if len(self.__components ) == 0: raise Exception('Vector is empty' ) A = [c**2 for c in self.__components] return math.sqrt(sum(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Vector , UpperCamelCase__ : bool = False ): A = self * other A = self.euclidean_length() * other.euclidean_length() if deg: return math.degrees(math.acos(num / den ) ) else: return math.acos(num / den ) def __UpperCamelCase (lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) return Vector([0] * dimension ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, lowerCAmelCase )) A = [0] * dimension A = 1 return Vector(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : float, lowerCAmelCase : Vector, lowerCAmelCase : Vector ) -> Vector: assert ( isinstance(lowerCAmelCase, lowerCAmelCase ) and isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, (int, float) )) ) return x * scalar + y def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: random.seed(lowerCAmelCase ) A = [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] return Vector(lowerCAmelCase ) class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : list[list[float]] , UpperCamelCase__ : int , UpperCamelCase__ : int ): A = matrix A = w A = h def __str__( self : int ): A = '' for i in range(self.__height ): ans += "|" for j in range(self.__width ): if j < self.__width - 1: ans += str(self.__matrix[i][j] ) + "," else: ans += str(self.__matrix[i][j] ) + "|\n" return ans def __add__( self : Optional[Any] , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] + other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrix must have the same dimension!' ) def __sub__( self : Dict , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] - other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrices must have the same dimension!' ) @overload def __mul__( self : int , UpperCamelCase__ : float ): ... @overload def __mul__( self : Union[str, Any] , UpperCamelCase__ : Vector ): ... def __mul__( self : Tuple , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): # matrix-vector if len(UpperCamelCase__ ) == self.__width: A = zero_vector(self.__height ) for i in range(self.__height ): A = [ self.__matrix[i][j] * other.component(UpperCamelCase__ ) for j in range(self.__width ) ] ans.change_component(UpperCamelCase__ , sum(UpperCamelCase__ ) ) return ans else: raise Exception( 'vector must have the same size as the ' 'number of columns of the matrix!' ) elif isinstance(UpperCamelCase__ , (int, float) ): # matrix-scalar A = [ [self.__matrix[i][j] * other for j in range(self.__width )] for i in range(self.__height ) ] return Matrix(UpperCamelCase__ , self.__width , self.__height ) return None def UpperCamelCase ( self : Optional[int] ): return self.__height def UpperCamelCase ( self : List[Any] ): return self.__width def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int ): if 0 <= x < self.__height and 0 <= y < self.__width: return self.__matrix[x][y] else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : float ): if 0 <= x < self.__height and 0 <= y < self.__width: A = value else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) A = self.__matrix[:x] + self.__matrix[x + 1 :] for i in range(len(UpperCamelCase__ ) ): A = minor[i][:y] + minor[i][y + 1 :] return Matrix(UpperCamelCase__ , self.__width - 1 , self.__height - 1 ).determinant() def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if 0 <= x < self.__height and 0 <= y < self.__width: return (-1) ** (x + y) * self.minor(UpperCamelCase__ , UpperCamelCase__ ) else: raise Exception('Indices out of bounds' ) def UpperCamelCase ( self : Tuple ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if self.__height < 1: raise Exception('Matrix has no element' ) elif self.__height == 1: return self.__matrix[0][0] elif self.__height == 2: return ( self.__matrix[0][0] * self.__matrix[1][1] - self.__matrix[0][1] * self.__matrix[1][0] ) else: A = [ self.__matrix[0][y] * self.cofactor(0 , UpperCamelCase__ ) for y in range(self.__width ) ] return sum(UpperCamelCase__ ) def __UpperCamelCase (lowerCAmelCase : int ) -> Matrix: A = [[0] * n for _ in range(lowerCAmelCase )] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Matrix: random.seed(lowerCAmelCase ) A = [ [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] for _ in range(lowerCAmelCase ) ] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )
699
1
import unittest from transformers import JukeboxTokenizer from transformers.testing_utils import require_torch class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = JukeboxTokenizer SCREAMING_SNAKE_CASE : Optional[Any] = { '''artist''': '''Zac Brown Band''', '''genres''': '''Country''', '''lyrics''': '''I met a traveller from an antique land, Who said "Two vast and trunkless legs of stone Stand in the desert. . . . Near them, on the sand, Half sunk a shattered visage lies, whose frown, And wrinkled lip, and sneer of cold command, Tell that its sculptor well those passions read Which yet survive, stamped on these lifeless things, The hand that mocked them, and the heart that fed; And on the pedestal, these words appear: My name is Ozymandias, King of Kings; Look on my Works, ye Mighty, and despair! Nothing beside remains. Round the decay Of that colossal Wreck, boundless and bare The lone and level sands stretch far away ''', } @require_torch def UpperCamelCase ( self : Optional[Any] ): import torch A = JukeboxTokenizer.from_pretrained('openai/jukebox-1b-lyrics' ) A = tokenizer(**self.metas )['input_ids'] # fmt: off A = [ torch.tensor([[ 0, 0, 0, 7169, 507, 9, 76, 39, 31, 46, 76, 27, 76, 46, 44, 27, 48, 31, 38, 38, 31, 44, 76, 32, 44, 41, 39, 76, 27, 40, 76, 27, 40, 46, 35, 43, 47, 31, 76, 38, 27, 40, 30, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76, 23, 34, 41, 76, 45, 27, 35, 30, 76, 71, 20, 49, 41, 76, 48, 27, 45, 46, 76, 27, 40, 30, 76, 46, 44, 47, 40, 37, 38, 31, 45, 45, 76, 38, 31, 33, 45, 76, 41, 32, 76, 45, 46, 41, 40, 31, 78, 76, 76, 76, 76, 76, 76, 76, 76, 19, 46, 27, 40, 30, 76, 35, 40, 76, 46, 34, 31, 76, 30, 31, 45, 31, 44, 46, 63, 76, 63, 76, 63, 76, 63, 76, 14, 31, 27, 44, 76, 46, 34, 31, 39, 64, 76, 41, 40, 76, 46, 34, 31, 76, 45, 27, 40, 30, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76, 8, 27, 38, 32, 76, 45, 47, 40, 37, 76, 27, 76, 45, 34, 27, 46, 46, 31, 44, 31, 30, 76, 48, 35, 45, 27, 33, 31, 76, 38, 35, 31, 45, 64, 76, 49, 34, 41, 45, 31, 76, 32, 44, 41, 49, 40, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76, 1, 40, 30, 76, 49, 44, 35, 40, 37, 38, 31, 30, 76, 38, 35, 42, 64, 76, 27, 40, 30, 76, 45, 40, 31, 31, 44, 76, 41, 32, 76, 29, 41, 38, 30, 76, 29, 41, 39, 39, 27, 40, 30, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76, 20, 31, 38, 38, 76, 46, 34, 27, 46, 76, 35, 46, 45, 76, 45, 29, 47, 38, 42, 46, 41, 44, 76, 49, 31, 38, 38, 76, 46, 34, 41, 45, 31, 76, 42, 27, 45, 45, 35, 41, 40, 45, 76, 44, 31, 27, 30, 78, 76, 76, 76, 76, 76, 76, 76, 76, 23, 34, 35, 29, 34, 76, 51, 31, 46, 76, 45, 47, 44, 48, 35, 48, 31, 64, 76, 45, 46, 27, 39, 42, 31, 30, 76, 41, 40, 76, 46, 34, 31, 45, 31, 76, 38, 35, 32, 31, 38, 31, 45, 45, 76, 46, 34, 35, 40, 33, 45, 64, 78, 76, 76, 76, 76, 76, 76, 76, 76, 20, 34, 31, 76, 34, 27, 40, 30, 76, 46, 34, 27, 46, 76, 39, 41, 29, 37, 31, 30, 76, 46, 34, 31, 39, 64, 76, 27, 40, 30, 76, 46, 34, 31, 76, 34, 31, 27, 44, 46, 76, 46, 34, 27, 46, 76, 32, 31, 30, 66, 78, 76, 76, 76, 76, 76, 76, 76, 76, 1, 40, 30, 76, 41, 40, 76, 46, 34, 31, 76, 42, 31, 30, 31, 45, 46, 27, 38, 64, 76, 46, 34, 31, 45, 31, 76, 49, 41, 44, 30, 45, 76, 27, 42, 42, 31, 27, 44, 65, 78, 76, 76, 76, 76, 76, 76, 76, 76, 13, 51, 76, 40, 27, 39, 31, 76, 35, 45, 76, 15, 52, 51, 39, 27, 40, 30, 35, 27, 45, 64, 76, 11, 35, 40, 33, 76, 41, 32, 76, 11, 35, 40, 33, 45, 66, 78, 76, 76, 76, 76, 76, 76, 76, 76, 12, 41, 41, 37, 76, 41, 40, 76, 39, 51, 76, 23, 41, 44, 37, 45, 64, 76, 51, 31, 76, 13, 35, 33, 34, 46, 51, 64, 76, 27, 40, 30, 76, 30, 31, 45, 42, 27, 35, 44, 67, 78, 76, 76, 76, 76, 76, 76, 76, 76, 14, 41, 46, 34, 35, 40, 33, 76, 28, 31, 45, 35, 30, 31, 76, 44, 31, 39, 27, 35, 40, 45, 63, 76, 18, 41, 47, 40, 30, 76, 46, 34, 31, 76, 30, 31, 29, 27, 51, 78, 76, 76, 76, 76, 76, 76, 76, 76, 15, 32, 76, 46, 34, 27, 46, 76, 29, 41, 38, 41, 45, 45, 27, 38, 76, 23, 44, 31, 29, 37, 64, 76, 28, 41, 47, 40, 30, 38, 31, 45, 45, 76, 27, 40, 30, 76, 28, 27, 44, 31, 78, 76, 76, 76, 76, 76, 76, 76, 76, 20, 34, 31, 76, 38, 41, 40, 31, 76, 27, 40, 30, 76, 38, 31, 48, 31, 38, 76, 45, 27, 40, 30, 45, 76, 45, 46, 44, 31, 46, 29, 34, 76, 32, 27, 44, 76, 27, 49, 27, 51, 78, 76, 76, 76, 76, 76, 76, 76, 76]] ), torch.tensor([[0, 0, 0, 1069, 11]] ), torch.tensor([[0, 0, 0, 1069, 11]] ), ] # fmt: on self.assertTrue(torch.allclose(tokens[0] , EXPECTED_OUTPUT[0] ) ) self.assertTrue(torch.allclose(tokens[1] , EXPECTED_OUTPUT[1] ) ) self.assertTrue(torch.allclose(tokens[2] , EXPECTED_OUTPUT[2] ) ) @require_torch def UpperCamelCase ( self : Optional[Any] ): import torch A = JukeboxTokenizer.from_pretrained('openai/jukebox-5b-lyrics' ) A = tokenizer(**self.metas )['input_ids'] # fmt: off A = [ torch.tensor([[ 0, 0, 0, 1069, 11, -1, -1, -1, -1, 9, 77, 39, 31, 46, 77, 27, 77, 46, 44, 27, 48, 31, 38, 38, 31, 44, 77, 32, 44, 41, 39, 77, 27, 40, 77, 27, 40, 46, 35, 43, 47, 31, 77, 38, 27, 40, 30, 64, 79, 77, 77, 77, 77, 77, 77, 77, 77, 23, 34, 41, 77, 45, 27, 35, 30, 77, 72, 20, 49, 41, 77, 48, 27, 45, 46, 77, 27, 40, 30, 77, 46, 44, 47, 40, 37, 38, 31, 45, 45, 77, 38, 31, 33, 45, 77, 41, 32, 77, 45, 46, 41, 40, 31, 79, 77, 77, 77, 77, 77, 77, 77, 77, 19, 46, 27, 40, 30, 77, 35, 40, 77, 46, 34, 31, 77, 30, 31, 45, 31, 44, 46, 63, 77, 63, 77, 63, 77, 63, 77, 14, 31, 27, 44, 77, 46, 34, 31, 39, 64, 77, 41, 40, 77, 46, 34, 31, 77, 45, 27, 40, 30, 64, 79, 77, 77, 77, 77, 77, 77, 77, 77, 8, 27, 38, 32, 77, 45, 47, 40, 37, 77, 27, 77, 45, 34, 27, 46, 46, 31, 44, 31, 30, 77, 48, 35, 45, 27, 33, 31, 77, 38, 35, 31, 45, 64, 77, 49, 34, 41, 45, 31, 77, 32, 44, 41, 49, 40, 64, 79, 77, 77, 77, 77, 77, 77, 77, 77, 1, 40, 30, 77, 49, 44, 35, 40, 37, 38, 31, 30, 77, 38, 35, 42, 64, 77, 27, 40, 30, 77, 45, 40, 31, 31, 44, 77, 41, 32, 77, 29, 41, 38, 30, 77, 29, 41, 39, 39, 27, 40, 30, 64, 79, 77, 77, 77, 77, 77, 77, 77, 77, 20, 31, 38, 38, 77, 46, 34, 27, 46, 77, 35, 46, 45, 77, 45, 29, 47, 38, 42, 46, 41, 44, 77, 49, 31, 38, 38, 77, 46, 34, 41, 45, 31, 77, 42, 27, 45, 45, 35, 41, 40, 45, 77, 44, 31, 27, 30, 79, 77, 77, 77, 77, 77, 77, 77, 77, 23, 34, 35, 29, 34, 77, 51, 31, 46, 77, 45, 47, 44, 48, 35, 48, 31, 64, 77, 45, 46, 27, 39, 42, 31, 30, 77, 41, 40, 77, 46, 34, 31, 45, 31, 77, 38, 35, 32, 31, 38, 31, 45, 45, 77, 46, 34, 35, 40, 33, 45, 64, 79, 77, 77, 77, 77, 77, 77, 77, 77, 20, 34, 31, 77, 34, 27, 40, 30, 77, 46, 34, 27, 46, 77, 39, 41, 29, 37, 31, 30, 77, 46, 34, 31, 39, 64, 77, 27, 40, 30, 77, 46, 34, 31, 77, 34, 31, 27, 44, 46, 77, 46, 34, 27, 46, 77, 32, 31, 30, 66, 79, 77, 77, 77, 77, 77, 77, 77, 77, 1, 40, 30, 77, 41, 40, 77, 46, 34, 31, 77, 42, 31, 30, 31, 45, 46, 27, 38, 64, 77, 46, 34, 31, 45, 31, 77, 49, 41, 44, 30, 45, 77, 27, 42, 42, 31, 27, 44, 65, 79, 77, 77, 77, 77, 77, 77, 77, 77, 13, 51, 77, 40, 27, 39, 31, 77, 35, 45, 77, 15, 52, 51, 39, 27, 40, 30, 35, 27, 45, 64, 77, 11, 35, 40, 33, 77, 41, 32, 77, 11, 35, 40, 33, 45, 66, 79, 77, 77, 77, 77, 77, 77, 77, 77, 12, 41, 41, 37, 77, 41, 40, 77, 39, 51, 77, 23, 41, 44, 37, 45, 64, 77, 51, 31, 77, 13, 35, 33, 34, 46, 51, 64, 77, 27, 40, 30, 77, 30, 31, 45, 42, 27, 35, 44, 67, 79, 77, 77, 77, 77, 77, 77, 77, 77, 14, 41, 46, 34, 35, 40, 33, 77, 28, 31, 45, 35, 30, 31, 77, 44, 31, 39, 27, 35, 40, 45, 63, 77, 18, 41, 47, 40, 30, 77, 46, 34, 31, 77, 30, 31, 29, 27, 51, 79, 77, 77, 77, 77, 77, 77, 77, 77, 15, 32, 77, 46, 34, 27, 46, 77, 29, 41, 38, 41, 45, 45, 27, 38, 77, 23, 44, 31, 29, 37, 64, 77, 28, 41, 47, 40, 30, 38, 31, 45, 45, 77, 27, 40, 30, 77, 28, 27, 44, 31, 79, 77, 77, 77, 77, 77, 77, 77, 77, 20, 34, 31, 77, 38, 41, 40, 31, 77, 27, 40, 30, 77, 38, 31, 48, 31, 38, 77, 45, 27, 40, 30, 45, 77, 45, 46, 44, 31, 46, 29, 34, 77, 32, 27, 44, 77, 27, 49, 27, 51, 79, 77, 77, 77, 77, 77, 77, 77, 77]] ), torch.tensor([[0, 0, 0, 1069, 11, -1, -1, -1, -1]] ), torch.tensor([[0, 0, 0, 1069, 11, -1, -1, -1, -1]] ), ] # fmt: on self.assertTrue(torch.allclose(tokens[0] , EXPECTED_OUTPUT[0] ) ) self.assertTrue(torch.allclose(tokens[1] , EXPECTED_OUTPUT[1] ) ) self.assertTrue(torch.allclose(tokens[2] , EXPECTED_OUTPUT[2] ) )
699
from collections import OrderedDict from typing import Any, Mapping, Optional from ... import PreTrainedTokenizer from ...configuration_utils import PretrainedConfig from ...file_utils import TensorType, is_torch_available from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast from ...onnx.utils import compute_effective_axis_dimension from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "facebook/blenderbot_small-90M": "https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/config.json", # See all BlenderbotSmall models at https://huggingface.co/models?filter=blenderbot_small } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = '''blenderbot-small''' SCREAMING_SNAKE_CASE : Any = ['''past_key_values'''] SCREAMING_SNAKE_CASE : List[str] = {'''num_attention_heads''': '''encoder_attention_heads''', '''hidden_size''': '''d_model'''} def __init__( self : List[str] , UpperCamelCase__ : Optional[Any]=50265 , UpperCamelCase__ : Optional[int]=512 , UpperCamelCase__ : int=8 , UpperCamelCase__ : Optional[int]=2048 , UpperCamelCase__ : Optional[Any]=16 , UpperCamelCase__ : Optional[Any]=8 , UpperCamelCase__ : List[Any]=2048 , UpperCamelCase__ : int=16 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : int=True , UpperCamelCase__ : Optional[Any]="gelu" , UpperCamelCase__ : Any=512 , UpperCamelCase__ : Union[str, Any]=0.1 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Union[str, Any]=0.0 , UpperCamelCase__ : Dict=0.02 , UpperCamelCase__ : Optional[Any]=1 , UpperCamelCase__ : Any=False , UpperCamelCase__ : Dict=0 , UpperCamelCase__ : Optional[int]=1 , UpperCamelCase__ : List[str]=2 , UpperCamelCase__ : Dict=2 , **UpperCamelCase__ : List[str] , ): A = vocab_size A = max_position_embeddings A = d_model A = encoder_ffn_dim A = encoder_layers A = encoder_attention_heads A = decoder_ffn_dim A = decoder_layers A = decoder_attention_heads A = dropout A = attention_dropout A = activation_dropout A = activation_function A = init_std A = encoder_layerdrop A = decoder_layerdrop A = use_cache A = encoder_layers A = scale_embedding # scale factor will be sqrt(d_model) if True super().__init__( pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , is_encoder_decoder=UpperCamelCase__ , decoder_start_token_id=UpperCamelCase__ , forced_eos_token_id=UpperCamelCase__ , **UpperCamelCase__ , ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' @property def UpperCamelCase ( self : List[Any] ): if self.task in ["default", "seq2seq-lm"]: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A = {0: 'batch'} A = {0: 'batch', 1: 'past_decoder_sequence + sequence'} else: A = {0: 'batch', 1: 'decoder_sequence'} A = {0: 'batch', 1: 'decoder_sequence'} if self.use_past: self.fill_with_past_key_values_(UpperCamelCase__ , direction='inputs' ) elif self.task == "causal-lm": # TODO: figure this case out. A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} else: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ('decoder_input_ids', {0: 'batch', 1: 'decoder_sequence'}), ('decoder_attention_mask', {0: 'batch', 1: 'decoder_sequence'}), ] ) return common_inputs @property def UpperCamelCase ( self : int ): if self.task in ["default", "seq2seq-lm"]: A = super().outputs else: A = super(UpperCamelCase__ , self ).outputs if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} return common_outputs def UpperCamelCase ( self : int , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) # Generate decoder inputs A = seq_length if not self.use_past else 1 A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) A = {f'''decoder_{name}''': tensor for name, tensor in decoder_inputs.items()} A = dict(**UpperCamelCase__ , **UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape A = common_inputs['decoder_input_ids'].shape[1] A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, encoder_seq_length, self._config.hidden_size // num_encoder_attention_heads, ) A = decoder_seq_length + 3 A = ( batch, num_decoder_attention_heads, decoder_past_length, self._config.hidden_size // num_decoder_attention_heads, ) A = torch.cat( [common_inputs['decoder_attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ )] , dim=1 ) A = [] # If the number of encoder and decoder layers are present in the model configuration, both are considered A , A = self.num_layers A = min(UpperCamelCase__ , UpperCamelCase__ ) A = max(UpperCamelCase__ , UpperCamelCase__ ) - min_num_layers A = 'encoder' if num_encoder_layers > num_decoder_layers else 'decoder' for _ in range(UpperCamelCase__ ): common_inputs["past_key_values"].append( ( torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), ) ) # TODO: test this. A = encoder_shape if remaining_side_name == 'encoder' else decoder_shape for _ in range(UpperCamelCase__ , UpperCamelCase__ ): common_inputs["past_key_values"].append((torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) ) return common_inputs def UpperCamelCase ( self : Tuple , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape # Not using the same length for past_key_values A = seqlen + 2 A , A = self.num_layers A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, past_key_values_length, self._config.hidden_size // num_encoder_attention_heads, ) A = common_inputs['attention_mask'].dtype A = torch.cat( [common_inputs['attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ , dtype=UpperCamelCase__ )] , dim=1 ) A = [ (torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) for _ in range(UpperCamelCase__ ) ] return common_inputs def UpperCamelCase ( self : List[str] , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): # Copied from OnnxConfig.generate_dummy_inputs # Did not use super(OnnxConfigWithPast, self).generate_dummy_inputs for code clarity. # If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 ) # If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX A = tokenizer.num_special_tokens_to_add(UpperCamelCase__ ) A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=UpperCamelCase__ ) # Generate dummy inputs according to compute batch and sequence A = [' '.join([tokenizer.unk_token] ) * seq_length] * batch_size A = dict(tokenizer(UpperCamelCase__ , return_tensors=UpperCamelCase__ ) ) return common_inputs def UpperCamelCase ( self : Any , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): if self.task in ["default", "seq2seq-lm"]: A = self._generate_dummy_inputs_for_default_and_seqaseq_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) elif self.task == "causal-lm": A = self._generate_dummy_inputs_for_causal_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) else: A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) return common_inputs def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : List[str] , UpperCamelCase__ : str , UpperCamelCase__ : Tuple ): if self.task in ["default", "seq2seq-lm"]: A = super()._flatten_past_key_values_(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) else: A = super(UpperCamelCase__ , self )._flatten_past_key_values_( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
699
1
import operator as op _UpperCAmelCase = "scaler.pt" _UpperCAmelCase = "pytorch_model" _UpperCAmelCase = "random_states" _UpperCAmelCase = "optimizer" _UpperCAmelCase = "scheduler" _UpperCAmelCase = "pytorch_model.bin" _UpperCAmelCase = "pytorch_model.bin.index.json" _UpperCAmelCase = "model.safetensors" _UpperCAmelCase = "model.safetensors.index.json" _UpperCAmelCase = "1.10.2" _UpperCAmelCase = "py38" _UpperCAmelCase = "4.17.0" _UpperCAmelCase = ["ml.p3.16xlarge", "ml.p3dn.24xlarge", "ml.p4dn.24xlarge"] _UpperCAmelCase = ["FULL_SHARD", "SHARD_GRAD_OP", "NO_SHARD", "HYBRID_SHARD", "HYBRID_SHARD_ZERO2"] _UpperCAmelCase = ["TRANSFORMER_BASED_WRAP", "SIZE_BASED_WRAP", "NO_WRAP"] _UpperCAmelCase = ["BACKWARD_PRE", "BACKWARD_POST", "NO_PREFETCH"] _UpperCAmelCase = ["FULL_STATE_DICT", "LOCAL_STATE_DICT", "SHARDED_STATE_DICT"] _UpperCAmelCase = "2.0.1" _UpperCAmelCase = ["pdsh", "standard", "openmpi", "mvapich"] _UpperCAmelCase = ["default", "reduce-overhead", "max-autotune"] _UpperCAmelCase = {">": op.gt, ">=": op.ge, "==": op.eq, "!=": op.ne, "<=": op.le, "<": op.lt} # These are the args for `torch.distributed.launch` for pytorch < 1.9 _UpperCAmelCase = [ "nnodes", "nproc_per_node", "rdzv_backend", "rdzv_endpoint", "rdzv_id", "rdzv_conf", "standalone", "max_restarts", "monitor_interval", "start_method", "role", "module", "m", "no_python", "run_path", "log_dir", "r", "redirects", "t", "tee", "node_rank", "master_addr", "master_port", ] _UpperCAmelCase = ["DEEPSPEED", "MULTI_GPU", "FSDP", "MEGATRON_LM"] _UpperCAmelCase = ["DEEPSPEED", "MULTI_XPU", "FSDP"]
699
from typing import List, Optional, Union from ...processing_utils import ProcessorMixin from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy from ...utils import TensorType class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = ['''image_processor''', '''tokenizer'''] SCREAMING_SNAKE_CASE : List[str] = '''BridgeTowerImageProcessor''' SCREAMING_SNAKE_CASE : Tuple = ('''RobertaTokenizer''', '''RobertaTokenizerFast''') def __init__( self : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int] ): super().__init__(UpperCamelCase__ , UpperCamelCase__ ) def __call__( self : Optional[int] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None , UpperCamelCase__ : bool = True , UpperCamelCase__ : Union[bool, str, PaddingStrategy] = False , UpperCamelCase__ : Union[bool, str, TruncationStrategy] = None , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : Optional[Union[str, TensorType]] = None , **UpperCamelCase__ : List[Any] , ): A = self.tokenizer( text=UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , padding=UpperCamelCase__ , truncation=UpperCamelCase__ , max_length=UpperCamelCase__ , stride=UpperCamelCase__ , pad_to_multiple_of=UpperCamelCase__ , return_token_type_ids=UpperCamelCase__ , return_attention_mask=UpperCamelCase__ , return_overflowing_tokens=UpperCamelCase__ , return_special_tokens_mask=UpperCamelCase__ , return_offsets_mapping=UpperCamelCase__ , return_length=UpperCamelCase__ , verbose=UpperCamelCase__ , return_tensors=UpperCamelCase__ , **UpperCamelCase__ , ) # add pixel_values + pixel_mask A = self.image_processor( UpperCamelCase__ , return_tensors=UpperCamelCase__ , do_normalize=UpperCamelCase__ , do_center_crop=UpperCamelCase__ , **UpperCamelCase__ ) encoding.update(UpperCamelCase__ ) return encoding def UpperCamelCase ( self : Dict , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : Any ): return self.tokenizer.batch_decode(*UpperCamelCase__ , **UpperCamelCase__ ) def UpperCamelCase ( self : int , *UpperCamelCase__ : int , **UpperCamelCase__ : List[str] ): return self.tokenizer.decode(*UpperCamelCase__ , **UpperCamelCase__ ) @property def UpperCamelCase ( self : Any ): A = self.tokenizer.model_input_names A = self.image_processor.model_input_names return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
699
1
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = { "configuration_roberta": ["ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP", "RobertaConfig", "RobertaOnnxConfig"], "tokenization_roberta": ["RobertaTokenizer"], } try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["RobertaTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST", "RobertaForCausalLM", "RobertaForMaskedLM", "RobertaForMultipleChoice", "RobertaForQuestionAnswering", "RobertaForSequenceClassification", "RobertaForTokenClassification", "RobertaModel", "RobertaPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST", "TFRobertaForCausalLM", "TFRobertaForMaskedLM", "TFRobertaForMultipleChoice", "TFRobertaForQuestionAnswering", "TFRobertaForSequenceClassification", "TFRobertaForTokenClassification", "TFRobertaMainLayer", "TFRobertaModel", "TFRobertaPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxRobertaForCausalLM", "FlaxRobertaForMaskedLM", "FlaxRobertaForMultipleChoice", "FlaxRobertaForQuestionAnswering", "FlaxRobertaForSequenceClassification", "FlaxRobertaForTokenClassification", "FlaxRobertaModel", "FlaxRobertaPreTrainedModel", ] if TYPE_CHECKING: from .configuration_roberta import ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP, RobertaConfig, RobertaOnnxConfig from .tokenization_roberta import RobertaTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_roberta_fast import RobertaTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_roberta import ( ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST, RobertaForCausalLM, RobertaForMaskedLM, RobertaForMultipleChoice, RobertaForQuestionAnswering, RobertaForSequenceClassification, RobertaForTokenClassification, RobertaModel, RobertaPreTrainedModel, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_roberta import ( TF_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST, TFRobertaForCausalLM, TFRobertaForMaskedLM, TFRobertaForMultipleChoice, TFRobertaForQuestionAnswering, TFRobertaForSequenceClassification, TFRobertaForTokenClassification, TFRobertaMainLayer, TFRobertaModel, TFRobertaPreTrainedModel, ) try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_roberta import ( FlaxRobertaForCausalLM, FlaxRobertaForMaskedLM, FlaxRobertaForMultipleChoice, FlaxRobertaForQuestionAnswering, FlaxRobertaForSequenceClassification, FlaxRobertaForTokenClassification, FlaxRobertaModel, FlaxRobertaPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> str: 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=10))
699
1
from __future__ import annotations from decimal import Decimal from math import * # noqa: F403 from sympy import diff def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : float | Decimal, lowerCAmelCase : float = 10**-10 ) -> float: A = a while True: A = Decimal(lowerCAmelCase ) - ( Decimal(eval(lowerCAmelCase ) ) / Decimal(eval(str(diff(lowerCAmelCase ) ) ) ) # noqa: S307 ) # This number dictates the accuracy of the answer if abs(eval(lowerCAmelCase ) ) < precision: # noqa: S307 return float(lowerCAmelCase ) # Let's Execute if __name__ == "__main__": # Find root of trigonometric function # Find value of pi print(F'''The root of sin(x) = 0 is {newton_raphson("sin(x)", 2)}''') # Find root of polynomial print(F'''The root of x**2 - 5*x + 2 = 0 is {newton_raphson("x**2 - 5*x + 2", 0.4)}''') # Find Square Root of 5 print(F'''The root of log(x) - 1 = 0 is {newton_raphson("log(x) - 1", 2)}''') # Exponential Roots print(F'''The root of exp(x) - 1 = 0 is {newton_raphson("exp(x) - 1", 0)}''')
699
