code
stringlengths 87
55.2k
| code_codestyle
int64 0
349
| style_context
stringlengths 135
49.1k
| style_context_codestyle
int64 0
349
| label
int64 0
1
|
|---|---|---|---|---|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_flax_available, is_torch_available
_lowerCAmelCase = {'''configuration_speech_encoder_decoder''': ['''SpeechEncoderDecoderConfig''']}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = ['''SpeechEncoderDecoderModel''']
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = ['''FlaxSpeechEncoderDecoderModel''']
if TYPE_CHECKING:
from .configuration_speech_encoder_decoder import SpeechEncoderDecoderConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_speech_encoder_decoder import SpeechEncoderDecoderModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_speech_encoder_decoder import FlaxSpeechEncoderDecoderModel
else:
import sys
_lowerCAmelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 37
|
'''simple docstring'''
from collections import UserDict
from typing import List, Union
from ..utils import (
add_end_docstrings,
is_tf_available,
is_torch_available,
is_vision_available,
logging,
requires_backends,
)
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
from ..tf_utils import stable_softmax
_lowerCAmelCase = logging.get_logger(__name__)
@add_end_docstrings(SCREAMING_SNAKE_CASE_ )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,**__UpperCAmelCase ) -> Tuple:
super().__init__(**__UpperCAmelCase )
requires_backends(self ,"""vision""" )
self.check_model_type(
TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if self.framework == """tf"""
else MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING )
def __call__( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> str:
return super().__call__(__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[Any] = {}
if "candidate_labels" in kwargs:
lowerCAmelCase__ : int = kwargs["""candidate_labels"""]
if "hypothesis_template" in kwargs:
lowerCAmelCase__ : Optional[int] = kwargs["""hypothesis_template"""]
return preprocess_params, {}, {}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase="This is a photo of {}." ) -> int:
lowerCAmelCase__ : str = load_image(__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.image_processor(images=[image] ,return_tensors=self.framework )
lowerCAmelCase__ : List[Any] = candidate_labels
lowerCAmelCase__ : List[str] = [hypothesis_template.format(__UpperCAmelCase ) for x in candidate_labels]
lowerCAmelCase__ : Optional[Any] = self.tokenizer(__UpperCAmelCase ,return_tensors=self.framework ,padding=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = [text_inputs]
return inputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = model_inputs.pop("""candidate_labels""" )
lowerCAmelCase__ : Union[str, Any] = model_inputs.pop("""text_inputs""" )
if isinstance(text_inputs[0] ,__UpperCAmelCase ):
lowerCAmelCase__ : int = text_inputs[0]
else:
# Batching case.
lowerCAmelCase__ : Dict = text_inputs[0][0]
lowerCAmelCase__ : Any = self.model(**__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = {
"""candidate_labels""": candidate_labels,
"""logits""": outputs.logits_per_image,
}
return model_outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Union[str, Any] = model_outputs.pop("""candidate_labels""" )
lowerCAmelCase__ : List[str] = model_outputs["""logits"""][0]
if self.framework == "pt":
lowerCAmelCase__ : List[str] = logits.softmax(dim=-1 ).squeeze(-1 )
lowerCAmelCase__ : Optional[Any] = probs.tolist()
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Dict = [scores]
elif self.framework == "tf":
lowerCAmelCase__ : Any = stable_softmax(__UpperCAmelCase ,axis=-1 )
lowerCAmelCase__ : List[Any] = probs.numpy().tolist()
else:
raise ValueError(F"""Unsupported framework: {self.framework}""" )
lowerCAmelCase__ : Tuple = [
{"""score""": score, """label""": candidate_label}
for score, candidate_label in sorted(zip(__UpperCAmelCase ,__UpperCAmelCase ) ,key=lambda __UpperCAmelCase : -x[0] )
]
return result
| 37
| 1
|
'''simple docstring'''
import uuid
from typing import Any, Dict, List, Optional, Union
from ..utils import add_end_docstrings, is_tf_available, is_torch_available, logging
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_tf_available():
import tensorflow as tf
if is_torch_available():
import torch
_lowerCAmelCase = logging.get_logger(__name__)
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ) -> str:
if not conversation_id:
lowerCAmelCase__ : List[str] = uuid.uuida()
if past_user_inputs is None:
lowerCAmelCase__ : List[Any] = []
if generated_responses is None:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : uuid.UUID = conversation_id
lowerCAmelCase__ : List[str] = past_user_inputs
lowerCAmelCase__ : List[str] = generated_responses
lowerCAmelCase__ : Optional[str] = text
def __eq__( self ,__UpperCAmelCase ) -> Dict:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
return False
if self.uuid == other.uuid:
return True
return (
self.new_user_input == other.new_user_input
and self.past_user_inputs == other.past_user_inputs
and self.generated_responses == other.generated_responses
)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Optional[Any]:
if self.new_user_input:
if overwrite:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" was overwritten """
F"""with: \"{text}\".""" )
lowerCAmelCase__ : Optional[int] = text
else:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" new input """
F"""ignored: \"{text}\". Set `overwrite` to True to overwrite unprocessed user input""" )
else:
lowerCAmelCase__ : Optional[Any] = text
def UpperCAmelCase_ ( self ) -> List[Any]:
if self.new_user_input:
self.past_user_inputs.append(self.new_user_input )
lowerCAmelCase__ : Union[str, Any] = None
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
self.generated_responses.append(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
for user_input, generated_response in zip(self.past_user_inputs ,self.generated_responses ):
yield True, user_input
yield False, generated_response
if self.new_user_input:
yield True, self.new_user_input
def __repr__( self ) -> Tuple:
lowerCAmelCase__ : Tuple = F"""Conversation id: {self.uuid} \n"""
for is_user, text in self.iter_texts():
lowerCAmelCase__ : Any = """user""" if is_user else """bot"""
output += F"""{name} >> {text} \n"""
return output
@add_end_docstrings(
SCREAMING_SNAKE_CASE_ , R'''
min_length_for_response (`int`, *optional*, defaults to 32):
The minimum length (in number of tokens) for a response.
minimum_tokens (`int`, *optional*, defaults to 10):
The minimum length of tokens to leave for a response.
''' , )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Tuple:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
if self.tokenizer.pad_token_id is None:
lowerCAmelCase__ : Tuple = self.tokenizer.eos_token
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : List[Any] = {}
lowerCAmelCase__ : Optional[int] = {}
lowerCAmelCase__ : List[str] = {}
if min_length_for_response is not None:
lowerCAmelCase__ : Any = min_length_for_response
if minimum_tokens is not None:
lowerCAmelCase__ : Optional[int] = minimum_tokens
if "max_length" in generate_kwargs:
lowerCAmelCase__ : Optional[Any] = generate_kwargs["""max_length"""]
# self.max_length = generate_kwargs.get("max_length", self.model.config.max_length)
if clean_up_tokenization_spaces is not None:
lowerCAmelCase__ : int = clean_up_tokenization_spaces
if generate_kwargs:
forward_params.update(__UpperCAmelCase )
return preprocess_params, forward_params, postprocess_params
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase=0 ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Optional[int] = super().__call__(__UpperCAmelCase ,num_workers=__UpperCAmelCase ,**__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) and len(__UpperCAmelCase ) == 1:
return outputs[0]
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=32 ) -> Dict[str, Any]:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
raise ValueError("""ConversationalPipeline, expects Conversation as inputs""" )
if conversation.new_user_input is None:
raise ValueError(
F"""Conversation with UUID {type(conversation.uuid )} does not contain new user input to process. """
"""Add user inputs with the conversation's `add_user_input` method""" )
if hasattr(self.tokenizer ,"""_build_conversation_input_ids""" ):
lowerCAmelCase__ : str = self.tokenizer._build_conversation_input_ids(__UpperCAmelCase )
else:
# If the tokenizer cannot handle conversations, we default to only the old version
lowerCAmelCase__ : List[Any] = self._legacy_parse_and_tokenize(__UpperCAmelCase )
if self.framework == "pt":
lowerCAmelCase__ : List[Any] = torch.LongTensor([input_ids] )
elif self.framework == "tf":
lowerCAmelCase__ : Dict = tf.constant([input_ids] )
return {"input_ids": input_ids, "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=10 ,**__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[Any] = generate_kwargs.get("""max_length""" ,self.model.config.max_length )
lowerCAmelCase__ : Optional[Any] = model_inputs["""input_ids"""].shape[1]
if max_length - minimum_tokens < n:
logger.warning(F"""Conversation input is to long ({n}), trimming it to ({max_length} - {minimum_tokens})""" )
lowerCAmelCase__ : str = max_length - minimum_tokens
lowerCAmelCase__ : Union[str, Any] = model_inputs["""input_ids"""][:, -trim:]
if "attention_mask" in model_inputs:
lowerCAmelCase__ : Tuple = model_inputs["""attention_mask"""][:, -trim:]
lowerCAmelCase__ : str = model_inputs.pop("""conversation""" )
lowerCAmelCase__ : Union[str, Any] = max_length
lowerCAmelCase__ : Any = self.model.generate(**__UpperCAmelCase ,**__UpperCAmelCase )
if self.model.config.is_encoder_decoder:
lowerCAmelCase__ : int = 1
else:
lowerCAmelCase__ : int = n
return {"output_ids": output_ids[:, start_position:], "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=True ) -> List[str]:
lowerCAmelCase__ : Optional[int] = model_outputs["""output_ids"""]
lowerCAmelCase__ : Tuple = self.tokenizer.decode(
output_ids[0] ,skip_special_tokens=__UpperCAmelCase ,clean_up_tokenization_spaces=__UpperCAmelCase ,)
lowerCAmelCase__ : Union[str, Any] = model_outputs["""conversation"""]
conversation.mark_processed()
conversation.append_response(__UpperCAmelCase )
return conversation
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Dict = self.tokenizer.eos_token_id
lowerCAmelCase__ : int = []
for is_user, text in conversation.iter_texts():
if eos_token_id is not None:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) + [eos_token_id] )
else:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) )
if len(__UpperCAmelCase ) > self.tokenizer.model_max_length:
lowerCAmelCase__ : Optional[Any] = input_ids[-self.tokenizer.model_max_length :]
return input_ids
| 37
|
'''simple docstring'''
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , )
@pytest.mark.usefixtures('''sm_env''' )
@parameterized_class(
[
{
'''framework''': '''pytorch''',
'''script''': '''run_glue.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_5_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''pytorch''',
'''script''': '''run_ddp.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''tensorflow''',
'''script''': '''run_tf_dist.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.6, '''eval_loss''': 0.7},
},
] )
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.framework == "pytorch":
subprocess.run(
F"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() ,encoding="""utf-8""" ,check=__UpperCAmelCase ,)
assert hasattr(self ,"""env""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = F"""{self.env.base_job_name}-{instance_count}-{'ddp' if 'ddp' in self.script else 'smd'}"""
# distributed data settings
lowerCAmelCase__ : Any = {"""smdistributed""": {"""dataparallel""": {"""enabled""": True}}} if self.script != """run_ddp.py""" else None
# creates estimator
return HuggingFace(
entry_point=self.script ,source_dir=self.env.test_path ,role=self.env.role ,image_uri=self.env.image_uri ,base_job_name=__UpperCAmelCase ,instance_count=__UpperCAmelCase ,instance_type=self.instance_type ,debugger_hook_config=__UpperCAmelCase ,hyperparameters={**self.env.distributed_hyperparameters, """model_name_or_path""": self.model_name_or_path} ,metric_definitions=self.env.metric_definitions ,distribution=__UpperCAmelCase ,py_version="""py36""" ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
TrainingJobAnalytics(__UpperCAmelCase ).export_csv(F"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(2,)] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
# create estimator
lowerCAmelCase__ : List[Any] = self.create_estimator(__UpperCAmelCase )
# run training
estimator.fit()
# result dataframe
lowerCAmelCase__ : Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
lowerCAmelCase__ : int = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
lowerCAmelCase__ : List[Any] = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
lowerCAmelCase__ : List[str] = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" ,99_9999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(F"""{estimator.latest_training_job.name}.json""" ,"""w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} ,__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
import json
import os
from pathlib import Path
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple, Union
import sentencepiece
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = '''▁'''
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''spm_file''': '''sentencepiece.bpe.model''',
}
_lowerCAmelCase = {
'''vocab_file''': {
'''facebook/s2t-small-librispeech-asr''': (
'''https://huggingface.co/facebook/s2t-small-librispeech-asr/resolve/main/vocab.json'''
),
},
'''spm_file''': {
'''facebook/s2t-small-librispeech-asr''': (
'''https://huggingface.co/facebook/s2t-small-librispeech-asr/resolve/main/sentencepiece.bpe.model'''
)
},
}
_lowerCAmelCase = {
'''facebook/s2t-small-librispeech-asr''': 1024,
}
_lowerCAmelCase = ['''pt''', '''fr''', '''ru''', '''nl''', '''ro''', '''it''', '''es''', '''de''']
_lowerCAmelCase = {'''mustc''': MUSTC_LANGS}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Union[str, Any] = VOCAB_FILES_NAMES
__lowercase : List[str] = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = MAX_MODEL_INPUT_SIZES
__lowercase : Any = ['''input_ids''', '''attention_mask''']
__lowercase : List[int] = []
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<s>" ,__UpperCAmelCase="</s>" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
lowerCAmelCase__ : Optional[Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,do_upper_case=__UpperCAmelCase ,do_lower_case=__UpperCAmelCase ,tgt_lang=__UpperCAmelCase ,lang_codes=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : List[str] = do_upper_case
lowerCAmelCase__ : List[Any] = do_lower_case
lowerCAmelCase__ : Union[str, Any] = load_json(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = {v: k for k, v in self.encoder.items()}
lowerCAmelCase__ : List[Any] = spm_file
lowerCAmelCase__ : str = load_spm(__UpperCAmelCase ,self.sp_model_kwargs )
if lang_codes is not None:
lowerCAmelCase__ : Union[str, Any] = lang_codes
lowerCAmelCase__ : List[Any] = LANGUAGES[lang_codes]
lowerCAmelCase__ : List[str] = [F"""<lang:{lang}>""" for lang in self.langs]
lowerCAmelCase__ : Any = {lang: self.sp_model.PieceToId(F"""<lang:{lang}>""" ) for lang in self.langs}
lowerCAmelCase__ : List[str] = self.lang_tokens
lowerCAmelCase__ : Tuple = tgt_lang if tgt_lang is not None else self.langs[0]
self.set_tgt_lang_special_tokens(self._tgt_lang )
else:
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> int:
return len(self.encoder )
@property
def UpperCAmelCase_ ( self ) -> str:
return self._tgt_lang
@tgt_lang.setter
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> None:
lowerCAmelCase__ : Union[str, Any] = new_tgt_lang
self.set_tgt_lang_special_tokens(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> None:
lowerCAmelCase__ : str = self.lang_code_to_id[tgt_lang]
lowerCAmelCase__ : int = [lang_code_id]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.encoder.get(__UpperCAmelCase ,self.encoder[self.unk_token] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[Any] = []
lowerCAmelCase__ : Dict = """"""
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
lowerCAmelCase__ : Any = self.sp_model.decode(__UpperCAmelCase )
out_string += (decoded.upper() if self.do_upper_case else decoded) + token + " "
lowerCAmelCase__ : Any = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = self.sp_model.decode(__UpperCAmelCase )
out_string += decoded.upper() if self.do_upper_case else decoded
return out_string.strip()
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ) -> List[int]:
if token_ids_a is None:
return self.prefix_tokens + token_ids_a + [self.eos_token_id]
# We don't expect to process pairs, but leave the pair logic for API consistency
return self.prefix_tokens + token_ids_a + token_ids_a + [self.eos_token_id]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=__UpperCAmelCase ,token_ids_a=__UpperCAmelCase ,already_has_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : Any = [1] * len(self.prefix_tokens )
lowerCAmelCase__ : int = [1]
if token_ids_a is None:
return prefix_ones + ([0] * len(__UpperCAmelCase )) + suffix_ones
return prefix_ones + ([0] * len(__UpperCAmelCase )) + ([0] * len(__UpperCAmelCase )) + suffix_ones
def UpperCAmelCase_ ( self ) -> Dict:
lowerCAmelCase__ : int = self.encoder.copy()
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ) -> Dict:
lowerCAmelCase__ : Optional[Any] = self.__dict__.copy()
lowerCAmelCase__ : str = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> None:
lowerCAmelCase__ : List[Any] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Optional[Any] = {}
lowerCAmelCase__ : Optional[Any] = load_spm(self.spm_file ,self.sp_model_kwargs )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
lowerCAmelCase__ : Optional[int] = Path(__UpperCAmelCase )
assert save_dir.is_dir(), F"""{save_directory} should be a directory"""
lowerCAmelCase__ : Dict = save_dir / (
(filename_prefix + """-""" if filename_prefix else """""") + self.vocab_files_names["""vocab_file"""]
)
lowerCAmelCase__ : int = save_dir / (
(filename_prefix + """-""" if filename_prefix else """""") + self.vocab_files_names["""spm_file"""]
)
save_json(self.encoder ,__UpperCAmelCase )
if os.path.abspath(self.spm_file ) != os.path.abspath(__UpperCAmelCase ) and os.path.isfile(self.spm_file ):
copyfile(self.spm_file ,__UpperCAmelCase )
elif not os.path.isfile(self.spm_file ):
with open(__UpperCAmelCase ,"""wb""" ) as fi:
lowerCAmelCase__ : Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (str(__UpperCAmelCase ), str(__UpperCAmelCase ))
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = sentencepiece.SentencePieceProcessor(**UpperCamelCase )
spm.Load(str(UpperCamelCase ) )
return spm
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
with open(UpperCamelCase , """r""" ) as f:
return json.load(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
with open(UpperCamelCase , """w""" ) as f:
json.dump(UpperCamelCase , UpperCamelCase , indent=2 )
| 37
|
'''simple docstring'''
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''albert-base-v1''': '''https://huggingface.co/albert-base-v1/resolve/main/spiece.model''',
'''albert-large-v1''': '''https://huggingface.co/albert-large-v1/resolve/main/spiece.model''',
'''albert-xlarge-v1''': '''https://huggingface.co/albert-xlarge-v1/resolve/main/spiece.model''',
'''albert-xxlarge-v1''': '''https://huggingface.co/albert-xxlarge-v1/resolve/main/spiece.model''',
'''albert-base-v2''': '''https://huggingface.co/albert-base-v2/resolve/main/spiece.model''',
'''albert-large-v2''': '''https://huggingface.co/albert-large-v2/resolve/main/spiece.model''',
'''albert-xlarge-v2''': '''https://huggingface.co/albert-xlarge-v2/resolve/main/spiece.model''',
'''albert-xxlarge-v2''': '''https://huggingface.co/albert-xxlarge-v2/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''albert-base-v1''': 512,
'''albert-large-v1''': 512,
'''albert-xlarge-v1''': 512,
'''albert-xxlarge-v1''': 512,
'''albert-base-v2''': 512,
'''albert-large-v2''': 512,
'''albert-xlarge-v2''': 512,
'''albert-xxlarge-v2''': 512,
}
_lowerCAmelCase = '''▁'''
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Any = VOCAB_FILES_NAMES
__lowercase : Optional[Any] = PRETRAINED_VOCAB_FILES_MAP
__lowercase : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=True ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[MASK]" ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
# Mask token behave like a normal word, i.e. include the space before it and
# is included in the raw text, there should be a match in a non-normalized sentence.
lowerCAmelCase__ : Tuple = (
AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ,normalized=__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase )
else mask_token
)
lowerCAmelCase__ : Any = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : str = do_lower_case
lowerCAmelCase__ : int = remove_space
lowerCAmelCase__ : Tuple = keep_accents
lowerCAmelCase__ : Any = vocab_file
lowerCAmelCase__ : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> Optional[int]:
return len(self.sp_model )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Union[str, Any] = {self.convert_ids_to_tokens(__UpperCAmelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = self.__dict__.copy()
lowerCAmelCase__ : Optional[Any] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Union[str, Any] = {}
lowerCAmelCase__ : Any = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[int]:
if self.remove_space:
lowerCAmelCase__ : int = """ """.join(inputs.strip().split() )
else:
lowerCAmelCase__ : str = inputs
lowerCAmelCase__ : Tuple = outputs.replace("""``""" ,"""\"""" ).replace("""''""" ,"""\"""" )
if not self.keep_accents:
lowerCAmelCase__ : Any = unicodedata.normalize("""NFKD""" ,__UpperCAmelCase )
lowerCAmelCase__ : Dict = """""".join([c for c in outputs if not unicodedata.combining(__UpperCAmelCase )] )
if self.do_lower_case:
lowerCAmelCase__ : Tuple = outputs.lower()
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
lowerCAmelCase__ : str = []
for piece in pieces:
if len(__UpperCAmelCase ) > 1 and piece[-1] == str(""",""" ) and piece[-2].isdigit():
lowerCAmelCase__ : List[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(__UpperCAmelCase ,"""""" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
lowerCAmelCase__ : str = cur_pieces[1:]
else:
lowerCAmelCase__ : int = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(__UpperCAmelCase )
else:
new_pieces.append(__UpperCAmelCase )
return new_pieces
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
return self.sp_model.IdToPiece(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : Tuple = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : List[str] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string.strip()
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : int = [self.sep_token_id]
lowerCAmelCase__ : Dict = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
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 not None:
return [1] + ([0] * len(__UpperCAmelCase )) + [1] + ([0] * len(__UpperCAmelCase )) + [1]
return [1] + ([0] * len(__UpperCAmelCase )) + [1]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : List[str] = [self.sep_token_id]
lowerCAmelCase__ : Tuple = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : int = 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:
lowerCAmelCase__ : List[Any] = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
| 37
| 1
|
'''simple docstring'''
from __future__ import annotations
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = 0.00
lowerCAmelCase__ : List[Any] = 0
for resistor in resistors:
if resistor <= 0:
lowerCAmelCase__ : Dict = f"""Resistor at index {index} has a negative or zero value!"""
raise ValueError(UpperCamelCase )
first_sum += 1 / float(UpperCamelCase )
index += 1
return 1 / first_sum
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = 0.00
lowerCAmelCase__ : List[str] = 0
for resistor in resistors:
sum_r += resistor
if resistor < 0:
lowerCAmelCase__ : Union[str, Any] = f"""Resistor at index {index} has a negative value!"""
raise ValueError(UpperCamelCase )
index += 1
return sum_r
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''merges_file''': '''merges.txt''',
}
_lowerCAmelCase = {
'''vocab_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json'''},
'''merges_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt'''},
}
_lowerCAmelCase = {
'''ctrl''': 256,
}
_lowerCAmelCase = {
'''Pregnancy''': 16_8629,
'''Christianity''': 7675,
'''Explain''': 10_6423,
'''Fitness''': 6_3440,
'''Saving''': 6_3163,
'''Ask''': 2_7171,
'''Ass''': 9_5985,
'''Joke''': 16_3509,
'''Questions''': 4_5622,
'''Thoughts''': 4_9605,
'''Retail''': 5_2342,
'''Feminism''': 16_4338,
'''Writing''': 1_1992,
'''Atheism''': 19_2263,
'''Netflix''': 4_8616,
'''Computing''': 3_9639,
'''Opinion''': 4_3213,
'''Alone''': 4_4967,
'''Funny''': 5_8917,
'''Gaming''': 4_0358,
'''Human''': 4088,
'''India''': 1331,
'''Joker''': 7_7138,
'''Diet''': 3_6206,
'''Legal''': 1_1859,
'''Norman''': 4939,
'''Tip''': 7_2689,
'''Weight''': 5_2343,
'''Movies''': 4_6273,
'''Running''': 2_3425,
'''Science''': 2090,
'''Horror''': 3_7793,
'''Confession''': 6_0572,
'''Finance''': 1_2250,
'''Politics''': 1_6360,
'''Scary''': 19_1985,
'''Support''': 1_2654,
'''Technologies''': 3_2516,
'''Teenage''': 6_6160,
'''Event''': 3_2769,
'''Learned''': 6_7460,
'''Notion''': 18_2770,
'''Wikipedia''': 3_7583,
'''Books''': 6665,
'''Extract''': 7_6050,
'''Confessions''': 10_2701,
'''Conspiracy''': 7_5932,
'''Links''': 6_3674,
'''Narcissus''': 15_0425,
'''Relationship''': 5_4766,
'''Relationships''': 13_4796,
'''Reviews''': 4_1671,
'''News''': 4256,
'''Translation''': 2_6820,
'''multilingual''': 12_8406,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = set()
lowerCAmelCase__ : str = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCAmelCase__ : List[Any] = char
lowerCAmelCase__ : Optional[Any] = set(UpperCamelCase )
return pairs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Any = CONTROL_CODES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<unk>" ,**__UpperCAmelCase ) -> Optional[int]:
super().__init__(unk_token=__UpperCAmelCase ,**__UpperCAmelCase )
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as vocab_handle:
lowerCAmelCase__ : int = json.load(__UpperCAmelCase )
lowerCAmelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as merges_handle:
lowerCAmelCase__ : Union[str, Any] = merges_handle.read().split("""\n""" )[1:-1]
lowerCAmelCase__ : Optional[Any] = [tuple(merge.split() ) for merge in merges]
lowerCAmelCase__ : int = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return len(self.encoder )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return dict(self.encoder ,**self.added_tokens_encoder )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if token in self.cache:
return self.cache[token]
lowerCAmelCase__ : Optional[Any] = tuple(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tuple(list(word[:-1] ) + [word[-1] + """</w>"""] )
lowerCAmelCase__ : Any = get_pairs(__UpperCAmelCase )
if not pairs:
return token
while True:
lowerCAmelCase__ : List[str] = min(__UpperCAmelCase ,key=lambda __UpperCAmelCase : self.bpe_ranks.get(__UpperCAmelCase ,float("""inf""" ) ) )
if bigram not in self.bpe_ranks:
break
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = bigram
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : Dict = 0
while i < len(__UpperCAmelCase ):
try:
lowerCAmelCase__ : Optional[Any] = word.index(__UpperCAmelCase ,__UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCAmelCase__ : Dict = 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
lowerCAmelCase__ : Tuple = tuple(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = new_word
if len(__UpperCAmelCase ) == 1:
break
else:
lowerCAmelCase__ : Dict = get_pairs(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """@@ """.join(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = word[:-4]
lowerCAmelCase__ : Optional[Any] = word
return word
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : Any = re.findall(R"""\S+\n?""" ,__UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(__UpperCAmelCase ).split(""" """ ) ) )
return split_tokens
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
return self.encoder.get(__UpperCAmelCase ,self.encoder.get(self.unk_token ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = """ """.join(__UpperCAmelCase ).replace("""@@ """ ,"""""" ).strip()
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : List[Any] = os.path.join(
__UpperCAmelCase ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : Optional[int] = 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""" )
lowerCAmelCase__ : int = 0
with open(__UpperCAmelCase ,"""w""" ,encoding="""utf-8""" ) as writer:
writer.write("""#version: 0.2\n""" )
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 {merge_file}: BPE merge indices are not consecutive."""