import math import torch from torch import nn from ..configuration_utils import ConfigMixin, register_to_config from .attention_processor import Attention from .embeddings import get_timestep_embedding from .modeling_utils import ModelMixin class _UpperCAmelCase ( __lowercase , __lowercase ): '''simple docstring''' @register_to_config def __init__( self : Any , UpperCamelCase__ : int = 128 , UpperCamelCase__ : int = 256 , UpperCamelCase__ : float = 2_000.0 , UpperCamelCase__ : int = 768 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 64 , UpperCamelCase__ : int = 2048 , UpperCamelCase__ : float = 0.1 , ): super().__init__() A = nn.Sequential( nn.Linear(UpperCamelCase__ , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , nn.Linear(d_model * 4 , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , ) A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = False A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.ModuleList() for lyr_num in range(UpperCamelCase__ ): # FiLM conditional T5 decoder A = DecoderLayer(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) self.decoders.append(UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Any , UpperCamelCase__ : int ): A = torch.mul(query_input.unsqueeze(-1 ) , key_input.unsqueeze(-2 ) ) return mask.unsqueeze(-3 ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[int] ): A , A , A = decoder_input_tokens.shape assert decoder_noise_time.shape == (batch,) # decoder_noise_time is in [0, 1), so rescale to expected timing range. A = get_timestep_embedding( decoder_noise_time * self.config.max_decoder_noise_time , embedding_dim=self.config.d_model , max_period=self.config.max_decoder_noise_time , ).to(dtype=self.dtype ) A = self.conditioning_emb(UpperCamelCase__ ).unsqueeze(1 ) assert conditioning_emb.shape == (batch, 1, self.config.d_model * 4) A = decoder_input_tokens.shape[1] # If we want to use relative positions for audio context, we can just offset # this sequence by the length of encodings_and_masks. A = torch.broadcast_to( torch.arange(UpperCamelCase__ , device=decoder_input_tokens.device ) , (batch, seq_length) , ) A = self.position_encoding(UpperCamelCase__ ) A = self.continuous_inputs_projection(UpperCamelCase__ ) inputs += position_encodings A = self.dropout(UpperCamelCase__ ) # decoder: No padding present. A = torch.ones( decoder_input_tokens.shape[:2] , device=decoder_input_tokens.device , dtype=inputs.dtype ) # Translate encoding masks to encoder-decoder masks. A = [(x, self.encoder_decoder_mask(UpperCamelCase__ , UpperCamelCase__ )) for x, y in encodings_and_masks] # cross attend style: concat encodings A = torch.cat([x[0] for x in encodings_and_encdec_masks] , dim=1 ) A = torch.cat([x[1] for x in encodings_and_encdec_masks] , dim=-1 ) for lyr in self.decoders: A = lyr( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , )[0] A = self.decoder_norm(UpperCamelCase__ ) A = self.post_dropout(UpperCamelCase__ ) A = self.spec_out(UpperCamelCase__ ) return spec_out class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Union[str, Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Union[str, Any]=1e-6 ): super().__init__() A = nn.ModuleList() # cond self attention: layer 0 self.layer.append( TaLayerSelfAttentionCond(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) ) # cross attention: layer 1 self.layer.append( TaLayerCrossAttention( d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ , ) ) # Film Cond MLP + dropout: last layer self.layer.append( TaLayerFFCond(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ ) ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : int=None , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[Any]=None , ): A = self.layer[0]( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) if encoder_hidden_states is not None: A = torch.where(encoder_attention_mask > 0 , 0 , -1e1_0 ).to( encoder_hidden_states.dtype ) A = self.layer[1]( UpperCamelCase__ , key_value_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) # Apply Film Conditional Feed Forward layer A = self.layer[-1](UpperCamelCase__ , UpperCamelCase__ ) return (hidden_states,) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Dict ): super().__init__() A = TaLayerNorm(UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : str , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Tuple=None , ): # pre_self_attention_layer_norm A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.FiLMLayer(UpperCamelCase__ , UpperCamelCase__ ) # Self-attention block A = self.attention(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[Any] ): super().__init__() A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[str]=None , ): A = self.layer_norm(UpperCamelCase__ ) A = self.attention( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=attention_mask.squeeze(1 ) , ) A = hidden_states + self.dropout(UpperCamelCase__ ) return layer_output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any ): super().__init__() A = TaDenseGatedActDense(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any=None ): A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.film(UpperCamelCase__ , UpperCamelCase__ ) A = self.DenseReluDense(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) A = NewGELUActivation() def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : List[Any] ): A = self.act(self.wi_a(UpperCamelCase__ ) ) A = self.wi_a(UpperCamelCase__ ) A = hidden_gelu * hidden_linear A = self.dropout(UpperCamelCase__ ) A = self.wo(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Tuple=1e-6 ): super().__init__() A = nn.Parameter(torch.ones(UpperCamelCase__ ) ) A = eps def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int ): # T5 uses a layer_norm which only scales and doesn't shift, which is also known as Root Mean # Square Layer Normalization https://arxiv.org/abs/1910.07467 thus variance is calculated # w/o mean and there is no bias. Additionally we want to make sure that the accumulation for # half-precision inputs is done in fp32 A = hidden_states.to(torch.floataa ).pow(2 ).mean(-1 , keepdim=UpperCamelCase__ ) A = hidden_states * torch.rsqrt(variance + self.variance_epsilon ) # convert into half-precision if necessary if self.weight.dtype in [torch.floataa, torch.bfloataa]: A = hidden_states.to(self.weight.dtype ) return self.weight * hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def UpperCamelCase ( self : Any , UpperCamelCase__ : torch.Tensor ): return 0.5 * input * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi ) * (input + 0.044_715 * torch.pow(UpperCamelCase__ , 3.0 )) )) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , out_features * 2 , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : List[Any] ): A = self.scale_bias(UpperCamelCase__ ) A , A = torch.chunk(UpperCamelCase__ , 2 , -1 ) A = x * (1 + scale) + shift return x
699
1
import unittest from queue import Empty from threading import Thread from transformers import AutoTokenizer, TextIteratorStreamer, TextStreamer, is_torch_available from transformers.testing_utils import CaptureStdout, require_torch, torch_device from ..test_modeling_common import ids_tensor if is_torch_available(): import torch from transformers import AutoModelForCausalLM @require_torch class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : int ): A = AutoTokenizer.from_pretrained('hf-internal-testing/tiny-random-gpt2' ) A = AutoModelForCausalLM.from_pretrained('hf-internal-testing/tiny-random-gpt2' ).to(UpperCamelCase__ ) A = -1 A = ids_tensor((1, 5) , vocab_size=model.config.vocab_size ).to(UpperCamelCase__ ) A = model.generate(UpperCamelCase__ , max_new_tokens=10 , do_sample=UpperCamelCase__ ) A = tokenizer.decode(greedy_ids[0] ) with CaptureStdout() as cs: A = TextStreamer(UpperCamelCase__ ) model.generate(UpperCamelCase__ , max_new_tokens=10 , do_sample=UpperCamelCase__ , streamer=UpperCamelCase__ ) # The greedy text should be printed to stdout, except for the final "\n" in the streamer A = cs.out[:-1] self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] ): A = AutoTokenizer.from_pretrained('hf-internal-testing/tiny-random-gpt2' ) A = AutoModelForCausalLM.from_pretrained('hf-internal-testing/tiny-random-gpt2' ).to(UpperCamelCase__ ) A = -1 A = ids_tensor((1, 5) , vocab_size=model.config.vocab_size ).to(UpperCamelCase__ ) A = model.generate(UpperCamelCase__ , max_new_tokens=10 , do_sample=UpperCamelCase__ ) A = tokenizer.decode(greedy_ids[0] ) A = TextIteratorStreamer(UpperCamelCase__ ) A = {'input_ids': input_ids, 'max_new_tokens': 10, 'do_sample': False, 'streamer': streamer} A = Thread(target=model.generate , kwargs=UpperCamelCase__ ) thread.start() A = '' for new_text in streamer: streamer_text += new_text self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : str ): A = AutoTokenizer.from_pretrained('hf-internal-testing/tiny-random-gpt2' ) A = AutoModelForCausalLM.from_pretrained('hf-internal-testing/tiny-random-gpt2' ).to(UpperCamelCase__ ) A = -1 A = ids_tensor((1, 5) , vocab_size=model.config.vocab_size ).to(UpperCamelCase__ ) A = model.generate(UpperCamelCase__ , max_new_tokens=10 , do_sample=UpperCamelCase__ ) A = greedy_ids[:, input_ids.shape[1] :] A = tokenizer.decode(new_greedy_ids[0] ) with CaptureStdout() as cs: A = TextStreamer(UpperCamelCase__ , skip_prompt=UpperCamelCase__ ) model.generate(UpperCamelCase__ , max_new_tokens=10 , do_sample=UpperCamelCase__ , streamer=UpperCamelCase__ ) # The greedy text should be printed to stdout, except for the final "\n" in the streamer A = cs.out[:-1] self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Dict ): # Tests that we can pass `decode_kwargs` to the streamer to control how the tokens are decoded. Must be tested # with actual models -- the dummy models' tokenizers are not aligned with their models, and # `skip_special_tokens=True` has no effect on them A = AutoTokenizer.from_pretrained('distilgpt2' ) A = AutoModelForCausalLM.from_pretrained('distilgpt2' ).to(UpperCamelCase__ ) A = -1 A = torch.ones((1, 5) , device=UpperCamelCase__ ).long() * model.config.bos_token_id with CaptureStdout() as cs: A = TextStreamer(UpperCamelCase__ , skip_special_tokens=UpperCamelCase__ ) model.generate(UpperCamelCase__ , max_new_tokens=1 , do_sample=UpperCamelCase__ , streamer=UpperCamelCase__ ) # The prompt contains a special token, so the streamer should not print it. As such, the output text, when # re-tokenized, must only contain one token A = cs.out[:-1] # Remove the final "\n" A = tokenizer(UpperCamelCase__ , return_tensors='pt' ) self.assertEqual(streamer_text_tokenized.input_ids.shape , (1, 1) ) def UpperCamelCase ( self : Union[str, Any] ): A = AutoTokenizer.from_pretrained('hf-internal-testing/tiny-random-gpt2' ) A = AutoModelForCausalLM.from_pretrained('hf-internal-testing/tiny-random-gpt2' ).to(UpperCamelCase__ ) A = -1 A = ids_tensor((1, 5) , vocab_size=model.config.vocab_size ).to(UpperCamelCase__ ) A = TextIteratorStreamer(UpperCamelCase__ , timeout=0.001 ) A = {'input_ids': input_ids, 'max_new_tokens': 10, 'do_sample': False, 'streamer': streamer} A = Thread(target=model.generate , kwargs=UpperCamelCase__ ) thread.start() # The streamer will timeout after 0.001 seconds, so an exception will be raised with self.assertRaises(UpperCamelCase__ ): A = '' for new_text in streamer: streamer_text += new_text
699
import numpy as np import torch from imwatermark import WatermarkEncoder # Copied from https://github.com/Stability-AI/generative-models/blob/613af104c6b85184091d42d374fef420eddb356d/scripts/demo/streamlit_helpers.py#L66 _UpperCAmelCase = 0b10_11_00_11_11_10_11_00_10_01_00_00_01_11_10_11_10_11_00_01_10_01_11_10 # bin(x)[2:] gives bits of x as str, use int to convert them to 0/1 _UpperCAmelCase = [int(bit) for bit in bin(WATERMARK_MESSAGE)[2:]] class _UpperCAmelCase : '''simple docstring''' def __init__( self : Union[str, Any] ): A = WATERMARK_BITS A = WatermarkEncoder() self.encoder.set_watermark('bits' , self.watermark ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : torch.FloatTensor ): # can't encode images that are smaller than 256 if images.shape[-1] < 256: return images A = (255 * (images / 2 + 0.5)).cpu().permute(0 , 2 , 3 , 1 ).float().numpy() A = [self.encoder.encode(UpperCamelCase__ , 'dwtDct' ) for image in images] A = torch.from_numpy(np.array(UpperCamelCase__ ) ).permute(0 , 3 , 1 , 2 ) A = torch.clamp(2 * (images / 255 - 0.5) , min=-1.0 , max=1.0 ) return images
699
1
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) if is_sentencepiece_available(): from ..ta.tokenization_ta import TaTokenizer else: from ...utils.dummy_sentencepiece_objects import TaTokenizer _UpperCAmelCase = TaTokenizer if is_tokenizers_available(): from ..ta.tokenization_ta_fast import TaTokenizerFast else: from ...utils.dummy_tokenizers_objects import TaTokenizerFast _UpperCAmelCase = TaTokenizerFast _UpperCAmelCase = {"configuration_mt5": ["MT5Config", "MT5OnnxConfig"]} try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "MT5EncoderModel", "MT5ForConditionalGeneration", "MT5ForQuestionAnswering", "MT5Model", "MT5PreTrainedModel", "MT5Stack", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["TFMT5EncoderModel", "TFMT5ForConditionalGeneration", "TFMT5Model"] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["FlaxMT5EncoderModel", "FlaxMT5ForConditionalGeneration", "FlaxMT5Model"] if TYPE_CHECKING: from .configuration_mta import MTaConfig, MTaOnnxConfig try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_mta import ( MTaEncoderModel, MTaForConditionalGeneration, MTaForQuestionAnswering, MTaModel, MTaPreTrainedModel, MTaStack, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel else: import sys _UpperCAmelCase = _LazyModule( __name__, globals()["__file__"], _import_structure, extra_objects={"MT5Tokenizer": MTaTokenizer, "MT5TokenizerFast": MTaTokenizerFast}, module_spec=__spec__, )
699
import argparse import fairseq import torch from transformers import UniSpeechSatConfig, UniSpeechSatForCTC, UniSpeechSatForPreTraining, logging logging.set_verbosity_info() _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "post_extract_proj": "feature_projection.projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.layer_norm": "encoder.layer_norm", "encoder.layer_norm_for_extract": "layer_norm_for_extract", "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": "lm_head", "label_embs_concat": "label_embeddings_concat", "mask_emb": "masked_spec_embed", "spk_proj": "speaker_proj", } _UpperCAmelCase = [ "lm_head", "quantizer.weight_proj", "quantizer.codevectors", "project_q", "project_hid", "label_embeddings_concat", "speaker_proj", "layer_norm_for_extract", ] def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Dict, lowerCAmelCase : Optional[int], lowerCAmelCase : List[Any], lowerCAmelCase : str ) -> int: 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 if hf_shape != value.shape: raise ValueError( 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 __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Optional[int] ) -> Dict: A = [] A = fairseq_model.state_dict() A = hf_model.unispeech_sat.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 = 'unispeech_sat.' + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key if key in name or key.split('w2v_model.' )[-1] == name.split('.' )[0]: if "layer_norm_for_extract" in name and (".".join(name.split('.' )[:-1] ) != key): # special case since naming is very similar continue 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 "bias" in name: A = 'bias' elif "weight" in name: # TODO: don't match quantizer.weight_proj A = 'weight' 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 __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : Tuple, lowerCAmelCase : List[Any], lowerCAmelCase : int ) -> Dict: 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.bias.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.weight.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape: raise ValueError( f'''{full_name} has size {value.shape}, but''' f''' {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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape: raise ValueError( 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 ) @torch.no_grad() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : Dict, lowerCAmelCase : Union[str, Any]=None, lowerCAmelCase : str=None, lowerCAmelCase : List[Any]=True ) -> Union[str, Any]: if config_path is not None: A = UniSpeechSatConfig.from_pretrained(lowerCAmelCase ) else: A = UniSpeechSatConfig() A = '' if is_finetuned: A = UniSpeechSatForCTC(lowerCAmelCase ) else: A = UniSpeechSatForPreTraining(lowerCAmelCase ) A , A , A = fairseq.checkpoint_utils.load_model_ensemble_and_task( [checkpoint_path], arg_overrides={'data': '/'.join(dict_path.split('/' )[:-1] )} ) A = model[0].eval() recursively_load_weights(lowerCAmelCase, lowerCAmelCase ) hf_wavavec.save_pretrained(lowerCAmelCase ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--dict_path", default=None, type=str, help="Path to dict of fine-tuned model") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") parser.add_argument( "--not_finetuned", action="store_true", help="Whether the model to convert is a fine-tuned model or not" ) _UpperCAmelCase = parser.parse_args() convert_unispeech_sat_checkpoint( args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned )
699
1
import json import os import shutil import tempfile import unittest import numpy as np import pytest from transformers import CLIPTokenizer, CLIPTokenizerFast from transformers.models.clip.tokenization_clip import VOCAB_FILES_NAMES from transformers.testing_utils import require_vision from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available if is_vision_available(): from PIL import Image from transformers import CLIPSegProcessor, ViTImageProcessor @require_vision class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : List[str] ): A = tempfile.mkdtemp() # fmt: off A = ['l', 'o', 'w', 'e', 'r', 's', 't', 'i', 'd', 'n', 'lo', 'l</w>', 'w</w>', 'r</w>', 't</w>', 'low</w>', 'er</w>', 'lowest</w>', 'newer</w>', 'wider', '<unk>', '<|startoftext|>', '<|endoftext|>'] # fmt: on A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = ['#version: 0.2', 'l o', 'lo w</w>', 'e r</w>', ''] A = {'unk_token': '<unk>'} 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' , encoding='utf-8' ) as fp: fp.write(json.dumps(UpperCamelCase__ ) + '\n' ) with open(self.merges_file , 'w' , encoding='utf-8' ) as fp: fp.write('\n'.join(UpperCamelCase__ ) ) A = { 'do_resize': True, 'size': 20, 'do_center_crop': True, 'crop_size': 18, 'do_normalize': True, 'image_mean': [0.48_145_466, 0.4_578_275, 0.40_821_073], 'image_std': [0.26_862_954, 0.26_130_258, 0.27_577_711], } A = os.path.join(self.tmpdirname , UpperCamelCase__ ) with open(self.image_processor_file , 'w' , encoding='utf-8' ) as fp: json.dump(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[int] , **UpperCamelCase__ : Optional[int] ): return CLIPTokenizer.from_pretrained(self.tmpdirname , **UpperCamelCase__ ) def UpperCamelCase ( self : str , **UpperCamelCase__ : Union[str, Any] ): return CLIPTokenizerFast.from_pretrained(self.tmpdirname , **UpperCamelCase__ ) def UpperCamelCase ( self : Any , **UpperCamelCase__ : List[str] ): return ViTImageProcessor.from_pretrained(self.tmpdirname , **UpperCamelCase__ ) def UpperCamelCase ( self : List[str] ): shutil.rmtree(self.tmpdirname ) def UpperCamelCase ( self : Optional[Any] ): A = [np.random.randint(255 , size=(3, 30, 400) , dtype=np.uinta )] A = [Image.fromarray(np.moveaxis(UpperCamelCase__ , 0 , -1 ) ) for x in image_inputs] return image_inputs def UpperCamelCase ( self : List[str] ): A = self.get_tokenizer() A = self.get_rust_tokenizer() A = self.get_image_processor() A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) processor_slow.save_pretrained(self.tmpdirname ) A = CLIPSegProcessor.from_pretrained(self.tmpdirname , use_fast=UpperCamelCase__ ) A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) processor_fast.save_pretrained(self.tmpdirname ) A = CLIPSegProcessor.from_pretrained(self.tmpdirname ) self.assertEqual(processor_slow.tokenizer.get_vocab() , tokenizer_slow.get_vocab() ) self.assertEqual(processor_fast.tokenizer.get_vocab() , tokenizer_fast.get_vocab() ) self.assertEqual(tokenizer_slow.get_vocab() , tokenizer_fast.get_vocab() ) self.assertIsInstance(processor_slow.tokenizer , UpperCamelCase__ ) self.assertIsInstance(processor_fast.tokenizer , UpperCamelCase__ ) self.assertEqual(processor_slow.image_processor.to_json_string() , image_processor.to_json_string() ) self.assertEqual(processor_fast.image_processor.to_json_string() , image_processor.to_json_string() ) self.assertIsInstance(processor_slow.image_processor , UpperCamelCase__ ) self.assertIsInstance(processor_fast.image_processor , UpperCamelCase__ ) def UpperCamelCase ( self : Any ): A = CLIPSegProcessor(tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() ) processor.save_pretrained(self.tmpdirname ) A = self.get_tokenizer(bos_token='(BOS)' , eos_token='(EOS)' ) A = self.get_image_processor(do_normalize=UpperCamelCase__ , padding_value=1.0 ) A = CLIPSegProcessor.from_pretrained( self.tmpdirname , bos_token='(BOS)' , eos_token='(EOS)' , do_normalize=UpperCamelCase__ , padding_value=1.0 ) self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() ) self.assertIsInstance(processor.tokenizer , UpperCamelCase__ ) self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() ) self.assertIsInstance(processor.image_processor , UpperCamelCase__ ) def UpperCamelCase ( self : List[Any] ): A = self.get_image_processor() A = self.get_tokenizer() A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) A = self.prepare_image_inputs() A = image_processor(UpperCamelCase__ , return_tensors='np' ) A = processor(images=UpperCamelCase__ , return_tensors='np' ) for key in input_feat_extract.keys(): self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 ) def UpperCamelCase ( self : List[Any] ): A = self.get_image_processor() A = self.get_tokenizer() A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) A = 'lower newer' A = processor(text=UpperCamelCase__ ) A = tokenizer(UpperCamelCase__ ) for key in encoded_tok.keys(): self.assertListEqual(encoded_tok[key] , encoded_processor[key] ) def UpperCamelCase ( self : Tuple ): A = self.get_image_processor() A = self.get_tokenizer() A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) A = 'lower newer' A = self.prepare_image_inputs() A = processor(text=UpperCamelCase__ , images=UpperCamelCase__ ) self.assertListEqual(list(inputs.keys() ) , ['input_ids', 'attention_mask', 'pixel_values'] ) # test if it raises when no input is passed with pytest.raises(UpperCamelCase__ ): processor() def UpperCamelCase ( self : Optional[Any] ): A = self.get_image_processor() A = self.get_tokenizer() A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) A = self.prepare_image_inputs() A = self.prepare_image_inputs() A = processor(images=UpperCamelCase__ , visual_prompt=UpperCamelCase__ ) self.assertListEqual(list(inputs.keys() ) , ['pixel_values', 'conditional_pixel_values'] ) # test if it raises when no input is passed with pytest.raises(UpperCamelCase__ ): processor() def UpperCamelCase ( self : Union[str, Any] ): A = self.get_image_processor() A = self.get_tokenizer() A = CLIPSegProcessor(tokenizer=UpperCamelCase__ , image_processor=UpperCamelCase__ ) A = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]] A = processor.batch_decode(UpperCamelCase__ ) A = tokenizer.batch_decode(UpperCamelCase__ ) self.assertListEqual(UpperCamelCase__ , UpperCamelCase__ )
699
from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar _UpperCAmelCase = TypeVar("T") class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : T ): A = data A = None def __str__( self : Optional[int] ): return f'''{self.data}''' class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple ): A = None def __iter__( self : int ): A = self.top while node: yield node.data A = node.next def __str__( self : Any ): return "->".join([str(UpperCamelCase__ ) for item in self] ) def __len__( self : Dict ): return len(tuple(iter(self ) ) ) def UpperCamelCase ( self : List[str] ): return self.top is None def UpperCamelCase ( self : Dict , UpperCamelCase__ : T ): A = Node(UpperCamelCase__ ) if not self.is_empty(): A = self.top A = node def UpperCamelCase ( self : Dict ): if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top , UpperCamelCase__ ) A = self.top A = self.top.next return pop_node.data def UpperCamelCase ( self : List[str] ): if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def UpperCamelCase ( self : List[str] ): A = None if __name__ == "__main__": from doctest import testmod testmod()
699
1
from collections.abc import Generator from math import sin def __UpperCamelCase (lowerCAmelCase : bytes ) -> bytes: if len(lowerCAmelCase ) != 32: raise ValueError('Input must be of length 32' ) A = b'' for i in [3, 2, 1, 0]: little_endian += string_aa[8 * i : 8 * i + 8] return little_endian def __UpperCamelCase (lowerCAmelCase : int ) -> bytes: if i < 0: raise ValueError('Input must be non-negative' ) A = format(lowerCAmelCase, '08x' )[-8:] A = 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 (lowerCAmelCase : bytes ) -> bytes: A = b'' for char in message: bit_string += format(lowerCAmelCase, '08b' ).encode('utf-8' ) A = format(len(lowerCAmelCase ), '064b' ).encode('utf-8' ) # Pad bit_string to a multiple of 512 chars bit_string += b"1" while len(lowerCAmelCase ) % 512 != 448: bit_string += b"0" bit_string += to_little_endian(start_len[32:] ) + to_little_endian(start_len[:32] ) return bit_string def __UpperCamelCase (lowerCAmelCase : bytes ) -> Generator[list[int], None, None]: if len(lowerCAmelCase ) % 512 != 0: raise ValueError('Input must have length that\'s a multiple of 512' ) for pos in range(0, len(lowerCAmelCase ), 512 ): A = bit_string[pos : pos + 512] A = [] for i in range(0, 512, 32 ): block_words.append(int(to_little_endian(block[i : i + 32] ), 2 ) ) yield block_words def __UpperCamelCase (lowerCAmelCase : int ) -> int: if i < 0: raise ValueError('Input must be non-negative' ) A = format(lowerCAmelCase, '032b' ) A = '' for c in i_str: new_str += "1" if c == "0" else "0" return int(lowerCAmelCase, 2 ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> int: return (a + b) % 2**32 def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> int: if i < 0: raise ValueError('Input must be non-negative' ) if shift < 0: raise ValueError('Shift must be non-negative' ) return ((i << shift) ^ (i >> (32 - shift))) % 2**32 def __UpperCamelCase (lowerCAmelCase : bytes ) -> bytes: A = preprocess(lowerCAmelCase ) A = [int(2**32 * abs(sin(i + 1 ) ) ) for i in range(64 )] # Starting states A = 0X6_7_4_5_2_3_0_1 A = 0Xe_f_c_d_a_b_8_9 A = 0X9_8_b_a_d_c_f_e A = 0X1_0_3_2_5_4_7_6 A = [ 7, 12, 17, 22, 7, 12, 17, 22, 7, 12, 17, 22, 7, 12, 17, 22, 5, 9, 14, 20, 5, 9, 14, 20, 5, 9, 14, 20, 5, 9, 14, 20, 4, 11, 16, 23, 4, 11, 16, 23, 4, 11, 16, 23, 4, 11, 16, 23, 6, 10, 15, 21, 6, 10, 15, 21, 6, 10, 15, 21, 6, 10, 15, 21, ] # Process bit string in chunks, each with 16 32-char words for block_words in get_block_words(lowerCAmelCase ): A = aa A = ba A = ca A = da # Hash current chunk for i in range(64 ): if i <= 15: # f = (b & c) | (not_32(b) & d) # Alternate definition for f A = d ^ (b & (c ^ d)) A = i elif i <= 31: # f = (d & b) | (not_32(d) & c) # Alternate definition for f A = c ^ (d & (b ^ c)) A = (5 * i + 1) % 16 elif i <= 47: A = b ^ c ^ d A = (3 * i + 5) % 16 else: A = c ^ (b | not_aa(lowerCAmelCase )) A = (7 * i) % 16 A = (f + a + added_consts[i] + block_words[g]) % 2**32 A = d A = c A = b A = sum_aa(lowerCAmelCase, left_rotate_aa(lowerCAmelCase, shift_amounts[i] ) ) # Add hashed chunk to running total A = sum_aa(lowerCAmelCase, lowerCAmelCase ) A = sum_aa(lowerCAmelCase, lowerCAmelCase ) A = sum_aa(lowerCAmelCase, lowerCAmelCase ) A = sum_aa(lowerCAmelCase, lowerCAmelCase ) A = reformat_hex(lowerCAmelCase ) + reformat_hex(lowerCAmelCase ) + reformat_hex(lowerCAmelCase ) + reformat_hex(lowerCAmelCase ) return digest if __name__ == "__main__": import doctest doctest.testmod()
699
from __future__ import annotations import math def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : bool, lowerCAmelCase : list[int], lowerCAmelCase : float ) -> int: if depth < 0: raise ValueError('Depth cannot be less than 0' ) if not scores: raise ValueError('Scores cannot be empty' ) if depth == height: return scores[node_index] return ( max( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) if is_max else min( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) ) def __UpperCamelCase () -> None: A = [90, 23, 6, 33, 21, 65, 123, 34_423] A = math.log(len(lowerCAmelCase ), 2 ) print(f'''Optimal value : {minimax(0, 0, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )}''' ) if __name__ == "__main__": import doctest doctest.testmod() main()
699
1
import sys def __UpperCamelCase (lowerCAmelCase : Dict ) -> Dict: A = len(lowerCAmelCase ) A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] for chain_length in range(2, lowerCAmelCase ): for a in range(1, n - chain_length + 1 ): A = a + chain_length - 1 A = sys.maxsize for c in range(lowerCAmelCase, lowerCAmelCase ): A = ( matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b] ) if cost < matrix[a][b]: A = cost A = c return matrix, sol def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Union[str, Any], lowerCAmelCase : Union[str, Any] ) -> List[str]: if i == j: print('A' + str(lowerCAmelCase ), end=' ' ) else: print('(', end=' ' ) print_optiomal_solution(lowerCAmelCase, lowerCAmelCase, optimal_solution[i][j] ) print_optiomal_solution(lowerCAmelCase, optimal_solution[i][j] + 1, lowerCAmelCase ) print(')', end=' ' ) def __UpperCamelCase () -> List[str]: A = [30, 35, 15, 5, 10, 20, 25] A = len(lowerCAmelCase ) # Size of matrix created from above array will be # 30*35 35*15 15*5 5*10 10*20 20*25 A , A = matrix_chain_order(lowerCAmelCase ) print('No. of Operation required: ' + str(matrix[1][n - 1] ) ) print_optiomal_solution(lowerCAmelCase, 1, n - 1 ) if __name__ == "__main__": main()
699
import logging import os from dataclasses import dataclass, field from typing import Dict, Optional import datasets import numpy as np import tensorflow as tf from transformers import ( AutoConfig, AutoTokenizer, EvalPrediction, HfArgumentParser, PreTrainedTokenizer, TFAutoModelForSequenceClassification, TFTrainer, TFTrainingArguments, ) from transformers.utils import logging as hf_logging hf_logging.set_verbosity_info() hf_logging.enable_default_handler() hf_logging.enable_explicit_format() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : PreTrainedTokenizer, lowerCAmelCase : int, lowerCAmelCase : Optional[int] = None, ) -> Dict: A = {} if train_file is not None: A = [train_file] if eval_file is not None: A = [eval_file] if test_file is not None: A = [test_file] A = datasets.load_dataset('csv', data_files=lowerCAmelCase ) A = list(ds[list(files.keys() )[0]].features.keys() ) A = features_name.pop(lowerCAmelCase ) A = list(set(ds[list(files.keys() )[0]][label_name] ) ) A = {label: i for i, label in enumerate(lowerCAmelCase )} A = tokenizer.model_input_names A = {} if len(lowerCAmelCase ) == 1: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( example[features_name[0]], truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length' ), batched=lowerCAmelCase, ) elif len(lowerCAmelCase ) == 2: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( (example[features_name[0]], example[features_name[1]]), truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length', ), batched=lowerCAmelCase, ) def gen_train(): for ex in transformed_ds[datasets.Split.TRAIN]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_val(): for ex in transformed_ds[datasets.Split.VALIDATION]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_test(): for ex in transformed_ds[datasets.Split.TEST]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TRAIN in transformed_ds else None ) if train_ds is not None: A = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.VALIDATION in transformed_ds else None ) if val_ds is not None: A = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TEST in transformed_ds else None ) if test_ds is not None: A = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) ) return train_ds, val_ds, test_ds, labelaid _UpperCAmelCase = logging.getLogger(__name__) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : int = field(metadata={'''help''': '''Which column contains the label'''} ) SCREAMING_SNAKE_CASE : str = field(default=__lowercase , metadata={'''help''': '''The path of the training file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the development file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the test file'''} ) SCREAMING_SNAKE_CASE : int = field( default=1_28 , metadata={ '''help''': ( '''The maximum total input sequence length after tokenization. Sequences longer ''' '''than this will be truncated, sequences shorter will be padded.''' ) } , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Overwrite the cached training and evaluation sets'''} ) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : str = field( metadata={'''help''': '''Path to pretrained model or model identifier from huggingface.co/models'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained config name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained tokenizer name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : bool = field(default=__lowercase , metadata={'''help''': '''Set this flag to use fast tokenization.'''} ) # If you want to tweak more attributes on your tokenizer, you should do it in a distinct script, # or just modify its tokenizer_config.json. SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Where do you want to store the pretrained models downloaded from huggingface.co'''} , ) def __UpperCamelCase () -> Any: # See all possible arguments in src/transformers/training_args.py # or by passing the --help flag to this script. # We now keep distinct sets of args, for a cleaner separation of concerns. A = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) ) A , A , A = parser.parse_args_into_dataclasses() if ( os.path.exists(training_args.output_dir ) and os.listdir(training_args.output_dir ) and training_args.do_train and not training_args.overwrite_output_dir ): raise ValueError( f'''Output directory ({training_args.output_dir}) already exists and is not empty. Use''' ' --overwrite_output_dir to overcome.' ) # Setup logging logging.basicConfig( format='%(asctime)s - %(levelname)s - %(name)s - %(message)s', datefmt='%m/%d/%Y %H:%M:%S', level=logging.INFO, ) logger.info( f'''n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, ''' f'''16-bits training: {training_args.fpaa}''' ) logger.info(f'''Training/evaluation parameters {training_args}''' ) # Load pretrained model and tokenizer # # Distributed training: # The .from_pretrained methods guarantee that only one local process can concurrently # download model & vocab. A = AutoTokenizer.from_pretrained( model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path, cache_dir=model_args.cache_dir, ) A , A , A , A = get_tfds( train_file=data_args.train_file, eval_file=data_args.dev_file, test_file=data_args.test_file, tokenizer=lowerCAmelCase, label_column_id=data_args.label_column_id, max_seq_length=data_args.max_seq_length, ) A = AutoConfig.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, num_labels=len(lowerCAmelCase ), labelaid=lowerCAmelCase, idalabel={id: label for label, id in labelaid.items()}, finetuning_task='text-classification', cache_dir=model_args.cache_dir, ) with training_args.strategy.scope(): A = TFAutoModelForSequenceClassification.from_pretrained( model_args.model_name_or_path, from_pt=bool('.bin' in model_args.model_name_or_path ), config=lowerCAmelCase, cache_dir=model_args.cache_dir, ) def compute_metrics(lowerCAmelCase : EvalPrediction ) -> Dict: A = np.argmax(p.predictions, axis=1 ) return {"acc": (preds == p.label_ids).mean()} # Initialize our Trainer A = TFTrainer( model=lowerCAmelCase, args=lowerCAmelCase, train_dataset=lowerCAmelCase, eval_dataset=lowerCAmelCase, compute_metrics=lowerCAmelCase, ) # Training if training_args.do_train: trainer.train() trainer.save_model() tokenizer.save_pretrained(training_args.output_dir ) # Evaluation A = {} if training_args.do_eval: logger.info('*** Evaluate ***' ) A = trainer.evaluate() A = os.path.join(training_args.output_dir, 'eval_results.txt' ) with open(lowerCAmelCase, 'w' ) as writer: logger.info('***** Eval results *****' ) for key, value in result.items(): logger.info(f''' {key} = {value}''' ) writer.write(f'''{key} = {value}\n''' ) results.update(lowerCAmelCase ) return results if __name__ == "__main__": main()
699
1
import warnings from ...utils import logging from .image_processing_segformer import SegformerImageProcessor _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : Optional[int] , *UpperCamelCase__ : List[str] , **UpperCamelCase__ : Union[str, Any] ): warnings.warn( 'The class SegformerFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use SegformerImageProcessor instead.' , UpperCamelCase__ , ) super().__init__(*UpperCamelCase__ , **UpperCamelCase__ )
699
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = {"configuration_xglm": ["XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "XGLMConfig"]} try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizer"] try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "XGLMForCausalLM", "XGLMModel", "XGLMPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxXGLMForCausalLM", "FlaxXGLMModel", "FlaxXGLMPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "TFXGLMForCausalLM", "TFXGLMModel", "TFXGLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm import XGLMTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm_fast import XGLMTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_xglm import ( TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, TFXGLMForCausalLM, TFXGLMModel, TFXGLMPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure)
699
1
import json from typing import List, Optional, Tuple from tokenizers import normalizers from ...tokenization_utils_fast import PreTrainedTokenizerFast from ...utils import logging from .tokenization_mobilebert import MobileBertTokenizer _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = {"vocab_file": "vocab.txt", "tokenizer_file": "tokenizer.json"} _UpperCAmelCase = { "vocab_file": {"mobilebert-uncased": "https://huggingface.co/google/mobilebert-uncased/resolve/main/vocab.txt"}, "tokenizer_file": { "mobilebert-uncased": "https://huggingface.co/google/mobilebert-uncased/resolve/main/tokenizer.json" }, } _UpperCAmelCase = {"mobilebert-uncased": 512} _UpperCAmelCase = {} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[Any] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : List[Any] = PRETRAINED_INIT_CONFIGURATION SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Tuple = MobileBertTokenizer def __init__( self : Union[str, Any] , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : str=True , UpperCamelCase__ : List[str]="[UNK]" , UpperCamelCase__ : List[str]="[SEP]" , UpperCamelCase__ : Optional[Any]="[PAD]" , UpperCamelCase__ : Dict="[CLS]" , UpperCamelCase__ : Optional[int]="[MASK]" , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : Union[str, Any]=None , **UpperCamelCase__ : Union[str, Any] , ): super().__init__( UpperCamelCase__ , tokenizer_file=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , unk_token=UpperCamelCase__ , sep_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , cls_token=UpperCamelCase__ , mask_token=UpperCamelCase__ , tokenize_chinese_chars=UpperCamelCase__ , strip_accents=UpperCamelCase__ , **UpperCamelCase__ , ) A = json.loads(self.backend_tokenizer.normalizer.__getstate__() ) if ( normalizer_state.get('lowercase' , UpperCamelCase__ ) != do_lower_case or normalizer_state.get('strip_accents' , UpperCamelCase__ ) != strip_accents or normalizer_state.get('handle_chinese_chars' , UpperCamelCase__ ) != tokenize_chinese_chars ): A = getattr(UpperCamelCase__ , normalizer_state.pop('type' ) ) A = do_lower_case A = strip_accents A = tokenize_chinese_chars A = normalizer_class(**UpperCamelCase__ ) A = do_lower_case def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Any , UpperCamelCase__ : int=None ): A = [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 UpperCamelCase ( self : int , UpperCamelCase__ : List[int] , UpperCamelCase__ : Optional[List[int]] = None ): A = [self.sep_token_id] A = [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 UpperCamelCase ( self : List[str] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): A = self._tokenizer.model.save(UpperCamelCase__ , name=UpperCamelCase__ ) return tuple(UpperCamelCase__ )
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Optional[int]: if b == 0: return 1 if (b % 2) == 0: return actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) else: return a * actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> float: if b < 0: return 1 / actual_power(lowerCAmelCase, lowerCAmelCase ) return actual_power(lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": print(power(-2, -3))
699
1
from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar _UpperCAmelCase = TypeVar("T") class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : T ): A = data A = None def __str__( self : Optional[int] ): return f'''{self.data}''' class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple ): A = None def __iter__( self : int ): A = self.top while node: yield node.data A = node.next def __str__( self : Any ): return "->".join([str(UpperCamelCase__ ) for item in self] ) def __len__( self : Dict ): return len(tuple(iter(self ) ) ) def UpperCamelCase ( self : List[str] ): return self.top is None def UpperCamelCase ( self : Dict , UpperCamelCase__ : T ): A = Node(UpperCamelCase__ ) if not self.is_empty(): A = self.top A = node def UpperCamelCase ( self : Dict ): if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top , UpperCamelCase__ ) A = self.top A = self.top.next return pop_node.data def UpperCamelCase ( self : List[str] ): if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def UpperCamelCase ( self : List[str] ): A = None if __name__ == "__main__": from doctest import testmod testmod()
699
def __UpperCamelCase (lowerCAmelCase : list[int] ) -> int: if not numbers: return 0 if not isinstance(lowerCAmelCase, (list, tuple) ) or not all( isinstance(lowerCAmelCase, lowerCAmelCase ) for number in numbers ): raise ValueError('numbers must be an iterable of integers' ) A = A = A = numbers[0] for i in range(1, len(lowerCAmelCase ) ): # update the maximum and minimum subarray products A = numbers[i] if number < 0: A , A = min_till_now, max_till_now A = max(lowerCAmelCase, max_till_now * number ) A = min(lowerCAmelCase, min_till_now * number ) # update the maximum product found till now A = max(lowerCAmelCase, lowerCAmelCase ) return max_prod
699
1
import unittest import torch from diffusers import DDIMScheduler, DDPMScheduler, UNetaDModel from diffusers.training_utils import set_seed from diffusers.utils.testing_utils import slow _UpperCAmelCase = False class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Union[str, Any]=32 ): set_seed(0 ) A = UNetaDModel(sample_size=UpperCamelCase__ , in_channels=3 , out_channels=3 ) A = torch.optim.SGD(model.parameters() , lr=0.0_001 ) return model, optimizer @slow def UpperCamelCase ( self : Dict ): A = 'cpu' # ensure full determinism without setting the CUBLAS_WORKSPACE_CONFIG env variable A = DDPMScheduler( num_train_timesteps=1000 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule='linear' , clip_sample=UpperCamelCase__ , ) A = DDIMScheduler( num_train_timesteps=1000 , beta_start=0.0_001 , beta_end=0.02 , beta_schedule='linear' , clip_sample=UpperCamelCase__ , ) assert ddpm_scheduler.config.num_train_timesteps == ddim_scheduler.config.num_train_timesteps # shared batches for DDPM and DDIM set_seed(0 ) A = [torch.randn((4, 3, 32, 32) ).clip(-1 , 1 ).to(UpperCamelCase__ ) for _ in range(4 )] A = [torch.randn((4, 3, 32, 32) ).to(UpperCamelCase__ ) for _ in range(4 )] A = [torch.randint(0 , 1000 , (4,) ).long().to(UpperCamelCase__ ) for _ in range(4 )] # train with a DDPM scheduler A , A = self.get_model_optimizer(resolution=32 ) model.train().to(UpperCamelCase__ ) for i in range(4 ): optimizer.zero_grad() A = ddpm_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] ) A = model(UpperCamelCase__ , timesteps[i] ).sample A = torch.nn.functional.mse_loss(UpperCamelCase__ , noise[i] ) loss.backward() optimizer.step() del model, optimizer # recreate the model and optimizer, and retry with DDIM A , A = self.get_model_optimizer(resolution=32 ) model.train().to(UpperCamelCase__ ) for i in range(4 ): optimizer.zero_grad() A = ddim_scheduler.add_noise(clean_images[i] , noise[i] , timesteps[i] ) A = model(UpperCamelCase__ , timesteps[i] ).sample A = torch.nn.functional.mse_loss(UpperCamelCase__ , noise[i] ) loss.backward() optimizer.step() del model, optimizer self.assertTrue(torch.allclose(UpperCamelCase__ , UpperCamelCase__ , atol=1e-5 ) ) self.assertTrue(torch.allclose(UpperCamelCase__ , UpperCamelCase__ , atol=1e-5 ) )
699
from .glue import glue_convert_examples_to_features, glue_output_modes, glue_processors, glue_tasks_num_labels from .squad import SquadExample, SquadFeatures, SquadVaProcessor, SquadVaProcessor, squad_convert_examples_to_features from .utils import DataProcessor, InputExample, InputFeatures, SingleSentenceClassificationProcessor from .xnli import xnli_output_modes, xnli_processors, xnli_tasks_num_labels
699
1
import re import jax.numpy as jnp from flax.traverse_util import flatten_dict, unflatten_dict from jax.random import PRNGKey from ..utils import logging _UpperCAmelCase = logging.get_logger(__name__) def __UpperCamelCase (lowerCAmelCase : List[str] ) -> Dict: A = r'\w+[.]\d+' A = re.findall(lowerCAmelCase, lowerCAmelCase ) for pat in pats: A = key.replace(lowerCAmelCase, '_'.join(pat.split('.' ) ) ) return key def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Dict, lowerCAmelCase : Dict ) -> Any: A = pt_tuple_key[:-1] + ('scale',) if ( any('norm' in str_ for str_ in pt_tuple_key ) and (pt_tuple_key[-1] == "bias") and (pt_tuple_key[:-1] + ("bias",) not in random_flax_state_dict) and (pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict) ): A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor elif pt_tuple_key[-1] in ["weight", "gamma"] and pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor # embedding if pt_tuple_key[-1] == "weight" and pt_tuple_key[:-1] + ("embedding",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('embedding',) return renamed_pt_tuple_key, pt_tensor # conv layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight" and pt_tensor.ndim == 4: A = pt_tensor.transpose(2, 3, 1, 0 ) return renamed_pt_tuple_key, pt_tensor # linear layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight": A = pt_tensor.T return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm weight A = pt_tuple_key[:-1] + ('weight',) if pt_tuple_key[-1] == "gamma": return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm bias A = pt_tuple_key[:-1] + ('bias',) if pt_tuple_key[-1] == "beta": return renamed_pt_tuple_key, pt_tensor return pt_tuple_key, pt_tensor def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Any, lowerCAmelCase : str=42 ) -> Any: # Step 1: Convert pytorch tensor to numpy A = {k: v.numpy() for k, v in pt_state_dict.items()} # Step 2: Since the model is stateless, get random Flax params A = flax_model.init_weights(PRNGKey(lowerCAmelCase ) ) A = flatten_dict(lowerCAmelCase ) A = {} # Need to change some parameters name to match Flax names for pt_key, pt_tensor in pt_state_dict.items(): A = rename_key(lowerCAmelCase ) A = tuple(renamed_pt_key.split('.' ) ) # Correctly rename weight parameters A , A = rename_key_and_reshape_tensor(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if flax_key in random_flax_state_dict: if flax_tensor.shape != random_flax_state_dict[flax_key].shape: raise ValueError( f'''PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape ''' f'''{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.''' ) # also add unexpected weight so that warning is thrown A = jnp.asarray(lowerCAmelCase ) return unflatten_dict(lowerCAmelCase )
699
from ...configuration_utils import PretrainedConfig from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "microsoft/biogpt": "https://huggingface.co/microsoft/biogpt/resolve/main/config.json", # See all BioGPT models at https://huggingface.co/models?filter=biogpt } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = '''biogpt''' def __init__( self : Optional[Any] , UpperCamelCase__ : str=42384 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : Dict=24 , UpperCamelCase__ : Any=16 , UpperCamelCase__ : str=4096 , UpperCamelCase__ : Tuple="gelu" , UpperCamelCase__ : List[str]=0.1 , UpperCamelCase__ : Any=0.1 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : List[Any]=0.02 , UpperCamelCase__ : Dict=1e-1_2 , UpperCamelCase__ : Any=True , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Any=1 , UpperCamelCase__ : List[str]=0 , UpperCamelCase__ : Optional[Any]=2 , **UpperCamelCase__ : List[Any] , ): A = vocab_size A = max_position_embeddings A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = initializer_range A = layer_norm_eps A = scale_embedding A = use_cache A = layerdrop A = activation_dropout super().__init__(pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , **UpperCamelCase__ )
699
1
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> str: if a < 0 or b < 0: raise ValueError('the value of both inputs must be positive' ) A = str(bin(lowerCAmelCase ) )[2:] # remove the leading "0b" A = str(bin(lowerCAmelCase ) )[2:] A = max(len(lowerCAmelCase ), len(lowerCAmelCase ) ) return "0b" + "".join( str(int('1' in (char_a, char_b) ) ) for char_a, char_b in zip(a_binary.zfill(lowerCAmelCase ), b_binary.zfill(lowerCAmelCase ) ) ) if __name__ == "__main__": import doctest doctest.testmod()
699
import sys def __UpperCamelCase (lowerCAmelCase : Dict ) -> Dict: A = len(lowerCAmelCase ) A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] for chain_length in range(2, lowerCAmelCase ): for a in range(1, n - chain_length + 1 ): A = a + chain_length - 1 A = sys.maxsize for c in range(lowerCAmelCase, lowerCAmelCase ): A = ( matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b] ) if cost < matrix[a][b]: A = cost A = c return matrix, sol def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Union[str, Any], lowerCAmelCase : Union[str, Any] ) -> List[str]: if i == j: print('A' + str(lowerCAmelCase ), end=' ' ) else: print('(', end=' ' ) print_optiomal_solution(lowerCAmelCase, lowerCAmelCase, optimal_solution[i][j] ) print_optiomal_solution(lowerCAmelCase, optimal_solution[i][j] + 1, lowerCAmelCase ) print(')', end=' ' ) def __UpperCamelCase () -> List[str]: A = [30, 35, 15, 5, 10, 20, 25] A = len(lowerCAmelCase ) # Size of matrix created from above array will be # 30*35 35*15 15*5 5*10 10*20 20*25 A , A = matrix_chain_order(lowerCAmelCase ) print('No. of Operation required: ' + str(matrix[1][n - 1] ) ) print_optiomal_solution(lowerCAmelCase, 1, n - 1 ) if __name__ == "__main__": main()
699
1
from cva import destroyAllWindows, imread, imshow, waitKey def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> Union[str, Any]: # getting number of pixels in the image A , A = img.shape[0], img.shape[1] # converting each pixel's color to its negative for i in range(lowerCAmelCase ): for j in range(lowerCAmelCase ): A = [255, 255, 255] - img[i][j] return img if __name__ == "__main__": # read original image _UpperCAmelCase = imread("image_data/lena.jpg", 1) # convert to its negative _UpperCAmelCase = convert_to_negative(img) # show result image imshow("negative of original image", img) waitKey(0) destroyAllWindows()
699
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> bool: return all(number % divisor != 0 for divisor in range(2, isqrt(lowerCAmelCase ) + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int = 10**6 ) -> int: A = 0 A = 1 A = 7 while prime_candidate < max_prime: primes_count += is_prime(lowerCAmelCase ) cube_index += 1 prime_candidate += 6 * cube_index return primes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
1
import unittest from transformers import BertGenerationTokenizer from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_torch, slow from transformers.utils import cached_property from ...test_tokenization_common import TokenizerTesterMixin _UpperCAmelCase = "▁" _UpperCAmelCase = get_tests_dir("fixtures/test_sentencepiece.model") @require_sentencepiece class _UpperCAmelCase ( __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Tuple = BertGenerationTokenizer SCREAMING_SNAKE_CASE : List[str] = False SCREAMING_SNAKE_CASE : List[Any] = True def UpperCamelCase ( self : Any ): super().setUp() A = BertGenerationTokenizer(UpperCamelCase__ , keep_accents=UpperCamelCase__ ) tokenizer.save_pretrained(self.tmpdirname ) def UpperCamelCase ( self : Dict ): A = '<s>' A = 1 self.assertEqual(self.get_tokenizer()._convert_token_to_id(UpperCamelCase__ ) , UpperCamelCase__ ) self.assertEqual(self.get_tokenizer()._convert_id_to_token(UpperCamelCase__ ) , UpperCamelCase__ ) def UpperCamelCase ( self : Union[str, Any] ): A = list(self.get_tokenizer().get_vocab().keys() ) self.assertEqual(vocab_keys[0] , '<unk>' ) self.assertEqual(vocab_keys[1] , '<s>' ) self.assertEqual(vocab_keys[-1] , '<pad>' ) self.assertEqual(len(UpperCamelCase__ ) , 1002 ) def UpperCamelCase ( self : List[Any] ): self.assertEqual(self.get_tokenizer().vocab_size , 1000 ) def UpperCamelCase ( self : List[str] ): A = BertGenerationTokenizer(UpperCamelCase__ , keep_accents=UpperCamelCase__ ) A = tokenizer.tokenize('This is a test' ) self.assertListEqual(UpperCamelCase__ , ['▁This', '▁is', '▁a', '▁t', 'est'] ) self.assertListEqual( tokenizer.convert_tokens_to_ids(UpperCamelCase__ ) , [285, 46, 10, 170, 382] , ) A = tokenizer.tokenize('I was born in 92000, and this is falsé.' ) self.assertListEqual( UpperCamelCase__ , [ SPIECE_UNDERLINE + 'I', SPIECE_UNDERLINE + 'was', SPIECE_UNDERLINE + 'b', 'or', 'n', SPIECE_UNDERLINE + 'in', SPIECE_UNDERLINE + '', '9', '2', '0', '0', '0', ',', SPIECE_UNDERLINE + 'and', SPIECE_UNDERLINE + 'this', SPIECE_UNDERLINE + 'is', SPIECE_UNDERLINE + 'f', 'al', 's', 'é', '.', ] , ) A = tokenizer.convert_tokens_to_ids(UpperCamelCase__ ) self.assertListEqual( UpperCamelCase__ , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , ) A = tokenizer.convert_ids_to_tokens(UpperCamelCase__ ) self.assertListEqual( UpperCamelCase__ , [ SPIECE_UNDERLINE + 'I', SPIECE_UNDERLINE + 'was', SPIECE_UNDERLINE + 'b', 'or', 'n', SPIECE_UNDERLINE + 'in', SPIECE_UNDERLINE + '', '<unk>', '2', '0', '0', '0', ',', SPIECE_UNDERLINE + 'and', SPIECE_UNDERLINE + 'this', SPIECE_UNDERLINE + 'is', SPIECE_UNDERLINE + 'f', 'al', 's', '<unk>', '.', ] , ) @cached_property def UpperCamelCase ( self : List[Any] ): return BertGenerationTokenizer.from_pretrained('google/bert_for_seq_generation_L-24_bbc_encoder' ) @slow def UpperCamelCase ( self : Optional[int] ): A = 'Hello World!' A = [18536, 2260, 101] self.assertListEqual(UpperCamelCase__ , self.big_tokenizer.encode(UpperCamelCase__ ) ) @slow def UpperCamelCase ( self : int ): A = ( 'This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) " [ ] ! : - . Also we will' ' add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth' ) A = [ 871, 419, 358, 946, 991, 2521, 452, 358, 1357, 387, 7751, 3536, 112, 985, 456, 126, 865, 938, 5400, 5734, 458, 1368, 467, 786, 2462, 5246, 1159, 633, 865, 4519, 457, 582, 852, 2557, 427, 916, 508, 405, 34324, 497, 391, 408, 11342, 1244, 385, 100, 938, 985, 456, 574, 362, 12597, 3200, 3129, 1172, ] self.assertListEqual(UpperCamelCase__ , self.big_tokenizer.encode(UpperCamelCase__ ) ) @require_torch @slow def UpperCamelCase ( self : Optional[int] ): import torch from transformers import BertGenerationConfig, BertGenerationEncoder # Build sequence A = list(self.big_tokenizer.get_vocab().keys() )[:10] A = ' '.join(UpperCamelCase__ ) A = self.big_tokenizer.encode_plus(UpperCamelCase__ , return_tensors='pt' , return_token_type_ids=UpperCamelCase__ ) A = self.big_tokenizer.batch_encode_plus( [sequence + ' ' + sequence] , return_tensors='pt' , return_token_type_ids=UpperCamelCase__ ) A = BertGenerationConfig() A = BertGenerationEncoder(UpperCamelCase__ ) assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size with torch.no_grad(): model(**UpperCamelCase__ ) model(**UpperCamelCase__ ) @slow def UpperCamelCase ( self : List[str] ): # fmt: off A = {'input_ids': [[39286, 458, 36335, 2001, 456, 13073, 13266, 455, 113, 7746, 1741, 11157, 391, 13073, 13266, 455, 113, 3967, 35412, 113, 4936, 109, 3870, 2377, 113, 30084, 45720, 458, 134, 17496, 112, 503, 11672, 113, 118, 112, 5665, 13347, 38687, 112, 1496, 31389, 112, 3268, 47264, 134, 962, 112, 16377, 8035, 23130, 430, 12169, 15518, 28592, 458, 146, 41697, 109, 391, 12169, 15518, 16689, 458, 146, 41358, 109, 452, 726, 4034, 111, 763, 35412, 5082, 388, 1903, 111, 9051, 391, 2870, 48918, 1900, 1123, 550, 998, 112, 9586, 15985, 455, 391, 410, 22955, 37636, 114], [448, 17496, 419, 3663, 385, 763, 113, 27533, 2870, 3283, 13043, 1639, 24713, 523, 656, 24013, 18550, 2521, 517, 27014, 21244, 420, 1212, 1465, 391, 927, 4833, 388, 578, 11786, 114, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [484, 2169, 7687, 21932, 18146, 726, 363, 17032, 3391, 114, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], 'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 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, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 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=UpperCamelCase__ , model_name='google/bert_for_seq_generation_L-24_bbc_encoder' , revision='c817d1fd1be2ffa69431227a1fe320544943d4db' , )
699
import warnings from ...utils import logging from .image_processing_imagegpt import ImageGPTImageProcessor _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , *UpperCamelCase__ : List[Any] , **UpperCamelCase__ : Tuple ): warnings.warn( 'The class ImageGPTFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use ImageGPTImageProcessor instead.' , UpperCamelCase__ , ) super().__init__(*UpperCamelCase__ , **UpperCamelCase__ )
699
1
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = { "configuration_distilbert": [ "DISTILBERT_PRETRAINED_CONFIG_ARCHIVE_MAP", "DistilBertConfig", "DistilBertOnnxConfig", ], "tokenization_distilbert": ["DistilBertTokenizer"], } try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["DistilBertTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST", "DistilBertForMaskedLM", "DistilBertForMultipleChoice", "DistilBertForQuestionAnswering", "DistilBertForSequenceClassification", "DistilBertForTokenClassification", "DistilBertModel", "DistilBertPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST", "TFDistilBertForMaskedLM", "TFDistilBertForMultipleChoice", "TFDistilBertForQuestionAnswering", "TFDistilBertForSequenceClassification", "TFDistilBertForTokenClassification", "TFDistilBertMainLayer", "TFDistilBertModel", "TFDistilBertPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxDistilBertForMaskedLM", "FlaxDistilBertForMultipleChoice", "FlaxDistilBertForQuestionAnswering", "FlaxDistilBertForSequenceClassification", "FlaxDistilBertForTokenClassification", "FlaxDistilBertModel", "FlaxDistilBertPreTrainedModel", ] if TYPE_CHECKING: from .configuration_distilbert import ( DISTILBERT_PRETRAINED_CONFIG_ARCHIVE_MAP, DistilBertConfig, DistilBertOnnxConfig, ) from .tokenization_distilbert import DistilBertTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_distilbert_fast import DistilBertTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_distilbert import ( DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST, DistilBertForMaskedLM, DistilBertForMultipleChoice, DistilBertForQuestionAnswering, DistilBertForSequenceClassification, DistilBertForTokenClassification, DistilBertModel, DistilBertPreTrainedModel, ) try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_distilbert import ( TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST, TFDistilBertForMaskedLM, TFDistilBertForMultipleChoice, TFDistilBertForQuestionAnswering, TFDistilBertForSequenceClassification, TFDistilBertForTokenClassification, TFDistilBertMainLayer, TFDistilBertModel, TFDistilBertPreTrainedModel, ) try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_distilbert import ( FlaxDistilBertForMaskedLM, FlaxDistilBertForMultipleChoice, FlaxDistilBertForQuestionAnswering, FlaxDistilBertForSequenceClassification, FlaxDistilBertForTokenClassification, FlaxDistilBertModel, FlaxDistilBertPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
699