""" Please check that the tokenizer is not corrupted!""" )
lowerCAmelCase__ : Dict = token_index
writer.write(""" """.join(__UpperCAmelCase ) + """\n""" )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 37
| 1
|
'''simple docstring'''
import argparse
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
########################################################################
# This is a fully working simple example to use Accelerate,
# specifically showcasing the experiment tracking capability,
# and builds off the `nlp_example.py` script.
#
# This example trains a Bert base model on GLUE MRPC
# in any of the following settings (with the same script):
# - single CPU or single GPU
# - multi GPUS (using PyTorch distributed mode)
# - (multi) TPUs
# - fp16 (mixed-precision) or fp32 (normal precision)
#
# To help focus on the differences in the code, building `DataLoaders`
# was refactored into its own function.
# New additions from the base script can be found quickly by
# looking for the # New Code # tags
#
# To run it in each of these various modes, follow the instructions
# in the readme for examples:
# https://github.com/huggingface/accelerate/tree/main/examples
#
########################################################################
_lowerCAmelCase = 16
_lowerCAmelCase = 32
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase = 16 ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = AutoTokenizer.from_pretrained("""bert-base-cased""" )
lowerCAmelCase__ : int = load_dataset("""glue""" , """mrpc""" )
def tokenize_function(UpperCamelCase ):
# max_length=None => use the model max length (it's actually the default)
lowerCAmelCase__ : List[Any] = tokenizer(examples["""sentence1"""] , examples["""sentence2"""] , truncation=UpperCamelCase , max_length=UpperCamelCase )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
# starting with the main process first:
with accelerator.main_process_first():
lowerCAmelCase__ : Optional[int] = datasets.map(
UpperCamelCase , batched=UpperCamelCase , remove_columns=["""idx""", """sentence1""", """sentence2"""] , )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
lowerCAmelCase__ : Optional[int] = tokenized_datasets.rename_column("""label""" , """labels""" )
def collate_fn(UpperCamelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
lowerCAmelCase__ : List[str] = 128 if accelerator.distributed_type == DistributedType.TPU else None
# When using mixed precision we want round multiples of 8/16
if accelerator.mixed_precision == "fp8":
lowerCAmelCase__ : Optional[int] = 16
elif accelerator.mixed_precision != "no":
lowerCAmelCase__ : Dict = 8
else:
lowerCAmelCase__ : Any = None
return tokenizer.pad(
UpperCamelCase , padding="""longest""" , max_length=UpperCamelCase , pad_to_multiple_of=UpperCamelCase , return_tensors="""pt""" , )
# Instantiate dataloaders.
lowerCAmelCase__ : List[Any] = DataLoader(
tokenized_datasets["""train"""] , shuffle=UpperCamelCase , collate_fn=UpperCamelCase , batch_size=UpperCamelCase )
lowerCAmelCase__ : List[Any] = DataLoader(
tokenized_datasets["""validation"""] , shuffle=UpperCamelCase , collate_fn=UpperCamelCase , batch_size=UpperCamelCase )
return train_dataloader, eval_dataloader
# For testing only
if os.environ.get('''TESTING_MOCKED_DATALOADERS''', None) == "1":
from accelerate.test_utils.training import mocked_dataloaders
_lowerCAmelCase = mocked_dataloaders # noqa: F811
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
if os.environ.get("""TESTING_MOCKED_DATALOADERS""" , UpperCamelCase ) == "1":
lowerCAmelCase__ : Tuple = 2
# Initialize Accelerator
# New Code #
# We pass in "all" to `log_with` to grab all available trackers in the environment
# Note: If using a custom `Tracker` class, should be passed in here such as:
# >>> log_with = ["all", MyCustomTrackerClassInstance()]
if args.with_tracking:
lowerCAmelCase__ : Dict = Accelerator(
cpu=args.cpu , mixed_precision=args.mixed_precision , log_with="""all""" , project_dir=args.project_dir )
else:
lowerCAmelCase__ : Tuple = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision )
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
lowerCAmelCase__ : List[str] = config["""lr"""]
lowerCAmelCase__ : Any = int(config["""num_epochs"""] )
lowerCAmelCase__ : List[str] = int(config["""seed"""] )
lowerCAmelCase__ : List[str] = int(config["""batch_size"""] )
set_seed(UpperCamelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = get_dataloaders(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : str = evaluate.load("""glue""" , """mrpc""" )
# If the batch size is too big we use gradient accumulation
lowerCAmelCase__ : Optional[int] = 1
if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU:
lowerCAmelCase__ : Dict = batch_size // MAX_GPU_BATCH_SIZE
lowerCAmelCase__ : str = MAX_GPU_BATCH_SIZE
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
lowerCAmelCase__ : List[str] = AutoModelForSequenceClassification.from_pretrained("""bert-base-cased""" , return_dict=UpperCamelCase )
# We could avoid this line since the accelerator is set with `device_placement=True` (default value).
# Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer
# creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that).
lowerCAmelCase__ : str = model.to(accelerator.device )
# Instantiate optimizer
lowerCAmelCase__ : str = AdamW(params=model.parameters() , lr=UpperCamelCase )
# Instantiate scheduler
lowerCAmelCase__ : Dict = get_linear_schedule_with_warmup(
optimizer=UpperCamelCase , num_warmup_steps=100 , num_training_steps=(len(UpperCamelCase ) * num_epochs) // gradient_accumulation_steps , )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = accelerator.prepare(
UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase )
# New Code #
# We need to initialize the trackers we use. Overall configurations can also be stored
if args.with_tracking:
lowerCAmelCase__ : Dict = os.path.split(UpperCamelCase )[-1].split(""".""" )[0]
accelerator.init_trackers(UpperCamelCase , UpperCamelCase )
# Now we train the model
for epoch in range(UpperCamelCase ):
model.train()
# New Code #
# For our tracking example, we will log the total loss of each epoch
if args.with_tracking:
lowerCAmelCase__ : Dict = 0
for step, batch in enumerate(UpperCamelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
lowerCAmelCase__ : Tuple = model(**UpperCamelCase )
lowerCAmelCase__ : List[str] = outputs.loss
# New Code #
if args.with_tracking:
total_loss += loss.detach().float()
lowerCAmelCase__ : Optional[Any] = loss / gradient_accumulation_steps
accelerator.backward(UpperCamelCase )
if step % gradient_accumulation_steps == 0:
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
model.eval()
for step, batch in enumerate(UpperCamelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True` (the default).
batch.to(accelerator.device )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(**UpperCamelCase )
lowerCAmelCase__ : Optional[int] = outputs.logits.argmax(dim=-1 )
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = accelerator.gather_for_metrics((predictions, batch["""labels"""]) )
metric.add_batch(
predictions=UpperCamelCase , references=UpperCamelCase , )
lowerCAmelCase__ : Optional[int] = metric.compute()
# Use accelerator.print to print only on the main process.
accelerator.print(f"""epoch {epoch}:""" , UpperCamelCase )
# New Code #
# To actually log, we call `Accelerator.log`
# The values passed can be of `str`, `int`, `float` or `dict` of `str` to `float`/`int`
if args.with_tracking:
accelerator.log(
{
"""accuracy""": eval_metric["""accuracy"""],
"""f1""": eval_metric["""f1"""],
"""train_loss""": total_loss.item() / len(UpperCamelCase ),
"""epoch""": epoch,
} , step=UpperCamelCase , )
# New Code #
# When a run is finished, you should call `accelerator.end_training()`
# to close all of the open trackers
if args.with_tracking:
accelerator.end_training()
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = argparse.ArgumentParser(description="""Simple example of training script.""" )
parser.add_argument(
"""--mixed_precision""" , type=UpperCamelCase , default=UpperCamelCase , choices=["""no""", """fp16""", """bf16""", """fp8"""] , help="""Whether to use mixed precision. Choose"""
"""between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10."""
"""and an Nvidia Ampere GPU.""" , )
parser.add_argument("""--cpu""" , action="""store_true""" , help="""If passed, will train on the CPU.""" )
parser.add_argument(
"""--with_tracking""" , action="""store_true""" , help="""Whether to load in all available experiment trackers from the environment and use them for logging.""" , )
parser.add_argument(
"""--project_dir""" , type=UpperCamelCase , default="""logs""" , help="""Location on where to store experiment tracking logs` and relevent project information""" , )
lowerCAmelCase__ : List[Any] = parser.parse_args()
lowerCAmelCase__ : Any = {"""lr""": 2e-5, """num_epochs""": 3, """seed""": 42, """batch_size""": 16}
training_function(UpperCamelCase , UpperCamelCase )
if __name__ == "__main__":
main()
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : str = set()
# Replace all the whitespace in our sentence
lowerCAmelCase__ : Tuple = input_str.replace(""" """ , """""" )
for alpha in input_str:
if "a" <= alpha.lower() <= "z":
frequency.add(alpha.lower() )
return len(UpperCamelCase ) == 26
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : Any = [False] * 26
for char in input_str:
if char.islower():
lowerCAmelCase__ : Optional[Any] = True
elif char.isupper():
lowerCAmelCase__ : Any = True
return all(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
return len({char for char in input_str.lower() if char.isalpha()} ) == 26
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from timeit import timeit
lowerCAmelCase__ : Union[str, Any] = """from __main__ import is_pangram, is_pangram_faster, is_pangram_fastest"""
print(timeit("""is_pangram()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_faster()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_fastest()""" , setup=UpperCamelCase ) )
# 5.348480500048026, 2.6477354579837993, 1.8470395830227062
# 5.036091582966037, 2.644472333951853, 1.8869528750656173
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import argparse
import glob
import logging
import os
import time
from argparse import Namespace
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from torch.utils.data import DataLoader, TensorDataset
from transformers import glue_compute_metrics as compute_metrics
from transformers import glue_convert_examples_to_features as convert_examples_to_features
from transformers import glue_output_modes, glue_tasks_num_labels
from transformers import glue_processors as processors
_lowerCAmelCase = logging.getLogger(__name__)
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = '''sequence-classification'''
def __init__( self ,__UpperCAmelCase ) -> Dict:
if type(__UpperCAmelCase ) == dict:
lowerCAmelCase__ : Optional[Any] = Namespace(**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = glue_output_modes[hparams.task]
lowerCAmelCase__ : Any = glue_tasks_num_labels[hparams.task]
super().__init__(__UpperCAmelCase ,__UpperCAmelCase ,self.mode )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> Union[str, Any]:
return self.model(**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type not in ["distilbert", "bart"]:
lowerCAmelCase__ : List[str] = batch[2] if self.config.model_type in ["""bert""", """xlnet""", """albert"""] else None
lowerCAmelCase__ : Optional[Any] = self(**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = outputs[0]
lowerCAmelCase__ : Dict = self.trainer.lr_schedulers[0]["""scheduler"""]
lowerCAmelCase__ : List[Any] = {"""loss""": loss, """rate""": lr_scheduler.get_last_lr()[-1]}
return {"loss": loss, "log": tensorboard_logs}
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : Any = self.hparams
lowerCAmelCase__ : List[str] = processors[args.task]()
lowerCAmelCase__ : str = processor.get_labels()
for mode in ["train", "dev"]:
lowerCAmelCase__ : Tuple = self._feature_file(__UpperCAmelCase )
if os.path.exists(__UpperCAmelCase ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" ,__UpperCAmelCase )
else:
logger.info("""Creating features from dataset file at %s""" ,args.data_dir )
lowerCAmelCase__ : List[Any] = (
processor.get_dev_examples(args.data_dir )
if mode == """dev"""
else processor.get_train_examples(args.data_dir )
)
lowerCAmelCase__ : Optional[Any] = convert_examples_to_features(
__UpperCAmelCase ,self.tokenizer ,max_length=args.max_seq_length ,label_list=self.labels ,output_mode=args.glue_output_mode ,)
logger.info("""Saving features into cached file %s""" ,__UpperCAmelCase )
torch.save(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> DataLoader:
lowerCAmelCase__ : Optional[int] = """dev""" if mode == """test""" else mode
lowerCAmelCase__ : Optional[int] = self._feature_file(__UpperCAmelCase )
logger.info("""Loading features from cached file %s""" ,__UpperCAmelCase )
lowerCAmelCase__ : Tuple = torch.load(__UpperCAmelCase )
lowerCAmelCase__ : str = torch.tensor([f.input_ids for f in features] ,dtype=torch.long )
lowerCAmelCase__ : str = torch.tensor([f.attention_mask for f in features] ,dtype=torch.long )
lowerCAmelCase__ : List[str] = torch.tensor([f.token_type_ids for f in features] ,dtype=torch.long )
if self.hparams.glue_output_mode == "classification":
lowerCAmelCase__ : Optional[int] = torch.tensor([f.label for f in features] ,dtype=torch.long )
elif self.hparams.glue_output_mode == "regression":
lowerCAmelCase__ : Any = torch.tensor([f.label for f in features] ,dtype=torch.float )
return DataLoader(
TensorDataset(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) ,batch_size=__UpperCAmelCase ,shuffle=__UpperCAmelCase ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type not in ["distilbert", "bart"]:
lowerCAmelCase__ : Optional[int] = batch[2] if self.config.model_type in ["""bert""", """xlnet""", """albert"""] else None
lowerCAmelCase__ : Dict = self(**__UpperCAmelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = outputs[:2]
lowerCAmelCase__ : Dict = logits.detach().cpu().numpy()
lowerCAmelCase__ : Any = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> tuple:
lowerCAmelCase__ : Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean().detach().cpu().item()
lowerCAmelCase__ : Dict = np.concatenate([x["""pred"""] for x in outputs] ,axis=0 )
if self.hparams.glue_output_mode == "classification":
lowerCAmelCase__ : Union[str, Any] = np.argmax(__UpperCAmelCase ,axis=1 )
elif self.hparams.glue_output_mode == "regression":
lowerCAmelCase__ : Tuple = np.squeeze(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = np.concatenate([x["""target"""] for x in outputs] ,axis=0 )
lowerCAmelCase__ : Dict = [[] for _ in range(out_label_ids.shape[0] )]
lowerCAmelCase__ : List[Any] = [[] for _ in range(out_label_ids.shape[0] )]
lowerCAmelCase__ : List[Any] = {**{"""val_loss""": val_loss_mean}, **compute_metrics(self.hparams.task ,__UpperCAmelCase ,__UpperCAmelCase )}
lowerCAmelCase__ : Dict = dict(results.items() )
lowerCAmelCase__ : Optional[Any] = results
return ret, preds_list, out_label_list
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> dict:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self._eval_end(__UpperCAmelCase )
lowerCAmelCase__ : int = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> dict:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = self._eval_end(__UpperCAmelCase )
lowerCAmelCase__ : str = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def UpperCAmelCase_ ( __UpperCAmelCase ,__UpperCAmelCase ) -> int:
BaseTransformer.add_model_specific_args(__UpperCAmelCase ,__UpperCAmelCase )
parser.add_argument(
"""--max_seq_length""" ,default=128 ,type=__UpperCAmelCase ,help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) ,)
parser.add_argument(
"""--task""" ,default="""""" ,type=__UpperCAmelCase ,required=__UpperCAmelCase ,help="""The GLUE task to run""" ,)
parser.add_argument(
"""--gpus""" ,default=0 ,type=__UpperCAmelCase ,help="""The number of GPUs allocated for this, it is by default 0 meaning none""" ,)
parser.add_argument(
"""--overwrite_cache""" ,action="""store_true""" ,help="""Overwrite the cached training and evaluation sets""" )
return parser
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Any = argparse.ArgumentParser()
add_generic_args(UpperCamelCase , os.getcwd() )
lowerCAmelCase__ : Optional[int] = GLUETransformer.add_model_specific_args(UpperCamelCase , os.getcwd() )
lowerCAmelCase__ : Optional[Any] = parser.parse_args()
# If output_dir not provided, a folder will be generated in pwd
if args.output_dir is None:
lowerCAmelCase__ : Union[str, Any] = os.path.join(
"""./results""" , f"""{args.task}_{time.strftime('%Y%m%d_%H%M%S' )}""" , )
os.makedirs(args.output_dir )
lowerCAmelCase__ : str = GLUETransformer(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = generic_train(UpperCamelCase , UpperCamelCase )
# Optionally, predict on dev set and write to output_dir
if args.do_predict:
lowerCAmelCase__ : str = sorted(glob.glob(os.path.join(args.output_dir , """checkpoint-epoch=*.ckpt""" ) , recursive=UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = model.load_from_checkpoint(checkpoints[-1] )
return trainer.test(UpperCamelCase )
if __name__ == "__main__":
main()
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = abs(UpperCamelCase )
lowerCAmelCase__ : List[Any] = 0
while n > 0:
res += n % 10
n //= 10
return res
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = abs(UpperCamelCase )
return n if n < 10 else n % 10 + sum_of_digits(n // 10 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return sum(int(UpperCamelCase ) for c in str(abs(UpperCamelCase ) ) )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(UpperCamelCase , UpperCamelCase ) -> None:
lowerCAmelCase__ : str = f"""{func.__name__}({value})"""
lowerCAmelCase__ : str = timeit(f"""__main__.{call}""" , setup="""import __main__""" )
print(f"""{call:56} = {func(UpperCamelCase )} -- {timing:.4f} seconds""" )
for value in (262144, 1125899906842624, 1267650600228229401496703205376):
for func in (sum_of_digits, sum_of_digits_recursion, sum_of_digits_compact):
benchmark_a_function(UpperCamelCase , UpperCamelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import json
import os
import re
import sys
import urllib.request
import requests
from bsa import BeautifulSoup
_lowerCAmelCase = {
'''User-Agent''': '''Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'''
''' (KHTML, like Gecko) Chrome/70.0.3538.102 Safari/537.36 Edge/18.19582'''
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "dhaka" , UpperCamelCase = 5 ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = min(UpperCamelCase , 50 ) # Prevent abuse!
lowerCAmelCase__ : List[Any] = {
"""q""": query,
"""tbm""": """isch""",
"""hl""": """en""",
"""ijn""": """0""",
}
lowerCAmelCase__ : List[Any] = requests.get("""https://www.google.com/search""" , params=UpperCamelCase , headers=UpperCamelCase )
lowerCAmelCase__ : List[Any] = BeautifulSoup(html.text , """html.parser""" )
lowerCAmelCase__ : int = """""".join(
re.findall(R"""AF_initDataCallback\(([^<]+)\);""" , str(soup.select("""script""" ) ) ) )
lowerCAmelCase__ : Dict = json.dumps(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = json.loads(UpperCamelCase )
lowerCAmelCase__ : Dict = re.findall(
R"""\[\"GRID_STATE0\",null,\[\[1,\[0,\".*?\",(.*),\"All\",""" , UpperCamelCase , )
if not matched_google_image_data:
return 0
lowerCAmelCase__ : Tuple = re.sub(
R"""\[\"(https\:\/\/encrypted-tbn0\.gstatic\.com\/images\?.*?)\",\d+,\d+\]""" , """""" , str(UpperCamelCase ) , )
lowerCAmelCase__ : Union[str, Any] = re.findall(
R"""(?:'|,),\[\"(https:|http.*?)\",\d+,\d+\]""" , UpperCamelCase , )
for index, fixed_full_res_image in enumerate(UpperCamelCase ):
if index >= max_images:
return index
lowerCAmelCase__ : Optional[int] = bytes(UpperCamelCase , """ascii""" ).decode(
"""unicode-escape""" )
lowerCAmelCase__ : Optional[Any] = bytes(UpperCamelCase , """ascii""" ).decode(
"""unicode-escape""" )
lowerCAmelCase__ : Tuple = urllib.request.build_opener()
lowerCAmelCase__ : int = [
(
"""User-Agent""",
"""Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"""
""" (KHTML, like Gecko) Chrome/70.0.3538.102 Safari/537.36 Edge/18.19582""",
)
]
urllib.request.install_opener(UpperCamelCase )
lowerCAmelCase__ : List[Any] = f"""query_{query.replace(' ' , '_' )}"""
if not os.path.exists(UpperCamelCase ):
os.makedirs(UpperCamelCase )
urllib.request.urlretrieve( # noqa: S310
UpperCamelCase , f"""{path_name}/original_size_img_{index}.jpg""" )
return index
if __name__ == "__main__":
try:
_lowerCAmelCase = download_images_from_google_query(sys.argv[1])
print(F"""{image_count} images were downloaded to disk.""")
except IndexError:
print('''Please provide a search term.''')
raise
| 37
|
'''simple docstring'''
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team.
#
# 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.
# this script dumps information about the environment
import os
import sys
import transformers
_lowerCAmelCase = '''3'''
print('''Python version:''', sys.version)
print('''transformers version:''', transformers.__version__)
try:
import torch
print('''Torch version:''', torch.__version__)
print('''Cuda available:''', torch.cuda.is_available())
print('''Cuda version:''', torch.version.cuda)
print('''CuDNN version:''', torch.backends.cudnn.version())
print('''Number of GPUs available:''', torch.cuda.device_count())
print('''NCCL version:''', torch.cuda.nccl.version())
except ImportError:
print('''Torch version:''', None)
try:
import deepspeed
print('''DeepSpeed version:''', deepspeed.__version__)
except ImportError:
print('''DeepSpeed version:''', None)
try:
import tensorflow as tf
print('''TensorFlow version:''', tf.__version__)
print('''TF GPUs available:''', bool(tf.config.list_physical_devices('''GPU''')))
print('''Number of TF GPUs available:''', len(tf.config.list_physical_devices('''GPU''')))
except ImportError:
print('''TensorFlow version:''', None)
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = len(UpperCamelCase )
for i in range(1 , UpperCamelCase ):
lowerCAmelCase__ : Optional[int] = collection[i]
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Dict = i - 1
while low <= high:
lowerCAmelCase__ : List[str] = (low + high) // 2
if val < collection[mid]:
lowerCAmelCase__ : List[str] = mid - 1
else:
lowerCAmelCase__ : Optional[Any] = mid + 1
for j in range(UpperCamelCase , UpperCamelCase , -1 ):
lowerCAmelCase__ : Optional[Any] = collection[j - 1]
lowerCAmelCase__ : Any = val
return collection
if __name__ == "__main__":
_lowerCAmelCase = input('''Enter numbers separated by a comma:\n''').strip()
_lowerCAmelCase = [int(item) for item in user_input.split(''',''')]
print(binary_insertion_sort(unsorted))
| 37
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xlm-roberta-xl''': '''https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json''',
'''facebook/xlm-roberta-xxl''': '''https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json''',
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''xlm-roberta-xl'''
def __init__( self ,__UpperCAmelCase=25_0880 ,__UpperCAmelCase=2560 ,__UpperCAmelCase=36 ,__UpperCAmelCase=32 ,__UpperCAmelCase=1_0240 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=514 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-05 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = vocab_size
lowerCAmelCase__ : int = hidden_size
lowerCAmelCase__ : int = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : str = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : List[str] = type_vocab_size
lowerCAmelCase__ : List[Any] = initializer_range
lowerCAmelCase__ : Tuple = layer_norm_eps
lowerCAmelCase__ : int = position_embedding_type
lowerCAmelCase__ : Optional[Any] = use_cache
lowerCAmelCase__ : Optional[Any] = classifier_dropout
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Dict = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Any = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
| 1
|
'''simple docstring'''
from copy import deepcopy
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ) -> None:
if arr is None and size is not None:
lowerCAmelCase__ : List[str] = size
lowerCAmelCase__ : str = [0] * size
elif arr is not None:
self.init(__UpperCAmelCase )
else:
raise ValueError("""Either arr or size must be specified""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> None:
lowerCAmelCase__ : Any = len(__UpperCAmelCase )
lowerCAmelCase__ : Any = deepcopy(__UpperCAmelCase )
for i in range(1 ,self.size ):
lowerCAmelCase__ : Optional[Any] = self.next_(__UpperCAmelCase )
if j < self.size:
self.tree[j] += self.tree[i]
def UpperCAmelCase_ ( self ) -> list[int]:
lowerCAmelCase__ : Dict = self.tree[:]
for i in range(self.size - 1 ,0 ,-1 ):
lowerCAmelCase__ : Union[str, Any] = 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 ,__UpperCAmelCase ,__UpperCAmelCase ) -> None:
if index == 0:
self.tree[0] += value
return
while index < self.size:
self.tree[index] += value
lowerCAmelCase__ : Optional[Any] = self.next_(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> None:
self.add(__UpperCAmelCase ,value - self.get(__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if right == 0:
return 0
lowerCAmelCase__ : List[Any] = self.tree[0]
right -= 1 # make right inclusive
while right > 0:
result += self.tree[right]
lowerCAmelCase__ : Union[str, Any] = self.prev(__UpperCAmelCase )
return result
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> int:
return self.prefix(__UpperCAmelCase ) - self.prefix(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return self.query(__UpperCAmelCase ,index + 1 )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
value -= self.tree[0]
if value < 0:
return -1
lowerCAmelCase__ : Any = 1 # Largest power of 2 <= size
while j * 2 < self.size:
j *= 2
lowerCAmelCase__ : int = 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()
| 37
|
'''simple docstring'''
from __future__ import annotations
import math
from collections import Counter
from string import ascii_lowercase
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = analyze_text(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = list(""" """ + ascii_lowercase )
# what is our total sum of probabilities.
lowerCAmelCase__ : List[Any] = sum(single_char_strings.values() )
# one length string
lowerCAmelCase__ : Optional[int] = 0
# for each alpha we go in our dict and if it is in it we calculate entropy
for ch in my_alphas:
if ch in single_char_strings:
lowerCAmelCase__ : List[Any] = single_char_strings[ch]
lowerCAmelCase__ : List[Any] = my_str / all_sum
my_fir_sum += prob * math.loga(UpperCamelCase ) # entropy formula.