from typing import Any, Dict, Optional import torch import torch.nn.functional as F from torch import nn from ..utils import maybe_allow_in_graph from .activations import get_activation from .attention_processor import Attention from .embeddings import CombinedTimestepLabelEmbeddings @maybe_allow_in_graph class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int]=0.0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : str = "layer_norm" , UpperCamelCase__ : bool = False , ): super().__init__() A = only_cross_attention A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm_zero' A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm' if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None: raise ValueError( f'''`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to''' f''' define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}.''' ) # Define 3 blocks. Each block has its own normalization layer. # 1. Self-Attn if self.use_ada_layer_norm: A = AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A = AdaLayerNormZero(UpperCamelCase__ , UpperCamelCase__ ) else: A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = Attention( query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if only_cross_attention else None , upcast_attention=UpperCamelCase__ , ) # 2. Cross-Attn if cross_attention_dim is not None or double_self_attention: # We currently only use AdaLayerNormZero for self attention where there will only be one attention block. # I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during # the second cross attention block. A = ( AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) ) A = Attention( query_dim=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if not double_self_attention else None , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , upcast_attention=UpperCamelCase__ , ) # is self-attn if encoder_hidden_states is none else: A = None A = None # 3. Feed-forward A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = FeedForward(UpperCamelCase__ , dropout=UpperCamelCase__ , activation_fn=UpperCamelCase__ , final_dropout=UpperCamelCase__ ) # let chunk size default to None A = None A = 0 def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int ): # Sets chunk feed-forward A = chunk_size A = dim def UpperCamelCase ( self : Dict , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , UpperCamelCase__ : Dict[str, Any] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , ): # Notice that normalization is always applied before the real computation in the following blocks. # 1. Self-Attention if self.use_ada_layer_norm: A = self.norma(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A , A , A , A , A = self.norma( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=hidden_states.dtype ) else: A = self.norma(UpperCamelCase__ ) A = cross_attention_kwargs if cross_attention_kwargs is not None else {} A = self.attna( UpperCamelCase__ , encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) if self.use_ada_layer_norm_zero: A = gate_msa.unsqueeze(1 ) * attn_output A = attn_output + hidden_states # 2. Cross-Attention if self.attna is not None: A = ( self.norma(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else self.norma(UpperCamelCase__ ) ) A = self.attna( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) A = attn_output + hidden_states # 3. Feed-forward A = self.norma(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None] if self._chunk_size is not None: # "feed_forward_chunk_size" can be used to save memory if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0: raise ValueError( f'''`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`.''' ) A = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size A = torch.cat( [self.ff(UpperCamelCase__ ) for hid_slice in norm_hidden_states.chunk(UpperCamelCase__ , dim=self._chunk_dim )] , dim=self._chunk_dim , ) else: A = self.ff(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = gate_mlp.unsqueeze(1 ) * ff_output A = ff_output + hidden_states return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 4 , UpperCamelCase__ : float = 0.0 , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : bool = False , ): super().__init__() A = int(dim * mult ) A = dim_out if dim_out is not None else dim if activation_fn == "gelu": A = GELU(UpperCamelCase__ , UpperCamelCase__ ) if activation_fn == "gelu-approximate": A = GELU(UpperCamelCase__ , UpperCamelCase__ , approximate='tanh' ) elif activation_fn == "geglu": A = GEGLU(UpperCamelCase__ , UpperCamelCase__ ) elif activation_fn == "geglu-approximate": A = ApproximateGELU(UpperCamelCase__ , UpperCamelCase__ ) A = nn.ModuleList([] ) # project in self.net.append(UpperCamelCase__ ) # project dropout self.net.append(nn.Dropout(UpperCamelCase__ ) ) # project out self.net.append(nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) ) # FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout if final_dropout: self.net.append(nn.Dropout(UpperCamelCase__ ) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int ): for module in self.net: A = module(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : str = "none" ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) A = approximate def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ , approximate=self.approximate ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) , approximate=self.approximate ).to(dtype=gate.dtype ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int ): A = self.proj(UpperCamelCase__ ) A = self.gelu(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , dim_out * 2 ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) ).to(dtype=gate.dtype ) def UpperCamelCase ( self : str , UpperCamelCase__ : str ): A , A = self.proj(UpperCamelCase__ ).chunk(2 , dim=-1 ) return hidden_states * self.gelu(UpperCamelCase__ ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[int] ): A = self.proj(UpperCamelCase__ ) return x * torch.sigmoid(1.702 * x ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Tuple ): super().__init__() A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , embedding_dim * 2 ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[Any] ): A = self.linear(self.silu(self.emb(UpperCamelCase__ ) ) ) A , A = torch.chunk(UpperCamelCase__ , 2 ) A = self.norm(UpperCamelCase__ ) * (1 + scale) + shift return x class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : int , UpperCamelCase__ : List[str] ): super().__init__() A = CombinedTimestepLabelEmbeddings(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , 6 * embedding_dim , bias=UpperCamelCase__ ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ , eps=1e-6 ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : str , UpperCamelCase__ : Tuple=None ): A = self.linear(self.silu(self.emb(UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=UpperCamelCase__ ) ) ) A , A , A , A , A , A = emb.chunk(6 , dim=1 ) A = self.norm(UpperCamelCase__ ) * (1 + scale_msa[:, None]) + shift_msa[:, None] return x, gate_msa, shift_mlp, scale_mlp, gate_mlp class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[str] = None , UpperCamelCase__ : float = 1e-5 ): super().__init__() A = num_groups A = eps if act_fn is None: A = None else: A = get_activation(UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , out_dim * 2 ) def UpperCamelCase ( self : Any , UpperCamelCase__ : str , UpperCamelCase__ : str ): if self.act: A = self.act(UpperCamelCase__ ) A = self.linear(UpperCamelCase__ ) A = emb[:, :, None, None] A , A = emb.chunk(2 , dim=1 ) A = F.group_norm(UpperCamelCase__ , self.num_groups , eps=self.eps ) A = x * (1 + scale) + shift return x
699
1
import numpy class _UpperCAmelCase : '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : numpy.ndarray , UpperCamelCase__ : numpy.ndarray ): A = input_array # Random initial weights are assigned where first argument is the # number of nodes in previous layer and second argument is the # number of nodes in the next layer. # Random initial weights are assigned. # self.input_array.shape[1] is used to represent number of nodes in input layer. # First hidden layer consists of 4 nodes. A = numpy.random.rand( self.input_array.shape[1] , 4 ) # Random initial values for the first hidden layer. # First hidden layer has 4 nodes. # Second hidden layer has 3 nodes. A = numpy.random.rand( 4 , 3 ) # Random initial values for the second hidden layer. # Second hidden layer has 3 nodes. # Output layer has 1 node. A = numpy.random.rand(3 , 1 ) # Real output values provided. A = output_array # Predicted output values by the neural network. # Predicted_output array initially consists of zeroes. A = numpy.zeros(output_array.shape ) def UpperCamelCase ( self : Dict ): A = sigmoid( numpy.dot(self.input_array , self.input_layer_and_first_hidden_layer_weights ) ) # layer_between_first_hidden_layer_and_second_hidden_layer is the layer # connecting the first hidden set of nodes with the second hidden set of nodes. A = sigmoid( numpy.dot( self.layer_between_input_and_first_hidden_layer , self.first_hidden_layer_and_second_hidden_layer_weights , ) ) # layer_between_second_hidden_layer_and_output is the layer connecting # second hidden layer with the output node. A = sigmoid( numpy.dot( self.layer_between_first_hidden_layer_and_second_hidden_layer , self.second_hidden_layer_and_output_layer_weights , ) ) return self.layer_between_second_hidden_layer_and_output def UpperCamelCase ( self : Union[str, Any] ): A = numpy.dot( self.layer_between_first_hidden_layer_and_second_hidden_layer.T , 2 * (self.output_array - self.predicted_output) * sigmoid_derivative(self.predicted_output ) , ) A = numpy.dot( self.layer_between_input_and_first_hidden_layer.T , numpy.dot( 2 * (self.output_array - self.predicted_output) * sigmoid_derivative(self.predicted_output ) , self.second_hidden_layer_and_output_layer_weights.T , ) * sigmoid_derivative( self.layer_between_first_hidden_layer_and_second_hidden_layer ) , ) A = numpy.dot( self.input_array.T , numpy.dot( numpy.dot( 2 * (self.output_array - self.predicted_output) * sigmoid_derivative(self.predicted_output ) , self.second_hidden_layer_and_output_layer_weights.T , ) * sigmoid_derivative( self.layer_between_first_hidden_layer_and_second_hidden_layer ) , self.first_hidden_layer_and_second_hidden_layer_weights.T , ) * sigmoid_derivative(self.layer_between_input_and_first_hidden_layer ) , ) self.input_layer_and_first_hidden_layer_weights += ( updated_input_layer_and_first_hidden_layer_weights ) self.first_hidden_layer_and_second_hidden_layer_weights += ( updated_first_hidden_layer_and_second_hidden_layer_weights ) self.second_hidden_layer_and_output_layer_weights += ( updated_second_hidden_layer_and_output_layer_weights ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : numpy.ndarray , UpperCamelCase__ : int , UpperCamelCase__ : bool ): for iteration in range(1 , iterations + 1 ): A = self.feedforward() self.back_propagation() if give_loss: A = numpy.mean(numpy.square(output - self.feedforward() ) ) print(f'''Iteration {iteration} Loss: {loss}''' ) def UpperCamelCase ( self : int , UpperCamelCase__ : numpy.ndarray ): A = input_arr A = sigmoid( numpy.dot(self.array , self.input_layer_and_first_hidden_layer_weights ) ) A = sigmoid( numpy.dot( self.layer_between_input_and_first_hidden_layer , self.first_hidden_layer_and_second_hidden_layer_weights , ) ) A = sigmoid( numpy.dot( self.layer_between_first_hidden_layer_and_second_hidden_layer , self.second_hidden_layer_and_output_layer_weights , ) ) return int(self.layer_between_second_hidden_layer_and_output > 0.6 ) def __UpperCamelCase (lowerCAmelCase : numpy.ndarray ) -> numpy.ndarray: return 1 / (1 + numpy.exp(-value )) def __UpperCamelCase (lowerCAmelCase : numpy.ndarray ) -> numpy.ndarray: return (value) * (1 - (value)) def __UpperCamelCase () -> int: A = numpy.array( ( [0, 0, 0], [0, 0, 1], [0, 1, 0], [0, 1, 1], [1, 0, 0], [1, 0, 1], [1, 1, 0], [1, 1, 1], ), dtype=numpy.floataa, ) # True output values for the given input values. A = numpy.array(([0], [1], [1], [0], [1], [0], [0], [1]), dtype=numpy.floataa ) # Calling neural network class. A = TwoHiddenLayerNeuralNetwork( input_array=lowerCAmelCase, output_array=lowerCAmelCase ) # Calling training function. # Set give_loss to True if you want to see loss in every iteration. neural_network.train(output=lowerCAmelCase, iterations=10, give_loss=lowerCAmelCase ) return neural_network.predict(numpy.array(([1, 1, 1]), dtype=numpy.floataa ) ) if __name__ == "__main__": example()
699
import json import os from typing import Dict, List, Optional, Tuple from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "vocab_file": "vocab.json", "tokenizer_config_file": "tokenizer_config.json", "merges_file": "merges.txt", } _UpperCAmelCase = { "vocab_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/vocab.json" ), }, "tokenizer_config_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/tokenizer_config.json" ), }, "merges_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/merges.txt" ), }, } _UpperCAmelCase = "</w>" _UpperCAmelCase = "@@ " def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> List[str]: A = set() A = word[0] for char in word[1:]: pairs.add((prev_char, char) ) A = char return pairs # Speech2Text2 has no max input length _UpperCAmelCase = {"facebook/s2t-wav2vec2-large-en-de": 1_024} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] def __init__( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="<pad>" , UpperCamelCase__ : int="</s>" , UpperCamelCase__ : Tuple="<unk>" , UpperCamelCase__ : List[str]=False , UpperCamelCase__ : List[str]=None , **UpperCamelCase__ : Optional[int] , ): super().__init__( unk_token=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , **UpperCamelCase__ , ) A = do_lower_case with open(UpperCamelCase__ , encoding='utf-8' ) as vocab_handle: A = json.load(UpperCamelCase__ ) A = {v: k for k, v in self.encoder.items()} if merges_file is None: logger.info(f'''No merges files provided. {self.__class__.__name__} can only be used for decoding.''' ) A = None A = None else: with open(UpperCamelCase__ , encoding='utf-8' ) as merges_handle: A = merges_handle.read().split('\n' )[:-1] A = [tuple(merge.split()[:2] ) for merge in merges] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = {} @property def UpperCamelCase ( self : Union[str, Any] ): return len(self.decoder ) def UpperCamelCase ( self : Optional[Any] ): return dict(self.encoder , **self.added_tokens_encoder ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): A = tuple(token[:-1] ) + (token[-1] + BPE_TOKEN_MERGES,) if token in self.cache: return self.cache[token] A = get_pairs(UpperCamelCase__ ) if not pairs: return token while True: A = min(UpperCamelCase__ , key=lambda UpperCamelCase__ : self.bpe_ranks.get(UpperCamelCase__ , float('inf' ) ) ) if bigram not in self.bpe_ranks: break A , A = bigram A = [] A = 0 while i < len(UpperCamelCase__ ): try: A = word.index(UpperCamelCase__ , UpperCamelCase__ ) except ValueError: new_word.extend(word[i:] ) break else: new_word.extend(word[i:j] ) A = j if word[i] == first and i < len(UpperCamelCase__ ) - 1 and word[i + 1] == second: new_word.append(first + second ) i += 2 else: new_word.append(word[i] ) i += 1 A = tuple(UpperCamelCase__ ) A = new_word if len(UpperCamelCase__ ) == 1: break else: A = get_pairs(UpperCamelCase__ ) A = ' '.join(UpperCamelCase__ ) if word == "\n " + BPE_TOKEN_MERGES: A = '\n' + BPE_TOKEN_MERGES if word.endswith(UpperCamelCase__ ): A = word.replace(UpperCamelCase__ , '' ) A = word.replace(' ' , UpperCamelCase__ ) A = word return word def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict ): if self.bpe_ranks is None: raise ValueError( 'This tokenizer was instantiated without a `merges.txt` file, so' ' that it can only be used for decoding, not for encoding.' 'Make sure to provide `merges.txt` file at instantiation to enable ' 'encoding.' ) if self.do_lower_case: A = text.lower() A = text.split() A = [] for token in text: if token: split_tokens.extend(list(self.bpe(UpperCamelCase__ ).split(' ' ) ) ) return split_tokens def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : str ): return self.encoder.get(UpperCamelCase__ , self.encoder.get(self.unk_token ) ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): A = self.decoder.get(UpperCamelCase__ , self.unk_token ) return result def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] ): A = ' '.join(UpperCamelCase__ ) # make sure @@ tokens are concatenated A = ''.join(string.split(UpperCamelCase__ ) ) return string def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] ) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as f: f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCamelCase__ , ensure_ascii=UpperCamelCase__ ) + '\n' ) A = 0 if self.bpe_ranks is None: return (vocab_file,) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as writer: for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCamelCase__ : kv[1] ): if index != token_index: logger.warning( f'''Saving vocabulary to {merges_file}: BPE merge indices are not consecutive.''' ' Please check that the tokenizer is not corrupted!' ) A = token_index writer.write(' '.join(UpperCamelCase__ ) + '\n' ) index += 1 return (vocab_file, merges_file)
699
1
from __future__ import annotations def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str ) -> bool: A = get_failure_array(lowerCAmelCase ) # 2) Step through text searching for pattern A , A = 0, 0 # index into text, pattern while i < len(lowerCAmelCase ): if pattern[j] == text[i]: if j == (len(lowerCAmelCase ) - 1): return True j += 1 # if this is a prefix in our pattern # just go back far enough to continue elif j > 0: A = failure[j - 1] continue i += 1 return False def __UpperCamelCase (lowerCAmelCase : str ) -> list[int]: A = [0] A = 0 A = 1 while j < len(lowerCAmelCase ): if pattern[i] == pattern[j]: i += 1 elif i > 0: A = failure[i - 1] continue j += 1 failure.append(lowerCAmelCase ) return failure if __name__ == "__main__": # Test 1) _UpperCAmelCase = "abc1abc12" _UpperCAmelCase = "alskfjaldsabc1abc1abc12k23adsfabcabc" _UpperCAmelCase = "alskfjaldsk23adsfabcabc" assert kmp(pattern, texta) and not kmp(pattern, texta) # Test 2) _UpperCAmelCase = "ABABX" _UpperCAmelCase = "ABABZABABYABABX" assert kmp(pattern, text) # Test 3) _UpperCAmelCase = "AAAB" _UpperCAmelCase = "ABAAAAAB" assert kmp(pattern, text) # Test 4) _UpperCAmelCase = "abcdabcy" _UpperCAmelCase = "abcxabcdabxabcdabcdabcy" assert kmp(pattern, text) # Test 5) _UpperCAmelCase = "aabaabaaa" assert get_failure_array(pattern) == [0, 1, 0, 1, 2, 3, 4, 5, 2]
699
# Copyright 2023 The HuggingFace Inc. team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import torch from ..models.auto import AutoModelForSequenceClassification, AutoTokenizer from .base import PipelineTool class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = '''facebook/bart-large-mnli''' SCREAMING_SNAKE_CASE : Union[str, Any] = ( '''This is a tool that classifies an English text using provided labels. It takes two inputs: `text`, which ''' '''should be the text to classify, and `labels`, which should be the list of labels to use for classification. ''' '''It returns the most likely label in the list of provided `labels` for the input text.''' ) SCREAMING_SNAKE_CASE : Any = '''text_classifier''' SCREAMING_SNAKE_CASE : Any = AutoTokenizer SCREAMING_SNAKE_CASE : Dict = AutoModelForSequenceClassification SCREAMING_SNAKE_CASE : List[Any] = ['''text''', ['''text''']] SCREAMING_SNAKE_CASE : Dict = ['''text'''] def UpperCamelCase ( self : List[str] ): super().setup() A = self.model.config A = -1 for idx, label in config.idalabel.items(): if label.lower().startswith('entail' ): A = int(UpperCamelCase__ ) if self.entailment_id == -1: raise ValueError('Could not determine the entailment ID from the model config, please pass it at init.' ) def UpperCamelCase ( self : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict ): A = labels return self.pre_processor( [text] * len(UpperCamelCase__ ) , [f'''This example is {label}''' for label in labels] , return_tensors='pt' , padding='max_length' , ) def UpperCamelCase ( self : int , UpperCamelCase__ : List[str] ): A = outputs.logits A = torch.argmax(logits[:, 2] ).item() return self._labels[label_id]
699
1
import csv from collections import defaultdict from dataclasses import dataclass, field from typing import List, Optional import matplotlib.pyplot as plt import numpy as np from matplotlib.ticker import ScalarFormatter from transformers import HfArgumentParser def __UpperCamelCase (lowerCAmelCase : str=None, lowerCAmelCase : Optional[Any]=None ) -> Any: return field(default_factory=lambda: default, metadata=lowerCAmelCase ) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : str = field( metadata={'''help''': '''The csv file to plot.'''} , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Whether to plot along batch size or sequence length. Defaults to sequence length.'''} , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Whether the csv file has time results or memory results. Defaults to memory results.'''} , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Disable logarithmic scale when plotting'''} , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={ '''help''': '''Whether the csv file has training results or inference results. Defaults to inference results.''' } , ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Filename under which the plot will be saved. If unused no plot is saved.'''} , ) SCREAMING_SNAKE_CASE : Optional[List[str]] = list_field( default=__lowercase , metadata={'''help''': '''List of model names that are used instead of the ones in the csv file.'''} ) def __UpperCamelCase (lowerCAmelCase : Any ) -> List[Any]: try: int(lowerCAmelCase ) return True except ValueError: return False def __UpperCamelCase (lowerCAmelCase : int ) -> Any: try: float(lowerCAmelCase ) return True except ValueError: return False class _UpperCAmelCase : '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Tuple ): A = args A = defaultdict(lambda: {"bsz": [], "seq_len": [], "result": {}} ) with open(self.args.csv_file , newline='' ) as csv_file: A = csv.DictReader(UpperCamelCase__ ) for row in reader: A = row['model'] self.result_dict[model_name]["bsz"].append(int(row['batch_size'] ) ) self.result_dict[model_name]["seq_len"].append(int(row['sequence_length'] ) ) if can_convert_to_int(row['result'] ): # value is not None A = int(row['result'] ) elif can_convert_to_float(row['result'] ): # value is not None A = float(row['result'] ) def UpperCamelCase ( self : Optional[Any] ): A , A = plt.subplots() A = 'Time usage' if self.args.is_time else 'Memory usage' A = title_str + ' for training' if self.args.is_train else title_str + ' for inference' if not self.args.no_log_scale: # set logarithm scales ax.set_xscale('log' ) ax.set_yscale('log' ) for axis in [ax.xaxis, ax.yaxis]: axis.set_major_formatter(ScalarFormatter() ) for model_name_idx, model_name in enumerate(self.result_dict.keys() ): A = sorted(set(self.result_dict[model_name]['bsz'] ) ) A = sorted(set(self.result_dict[model_name]['seq_len'] ) ) A = self.result_dict[model_name]['result'] ((A) , (A)) = ( (batch_sizes, sequence_lengths) if self.args.plot_along_batch else (sequence_lengths, batch_sizes) ) A = ( model_name if self.args.short_model_names is None else self.args.short_model_names[model_name_idx] ) for inner_loop_value in inner_loop_array: if self.args.plot_along_batch: A = np.asarray( [results[(x, inner_loop_value)] for x in x_axis_array if (x, inner_loop_value) in results] , dtype=UpperCamelCase__ , ) else: A = np.asarray( [results[(inner_loop_value, x)] for x in x_axis_array if (inner_loop_value, x) in results] , dtype=np.floataa , ) ((A) , (A)) = ( ('batch_size', 'len') if self.args.plot_along_batch else ('in #tokens', 'bsz') ) A = np.asarray(UpperCamelCase__ , UpperCamelCase__ )[: len(UpperCamelCase__ )] plt.scatter( UpperCamelCase__ , UpperCamelCase__ , label=f'''{label_model_name} - {inner_loop_label}: {inner_loop_value}''' ) plt.plot(UpperCamelCase__ , UpperCamelCase__ , '--' ) title_str += f''' {label_model_name} vs.''' A = title_str[:-4] A = 'Time in s' if self.args.is_time else 'Memory in MB' # plot plt.title(UpperCamelCase__ ) plt.xlabel(UpperCamelCase__ ) plt.ylabel(UpperCamelCase__ ) plt.legend() if self.args.figure_png_file is not None: plt.savefig(self.args.figure_png_file ) else: plt.show() def __UpperCamelCase () -> str: A = HfArgumentParser(lowerCAmelCase ) A = parser.parse_args_into_dataclasses()[0] A = Plot(args=lowerCAmelCase ) plot.plot() if __name__ == "__main__": main()
699
import re import jax.numpy as jnp from flax.traverse_util import flatten_dict, unflatten_dict from jax.random import PRNGKey from ..utils import logging _UpperCAmelCase = logging.get_logger(__name__) def __UpperCamelCase (lowerCAmelCase : List[str] ) -> Dict: A = r'\w+[.]\d+' A = re.findall(lowerCAmelCase, lowerCAmelCase ) for pat in pats: A = key.replace(lowerCAmelCase, '_'.join(pat.split('.' ) ) ) return key def __UpperCamelCase (lowerCAmelCase : Optional[int], lowerCAmelCase : Dict, lowerCAmelCase : Dict ) -> Any: A = pt_tuple_key[:-1] + ('scale',) if ( any('norm' in str_ for str_ in pt_tuple_key ) and (pt_tuple_key[-1] == "bias") and (pt_tuple_key[:-1] + ("bias",) not in random_flax_state_dict) and (pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict) ): A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor elif pt_tuple_key[-1] in ["weight", "gamma"] and pt_tuple_key[:-1] + ("scale",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('scale',) return renamed_pt_tuple_key, pt_tensor # embedding if pt_tuple_key[-1] == "weight" and pt_tuple_key[:-1] + ("embedding",) in random_flax_state_dict: A = pt_tuple_key[:-1] + ('embedding',) return renamed_pt_tuple_key, pt_tensor # conv layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight" and pt_tensor.ndim == 4: A = pt_tensor.transpose(2, 3, 1, 0 ) return renamed_pt_tuple_key, pt_tensor # linear layer A = pt_tuple_key[:-1] + ('kernel',) if pt_tuple_key[-1] == "weight": A = pt_tensor.T return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm weight A = pt_tuple_key[:-1] + ('weight',) if pt_tuple_key[-1] == "gamma": return renamed_pt_tuple_key, pt_tensor # old PyTorch layer norm bias A = pt_tuple_key[:-1] + ('bias',) if pt_tuple_key[-1] == "beta": return renamed_pt_tuple_key, pt_tensor return pt_tuple_key, pt_tensor def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Any, lowerCAmelCase : str=42 ) -> Any: # Step 1: Convert pytorch tensor to numpy A = {k: v.numpy() for k, v in pt_state_dict.items()} # Step 2: Since the model is stateless, get random Flax params A = flax_model.init_weights(PRNGKey(lowerCAmelCase ) ) A = flatten_dict(lowerCAmelCase ) A = {} # Need to change some parameters name to match Flax names for pt_key, pt_tensor in pt_state_dict.items(): A = rename_key(lowerCAmelCase ) A = tuple(renamed_pt_key.split('.' ) ) # Correctly rename weight parameters A , A = rename_key_and_reshape_tensor(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) if flax_key in random_flax_state_dict: if flax_tensor.shape != random_flax_state_dict[flax_key].shape: raise ValueError( f'''PyTorch checkpoint seems to be incorrect. Weight {pt_key} was expected to be of shape ''' f'''{random_flax_state_dict[flax_key].shape}, but is {flax_tensor.shape}.''' ) # also add unexpected weight so that warning is thrown A = jnp.asarray(lowerCAmelCase ) return unflatten_dict(lowerCAmelCase )
699
1
from torch import nn class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Tuple ): super().__init__() A = class_size A = embed_size # self.mlp1 = nn.Linear(embed_size, embed_size) # self.mlp2 = (nn.Linear(embed_size, class_size)) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): # hidden_state = nn.functional.relu(self.mlp1(hidden_state)) # hidden_state = self.mlp2(hidden_state) A = self.mlp(UpperCamelCase__ ) return logits
699
from __future__ import annotations import math import random from collections.abc import Collection from typing import overload class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Collection[float] | None = None ): if components is None: A = [] A = list(UpperCamelCase__ ) def __len__( self : List[Any] ): return len(self.__components ) def __str__( self : str ): return "(" + ",".join(map(UpperCamelCase__ , self.__components ) ) + ")" def __add__( self : str , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] + other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: raise Exception('must have the same size' ) def __sub__( self : Dict , UpperCamelCase__ : Vector ): A = len(self ) if size == len(UpperCamelCase__ ): A = [self.__components[i] - other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return Vector(UpperCamelCase__ ) else: # error case raise Exception('must have the same size' ) @overload def __mul__( self : Tuple , UpperCamelCase__ : float ): ... @overload def __mul__( self : Dict , UpperCamelCase__ : Vector ): ... def __mul__( self : Union[str, Any] , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , (float, int) ): A = [c * other for c in self.__components] return Vector(UpperCamelCase__ ) elif isinstance(UpperCamelCase__ , UpperCamelCase__ ) and len(self ) == len(UpperCamelCase__ ): A = len(self ) A = [self.__components[i] * other.component(UpperCamelCase__ ) for i in range(UpperCamelCase__ )] return sum(UpperCamelCase__ ) else: # error case raise Exception('invalid operand!' ) def UpperCamelCase ( self : Union[str, Any] ): return Vector(self.__components ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ) and -len(self.__components ) <= i < len(self.__components ): return self.__components[i] else: raise Exception('index out of range' ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : float ): assert -len(self.__components ) <= pos < len(self.__components ) A = value def UpperCamelCase ( self : str ): if len(self.__components ) == 0: raise Exception('Vector is empty' ) A = [c**2 for c in self.__components] return math.sqrt(sum(UpperCamelCase__ ) ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Vector , UpperCamelCase__ : bool = False ): A = self * other A = self.euclidean_length() * other.euclidean_length() if deg: return math.degrees(math.acos(num / den ) ) else: return math.acos(num / den ) def __UpperCamelCase (lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) return Vector([0] * dimension ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: assert isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, lowerCAmelCase )) A = [0] * dimension A = 1 return Vector(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : float, lowerCAmelCase : Vector, lowerCAmelCase : Vector ) -> Vector: assert ( isinstance(lowerCAmelCase, lowerCAmelCase ) and isinstance(lowerCAmelCase, lowerCAmelCase ) and (isinstance(lowerCAmelCase, (int, float) )) ) return x * scalar + y def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Vector: random.seed(lowerCAmelCase ) A = [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] return Vector(lowerCAmelCase ) class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : list[list[float]] , UpperCamelCase__ : int , UpperCamelCase__ : int ): A = matrix A = w A = h def __str__( self : int ): A = '' for i in range(self.__height ): ans += "|" for j in range(self.__width ): if j < self.__width - 1: ans += str(self.__matrix[i][j] ) + "," else: ans += str(self.__matrix[i][j] ) + "|\n" return ans def __add__( self : Optional[Any] , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] + other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrix must have the same dimension!' ) def __sub__( self : Dict , UpperCamelCase__ : Matrix ): if self.__width == other.width() and self.__height == other.height(): A = [] for i in range(self.__height ): A = [ self.__matrix[i][j] - other.component(UpperCamelCase__ , UpperCamelCase__ ) for j in range(self.__width ) ] matrix.append(UpperCamelCase__ ) return Matrix(UpperCamelCase__ , self.__width , self.__height ) else: raise Exception('matrices must have the same dimension!' ) @overload def __mul__( self : int , UpperCamelCase__ : float ): ... @overload def __mul__( self : Union[str, Any] , UpperCamelCase__ : Vector ): ... def __mul__( self : Tuple , UpperCamelCase__ : float | Vector ): if isinstance(UpperCamelCase__ , UpperCamelCase__ ): # matrix-vector if len(UpperCamelCase__ ) == self.__width: A = zero_vector(self.__height ) for i in range(self.__height ): A = [ self.__matrix[i][j] * other.component(UpperCamelCase__ ) for j in range(self.__width ) ] ans.change_component(UpperCamelCase__ , sum(UpperCamelCase__ ) ) return ans else: raise Exception( 'vector must have the same size as the ' 'number of columns of the matrix!' ) elif isinstance(UpperCamelCase__ , (int, float) ): # matrix-scalar A = [ [self.__matrix[i][j] * other for j in range(self.__width )] for i in range(self.__height ) ] return Matrix(UpperCamelCase__ , self.__width , self.__height ) return None def UpperCamelCase ( self : Optional[int] ): return self.__height def UpperCamelCase ( self : List[Any] ): return self.__width def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int , UpperCamelCase__ : int ): if 0 <= x < self.__height and 0 <= y < self.__width: return self.__matrix[x][y] else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : float ): if 0 <= x < self.__height and 0 <= y < self.__width: A = value else: raise Exception('change_component: indices out of bounds' ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) A = self.__matrix[:x] + self.__matrix[x + 1 :] for i in range(len(UpperCamelCase__ ) ): A = minor[i][:y] + minor[i][y + 1 :] return Matrix(UpperCamelCase__ , self.__width - 1 , self.__height - 1 ).determinant() def UpperCamelCase ( self : str , UpperCamelCase__ : int , UpperCamelCase__ : int ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if 0 <= x < self.__height and 0 <= y < self.__width: return (-1) ** (x + y) * self.minor(UpperCamelCase__ , UpperCamelCase__ ) else: raise Exception('Indices out of bounds' ) def UpperCamelCase ( self : Tuple ): if self.__height != self.__width: raise Exception('Matrix is not square' ) if self.__height < 1: raise Exception('Matrix has no element' ) elif self.__height == 1: return self.__matrix[0][0] elif self.__height == 2: return ( self.__matrix[0][0] * self.__matrix[1][1] - self.__matrix[0][1] * self.__matrix[1][0] ) else: A = [ self.__matrix[0][y] * self.cofactor(0 , UpperCamelCase__ ) for y in range(self.__width ) ] return sum(UpperCamelCase__ ) def __UpperCamelCase (lowerCAmelCase : int ) -> Matrix: A = [[0] * n for _ in range(lowerCAmelCase )] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : int ) -> Matrix: random.seed(lowerCAmelCase ) A = [ [random.randint(lowerCAmelCase, lowerCAmelCase ) for _ in range(lowerCAmelCase )] for _ in range(lowerCAmelCase ) ] return Matrix(lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )
699
1
from typing import Dict, List, Optional, Tuple, Union import torch from ...models import AutoencoderKL, TransformeraDModel from ...schedulers import KarrasDiffusionSchedulers from ...utils import randn_tensor from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : TransformeraDModel , UpperCamelCase__ : AutoencoderKL , UpperCamelCase__ : KarrasDiffusionSchedulers , UpperCamelCase__ : Optional[Dict[int, str]] = None , ): super().__init__() self.register_modules(transformer=UpperCamelCase__ , vae=UpperCamelCase__ , scheduler=UpperCamelCase__ ) # create a imagenet -> id dictionary for easier use A = {} if idalabel is not None: for key, value in idalabel.items(): for label in value.split(',' ): A = int(UpperCamelCase__ ) A = dict(sorted(self.labels.items() ) ) def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, List[str]] ): if not isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = list(UpperCamelCase__ ) for l in label: if l not in self.labels: raise ValueError( f'''{l} does not exist. Please make sure to select one of the following labels: \n {self.labels}.''' ) return [self.labels[l] for l in label] @torch.no_grad() def __call__( self : Dict , UpperCamelCase__ : List[int] , UpperCamelCase__ : float = 4.0 , UpperCamelCase__ : Optional[Union[torch.Generator, List[torch.Generator]]] = None , UpperCamelCase__ : int = 50 , UpperCamelCase__ : Optional[str] = "pil" , UpperCamelCase__ : bool = True , ): A = len(UpperCamelCase__ ) A = self.transformer.config.sample_size A = self.transformer.config.in_channels A = randn_tensor( shape=(batch_size, latent_channels, latent_size, latent_size) , generator=UpperCamelCase__ , device=self.device , dtype=self.transformer.dtype , ) A = torch.cat([latents] * 2 ) if guidance_scale > 1 else latents A = torch.tensor(UpperCamelCase__ , device=self.device ).reshape(-1 ) A = torch.tensor([1000] * batch_size , device=self.device ) A = torch.cat([class_labels, class_null] , 0 ) if guidance_scale > 1 else class_labels # set step values self.scheduler.set_timesteps(UpperCamelCase__ ) for t in self.progress_bar(self.scheduler.timesteps ): if guidance_scale > 1: A = latent_model_input[: len(UpperCamelCase__ ) // 2] A = torch.cat([half, half] , dim=0 ) A = self.scheduler.scale_model_input(UpperCamelCase__ , UpperCamelCase__ ) A = t if not torch.is_tensor(UpperCamelCase__ ): # TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can # This would be a good case for the `match` statement (Python 3.10+) A = latent_model_input.device.type == 'mps' if isinstance(UpperCamelCase__ , UpperCamelCase__ ): A = torch.floataa if is_mps else torch.floataa else: A = torch.intaa if is_mps else torch.intaa A = torch.tensor([timesteps] , dtype=UpperCamelCase__ , device=latent_model_input.device ) elif len(timesteps.shape ) == 0: A = timesteps[None].to(latent_model_input.device ) # broadcast to batch dimension in a way that's compatible with ONNX/Core ML A = timesteps.expand(latent_model_input.shape[0] ) # predict noise model_output A = self.transformer( UpperCamelCase__ , timestep=UpperCamelCase__ , class_labels=UpperCamelCase__ ).sample # perform guidance if guidance_scale > 1: A , A = noise_pred[:, :latent_channels], noise_pred[:, latent_channels:] A , A = torch.split(UpperCamelCase__ , len(UpperCamelCase__ ) // 2 , dim=0 ) A = uncond_eps + guidance_scale * (cond_eps - uncond_eps) A = torch.cat([half_eps, half_eps] , dim=0 ) A = torch.cat([eps, rest] , dim=1 ) # learned sigma if self.transformer.config.out_channels // 2 == latent_channels: A , A = torch.split(UpperCamelCase__ , UpperCamelCase__ , dim=1 ) else: A = noise_pred # compute previous image: x_t -> x_t-1 A = self.scheduler.step(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ).prev_sample if guidance_scale > 1: A , A = latent_model_input.chunk(2 , dim=0 ) else: A = latent_model_input A = 1 / self.vae.config.scaling_factor * latents A = self.vae.decode(UpperCamelCase__ ).sample A = (samples / 2 + 0.5).clamp(0 , 1 ) # we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16 A = samples.cpu().permute(0 , 2 , 3 , 1 ).float().numpy() if output_type == "pil": A = self.numpy_to_pil(UpperCamelCase__ ) if not return_dict: return (samples,) return ImagePipelineOutput(images=UpperCamelCase__ )
699
from collections import OrderedDict from typing import Any, Mapping, Optional from ... import PreTrainedTokenizer from ...configuration_utils import PretrainedConfig from ...file_utils import TensorType, is_torch_available from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast from ...onnx.utils import compute_effective_axis_dimension from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "facebook/blenderbot_small-90M": "https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/config.json", # See all BlenderbotSmall models at https://huggingface.co/models?filter=blenderbot_small } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = '''blenderbot-small''' SCREAMING_SNAKE_CASE : Any = ['''past_key_values'''] SCREAMING_SNAKE_CASE : List[str] = {'''num_attention_heads''': '''encoder_attention_heads''', '''hidden_size''': '''d_model'''} def __init__( self : List[str] , UpperCamelCase__ : Optional[Any]=50265 , UpperCamelCase__ : Optional[int]=512 , UpperCamelCase__ : int=8 , UpperCamelCase__ : Optional[int]=2048 , UpperCamelCase__ : Optional[Any]=16 , UpperCamelCase__ : Optional[Any]=8 , UpperCamelCase__ : List[Any]=2048 , UpperCamelCase__ : int=16 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : int=True , UpperCamelCase__ : Optional[Any]="gelu" , UpperCamelCase__ : Any=512 , UpperCamelCase__ : Union[str, Any]=0.1 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : Union[str, Any]=0.0 , UpperCamelCase__ : Dict=0.02 , UpperCamelCase__ : Optional[Any]=1 , UpperCamelCase__ : Any=False , UpperCamelCase__ : Dict=0 , UpperCamelCase__ : Optional[int]=1 , UpperCamelCase__ : List[str]=2 , UpperCamelCase__ : Dict=2 , **UpperCamelCase__ : List[str] , ): A = vocab_size A = max_position_embeddings A = d_model A = encoder_ffn_dim A = encoder_layers A = encoder_attention_heads A = decoder_ffn_dim A = decoder_layers A = decoder_attention_heads A = dropout A = attention_dropout A = activation_dropout A = activation_function A = init_std A = encoder_layerdrop A = decoder_layerdrop A = use_cache A = encoder_layers A = scale_embedding # scale factor will be sqrt(d_model) if True super().__init__( pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , is_encoder_decoder=UpperCamelCase__ , decoder_start_token_id=UpperCamelCase__ , forced_eos_token_id=UpperCamelCase__ , **UpperCamelCase__ , ) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' @property def UpperCamelCase ( self : List[Any] ): if self.task in ["default", "seq2seq-lm"]: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A = {0: 'batch'} A = {0: 'batch', 1: 'past_decoder_sequence + sequence'} else: A = {0: 'batch', 1: 'decoder_sequence'} A = {0: 'batch', 1: 'decoder_sequence'} if self.use_past: self.fill_with_past_key_values_(UpperCamelCase__ , direction='inputs' ) elif self.task == "causal-lm": # TODO: figure this case out. A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ] ) if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} else: A = OrderedDict( [ ('input_ids', {0: 'batch', 1: 'encoder_sequence'}), ('attention_mask', {0: 'batch', 1: 'encoder_sequence'}), ('decoder_input_ids', {0: 'batch', 1: 'decoder_sequence'}), ('decoder_attention_mask', {0: 'batch', 1: 'decoder_sequence'}), ] ) return common_inputs @property def UpperCamelCase ( self : int ): if self.task in ["default", "seq2seq-lm"]: A = super().outputs else: A = super(UpperCamelCase__ , self ).outputs if self.use_past: A , A = self.num_layers for i in range(UpperCamelCase__ ): A = {0: 'batch', 2: 'past_sequence + sequence'} A = {0: 'batch', 2: 'past_sequence + sequence'} return common_outputs def UpperCamelCase ( self : int , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) # Generate decoder inputs A = seq_length if not self.use_past else 1 A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) A = {f'''decoder_{name}''': tensor for name, tensor in decoder_inputs.items()} A = dict(**UpperCamelCase__ , **UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape A = common_inputs['decoder_input_ids'].shape[1] A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, encoder_seq_length, self._config.hidden_size // num_encoder_attention_heads, ) A = decoder_seq_length + 3 A = ( batch, num_decoder_attention_heads, decoder_past_length, self._config.hidden_size // num_decoder_attention_heads, ) A = torch.cat( [common_inputs['decoder_attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ )] , dim=1 ) A = [] # If the number of encoder and decoder layers are present in the model configuration, both are considered A , A = self.num_layers A = min(UpperCamelCase__ , UpperCamelCase__ ) A = max(UpperCamelCase__ , UpperCamelCase__ ) - min_num_layers A = 'encoder' if num_encoder_layers > num_decoder_layers else 'decoder' for _ in range(UpperCamelCase__ ): common_inputs["past_key_values"].append( ( torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ ), ) ) # TODO: test this. A = encoder_shape if remaining_side_name == 'encoder' else decoder_shape for _ in range(UpperCamelCase__ , UpperCamelCase__ ): common_inputs["past_key_values"].append((torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) ) return common_inputs def UpperCamelCase ( self : Tuple , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) if self.use_past: if not is_torch_available(): raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' ) else: import torch A , A = common_inputs['input_ids'].shape # Not using the same length for past_key_values A = seqlen + 2 A , A = self.num_layers A , A = self.num_attention_heads A = ( batch, num_encoder_attention_heads, past_key_values_length, self._config.hidden_size // num_encoder_attention_heads, ) A = common_inputs['attention_mask'].dtype A = torch.cat( [common_inputs['attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ , dtype=UpperCamelCase__ )] , dim=1 ) A = [ (torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) for _ in range(UpperCamelCase__ ) ] return common_inputs def UpperCamelCase ( self : List[str] , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): # Copied from OnnxConfig.generate_dummy_inputs # Did not use super(OnnxConfigWithPast, self).generate_dummy_inputs for code clarity. # If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 ) # If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX A = tokenizer.num_special_tokens_to_add(UpperCamelCase__ ) A = compute_effective_axis_dimension( UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=UpperCamelCase__ ) # Generate dummy inputs according to compute batch and sequence A = [' '.join([tokenizer.unk_token] ) * seq_length] * batch_size A = dict(tokenizer(UpperCamelCase__ , return_tensors=UpperCamelCase__ ) ) return common_inputs def UpperCamelCase ( self : Any , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): if self.task in ["default", "seq2seq-lm"]: A = self._generate_dummy_inputs_for_default_and_seqaseq_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) elif self.task == "causal-lm": A = self._generate_dummy_inputs_for_causal_lm( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) else: A = self._generate_dummy_inputs_for_sequence_classification_and_question_answering( UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) return common_inputs def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : List[str] , UpperCamelCase__ : str , UpperCamelCase__ : Tuple ): if self.task in ["default", "seq2seq-lm"]: A = super()._flatten_past_key_values_(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) else: A = super(UpperCamelCase__ , self )._flatten_past_key_values_( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
699
1
import inspect import unittest import warnings from transformers import DeiTConfig from transformers.models.auto import get_values from transformers.testing_utils import ( require_accelerate, require_torch, require_torch_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 torch import nn from transformers import ( MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING, MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING, MODEL_MAPPING, DeiTForImageClassification, DeiTForImageClassificationWithTeacher, DeiTForMaskedImageModeling, DeiTModel, ) from transformers.models.deit.modeling_deit import DEIT_PRETRAINED_MODEL_ARCHIVE_LIST if is_vision_available(): from PIL import Image from transformers import DeiTImageProcessor class _UpperCAmelCase : '''simple docstring''' def __init__( self : List[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : int=13 , UpperCamelCase__ : Optional[Any]=30 , UpperCamelCase__ : Dict=2 , UpperCamelCase__ : Optional[Any]=3 , UpperCamelCase__ : Dict=True , UpperCamelCase__ : Tuple=True , UpperCamelCase__ : List[str]=32 , UpperCamelCase__ : Union[str, Any]=5 , UpperCamelCase__ : Dict=4 , UpperCamelCase__ : Optional[Any]=37 , UpperCamelCase__ : List[Any]="gelu" , UpperCamelCase__ : List[str]=0.1 , UpperCamelCase__ : Dict=0.1 , UpperCamelCase__ : Tuple=10 , UpperCamelCase__ : Union[str, Any]=0.02 , UpperCamelCase__ : int=3 , UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : Optional[Any]=2 , ): A = parent A = batch_size A = image_size A = patch_size A = num_channels A = is_training A = use_labels A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = type_sequence_label_size A = initializer_range A = scope A = encoder_stride # in DeiT, the seq length equals the number of patches + 2 (we add 2 for the [CLS] and distilation tokens) A = (image_size // patch_size) ** 2 A = num_patches + 2 def UpperCamelCase ( self : List[Any] ): A = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] ) A = None if self.use_labels: A = ids_tensor([self.batch_size] , self.type_sequence_label_size ) A = self.get_config() return config, pixel_values, labels def UpperCamelCase ( self : List[str] ): return DeiTConfig( image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=UpperCamelCase__ , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Any , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[int] ): A = DeiTModel(config=UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ ) self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[int] ): A = DeiTForMaskedImageModeling(config=UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ ) self.parent.assertEqual( result.reconstruction.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) ) # test greyscale images A = 1 A = DeiTForMaskedImageModeling(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] ) A = model(UpperCamelCase__ ) self.parent.assertEqual(result.reconstruction.shape , (self.batch_size, 1, self.image_size, self.image_size) ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : str ): A = self.type_sequence_label_size A = DeiTForImageClassification(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , labels=UpperCamelCase__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) ) # test greyscale images A = 1 A = DeiTForImageClassification(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] ) A = model(UpperCamelCase__ , labels=UpperCamelCase__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) ) def UpperCamelCase ( self : Any ): A = self.prepare_config_and_inputs() ( ( A ) , ( A ) , ( A ) , ) = config_and_inputs A = {'pixel_values': pixel_values} return config, inputs_dict @require_torch class _UpperCAmelCase ( __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ( ( DeiTModel, DeiTForImageClassification, DeiTForImageClassificationWithTeacher, DeiTForMaskedImageModeling, ) if is_torch_available() else () ) SCREAMING_SNAKE_CASE : Union[str, Any] = ( { '''feature-extraction''': DeiTModel, '''image-classification''': (DeiTForImageClassification, DeiTForImageClassificationWithTeacher), } if is_torch_available() else {} ) SCREAMING_SNAKE_CASE : int = False SCREAMING_SNAKE_CASE : List[Any] = False SCREAMING_SNAKE_CASE : Optional[Any] = False def UpperCamelCase ( self : List[str] ): A = DeiTModelTester(self ) A = ConfigTester(self , config_class=UpperCamelCase__ , has_text_modality=UpperCamelCase__ , hidden_size=37 ) def UpperCamelCase ( self : List[Any] ): self.config_tester.run_common_tests() @unittest.skip(reason='DeiT does not use inputs_embeds' ) def UpperCamelCase ( self : str ): pass def UpperCamelCase ( self : Union[str, Any] ): A , A = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: A = model_class(UpperCamelCase__ ) self.assertIsInstance(model.get_input_embeddings() , (nn.Module) ) A = model.get_output_embeddings() self.assertTrue(x is None or isinstance(UpperCamelCase__ , nn.Linear ) ) def UpperCamelCase ( self : Tuple ): A , A = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: A = model_class(UpperCamelCase__ ) 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] , UpperCamelCase__ ) def UpperCamelCase ( self : int ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_model(*UpperCamelCase__ ) def UpperCamelCase ( self : Tuple ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_masked_image_modeling(*UpperCamelCase__ ) def UpperCamelCase ( self : str ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_image_classification(*UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any]=False ): A = super()._prepare_for_class(UpperCamelCase__ , UpperCamelCase__ , return_labels=UpperCamelCase__ ) if return_labels: if model_class.__name__ == "DeiTForImageClassificationWithTeacher": del inputs_dict["labels"] return inputs_dict def UpperCamelCase ( self : List[Any] ): if not self.model_tester.is_training: return A , A = self.model_tester.prepare_config_and_inputs_for_common() A = True for model_class in self.all_model_classes: # DeiTForImageClassificationWithTeacher supports inference-only if ( model_class in get_values(UpperCamelCase__ ) or model_class.__name__ == "DeiTForImageClassificationWithTeacher" ): continue A = model_class(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.train() A = self._prepare_for_class(UpperCamelCase__ , UpperCamelCase__ , return_labels=UpperCamelCase__ ) A = model(**UpperCamelCase__ ).loss loss.backward() def UpperCamelCase ( self : List[Any] ): A , A = self.model_tester.prepare_config_and_inputs_for_common() if not self.model_tester.is_training: return A = False A = True for model_class in self.all_model_classes: if model_class in get_values(UpperCamelCase__ ) or not model_class.supports_gradient_checkpointing: continue # DeiTForImageClassificationWithTeacher supports inference-only if model_class.__name__ == "DeiTForImageClassificationWithTeacher": continue A = model_class(UpperCamelCase__ ) model.gradient_checkpointing_enable() model.to(UpperCamelCase__ ) model.train() A = self._prepare_for_class(UpperCamelCase__ , UpperCamelCase__ , return_labels=UpperCamelCase__ ) A = model(**UpperCamelCase__ ).loss loss.backward() def UpperCamelCase ( self : Any ): A , A = self.model_tester.prepare_config_and_inputs_for_common() A = [ {'title': 'multi_label_classification', 'num_labels': 2, 'dtype': torch.float}, {'title': 'single_label_classification', 'num_labels': 1, 'dtype': torch.long}, {'title': 'regression', 'num_labels': 1, 'dtype': torch.float}, ] for model_class in self.all_model_classes: if ( model_class not in [ *get_values(UpperCamelCase__ ), *get_values(UpperCamelCase__ ), ] or model_class.__name__ == "DeiTForImageClassificationWithTeacher" ): continue for problem_type in problem_types: with self.subTest(msg=f'''Testing {model_class} with {problem_type["title"]}''' ): A = problem_type['title'] A = problem_type['num_labels'] A = model_class(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.train() A = self._prepare_for_class(UpperCamelCase__ , UpperCamelCase__ , return_labels=UpperCamelCase__ ) if problem_type["num_labels"] > 1: A = inputs['labels'].unsqueeze(1 ).repeat(1 , problem_type['num_labels'] ) A = inputs['labels'].to(problem_type['dtype'] ) # This tests that we do not trigger the warning form PyTorch "Using a target size that is different # to the input size. This will likely lead to incorrect results due to broadcasting. Please ensure # they have the same size." which is a symptom something in wrong for the regression problem. # See https://github.com/huggingface/transformers/issues/11780 with warnings.catch_warnings(record=UpperCamelCase__ ) as warning_list: A = model(**UpperCamelCase__ ).loss for w in warning_list: if "Using a target size that is different to the input size" in str(w.message ): raise ValueError( f'''Something is going wrong in the regression problem: intercepted {w.message}''' ) loss.backward() @slow def UpperCamelCase ( self : Union[str, Any] ): for model_name in DEIT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]: A = DeiTModel.from_pretrained(UpperCamelCase__ ) self.assertIsNotNone(UpperCamelCase__ ) def __UpperCamelCase () -> List[str]: A = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' ) return image @require_torch @require_vision class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' @cached_property def UpperCamelCase ( self : Union[str, Any] ): return ( DeiTImageProcessor.from_pretrained('facebook/deit-base-distilled-patch16-224' ) if is_vision_available() else None ) @slow def UpperCamelCase ( self : Union[str, Any] ): A = DeiTForImageClassificationWithTeacher.from_pretrained('facebook/deit-base-distilled-patch16-224' ).to( UpperCamelCase__ ) A = self.default_image_processor A = prepare_img() A = image_processor(images=UpperCamelCase__ , return_tensors='pt' ).to(UpperCamelCase__ ) # forward pass with torch.no_grad(): A = model(**UpperCamelCase__ ) # verify the logits A = torch.Size((1, 1000) ) self.assertEqual(outputs.logits.shape , UpperCamelCase__ ) A = torch.tensor([-1.0_266, 0.1_912, -1.2_861] ).to(UpperCamelCase__ ) self.assertTrue(torch.allclose(outputs.logits[0, :3] , UpperCamelCase__ , atol=1e-4 ) ) @slow @require_accelerate @require_torch_gpu def UpperCamelCase ( self : Optional[int] ): A = DeiTModel.from_pretrained( 'facebook/deit-base-distilled-patch16-224' , torch_dtype=torch.floataa , device_map='auto' ) A = self.default_image_processor A = prepare_img() A = image_processor(images=UpperCamelCase__ , return_tensors='pt' ) A = inputs.pixel_values.to(UpperCamelCase__ ) # forward pass to make sure inference works in fp16 with torch.no_grad(): A = model(UpperCamelCase__ )
699
from typing import List, Optional, Union from ...processing_utils import ProcessorMixin from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy from ...utils import TensorType class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = ['''image_processor''', '''tokenizer'''] SCREAMING_SNAKE_CASE : List[str] = '''BridgeTowerImageProcessor''' SCREAMING_SNAKE_CASE : Tuple = ('''RobertaTokenizer''', '''RobertaTokenizerFast''') def __init__( self : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int] ): super().__init__(UpperCamelCase__ , UpperCamelCase__ ) def __call__( self : Optional[int] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None , UpperCamelCase__ : bool = True , UpperCamelCase__ : Union[bool, str, PaddingStrategy] = False , UpperCamelCase__ : Union[bool, str, TruncationStrategy] = None , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : Optional[bool] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : Optional[Union[str, TensorType]] = None , **UpperCamelCase__ : List[Any] , ): A = self.tokenizer( text=UpperCamelCase__ , add_special_tokens=UpperCamelCase__ , padding=UpperCamelCase__ , truncation=UpperCamelCase__ , max_length=UpperCamelCase__ , stride=UpperCamelCase__ , pad_to_multiple_of=UpperCamelCase__ , return_token_type_ids=UpperCamelCase__ , return_attention_mask=UpperCamelCase__ , return_overflowing_tokens=UpperCamelCase__ , return_special_tokens_mask=UpperCamelCase__ , return_offsets_mapping=UpperCamelCase__ , return_length=UpperCamelCase__ , verbose=UpperCamelCase__ , return_tensors=UpperCamelCase__ , **UpperCamelCase__ , ) # add pixel_values + pixel_mask A = self.image_processor( UpperCamelCase__ , return_tensors=UpperCamelCase__ , do_normalize=UpperCamelCase__ , do_center_crop=UpperCamelCase__ , **UpperCamelCase__ ) encoding.update(UpperCamelCase__ ) return encoding def UpperCamelCase ( self : Dict , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : Any ): return self.tokenizer.batch_decode(*UpperCamelCase__ , **UpperCamelCase__ ) def UpperCamelCase ( self : int , *UpperCamelCase__ : int , **UpperCamelCase__ : List[str] ): return self.tokenizer.decode(*UpperCamelCase__ , **UpperCamelCase__ ) @property def UpperCamelCase ( self : Any ): A = self.tokenizer.model_input_names A = self.image_processor.model_input_names return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
699
1