# print entropy
print(f"""{round(-1 * my_fir_sum ):.1f}""" )
# two len string
lowerCAmelCase__ : Dict = sum(two_char_strings.values() )
lowerCAmelCase__ : int = 0
# for each alpha (two in size) calculate entropy.
for cha in my_alphas:
for cha in my_alphas:
lowerCAmelCase__ : Union[str, Any] = cha + cha
if sequence in two_char_strings:
lowerCAmelCase__ : Dict = two_char_strings[sequence]
lowerCAmelCase__ : Tuple = int(UpperCamelCase ) / all_sum
my_sec_sum += prob * math.loga(UpperCamelCase )
# print second entropy
print(f"""{round(-1 * my_sec_sum ):.1f}""" )
# print the difference between them
print(f"""{round((-1 * my_sec_sum) - (-1 * my_fir_sum) ):.1f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = Counter() # type: ignore
lowerCAmelCase__ : Tuple = Counter() # type: ignore
single_char_strings[text[-1]] += 1
# first case when we have space at start.
two_char_strings[" " + text[0]] += 1
for i in range(0 , len(UpperCamelCase ) - 1 ):
single_char_strings[text[i]] += 1
two_char_strings[text[i : i + 2]] += 1
return single_char_strings, two_char_strings
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
import doctest
doctest.testmod()
# text = (
# "Had repulsive dashwoods suspicion sincerity but advantage now him. Remark "
# "easily garret nor nay. Civil those mrs enjoy shy fat merry. You greatest "
# "jointure saw horrible. He private he on be imagine suppose. Fertile "
# "beloved evident through no service elderly is. Blind there if every no so "
# "at. Own neglected you preferred way sincerity delivered his attempted. To "
# "of message cottage windows do besides against uncivil. Delightful "
# "unreserved impossible few estimating men favourable see entreaties. She "
# "propriety immediate was improving. He or entrance humoured likewise "
# "moderate. Much nor game son say feel. Fat make met can must form into "
# "gate. Me we offending prevailed discovery. "
# )
# calculate_prob(text)
if __name__ == "__main__":
main()
| 37
| 1
|
'''simple docstring'''
import argparse
import os
import pickle
import sys
import torch
from transformers import TransfoXLConfig, TransfoXLLMHeadModel, load_tf_weights_in_transfo_xl
from transformers.models.transfo_xl import tokenization_transfo_xl as data_utils
from transformers.models.transfo_xl.tokenization_transfo_xl import CORPUS_NAME, VOCAB_FILES_NAMES
from transformers.utils import CONFIG_NAME, WEIGHTS_NAME, logging
logging.set_verbosity_info()
# We do this to be able to load python 2 datasets pickles
# See e.g. https://stackoverflow.com/questions/2121874/python-pickling-after-changing-a-modules-directory/2121918#2121918
_lowerCAmelCase = data_utils.TransfoXLTokenizer
_lowerCAmelCase = data_utils.TransfoXLCorpus
_lowerCAmelCase = data_utils
_lowerCAmelCase = data_utils
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
if transfo_xl_dataset_file:
# Convert a pre-processed corpus (see original TensorFlow repo)
with open(UpperCamelCase , """rb""" ) as fp:
lowerCAmelCase__ : Union[str, Any] = pickle.load(UpperCamelCase , encoding="""latin1""" )
# Save vocabulary and dataset cache as Dictionaries (should be better than pickles for the long-term)
lowerCAmelCase__ : List[str] = pytorch_dump_folder_path + """/""" + VOCAB_FILES_NAMES["""pretrained_vocab_file"""]
print(f"""Save vocabulary to {pytorch_vocab_dump_path}""" )
lowerCAmelCase__ : Dict = corpus.vocab.__dict__
torch.save(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Dict = corpus.__dict__
corpus_dict_no_vocab.pop("""vocab""" , UpperCamelCase )
lowerCAmelCase__ : List[Any] = pytorch_dump_folder_path + """/""" + CORPUS_NAME
print(f"""Save dataset to {pytorch_dataset_dump_path}""" )
torch.save(UpperCamelCase , UpperCamelCase )
if tf_checkpoint_path:
# Convert a pre-trained TensorFlow model
lowerCAmelCase__ : str = os.path.abspath(UpperCamelCase )
lowerCAmelCase__ : List[str] = os.path.abspath(UpperCamelCase )
print(f"""Converting Transformer XL checkpoint from {tf_path} with config at {config_path}.""" )
# Initialise PyTorch model
if transfo_xl_config_file == "":
lowerCAmelCase__ : Optional[Any] = TransfoXLConfig()
else:
lowerCAmelCase__ : Optional[int] = TransfoXLConfig.from_json_file(UpperCamelCase )
print(f"""Building PyTorch model from configuration: {config}""" )
lowerCAmelCase__ : str = TransfoXLLMHeadModel(UpperCamelCase )
lowerCAmelCase__ : Tuple = load_tf_weights_in_transfo_xl(UpperCamelCase , UpperCamelCase , UpperCamelCase )
# Save pytorch-model
lowerCAmelCase__ : Optional[int] = os.path.join(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Dict = os.path.join(UpperCamelCase , UpperCamelCase )
print(f"""Save PyTorch model to {os.path.abspath(UpperCamelCase )}""" )
torch.save(model.state_dict() , UpperCamelCase )
print(f"""Save configuration file to {os.path.abspath(UpperCamelCase )}""" )
with open(UpperCamelCase , """w""" , encoding="""utf-8""" ) as f:
f.write(config.to_json_string() )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--pytorch_dump_folder_path''',
default=None,
type=str,
required=True,
help='''Path to the folder to store the PyTorch model or dataset/vocab.''',
)
parser.add_argument(
'''--tf_checkpoint_path''',
default='''''',
type=str,
help='''An optional path to a TensorFlow checkpoint path to be converted.''',
)
parser.add_argument(
'''--transfo_xl_config_file''',
default='''''',
type=str,
help=(
'''An optional config json file corresponding to the pre-trained BERT model. \n'''
'''This specifies the model architecture.'''
),
)
parser.add_argument(
'''--transfo_xl_dataset_file''',
default='''''',
type=str,
help='''An optional dataset file to be converted in a vocabulary.''',
)
_lowerCAmelCase = parser.parse_args()
convert_transfo_xl_checkpoint_to_pytorch(
args.tf_checkpoint_path,
args.transfo_xl_config_file,
args.pytorch_dump_folder_path,
args.transfo_xl_dataset_file,
)
| 37
|
'''simple docstring'''
import argparse
import json
import torch
from diffusers import DDPMScheduler, LDMPipeline, UNetaDModel, VQModel
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=1 ):
"""simple docstring"""
if n_shave_prefix_segments >= 0:
return ".".join(path.split(""".""" )[n_shave_prefix_segments:] )
else:
return ".".join(path.split(""".""" )[:n_shave_prefix_segments] )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = []
for old_item in old_list:
lowerCAmelCase__ : Optional[Any] = old_item.replace("""in_layers.0""" , """norm1""" )
lowerCAmelCase__ : Optional[int] = new_item.replace("""in_layers.2""" , """conv1""" )
lowerCAmelCase__ : Dict = new_item.replace("""out_layers.0""" , """norm2""" )
lowerCAmelCase__ : str = new_item.replace("""out_layers.3""" , """conv2""" )
lowerCAmelCase__ : str = new_item.replace("""emb_layers.1""" , """time_emb_proj""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""skip_connection""" , """conv_shortcut""" )
lowerCAmelCase__ : Union[str, Any] = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : int = []
for old_item in old_list:
lowerCAmelCase__ : List[str] = old_item
lowerCAmelCase__ : int = new_item.replace("""norm.weight""" , """group_norm.weight""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""norm.bias""" , """group_norm.bias""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""proj_out.weight""" , """proj_attn.weight""" )
lowerCAmelCase__ : int = new_item.replace("""proj_out.bias""" , """proj_attn.bias""" )
lowerCAmelCase__ : str = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=None ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "Paths should be a list of dicts containing 'old' and 'new' keys."
# Splits the attention layers into three variables.
if attention_paths_to_split is not None:
for path, path_map in attention_paths_to_split.items():
lowerCAmelCase__ : Any = old_checkpoint[path]
lowerCAmelCase__ : int = old_tensor.shape[0] // 3
lowerCAmelCase__ : int = (-1, channels) if len(old_tensor.shape ) == 3 else (-1)
lowerCAmelCase__ : Tuple = old_tensor.shape[0] // config["""num_head_channels"""] // 3
lowerCAmelCase__ : List[Any] = old_tensor.reshape((num_heads, 3 * channels // num_heads) + old_tensor.shape[1:] )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : str = old_tensor.split(channels // num_heads , dim=1 )
lowerCAmelCase__ : int = query.reshape(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = key.reshape(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = value.reshape(UpperCamelCase )
for path in paths:
lowerCAmelCase__ : Any = path["""new"""]
# These have already been assigned
if attention_paths_to_split is not None and new_path in attention_paths_to_split:
continue
# Global renaming happens here
lowerCAmelCase__ : Any = new_path.replace("""middle_block.0""" , """mid_block.resnets.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.1""" , """mid_block.attentions.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.2""" , """mid_block.resnets.1""" )
if additional_replacements is not None:
for replacement in additional_replacements:
lowerCAmelCase__ : Any = new_path.replace(replacement["""old"""] , replacement["""new"""] )
# proj_attn.weight has to be converted from conv 1D to linear
if "proj_attn.weight" in new_path:
lowerCAmelCase__ : List[Any] = old_checkpoint[path["""old"""]][:, :, 0]
else:
lowerCAmelCase__ : Dict = old_checkpoint[path["""old"""]]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = {}
lowerCAmelCase__ : str = checkpoint["""time_embed.0.weight"""]
lowerCAmelCase__ : List[Any] = checkpoint["""time_embed.0.bias"""]
lowerCAmelCase__ : int = checkpoint["""time_embed.2.weight"""]
lowerCAmelCase__ : List[str] = checkpoint["""time_embed.2.bias"""]
lowerCAmelCase__ : str = checkpoint["""input_blocks.0.0.weight"""]
lowerCAmelCase__ : Any = checkpoint["""input_blocks.0.0.bias"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.weight"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.bias"""]
lowerCAmelCase__ : str = checkpoint["""out.2.weight"""]
lowerCAmelCase__ : Tuple = checkpoint["""out.2.bias"""]
# Retrieves the keys for the input blocks only
lowerCAmelCase__ : Optional[Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """input_blocks""" in layer} )
lowerCAmelCase__ : Optional[Any] = {
layer_id: [key for key in checkpoint if f"""input_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the middle blocks only
lowerCAmelCase__ : Union[str, Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """middle_block""" in layer} )
lowerCAmelCase__ : Union[str, Any] = {
layer_id: [key for key in checkpoint if f"""middle_block.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the output blocks only
lowerCAmelCase__ : List[str] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """output_blocks""" in layer} )
lowerCAmelCase__ : List[Any] = {
layer_id: [key for key in checkpoint if f"""output_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
for i in range(1 , UpperCamelCase ):
lowerCAmelCase__ : Dict = (i - 1) // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Tuple = (i - 1) % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Optional[int] = [key for key in input_blocks[i] if f"""input_blocks.{i}.0""" in key]
lowerCAmelCase__ : Optional[Any] = [key for key in input_blocks[i] if f"""input_blocks.{i}.1""" in key]
if f"""input_blocks.{i}.0.op.weight""" in checkpoint:
lowerCAmelCase__ : Optional[int] = checkpoint[
f"""input_blocks.{i}.0.op.weight"""
]
lowerCAmelCase__ : Tuple = checkpoint[
f"""input_blocks.{i}.0.op.bias"""
]
continue
lowerCAmelCase__ : Optional[Any] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Dict = {"""old""": f"""input_blocks.{i}.0""", """new""": f"""down_blocks.{block_id}.resnets.{layer_in_block_id}"""}
lowerCAmelCase__ : Optional[Any] = {"""old""": """resnets.2.op""", """new""": """downsamplers.0.op"""}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path, resnet_op] , config=UpperCamelCase )
if len(UpperCamelCase ):
lowerCAmelCase__ : Optional[Any] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""input_blocks.{i}.1""",
"""new""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : List[str] = {
f"""input_blocks.{i}.1.qkv.bias""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""input_blocks.{i}.1.qkv.weight""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase , )
lowerCAmelCase__ : Dict = middle_blocks[0]
lowerCAmelCase__ : Union[str, Any] = middle_blocks[1]
lowerCAmelCase__ : Dict = middle_blocks[2]
lowerCAmelCase__ : Any = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Dict = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {
"""middle_block.1.qkv.bias""": {
"""key""": """mid_block.attentions.0.key.bias""",
"""query""": """mid_block.attentions.0.query.bias""",
"""value""": """mid_block.attentions.0.value.bias""",
},
"""middle_block.1.qkv.weight""": {
"""key""": """mid_block.attentions.0.key.weight""",
"""query""": """mid_block.attentions.0.query.weight""",
"""value""": """mid_block.attentions.0.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase )
for i in range(UpperCamelCase ):
lowerCAmelCase__ : Tuple = i // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : List[str] = i % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : int = [shave_segments(UpperCamelCase , 2 ) for name in output_blocks[i]]
lowerCAmelCase__ : Union[str, Any] = {}
for layer in output_block_layers:
lowerCAmelCase__ , lowerCAmelCase__ : Any = layer.split(""".""" )[0], shave_segments(UpperCamelCase , 1 )
if layer_id in output_block_list:
output_block_list[layer_id].append(UpperCamelCase )
else:
lowerCAmelCase__ : str = [layer_name]
if len(UpperCamelCase ) > 1:
lowerCAmelCase__ : str = [key for key in output_blocks[i] if f"""output_blocks.{i}.0""" in key]
lowerCAmelCase__ : Dict = [key for key in output_blocks[i] if f"""output_blocks.{i}.1""" in key]
lowerCAmelCase__ : Optional[int] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {"""old""": f"""output_blocks.{i}.0""", """new""": f"""up_blocks.{block_id}.resnets.{layer_in_block_id}"""}
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , config=UpperCamelCase )
if ["conv.weight", "conv.bias"] in output_block_list.values():
lowerCAmelCase__ : List[Any] = list(output_block_list.values() ).index(["""conv.weight""", """conv.bias"""] )
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.weight"""
]
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.bias"""
]
# Clear attentions as they have been attributed above.
if len(UpperCamelCase ) == 2:
lowerCAmelCase__ : Tuple = []
if len(UpperCamelCase ):
lowerCAmelCase__ : Dict = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""output_blocks.{i}.1""",
"""new""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : Tuple = {
f"""output_blocks.{i}.1.qkv.bias""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""output_blocks.{i}.1.qkv.weight""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=to_split if any("""qkv""" in key for key in attentions ) else None , config=UpperCamelCase , )
else:
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase , n_shave_prefix_segments=1 )
for path in resnet_0_paths:
lowerCAmelCase__ : Tuple = """.""".join(["""output_blocks""", str(UpperCamelCase ), path["""old"""]] )
lowerCAmelCase__ : List[Any] = """.""".join(["""up_blocks""", str(UpperCamelCase ), """resnets""", str(UpperCamelCase ), path["""new"""]] )
lowerCAmelCase__ : Union[str, Any] = checkpoint[old_path]
return new_checkpoint
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the checkpoint to convert.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help='''The config json file corresponding to the architecture.''',
)
parser.add_argument('''--dump_path''', default=None, type=str, required=True, help='''Path to the output model.''')
_lowerCAmelCase = parser.parse_args()
_lowerCAmelCase = torch.load(args.checkpoint_path)
with open(args.config_file) as f:
_lowerCAmelCase = json.loads(f.read())
_lowerCAmelCase = convert_ldm_checkpoint(checkpoint, config)
if "ldm" in config:
del config["ldm"]
_lowerCAmelCase = UNetaDModel(**config)
model.load_state_dict(converted_checkpoint)
try:
_lowerCAmelCase = DDPMScheduler.from_config('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = VQModel.from_pretrained('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = LDMPipeline(unet=model, scheduler=scheduler, vae=vqvae)
pipe.save_pretrained(args.dump_path)
except: # noqa: E722
model.save_pretrained(args.dump_path)
| 37
| 1
|
'''simple docstring'''
import argparse
import torch
from transformers import (
SpeechTaConfig,
SpeechTaFeatureExtractor,
SpeechTaForSpeechToSpeech,
SpeechTaForSpeechToText,
SpeechTaForTextToSpeech,
SpeechTaProcessor,
SpeechTaTokenizer,
logging,
)
from transformers.tokenization_utils import AddedToken
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger('''transformers.models.speecht5''')
_lowerCAmelCase = {
'''speech_encoder_prenet.layer_norm''': '''speecht5.encoder.prenet.feature_projection.layer_norm''',
'''speech_encoder_prenet.post_extract_proj''': '''speecht5.encoder.prenet.feature_projection.projection''',
'''speech_encoder_prenet.pos_conv.0''': '''speecht5.encoder.prenet.pos_conv_embed.conv''',
'''speech_encoder_prenet.mask_emb''': '''speecht5.encoder.prenet.masked_spec_embed''',
}
_lowerCAmelCase = {
'''text_encoder_prenet.encoder_prenet.0''': '''speecht5.encoder.prenet.embed_tokens''',
'''text_encoder_prenet.encoder_prenet.1.alpha''': '''speecht5.encoder.prenet.encode_positions.alpha''',
}
_lowerCAmelCase = {
'''speech_decoder_prenet.decoder_prenet.0.0.prenet.0.0''': '''speecht5.decoder.prenet.layers.0''',
'''speech_decoder_prenet.decoder_prenet.0.0.prenet.1.0''': '''speecht5.decoder.prenet.layers.1''',
'''speech_decoder_prenet.decoder_prenet.0.1''': '''speecht5.decoder.prenet.final_layer''',
'''speech_decoder_prenet.decoder_prenet.1.alpha''': '''speecht5.decoder.prenet.encode_positions.alpha''',
'''speech_decoder_prenet.spkembs_layer.0''': '''speecht5.decoder.prenet.speaker_embeds_layer''',
}
_lowerCAmelCase = {
'''speech_decoder_postnet.feat_out''': '''speech_decoder_postnet.feat_out''',
'''speech_decoder_postnet.prob_out''': '''speech_decoder_postnet.prob_out''',
'''speech_decoder_postnet.postnet.postnet.0.0''': '''speech_decoder_postnet.layers.0.conv''',
'''speech_decoder_postnet.postnet.postnet.0.1''': '''speech_decoder_postnet.layers.0.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.1.0''': '''speech_decoder_postnet.layers.1.conv''',
'''speech_decoder_postnet.postnet.postnet.1.1''': '''speech_decoder_postnet.layers.1.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.2.0''': '''speech_decoder_postnet.layers.2.conv''',
'''speech_decoder_postnet.postnet.postnet.2.1''': '''speech_decoder_postnet.layers.2.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.3.0''': '''speech_decoder_postnet.layers.3.conv''',
'''speech_decoder_postnet.postnet.postnet.3.1''': '''speech_decoder_postnet.layers.3.batch_norm''',
'''speech_decoder_postnet.postnet.postnet.4.0''': '''speech_decoder_postnet.layers.4.conv''',
'''speech_decoder_postnet.postnet.postnet.4.1''': '''speech_decoder_postnet.layers.4.batch_norm''',
}
_lowerCAmelCase = {
'''text_decoder_prenet.embed_tokens''': '''speecht5.decoder.prenet.embed_tokens''',
}
_lowerCAmelCase = {
'''text_decoder_postnet.output_projection''': '''text_decoder_postnet.lm_head''',
}
_lowerCAmelCase = {
'''encoder.layers.*.self_attn.k_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.k_proj''',
'''encoder.layers.*.self_attn.v_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.v_proj''',
'''encoder.layers.*.self_attn.q_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.q_proj''',
'''encoder.layers.*.self_attn.out_proj''': '''speecht5.encoder.wrapped_encoder.layers.*.attention.out_proj''',
'''encoder.layers.*.self_attn_layer_norm''': '''speecht5.encoder.wrapped_encoder.layers.*.layer_norm''',
'''encoder.layers.*.fc1''': '''speecht5.encoder.wrapped_encoder.layers.*.feed_forward.intermediate_dense''',
'''encoder.layers.*.fc2''': '''speecht5.encoder.wrapped_encoder.layers.*.feed_forward.output_dense''',
'''encoder.layers.*.final_layer_norm''': '''speecht5.encoder.wrapped_encoder.layers.*.final_layer_norm''',
'''encoder.layer_norm''': '''speecht5.encoder.wrapped_encoder.layer_norm''',
'''encoder.pos_emb.pe_k''': '''speecht5.encoder.wrapped_encoder.embed_positions.pe_k''',
}
_lowerCAmelCase = {
'''decoder.layers.*.self_attn.k_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.k_proj''',
'''decoder.layers.*.self_attn.v_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.v_proj''',
'''decoder.layers.*.self_attn.q_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.q_proj''',
'''decoder.layers.*.self_attn.out_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn.out_proj''',
'''decoder.layers.*.self_attn_layer_norm''': '''speecht5.decoder.wrapped_decoder.layers.*.self_attn_layer_norm''',
'''decoder.layers.*.encoder_attn.k_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.k_proj''',
'''decoder.layers.*.encoder_attn.v_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.v_proj''',
'''decoder.layers.*.encoder_attn.q_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.q_proj''',
'''decoder.layers.*.encoder_attn.out_proj''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn.out_proj''',
'''decoder.layers.*.encoder_attn_layer_norm''': '''speecht5.decoder.wrapped_decoder.layers.*.encoder_attn_layer_norm''',
'''decoder.layers.*.fc1''': '''speecht5.decoder.wrapped_decoder.layers.*.feed_forward.intermediate_dense''',
'''decoder.layers.*.fc2''': '''speecht5.decoder.wrapped_decoder.layers.*.feed_forward.output_dense''',
'''decoder.layers.*.final_layer_norm''': '''speecht5.decoder.wrapped_decoder.layers.*.final_layer_norm''',
}
_lowerCAmelCase = {
**MAPPING_SPEECH_ENCODER_PRENET,
**MAPPING_ENCODER,
**MAPPING_DECODER,
**MAPPING_TEXT_DECODER_PRENET,
**MAPPING_TEXT_DECODER_POSTNET,
}
_lowerCAmelCase = {
**MAPPING_TEXT_ENCODER_PRENET,
**MAPPING_ENCODER,
**MAPPING_DECODER,
**MAPPING_SPEECH_DECODER_PRENET,
**MAPPING_SPEECH_DECODER_POSTNET,
}
_lowerCAmelCase = {
**MAPPING_SPEECH_ENCODER_PRENET,
**MAPPING_ENCODER,
**MAPPING_DECODER,
**MAPPING_SPEECH_DECODER_PRENET,
**MAPPING_SPEECH_DECODER_POSTNET,
}
_lowerCAmelCase = []
_lowerCAmelCase = [
'''encoder.version''',
'''encoder.layers.*.norm_k.weight''',
'''encoder.layers.*.norm_k.bias''',
'''decoder.version''',
'''decoder.layers.*.norm_k.weight''',
'''decoder.layers.*.norm_k.bias''',
'''decoder.pos_emb.pe_k''',
'''speech_encoder_prenet.embed_positions._float_tensor''',
'''text_decoder_prenet.embed_positions._float_tensor''',
]
_lowerCAmelCase = IGNORE_KEYS + [
'''encoder.proj''',
'''text_encoder_prenet.*''',
'''speech_decoder_prenet.*''',
'''speech_decoder_postnet.*''',
]
_lowerCAmelCase = IGNORE_KEYS + [
'''encoder.proj''',
'''speech_encoder_prenet.*''',
'''text_decoder_prenet.*''',
'''text_decoder_postnet.*''',
]
_lowerCAmelCase = IGNORE_KEYS + [
'''encoder.proj''',
'''text_encoder_prenet.*''',
'''text_decoder_prenet.*''',
'''text_decoder_postnet.*''',
]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for attribute in key.split(""".""" ):
lowerCAmelCase__ : Optional[Any] = getattr(UpperCamelCase , UpperCamelCase )
if weight_type is not None:
lowerCAmelCase__ : Optional[int] = getattr(UpperCamelCase , UpperCamelCase ).shape
else:
lowerCAmelCase__ : Optional[int] = 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":
lowerCAmelCase__ : int = value
elif weight_type == "weight_g":
lowerCAmelCase__ : Optional[int] = value
elif weight_type == "weight_v":
lowerCAmelCase__ : Union[str, Any] = value
elif weight_type == "bias":
lowerCAmelCase__ : str = value
elif weight_type == "running_mean":
lowerCAmelCase__ : str = value
elif weight_type == "running_var":
lowerCAmelCase__ : int = value
elif weight_type == "num_batches_tracked":
lowerCAmelCase__ : Dict = value
else:
lowerCAmelCase__ : str = value
logger.info(f"""{key + ('.' + weight_type if weight_type is not None else '')} was initialized from {full_name}.""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for key in ignore_keys:
if key.endswith(""".*""" ):
if name.startswith(key[:-1] ):
return True
elif ".*." in key:
lowerCAmelCase__ , lowerCAmelCase__ : Any = key.split(""".*.""" )
if prefix in name and suffix in name:
return True
elif key in name:
return True
return False
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Dict = []
if task == "s2t":
lowerCAmelCase__ : int = hf_model.speechta.encoder.prenet.feature_encoder
lowerCAmelCase__ : List[str] = MAPPING_S2T
lowerCAmelCase__ : List[str] = IGNORE_KEYS_S2T
elif task == "t2s":
lowerCAmelCase__ : Optional[Any] = None
lowerCAmelCase__ : Optional[int] = MAPPING_T2S
lowerCAmelCase__ : str = IGNORE_KEYS_T2S
elif task == "s2s":
lowerCAmelCase__ : int = hf_model.speechta.encoder.prenet.feature_encoder
lowerCAmelCase__ : Optional[Any] = MAPPING_S2S
lowerCAmelCase__ : Optional[int] = IGNORE_KEYS_S2S
else:
raise ValueError(f"""Unsupported task: {task}""" )
for name, value in fairseq_dict.items():
if should_ignore(UpperCamelCase , UpperCamelCase ):
logger.info(f"""{name} was ignored""" )
continue
lowerCAmelCase__ : int = False
if "conv_layers" in name:
load_conv_layer(
UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , hf_model.config.feat_extract_norm == """group""" , )
lowerCAmelCase__ : str = True
else:
for key, mapped_key in MAPPING.items():
# mapped_key = "speecht5." + mapped_key if mapped_key not in TOP_LEVEL_KEYS else mapped_key
if "*" in key:
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = key.split(""".*.""" )
if prefix in name and suffix in name:
lowerCAmelCase__ : Any = suffix
# if key in name or key.split("w2v_model.")[-1] == name.split(".")[0]:
if key in name:
lowerCAmelCase__ : List[str] = True
if "*" in mapped_key:
lowerCAmelCase__ : int = name.split(UpperCamelCase )[0].split(""".""" )[-2]
lowerCAmelCase__ : Dict = mapped_key.replace("""*""" , UpperCamelCase )
if "weight_g" in name:
lowerCAmelCase__ : int = """weight_g"""
elif "weight_v" in name:
lowerCAmelCase__ : Any = """weight_v"""
elif "bias" in name:
lowerCAmelCase__ : List[Any] = """bias"""
elif "weight" in name:
lowerCAmelCase__ : int = """weight"""
elif "running_mean" in name:
lowerCAmelCase__ : List[Any] = """running_mean"""
elif "running_var" in name:
lowerCAmelCase__ : Any = """running_var"""
elif "num_batches_tracked" in name:
lowerCAmelCase__ : Optional[Any] = """num_batches_tracked"""
else:
lowerCAmelCase__ : Optional[Any] = None
set_recursively(UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase )
continue
if not is_used:
unused_weights.append(UpperCamelCase )
logger.warning(f"""Unused weights: {unused_weights}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = full_name.split("""conv_layers.""" )[-1]
lowerCAmelCase__ : Dict = name.split(""".""" )
lowerCAmelCase__ : str = int(items[0] )
lowerCAmelCase__ : str = 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.""" )
lowerCAmelCase__ : Union[str, Any] = 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.""" )
lowerCAmelCase__ : Any = 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.conv_layers[layer_id].layer_norm.bias.data.shape} was found.""" )
lowerCAmelCase__ : List[str] = value
logger.info(f"""Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.""" )
elif "weight" in name:
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.conv_layers[layer_id].layer_norm.weight.data.shape} was found.""" )
lowerCAmelCase__ : Any = value
logger.info(f"""Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.""" )
else:
unused_weights.append(UpperCamelCase )
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=None , ):
"""simple docstring"""
if config_path is not None:
lowerCAmelCase__ : str = SpeechTaConfig.from_pretrained(UpperCamelCase )
else:
lowerCAmelCase__ : Dict = SpeechTaConfig()
if task == "s2t":
lowerCAmelCase__ : Optional[int] = config.max_text_positions
lowerCAmelCase__ : int = SpeechTaForSpeechToText(UpperCamelCase )
elif task == "t2s":
lowerCAmelCase__ : Tuple = 1876
lowerCAmelCase__ : int = 600
lowerCAmelCase__ : List[Any] = config.max_speech_positions
lowerCAmelCase__ : Dict = SpeechTaForTextToSpeech(UpperCamelCase )
elif task == "s2s":
lowerCAmelCase__ : List[Any] = 1876
lowerCAmelCase__ : List[Any] = config.max_speech_positions
lowerCAmelCase__ : Dict = SpeechTaForSpeechToSpeech(UpperCamelCase )
else:
raise ValueError(f"""Unknown task name: {task}""" )
if vocab_path:
lowerCAmelCase__ : Dict = SpeechTaTokenizer(UpperCamelCase , model_max_length=config.max_text_positions )
# Mask token behaves like a normal word, i.e. include the space before it
lowerCAmelCase__ : List[str] = AddedToken("""<mask>""" , lstrip=UpperCamelCase , rstrip=UpperCamelCase )
lowerCAmelCase__ : Tuple = mask_token
tokenizer.add_special_tokens({"""mask_token""": mask_token} )
tokenizer.add_tokens(["""<ctc_blank>"""] )
lowerCAmelCase__ : str = SpeechTaFeatureExtractor()
lowerCAmelCase__ : int = SpeechTaProcessor(tokenizer=UpperCamelCase , feature_extractor=UpperCamelCase )
processor.save_pretrained(UpperCamelCase )
lowerCAmelCase__ : Dict = torch.load(UpperCamelCase )
recursively_load_weights(fairseq_checkpoint["""model"""] , UpperCamelCase , UpperCamelCase )
model.save_pretrained(UpperCamelCase )
if repo_id:
print("""Pushing to the hub...""" )
processor.push_to_hub(UpperCamelCase )
model.push_to_hub(UpperCamelCase )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--task''',
default='''s2t''',
type=str,
help='''Type of the SpeechT5 model you\'d like to convert. Should be one of \'s2t\', \'t2s\', \'s2s\'.''',
)
parser.add_argument('''--checkpoint_path''', required=True, default=None, type=str, help='''Path to fairseq checkpoint''')
parser.add_argument('''--vocab_path''', default=None, type=str, help='''Path to SentencePiece model''')
parser.add_argument('''--config_path''', default=None, type=str, help='''Path to hf config.json of model to convert''')
parser.add_argument(
'''--pytorch_dump_folder_path''', required=True, default=None, type=str, help='''Path to the output PyTorch model.'''
)
parser.add_argument(
'''--push_to_hub''', default=None, type=str, help='''Where to upload the converted model on the 🤗 hub.'''
)
_lowerCAmelCase = parser.parse_args()
convert_speechta_checkpoint(
args.task,
args.checkpoint_path,
args.pytorch_dump_folder_path,
args.config_path,
args.vocab_path,
args.push_to_hub,
)
| 37
|
'''simple docstring'''
from math import sqrt
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' must been an int and positive"
lowerCAmelCase__ : int = True
# 0 and 1 are none primes.
if number <= 1:
lowerCAmelCase__ : Optional[Any] = False
for divisor in range(2 , int(round(sqrt(UpperCamelCase ) ) ) + 1 ):
# if 'number' divisible by 'divisor' then sets 'status'
# of false and break up the loop.
if number % divisor == 0:
lowerCAmelCase__ : Any = False
break
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'status' must been from type bool"
return status
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
# beginList: contains all natural numbers from 2 up to N
lowerCAmelCase__ : List[str] = list(range(2 , n + 1 ) )
lowerCAmelCase__ : str = [] # this list will be returns.
# actual sieve of erathostenes
for i in range(len(UpperCamelCase ) ):
for j in range(i + 1 , len(UpperCamelCase ) ):
if (begin_list[i] != 0) and (begin_list[j] % begin_list[i] == 0):
lowerCAmelCase__ : List[Any] = 0
# filters actual prime numbers.
lowerCAmelCase__ : List[Any] = [x for x in begin_list if x != 0]
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
lowerCAmelCase__ : List[str] = []
# iterates over all numbers between 2 up to N+1
# if a number is prime then appends to list 'ans'
for number in range(2 , n + 1 ):
if is_prime(UpperCamelCase ):
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and number >= 0, "'number' must been an int and >= 0"
lowerCAmelCase__ : Optional[Any] = [] # this list will be returns of the function.
# potential prime number factors.
lowerCAmelCase__ : Dict = 2
lowerCAmelCase__ : Dict = number
if number == 0 or number == 1:
ans.append(UpperCamelCase )
# if 'number' not prime then builds the prime factorization of 'number'
elif not is_prime(UpperCamelCase ):
while quotient != 1:
if is_prime(UpperCamelCase ) and (quotient % factor == 0):
ans.append(UpperCamelCase )
quotient /= factor
else:
factor += 1
else:
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : Optional[int] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = max(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = min(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 == 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 == 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 != 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 != 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase ) and (number > 2) and is_even(UpperCamelCase )
), "'number' must been an int, even and > 2"
lowerCAmelCase__ : Dict = [] # this list will returned
# creates a list of prime numbers between 2 up to 'number'
lowerCAmelCase__ : Dict = get_prime_numbers(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = len(UpperCamelCase )
# run variable for while-loops.
lowerCAmelCase__ : List[str] = 0
lowerCAmelCase__ : List[Any] = None
# exit variable. for break up the loops
lowerCAmelCase__ : Any = True
while i < len_pn and loop:
lowerCAmelCase__ : List[Any] = i + 1
while j < len_pn and loop:
if prime_numbers[i] + prime_numbers[j] == number:
lowerCAmelCase__ : Optional[Any] = False
ans.append(prime_numbers[i] )
ans.append(prime_numbers[j] )
j += 1
i += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (len(UpperCamelCase ) == 2)
and (ans[0] + ans[1] == number)
and is_prime(ans[0] )
and is_prime(ans[1] )
), "'ans' must contains two primes. And sum of elements must been eq 'number'"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 0)
and (numbera >= 0)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 0
while numbera != 0:
lowerCAmelCase__ : Any = numbera % numbera
lowerCAmelCase__ : str = numbera
lowerCAmelCase__ : List[str] = rest
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
numbera >= 0
), "'number' must been from type int and positive"
return numbera
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 1)
and (numbera >= 1)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 1 # actual answer that will be return.
# for kgV (x,1)
if numbera > 1 and numbera > 1:
# builds the prime factorization of 'number1' and 'number2'
lowerCAmelCase__ : int = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = prime_factorization(UpperCamelCase )
elif numbera == 1 or numbera == 1:
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : Dict = []
lowerCAmelCase__ : List[str] = max(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = 0
lowerCAmelCase__ : str = 0
lowerCAmelCase__ : List[Any] = [] # captured numbers int both 'primeFac1' and 'primeFac2'
# iterates through primeFac1
for n in prime_fac_a:
if n not in done:
if n in prime_fac_a:
lowerCAmelCase__ : int = prime_fac_a.count(UpperCamelCase )
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(max(UpperCamelCase , UpperCamelCase ) ):
ans *= n
else:
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# iterates through primeFac2
for n in prime_fac_a:
if n not in done:
lowerCAmelCase__ : Optional[int] = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
ans >= 0
), "'ans' must been from type int and positive"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'number' must been a positive int"
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Tuple = 2 # this variable holds the answer
while index < n:
index += 1
ans += 1 # counts to the next number
# if ans not prime then
# runs to the next prime number.
while not is_prime(UpperCamelCase ):
ans += 1
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and is_prime(
UpperCamelCase ), "'ans' must been a prime number and from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
is_prime(UpperCamelCase ) and is_prime(UpperCamelCase ) and (p_number_a < p_number_a)
), "The arguments must been prime numbers and 'pNumber1' < 'pNumber2'"
lowerCAmelCase__ : Dict = p_number_a + 1 # jump to the next number
lowerCAmelCase__ : List[Any] = [] # this list will be returns.
# if number is not prime then
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
while number < p_number_a:
ans.append(UpperCamelCase )
number += 1
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and ans[0] != p_number_a
and ans[len(UpperCamelCase ) - 1] != p_number_a
), "'ans' must been a list without the arguments"
# 'ans' contains not 'pNumber1' and 'pNumber2' !
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 1), "'n' must been int and >= 1"
lowerCAmelCase__ : List[Any] = [] # will be returned.
for divisor in range(1 , n + 1 ):
if n % divisor == 0:
ans.append(UpperCamelCase )
# precondition
assert ans[0] == 1 and ans[len(UpperCamelCase ) - 1] == n, "Error in function getDivisiors(...)"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number > 1
), "'number' must been an int and >= 1"
lowerCAmelCase__ : Optional[int] = get_divisors(UpperCamelCase )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (divisors[0] == 1)
and (divisors[len(UpperCamelCase ) - 1] == number)
), "Error in help-function getDivisiors(...)"
# summed all divisors up to 'number' (exclusive), hence [:-1]
return sum(divisors[:-1] ) == number
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (denominator != 0)
), "The arguments must been from type int and 'denominator' != 0"
# build the greatest common divisor of numerator and denominator.
lowerCAmelCase__ : int = gcd(abs(UpperCamelCase ) , abs(UpperCamelCase ) )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (numerator % gcd_of_fraction == 0)
and (denominator % gcd_of_fraction == 0)
), "Error in function gcd(...,...)"
return (numerator // gcd_of_fraction, denominator // gcd_of_fraction)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been a int and >= 0"
lowerCAmelCase__ : str = 1 # this will be return.
for factor in range(1 , n + 1 ):
ans *= factor
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
lowerCAmelCase__ : Any = 1
lowerCAmelCase__ : Optional[Any] = 1 # this will be return
for _ in range(n - 1 ):
lowerCAmelCase__ : Dict = ans
ans += fiba
lowerCAmelCase__ : str = tmp
return ans
| 37
| 1
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''google/fnet-base''': '''https://huggingface.co/google/fnet-base/resolve/main/config.json''',
'''google/fnet-large''': '''https://huggingface.co/google/fnet-large/resolve/main/config.json'''
# See all FNet models at https://huggingface.co/models?filter=fnet
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[Any] = '''fnet'''
def __init__( self ,__UpperCAmelCase=3_2000 ,__UpperCAmelCase=768 ,__UpperCAmelCase=12 ,__UpperCAmelCase=3072 ,__UpperCAmelCase="gelu_new" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=512 ,__UpperCAmelCase=4 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-12 ,__UpperCAmelCase=False ,__UpperCAmelCase=512 ,__UpperCAmelCase=3 ,__UpperCAmelCase=1 ,__UpperCAmelCase=2 ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = vocab_size
lowerCAmelCase__ : Union[str, Any] = max_position_embeddings
lowerCAmelCase__ : Tuple = hidden_size
lowerCAmelCase__ : List[Any] = num_hidden_layers
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : str = hidden_act
lowerCAmelCase__ : str = hidden_dropout_prob
lowerCAmelCase__ : List[str] = initializer_range
lowerCAmelCase__ : str = type_vocab_size
lowerCAmelCase__ : Optional[int] = layer_norm_eps
lowerCAmelCase__ : int = use_tpu_fourier_optimizations
lowerCAmelCase__ : int = tpu_short_seq_length
| 37
|
'''simple docstring'''
from sklearn.metrics import fa_score, matthews_corrcoef
import datasets
from .record_evaluation import evaluate as evaluate_record
_lowerCAmelCase = '''\
@article{wang2019superglue,
title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
journal={arXiv preprint arXiv:1905.00537},
year={2019}
}
'''
_lowerCAmelCase = '''\
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
GLUE with a new set of more difficult language understanding tasks, improved
resources, and a new public leaderboard.
'''
_lowerCAmelCase = '''
Compute SuperGLUE evaluation metric associated to each SuperGLUE dataset.
Args:
predictions: list of predictions to score. Depending on the SuperGlUE subset:
- for \'record\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'prediction_text\': the predicted answer text
- for \'multirc\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question-answer pair as specified by the dataset
- \'prediction\': the predicted answer label
- otherwise: list of predicted labels
references: list of reference labels. Depending on the SuperGLUE subset:
- for \'record\': list of question-answers dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'answers\': list of possible answers
- otherwise: list of reference labels
Returns: depending on the SuperGLUE subset:
- for \'record\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1\': F1 score
- for \'multirc\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1_m\': Per-question macro-F1 score
- \'f1_a\': Average F1 score over all answers
- for \'axb\':
\'matthews_correlation\': Matthew Correlation
- for \'cb\':
- \'accuracy\': Accuracy
- \'f1\': F1 score
- for all others:
- \'accuracy\': Accuracy
Examples:
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'copa\') # any of ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'cb\')
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'record\')
>>> predictions = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'prediction_text\': \'answer\'}]
>>> references = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'answers\': [\'answer\', \'another_answer\']}]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'multirc\')
>>> predictions = [{\'idx\': {\'answer\': 0, \'paragraph\': 0, \'question\': 0}, \'prediction\': 0}, {\'idx\': {\'answer\': 1, \'paragraph\': 2, \'question\': 3}, \'prediction\': 1}]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1_m\': 1.0, \'f1_a\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'axb\')
>>> references = [0, 1]
>>> predictions = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'matthews_correlation\': 1.0}
'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return float((preds == labels).mean() )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase="binary" ):
"""simple docstring"""
lowerCAmelCase__ : Any = simple_accuracy(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = float(fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average=UpperCamelCase ) )
return {
"accuracy": acc,
"f1": fa,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
for id_pred, label in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : str = f"""{id_pred['idx']['paragraph']}-{id_pred['idx']['question']}"""
lowerCAmelCase__ : Dict = id_pred["""prediction"""]
if question_id in question_map:
question_map[question_id].append((pred, label) )
else:
lowerCAmelCase__ : Optional[int] = [(pred, label)]
lowerCAmelCase__ , lowerCAmelCase__ : int = [], []
for question, preds_labels in question_map.items():
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = zip(*UpperCamelCase )
lowerCAmelCase__ : List[Any] = fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average="""macro""" )
fas.append(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = int(sum(pred == label for pred, label in preds_labels ) == len(UpperCamelCase ) )
ems.append(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = float(sum(UpperCamelCase ) / len(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = sum(UpperCamelCase ) / len(UpperCamelCase )
lowerCAmelCase__ : Dict = float(fa_score(y_true=UpperCamelCase , y_pred=[id_pred["""prediction"""] for id_pred in ids_preds] ) )
return {"exact_match": em, "f1_m": fa_m, "f1_a": fa_a}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.config_name not in [
"boolq",
"cb",
"copa",
"multirc",
"record",
"rte",
"wic",
"wsc",
"wsc.fixed",
"axb",
"axg",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(self._get_feature_types() ) ,codebase_urls=[] ,reference_urls=[] ,format="""numpy""" if not self.config_name == """record""" and not self.config_name == """multirc""" else None ,)
def UpperCAmelCase_ ( self ) -> str:
if self.config_name == "record":
return {
"predictions": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"prediction_text": datasets.Value("""string""" ),
},
"references": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"answers": datasets.Sequence(datasets.Value("""string""" ) ),
},
}
elif self.config_name == "multirc":
return {
"predictions": {
"idx": {
"answer": datasets.Value("""int64""" ),
"paragraph": datasets.Value("""int64""" ),
"question": datasets.Value("""int64""" ),
},
"prediction": datasets.Value("""int64""" ),
},
"references": datasets.Value("""int64""" ),
}
else:
return {
"predictions": datasets.Value("""int64""" ),
"references": datasets.Value("""int64""" ),
}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
if self.config_name == "axb":
return {"matthews_correlation": matthews_corrcoef(__UpperCAmelCase ,__UpperCAmelCase )}
elif self.config_name == "cb":
return acc_and_fa(__UpperCAmelCase ,__UpperCAmelCase ,fa_avg="""macro""" )
elif self.config_name == "record":
lowerCAmelCase__ : Optional[Any] = [
{
"""qas""": [
{"""id""": ref["""idx"""]["""query"""], """answers""": [{"""text""": ans} for ans in ref["""answers"""]]}
for ref in references
]
}
]
lowerCAmelCase__ : Union[str, Any] = {pred["""idx"""]["""query"""]: pred["""prediction_text"""] for pred in predictions}
return evaluate_record(__UpperCAmelCase ,__UpperCAmelCase )[0]
elif self.config_name == "multirc":
return evaluate_multirc(__UpperCAmelCase ,__UpperCAmelCase )
elif self.config_name in ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]:
return {"accuracy": simple_accuracy(__UpperCAmelCase ,__UpperCAmelCase )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
| 37
| 1
|
'''simple docstring'''
from __future__ import annotations
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = position
lowerCAmelCase__ : int = [
(y + 1, x + 2),
(y - 1, x + 2),
(y + 1, x - 2),
(y - 1, x - 2),
(y + 2, x + 1),
(y + 2, x - 1),
(y - 2, x + 1),
(y - 2, x - 1),
]
lowerCAmelCase__ : int = []
for position in positions:
lowerCAmelCase__ , lowerCAmelCase__ : Dict = position
if 0 <= y_test < n and 0 <= x_test < n:
permissible_positions.append(UpperCamelCase )
return permissible_positions
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return not any(elem == 0 for row in board for elem in row )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
if is_complete(UpperCamelCase ):
return True
for position in get_valid_pos(UpperCamelCase , len(UpperCamelCase ) ):
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = position
if board[y][x] == 0:
lowerCAmelCase__ : List[str] = curr + 1
if open_knight_tour_helper(UpperCamelCase , UpperCamelCase , curr + 1 ):
return True
lowerCAmelCase__ : Optional[int] = 0
return False
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = [[0 for i in range(UpperCamelCase )] for j in range(UpperCamelCase )]
for i in range(UpperCamelCase ):
for j in range(UpperCamelCase ):
lowerCAmelCase__ : int = 1
if open_knight_tour_helper(UpperCamelCase , (i, j) , 1 ):
return board
lowerCAmelCase__ : Any = 0
lowerCAmelCase__ : Optional[int] = f"""Open Kight Tour cannot be performed on a board of size {n}"""
raise ValueError(UpperCamelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
import unittest
from pathlib import Path
from tempfile import NamedTemporaryFile, TemporaryDirectory
from transformers import BertConfig, BertTokenizerFast, FeatureExtractionPipeline
from transformers.convert_graph_to_onnx import (
convert,
ensure_valid_input,
generate_identified_filename,
infer_shapes,
quantize,
)
from transformers.testing_utils import require_tf, require_tokenizers, require_torch, slow
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[Any]:
return None
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
return None
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
__lowercase : Dict = [
# (model_name, model_kwargs)
('''bert-base-cased''', {}),
('''gpt2''', {'''use_cache''': False}), # We don't support exporting GPT2 past keys anymore
]
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> int:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Any:
from transformers import BertModel
lowerCAmelCase__ : Optional[int] = ["""[UNK]""", """[SEP]""", """[CLS]""", """[PAD]""", """[MASK]""", """some""", """other""", """words"""]
with NamedTemporaryFile(mode="""w+t""" ) as vocab_file:
vocab_file.write("""\n""".join(__UpperCAmelCase ) )
vocab_file.flush()
lowerCAmelCase__ : Dict = BertTokenizerFast(vocab_file.name )
with TemporaryDirectory() as bert_save_dir:
lowerCAmelCase__ : Tuple = BertModel(BertConfig(vocab_size=len(__UpperCAmelCase ) ) )
model.save_pretrained(__UpperCAmelCase )
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,__UpperCAmelCase )
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> List[str]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Dict = self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = quantize(Path(__UpperCAmelCase ) )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> List[Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Any = self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = quantize(__UpperCAmelCase )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[Any]:
try:
# Compute path
with TemporaryDirectory() as tempdir:
lowerCAmelCase__ : Optional[int] = Path(__UpperCAmelCase ).joinpath("""model.onnx""" )
# Remove folder if exists
if path.parent.exists():
path.parent.rmdir()
# Export
convert(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,**__UpperCAmelCase )
return path
except Exception as e:
self.fail(__UpperCAmelCase )
@require_torch
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
from transformers import BertModel
lowerCAmelCase__ : List[Any] = BertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Union[str, Any] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""pt""" )
@require_tf
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Optional[int]:
from transformers import TFBertModel
lowerCAmelCase__ : int = TFBertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Optional[int] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""tf""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = FeatureExtractionPipeline(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = ["""input_ids""", """token_type_ids""", """attention_mask""", """output_0""", """output_1"""]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = infer_shapes(__UpperCAmelCase ,__UpperCAmelCase )
# Assert all variables are present
self.assertEqual(len(__UpperCAmelCase ) ,len(__UpperCAmelCase ) )
self.assertTrue(all(var_name in shapes for var_name in variable_names ) )
self.assertSequenceEqual(variable_names[:3] ,__UpperCAmelCase )
self.assertSequenceEqual(variable_names[3:] ,__UpperCAmelCase )
# Assert inputs are {0: batch, 1: sequence}
for var_name in ["input_ids", "token_type_ids", "attention_mask"]:
self.assertDictEqual(shapes[var_name] ,{0: """batch""", 1: """sequence"""} )
# Assert outputs are {0: batch, 1: sequence} and {0: batch}
self.assertDictEqual(shapes["""output_0"""] ,{0: """batch""", 1: """sequence"""} )
self.assertDictEqual(shapes["""output_1"""] ,{0: """batch"""} )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = ["""input_ids""", """attention_mask""", """token_type_ids"""]
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": [1, 2, 3, 4], """attention_mask""": [0, 0, 0, 0], """token_type_ids""": [1, 1, 1, 1]}
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = ensure_valid_input(FuncContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the same number of args (all are valid)
self.assertEqual(len(__UpperCAmelCase ) ,3 )
# Should have exactly the same input names
self.assertEqual(set(__UpperCAmelCase ) ,set(__UpperCAmelCase ) )
# Parameter should be reordered according to their respective place in the function:
# (input_ids, token_type_ids, attention_mask)
self.assertEqual(__UpperCAmelCase ,(tokens["""input_ids"""], tokens["""token_type_ids"""], tokens["""attention_mask"""]) )
# Generated args are interleaved with another args (for instance parameter "past" in GPT2)
lowerCAmelCase__ , lowerCAmelCase__ : int = ensure_valid_input(FuncNonContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the one arg (all before the one not provided "some_other_args")
self.assertEqual(len(__UpperCAmelCase ) ,1 )
self.assertEqual(len(__UpperCAmelCase ) ,1 )
# Should have only "input_ids"
self.assertEqual(inputs_args[0] ,tokens["""input_ids"""] )