from typing import Dict, Optional import numpy as np import datasets _UpperCAmelCase = "\nIoU is the area of overlap between the predicted segmentation and the ground truth divided by the area of union\nbetween the predicted segmentation and the ground truth. For binary (two classes) or multi-class segmentation,\nthe mean IoU of the image is calculated by taking the IoU of each class and averaging them.\n" _UpperCAmelCase = "\nArgs:\n predictions (`List[ndarray]`):\n List of predicted segmentation maps, each of shape (height, width). Each segmentation map can be of a different size.\n references (`List[ndarray]`):\n List of ground truth segmentation maps, each of shape (height, width). Each segmentation map can be of a different size.\n num_labels (`int`):\n Number of classes (categories).\n ignore_index (`int`):\n Index that will be ignored during evaluation.\n nan_to_num (`int`, *optional*):\n If specified, NaN values will be replaced by the number defined by the user.\n label_map (`dict`, *optional*):\n If specified, dictionary mapping old label indices to new label indices.\n reduce_labels (`bool`, *optional*, defaults to `False`):\n Whether or not to reduce all label values of segmentation maps by 1. Usually used for datasets where 0 is used for background,\n and background itself is not included in all classes of a dataset (e.g. ADE20k). The background label will be replaced by 255.\n\nReturns:\n `Dict[str, float | ndarray]` comprising various elements:\n - *mean_iou* (`float`):\n Mean Intersection-over-Union (IoU averaged over all categories).\n - *mean_accuracy* (`float`):\n Mean accuracy (averaged over all categories).\n - *overall_accuracy* (`float`):\n Overall accuracy on all images.\n - *per_category_accuracy* (`ndarray` of shape `(num_labels,)`):\n Per category accuracy.\n - *per_category_iou* (`ndarray` of shape `(num_labels,)`):\n Per category IoU.\n\nExamples:\n\n >>> import numpy as np\n\n >>> mean_iou = datasets.load_metric(\"mean_iou\")\n\n >>> # suppose one has 3 different segmentation maps predicted\n >>> predicted_1 = np.array([[1, 2], [3, 4], [5, 255]])\n >>> actual_1 = np.array([[0, 3], [5, 4], [6, 255]])\n\n >>> predicted_2 = np.array([[2, 7], [9, 2], [3, 6]])\n >>> actual_2 = np.array([[1, 7], [9, 2], [3, 6]])\n\n >>> predicted_3 = np.array([[2, 2, 3], [8, 2, 4], [3, 255, 2]])\n >>> actual_3 = np.array([[1, 2, 2], [8, 2, 1], [3, 255, 1]])\n\n >>> predicted = [predicted_1, predicted_2, predicted_3]\n >>> ground_truth = [actual_1, actual_2, actual_3]\n\n >>> results = mean_iou.compute(predictions=predicted, references=ground_truth, num_labels=10, ignore_index=255, reduce_labels=False)\n >>> print(results) # doctest: +NORMALIZE_WHITESPACE\n {'mean_iou': 0.47750000000000004, 'mean_accuracy': 0.5916666666666666, 'overall_accuracy': 0.5263157894736842, 'per_category_iou': array([0. , 0. , 0.375, 0.4 , 0.5 , 0. , 0.5 , 1. , 1. , 1. ]), 'per_category_accuracy': array([0. , 0. , 0.75 , 0.66666667, 1. , 0. , 0.5 , 1. , 1. , 1. ])}\n" _UpperCAmelCase = "\\n@software{MMSegmentation_Contributors_OpenMMLab_Semantic_Segmentation_2020,\nauthor = {{MMSegmentation Contributors}},\nlicense = {Apache-2.0},\nmonth = {7},\ntitle = {{OpenMMLab Semantic Segmentation Toolbox and Benchmark}},\nurl = {https://github.com/open-mmlab/mmsegmentation},\nyear = {2020}\n}" def __UpperCamelCase (lowerCAmelCase : Union[str, Any], lowerCAmelCase : Tuple, lowerCAmelCase : Tuple, lowerCAmelCase : bool, lowerCAmelCase : Optional[Dict[int, int]] = None, lowerCAmelCase : bool = False, ) -> Optional[Any]: if label_map is not None: for old_id, new_id in label_map.items(): A = new_id # turn into Numpy arrays A = np.array(lowerCAmelCase ) A = np.array(lowerCAmelCase ) if reduce_labels: A = 255 A = label - 1 A = 255 A = label != ignore_index A = np.not_equal(lowerCAmelCase, lowerCAmelCase ) A = pred_label[mask] A = np.array(lowerCAmelCase )[mask] A = pred_label[pred_label == label] A = np.histogram(lowerCAmelCase, bins=lowerCAmelCase, range=(0, num_labels - 1) )[0] A = np.histogram(lowerCAmelCase, bins=lowerCAmelCase, range=(0, num_labels - 1) )[0] A = np.histogram(lowerCAmelCase, bins=lowerCAmelCase, range=(0, num_labels - 1) )[0] A = area_pred_label + area_label - area_intersect return area_intersect, area_union, area_pred_label, area_label def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : List[str], lowerCAmelCase : str, lowerCAmelCase : bool, lowerCAmelCase : Optional[Dict[int, int]] = None, lowerCAmelCase : bool = False, ) -> str: A = np.zeros((num_labels,), dtype=np.floataa ) A = np.zeros((num_labels,), dtype=np.floataa ) A = np.zeros((num_labels,), dtype=np.floataa ) A = np.zeros((num_labels,), dtype=np.floataa ) for result, gt_seg_map in zip(lowerCAmelCase, lowerCAmelCase ): A , A , A , A = intersect_and_union( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) total_area_intersect += area_intersect total_area_union += area_union total_area_pred_label += area_pred_label total_area_label += area_label return total_area_intersect, total_area_union, total_area_pred_label, total_area_label def __UpperCamelCase (lowerCAmelCase : Tuple, lowerCAmelCase : Union[str, Any], lowerCAmelCase : int, lowerCAmelCase : bool, lowerCAmelCase : Optional[int] = None, lowerCAmelCase : Optional[Dict[int, int]] = None, lowerCAmelCase : bool = False, ) -> Any: A , A , A , A = total_intersect_and_union( lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ) # compute metrics A = {} A = total_area_intersect.sum() / total_area_label.sum() A = total_area_intersect / total_area_union A = total_area_intersect / total_area_label A = np.nanmean(lowerCAmelCase ) A = np.nanmean(lowerCAmelCase ) A = all_acc A = iou A = acc if nan_to_num is not None: A = {metric: np.nan_to_num(lowerCAmelCase, nan=lowerCAmelCase ) for metric, metric_value in metrics.items()} return metrics @datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION ) class _UpperCAmelCase ( datasets.Metric ): '''simple docstring''' def UpperCamelCase ( self : Dict ): return datasets.MetricInfo( description=_DESCRIPTION , citation=_CITATION , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features( # 1st Seq - height dim, 2nd - width dim { 'predictions': datasets.Sequence(datasets.Sequence(datasets.Value('uint16' ) ) ), 'references': datasets.Sequence(datasets.Sequence(datasets.Value('uint16' ) ) ), } ) , reference_urls=[ 'https://github.com/open-mmlab/mmsegmentation/blob/71c201b1813267d78764f306a297ca717827c4bf/mmseg/core/evaluation/metrics.py' ] , ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : bool , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : Optional[Dict[int, int]] = None , UpperCamelCase__ : bool = False , ): A = mean_iou( results=UpperCamelCase__ , gt_seg_maps=UpperCamelCase__ , num_labels=UpperCamelCase__ , ignore_index=UpperCamelCase__ , nan_to_num=UpperCamelCase__ , label_map=UpperCamelCase__ , reduce_labels=UpperCamelCase__ , ) return iou_result
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> str: 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=10))
699
1
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 _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Tuple = '''longformer''' def __init__( self : Dict , UpperCamelCase__ : Union[List[int], int] = 512 , UpperCamelCase__ : int = 2 , UpperCamelCase__ : int = 1 , UpperCamelCase__ : int = 0 , UpperCamelCase__ : int = 2 , UpperCamelCase__ : int = 30522 , UpperCamelCase__ : int = 768 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 3072 , UpperCamelCase__ : str = "gelu" , UpperCamelCase__ : float = 0.1 , UpperCamelCase__ : float = 0.1 , UpperCamelCase__ : int = 512 , UpperCamelCase__ : int = 2 , UpperCamelCase__ : float = 0.02 , UpperCamelCase__ : float = 1e-1_2 , UpperCamelCase__ : bool = False , **UpperCamelCase__ : Dict , ): super().__init__(pad_token_id=UpperCamelCase__ , **UpperCamelCase__ ) A = attention_window A = sep_token_id A = bos_token_id A = eos_token_id A = vocab_size A = hidden_size A = num_hidden_layers A = num_attention_heads A = hidden_act A = intermediate_size A = hidden_dropout_prob A = attention_probs_dropout_prob A = max_position_embeddings A = type_vocab_size A = initializer_range A = layer_norm_eps A = onnx_export class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : "PretrainedConfig" , UpperCamelCase__ : str = "default" , UpperCamelCase__ : "List[PatchingSpec]" = None ): super().__init__(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ ) A = True @property def UpperCamelCase ( self : Optional[Any] ): if self.task == "multiple-choice": A = {0: 'batch', 1: 'choice', 2: 'sequence'} else: A = {0: 'batch', 1: 'sequence'} return OrderedDict( [ ('input_ids', dynamic_axis), ('attention_mask', dynamic_axis), ('global_attention_mask', dynamic_axis), ] ) @property def UpperCamelCase ( self : Union[str, Any] ): A = super().outputs if self.task == "default": A = {0: 'batch'} return outputs @property def UpperCamelCase ( self : Tuple ): return 1e-4 @property def UpperCamelCase ( self : Any ): # needs to be >= 14 to support tril operator return max(super().default_onnx_opset , 14 ) def UpperCamelCase ( self : str , UpperCamelCase__ : "PreTrainedTokenizerBase" , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ): A = super().generate_dummy_inputs( preprocessor=UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ ) 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 A = torch.zeros_like(inputs['input_ids'] ) # make every second token global A = 1 return inputs
699
import math import torch from torch import nn from ..configuration_utils import ConfigMixin, register_to_config from .attention_processor import Attention from .embeddings import get_timestep_embedding from .modeling_utils import ModelMixin class _UpperCAmelCase ( __lowercase , __lowercase ): '''simple docstring''' @register_to_config def __init__( self : Any , UpperCamelCase__ : int = 128 , UpperCamelCase__ : int = 256 , UpperCamelCase__ : float = 2_000.0 , UpperCamelCase__ : int = 768 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 12 , UpperCamelCase__ : int = 64 , UpperCamelCase__ : int = 2048 , UpperCamelCase__ : float = 0.1 , ): super().__init__() A = nn.Sequential( nn.Linear(UpperCamelCase__ , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , nn.Linear(d_model * 4 , d_model * 4 , bias=UpperCamelCase__ ) , nn.SiLU() , ) A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = False A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.ModuleList() for lyr_num in range(UpperCamelCase__ ): # FiLM conditional T5 decoder A = DecoderLayer(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) self.decoders.append(UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ ) A = nn.Dropout(p=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Any , UpperCamelCase__ : int ): A = torch.mul(query_input.unsqueeze(-1 ) , key_input.unsqueeze(-2 ) ) return mask.unsqueeze(-3 ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Optional[int] ): A , A , A = decoder_input_tokens.shape assert decoder_noise_time.shape == (batch,) # decoder_noise_time is in [0, 1), so rescale to expected timing range. A = get_timestep_embedding( decoder_noise_time * self.config.max_decoder_noise_time , embedding_dim=self.config.d_model , max_period=self.config.max_decoder_noise_time , ).to(dtype=self.dtype ) A = self.conditioning_emb(UpperCamelCase__ ).unsqueeze(1 ) assert conditioning_emb.shape == (batch, 1, self.config.d_model * 4) A = decoder_input_tokens.shape[1] # If we want to use relative positions for audio context, we can just offset # this sequence by the length of encodings_and_masks. A = torch.broadcast_to( torch.arange(UpperCamelCase__ , device=decoder_input_tokens.device ) , (batch, seq_length) , ) A = self.position_encoding(UpperCamelCase__ ) A = self.continuous_inputs_projection(UpperCamelCase__ ) inputs += position_encodings A = self.dropout(UpperCamelCase__ ) # decoder: No padding present. A = torch.ones( decoder_input_tokens.shape[:2] , device=decoder_input_tokens.device , dtype=inputs.dtype ) # Translate encoding masks to encoder-decoder masks. A = [(x, self.encoder_decoder_mask(UpperCamelCase__ , UpperCamelCase__ )) for x, y in encodings_and_masks] # cross attend style: concat encodings A = torch.cat([x[0] for x in encodings_and_encdec_masks] , dim=1 ) A = torch.cat([x[1] for x in encodings_and_encdec_masks] , dim=-1 ) for lyr in self.decoders: A = lyr( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , encoder_attention_mask=UpperCamelCase__ , )[0] A = self.decoder_norm(UpperCamelCase__ ) A = self.post_dropout(UpperCamelCase__ ) A = self.spec_out(UpperCamelCase__ ) return spec_out class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Union[str, Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Union[str, Any]=1e-6 ): super().__init__() A = nn.ModuleList() # cond self attention: layer 0 self.layer.append( TaLayerSelfAttentionCond(d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) ) # cross attention: layer 1 self.layer.append( TaLayerCrossAttention( d_model=UpperCamelCase__ , d_kv=UpperCamelCase__ , num_heads=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ , ) ) # Film Cond MLP + dropout: last layer self.layer.append( TaLayerFFCond(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ , layer_norm_epsilon=UpperCamelCase__ ) ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Optional[int]=None , UpperCamelCase__ : int=None , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[Any]=None , ): A = self.layer[0]( UpperCamelCase__ , conditioning_emb=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) if encoder_hidden_states is not None: A = torch.where(encoder_attention_mask > 0 , 0 , -1e1_0 ).to( encoder_hidden_states.dtype ) A = self.layer[1]( UpperCamelCase__ , key_value_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , ) # Apply Film Conditional Feed Forward layer A = self.layer[-1](UpperCamelCase__ , UpperCamelCase__ ) return (hidden_states,) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : Tuple , UpperCamelCase__ : str , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Dict ): super().__init__() A = TaLayerNorm(UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : str , UpperCamelCase__ : Optional[Any]=None , UpperCamelCase__ : Tuple=None , ): # pre_self_attention_layer_norm A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.FiLMLayer(UpperCamelCase__ , UpperCamelCase__ ) # Self-attention block A = self.attention(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : List[str] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[Any] ): super().__init__() A = Attention(query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , out_bias=UpperCamelCase__ , scale_qk=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Dict , UpperCamelCase__ : List[Any]=None , UpperCamelCase__ : List[str]=None , ): A = self.layer_norm(UpperCamelCase__ ) A = self.attention( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=attention_mask.squeeze(1 ) , ) A = hidden_states + self.dropout(UpperCamelCase__ ) return layer_output class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any ): super().__init__() A = TaDenseGatedActDense(d_model=UpperCamelCase__ , d_ff=UpperCamelCase__ , dropout_rate=UpperCamelCase__ ) A = TaFiLMLayer(in_features=d_model * 4 , out_features=UpperCamelCase__ ) A = TaLayerNorm(UpperCamelCase__ , eps=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) def UpperCamelCase ( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Any=None ): A = self.layer_norm(UpperCamelCase__ ) if conditioning_emb is not None: A = self.film(UpperCamelCase__ , UpperCamelCase__ ) A = self.DenseReluDense(UpperCamelCase__ ) A = hidden_states + self.dropout(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Any , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ , bias=UpperCamelCase__ ) A = nn.Dropout(UpperCamelCase__ ) A = NewGELUActivation() def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : List[Any] ): A = self.act(self.wi_a(UpperCamelCase__ ) ) A = self.wi_a(UpperCamelCase__ ) A = hidden_gelu * hidden_linear A = self.dropout(UpperCamelCase__ ) A = self.wo(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Tuple=1e-6 ): super().__init__() A = nn.Parameter(torch.ones(UpperCamelCase__ ) ) A = eps def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : int ): # T5 uses a layer_norm which only scales and doesn't shift, which is also known as Root Mean # Square Layer Normalization https://arxiv.org/abs/1910.07467 thus variance is calculated # w/o mean and there is no bias. Additionally we want to make sure that the accumulation for # half-precision inputs is done in fp32 A = hidden_states.to(torch.floataa ).pow(2 ).mean(-1 , keepdim=UpperCamelCase__ ) A = hidden_states * torch.rsqrt(variance + self.variance_epsilon ) # convert into half-precision if necessary if self.weight.dtype in [torch.floataa, torch.bfloataa]: A = hidden_states.to(self.weight.dtype ) return self.weight * hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def UpperCamelCase ( self : Any , UpperCamelCase__ : torch.Tensor ): return 0.5 * input * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi ) * (input + 0.044_715 * torch.pow(UpperCamelCase__ , 3.0 )) )) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Dict , UpperCamelCase__ : Dict , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , out_features * 2 , bias=UpperCamelCase__ ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : List[Any] ): A = self.scale_bias(UpperCamelCase__ ) A , A = torch.chunk(UpperCamelCase__ , 2 , -1 ) A = x * (1 + scale) + shift return x
699
1
import os import re import shutil import sys import tempfile import unittest import black _UpperCAmelCase = os.path.abspath(os.path.dirname(os.path.dirname(os.path.dirname(__file__)))) sys.path.append(os.path.join(git_repo_path, "utils")) import check_copies # noqa: E402 # This is the reference code that will be used in the tests. # If BertLMPredictionHead is changed in modeling_bert.py, this code needs to be manually updated. _UpperCAmelCase = " def __init__(self, config):\n super().__init__()\n self.transform = BertPredictionHeadTransform(config)\n\n # The output weights are the same as the input embeddings, but there is\n # an output-only bias for each token.\n self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=False)\n\n self.bias = nn.Parameter(torch.zeros(config.vocab_size))\n\n # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings`\n self.decoder.bias = self.bias\n\n def forward(self, hidden_states):\n hidden_states = self.transform(hidden_states)\n hidden_states = self.decoder(hidden_states)\n return hidden_states\n" class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : Dict ): A = tempfile.mkdtemp() os.makedirs(os.path.join(self.transformer_dir , 'models/bert/' ) ) A = self.transformer_dir shutil.copy( os.path.join(UpperCamelCase__ , 'src/transformers/models/bert/modeling_bert.py' ) , os.path.join(self.transformer_dir , 'models/bert/modeling_bert.py' ) , ) def UpperCamelCase ( self : Optional[Any] ): A = 'src/transformers' shutil.rmtree(self.transformer_dir ) def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : int , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : str , UpperCamelCase__ : int=None ): A = comment + f'''\nclass {class_name}(nn.Module):\n''' + class_code if overwrite_result is not None: A = comment + f'''\nclass {class_name}(nn.Module):\n''' + overwrite_result A = black.Mode(target_versions={black.TargetVersion.PYaa} , line_length=119 ) A = black.format_str(UpperCamelCase__ , mode=UpperCamelCase__ ) A = os.path.join(self.transformer_dir , 'new_code.py' ) with open(UpperCamelCase__ , 'w' , newline='\n' ) as f: f.write(UpperCamelCase__ ) if overwrite_result is None: self.assertTrue(len(check_copies.is_copy_consistent(UpperCamelCase__ ) ) == 0 ) else: check_copies.is_copy_consistent(f.name , overwrite=UpperCamelCase__ ) with open(UpperCamelCase__ , 'r' ) as f: self.assertTrue(f.read() , UpperCamelCase__ ) def UpperCamelCase ( self : int ): A = check_copies.find_code_in_transformers('models.bert.modeling_bert.BertLMPredictionHead' ) self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Union[str, Any] ): # Base copy consistency self.check_copy_consistency( '# Copied from transformers.models.bert.modeling_bert.BertLMPredictionHead' , 'BertLMPredictionHead' , REFERENCE_CODE + '\n' , ) # With no empty line at the end self.check_copy_consistency( '# Copied from transformers.models.bert.modeling_bert.BertLMPredictionHead' , 'BertLMPredictionHead' , UpperCamelCase__ , ) # Copy consistency with rename self.check_copy_consistency( '# Copied from transformers.models.bert.modeling_bert.BertLMPredictionHead with Bert->TestModel' , 'TestModelLMPredictionHead' , re.sub('Bert' , 'TestModel' , UpperCamelCase__ ) , ) # Copy consistency with a really long name A = 'TestModelWithAReallyLongNameBecauseSomePeopleLikeThatForSomeReason' self.check_copy_consistency( f'''# Copied from transformers.models.bert.modeling_bert.BertLMPredictionHead with Bert->{long_class_name}''' , f'''{long_class_name}LMPredictionHead''' , re.sub('Bert' , UpperCamelCase__ , UpperCamelCase__ ) , ) # Copy consistency with overwrite self.check_copy_consistency( '# Copied from transformers.models.bert.modeling_bert.BertLMPredictionHead with Bert->TestModel' , 'TestModelLMPredictionHead' , UpperCamelCase__ , overwrite_result=re.sub('Bert' , 'TestModel' , UpperCamelCase__ ) , ) def UpperCamelCase ( self : int ): A = check_copies.LOCALIZED_READMES['README_zh-hans.md'] A = ( '1. **[ALBERT](https://huggingface.co/transformers/model_doc/albert.html)** (from Google Research and the' ' Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for' ' Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong' ' Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.\n1.' ' **[DistilBERT](https://huggingface.co/transformers/model_doc/distilbert.html)** (from HuggingFace),' ' released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and' ' lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same' ' method has been applied to compress GPT2 into' ' [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into' ' [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation),' ' Multilingual BERT into' ' [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German' ' version of DistilBERT.\n1. **[ELECTRA](https://huggingface.co/transformers/model_doc/electra.html)**' ' (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders' ' as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang' ' Luong, Quoc V. Le, Christopher D. Manning.' ) A = ( '1. **[ALBERT](https://huggingface.co/transformers/model_doc/albert.html)** (来自 Google Research and the' ' Toyota Technological Institute at Chicago) 伴随论文 [ALBERT: A Lite BERT for Self-supervised Learning of' ' Language Representations](https://arxiv.org/abs/1909.11942), 由 Zhenzhong Lan, Mingda Chen, Sebastian' ' Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut 发布。\n' ) A = ( '1. **[ALBERT](https://huggingface.co/transformers/model_doc/albert.html)** (来自 Google Research and the' ' Toyota Technological Institute at Chicago) 伴随论文 [ALBERT: A Lite BERT for Self-supervised Learning of' ' Language Representations](https://arxiv.org/abs/1909.11942), 由 Zhenzhong Lan, Mingda Chen, Sebastian' ' Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut 发布。\n1.' ' **[DistilBERT](https://huggingface.co/transformers/model_doc/distilbert.html)** (来自 HuggingFace) 伴随论文' ' [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and' ' lighter](https://arxiv.org/abs/1910.01108) 由 Victor Sanh, Lysandre Debut and Thomas Wolf 发布。 The same' ' method has been applied to compress GPT2 into' ' [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into' ' [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation),' ' Multilingual BERT into' ' [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German' ' version of DistilBERT.\n1. **[ELECTRA](https://huggingface.co/transformers/model_doc/electra.html)** (来自' ' Google Research/Stanford University) 伴随论文 [ELECTRA: Pre-training text encoders as discriminators rather' ' than generators](https://arxiv.org/abs/2003.10555) 由 Kevin Clark, Minh-Thang Luong, Quoc V. Le,' ' Christopher D. Manning 发布。\n' ) A , A = check_copies.convert_to_localized_md( UpperCamelCase__ , UpperCamelCase__ , localized_readme['format_model_list'] ) self.assertFalse(UpperCamelCase__ ) self.assertEqual(UpperCamelCase__ , UpperCamelCase__ ) A , A = check_copies.convert_to_localized_md( UpperCamelCase__ , UpperCamelCase__ , localized_readme['format_model_list'] ) # Check whether the number of models is equal to README.md after conversion. self.assertTrue(UpperCamelCase__ ) A = ( '1. **[ALBERT](https://huggingface.co/transformers/model_doc/albert.html)** (from Google Research and the' ' Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for' ' Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong' ' Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.' ) A = ( '1. **[ALBERT](https://huggingface.co/transformers/main/model_doc/albert.html)** (来自 Google Research and' ' the Toyota Technological Institute at Chicago) 伴随论文 [ALBERT: A Lite BERT for Self-supervised Learning of' ' Language Representations](https://arxiv.org/abs/1909.11942), 由 Zhenzhong Lan, Mingda Chen, Sebastian' ' Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut 发布。\n' ) A = ( '1. **[ALBERT](https://huggingface.co/transformers/model_doc/albert.html)** (来自 Google Research and the' ' Toyota Technological Institute at Chicago) 伴随论文 [ALBERT: A Lite BERT for Self-supervised Learning of' ' Language Representations](https://arxiv.org/abs/1909.11942), 由 Zhenzhong Lan, Mingda Chen, Sebastian' ' Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut 发布。\n' ) A , A = check_copies.convert_to_localized_md( UpperCamelCase__ , UpperCamelCase__ , localized_readme['format_model_list'] ) # Check if the model link is synchronized. self.assertEqual(UpperCamelCase__ , UpperCamelCase__ )
699
import numpy as np import torch from imwatermark import WatermarkEncoder # Copied from https://github.com/Stability-AI/generative-models/blob/613af104c6b85184091d42d374fef420eddb356d/scripts/demo/streamlit_helpers.py#L66 _UpperCAmelCase = 0b10_11_00_11_11_10_11_00_10_01_00_00_01_11_10_11_10_11_00_01_10_01_11_10 # bin(x)[2:] gives bits of x as str, use int to convert them to 0/1 _UpperCAmelCase = [int(bit) for bit in bin(WATERMARK_MESSAGE)[2:]] class _UpperCAmelCase : '''simple docstring''' def __init__( self : Union[str, Any] ): A = WATERMARK_BITS A = WatermarkEncoder() self.encoder.set_watermark('bits' , self.watermark ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : torch.FloatTensor ): # can't encode images that are smaller than 256 if images.shape[-1] < 256: return images A = (255 * (images / 2 + 0.5)).cpu().permute(0 , 2 , 3 , 1 ).float().numpy() A = [self.encoder.encode(UpperCamelCase__ , 'dwtDct' ) for image in images] A = torch.from_numpy(np.array(UpperCamelCase__ ) ).permute(0 , 3 , 1 , 2 ) A = torch.clamp(2 * (images / 255 - 0.5) , min=-1.0 , max=1.0 ) return images
699
1
from __future__ import annotations import math from collections.abc import Callable def __UpperCamelCase (lowerCAmelCase : Callable[[int | float], int | float], lowerCAmelCase : int | float, lowerCAmelCase : int | float, lowerCAmelCase : int = 100, ) -> float: A = x_start A = fnc(lowerCAmelCase ) A = 0.0 for _ in range(lowerCAmelCase ): # Approximates curve as a sequence of linear lines and sums their length A = (x_end - x_start) / steps + xa A = fnc(lowerCAmelCase ) length += math.hypot(xa - xa, fxa - fxa ) # Increment step A = xa A = fxa return length if __name__ == "__main__": def __UpperCamelCase (lowerCAmelCase : List[str] ) -> Dict: return math.sin(10 * x ) print("f(x) = sin(10 * x)") print("The length of the curve from x = -10 to x = 10 is:") _UpperCAmelCase = 10 while i <= 100_000: print(F'''With {i} steps: {line_length(f, -10, 10, i)}''') i *= 10
699
import argparse import fairseq import torch from transformers import UniSpeechSatConfig, UniSpeechSatForCTC, UniSpeechSatForPreTraining, logging logging.set_verbosity_info() _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "post_extract_proj": "feature_projection.projection", "encoder.pos_conv.0": "encoder.pos_conv_embed.conv", "self_attn.k_proj": "encoder.layers.*.attention.k_proj", "self_attn.v_proj": "encoder.layers.*.attention.v_proj", "self_attn.q_proj": "encoder.layers.*.attention.q_proj", "self_attn.out_proj": "encoder.layers.*.attention.out_proj", "self_attn_layer_norm": "encoder.layers.*.layer_norm", "fc1": "encoder.layers.*.feed_forward.intermediate_dense", "fc2": "encoder.layers.*.feed_forward.output_dense", "final_layer_norm": "encoder.layers.*.final_layer_norm", "encoder.layer_norm": "encoder.layer_norm", "encoder.layer_norm_for_extract": "layer_norm_for_extract", "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": "lm_head", "label_embs_concat": "label_embeddings_concat", "mask_emb": "masked_spec_embed", "spk_proj": "speaker_proj", } _UpperCAmelCase = [ "lm_head", "quantizer.weight_proj", "quantizer.codevectors", "project_q", "project_hid", "label_embeddings_concat", "speaker_proj", "layer_norm_for_extract", ] def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : Dict, lowerCAmelCase : Optional[int], lowerCAmelCase : List[Any], lowerCAmelCase : str ) -> int: 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 if hf_shape != value.shape: raise ValueError( 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 __UpperCamelCase (lowerCAmelCase : List[str], lowerCAmelCase : Optional[int] ) -> Dict: A = [] A = fairseq_model.state_dict() A = hf_model.unispeech_sat.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 = 'unispeech_sat.' + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key if key in name or key.split('w2v_model.' )[-1] == name.split('.' )[0]: if "layer_norm_for_extract" in name and (".".join(name.split('.' )[:-1] ) != key): # special case since naming is very similar continue 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 "bias" in name: A = 'bias' elif "weight" in name: # TODO: don't match quantizer.weight_proj A = 'weight' 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 __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : Tuple, lowerCAmelCase : List[Any], lowerCAmelCase : int ) -> Dict: 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.bias.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].conv.weight.data.shape: raise ValueError( 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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape: raise ValueError( f'''{full_name} has size {value.shape}, but''' f''' {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: if value.shape != feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape: raise ValueError( 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 ) @torch.no_grad() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : Dict, lowerCAmelCase : Union[str, Any]=None, lowerCAmelCase : str=None, lowerCAmelCase : List[Any]=True ) -> Union[str, Any]: if config_path is not None: A = UniSpeechSatConfig.from_pretrained(lowerCAmelCase ) else: A = UniSpeechSatConfig() A = '' if is_finetuned: A = UniSpeechSatForCTC(lowerCAmelCase ) else: A = UniSpeechSatForPreTraining(lowerCAmelCase ) A , A , A = fairseq.checkpoint_utils.load_model_ensemble_and_task( [checkpoint_path], arg_overrides={'data': '/'.join(dict_path.split('/' )[:-1] )} ) A = model[0].eval() recursively_load_weights(lowerCAmelCase, lowerCAmelCase ) hf_wavavec.save_pretrained(lowerCAmelCase ) if __name__ == "__main__": _UpperCAmelCase = argparse.ArgumentParser() parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.") parser.add_argument("--checkpoint_path", default=None, type=str, help="Path to fairseq checkpoint") parser.add_argument("--dict_path", default=None, type=str, help="Path to dict of fine-tuned model") parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert") parser.add_argument( "--not_finetuned", action="store_true", help="Whether the model to convert is a fine-tuned model or not" ) _UpperCAmelCase = parser.parse_args() convert_unispeech_sat_checkpoint( args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned )
699
1
def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str ) -> str: if not (isinstance(lowerCAmelCase, lowerCAmelCase ) and isinstance(lowerCAmelCase, lowerCAmelCase )): raise ValueError('longest_common_substring() takes two strings for inputs' ) A = len(lowerCAmelCase ) A = len(lowerCAmelCase ) A = [[0] * (texta_length + 1) for _ in range(texta_length + 1 )] A = 0 A = 0 for i in range(1, texta_length + 1 ): for j in range(1, texta_length + 1 ): if texta[i - 1] == texta[j - 1]: A = 1 + dp[i - 1][j - 1] if dp[i][j] > ans_length: A = i A = dp[i][j] return texta[ans_index - ans_length : ans_index] if __name__ == "__main__": import doctest doctest.testmod()