self.assertEqual(ordered_input_names[0] ,"""input_ids""" )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Dict = generate_identified_filename(Path("""/home/something/my_fake_model.onnx""" ) ,"""-test""" )
self.assertEqual("""/home/something/my_fake_model-test.onnx""" ,generated.as_posix() )
| 37
| 1
|
'''simple docstring'''
from math import ceil
from typing import List, Optional, Union
import numpy as np
from ...audio_utils import mel_filter_bank, spectrogram, window_function
from ...feature_extraction_sequence_utils import BatchFeature, SequenceFeatureExtractor
from ...utils import TensorType, logging
_lowerCAmelCase = logging.get_logger(__name__)
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : List[str] = ['''audio_values''', '''audio_mask''']
def __init__( self ,__UpperCAmelCase=2048 ,__UpperCAmelCase=1 ,__UpperCAmelCase=[16, 16] ,__UpperCAmelCase=128 ,__UpperCAmelCase=4_4100 ,__UpperCAmelCase=86 ,__UpperCAmelCase=2048 ,__UpperCAmelCase=0.0 ,**__UpperCAmelCase ,) -> Dict:
super().__init__(
feature_size=__UpperCAmelCase ,sampling_rate=__UpperCAmelCase ,padding_value=__UpperCAmelCase ,**__UpperCAmelCase ,)
lowerCAmelCase__ : Tuple = spectrogram_length
lowerCAmelCase__ : Optional[Any] = num_channels
lowerCAmelCase__ : List[str] = patch_size
lowerCAmelCase__ : List[str] = feature_size // self.patch_size[1]
lowerCAmelCase__ : Tuple = n_fft
lowerCAmelCase__ : Optional[Any] = sampling_rate // hop_length_to_sampling_rate
lowerCAmelCase__ : Any = sampling_rate
lowerCAmelCase__ : Optional[Any] = padding_value
lowerCAmelCase__ : int = mel_filter_bank(
num_frequency_bins=1 + n_fft // 2 ,num_mel_filters=__UpperCAmelCase ,min_frequency=0.0 ,max_frequency=2_2_0_5_0.0 ,sampling_rate=__UpperCAmelCase ,norm="""slaney""" ,mel_scale="""slaney""" ,).T
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> np.ndarray:
lowerCAmelCase__ : str = spectrogram(
__UpperCAmelCase ,window_function(self.n_fft ,"""hann""" ) ,frame_length=self.n_fft ,hop_length=self.hop_length ,power=2.0 ,mel_filters=self.mel_filters.T ,log_mel="""dB""" ,db_range=8_0.0 ,)
lowerCAmelCase__ : List[Any] = log_spec[:, :-1]
lowerCAmelCase__ : Any = log_spec - 2_0.0
lowerCAmelCase__ : Dict = np.clip(log_spec / 4_0.0 ,-2.0 ,0.0 ) + 1.0
return log_spec
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = True ,__UpperCAmelCase = None ,__UpperCAmelCase = False ,__UpperCAmelCase = False ,**__UpperCAmelCase ,) -> BatchFeature:
if sampling_rate is not None:
if sampling_rate != self.sampling_rate:
raise ValueError(
"""This feature extractor is set to support sampling rate"""
F""" of {self.sampling_rate}. Please make sure that the provided `raw_speech` input was sampled"""
F""" with {self.sampling_rate} and not {sampling_rate}.""" )
else:
logger.warning(
"""It is strongly recommended to pass the `sampling_rate` argument to this function. """
"""Failing to do so can result in silent errors that might be hard to debug.""" )
lowerCAmelCase__ : Union[str, Any] = isinstance(__UpperCAmelCase ,np.ndarray ) and len(raw_speech.shape ) > 1
if is_batched_numpy and len(raw_speech.shape ) > 2:
raise ValueError(F"""Only mono-channel audio is supported for input to {self}""" )
lowerCAmelCase__ : Optional[Any] = is_batched_numpy or (
isinstance(__UpperCAmelCase ,(list, tuple) ) and (isinstance(raw_speech[0] ,(np.ndarray, tuple, list) ))
)
if is_batched:
lowerCAmelCase__ : Optional[int] = [np.asarray([speech] ,dtype=np.floataa ).T for speech in raw_speech]
elif not is_batched and not isinstance(__UpperCAmelCase ,np.ndarray ):
lowerCAmelCase__ : Dict = np.asarray(__UpperCAmelCase ,dtype=np.floataa )
elif isinstance(__UpperCAmelCase ,np.ndarray ) and raw_speech.dtype is np.dtype(np.floataa ):
lowerCAmelCase__ : List[str] = raw_speech.astype(np.floataa )
# always return batch
if not is_batched:
lowerCAmelCase__ : int = [np.asarray([raw_speech] ).T]
# Convert audio signals to log mel spectrograms, truncate by time axis
lowerCAmelCase__ : Optional[Any] = [
self._np_extract_fbank_features(waveform.squeeze() ).T[: self.spectrogram_length] for waveform in raw_speech
]
if isinstance(audio_features[0] ,__UpperCAmelCase ):
lowerCAmelCase__ : Any = [np.asarray(__UpperCAmelCase ,dtype=np.floataa ) for feature in audio_features]
# Create audio attention mask
lowerCAmelCase__ : int = max(
[ceil(feature.shape[0] / self.patch_size[0] ) * self.freq_len for feature in audio_features] ) # The maximum number of audio patches in a batch
if return_attention_mask:
lowerCAmelCase__ : Optional[int] = [
(ceil(feature.shape[0] / self.patch_size[0] ) * self.freq_len) * [1]
+ (max_patch_len - ceil(feature.shape[0] / self.patch_size[0] ) * self.freq_len) * [0]
for feature in audio_features
]
lowerCAmelCase__ : Optional[int] = np.array(__UpperCAmelCase ).astype(np.floataa )
# convert into correct format for padding
lowerCAmelCase__ : int = max_patch_len // self.freq_len * self.patch_size[0] # The maximum audio size in a batch
lowerCAmelCase__ : Optional[int] = np.ones([len(__UpperCAmelCase ), 1, max_time_len, self.feature_size] ).astype(np.floataa )
lowerCAmelCase__ : Optional[int] = padded_audio_features * self.padding_value
for i in range(len(__UpperCAmelCase ) ):
lowerCAmelCase__ : int = audio_features[i]
lowerCAmelCase__ : Any = feature
# return as BatchFeature
if return_attention_mask:
lowerCAmelCase__ : Tuple = {"""audio_values""": padded_audio_features, """audio_mask""": audio_mask}
else:
lowerCAmelCase__ : Optional[Any] = {"""audio_values""": padded_audio_features}
lowerCAmelCase__ : Optional[int] = BatchFeature(data=__UpperCAmelCase ,tensor_type=__UpperCAmelCase )
return encoded_inputs
| 37
|
'''simple docstring'''
from maths.prime_factors import prime_factors
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if not isinstance(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : int = f"""Input value of [number={number}] must be an integer"""
raise TypeError(UpperCamelCase )
if number < 1:
raise ValueError("""Input must be a positive integer""" )
return -1 if len(prime_factors(UpperCamelCase ) ) % 2 else 1
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
| 1
|
'''simple docstring'''
import pickle
import numpy as np
from matplotlib import pyplot as plt
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=0.2 ,__UpperCAmelCase=0.2 ) -> str:
lowerCAmelCase__ : Dict = bp_numa
lowerCAmelCase__ : Optional[int] = bp_numa
lowerCAmelCase__ : Dict = bp_numa
lowerCAmelCase__ : List[Any] = conva_get[:2]
lowerCAmelCase__ : Tuple = conva_get[2]
lowerCAmelCase__ : int = size_pa
lowerCAmelCase__ : str = rate_w
lowerCAmelCase__ : Optional[int] = rate_t
lowerCAmelCase__ : int = [
np.mat(-1 * np.random.rand(self.conva[0] ,self.conva[0] ) + 0.5 )
for i in range(self.conva[1] )
]
lowerCAmelCase__ : Union[str, Any] = np.mat(-1 * np.random.rand(self.num_bpa ,self.num_bpa ) + 0.5 )
lowerCAmelCase__ : Dict = np.mat(-1 * np.random.rand(self.num_bpa ,self.num_bpa ) + 0.5 )
lowerCAmelCase__ : Union[str, Any] = -2 * np.random.rand(self.conva[1] ) + 1
lowerCAmelCase__ : Union[str, Any] = -2 * np.random.rand(self.num_bpa ) + 1
lowerCAmelCase__ : str = -2 * np.random.rand(self.num_bpa ) + 1
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
# save model dict with pickle
lowerCAmelCase__ : Optional[Any] = {
"""num_bp1""": self.num_bpa,
"""num_bp2""": self.num_bpa,
"""num_bp3""": self.num_bpa,
"""conv1""": self.conva,
"""step_conv1""": self.step_conva,
"""size_pooling1""": self.size_poolinga,
"""rate_weight""": self.rate_weight,
"""rate_thre""": self.rate_thre,
"""w_conv1""": self.w_conva,
"""wkj""": self.wkj,
"""vji""": self.vji,
"""thre_conv1""": self.thre_conva,
"""thre_bp2""": self.thre_bpa,
"""thre_bp3""": self.thre_bpa,
}
with open(__UpperCAmelCase ,"""wb""" ) as f:
pickle.dump(__UpperCAmelCase ,__UpperCAmelCase )
print(F"""Model saved： {save_path}""" )
@classmethod
def UpperCAmelCase_ ( cls ,__UpperCAmelCase ) -> List[Any]:
# read saved model
with open(__UpperCAmelCase ,"""rb""" ) as f:
lowerCAmelCase__ : List[str] = pickle.load(__UpperCAmelCase ) # noqa: S301
lowerCAmelCase__ : Any = model_dic.get("""conv1""" )
conv_get.append(model_dic.get("""step_conv1""" ) )
lowerCAmelCase__ : Optional[int] = model_dic.get("""size_pooling1""" )
lowerCAmelCase__ : Union[str, Any] = model_dic.get("""num_bp1""" )
lowerCAmelCase__ : Dict = model_dic.get("""num_bp2""" )
lowerCAmelCase__ : Optional[int] = model_dic.get("""num_bp3""" )
lowerCAmelCase__ : List[Any] = model_dic.get("""rate_weight""" )
lowerCAmelCase__ : Any = model_dic.get("""rate_thre""" )
# create model instance
lowerCAmelCase__ : Optional[int] = CNN(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase )
# modify model parameter
lowerCAmelCase__ : List[str] = model_dic.get("""w_conv1""" )
lowerCAmelCase__ : Dict = model_dic.get("""wkj""" )
lowerCAmelCase__ : str = model_dic.get("""vji""" )
lowerCAmelCase__ : Dict = model_dic.get("""thre_conv1""" )
lowerCAmelCase__ : Dict = model_dic.get("""thre_bp2""" )
lowerCAmelCase__ : Any = model_dic.get("""thre_bp3""" )
return conv_ins
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
return 1 / (1 + np.exp(-1 * x ))
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[Any]:
return round(__UpperCAmelCase ,3 )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> List[str]:
# convolution process
lowerCAmelCase__ : int = convs[0]
lowerCAmelCase__ : Union[str, Any] = convs[1]
lowerCAmelCase__ : Optional[Any] = np.shape(__UpperCAmelCase )[0]
# get the data slice of original image data, data_focus
lowerCAmelCase__ : Tuple = []
for i_focus in range(0 ,size_data - size_conv + 1 ,__UpperCAmelCase ):
for j_focus in range(0 ,size_data - size_conv + 1 ,__UpperCAmelCase ):
lowerCAmelCase__ : str = data[
i_focus : i_focus + size_conv, j_focus : j_focus + size_conv
]
data_focus.append(__UpperCAmelCase )
# calculate the feature map of every single kernel, and saved as list of matrix
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : str = int((size_data - size_conv) / conv_step + 1 )
for i_map in range(__UpperCAmelCase ):
lowerCAmelCase__ : List[str] = []
for i_focus in range(len(__UpperCAmelCase ) ):
lowerCAmelCase__ : Tuple = (
np.sum(np.multiply(data_focus[i_focus] ,w_convs[i_map] ) )
- thre_convs[i_map]
)
featuremap.append(self.sig(__UpperCAmelCase ) )
lowerCAmelCase__ : List[str] = np.asmatrix(__UpperCAmelCase ).reshape(
__UpperCAmelCase ,__UpperCAmelCase )
data_featuremap.append(__UpperCAmelCase )
# expanding the data slice to One dimenssion
lowerCAmelCase__ : int = []
for each_focus in data_focus:
focusa_list.extend(self.Expand_Mat(__UpperCAmelCase ) )
lowerCAmelCase__ : int = np.asarray(__UpperCAmelCase )
return focus_list, data_featuremap
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="average_pool" ) -> List[Any]:
# pooling process
lowerCAmelCase__ : str = len(featuremaps[0] )
lowerCAmelCase__ : List[Any] = int(size_map / size_pooling )
lowerCAmelCase__ : Any = []
for i_map in range(len(__UpperCAmelCase ) ):
lowerCAmelCase__ : Tuple = featuremaps[i_map]
lowerCAmelCase__ : Union[str, Any] = []
for i_focus in range(0 ,__UpperCAmelCase ,__UpperCAmelCase ):
for j_focus in range(0 ,__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : str = feature_map[
i_focus : i_focus + size_pooling,
j_focus : j_focus + size_pooling,
]
if pooling_type == "average_pool":
# average pooling
map_pooled.append(np.average(__UpperCAmelCase ) )
elif pooling_type == "max_pooling":
# max pooling
map_pooled.append(np.max(__UpperCAmelCase ) )
lowerCAmelCase__ : Tuple = np.asmatrix(__UpperCAmelCase ).reshape(__UpperCAmelCase ,__UpperCAmelCase )
featuremap_pooled.append(__UpperCAmelCase )
return featuremap_pooled
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
# expanding three dimension data to one dimension list
lowerCAmelCase__ : List[str] = []
for i in range(len(__UpperCAmelCase ) ):
lowerCAmelCase__ : List[Any] = np.shape(data[i] )
lowerCAmelCase__ : Optional[Any] = data[i].reshape(1 ,shapes[0] * shapes[1] )
lowerCAmelCase__ : List[Any] = data_listed.getA().tolist()[0]
data_expanded.extend(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = np.asarray(__UpperCAmelCase )
return data_expanded
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
# expanding matrix to one dimension list
lowerCAmelCase__ : Optional[Any] = np.asarray(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = np.shape(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = data_mat.reshape(1 ,shapes[0] * shapes[1] )
return data_expanded
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : int = []
lowerCAmelCase__ : Optional[Any] = 0
for i_map in range(__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = np.ones((size_map, size_map) )
for i in range(0 ,__UpperCAmelCase ,__UpperCAmelCase ):
for j in range(0 ,__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Optional[Any] = pd_pool[
i_pool
]
lowerCAmelCase__ : Optional[Any] = i_pool + 1
lowerCAmelCase__ : Dict = np.multiply(
__UpperCAmelCase ,np.multiply(out_map[i_map] ,(1 - out_map[i_map]) ) )
pd_all.append(__UpperCAmelCase )
return pd_all
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=bool ) -> Tuple:
# model traning
print("""----------------------Start Training-------------------------""" )
print((""" - - Shape: Train_Data """, np.shape(__UpperCAmelCase )) )
print((""" - - Shape: Teach_Data """, np.shape(__UpperCAmelCase )) )
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Union[str, Any] = []
lowerCAmelCase__ : Optional[int] = 1_0000
while rp < n_repeat and mse >= error_accuracy:
lowerCAmelCase__ : int = 0
print(F"""-------------Learning Time {rp}--------------""" )
for p in range(len(__UpperCAmelCase ) ):
# print('------------Learning Image: %d--------------'%p)
lowerCAmelCase__ : Any = np.asmatrix(datas_train[p] )
lowerCAmelCase__ : List[str] = np.asarray(datas_teach[p] )
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self.convolute(
__UpperCAmelCase ,self.conva ,self.w_conva ,self.thre_conva ,conv_step=self.step_conva ,)
lowerCAmelCase__ : Union[str, Any] = self.pooling(__UpperCAmelCase ,self.size_poolinga )
lowerCAmelCase__ : List[Any] = np.shape(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = self._expand(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = data_bp_input
lowerCAmelCase__ : int = np.dot(__UpperCAmelCase ,self.vji.T ) - self.thre_bpa
lowerCAmelCase__ : Tuple = self.sig(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = np.dot(__UpperCAmelCase ,self.wkj.T ) - self.thre_bpa
lowerCAmelCase__ : str = self.sig(__UpperCAmelCase )
# --------------Model Leaning ------------------------
# calculate error and gradient---------------
lowerCAmelCase__ : Optional[Any] = np.multiply(
(data_teach - bp_outa) ,np.multiply(__UpperCAmelCase ,(1 - bp_outa) ) )
lowerCAmelCase__ : List[str] = np.multiply(
np.dot(__UpperCAmelCase ,self.wkj ) ,np.multiply(__UpperCAmelCase ,(1 - bp_outa) ) )
lowerCAmelCase__ : Dict = np.dot(__UpperCAmelCase ,self.vji )
lowerCAmelCase__ : List[Any] = pd_i_all / (self.size_poolinga * self.size_poolinga)
lowerCAmelCase__ : Dict = pd_conva_pooled.T.getA().tolist()
lowerCAmelCase__ : Any = self._calculate_gradient_from_pool(
__UpperCAmelCase ,__UpperCAmelCase ,shape_featuremapa[0] ,shape_featuremapa[1] ,self.size_poolinga ,)
# weight and threshold learning process---------
# convolution layer
for k_conv in range(self.conva[1] ):
lowerCAmelCase__ : Dict = self._expand_mat(pd_conva_all[k_conv] )
lowerCAmelCase__ : Optional[Any] = self.rate_weight * np.dot(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : Any = self.w_conva[k_conv] + delta_w.reshape(
(self.conva[0], self.conva[0]) )
lowerCAmelCase__ : int = (
self.thre_conva[k_conv]
- np.sum(pd_conva_all[k_conv] ) * self.rate_thre
)
# all connected layer
lowerCAmelCase__ : str = self.wkj + pd_k_all.T * bp_outa * self.rate_weight
lowerCAmelCase__ : Tuple = self.vji + pd_j_all.T * bp_outa * self.rate_weight
lowerCAmelCase__ : List[Any] = self.thre_bpa - pd_k_all * self.rate_thre
lowerCAmelCase__ : str = self.thre_bpa - pd_j_all * self.rate_thre
# calculate the sum error of all single image
lowerCAmelCase__ : Union[str, Any] = np.sum(abs(data_teach - bp_outa ) )
error_count += errors
# print(' ----Teach ',data_teach)
# print(' ----BP_output ',bp_out3)
lowerCAmelCase__ : List[Any] = rp + 1
lowerCAmelCase__ : List[Any] = error_count / patterns
all_mse.append(__UpperCAmelCase )
def draw_error():
lowerCAmelCase__ : Optional[int] = [error_accuracy for i in range(int(n_repeat * 1.2 ) )]
plt.plot(__UpperCAmelCase ,"""+-""" )
plt.plot(__UpperCAmelCase ,"""r--""" )
plt.xlabel("""Learning Times""" )
plt.ylabel("""All_mse""" )
plt.grid(__UpperCAmelCase ,alpha=0.5 )
plt.show()
print("""------------------Training Complished---------------------""" )
print((""" - - Training epoch: """, rp, F""" - - Mse: {mse:.6f}""") )
if draw_e:
draw_error()
return mse
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
# model predict
lowerCAmelCase__ : str = []
print("""-------------------Start Testing-------------------------""" )
print((""" - - Shape: Test_Data """, np.shape(__UpperCAmelCase )) )
for p in range(len(__UpperCAmelCase ) ):
lowerCAmelCase__ : List[str] = np.asmatrix(datas_test[p] )
lowerCAmelCase__ , lowerCAmelCase__ : Dict = self.convolute(
__UpperCAmelCase ,self.conva ,self.w_conva ,self.thre_conva ,conv_step=self.step_conva ,)
lowerCAmelCase__ : str = self.pooling(__UpperCAmelCase ,self.size_poolinga )
lowerCAmelCase__ : Tuple = self._expand(__UpperCAmelCase )
lowerCAmelCase__ : Dict = data_bp_input
lowerCAmelCase__ : Optional[int] = bp_outa * self.vji.T - self.thre_bpa
lowerCAmelCase__ : Dict = self.sig(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = bp_outa * self.wkj.T - self.thre_bpa
lowerCAmelCase__ : Any = self.sig(__UpperCAmelCase )
produce_out.extend(bp_outa.getA().tolist() )
lowerCAmelCase__ : int = [list(map(self.do_round ,__UpperCAmelCase ) ) for each in produce_out]
return np.asarray(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
# return the data of image after convoluting process so we can check it out
lowerCAmelCase__ : str = np.asmatrix(__UpperCAmelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = self.convolute(
__UpperCAmelCase ,self.conva ,self.w_conva ,self.thre_conva ,conv_step=self.step_conva ,)
lowerCAmelCase__ : List[Any] = self.pooling(__UpperCAmelCase ,self.size_poolinga )
return data_conveda, data_pooleda
if __name__ == "__main__":
pass
| 37
|
'''simple docstring'''
import os
import re
import unicodedata
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import is_torch_available, logging
if is_torch_available():
import torch
if TYPE_CHECKING:
from transformers.pipelines.conversational import Conversation
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''AI-Sweden/gpt-sw3-126m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-126m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-350m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-350m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-1.6b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-1.6b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-6.7b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-6.7b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-20b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-20b/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''AI-Sweden/gpt-sw3-126m''': 2048,
'''AI-Sweden/gpt-sw3-350m''': 2048,
'''AI-Sweden/gpt-sw3-1.6b''': 2048,
'''AI-Sweden/gpt-sw3-6.7b''': 2048,
'''AI-Sweden/gpt-sw3-20b''': 2048,
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : str = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Optional[int] = ['''input_ids''', '''attention_mask''']
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
lowerCAmelCase__ : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
lowerCAmelCase__ : Dict = kwargs.get("""name_or_path""" )
if name_or_path is None:
logger.warning(
"""name_or_path not provided, will work for all GPTSw3 models except gpt-sw3-7b,"""
""" you are testing the model, this can safely be ignored""" )
lowerCAmelCase__ : Tuple = """None"""
# Default definitions for our 2 tokenizer versions, with None-checks to enable proper testing
lowerCAmelCase__ : Union[str, Any] = """<|endoftext|>""" if eos_token is None else eos_token
lowerCAmelCase__ : Dict = """<unk>""" if unk_token is None else unk_token
if "gpt-sw3-7b" in name_or_path:
lowerCAmelCase__ : Any = unk_token if pad_token is None else pad_token
lowerCAmelCase__ : Dict = eos_token if bos_token is None else bos_token
else:
lowerCAmelCase__ : List[str] = """<pad>""" if pad_token is None else pad_token
lowerCAmelCase__ : Optional[int] = """<s>""" if bos_token is None else bos_token
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : Optional[int] = do_lower_case
lowerCAmelCase__ : Dict = remove_space
lowerCAmelCase__ : Optional[Any] = keep_accents
lowerCAmelCase__ : int = vocab_file
lowerCAmelCase__ : str = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
# Used for whitespace normalization in input texts
# fmt : off
lowerCAmelCase__ : int = {""" """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """""", """"""}
# fmt : on
# Regular expression to remove non-printing characters (e.g. some unicode control chars) in preprocessing
lowerCAmelCase__ : List[str] = re.compile(
F"""[{''.join(map(__UpperCAmelCase ,list(range(0 ,9 ) ) + list(range(11 ,32 ) ) + list(range(127 ,160 ) ) + [160, 173, 8203] ) )}]""" )
def __getstate__( self ) -> Any:
lowerCAmelCase__ : int = self.__dict__.copy()
lowerCAmelCase__ : Optional[int] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Tuple = {}
lowerCAmelCase__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
@property
# Copied from transformers.models.albert.tokenization_albert.AlbertTokenizer.vocab_size
def UpperCAmelCase_ ( self ) -> int:
return len(self.sp_model )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : Tuple = self.non_printing_characters_re.sub("""""" ,__UpperCAmelCase )
# Normalize whitespaces
lowerCAmelCase__ : List[Any] = """""".join([char if char not in self.whitespaces else """ """ for char in text] )
# NFC Unicode normalization
lowerCAmelCase__ : List[Any] = unicodedata.normalize("""NFC""" ,__UpperCAmelCase )
return text
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[Any] = self.preprocess_text(__UpperCAmelCase )
return self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.IdToPiece(__UpperCAmelCase )
@staticmethod
def UpperCAmelCase_ ( __UpperCAmelCase ) -> str:
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : int = []
lowerCAmelCase__ : Optional[int] = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
# TODO: Check if this is needed, as it ensures that decode(encode(doc)) != doc by adding extra whitespace in the decoded document
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Optional[Any] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Any = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string
def UpperCAmelCase_ ( self ) -> Dict[str, int]:
lowerCAmelCase__ : Optional[int] = {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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : Optional[int] = 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:
lowerCAmelCase__ : str = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Union[List[int], List[List[int]], "torch.Tensor"]:
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : int = self.sp_model.encode(__UpperCAmelCase )
else:
lowerCAmelCase__ : int = [self.preprocess_text(__UpperCAmelCase ) for t in text]
lowerCAmelCase__ : Any = self.sp_model.encode(__UpperCAmelCase )
if return_tensors is True or return_tensors == "pt":
lowerCAmelCase__ : Tuple = torch.tensor(__UpperCAmelCase )
return token_ids
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.decode(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[int]:
lowerCAmelCase__ : List[Any] = [F"""User: {text}""" if is_user else F"""Bot: {text}""" for is_user, text in conversation.iter_texts()]
lowerCAmelCase__ : Any = (
F"""{self.eos_token}{self.bos_token}""" + F"""{self.bos_token}""".join(__UpperCAmelCase ) + F"""{self.bos_token}Bot:"""
)
return self.encode(text=__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ....tokenization_utils_fast import PreTrainedTokenizerFast
from ....utils import logging
from .tokenization_retribert import RetriBertTokenizer
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''vocab.txt''', '''tokenizer_file''': '''tokenizer.json'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''yjernite/retribert-base-uncased''': (
'''https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/vocab.txt'''
),
},
'''tokenizer_file''': {
'''yjernite/retribert-base-uncased''': (
'''https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/tokenizer.json'''
),
},
}
_lowerCAmelCase = {
'''yjernite/retribert-base-uncased''': 512,
}
_lowerCAmelCase = {
'''yjernite/retribert-base-uncased''': {'''do_lower_case''': True},
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Union[str, Any] = VOCAB_FILES_NAMES
__lowercase : int = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : str = PRETRAINED_INIT_CONFIGURATION
__lowercase : str = RetriBertTokenizer
__lowercase : str = ['''input_ids''', '''attention_mask''']
def __init__( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=True ,__UpperCAmelCase="[UNK]" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="[PAD]" ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[MASK]" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Tuple:
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 ,)
lowerCAmelCase__ : Optional[int] = 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
):
lowerCAmelCase__ : Optional[Any] = getattr(__UpperCAmelCase ,normalizer_state.pop("""type""" ) )
lowerCAmelCase__ : Union[str, Any] = do_lower_case
lowerCAmelCase__ : List[Any] = strip_accents
lowerCAmelCase__ : Optional[int] = tokenize_chinese_chars
lowerCAmelCase__ : Optional[int] = normalizer_class(**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = do_lower_case
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ) -> List[Any]:
lowerCAmelCase__ : int = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : int = [self.sep_token_id]
lowerCAmelCase__ : Optional[int] = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
lowerCAmelCase__ : Union[str, Any] = self._tokenizer.model.save(__UpperCAmelCase ,name=__UpperCAmelCase )
return tuple(__UpperCAmelCase )
| 37
|
'''simple docstring'''
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = XLMRobertaModel.from_pretrained("""xlm-roberta-base""" )
lowerCAmelCase__ : Optional[int] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : str = torch.Size((1, 12, 768) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.0_1_0_1, 0.1_2_1_8, -0.0_8_0_3, 0.0_8_0_1, 0.1_3_2_7, 0.0_7_7_6, -0.1_2_1_5, 0.2_3_8_3, 0.3_3_3_8, 0.3_1_0_6, 0.0_3_0_0, 0.0_2_5_2]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Optional[int] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
@slow
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : Union[str, Any] = XLMRobertaModel.from_pretrained("""xlm-roberta-large""" )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : Dict = torch.Size((1, 12, 1024) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[[-0.0_6_9_9, -0.0_3_1_8, 0.0_7_0_5, -0.1_2_4_1, 0.0_9_9_9, -0.0_5_2_0, 0.1_0_0_4, -0.1_8_3_8, -0.4_7_0_4, 0.1_4_3_7, 0.0_8_2_1, 0.0_1_2_6]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.large')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
| 37
| 1
|
'''simple docstring'''
import os
import re
import unicodedata
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import is_torch_available, logging
if is_torch_available():
import torch
if TYPE_CHECKING:
from transformers.pipelines.conversational import Conversation
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''AI-Sweden/gpt-sw3-126m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-126m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-350m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-350m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-1.6b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-1.6b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-6.7b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-6.7b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-20b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-20b/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''AI-Sweden/gpt-sw3-126m''': 2048,
'''AI-Sweden/gpt-sw3-350m''': 2048,
'''AI-Sweden/gpt-sw3-1.6b''': 2048,
'''AI-Sweden/gpt-sw3-6.7b''': 2048,
'''AI-Sweden/gpt-sw3-20b''': 2048,
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : str = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Optional[int] = ['''input_ids''', '''attention_mask''']
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
lowerCAmelCase__ : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
lowerCAmelCase__ : Dict = kwargs.get("""name_or_path""" )
if name_or_path is None:
logger.warning(
"""name_or_path not provided, will work for all GPTSw3 models except gpt-sw3-7b,"""
""" you are testing the model, this can safely be ignored""" )
lowerCAmelCase__ : Tuple = """None"""
# Default definitions for our 2 tokenizer versions, with None-checks to enable proper testing
lowerCAmelCase__ : Union[str, Any] = """<|endoftext|>""" if eos_token is None else eos_token
lowerCAmelCase__ : Dict = """<unk>""" if unk_token is None else unk_token
if "gpt-sw3-7b" in name_or_path:
lowerCAmelCase__ : Any = unk_token if pad_token is None else pad_token
lowerCAmelCase__ : Dict = eos_token if bos_token is None else bos_token
else:
lowerCAmelCase__ : List[str] = """<pad>""" if pad_token is None else pad_token
lowerCAmelCase__ : Optional[int] = """<s>""" if bos_token is None else bos_token
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : Optional[int] = do_lower_case
lowerCAmelCase__ : Dict = remove_space
lowerCAmelCase__ : Optional[Any] = keep_accents
lowerCAmelCase__ : int = vocab_file
lowerCAmelCase__ : str = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
# Used for whitespace normalization in input texts
# fmt : off
lowerCAmelCase__ : int = {""" """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """""", """"""}
# fmt : on
# Regular expression to remove non-printing characters (e.g. some unicode control chars) in preprocessing
lowerCAmelCase__ : List[str] = re.compile(
F"""[{''.join(map(__UpperCAmelCase ,list(range(0 ,9 ) ) + list(range(11 ,32 ) ) + list(range(127 ,160 ) ) + [160, 173, 8203] ) )}]""" )
def __getstate__( self ) -> Any:
lowerCAmelCase__ : int = self.__dict__.copy()
lowerCAmelCase__ : Optional[int] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Tuple = {}
lowerCAmelCase__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
@property
# Copied from transformers.models.albert.tokenization_albert.AlbertTokenizer.vocab_size
def UpperCAmelCase_ ( self ) -> int:
return len(self.sp_model )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : Tuple = self.non_printing_characters_re.sub("""""" ,__UpperCAmelCase )
# Normalize whitespaces
lowerCAmelCase__ : List[Any] = """""".join([char if char not in self.whitespaces else """ """ for char in text] )
# NFC Unicode normalization
lowerCAmelCase__ : List[Any] = unicodedata.normalize("""NFC""" ,__UpperCAmelCase )
return text
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[Any] = self.preprocess_text(__UpperCAmelCase )
return self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.IdToPiece(__UpperCAmelCase )
@staticmethod
def UpperCAmelCase_ ( __UpperCAmelCase ) -> str:
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : int = []
lowerCAmelCase__ : Optional[int] = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
# TODO: Check if this is needed, as it ensures that decode(encode(doc)) != doc by adding extra whitespace in the decoded document
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Optional[Any] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Any = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string
def UpperCAmelCase_ ( self ) -> Dict[str, int]:
lowerCAmelCase__ : Optional[int] = {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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : Optional[int] = 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:
lowerCAmelCase__ : str = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Union[List[int], List[List[int]], "torch.Tensor"]:
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : int = self.sp_model.encode(__UpperCAmelCase )
else:
lowerCAmelCase__ : int = [self.preprocess_text(__UpperCAmelCase ) for t in text]
lowerCAmelCase__ : Any = self.sp_model.encode(__UpperCAmelCase )
if return_tensors is True or return_tensors == "pt":
lowerCAmelCase__ : Tuple = torch.tensor(__UpperCAmelCase )
return token_ids
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.decode(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[int]:
lowerCAmelCase__ : List[Any] = [F"""User: {text}""" if is_user else F"""Bot: {text}""" for is_user, text in conversation.iter_texts()]
lowerCAmelCase__ : Any = (
F"""{self.eos_token}{self.bos_token}""" + F"""{self.bos_token}""".join(__UpperCAmelCase ) + F"""{self.bos_token}Bot:"""
)
return self.encode(text=__UpperCAmelCase )
| 37
|
'''simple docstring'''
from pickle import UnpicklingError
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict
from ..utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
with open(UpperCamelCase , """rb""" ) as flax_state_f:
lowerCAmelCase__ : Union[str, Any] = from_bytes(UpperCamelCase , flax_state_f.read() )
except UnpicklingError as e:
try:
with open(UpperCamelCase ) as f:
if f.read().startswith("""version""" ):
raise OSError(
"""You seem to have cloned a repository without having git-lfs installed. Please"""
""" install git-lfs and run `git lfs install` followed by `git lfs pull` in the"""
""" folder you cloned.""" )
else:
raise ValueError from e
except (UnicodeDecodeError, ValueError):
raise EnvironmentError(f"""Unable to convert {model_file} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(UpperCamelCase , UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
import torch # noqa: F401
except ImportError:
logger.error(
"""Loading Flax weights in PyTorch requires both PyTorch and Flax to be installed. Please see"""
""" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"""
""" instructions.""" )
raise
# check if we have bf16 weights
lowerCAmelCase__ : str = flatten_dict(jax.tree_util.tree_map(lambda UpperCamelCase : x.dtype == jnp.bfloataa , UpperCamelCase ) ).values()
if any(UpperCamelCase ):
# convert all weights to fp32 if they are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
"""Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` """
"""before loading those in PyTorch model.""" )
lowerCAmelCase__ : Dict = jax.tree_util.tree_map(
lambda UpperCamelCase : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , UpperCamelCase )
lowerCAmelCase__ : Any = """"""
lowerCAmelCase__ : Any = flatten_dict(UpperCamelCase , sep=""".""" )
lowerCAmelCase__ : Optional[int] = pt_model.state_dict()
# keep track of unexpected & missing keys
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : int = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
lowerCAmelCase__ : Union[str, Any] = flax_key_tuple.split(""".""" )
if flax_key_tuple_array[-1] == "kernel" and flax_tensor.ndim == 4:
lowerCAmelCase__ : Optional[int] = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = jnp.transpose(UpperCamelCase , (3, 2, 0, 1) )
elif flax_key_tuple_array[-1] == "kernel":
lowerCAmelCase__ : str = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = flax_tensor.T
elif flax_key_tuple_array[-1] == "scale":
lowerCAmelCase__ : int = flax_key_tuple_array[:-1] + ["""weight"""]
if "time_embedding" not in flax_key_tuple_array:
for i, flax_key_tuple_string in enumerate(UpperCamelCase ):
lowerCAmelCase__ : List[str] = (
flax_key_tuple_string.replace("""_0""" , """.0""" )
.replace("""_1""" , """.1""" )
.replace("""_2""" , """.2""" )
.replace("""_3""" , """.3""" )
.replace("""_4""" , """.4""" )
.replace("""_5""" , """.5""" )
.replace("""_6""" , """.6""" )
.replace("""_7""" , """.7""" )
.replace("""_8""" , """.8""" )
.replace("""_9""" , """.9""" )
)
lowerCAmelCase__ : Union[str, Any] = """.""".join(UpperCamelCase )
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
f"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
f"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
lowerCAmelCase__ : int = np.asarray(UpperCamelCase ) if not isinstance(UpperCamelCase , np.ndarray ) else flax_tensor
lowerCAmelCase__ : int = torch.from_numpy(UpperCamelCase )
# remove from missing keys
missing_keys.remove(UpperCamelCase )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(UpperCamelCase )
pt_model.load_state_dict(UpperCamelCase )
# re-transform missing_keys to list
lowerCAmelCase__ : Optional[int] = list(UpperCamelCase )
if len(UpperCamelCase ) > 0:
logger.warning(
"""Some weights of the Flax model were not used when initializing the PyTorch model"""
f""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
f""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
""" (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This"""
f""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
""" to be exactly identical (e.g. initializing a BertForSequenceClassification model from a"""
""" FlaxBertForSequenceClassification model).""" )
if len(UpperCamelCase ) > 0:
logger.warning(
f"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
f""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
""" use it for predictions and inference.""" )
return pt_model
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = 0
while b > 0:
if b & 1:
res += a
a += a
b >>= 1
return res
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = 0
while b > 0:
if b & 1:
lowerCAmelCase__ : Tuple = ((res % c) + (a % c)) % c
a += a
b >>= 1
return res
| 37
|
'''simple docstring'''
import inspect
import unittest
import numpy as np
from tests.test_modeling_common import floats_tensor
from transformers import DetrConfig, MaskFormerConfig, SwinConfig, is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device
from transformers.utils import cached_property
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaskFormerForInstanceSegmentation, MaskFormerModel
if is_vision_available():
from transformers import MaskFormerImageProcessor
if is_vision_available():
from PIL import Image
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=2 ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase=10 ,__UpperCAmelCase=3 ,__UpperCAmelCase=32 * 4 ,__UpperCAmelCase=32 * 6 ,__UpperCAmelCase=4 ,__UpperCAmelCase=32 ,) -> str:
lowerCAmelCase__ : Optional[int] = parent
lowerCAmelCase__ : Optional[int] = batch_size
lowerCAmelCase__ : Optional[int] = is_training
lowerCAmelCase__ : Dict = use_auxiliary_loss
lowerCAmelCase__ : Union[str, Any] = num_queries
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : List[str] = min_size
lowerCAmelCase__ : int = max_size
lowerCAmelCase__ : Optional[Any] = num_labels
lowerCAmelCase__ : List[Any] = mask_feature_size
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : str = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to(
__UpperCAmelCase )
lowerCAmelCase__ : str = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=__UpperCAmelCase )
lowerCAmelCase__ : Any = (
torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=__UpperCAmelCase ) > 0.5
).float()
lowerCAmelCase__ : Optional[int] = (torch.rand((self.batch_size, self.num_labels) ,device=__UpperCAmelCase ) > 0.5).long()
lowerCAmelCase__ : Any = self.get_config()
return config, pixel_values, pixel_mask, mask_labels, class_labels
def UpperCAmelCase_ ( self ) -> Dict:
return MaskFormerConfig.from_backbone_and_decoder_configs(
backbone_config=SwinConfig(
depths=[1, 1, 1, 1] ,) ,decoder_config=DetrConfig(
decoder_ffn_dim=128 ,num_queries=self.num_queries ,decoder_attention_heads=2 ,d_model=self.mask_feature_size ,) ,mask_feature_size=self.mask_feature_size ,fpn_feature_size=self.mask_feature_size ,num_channels=self.num_channels ,num_labels=self.num_labels ,)
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.prepare_config_and_inputs()
lowerCAmelCase__ : List[str] = {"""pixel_values""": pixel_values, """pixel_mask""": pixel_mask}
return config, inputs_dict
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Optional[int] = output.encoder_hidden_states
lowerCAmelCase__ : Optional[int] = output.pixel_decoder_hidden_states
lowerCAmelCase__ : Dict = output.transformer_decoder_hidden_states
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,config.decoder_config.decoder_layers )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=False ) -> Optional[Any]:
with torch.no_grad():
lowerCAmelCase__ : int = MaskFormerModel(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
lowerCAmelCase__ : str = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : int = model(__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
# the correct shape of output.transformer_decoder_hidden_states ensure the correcteness of the
# encoder and pixel decoder
self.parent.assertEqual(
output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.mask_feature_size) ,)
# let's ensure the other two hidden state exists
self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(output.encoder_last_hidden_state is not None )
if output_hidden_states:
self.check_output_hidden_state(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : Dict = MaskFormerForInstanceSegmentation(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
def comm_check_on_output(__UpperCAmelCase ):
# let's still check that all the required stuff is there
self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.encoder_last_hidden_state is not None )
# okay, now we need to check the logits shape
# due to the encoder compression, masks have a //4 spatial size
self.parent.assertEqual(
result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,)
# + 1 for null class
self.parent.assertEqual(
result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : Dict = model(__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = model(
pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
self.parent.assertTrue(result.loss is not None )
self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) )
@require_torch
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Optional[Any] = (MaskFormerModel, MaskFormerForInstanceSegmentation) if is_torch_available() else ()
__lowercase : int = (
{'''feature-extraction''': MaskFormerModel, '''image-segmentation''': MaskFormerForInstanceSegmentation}
if is_torch_available()
else {}
)
__lowercase : Union[str, Any] = False
__lowercase : Dict = False
__lowercase : Tuple = False
__lowercase : List[Any] = False
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : str = MaskFormerModelTester(self )
lowerCAmelCase__ : List[Any] = ConfigTester(self ,config_class=__UpperCAmelCase ,has_text_modality=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
self.config_tester.run_common_tests()
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ , lowerCAmelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_maskformer_instance_segmentation_head_model(*__UpperCAmelCase )
@unittest.skip(reason="""MaskFormer does not use inputs_embeds""" )
def UpperCAmelCase_ ( self ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormer does not have a get_input_embeddings method""" )
def UpperCAmelCase_ ( self ) -> str:
pass
@unittest.skip(reason="""MaskFormer is not a generative model""" )
def UpperCAmelCase_ ( self ) -> Any:
pass
@unittest.skip(reason="""MaskFormer does not use token embeddings""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
@require_torch_multi_gpu
@unittest.skip(
reason="""MaskFormer has some layers using `add_module` which doesn't work well with `nn.DataParallel`""" )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
pass
@unittest.skip("""Will be fixed soon by reducing the size of the model used for common tests.""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : str = model_class(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowerCAmelCase__ : Dict = [*signature.parameters.keys()]
lowerCAmelCase__ : Tuple = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] ,__UpperCAmelCase )
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model_name in ["facebook/maskformer-swin-small-coco"]:
lowerCAmelCase__ : List[str] = MaskFormerModel.from_pretrained(__UpperCAmelCase )
self.assertIsNotNone(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (self.model_tester.min_size,) * 2
lowerCAmelCase__ : Any = {
"""pixel_values""": torch.randn((2, 3, *size) ,device=__UpperCAmelCase ),
"""mask_labels""": torch.randn((2, 10, *size) ,device=__UpperCAmelCase ),
"""class_labels""": torch.zeros(2 ,10 ,device=__UpperCAmelCase ).long(),
}
lowerCAmelCase__ : Tuple = MaskFormerForInstanceSegmentation(MaskFormerConfig() ).to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : int = model_class(__UpperCAmelCase ).to(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase ,output_attentions=__UpperCAmelCase )
self.assertTrue(outputs.attentions is not None )
def UpperCAmelCase_ ( self ) -> int:
if not self.model_tester.is_training:
return
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Dict = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : List[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : List[str] = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase ).loss
loss.backward()
def UpperCAmelCase_ ( self ) -> List[str]:
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Tuple = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Tuple = True
lowerCAmelCase__ : Optional[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : Dict = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
lowerCAmelCase__ : str = outputs.pixel_decoder_hidden_states[0]
pixel_decoder_hidden_states.retain_grad()
# we requires_grad=True in inputs_embeds (line 2152), the original implementation don't
lowerCAmelCase__ : Union[str, Any] = outputs.transformer_decoder_hidden_states[0]
transformer_decoder_hidden_states.retain_grad()
lowerCAmelCase__ : List[Any] = outputs.attentions[0]
attentions.retain_grad()
outputs.loss.backward(retain_graph=__UpperCAmelCase )
self.assertIsNotNone(encoder_hidden_states.grad )
self.assertIsNotNone(pixel_decoder_hidden_states.grad )
self.assertIsNotNone(transformer_decoder_hidden_states.grad )
self.assertIsNotNone(attentions.grad )
_lowerCAmelCase = 1e-4
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_vision
@slow
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@cached_property
def UpperCAmelCase_ ( self ) -> List[Any]:
return (
MaskFormerImageProcessor.from_pretrained("""facebook/maskformer-swin-small-coco""" )
if is_vision_available()
else None
)
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Any = MaskFormerModel.from_pretrained("""facebook/maskformer-swin-small-coco""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = self.default_image_processor
lowerCAmelCase__ : str = prepare_img()
lowerCAmelCase__ : Optional[int] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[-0.0_4_8_2, 0.9_2_2_8, 0.4_9_5_1], [-0.2_5_4_7, 0.8_0_1_7, 0.8_5_2_7], [-0.0_0_6_9, 0.3_3_8_5, -0.0_0_8_9]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.encoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.8_4_2_2, -0.8_4_3_4, -0.9_7_1_8], [-1.0_1_4_4, -0.5_5_6_5, -0.4_1_9_5], [-1.0_0_3_8, -0.4_4_8_4, -0.1_9_6_1]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Optional[int] = torch.tensor(
[[0.2_8_5_2, -0.0_1_5_9, 0.9_7_3_5], [0.6_2_5_4, 0.1_8_5_8, 0.8_5_2_9], [-0.0_6_8_0, -0.4_1_1_6, 1.8_4_1_3]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : List[str] = prepare_img()
lowerCAmelCase__ : str = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[int] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : Optional[int] = [
[-1.3_7_3_7_1_2_4, -1.7_7_2_4_9_3_7, -1.9_3_6_4_2_3_3],
[-1.5_9_7_7_2_8_1, -1.9_8_6_7_9_3_9, -2.1_5_2_3_6_9_5],
[-1.5_7_9_5_3_9_8, -1.9_2_6_9_8_3_2, -2.0_9_3_9_4_2],
]
lowerCAmelCase__ : Optional[int] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Tuple = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[
[1.65_12E00, -5.25_72E00, -3.35_19E00],
[3.61_69E-02, -5.90_25E00, -2.93_13E00],
[1.07_66E-04, -7.76_30E00, -5.12_63E00],
] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-resnet101-coco-stuff""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : int = prepare_img()
lowerCAmelCase__ : Optional[Any] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : str = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : int = [[-0.9_0_4_6, -2.6_3_6_6, -4.6_0_6_2], [-3.4_1_7_9, -5.7_8_9_0, -8.8_0_5_7], [-4.9_1_7_9, -7.6_5_6_0, -1_0.7_7_1_1]]
lowerCAmelCase__ : List[str] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Tuple = torch.tensor(
[[4.7_1_8_8, -3.2_5_8_5, -2.8_8_5_7], [6.6_8_7_1, -2.9_1_8_1, -1.2_4_8_7], [7.2_4_4_9, -2.2_7_6_4, -2.1_8_7_4]] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : str = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Dict = self.default_image_processor
lowerCAmelCase__ : Union[str, Any] = image_processor(
[np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors="""pt""" ,)
lowerCAmelCase__ : Tuple = inputs["""pixel_values"""].to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""mask_labels"""]]
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""class_labels"""]]
with torch.no_grad():
lowerCAmelCase__ : Any = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
| 37
| 1
|
'''simple docstring'''
import argparse
from pathlib import Path
import torch
from transformers import OPTConfig, OPTModel
from transformers.utils import logging
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = torch.load(UpperCamelCase , map_location="""cpu""" )
if "model" in sd.keys():
lowerCAmelCase__ : str = torch.load(UpperCamelCase , map_location="""cpu""" )["""model"""]
# pop unnecessary weights
lowerCAmelCase__ : Dict = [
"""decoder.version""",
"""decoder.output_projection.weight""",
]
for key in keys_to_delete:
if key in sd:
sd.pop(UpperCamelCase )
lowerCAmelCase__ : List[str] = {
"""decoder.project_in_dim.weight""": """decoder.project_in.weight""",
"""decoder.project_out_dim.weight""": """decoder.project_out.weight""",
"""decoder.layer_norm.weight""": """decoder.final_layer_norm.weight""",
"""decoder.layer_norm.bias""": """decoder.final_layer_norm.bias""",
}
for old_key, new_key in keys_to_rename.items():
if old_key in sd:
lowerCAmelCase__ : Any = sd.pop(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = list(sd.keys() )
for key in keys:
if ".qkv_proj." in key:
lowerCAmelCase__ : Optional[int] = sd[key]
# We split QKV in separate Q,K,V
lowerCAmelCase__ : Optional[int] = key.replace(""".qkv_proj.""" , """.q_proj.""" )
lowerCAmelCase__ : List[str] = key.replace(""".qkv_proj.""" , """.k_proj.""" )
lowerCAmelCase__ : Optional[Any] = key.replace(""".qkv_proj.""" , """.v_proj.""" )
lowerCAmelCase__ : Union[str, Any] = value.shape[0]
assert depth % 3 == 0
# `SequeuceParallelTransformerBlock` has QKV weight is separated in K,V,Q despite the naming:
# https://cs.github.com/facebookresearch/metaseq/blob/51871bd73cd04c038f239ea2a26db1d7f6b37927/metaseq/modules/sequence_parallel_transformer_layer.py#L97
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : int = torch.split(UpperCamelCase , depth // 3 , dim=0 )
lowerCAmelCase__ : Dict = q
lowerCAmelCase__ : Dict = k
lowerCAmelCase__ : Any = v
del sd[key]
return sd
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=None ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = load_checkpoint(UpperCamelCase )
if config is not None:
lowerCAmelCase__ : Tuple = OPTConfig.from_pretrained(UpperCamelCase )
else:
lowerCAmelCase__ : List[str] = OPTConfig()
lowerCAmelCase__ : Optional[int] = OPTModel(UpperCamelCase ).half().eval()
model.load_state_dict(UpperCamelCase )
# Check results
Path(UpperCamelCase ).mkdir(exist_ok=UpperCamelCase )
model.save_pretrained(UpperCamelCase )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--fairseq_path''',
type=str,
help=(
'''path to fairseq checkpoint in correct format. You can find all checkpoints in the correct format here:'''
''' https://huggingface.co/models?other=opt_metasq'''
),
)
parser.add_argument('''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model.''')
parser.add_argument('''--hf_config''', default=None, type=str, help='''Define HF config.''')