699
from __future__ import annotations from collections.abc import Iterator from typing import Generic, TypeVar _UpperCAmelCase = TypeVar("T") class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple , UpperCamelCase__ : T ): A = data A = None def __str__( self : Optional[int] ): return f'''{self.data}''' class _UpperCAmelCase ( Generic[T] ): '''simple docstring''' def __init__( self : Tuple ): A = None def __iter__( self : int ): A = self.top while node: yield node.data A = node.next def __str__( self : Any ): return "->".join([str(UpperCamelCase__ ) for item in self] ) def __len__( self : Dict ): return len(tuple(iter(self ) ) ) def UpperCamelCase ( self : List[str] ): return self.top is None def UpperCamelCase ( self : Dict , UpperCamelCase__ : T ): A = Node(UpperCamelCase__ ) if not self.is_empty(): A = self.top A = node def UpperCamelCase ( self : Dict ): if self.is_empty(): raise IndexError('pop from empty stack' ) assert isinstance(self.top , UpperCamelCase__ ) A = self.top A = self.top.next return pop_node.data def UpperCamelCase ( self : List[str] ): if self.is_empty(): raise IndexError('peek from empty stack' ) assert self.top is not None return self.top.data def UpperCamelCase ( self : List[str] ): A = None if __name__ == "__main__": from doctest import testmod testmod()
699
1
from timeit import timeit def __UpperCamelCase (lowerCAmelCase : int ) -> int: if number < 0: raise ValueError('the value of input must not be negative' ) A = 0 while number: number &= number - 1 result += 1 return result def __UpperCamelCase (lowerCAmelCase : int ) -> int: if number < 0: raise ValueError('the value of input must not be negative' ) A = 0 while number: if number % 2 == 1: result += 1 number >>= 1 return result def __UpperCamelCase () -> None: def do_benchmark(lowerCAmelCase : int ) -> None: A = 'import __main__ as z' print(f'''Benchmark when {number = }:''' ) print(f'''{get_set_bits_count_using_modulo_operator(lowerCAmelCase ) = }''' ) A = timeit('z.get_set_bits_count_using_modulo_operator(25)', setup=lowerCAmelCase ) print(f'''timeit() runs in {timing} seconds''' ) print(f'''{get_set_bits_count_using_brian_kernighans_algorithm(lowerCAmelCase ) = }''' ) A = timeit( 'z.get_set_bits_count_using_brian_kernighans_algorithm(25)', setup=lowerCAmelCase, ) print(f'''timeit() runs in {timing} seconds''' ) for number in (25, 37, 58, 0): do_benchmark(lowerCAmelCase ) print() if __name__ == "__main__": import doctest doctest.testmod() benchmark()
699
from __future__ import annotations import math def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int, lowerCAmelCase : bool, lowerCAmelCase : list[int], lowerCAmelCase : float ) -> int: if depth < 0: raise ValueError('Depth cannot be less than 0' ) if not scores: raise ValueError('Scores cannot be empty' ) if depth == height: return scores[node_index] return ( max( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) if is_max else min( minimax(depth + 1, node_index * 2, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), minimax(depth + 1, node_index * 2 + 1, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase ), ) ) def __UpperCamelCase () -> None: A = [90, 23, 6, 33, 21, 65, 123, 34_423] A = math.log(len(lowerCAmelCase ), 2 ) print(f'''Optimal value : {minimax(0, 0, lowerCAmelCase, lowerCAmelCase, lowerCAmelCase )}''' ) if __name__ == "__main__": import doctest doctest.testmod() main()
699
1
def __UpperCamelCase (lowerCAmelCase : int = 1_000 ) -> int: return sum(e for e in range(3, lowerCAmelCase ) if e % 3 == 0 or e % 5 == 0 ) if __name__ == "__main__": print(F'''{solution() = }''')
699
import logging import os from dataclasses import dataclass, field from typing import Dict, Optional import datasets import numpy as np import tensorflow as tf from transformers import ( AutoConfig, AutoTokenizer, EvalPrediction, HfArgumentParser, PreTrainedTokenizer, TFAutoModelForSequenceClassification, TFTrainer, TFTrainingArguments, ) from transformers.utils import logging as hf_logging hf_logging.set_verbosity_info() hf_logging.enable_default_handler() hf_logging.enable_explicit_format() def __UpperCamelCase (lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : str, lowerCAmelCase : PreTrainedTokenizer, lowerCAmelCase : int, lowerCAmelCase : Optional[int] = None, ) -> Dict: A = {} if train_file is not None: A = [train_file] if eval_file is not None: A = [eval_file] if test_file is not None: A = [test_file] A = datasets.load_dataset('csv', data_files=lowerCAmelCase ) A = list(ds[list(files.keys() )[0]].features.keys() ) A = features_name.pop(lowerCAmelCase ) A = list(set(ds[list(files.keys() )[0]][label_name] ) ) A = {label: i for i, label in enumerate(lowerCAmelCase )} A = tokenizer.model_input_names A = {} if len(lowerCAmelCase ) == 1: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( example[features_name[0]], truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length' ), batched=lowerCAmelCase, ) elif len(lowerCAmelCase ) == 2: for k in files.keys(): A = ds[k].map( lambda lowerCAmelCase : tokenizer.batch_encode_plus( (example[features_name[0]], example[features_name[1]]), truncation=lowerCAmelCase, max_length=lowerCAmelCase, padding='max_length', ), batched=lowerCAmelCase, ) def gen_train(): for ex in transformed_ds[datasets.Split.TRAIN]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_val(): for ex in transformed_ds[datasets.Split.VALIDATION]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) def gen_test(): for ex in transformed_ds[datasets.Split.TEST]: A = {k: v for k, v in ex.items() if k in input_names} A = labelaid[ex[label_name]] yield (d, label) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TRAIN in transformed_ds else None ) if train_ds is not None: A = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.VALIDATION in transformed_ds else None ) if val_ds is not None: A = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) ) A = ( tf.data.Dataset.from_generator( lowerCAmelCase, ({k: tf.intaa for k in input_names}, tf.intaa), ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )), ) if datasets.Split.TEST in transformed_ds else None ) if test_ds is not None: A = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) ) return train_ds, val_ds, test_ds, labelaid _UpperCAmelCase = logging.getLogger(__name__) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : int = field(metadata={'''help''': '''Which column contains the label'''} ) SCREAMING_SNAKE_CASE : str = field(default=__lowercase , metadata={'''help''': '''The path of the training file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the development file'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field(default=__lowercase , metadata={'''help''': '''The path of the test file'''} ) SCREAMING_SNAKE_CASE : int = field( default=1_28 , metadata={ '''help''': ( '''The maximum total input sequence length after tokenization. Sequences longer ''' '''than this will be truncated, sequences shorter will be padded.''' ) } , ) SCREAMING_SNAKE_CASE : bool = field( default=__lowercase , metadata={'''help''': '''Overwrite the cached training and evaluation sets'''} ) @dataclass class _UpperCAmelCase : '''simple docstring''' SCREAMING_SNAKE_CASE : str = field( metadata={'''help''': '''Path to pretrained model or model identifier from huggingface.co/models'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained config name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Pretrained tokenizer name or path if not the same as model_name'''} ) SCREAMING_SNAKE_CASE : bool = field(default=__lowercase , metadata={'''help''': '''Set this flag to use fast tokenization.'''} ) # If you want to tweak more attributes on your tokenizer, you should do it in a distinct script, # or just modify its tokenizer_config.json. SCREAMING_SNAKE_CASE : Optional[str] = field( default=__lowercase , metadata={'''help''': '''Where do you want to store the pretrained models downloaded from huggingface.co'''} , ) def __UpperCamelCase () -> Any: # See all possible arguments in src/transformers/training_args.py # or by passing the --help flag to this script. # We now keep distinct sets of args, for a cleaner separation of concerns. A = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) ) A , A , A = parser.parse_args_into_dataclasses() if ( os.path.exists(training_args.output_dir ) and os.listdir(training_args.output_dir ) and training_args.do_train and not training_args.overwrite_output_dir ): raise ValueError( f'''Output directory ({training_args.output_dir}) already exists and is not empty. Use''' ' --overwrite_output_dir to overcome.' ) # Setup logging logging.basicConfig( format='%(asctime)s - %(levelname)s - %(name)s - %(message)s', datefmt='%m/%d/%Y %H:%M:%S', level=logging.INFO, ) logger.info( f'''n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, ''' f'''16-bits training: {training_args.fpaa}''' ) logger.info(f'''Training/evaluation parameters {training_args}''' ) # Load pretrained model and tokenizer # # Distributed training: # The .from_pretrained methods guarantee that only one local process can concurrently # download model & vocab. A = AutoTokenizer.from_pretrained( model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path, cache_dir=model_args.cache_dir, ) A , A , A , A = get_tfds( train_file=data_args.train_file, eval_file=data_args.dev_file, test_file=data_args.test_file, tokenizer=lowerCAmelCase, label_column_id=data_args.label_column_id, max_seq_length=data_args.max_seq_length, ) A = AutoConfig.from_pretrained( model_args.config_name if model_args.config_name else model_args.model_name_or_path, num_labels=len(lowerCAmelCase ), labelaid=lowerCAmelCase, idalabel={id: label for label, id in labelaid.items()}, finetuning_task='text-classification', cache_dir=model_args.cache_dir, ) with training_args.strategy.scope(): A = TFAutoModelForSequenceClassification.from_pretrained( model_args.model_name_or_path, from_pt=bool('.bin' in model_args.model_name_or_path ), config=lowerCAmelCase, cache_dir=model_args.cache_dir, ) def compute_metrics(lowerCAmelCase : EvalPrediction ) -> Dict: A = np.argmax(p.predictions, axis=1 ) return {"acc": (preds == p.label_ids).mean()} # Initialize our Trainer A = TFTrainer( model=lowerCAmelCase, args=lowerCAmelCase, train_dataset=lowerCAmelCase, eval_dataset=lowerCAmelCase, compute_metrics=lowerCAmelCase, ) # Training if training_args.do_train: trainer.train() trainer.save_model() tokenizer.save_pretrained(training_args.output_dir ) # Evaluation A = {} if training_args.do_eval: logger.info('*** Evaluate ***' ) A = trainer.evaluate() A = os.path.join(training_args.output_dir, 'eval_results.txt' ) with open(lowerCAmelCase, 'w' ) as writer: logger.info('***** Eval results *****' ) for key, value in result.items(): logger.info(f''' {key} = {value}''' ) writer.write(f'''{key} = {value}\n''' ) results.update(lowerCAmelCase ) return results if __name__ == "__main__": main()
699
1
def __UpperCamelCase (lowerCAmelCase : list[int] ) -> int: if not numbers: return 0 if not isinstance(lowerCAmelCase, (list, tuple) ) or not all( isinstance(lowerCAmelCase, lowerCAmelCase ) for number in numbers ): raise ValueError('numbers must be an iterable of integers' ) A = A = A = numbers[0] for i in range(1, len(lowerCAmelCase ) ): # update the maximum and minimum subarray products A = numbers[i] if number < 0: A , A = min_till_now, max_till_now A = max(lowerCAmelCase, max_till_now * number ) A = min(lowerCAmelCase, min_till_now * number ) # update the maximum product found till now A = max(lowerCAmelCase, lowerCAmelCase ) return max_prod
699
from typing import TYPE_CHECKING from ...utils import ( OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_sentencepiece_available, is_tf_available, is_tokenizers_available, is_torch_available, ) _UpperCAmelCase = {"configuration_xglm": ["XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP", "XGLMConfig"]} try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizer"] try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = ["XGLMTokenizerFast"] try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "XGLMForCausalLM", "XGLMModel", "XGLMPreTrainedModel", ] try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "FlaxXGLMForCausalLM", "FlaxXGLMModel", "FlaxXGLMPreTrainedModel", ] try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: _UpperCAmelCase = [ "TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST", "TFXGLMForCausalLM", "TFXGLMModel", "TFXGLMPreTrainedModel", ] if TYPE_CHECKING: from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig try: if not is_sentencepiece_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm import XGLMTokenizer try: if not is_tokenizers_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .tokenization_xglm_fast import XGLMTokenizerFast try: if not is_torch_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel try: if not is_flax_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel try: if not is_tf_available(): raise OptionalDependencyNotAvailable() except OptionalDependencyNotAvailable: pass else: from .modeling_tf_xglm import ( TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, TFXGLMForCausalLM, TFXGLMModel, TFXGLMPreTrainedModel, ) else: import sys _UpperCAmelCase = _LazyModule(__name__, globals()["__file__"], _import_structure)
699
1
from ..utils import DummyObject, requires_backends class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[Any] = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : Dict , **UpperCamelCase__ : Any ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Tuple = ['''sentencepiece'''] def __init__( self : Union[str, Any] , *UpperCamelCase__ : Tuple , **UpperCamelCase__ : Optional[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ['''sentencepiece'''] def __init__( self : Optional[int] , *UpperCamelCase__ : str , **UpperCamelCase__ : List[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Tuple = ['''sentencepiece'''] def __init__( self : Optional[int] , *UpperCamelCase__ : int , **UpperCamelCase__ : Union[str, Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : str = ['''sentencepiece'''] def __init__( self : List[Any] , *UpperCamelCase__ : Optional[int] , **UpperCamelCase__ : List[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Dict = ['''sentencepiece'''] def __init__( self : List[str] , *UpperCamelCase__ : Union[str, Any] , **UpperCamelCase__ : Any ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : Dict , **UpperCamelCase__ : str ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[Any] = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : List[str] , **UpperCamelCase__ : Tuple ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = ['''sentencepiece'''] def __init__( self : int , *UpperCamelCase__ : str , **UpperCamelCase__ : Any ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[Any] = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : Tuple , **UpperCamelCase__ : Tuple ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ['''sentencepiece'''] def __init__( self : int , *UpperCamelCase__ : Dict , **UpperCamelCase__ : str ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = ['''sentencepiece'''] def __init__( self : str , *UpperCamelCase__ : Optional[int] , **UpperCamelCase__ : List[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = ['''sentencepiece'''] def __init__( self : List[Any] , *UpperCamelCase__ : Any , **UpperCamelCase__ : List[str] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : Optional[int] , **UpperCamelCase__ : Any ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = ['''sentencepiece'''] def __init__( self : Any , *UpperCamelCase__ : List[Any] , **UpperCamelCase__ : List[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[Any] = ['''sentencepiece'''] def __init__( self : List[Any] , *UpperCamelCase__ : Optional[int] , **UpperCamelCase__ : Optional[int] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[str] = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : List[str] , **UpperCamelCase__ : int ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = ['''sentencepiece'''] def __init__( self : Tuple , *UpperCamelCase__ : Optional[int] , **UpperCamelCase__ : Tuple ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = ['''sentencepiece'''] def __init__( self : List[str] , *UpperCamelCase__ : Any , **UpperCamelCase__ : List[str] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[Any] = ['''sentencepiece'''] def __init__( self : Optional[Any] , *UpperCamelCase__ : Union[str, Any] , **UpperCamelCase__ : Dict ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : List[Any] = ['''sentencepiece'''] def __init__( self : List[str] , *UpperCamelCase__ : Tuple , **UpperCamelCase__ : List[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[Any] = ['''sentencepiece'''] def __init__( self : Optional[int] , *UpperCamelCase__ : str , **UpperCamelCase__ : Optional[int] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ['''sentencepiece'''] def __init__( self : Tuple , *UpperCamelCase__ : Any , **UpperCamelCase__ : List[str] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = ['''sentencepiece'''] def __init__( self : int , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : Union[str, Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = ['''sentencepiece'''] def __init__( self : Any , *UpperCamelCase__ : Any , **UpperCamelCase__ : Optional[Any] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = ['''sentencepiece'''] def __init__( self : Union[str, Any] , *UpperCamelCase__ : Tuple , **UpperCamelCase__ : Tuple ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = ['''sentencepiece'''] def __init__( self : str , *UpperCamelCase__ : Tuple , **UpperCamelCase__ : Tuple ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : int = ['''sentencepiece'''] def __init__( self : str , *UpperCamelCase__ : Union[str, Any] , **UpperCamelCase__ : Optional[int] ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : str = ['''sentencepiece'''] def __init__( self : Union[str, Any] , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : int ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = ['''sentencepiece'''] def __init__( self : Optional[int] , *UpperCamelCase__ : Optional[Any] , **UpperCamelCase__ : str ): requires_backends(self , ['sentencepiece'] ) class _UpperCAmelCase ( metaclass=__lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[Any] = ['''sentencepiece'''] def __init__( self : Tuple , *UpperCamelCase__ : int , **UpperCamelCase__ : int ): requires_backends(self , ['sentencepiece'] )
699
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> Optional[int]: if b == 0: return 1 if (b % 2) == 0: return actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) else: return a * actual_power(lowerCAmelCase, int(b / 2 ) ) * actual_power(lowerCAmelCase, int(b / 2 ) ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : int ) -> float: if b < 0: return 1 / actual_power(lowerCAmelCase, lowerCAmelCase ) return actual_power(lowerCAmelCase, lowerCAmelCase ) if __name__ == "__main__": print(power(-2, -3))
699
1
import inspect import unittest from transformers import ViTHybridConfig from transformers.testing_utils import require_accelerate, require_torch, 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, _config_zero_init, floats_tensor, ids_tensor from ...test_pipeline_mixin import PipelineTesterMixin if is_torch_available(): import torch from torch import nn from transformers import ViTHybridForImageClassification, ViTHybridImageProcessor, ViTHybridModel from transformers.models.vit_hybrid.modeling_vit_hybrid import VIT_HYBRID_PRETRAINED_MODEL_ARCHIVE_LIST if is_vision_available(): from PIL import Image class _UpperCAmelCase : '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : List[str] , UpperCamelCase__ : Any=13 , UpperCamelCase__ : List[str]=64 , UpperCamelCase__ : Any=2 , UpperCamelCase__ : Tuple=3 , UpperCamelCase__ : List[Any]=True , UpperCamelCase__ : Dict=True , UpperCamelCase__ : List[Any]=32 , UpperCamelCase__ : int=5 , UpperCamelCase__ : int=4 , UpperCamelCase__ : Optional[Any]=37 , UpperCamelCase__ : Tuple="gelu" , UpperCamelCase__ : int=0.1 , UpperCamelCase__ : Tuple=0.1 , UpperCamelCase__ : Optional[Any]=10 , UpperCamelCase__ : Any=0.02 , UpperCamelCase__ : Union[str, Any]=[1, 16, 4, 4] , UpperCamelCase__ : int=None , ): A = parent A = batch_size A = image_size A = patch_size A = num_channels A = is_training A = use_labels A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = type_sequence_label_size A = initializer_range A = scope A = backbone_featmap_shape # in ViT hybrid, the seq length equals the number of patches + 1 (we add 1 for the [CLS] token) # the number of patches is based on the feature map of the backbone, which by default uses an output stride # of 32, which means that the feature map has a spatial resolution of 1/32 of the input image size A = (self.image_size // 32) ** 2 A = num_patches + 1 def UpperCamelCase ( self : List[Any] ): A = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] ) A = None if self.use_labels: A = ids_tensor([self.batch_size] , self.type_sequence_label_size ) A = self.get_config() return config, pixel_values, labels def UpperCamelCase ( self : Tuple ): A = { 'global_padding': 'same', 'layer_type': 'bottleneck', 'depths': [3, 4, 9], 'out_features': ['stage1', 'stage2', 'stage3'], 'embedding_dynamic_padding': True, 'hidden_sizes': [4, 8, 16, 32], 'num_groups': 2, } return ViTHybridConfig( image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , is_decoder=UpperCamelCase__ , initializer_range=self.initializer_range , backbone_featmap_shape=self.backbone_featmap_shape , backbone_config=UpperCamelCase__ , ) def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[Any] ): A = ViTHybridModel(config=UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ ) self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) ) def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : str ): A = self.type_sequence_label_size A = ViTHybridForImageClassification(UpperCamelCase__ ) model.to(UpperCamelCase__ ) model.eval() A = model(UpperCamelCase__ , labels=UpperCamelCase__ ) self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) ) def UpperCamelCase ( self : List[str] ): A = self.prepare_config_and_inputs() A , A , A = config_and_inputs A = {'pixel_values': pixel_values} return config, inputs_dict @require_torch class _UpperCAmelCase ( __lowercase , __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = (ViTHybridModel, ViTHybridForImageClassification) if is_torch_available() else () SCREAMING_SNAKE_CASE : List[str] = ( {'''feature-extraction''': ViTHybridModel, '''image-classification''': ViTHybridForImageClassification} if is_torch_available() else {} ) SCREAMING_SNAKE_CASE : Dict = False SCREAMING_SNAKE_CASE : List[Any] = False SCREAMING_SNAKE_CASE : List[Any] = False def UpperCamelCase ( self : Optional[Any] ): A = ViTHybridModelTester(self ) A = ConfigTester(self , config_class=UpperCamelCase__ , has_text_modality=UpperCamelCase__ , hidden_size=37 ) def UpperCamelCase ( self : Dict ): self.config_tester.run_common_tests() @unittest.skip(reason='ViT does not use inputs_embeds' ) def UpperCamelCase ( self : Any ): pass def UpperCamelCase ( self : List[str] ): A , A = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: A = model_class(UpperCamelCase__ ) self.assertIsInstance(model.get_input_embeddings() , (nn.Module) ) A = model.get_output_embeddings() self.assertTrue(x is None or isinstance(UpperCamelCase__ , nn.Linear ) ) def UpperCamelCase ( self : Optional[int] ): A , A = self.model_tester.prepare_config_and_inputs_for_common() for model_class in self.all_model_classes: A = model_class(UpperCamelCase__ ) 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] , UpperCamelCase__ ) def UpperCamelCase ( self : str ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_model(*UpperCamelCase__ ) def UpperCamelCase ( self : Union[str, Any] ): A = self.model_tester.prepare_config_and_inputs() self.model_tester.create_and_check_for_image_classification(*UpperCamelCase__ ) def UpperCamelCase ( self : str ): A , A = self.model_tester.prepare_config_and_inputs_for_common() A = _config_zero_init(UpperCamelCase__ ) for model_class in self.all_model_classes: A = model_class(config=UpperCamelCase__ ) # Skip the check for the backbone for name, module in model.named_modules(): if module.__class__.__name__ == "ViTHybridPatchEmbeddings": A = [f'''{name}.{key}''' for key in module.state_dict().keys()] break for name, param in model.named_parameters(): if param.requires_grad: if name in backbone_params: continue self.assertIn( ((param.data.mean() * 1e9).round() / 1e9).item() , [0.0, 1.0] , msg=f'''Parameter {name} of model {model_class} seems not properly initialized''' , ) @slow def UpperCamelCase ( self : Optional[int] ): for model_name in VIT_HYBRID_PRETRAINED_MODEL_ARCHIVE_LIST[:1]: A = ViTHybridModel.from_pretrained(UpperCamelCase__ ) self.assertIsNotNone(UpperCamelCase__ ) def __UpperCamelCase () -> Optional[Any]: A = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' ) return image @require_torch @require_vision class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' @cached_property def UpperCamelCase ( self : Optional[int] ): return ( ViTHybridImageProcessor.from_pretrained(VIT_HYBRID_PRETRAINED_MODEL_ARCHIVE_LIST[0] ) if is_vision_available() else None ) @slow def UpperCamelCase ( self : Any ): A = ViTHybridForImageClassification.from_pretrained(VIT_HYBRID_PRETRAINED_MODEL_ARCHIVE_LIST[0] ).to( UpperCamelCase__ ) A = self.default_image_processor A = prepare_img() A = image_processor(images=UpperCamelCase__ , return_tensors='pt' ).to(UpperCamelCase__ ) # forward pass with torch.no_grad(): A = model(**UpperCamelCase__ ) # verify the logits A = torch.Size((1, 1000) ) self.assertEqual(outputs.logits.shape , UpperCamelCase__ ) A = torch.tensor([-1.9_090, -0.4_993, -0.2_389] ).to(UpperCamelCase__ ) self.assertTrue(torch.allclose(outputs.logits[0, :3] , UpperCamelCase__ , atol=1e-4 ) ) @slow @require_accelerate def UpperCamelCase ( self : int ): A = ViTHybridImageProcessor.from_pretrained('google/vit-hybrid-base-bit-384' ) A = ViTHybridForImageClassification.from_pretrained('google/vit-hybrid-base-bit-384' , device_map='auto' ) A = prepare_img() A = image_processor(images=UpperCamelCase__ , return_tensors='pt' ) A = model(**UpperCamelCase__ ) A = outputs.logits # model predicts one of the 1000 ImageNet classes A = logits.argmax(-1 ).item() self.assertTrue(model.config.idalabel[predicted_class_idx] , 'tabby, tabby cat' )
699
def __UpperCamelCase (lowerCAmelCase : list[int] ) -> int: if not numbers: return 0 if not isinstance(lowerCAmelCase, (list, tuple) ) or not all( isinstance(lowerCAmelCase, lowerCAmelCase ) for number in numbers ): raise ValueError('numbers must be an iterable of integers' ) A = A = A = numbers[0] for i in range(1, len(lowerCAmelCase ) ): # update the maximum and minimum subarray products A = numbers[i] if number < 0: A , A = min_till_now, max_till_now A = max(lowerCAmelCase, max_till_now * number ) A = min(lowerCAmelCase, min_till_now * number ) # update the maximum product found till now A = max(lowerCAmelCase, lowerCAmelCase ) return max_prod
699
1
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> list[int]: A = [True] * max_number for i in range(2, isqrt(max_number - 1 ) + 1 ): if is_prime[i]: for j in range(i**2, lowerCAmelCase, lowerCAmelCase ): A = False return [i for i in range(2, lowerCAmelCase ) if is_prime[i]] def __UpperCamelCase (lowerCAmelCase : int = 10**8 ) -> int: A = calculate_prime_numbers(max_number // 2 ) A = 0 A = 0 A = len(lowerCAmelCase ) - 1 while left <= right: while prime_numbers[left] * prime_numbers[right] >= max_number: right -= 1 semiprimes_count += right - left + 1 left += 1 return semiprimes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
from .glue import glue_convert_examples_to_features, glue_output_modes, glue_processors, glue_tasks_num_labels from .squad import SquadExample, SquadFeatures, SquadVaProcessor, SquadVaProcessor, squad_convert_examples_to_features from .utils import DataProcessor, InputExample, InputFeatures, SingleSentenceClassificationProcessor from .xnli import xnli_output_modes, xnli_processors, xnli_tasks_num_labels
699
1
import sys import webbrowser import requests from bsa import BeautifulSoup from fake_useragent import UserAgent if __name__ == "__main__": print("Googling.....") _UpperCAmelCase = "https://www.google.com/search?q=" + " ".join(sys.argv[1:]) _UpperCAmelCase = requests.get(url, headers={"UserAgent": UserAgent().random}) # res.raise_for_status() with open("project1a.html", "wb") as out_file: # only for knowing the class for data in res.iter_content(10_000): out_file.write(data) _UpperCAmelCase = BeautifulSoup(res.text, "html.parser") _UpperCAmelCase = list(soup.select(".eZt8xd"))[:5] print(len(links)) for link in links: if link.text == "Maps": webbrowser.open(link.get("href")) else: webbrowser.open(F'''https://google.com{link.get("href")}''')
699
from ...configuration_utils import PretrainedConfig from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "microsoft/biogpt": "https://huggingface.co/microsoft/biogpt/resolve/main/config.json", # See all BioGPT models at https://huggingface.co/models?filter=biogpt } class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Optional[int] = '''biogpt''' def __init__( self : Optional[Any] , UpperCamelCase__ : str=42384 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : Dict=24 , UpperCamelCase__ : Any=16 , UpperCamelCase__ : str=4096 , UpperCamelCase__ : Tuple="gelu" , UpperCamelCase__ : List[str]=0.1 , UpperCamelCase__ : Any=0.1 , UpperCamelCase__ : Tuple=1024 , UpperCamelCase__ : List[Any]=0.02 , UpperCamelCase__ : Dict=1e-1_2 , UpperCamelCase__ : Any=True , UpperCamelCase__ : List[str]=True , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Any=1 , UpperCamelCase__ : List[str]=0 , UpperCamelCase__ : Optional[Any]=2 , **UpperCamelCase__ : List[Any] , ): A = vocab_size A = max_position_embeddings A = hidden_size A = num_hidden_layers A = num_attention_heads A = intermediate_size A = hidden_act A = hidden_dropout_prob A = attention_probs_dropout_prob A = initializer_range A = layer_norm_eps A = scale_embedding A = use_cache A = layerdrop A = activation_dropout super().__init__(pad_token_id=UpperCamelCase__ , bos_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , **UpperCamelCase__ )
699
1
def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : list[int], lowerCAmelCase : int ) -> int: def count_of_possible_combinations(lowerCAmelCase : int ) -> int: if target < 0: return 0 if target == 0: return 1 return sum(count_of_possible_combinations(target - item ) for item in array ) return count_of_possible_combinations(lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : list[int], lowerCAmelCase : int ) -> int: def count_of_possible_combinations_with_dp_array( lowerCAmelCase : int, lowerCAmelCase : list[int] ) -> int: if target < 0: return 0 if target == 0: return 1 if dp_array[target] != -1: return dp_array[target] A = sum( count_of_possible_combinations_with_dp_array(target - item, lowerCAmelCase ) for item in array ) A = answer return answer A = [-1] * (target + 1) return count_of_possible_combinations_with_dp_array(lowerCAmelCase, lowerCAmelCase ) def __UpperCamelCase (lowerCAmelCase : int, lowerCAmelCase : list[int], lowerCAmelCase : int ) -> int: A = [0] * (target + 1) A = 1 for i in range(1, target + 1 ): for j in range(lowerCAmelCase ): if i - array[j] >= 0: dp_array[i] += dp_array[i - array[j]] return dp_array[target] if __name__ == "__main__": import doctest doctest.testmod() _UpperCAmelCase = 3 _UpperCAmelCase = 5 _UpperCAmelCase = [1, 2, 5] print(combination_sum_iv(n, array, target))
699
import sys def __UpperCamelCase (lowerCAmelCase : Dict ) -> Dict: A = len(lowerCAmelCase ) A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] A = [[0 for x in range(lowerCAmelCase )] for x in range(lowerCAmelCase )] for chain_length in range(2, lowerCAmelCase ): for a in range(1, n - chain_length + 1 ): A = a + chain_length - 1 A = sys.maxsize for c in range(lowerCAmelCase, lowerCAmelCase ): A = ( matrix[a][c] + matrix[c + 1][b] + array[a - 1] * array[c] * array[b] ) if cost < matrix[a][b]: A = cost A = c return matrix, sol def __UpperCamelCase (lowerCAmelCase : Optional[Any], lowerCAmelCase : Union[str, Any], lowerCAmelCase : Union[str, Any] ) -> List[str]: if i == j: print('A' + str(lowerCAmelCase ), end=' ' ) else: print('(', end=' ' ) print_optiomal_solution(lowerCAmelCase, lowerCAmelCase, optimal_solution[i][j] ) print_optiomal_solution(lowerCAmelCase, optimal_solution[i][j] + 1, lowerCAmelCase ) print(')', end=' ' ) def __UpperCamelCase () -> List[str]: A = [30, 35, 15, 5, 10, 20, 25] A = len(lowerCAmelCase ) # Size of matrix created from above array will be # 30*35 35*15 15*5 5*10 10*20 20*25 A , A = matrix_chain_order(lowerCAmelCase ) print('No. of Operation required: ' + str(matrix[1][n - 1] ) ) print_optiomal_solution(lowerCAmelCase, 1, n - 1 ) if __name__ == "__main__": main()
699
1
import gc import random import unittest import numpy as np import torch from PIL import Image from diffusers import ( DDIMScheduler, KandinskyVaaInpaintPipeline, KandinskyVaaPriorPipeline, UNetaDConditionModel, VQModel, ) from diffusers.utils import floats_tensor, load_image, load_numpy, slow, torch_device from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu from ..test_pipelines_common import PipelineTesterMixin, assert_mean_pixel_difference enable_full_determinism() class _UpperCAmelCase ( __lowercase , unittest.TestCase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Any = KandinskyVaaInpaintPipeline SCREAMING_SNAKE_CASE : str = ['''image_embeds''', '''negative_image_embeds''', '''image''', '''mask_image'''] SCREAMING_SNAKE_CASE : Union[str, Any] = [ '''image_embeds''', '''negative_image_embeds''', '''image''', '''mask_image''', ] SCREAMING_SNAKE_CASE : str = [ '''generator''', '''height''', '''width''', '''latents''', '''guidance_scale''', '''num_inference_steps''', '''return_dict''', '''guidance_scale''', '''num_images_per_prompt''', '''output_type''', '''return_dict''', ] SCREAMING_SNAKE_CASE : str = False @property def UpperCamelCase ( self : Union[str, Any] ): return 32 @property def UpperCamelCase ( self : List[Any] ): return 32 @property def UpperCamelCase ( self : str ): return self.time_input_dim @property def UpperCamelCase ( self : Optional[Any] ): return self.time_input_dim * 4 @property def UpperCamelCase ( self : str ): return 100 @property def UpperCamelCase ( self : Any ): torch.manual_seed(0 ) A = { 'in_channels': 9, # Out channels is double in channels because predicts mean and variance 'out_channels': 8, 'addition_embed_type': 'image', 'down_block_types': ('ResnetDownsampleBlock2D', 'SimpleCrossAttnDownBlock2D'), 'up_block_types': ('SimpleCrossAttnUpBlock2D', 'ResnetUpsampleBlock2D'), 'mid_block_type': 'UNetMidBlock2DSimpleCrossAttn', 'block_out_channels': (self.block_out_channels_a, self.block_out_channels_a * 2), 'layers_per_block': 1, 'encoder_hid_dim': self.text_embedder_hidden_size, 'encoder_hid_dim_type': 'image_proj', 'cross_attention_dim': self.cross_attention_dim, 'attention_head_dim': 4, 'resnet_time_scale_shift': 'scale_shift', 'class_embed_type': None, } A = UNetaDConditionModel(**UpperCamelCase__ ) return model @property def UpperCamelCase ( self : Optional[Any] ): return { "block_out_channels": [32, 64], "down_block_types": ["DownEncoderBlock2D", "AttnDownEncoderBlock2D"], "in_channels": 3, "latent_channels": 4, "layers_per_block": 1, "norm_num_groups": 8, "norm_type": "spatial", "num_vq_embeddings": 12, "out_channels": 3, "up_block_types": [ "AttnUpDecoderBlock2D", "UpDecoderBlock2D", ], "vq_embed_dim": 4, } @property def UpperCamelCase ( self : Union[str, Any] ): torch.manual_seed(0 ) A = VQModel(**self.dummy_movq_kwargs ) return model def UpperCamelCase ( self : Optional[Any] ): A = self.dummy_unet A = self.dummy_movq A = DDIMScheduler( num_train_timesteps=1000 , beta_schedule='linear' , beta_start=0.00_085 , beta_end=0.012 , clip_sample=UpperCamelCase__ , set_alpha_to_one=UpperCamelCase__ , steps_offset=1 , prediction_type='epsilon' , thresholding=UpperCamelCase__ , ) A = { 'unet': unet, 'scheduler': scheduler, 'movq': movq, } return components def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : Tuple=0 ): A = floats_tensor((1, self.text_embedder_hidden_size) , rng=random.Random(UpperCamelCase__ ) ).to(UpperCamelCase__ ) A = floats_tensor((1, self.text_embedder_hidden_size) , rng=random.Random(seed + 1 ) ).to( UpperCamelCase__ ) # create init_image A = floats_tensor((1, 3, 64, 64) , rng=random.Random(UpperCamelCase__ ) ).to(UpperCamelCase__ ) A = image.cpu().permute(0 , 2 , 3 , 1 )[0] A = Image.fromarray(np.uinta(UpperCamelCase__ ) ).convert('RGB' ).resize((256, 256) ) # create mask A = np.ones((64, 64) , dtype=np.floataa ) A = 0 if str(UpperCamelCase__ ).startswith('mps' ): A = torch.manual_seed(UpperCamelCase__ ) else: A = torch.Generator(device=UpperCamelCase__ ).manual_seed(UpperCamelCase__ ) A = { 'image': init_image, 'mask_image': mask, 'image_embeds': image_embeds, 'negative_image_embeds': negative_image_embeds, 'generator': generator, 'height': 64, 'width': 64, 'num_inference_steps': 2, 'guidance_scale': 4.0, 'output_type': 'np', } return inputs def UpperCamelCase ( self : List[Any] ): A = 'cpu' A = self.get_dummy_components() A = self.pipeline_class(**UpperCamelCase__ ) A = pipe.to(UpperCamelCase__ ) pipe.set_progress_bar_config(disable=UpperCamelCase__ ) A = pipe(**self.get_dummy_inputs(UpperCamelCase__ ) ) A = output.images A = pipe( **self.get_dummy_inputs(UpperCamelCase__ ) , return_dict=UpperCamelCase__ , )[0] A = image[0, -3:, -3:, -1] A = image_from_tuple[0, -3:, -3:, -1] print(f'''image.shape {image.shape}''' ) assert image.shape == (1, 64, 64, 3) A = np.array( [0.50_775_903, 0.49_527_195, 0.48_824_543, 0.50_192_237, 0.48_644_906, 0.49_373_814, 0.4_780_598, 0.47_234_827, 0.48_327_848] ) assert ( np.abs(image_slice.flatten() - expected_slice ).max() < 1e-2 ), f''' expected_slice {expected_slice}, but got {image_slice.flatten()}''' assert ( np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1e-2 ), f''' expected_slice {expected_slice}, but got {image_from_tuple_slice.flatten()}''' def UpperCamelCase ( self : int ): super().test_inference_batch_single_identical(expected_max_diff=3e-3 ) @slow @require_torch_gpu class _UpperCAmelCase ( unittest.TestCase ): '''simple docstring''' def UpperCamelCase ( self : int ): # clean up the VRAM after each test super().tearDown() gc.collect() torch.cuda.empty_cache() def UpperCamelCase ( self : int ): A = load_numpy( 'https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main' '/kandinskyv22/kandinskyv22_inpaint_cat_with_hat_fp16.npy' ) A = load_image( 'https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main' '/kandinsky/cat.png' ) A = np.ones((768, 768) , dtype=np.floataa ) A = 0 A = 'a hat' A = KandinskyVaaPriorPipeline.from_pretrained( 'kandinsky-community/kandinsky-2-2-prior' , torch_dtype=torch.floataa ) pipe_prior.to(UpperCamelCase__ ) A = KandinskyVaaInpaintPipeline.from_pretrained( 'kandinsky-community/kandinsky-2-2-decoder-inpaint' , torch_dtype=torch.floataa ) A = pipeline.to(UpperCamelCase__ ) pipeline.set_progress_bar_config(disable=UpperCamelCase__ ) A = torch.Generator(device='cpu' ).manual_seed(0 ) A , A = pipe_prior( UpperCamelCase__ , generator=UpperCamelCase__ , num_inference_steps=5 , negative_prompt='' , ).to_tuple() A = pipeline( image=UpperCamelCase__ , mask_image=UpperCamelCase__ , image_embeds=UpperCamelCase__ , negative_image_embeds=UpperCamelCase__ , generator=UpperCamelCase__ , num_inference_steps=100 , height=768 , width=768 , output_type='np' , ) A = output.images[0] assert image.shape == (768, 768, 3) assert_mean_pixel_difference(UpperCamelCase__ , UpperCamelCase__ )
699
from math import isqrt def __UpperCamelCase (lowerCAmelCase : int ) -> bool: return all(number % divisor != 0 for divisor in range(2, isqrt(lowerCAmelCase ) + 1 ) ) def __UpperCamelCase (lowerCAmelCase : int = 10**6 ) -> int: A = 0 A = 1 A = 7 while prime_candidate < max_prime: primes_count += is_prime(lowerCAmelCase ) cube_index += 1 prime_candidate += 6 * cube_index return primes_count if __name__ == "__main__": print(F'''{solution() = }''')
699
1
def __UpperCamelCase (lowerCAmelCase : int ) -> list: # bit count represents no. of bits in the gray code if bit_count < 0: raise ValueError('The given input must be positive' ) # get the generated string sequence A = gray_code_sequence_string(lowerCAmelCase ) # # convert them to integers for i in range(len(lowerCAmelCase ) ): A = int(sequence[i], 2 ) return sequence def __UpperCamelCase (lowerCAmelCase : int ) -> list: # The approach is a recursive one # Base case achieved when either n = 0 or n=1 if bit_count == 0: return ["0"] if bit_count == 1: return ["0", "1"] A = 1 << bit_count # defines the length of the sequence # 1<< n is equivalent to 2^n # recursive answer will generate answer for n-1 bits A = gray_code_sequence_string(bit_count - 1 ) A = [] # append 0 to first half of the smaller sequence generated for i in range(seq_len // 2 ): A = '0' + smaller_sequence[i] sequence.append(lowerCAmelCase ) # append 1 to second half ... start from the end of the list for i in reversed(range(seq_len // 2 ) ): A = '1' + smaller_sequence[i] sequence.append(lowerCAmelCase ) return sequence if __name__ == "__main__": import doctest doctest.testmod()
699
import warnings from ...utils import logging from .image_processing_imagegpt import ImageGPTImageProcessor _UpperCAmelCase = logging.get_logger(__name__) class _UpperCAmelCase ( __lowercase ): '''simple docstring''' def __init__( self : List[str] , *UpperCamelCase__ : List[Any] , **UpperCamelCase__ : Tuple ): warnings.warn( 'The class ImageGPTFeatureExtractor is deprecated and will be removed in version 5 of Transformers.' ' Please use ImageGPTImageProcessor instead.' , UpperCamelCase__ , ) super().__init__(*UpperCamelCase__ , **UpperCamelCase__ )
699
1
def __UpperCamelCase () -> int: return [ a * b * (1_000 - a - b) for a in range(1, 999 ) for b in range(lowerCAmelCase, 999 ) if (a * a + b * b == (1_000 - a - b) ** 2) ][0] if __name__ == "__main__": print(F'''{solution() = }''')
699
from typing import Any, Dict, Optional import torch import torch.nn.functional as F from torch import nn from ..utils import maybe_allow_in_graph from .activations import get_activation from .attention_processor import Attention from .embeddings import CombinedTimestepLabelEmbeddings @maybe_allow_in_graph class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[int] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int]=0.0 , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = False , UpperCamelCase__ : bool = True , UpperCamelCase__ : str = "layer_norm" , UpperCamelCase__ : bool = False , ): super().__init__() A = only_cross_attention A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm_zero' A = (num_embeds_ada_norm is not None) and norm_type == 'ada_norm' if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None: raise ValueError( f'''`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to''' f''' define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}.''' ) # Define 3 blocks. Each block has its own normalization layer. # 1. Self-Attn if self.use_ada_layer_norm: A = AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A = AdaLayerNormZero(UpperCamelCase__ , UpperCamelCase__ ) else: A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = Attention( query_dim=UpperCamelCase__ , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if only_cross_attention else None , upcast_attention=UpperCamelCase__ , ) # 2. Cross-Attn if cross_attention_dim is not None or double_self_attention: # We currently only use AdaLayerNormZero for self attention where there will only be one attention block. # I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during # the second cross attention block. A = ( AdaLayerNorm(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) ) A = Attention( query_dim=UpperCamelCase__ , cross_attention_dim=cross_attention_dim if not double_self_attention else None , heads=UpperCamelCase__ , dim_head=UpperCamelCase__ , dropout=UpperCamelCase__ , bias=UpperCamelCase__ , upcast_attention=UpperCamelCase__ , ) # is self-attn if encoder_hidden_states is none else: A = None A = None # 3. Feed-forward A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) A = FeedForward(UpperCamelCase__ , dropout=UpperCamelCase__ , activation_fn=UpperCamelCase__ , final_dropout=UpperCamelCase__ ) # let chunk size default to None A = None A = 0 def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : int ): # Sets chunk feed-forward A = chunk_size A = dim def UpperCamelCase ( self : Dict , UpperCamelCase__ : torch.FloatTensor , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.FloatTensor] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , UpperCamelCase__ : Dict[str, Any] = None , UpperCamelCase__ : Optional[torch.LongTensor] = None , ): # Notice that normalization is always applied before the real computation in the following blocks. # 1. Self-Attention if self.use_ada_layer_norm: A = self.norma(UpperCamelCase__ , UpperCamelCase__ ) elif self.use_ada_layer_norm_zero: A , A , A , A , A = self.norma( UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=hidden_states.dtype ) else: A = self.norma(UpperCamelCase__ ) A = cross_attention_kwargs if cross_attention_kwargs is not None else {} A = self.attna( UpperCamelCase__ , encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) if self.use_ada_layer_norm_zero: A = gate_msa.unsqueeze(1 ) * attn_output A = attn_output + hidden_states # 2. Cross-Attention if self.attna is not None: A = ( self.norma(UpperCamelCase__ , UpperCamelCase__ ) if self.use_ada_layer_norm else self.norma(UpperCamelCase__ ) ) A = self.attna( UpperCamelCase__ , encoder_hidden_states=UpperCamelCase__ , attention_mask=UpperCamelCase__ , **UpperCamelCase__ , ) A = attn_output + hidden_states # 3. Feed-forward A = self.norma(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None] if self._chunk_size is not None: # "feed_forward_chunk_size" can be used to save memory if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0: raise ValueError( f'''`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`.''' ) A = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size A = torch.cat( [self.ff(UpperCamelCase__ ) for hid_slice in norm_hidden_states.chunk(UpperCamelCase__ , dim=self._chunk_dim )] , dim=self._chunk_dim , ) else: A = self.ff(UpperCamelCase__ ) if self.use_ada_layer_norm_zero: A = gate_mlp.unsqueeze(1 ) * ff_output A = ff_output + hidden_states return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[int] = None , UpperCamelCase__ : int = 4 , UpperCamelCase__ : float = 0.0 , UpperCamelCase__ : str = "geglu" , UpperCamelCase__ : bool = False , ): super().__init__() A = int(dim * mult ) A = dim_out if dim_out is not None else dim if activation_fn == "gelu": A = GELU(UpperCamelCase__ , UpperCamelCase__ ) if activation_fn == "gelu-approximate": A = GELU(UpperCamelCase__ , UpperCamelCase__ , approximate='tanh' ) elif activation_fn == "geglu": A = GEGLU(UpperCamelCase__ , UpperCamelCase__ ) elif activation_fn == "geglu-approximate": A = ApproximateGELU(UpperCamelCase__ , UpperCamelCase__ ) A = nn.ModuleList([] ) # project in self.net.append(UpperCamelCase__ ) # project dropout self.net.append(nn.Dropout(UpperCamelCase__ ) ) # project out self.net.append(nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) ) # FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout if final_dropout: self.net.append(nn.Dropout(UpperCamelCase__ ) ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : int ): for module in self.net: A = module(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : str = "none" ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) A = approximate def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ , approximate=self.approximate ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) , approximate=self.approximate ).to(dtype=gate.dtype ) def UpperCamelCase ( self : Any , UpperCamelCase__ : int ): A = self.proj(UpperCamelCase__ ) A = self.gelu(UpperCamelCase__ ) return hidden_states class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , dim_out * 2 ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Tuple ): if gate.device.type != "mps": return F.gelu(UpperCamelCase__ ) # mps: gelu is not implemented for float16 return F.gelu(gate.to(dtype=torch.floataa ) ).to(dtype=gate.dtype ) def UpperCamelCase ( self : str , UpperCamelCase__ : str ): A , A = self.proj(UpperCamelCase__ ).chunk(2 , dim=-1 ) return hidden_states * self.gelu(UpperCamelCase__ ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : int , UpperCamelCase__ : int , UpperCamelCase__ : int ): super().__init__() A = nn.Linear(UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Any , UpperCamelCase__ : Optional[int] ): A = self.proj(UpperCamelCase__ ) return x * torch.sigmoid(1.702 * x ) class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Tuple ): super().__init__() A = nn.Embedding(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , embedding_dim * 2 ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ ) def UpperCamelCase ( self : List[str] , UpperCamelCase__ : Dict , UpperCamelCase__ : Optional[Any] ): A = self.linear(self.silu(self.emb(UpperCamelCase__ ) ) ) A , A = torch.chunk(UpperCamelCase__ , 2 ) A = self.norm(UpperCamelCase__ ) * (1 + scale) + shift return x class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : str , UpperCamelCase__ : int , UpperCamelCase__ : List[str] ): super().__init__() A = CombinedTimestepLabelEmbeddings(UpperCamelCase__ , UpperCamelCase__ ) A = nn.SiLU() A = nn.Linear(UpperCamelCase__ , 6 * embedding_dim , bias=UpperCamelCase__ ) A = nn.LayerNorm(UpperCamelCase__ , elementwise_affine=UpperCamelCase__ , eps=1e-6 ) def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : str , UpperCamelCase__ : Tuple=None ): A = self.linear(self.silu(self.emb(UpperCamelCase__ , UpperCamelCase__ , hidden_dtype=UpperCamelCase__ ) ) ) A , A , A , A , A , A = emb.chunk(6 , dim=1 ) A = self.norm(UpperCamelCase__ ) * (1 + scale_msa[:, None]) + shift_msa[:, None] return x, gate_msa, shift_mlp, scale_mlp, gate_mlp class _UpperCAmelCase ( nn.Module ): '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : Optional[str] = None , UpperCamelCase__ : float = 1e-5 ): super().__init__() A = num_groups A = eps if act_fn is None: A = None else: A = get_activation(UpperCamelCase__ ) A = nn.Linear(UpperCamelCase__ , out_dim * 2 ) def UpperCamelCase ( self : Any , UpperCamelCase__ : str , UpperCamelCase__ : str ): if self.act: A = self.act(UpperCamelCase__ ) A = self.linear(UpperCamelCase__ ) A = emb[:, :, None, None] A , A = emb.chunk(2 , dim=1 ) A = F.group_norm(UpperCamelCase__ , self.num_groups , eps=self.eps ) A = x * (1 + scale) + shift return x
699
1
from collections.abc import Sequence from queue import Queue class _UpperCAmelCase : '''simple docstring''' def __init__( self : Optional[Any] , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Tuple , UpperCamelCase__ : int=None , UpperCamelCase__ : Union[str, Any]=None ): A = start A = end A = val A = (start + end) // 2 A = left A = right def __repr__( self : Union[str, Any] ): return f'''SegmentTreeNode(start={self.start}, end={self.end}, val={self.val})''' class _UpperCAmelCase : '''simple docstring''' def __init__( self : Any , UpperCamelCase__ : Sequence , UpperCamelCase__ : Optional[Any] ): A = collection A = function if self.collection: A = self._build_tree(0 , len(UpperCamelCase__ ) - 1 ) def UpperCamelCase ( self : str , UpperCamelCase__ : Dict , UpperCamelCase__ : List[Any] ): self._update_tree(self.root , UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Dict , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Dict ): return self._query_range(self.root , UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Union[str, Any] , UpperCamelCase__ : Any , UpperCamelCase__ : Any ): if start == end: return SegmentTreeNode(UpperCamelCase__ , UpperCamelCase__ , self.collection[start] ) A = (start + end) // 2 A = self._build_tree(UpperCamelCase__ , UpperCamelCase__ ) A = self._build_tree(mid + 1 , UpperCamelCase__ ) return SegmentTreeNode(UpperCamelCase__ , UpperCamelCase__ , self.fn(left.val , right.val ) , UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : Optional[Any] , UpperCamelCase__ : Any , UpperCamelCase__ : int ): if node.start == i and node.end == i: A = val return if i <= node.mid: self._update_tree(node.left , UpperCamelCase__ , UpperCamelCase__ ) else: self._update_tree(node.right , UpperCamelCase__ , UpperCamelCase__ ) A = self.fn(node.left.val , node.right.val ) def UpperCamelCase ( self : str , UpperCamelCase__ : Union[str, Any] , UpperCamelCase__ : str , UpperCamelCase__ : Union[str, Any] ): if node.start == i and node.end == j: return node.val if i <= node.mid: if j <= node.mid: # range in left child tree return self._query_range(node.left , UpperCamelCase__ , UpperCamelCase__ ) else: # range in left child tree and right child tree return self.fn( self._query_range(node.left , UpperCamelCase__ , node.mid ) , self._query_range(node.right , node.mid + 1 , UpperCamelCase__ ) , ) else: # range in right child tree return self._query_range(node.right , UpperCamelCase__ , UpperCamelCase__ ) def UpperCamelCase ( self : str ): if self.root is not None: A = Queue() queue.put(self.root ) while not queue.empty(): A = queue.get() yield node if node.left is not None: queue.put(node.left ) if node.right is not None: queue.put(node.right ) if __name__ == "__main__": import operator for fn in [operator.add, max, min]: print("*" * 50) _UpperCAmelCase = SegmentTree([2, 1, 5, 3, 4], fn) for node in arr.traverse(): print(node) print() arr.update(1, 5) for node in arr.traverse(): print(node) print() print(arr.query_range(3, 4)) # 7 print(arr.query_range(2, 2)) # 5 print(arr.query_range(1, 3)) # 13 print()
699
import json import os from typing import Dict, List, Optional, Tuple from ...tokenization_utils import PreTrainedTokenizer from ...utils import logging _UpperCAmelCase = logging.get_logger(__name__) _UpperCAmelCase = { "vocab_file": "vocab.json", "tokenizer_config_file": "tokenizer_config.json", "merges_file": "merges.txt", } _UpperCAmelCase = { "vocab_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/vocab.json" ), }, "tokenizer_config_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/tokenizer_config.json" ), }, "merges_file": { "facebook/s2t-wav2vec2-large-en-de": ( "https://huggingface.co/facebook/s2t-wav2vec2-large-en-de/resolve/main/merges.txt" ), }, } _UpperCAmelCase = "</w>" _UpperCAmelCase = "@@ " def __UpperCamelCase (lowerCAmelCase : Optional[int] ) -> List[str]: A = set() A = word[0] for char in word[1:]: pairs.add((prev_char, char) ) A = char return pairs # Speech2Text2 has no max input length _UpperCAmelCase = {"facebook/s2t-wav2vec2-large-en-de": 1_024} class _UpperCAmelCase ( __lowercase ): '''simple docstring''' SCREAMING_SNAKE_CASE : Union[str, Any] = VOCAB_FILES_NAMES SCREAMING_SNAKE_CASE : List[str] = PRETRAINED_VOCAB_FILES_MAP SCREAMING_SNAKE_CASE : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES SCREAMING_SNAKE_CASE : Any = ['''input_ids''', '''attention_mask'''] def __init__( self : Optional[Any] , UpperCamelCase__ : Optional[int] , UpperCamelCase__ : Optional[int]="<s>" , UpperCamelCase__ : str="<pad>" , UpperCamelCase__ : int="</s>" , UpperCamelCase__ : Tuple="<unk>" , UpperCamelCase__ : List[str]=False , UpperCamelCase__ : List[str]=None , **UpperCamelCase__ : Optional[int] , ): super().__init__( unk_token=UpperCamelCase__ , bos_token=UpperCamelCase__ , eos_token=UpperCamelCase__ , pad_token=UpperCamelCase__ , do_lower_case=UpperCamelCase__ , **UpperCamelCase__ , ) A = do_lower_case with open(UpperCamelCase__ , encoding='utf-8' ) as vocab_handle: A = json.load(UpperCamelCase__ ) A = {v: k for k, v in self.encoder.items()} if merges_file is None: logger.info(f'''No merges files provided. {self.__class__.__name__} can only be used for decoding.''' ) A = None A = None else: with open(UpperCamelCase__ , encoding='utf-8' ) as merges_handle: A = merges_handle.read().split('\n' )[:-1] A = [tuple(merge.split()[:2] ) for merge in merges] A = dict(zip(UpperCamelCase__ , range(len(UpperCamelCase__ ) ) ) ) A = {} @property def UpperCamelCase ( self : Union[str, Any] ): return len(self.decoder ) def UpperCamelCase ( self : Optional[Any] ): return dict(self.encoder , **self.added_tokens_encoder ) def UpperCamelCase ( self : Tuple , UpperCamelCase__ : Optional[int] ): A = tuple(token[:-1] ) + (token[-1] + BPE_TOKEN_MERGES,) if token in self.cache: return self.cache[token] A = get_pairs(UpperCamelCase__ ) if not pairs: return token while True: A = min(UpperCamelCase__ , key=lambda UpperCamelCase__ : self.bpe_ranks.get(UpperCamelCase__ , float('inf' ) ) ) if bigram not in self.bpe_ranks: break A , A = bigram A = [] A = 0 while i < len(UpperCamelCase__ ): try: A = word.index(UpperCamelCase__ , UpperCamelCase__ ) except ValueError: new_word.extend(word[i:] ) break else: new_word.extend(word[i:j] ) A = j if word[i] == first and i < len(UpperCamelCase__ ) - 1 and word[i + 1] == second: new_word.append(first + second ) i += 2 else: new_word.append(word[i] ) i += 1 A = tuple(UpperCamelCase__ ) A = new_word if len(UpperCamelCase__ ) == 1: break else: A = get_pairs(UpperCamelCase__ ) A = ' '.join(UpperCamelCase__ ) if word == "\n " + BPE_TOKEN_MERGES: A = '\n' + BPE_TOKEN_MERGES if word.endswith(UpperCamelCase__ ): A = word.replace(UpperCamelCase__ , '' ) A = word.replace(' ' , UpperCamelCase__ ) A = word return word def UpperCamelCase ( self : Optional[int] , UpperCamelCase__ : Dict ): if self.bpe_ranks is None: raise ValueError( 'This tokenizer was instantiated without a `merges.txt` file, so' ' that it can only be used for decoding, not for encoding.' 'Make sure to provide `merges.txt` file at instantiation to enable ' 'encoding.' ) if self.do_lower_case: A = text.lower() A = text.split() A = [] for token in text: if token: split_tokens.extend(list(self.bpe(UpperCamelCase__ ).split(' ' ) ) ) return split_tokens def UpperCamelCase ( self : Optional[Any] , UpperCamelCase__ : str ): return self.encoder.get(UpperCamelCase__ , self.encoder.get(self.unk_token ) ) def UpperCamelCase ( self : str , UpperCamelCase__ : int ): A = self.decoder.get(UpperCamelCase__ , self.unk_token ) return result def UpperCamelCase ( self : Tuple , UpperCamelCase__ : List[str] ): A = ' '.join(UpperCamelCase__ ) # make sure @@ tokens are concatenated A = ''.join(string.split(UpperCamelCase__ ) ) return string def UpperCamelCase ( self : List[Any] , UpperCamelCase__ : str , UpperCamelCase__ : Optional[str] = None ): if not os.path.isdir(UpperCamelCase__ ): logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' ) return A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'] ) A = os.path.join( UpperCamelCase__ , (filename_prefix + '-' if filename_prefix else '') + VOCAB_FILES_NAMES['merges_file'] ) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as f: f.write(json.dumps(self.encoder , indent=2 , sort_keys=UpperCamelCase__ , ensure_ascii=UpperCamelCase__ ) + '\n' ) A = 0 if self.bpe_ranks is None: return (vocab_file,) with open(UpperCamelCase__ , 'w' , encoding='utf-8' ) as writer: for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda UpperCamelCase__ : kv[1] ): if index != token_index: logger.warning( f'''Saving vocabulary to {merges_file}: BPE merge indices are not consecutive.''' ' Please check that the tokenizer is not corrupted!' ) A = token_index writer.write(' '.join(UpperCamelCase__ ) + '\n' ) index += 1 return (vocab_file, merges_file)
699
1