_lowerCAmelCase = parser.parse_args()
convert_opt_checkpoint(args.fairseq_path, args.pytorch_dump_folder_path, config=args.hf_config)
| 37
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''microsoft/focalnet-tiny''': '''https://huggingface.co/microsoft/focalnet-tiny/resolve/main/config.json''',
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[Any] = '''focalnet'''
def __init__( self ,__UpperCAmelCase=224 ,__UpperCAmelCase=4 ,__UpperCAmelCase=3 ,__UpperCAmelCase=96 ,__UpperCAmelCase=False ,__UpperCAmelCase=[192, 384, 768, 768] ,__UpperCAmelCase=[2, 2, 6, 2] ,__UpperCAmelCase=[2, 2, 2, 2] ,__UpperCAmelCase=[3, 3, 3, 3] ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=4.0 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=False ,__UpperCAmelCase=1E-4 ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-5 ,__UpperCAmelCase=32 ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Optional[Any]:
super().__init__(**__UpperCAmelCase )
lowerCAmelCase__ : Dict = image_size
lowerCAmelCase__ : int = patch_size
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : Dict = embed_dim
lowerCAmelCase__ : List[str] = use_conv_embed
lowerCAmelCase__ : List[Any] = hidden_sizes
lowerCAmelCase__ : Dict = depths
lowerCAmelCase__ : List[str] = focal_levels
lowerCAmelCase__ : List[str] = focal_windows
lowerCAmelCase__ : Dict = hidden_act
lowerCAmelCase__ : Dict = mlp_ratio
lowerCAmelCase__ : Tuple = hidden_dropout_prob
lowerCAmelCase__ : Tuple = drop_path_rate
lowerCAmelCase__ : Dict = use_layerscale
lowerCAmelCase__ : Optional[Any] = layerscale_value
lowerCAmelCase__ : str = use_post_layernorm
lowerCAmelCase__ : Union[str, Any] = use_post_layernorm_in_modulation
lowerCAmelCase__ : int = normalize_modulator
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : List[str] = layer_norm_eps
lowerCAmelCase__ : List[Any] = encoder_stride
lowerCAmelCase__ : Dict = ["""stem"""] + [F"""stage{idx}""" for idx in range(1 ,len(self.depths ) + 1 )]
lowerCAmelCase__ , lowerCAmelCase__ : Any = get_aligned_output_features_output_indices(
out_features=__UpperCAmelCase ,out_indices=__UpperCAmelCase ,stage_names=self.stage_names )
| 37
| 1
|
'''simple docstring'''
import numpy as np
import qiskit
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = 8 , UpperCamelCase = None ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = np.random.default_rng(seed=UpperCamelCase )
# Roughly 25% of the qubits will contribute to the key.
# So we take more than we need.
lowerCAmelCase__ : Tuple = 6 * key_len
# Measurement basis for Alice's qubits.
lowerCAmelCase__ : List[str] = rng.integers(2 , size=UpperCamelCase )
# The set of states Alice will prepare.
lowerCAmelCase__ : Optional[Any] = rng.integers(2 , size=UpperCamelCase )
# Measurement basis for Bob's qubits.
lowerCAmelCase__ : List[Any] = rng.integers(2 , size=UpperCamelCase )
# Quantum Circuit to simulate BB84
lowerCAmelCase__ : Any = qiskit.QuantumCircuit(UpperCamelCase , name="""BB84""" )
# Alice prepares her qubits according to rules above.
for index, _ in enumerate(UpperCamelCase ):
if alice_state[index] == 1:
bbaa_circ.x(UpperCamelCase )
if alice_basis[index] == 1:
bbaa_circ.h(UpperCamelCase )
bbaa_circ.barrier()
# Bob measures the received qubits according to rules above.
for index, _ in enumerate(UpperCamelCase ):
if bob_basis[index] == 1:
bbaa_circ.h(UpperCamelCase )
bbaa_circ.barrier()
bbaa_circ.measure_all()
# Simulate the quantum circuit.
lowerCAmelCase__ : Any = 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.
lowerCAmelCase__ : Optional[int] = qiskit.execute(UpperCamelCase , UpperCamelCase , shots=1 , seed_simulator=UpperCamelCase )
# Returns the result of measurement.
lowerCAmelCase__ : Union[str, Any] = job.result().get_counts(UpperCamelCase ).most_frequent()
# Extracting the generated key from the simulation results.
# Only keep measurement results where Alice and Bob chose the same basis.
lowerCAmelCase__ : str = """""".join(
[
result_bit
for alice_basis_bit, bob_basis_bit, result_bit in zip(
UpperCamelCase , UpperCamelCase , UpperCamelCase )
if alice_basis_bit == bob_basis_bit
] )
# Get final key. Pad with 0 if too short, otherwise truncate.
lowerCAmelCase__ : Tuple = gen_key[:key_len] if len(UpperCamelCase ) >= key_len else gen_key.ljust(UpperCamelCase , """0""" )
return key
if __name__ == "__main__":
print(F"""The generated key is : {bbaa(8, seed=0)}""")
from doctest import testmod
testmod()
| 37
|
'''simple docstring'''
import argparse
import os.path as osp
import re
import torch
from safetensors.torch import load_file, save_file
# =================#
# UNet Conversion #
# =================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''time_embed.0.weight''', '''time_embedding.linear_1.weight'''),
('''time_embed.0.bias''', '''time_embedding.linear_1.bias'''),
('''time_embed.2.weight''', '''time_embedding.linear_2.weight'''),
('''time_embed.2.bias''', '''time_embedding.linear_2.bias'''),
('''input_blocks.0.0.weight''', '''conv_in.weight'''),
('''input_blocks.0.0.bias''', '''conv_in.bias'''),
('''out.0.weight''', '''conv_norm_out.weight'''),
('''out.0.bias''', '''conv_norm_out.bias'''),
('''out.2.weight''', '''conv_out.weight'''),
('''out.2.bias''', '''conv_out.bias'''),
]
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''in_layers.0''', '''norm1'''),
('''in_layers.2''', '''conv1'''),
('''out_layers.0''', '''norm2'''),
('''out_layers.3''', '''conv2'''),
('''emb_layers.1''', '''time_emb_proj'''),
('''skip_connection''', '''conv_shortcut'''),
]
_lowerCAmelCase = []
# hardcoded number of downblocks and resnets/attentions...
# would need smarter logic for other networks.
for i in range(4):
# loop over downblocks/upblocks
for j in range(2):
# loop over resnets/attentions for downblocks
_lowerCAmelCase = F"""down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.0."""
unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
if i < 3:
# no attention layers in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.1."""
unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
for j in range(3):
# loop over resnets/attentions for upblocks
_lowerCAmelCase = F"""up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.0."""
unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
if i > 0:
# no attention layers in up_blocks.0
_lowerCAmelCase = F"""up_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.1."""
unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
if i < 3:
# no downsample in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0.conv."""
_lowerCAmelCase = F"""input_blocks.{3*(i+1)}.0.op."""
unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
# no upsample in up_blocks.3
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""output_blocks.{3*i + 2}.{1 if i == 0 else 2}."""
unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
_lowerCAmelCase = '''mid_block.attentions.0.'''
_lowerCAmelCase = '''middle_block.1.'''
unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
for j in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{j}."""
_lowerCAmelCase = F"""middle_block.{2*j}."""
unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = {k: k for k in unet_state_dict.keys()}
for sd_name, hf_name in unet_conversion_map:
lowerCAmelCase__ : Optional[int] = sd_name
for k, v in mapping.items():
if "resnets" in k:
for sd_part, hf_part in unet_conversion_map_resnet:
lowerCAmelCase__ : Any = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
for k, v in mapping.items():
for sd_part, hf_part in unet_conversion_map_layer:
lowerCAmelCase__ : List[Any] = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
lowerCAmelCase__ : Tuple = {v: unet_state_dict[k] for k, v in mapping.items()}
return new_state_dict
# ================#
# VAE Conversion #
# ================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''nin_shortcut''', '''conv_shortcut'''),
('''norm_out''', '''conv_norm_out'''),
('''mid.attn_1.''', '''mid_block.attentions.0.'''),
]
for i in range(4):
# down_blocks have two resnets
for j in range(2):
_lowerCAmelCase = F"""encoder.down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""encoder.down.{i}.block.{j}."""
vae_conversion_map.append((sd_down_prefix, hf_down_prefix))
if i < 3:
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0."""
_lowerCAmelCase = F"""down.{i}.downsample."""
vae_conversion_map.append((sd_downsample_prefix, hf_downsample_prefix))
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""up.{3-i}.upsample."""
vae_conversion_map.append((sd_upsample_prefix, hf_upsample_prefix))
# up_blocks have three resnets
# also, up blocks in hf are numbered in reverse from sd
for j in range(3):
_lowerCAmelCase = F"""decoder.up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""decoder.up.{3-i}.block.{j}."""
vae_conversion_map.append((sd_up_prefix, hf_up_prefix))
# this part accounts for mid blocks in both the encoder and the decoder
for i in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{i}."""
_lowerCAmelCase = F"""mid.block_{i+1}."""
vae_conversion_map.append((sd_mid_res_prefix, hf_mid_res_prefix))
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''norm.''', '''group_norm.'''),
('''q.''', '''query.'''),
('''k.''', '''key.'''),
('''v.''', '''value.'''),
('''proj_out.''', '''proj_attn.'''),
]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return w.reshape(*w.shape , 1 , 1 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = {k: k for k in vae_state_dict.keys()}
for k, v in mapping.items():
for sd_part, hf_part in vae_conversion_map:
lowerCAmelCase__ : str = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
for k, v in mapping.items():
if "attentions" in k:
for sd_part, hf_part in vae_conversion_map_attn:
lowerCAmelCase__ : Dict = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
lowerCAmelCase__ : Union[str, Any] = {v: vae_state_dict[k] for k, v in mapping.items()}
lowerCAmelCase__ : Tuple = ["""q""", """k""", """v""", """proj_out"""]
for k, v in new_state_dict.items():
for weight_name in weights_to_convert:
if f"""mid.attn_1.{weight_name}.weight""" in k:
print(f"""Reshaping {k} for SD format""" )
lowerCAmelCase__ : Optional[int] = reshape_weight_for_sd(UpperCamelCase )
return new_state_dict
# =========================#
# Text Encoder Conversion #
# =========================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''resblocks.''', '''text_model.encoder.layers.'''),
('''ln_1''', '''layer_norm1'''),
('''ln_2''', '''layer_norm2'''),
('''.c_fc.''', '''.fc1.'''),
('''.c_proj.''', '''.fc2.'''),
('''.attn''', '''.self_attn'''),
('''ln_final.''', '''transformer.text_model.final_layer_norm.'''),
('''token_embedding.weight''', '''transformer.text_model.embeddings.token_embedding.weight'''),
('''positional_embedding''', '''transformer.text_model.embeddings.position_embedding.weight'''),
]
_lowerCAmelCase = {re.escape(x[1]): x[0] for x in textenc_conversion_lst}
_lowerCAmelCase = re.compile('''|'''.join(protected.keys()))
# Ordering is from https://github.com/pytorch/pytorch/blob/master/test/cpp/api/modules.cpp
_lowerCAmelCase = {'''q''': 0, '''k''': 1, '''v''': 2}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = {}
lowerCAmelCase__ : int = {}
lowerCAmelCase__ : List[Any] = {}
for k, v in text_enc_dict.items():
if (
k.endswith(""".self_attn.q_proj.weight""" )
or k.endswith(""".self_attn.k_proj.weight""" )
or k.endswith(""".self_attn.v_proj.weight""" )
):
lowerCAmelCase__ : Optional[int] = k[: -len(""".q_proj.weight""" )]
lowerCAmelCase__ : Tuple = k[-len("""q_proj.weight""" )]
if k_pre not in capture_qkv_weight:
lowerCAmelCase__ : List[Any] = [None, None, None]
lowerCAmelCase__ : Dict = v
continue
if (
k.endswith(""".self_attn.q_proj.bias""" )
or k.endswith(""".self_attn.k_proj.bias""" )
or k.endswith(""".self_attn.v_proj.bias""" )
):
lowerCAmelCase__ : str = k[: -len(""".q_proj.bias""" )]
lowerCAmelCase__ : List[str] = k[-len("""q_proj.bias""" )]
if k_pre not in capture_qkv_bias:
lowerCAmelCase__ : Union[str, Any] = [None, None, None]
lowerCAmelCase__ : Any = v
continue
lowerCAmelCase__ : Dict = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = v
for k_pre, tensors in capture_qkv_weight.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : Any = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Tuple = torch.cat(UpperCamelCase )
for k_pre, tensors in capture_qkv_bias.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : str = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : List[Any] = torch.cat(UpperCamelCase )
return new_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return text_enc_dict
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument('''--model_path''', default=None, type=str, required=True, help='''Path to the model to convert.''')
parser.add_argument('''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the output model.''')
parser.add_argument('''--half''', action='''store_true''', help='''Save weights in half precision.''')
parser.add_argument(
'''--use_safetensors''', action='''store_true''', help='''Save weights use safetensors, default is ckpt.'''
)
_lowerCAmelCase = parser.parse_args()
assert args.model_path is not None, "Must provide a model path!"
assert args.checkpoint_path is not None, "Must provide a checkpoint path!"
# Path for safetensors
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''model.safetensors''')
# Load models from safetensors if it exists, if it doesn't pytorch
if osp.exists(unet_path):
_lowerCAmelCase = load_file(unet_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(unet_path, map_location='''cpu''')
if osp.exists(vae_path):
_lowerCAmelCase = load_file(vae_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(vae_path, map_location='''cpu''')
if osp.exists(text_enc_path):
_lowerCAmelCase = load_file(text_enc_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''pytorch_model.bin''')
_lowerCAmelCase = torch.load(text_enc_path, map_location='''cpu''')
# Convert the UNet model
_lowerCAmelCase = convert_unet_state_dict(unet_state_dict)
_lowerCAmelCase = {'''model.diffusion_model.''' + k: v for k, v in unet_state_dict.items()}
# Convert the VAE model
_lowerCAmelCase = convert_vae_state_dict(vae_state_dict)
_lowerCAmelCase = {'''first_stage_model.''' + k: v for k, v in vae_state_dict.items()}
# Easiest way to identify v2.0 model seems to be that the text encoder (OpenCLIP) is deeper
_lowerCAmelCase = '''text_model.encoder.layers.22.layer_norm2.bias''' in text_enc_dict
if is_vaa_model:
# Need to add the tag 'transformer' in advance so we can knock it out from the final layer-norm
_lowerCAmelCase = {'''transformer.''' + k: v for k, v in text_enc_dict.items()}
_lowerCAmelCase = convert_text_enc_state_dict_vaa(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.model.''' + k: v for k, v in text_enc_dict.items()}
else:
_lowerCAmelCase = convert_text_enc_state_dict(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.transformer.''' + k: v for k, v in text_enc_dict.items()}
# Put together new checkpoint
_lowerCAmelCase = {**unet_state_dict, **vae_state_dict, **text_enc_dict}
if args.half:
_lowerCAmelCase = {k: v.half() for k, v in state_dict.items()}
if args.use_safetensors:
save_file(state_dict, args.checkpoint_path)
else:
_lowerCAmelCase = {'''state_dict''': state_dict}
torch.save(state_dict, args.checkpoint_path)
| 37
| 1
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tensorflow_text_available, is_torch_available
_lowerCAmelCase = {
'''configuration_ernie''': ['''ERNIE_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''ErnieConfig''', '''ErnieOnnxConfig'''],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = [
'''ERNIE_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''ErnieForCausalLM''',
'''ErnieForMaskedLM''',
'''ErnieForMultipleChoice''',
'''ErnieForNextSentencePrediction''',
'''ErnieForPreTraining''',
'''ErnieForQuestionAnswering''',
'''ErnieForSequenceClassification''',
'''ErnieForTokenClassification''',
'''ErnieModel''',
'''ErniePreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_ernie import ERNIE_PRETRAINED_CONFIG_ARCHIVE_MAP, ErnieConfig, ErnieOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_ernie import (
ERNIE_PRETRAINED_MODEL_ARCHIVE_LIST,
ErnieForCausalLM,
ErnieForMaskedLM,
ErnieForMultipleChoice,
ErnieForNextSentencePrediction,
ErnieForPreTraining,
ErnieForQuestionAnswering,
ErnieForSequenceClassification,
ErnieForTokenClassification,
ErnieModel,
ErniePreTrainedModel,
)
else:
import sys
_lowerCAmelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 37
|
'''simple docstring'''
import os
from bleurt import score # From: git+https://github.com/google-research/bleurt.git
import datasets
_lowerCAmelCase = datasets.logging.get_logger(__name__)
_lowerCAmelCase = '''\
@inproceedings{bleurt,
title={BLEURT: Learning Robust Metrics for Text Generation},
author={Thibault Sellam and Dipanjan Das and Ankur P. Parikh},
booktitle={ACL},
year={2020},
url={https://arxiv.org/abs/2004.04696}
}
'''
_lowerCAmelCase = '''\
BLEURT a learnt evaluation metric for Natural Language Generation. It is built using multiple phases of transfer learning starting from a pretrained BERT model (Devlin et al. 2018)
and then employing another pre-training phrase using synthetic data. Finally it is trained on WMT human annotations. You may run BLEURT out-of-the-box or fine-tune
it for your specific application (the latter is expected to perform better).
See the project\'s README at https://github.com/google-research/bleurt#readme for more information.
'''
_lowerCAmelCase = '''
BLEURT score.
Args:
`predictions` (list of str): prediction/candidate sentences
`references` (list of str): reference sentences
`checkpoint` BLEURT checkpoint. Will default to BLEURT-tiny if None.
Returns:
\'scores\': List of scores.
Examples:
>>> predictions = ["hello there", "general kenobi"]
>>> references = ["hello there", "general kenobi"]
>>> bleurt = datasets.load_metric("bleurt")
>>> results = bleurt.compute(predictions=predictions, references=references)
>>> print([round(v, 2) for v in results["scores"]])
[1.03, 1.04]
'''
_lowerCAmelCase = {
'''bleurt-tiny-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-128.zip''',
'''bleurt-tiny-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-512.zip''',
'''bleurt-base-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-128.zip''',
'''bleurt-base-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-512.zip''',
'''bleurt-large-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-128.zip''',
'''bleurt-large-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-512.zip''',
'''BLEURT-20-D3''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D3.zip''',
'''BLEURT-20-D6''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D6.zip''',
'''BLEURT-20-D12''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D12.zip''',
'''BLEURT-20''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20.zip''',
}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,homepage="""https://github.com/google-research/bleurt""" ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(
{
"""predictions""": datasets.Value("""string""" ,id="""sequence""" ),
"""references""": datasets.Value("""string""" ,id="""sequence""" ),
} ) ,codebase_urls=["""https://github.com/google-research/bleurt"""] ,reference_urls=["""https://github.com/google-research/bleurt""", """https://arxiv.org/abs/2004.04696"""] ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
# check that config name specifies a valid BLEURT model
if self.config_name == "default":
logger.warning(
"""Using default BLEURT-Base checkpoint for sequence maximum length 128. """
"""You can use a bigger model for better results with e.g.: datasets.load_metric('bleurt', 'bleurt-large-512').""" )
lowerCAmelCase__ : str = """bleurt-base-128"""
if self.config_name.lower() in CHECKPOINT_URLS:
lowerCAmelCase__ : Union[str, Any] = self.config_name.lower()
elif self.config_name.upper() in CHECKPOINT_URLS:
lowerCAmelCase__ : List[Any] = self.config_name.upper()
else:
raise KeyError(
F"""{self.config_name} model not found. You should supply the name of a model checkpoint for bleurt in {CHECKPOINT_URLS.keys()}""" )
# download the model checkpoint specified by self.config_name and set up the scorer
lowerCAmelCase__ : int = dl_manager.download_and_extract(CHECKPOINT_URLS[checkpoint_name] )
lowerCAmelCase__ : Dict = score.BleurtScorer(os.path.join(__UpperCAmelCase ,__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Union[str, Any] = self.scorer.score(references=__UpperCAmelCase ,candidates=__UpperCAmelCase )
return {"scores": scores}
| 37
| 1
|
'''simple docstring'''
from typing import Callable, List, Optional, Tuple, Union
import torch
from transformers import CLIPTextModel, CLIPTokenizer
from ...configuration_utils import ConfigMixin, register_to_config
from ...models import ModelMixin, TransformeraDModel, VQModel
from ...schedulers import VQDiffusionScheduler
from ...utils import logging
from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
_lowerCAmelCase = logging.get_logger(__name__) # pylint: disable=invalid-name
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@register_to_config
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = None ) -> Dict:
super().__init__()
lowerCAmelCase__ : List[Any] = learnable
if self.learnable:
assert hidden_size is not None, "learnable=True requires `hidden_size` to be set"
assert length is not None, "learnable=True requires `length` to be set"
lowerCAmelCase__ : Dict = torch.zeros(__UpperCAmelCase ,__UpperCAmelCase )
else:
lowerCAmelCase__ : Optional[int] = None
lowerCAmelCase__ : Union[str, Any] = torch.nn.Parameter(__UpperCAmelCase )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : VQModel
__lowercase : CLIPTextModel
__lowercase : CLIPTokenizer
__lowercase : TransformeraDModel
__lowercase : LearnedClassifierFreeSamplingEmbeddings
__lowercase : VQDiffusionScheduler
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,) -> Union[str, Any]:
super().__init__()
self.register_modules(
vqvae=__UpperCAmelCase ,transformer=__UpperCAmelCase ,text_encoder=__UpperCAmelCase ,tokenizer=__UpperCAmelCase ,scheduler=__UpperCAmelCase ,learned_classifier_free_sampling_embeddings=__UpperCAmelCase ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : Optional[int] = len(__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else 1
# get prompt text embeddings
lowerCAmelCase__ : int = self.tokenizer(
__UpperCAmelCase ,padding="""max_length""" ,max_length=self.tokenizer.model_max_length ,return_tensors="""pt""" ,)
lowerCAmelCase__ : Dict = text_inputs.input_ids
if text_input_ids.shape[-1] > self.tokenizer.model_max_length:
lowerCAmelCase__ : List[Any] = self.tokenizer.batch_decode(text_input_ids[:, self.tokenizer.model_max_length :] )
logger.warning(
"""The following part of your input was truncated because CLIP can only handle sequences up to"""
F""" {self.tokenizer.model_max_length} tokens: {removed_text}""" )
lowerCAmelCase__ : Optional[Any] = text_input_ids[:, : self.tokenizer.model_max_length]
lowerCAmelCase__ : Union[str, Any] = self.text_encoder(text_input_ids.to(self.device ) )[0]
# NOTE: This additional step of normalizing the text embeddings is from VQ-Diffusion.
# While CLIP does normalize the pooled output of the text transformer when combining
# the image and text embeddings, CLIP does not directly normalize the last hidden state.
#
# CLIP normalizing the pooled output.
# https://github.com/huggingface/transformers/blob/d92e22d1f28324f513f3080e5c47c071a3916721/src/transformers/models/clip/modeling_clip.py#L1052-L1053
lowerCAmelCase__ : List[str] = prompt_embeds / prompt_embeds.norm(dim=-1 ,keepdim=__UpperCAmelCase )
# duplicate text embeddings for each generation per prompt
lowerCAmelCase__ : Any = prompt_embeds.repeat_interleave(__UpperCAmelCase ,dim=0 )
if do_classifier_free_guidance:
if self.learned_classifier_free_sampling_embeddings.learnable:
lowerCAmelCase__ : Dict = self.learned_classifier_free_sampling_embeddings.embeddings
lowerCAmelCase__ : Optional[Any] = negative_prompt_embeds.unsqueeze(0 ).repeat(__UpperCAmelCase ,1 ,1 )
else:
lowerCAmelCase__ : Optional[Any] = [""""""] * batch_size
lowerCAmelCase__ : int = text_input_ids.shape[-1]
lowerCAmelCase__ : str = self.tokenizer(
__UpperCAmelCase ,padding="""max_length""" ,max_length=__UpperCAmelCase ,truncation=__UpperCAmelCase ,return_tensors="""pt""" ,)
lowerCAmelCase__ : Union[str, Any] = self.text_encoder(uncond_input.input_ids.to(self.device ) )[0]
# See comment for normalizing text embeddings
lowerCAmelCase__ : Union[str, Any] = negative_prompt_embeds / negative_prompt_embeds.norm(dim=-1 ,keepdim=__UpperCAmelCase )
# duplicate unconditional embeddings for each generation per prompt, using mps friendly method
lowerCAmelCase__ : Optional[int] = negative_prompt_embeds.shape[1]
lowerCAmelCase__ : int = negative_prompt_embeds.repeat(1 ,__UpperCAmelCase ,1 )
lowerCAmelCase__ : Tuple = negative_prompt_embeds.view(batch_size * num_images_per_prompt ,__UpperCAmelCase ,-1 )
# For classifier free guidance, we need to do two forward passes.
# Here we concatenate the unconditional and text embeddings into a single batch
# to avoid doing two forward passes
lowerCAmelCase__ : List[Any] = torch.cat([negative_prompt_embeds, prompt_embeds] )
return prompt_embeds
@torch.no_grad()
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase = 100 ,__UpperCAmelCase = 5.0 ,__UpperCAmelCase = 1.0 ,__UpperCAmelCase = 1 ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = "pil" ,__UpperCAmelCase = True ,__UpperCAmelCase = None ,__UpperCAmelCase = 1 ,) -> Union[ImagePipelineOutput, Tuple]:
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = 1
elif isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : int = len(__UpperCAmelCase )
else:
raise ValueError(F"""`prompt` has to be of type `str` or `list` but is {type(__UpperCAmelCase )}""" )
lowerCAmelCase__ : Optional[Any] = batch_size * num_images_per_prompt
lowerCAmelCase__ : Union[str, Any] = guidance_scale > 1.0
lowerCAmelCase__ : Any = self._encode_prompt(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase )
if (callback_steps is None) or (
callback_steps is not None and (not isinstance(__UpperCAmelCase ,__UpperCAmelCase ) or callback_steps <= 0)
):
raise ValueError(
F"""`callback_steps` has to be a positive integer but is {callback_steps} of type"""
F""" {type(__UpperCAmelCase )}.""" )
# get the initial completely masked latents unless the user supplied it
lowerCAmelCase__ : Dict = (batch_size, self.transformer.num_latent_pixels)
if latents is None:
lowerCAmelCase__ : List[Any] = self.transformer.num_vector_embeds - 1
lowerCAmelCase__ : int = torch.full(__UpperCAmelCase ,__UpperCAmelCase ).to(self.device )
else:
if latents.shape != latents_shape:
raise ValueError(F"""Unexpected latents shape, got {latents.shape}, expected {latents_shape}""" )
if (latents < 0).any() or (latents >= self.transformer.num_vector_embeds).any():
raise ValueError(
"""Unexpected latents value(s). All latents be valid embedding indices i.e. in the range 0,"""
F""" {self.transformer.num_vector_embeds - 1} (inclusive).""" )
lowerCAmelCase__ : List[str] = latents.to(self.device )
# set timesteps
self.scheduler.set_timesteps(__UpperCAmelCase ,device=self.device )
lowerCAmelCase__ : Any = self.scheduler.timesteps.to(self.device )
lowerCAmelCase__ : int = latents
for i, t in enumerate(self.progress_bar(__UpperCAmelCase ) ):
# expand the sample if we are doing classifier free guidance
lowerCAmelCase__ : Union[str, Any] = torch.cat([sample] * 2 ) if do_classifier_free_guidance else sample
# predict the un-noised image
# model_output == `log_p_x_0`
lowerCAmelCase__ : List[str] = self.transformer(__UpperCAmelCase ,encoder_hidden_states=__UpperCAmelCase ,timestep=__UpperCAmelCase ).sample
if do_classifier_free_guidance:
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = model_output.chunk(2 )
lowerCAmelCase__ : List[Any] = model_output_uncond + guidance_scale * (model_output_text - model_output_uncond)
model_output -= torch.logsumexp(__UpperCAmelCase ,dim=1 ,keepdim=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = self.truncate(__UpperCAmelCase ,__UpperCAmelCase )
# remove `log(0)`'s (`-inf`s)
lowerCAmelCase__ : Optional[Any] = model_output.clamp(-70 )
# compute the previous noisy sample x_t -> x_t-1
lowerCAmelCase__ : List[str] = self.scheduler.step(__UpperCAmelCase ,timestep=__UpperCAmelCase ,sample=__UpperCAmelCase ,generator=__UpperCAmelCase ).prev_sample
# call the callback, if provided
if callback is not None and i % callback_steps == 0:
callback(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : Tuple = self.vqvae.config.vq_embed_dim
lowerCAmelCase__ : Optional[int] = (batch_size, self.transformer.height, self.transformer.width, embedding_channels)
lowerCAmelCase__ : List[str] = self.vqvae.quantize.get_codebook_entry(__UpperCAmelCase ,shape=__UpperCAmelCase )
lowerCAmelCase__ : int = self.vqvae.decode(__UpperCAmelCase ,force_not_quantize=__UpperCAmelCase ).sample
lowerCAmelCase__ : Dict = (image / 2 + 0.5).clamp(0 ,1 )
lowerCAmelCase__ : Union[str, Any] = image.cpu().permute(0 ,2 ,3 ,1 ).numpy()
if output_type == "pil":
lowerCAmelCase__ : Dict = self.numpy_to_pil(__UpperCAmelCase )
if not return_dict:
return (image,)
return ImagePipelineOutput(images=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> torch.FloatTensor:
lowerCAmelCase__ , lowerCAmelCase__ : Any = torch.sort(__UpperCAmelCase ,1 ,descending=__UpperCAmelCase )
lowerCAmelCase__ : Any = torch.exp(__UpperCAmelCase )
lowerCAmelCase__ : Dict = sorted_p_x_0.cumsum(dim=1 ) < truncation_rate
# Ensure that at least the largest probability is not zeroed out
lowerCAmelCase__ : Dict = torch.full_like(keep_mask[:, 0:1, :] ,__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = torch.cat((all_true, keep_mask) ,dim=1 )
lowerCAmelCase__ : str = keep_mask[:, :-1, :]
lowerCAmelCase__ : Optional[Any] = keep_mask.gather(1 ,indices.argsort(1 ) )
lowerCAmelCase__ : List[Any] = log_p_x_0.clone()
lowerCAmelCase__ : Optional[Any] = -torch.inf # -inf = log(0)
return rv
| 37
|
'''simple docstring'''
import uuid
from typing import Any, Dict, List, Optional, Union
from ..utils import add_end_docstrings, is_tf_available, is_torch_available, logging
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_tf_available():
import tensorflow as tf
if is_torch_available():
import torch
_lowerCAmelCase = logging.get_logger(__name__)
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ) -> str:
if not conversation_id:
lowerCAmelCase__ : List[str] = uuid.uuida()
if past_user_inputs is None:
lowerCAmelCase__ : List[Any] = []
if generated_responses is None:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : uuid.UUID = conversation_id
lowerCAmelCase__ : List[str] = past_user_inputs
lowerCAmelCase__ : List[str] = generated_responses
lowerCAmelCase__ : Optional[str] = text
def __eq__( self ,__UpperCAmelCase ) -> Dict:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
return False
if self.uuid == other.uuid:
return True
return (
self.new_user_input == other.new_user_input
and self.past_user_inputs == other.past_user_inputs
and self.generated_responses == other.generated_responses
)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Optional[Any]:
if self.new_user_input:
if overwrite:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" was overwritten """
F"""with: \"{text}\".""" )
lowerCAmelCase__ : Optional[int] = text
else:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" new input """
F"""ignored: \"{text}\". Set `overwrite` to True to overwrite unprocessed user input""" )
else:
lowerCAmelCase__ : Optional[Any] = text
def UpperCAmelCase_ ( self ) -> List[Any]:
if self.new_user_input:
self.past_user_inputs.append(self.new_user_input )
lowerCAmelCase__ : Union[str, Any] = None
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
self.generated_responses.append(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
for user_input, generated_response in zip(self.past_user_inputs ,self.generated_responses ):
yield True, user_input
yield False, generated_response
if self.new_user_input:
yield True, self.new_user_input
def __repr__( self ) -> Tuple:
lowerCAmelCase__ : Tuple = F"""Conversation id: {self.uuid} \n"""
for is_user, text in self.iter_texts():
lowerCAmelCase__ : Any = """user""" if is_user else """bot"""
output += F"""{name} >> {text} \n"""
return output
@add_end_docstrings(
SCREAMING_SNAKE_CASE_ , R'''
min_length_for_response (`int`, *optional*, defaults to 32):
The minimum length (in number of tokens) for a response.
minimum_tokens (`int`, *optional*, defaults to 10):
The minimum length of tokens to leave for a response.
''' , )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Tuple:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
if self.tokenizer.pad_token_id is None:
lowerCAmelCase__ : Tuple = self.tokenizer.eos_token
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : List[Any] = {}
lowerCAmelCase__ : Optional[int] = {}
lowerCAmelCase__ : List[str] = {}
if min_length_for_response is not None:
lowerCAmelCase__ : Any = min_length_for_response
if minimum_tokens is not None:
lowerCAmelCase__ : Optional[int] = minimum_tokens
if "max_length" in generate_kwargs:
lowerCAmelCase__ : Optional[Any] = generate_kwargs["""max_length"""]
# self.max_length = generate_kwargs.get("max_length", self.model.config.max_length)
if clean_up_tokenization_spaces is not None:
lowerCAmelCase__ : int = clean_up_tokenization_spaces
if generate_kwargs:
forward_params.update(__UpperCAmelCase )
return preprocess_params, forward_params, postprocess_params
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase=0 ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Optional[int] = super().__call__(__UpperCAmelCase ,num_workers=__UpperCAmelCase ,**__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) and len(__UpperCAmelCase ) == 1:
return outputs[0]
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=32 ) -> Dict[str, Any]:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
raise ValueError("""ConversationalPipeline, expects Conversation as inputs""" )
if conversation.new_user_input is None:
raise ValueError(
F"""Conversation with UUID {type(conversation.uuid )} does not contain new user input to process. """
"""Add user inputs with the conversation's `add_user_input` method""" )
if hasattr(self.tokenizer ,"""_build_conversation_input_ids""" ):
lowerCAmelCase__ : str = self.tokenizer._build_conversation_input_ids(__UpperCAmelCase )
else:
# If the tokenizer cannot handle conversations, we default to only the old version
lowerCAmelCase__ : List[Any] = self._legacy_parse_and_tokenize(__UpperCAmelCase )
if self.framework == "pt":
lowerCAmelCase__ : List[Any] = torch.LongTensor([input_ids] )
elif self.framework == "tf":
lowerCAmelCase__ : Dict = tf.constant([input_ids] )
return {"input_ids": input_ids, "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=10 ,**__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[Any] = generate_kwargs.get("""max_length""" ,self.model.config.max_length )
lowerCAmelCase__ : Optional[Any] = model_inputs["""input_ids"""].shape[1]
if max_length - minimum_tokens < n:
logger.warning(F"""Conversation input is to long ({n}), trimming it to ({max_length} - {minimum_tokens})""" )
lowerCAmelCase__ : str = max_length - minimum_tokens
lowerCAmelCase__ : Union[str, Any] = model_inputs["""input_ids"""][:, -trim:]
if "attention_mask" in model_inputs:
lowerCAmelCase__ : Tuple = model_inputs["""attention_mask"""][:, -trim:]
lowerCAmelCase__ : str = model_inputs.pop("""conversation""" )
lowerCAmelCase__ : Union[str, Any] = max_length
lowerCAmelCase__ : Any = self.model.generate(**__UpperCAmelCase ,**__UpperCAmelCase )
if self.model.config.is_encoder_decoder:
lowerCAmelCase__ : int = 1
else:
lowerCAmelCase__ : int = n
return {"output_ids": output_ids[:, start_position:], "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=True ) -> List[str]:
lowerCAmelCase__ : Optional[int] = model_outputs["""output_ids"""]
lowerCAmelCase__ : Tuple = self.tokenizer.decode(
output_ids[0] ,skip_special_tokens=__UpperCAmelCase ,clean_up_tokenization_spaces=__UpperCAmelCase ,)
lowerCAmelCase__ : Union[str, Any] = model_outputs["""conversation"""]
conversation.mark_processed()
conversation.append_response(__UpperCAmelCase )
return conversation
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Dict = self.tokenizer.eos_token_id
lowerCAmelCase__ : int = []
for is_user, text in conversation.iter_texts():
if eos_token_id is not None:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) + [eos_token_id] )
else:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) )
if len(__UpperCAmelCase ) > self.tokenizer.model_max_length:
lowerCAmelCase__ : Optional[Any] = input_ids[-self.tokenizer.model_max_length :]
return input_ids
| 37
| 1
|
'''simple docstring'''
import copy
import os
from typing import Union
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''google/pix2struct-textcaps-base''': (
'''https://huggingface.co/google/pix2struct-textcaps-base/resolve/main/config.json'''
),
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[int] = '''pix2struct_text_model'''
__lowercase : str = ['''past_key_values''']
__lowercase : Dict = {
'''hidden_size''': '''hidden_size''',
'''num_attention_heads''': '''num_heads''',
'''num_hidden_layers''': '''num_layers''',
}
def __init__( self ,__UpperCAmelCase=5_0244 ,__UpperCAmelCase=768 ,__UpperCAmelCase=64 ,__UpperCAmelCase=2048 ,__UpperCAmelCase=12 ,__UpperCAmelCase=12 ,__UpperCAmelCase=32 ,__UpperCAmelCase=128 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=1E-6 ,__UpperCAmelCase=1.0 ,__UpperCAmelCase="gelu_new" ,__UpperCAmelCase=0 ,__UpperCAmelCase=False ,__UpperCAmelCase=0 ,__UpperCAmelCase=1 ,__UpperCAmelCase=False ,__UpperCAmelCase=True ,**__UpperCAmelCase ,) -> Tuple:
lowerCAmelCase__ : Optional[Any] = vocab_size
lowerCAmelCase__ : List[Any] = hidden_size
lowerCAmelCase__ : List[Any] = d_kv
lowerCAmelCase__ : Optional[Any] = d_ff
lowerCAmelCase__ : Optional[int] = num_layers
lowerCAmelCase__ : Optional[Any] = num_heads
lowerCAmelCase__ : str = relative_attention_num_buckets
lowerCAmelCase__ : Dict = relative_attention_max_distance
lowerCAmelCase__ : Optional[Any] = dropout_rate
lowerCAmelCase__ : Optional[int] = layer_norm_epsilon
lowerCAmelCase__ : List[str] = initializer_factor
lowerCAmelCase__ : Dict = use_cache
lowerCAmelCase__ : int = eos_token_id
lowerCAmelCase__ : Optional[int] = decoder_start_token_id
# for backwards compatibility
lowerCAmelCase__ : Optional[int] = dense_act_fn
super().__init__(
pad_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,decoder_start_token_id=__UpperCAmelCase ,tie_word_embeddings=__UpperCAmelCase ,is_decoder=__UpperCAmelCase ,**__UpperCAmelCase ,)
@classmethod
def UpperCAmelCase_ ( cls ,__UpperCAmelCase ,**__UpperCAmelCase ) -> "PretrainedConfig":
cls._set_token_in_kwargs(__UpperCAmelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Dict = cls.get_config_dict(__UpperCAmelCase ,**__UpperCAmelCase )
# get the text config dict if we are loading from Pix2StructConfig
if config_dict.get("""model_type""" ) == "pix2struct":
lowerCAmelCase__ : List[str] = config_dict["""text_config"""]
if "model_type" in config_dict and hasattr(cls ,"""model_type""" ) and config_dict["model_type"] != cls.model_type:
logger.warning(
F"""You are using a model of type {config_dict['model_type']} to instantiate a model of type """
F"""{cls.model_type}. This is not supported for all configurations of models and can yield errors.""" )
return cls.from_dict(__UpperCAmelCase ,**__UpperCAmelCase )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Tuple = '''pix2struct_vision_model'''
def __init__( self ,__UpperCAmelCase=768 ,__UpperCAmelCase=768 ,__UpperCAmelCase=2048 ,__UpperCAmelCase=64 ,__UpperCAmelCase=12 ,__UpperCAmelCase=12 ,__UpperCAmelCase="gelu_new" ,__UpperCAmelCase=1E-6 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=1E-10 ,__UpperCAmelCase=1.0 ,__UpperCAmelCase=4096 ,__UpperCAmelCase=32 ,__UpperCAmelCase=128 ,**__UpperCAmelCase ,) -> Union[str, Any]:
super().__init__(**__UpperCAmelCase )
lowerCAmelCase__ : Any = hidden_size
lowerCAmelCase__ : Dict = patch_embed_hidden_size
lowerCAmelCase__ : List[Any] = d_ff
lowerCAmelCase__ : Dict = dropout_rate
lowerCAmelCase__ : Any = num_hidden_layers
lowerCAmelCase__ : List[str] = num_attention_heads
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : List[Any] = initializer_factor
lowerCAmelCase__ : Dict = attention_dropout
lowerCAmelCase__ : Any = layer_norm_eps
lowerCAmelCase__ : str = dense_act_fn
lowerCAmelCase__ : Tuple = seq_len
lowerCAmelCase__ : List[Any] = relative_attention_num_buckets
lowerCAmelCase__ : Any = relative_attention_max_distance
lowerCAmelCase__ : List[Any] = d_kv
@classmethod
def UpperCAmelCase_ ( cls ,__UpperCAmelCase ,**__UpperCAmelCase ) -> "PretrainedConfig":
cls._set_token_in_kwargs(__UpperCAmelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Any = cls.get_config_dict(__UpperCAmelCase ,**__UpperCAmelCase )
# get the vision config dict if we are loading from Pix2StructConfig
if config_dict.get("""model_type""" ) == "pix2struct":
lowerCAmelCase__ : Optional[Any] = config_dict["""vision_config"""]
if "model_type" in config_dict and hasattr(cls ,"""model_type""" ) and config_dict["model_type"] != cls.model_type:
logger.warning(
F"""You are using a model of type {config_dict['model_type']} to instantiate a model of type """
F"""{cls.model_type}. This is not supported for all configurations of models and can yield errors.""" )
return cls.from_dict(__UpperCAmelCase ,**__UpperCAmelCase )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Union[str, Any] = '''pix2struct'''
__lowercase : str = True
def __init__( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=1.0 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=True ,**__UpperCAmelCase ,) -> Tuple:
super().__init__(tie_word_embeddings=__UpperCAmelCase ,is_encoder_decoder=__UpperCAmelCase ,**__UpperCAmelCase )
if text_config is None:
lowerCAmelCase__ : Union[str, Any] = {}
logger.info("""text_config is None. Initializing the Pix2StructTextConfig with default values.""" )
if vision_config is None:
lowerCAmelCase__ : List[Any] = {}
logger.info("""vision_config is None. Initializing the Pix2StructVisionConfig with default values.""" )
lowerCAmelCase__ : Optional[Any] = PixaStructTextConfig(**__UpperCAmelCase )
lowerCAmelCase__ : Any = PixaStructVisionConfig(**__UpperCAmelCase )
lowerCAmelCase__ : Tuple = self.text_config.decoder_start_token_id
lowerCAmelCase__ : List[str] = self.text_config.pad_token_id
lowerCAmelCase__ : Any = self.text_config.eos_token_id
lowerCAmelCase__ : Dict = initializer_factor
lowerCAmelCase__ : List[str] = initializer_range
lowerCAmelCase__ : int = self.initializer_range
lowerCAmelCase__ : Tuple = self.initializer_range
lowerCAmelCase__ : Optional[int] = is_vqa
@classmethod
def UpperCAmelCase_ ( cls ,__UpperCAmelCase ,__UpperCAmelCase ,**__UpperCAmelCase ) -> Optional[int]:
return cls(text_config=text_config.to_dict() ,vision_config=vision_config.to_dict() ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : str = copy.deepcopy(self.__dict__ )
lowerCAmelCase__ : Any = self.text_config.to_dict()
lowerCAmelCase__ : Optional[Any] = self.vision_config.to_dict()
lowerCAmelCase__ : Any = self.__class__.model_type
return output
| 37
|
'''simple docstring'''
from collections import UserDict
from typing import List, Union
from ..utils import (
add_end_docstrings,
is_tf_available,
is_torch_available,
is_vision_available,
logging,
requires_backends,
)
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
from ..tf_utils import stable_softmax
_lowerCAmelCase = logging.get_logger(__name__)
@add_end_docstrings(SCREAMING_SNAKE_CASE_ )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,**__UpperCAmelCase ) -> Tuple:
super().__init__(**__UpperCAmelCase )
requires_backends(self ,"""vision""" )
self.check_model_type(
TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if self.framework == """tf"""
else MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING )
def __call__( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> str:
return super().__call__(__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[Any] = {}
if "candidate_labels" in kwargs:
lowerCAmelCase__ : int = kwargs["""candidate_labels"""]
if "hypothesis_template" in kwargs:
lowerCAmelCase__ : Optional[int] = kwargs["""hypothesis_template"""]
return preprocess_params, {}, {}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase="This is a photo of {}." ) -> int:
lowerCAmelCase__ : str = load_image(__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.image_processor(images=[image] ,return_tensors=self.framework )
lowerCAmelCase__ : List[Any] = candidate_labels
lowerCAmelCase__ : List[str] = [hypothesis_template.format(__UpperCAmelCase ) for x in candidate_labels]
lowerCAmelCase__ : Optional[Any] = self.tokenizer(__UpperCAmelCase ,return_tensors=self.framework ,padding=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = [text_inputs]
return inputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = model_inputs.pop("""candidate_labels""" )
lowerCAmelCase__ : Union[str, Any] = model_inputs.pop("""text_inputs""" )
if isinstance(text_inputs[0] ,__UpperCAmelCase ):
lowerCAmelCase__ : int = text_inputs[0]
else:
# Batching case.
lowerCAmelCase__ : Dict = text_inputs[0][0]
lowerCAmelCase__ : Any = self.model(**__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = {
"""candidate_labels""": candidate_labels,
"""logits""": outputs.logits_per_image,
}
return model_outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Union[str, Any] = model_outputs.pop("""candidate_labels""" )
lowerCAmelCase__ : List[str] = model_outputs["""logits"""][0]
if self.framework == "pt":
lowerCAmelCase__ : List[str] = logits.softmax(dim=-1 ).squeeze(-1 )
lowerCAmelCase__ : Optional[Any] = probs.tolist()
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Dict = [scores]
elif self.framework == "tf":
lowerCAmelCase__ : Any = stable_softmax(__UpperCAmelCase ,axis=-1 )
lowerCAmelCase__ : List[Any] = probs.numpy().tolist()
else:
raise ValueError(F"""Unsupported framework: {self.framework}""" )
lowerCAmelCase__ : Tuple = [
{"""score""": score, """label""": candidate_label}
for score, candidate_label in sorted(zip(__UpperCAmelCase ,__UpperCAmelCase ) ,key=lambda __UpperCAmelCase : -x[0] )
]
return result
| 37
| 1
|
'''simple docstring'''
import unittest
import numpy as np
import torch
from diffusers import VersatileDiffusionImageVariationPipeline
from diffusers.utils.testing_utils import load_image, require_torch_gpu, slow, torch_device
_lowerCAmelCase = False
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
pass
@slow
@require_torch_gpu
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = VersatileDiffusionImageVariationPipeline.from_pretrained("""shi-labs/versatile-diffusion""" )
pipe.to(__UpperCAmelCase )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/versatile_diffusion/benz.jpg""" )
lowerCAmelCase__ : Tuple = torch.manual_seed(0 )
lowerCAmelCase__ : Dict = pipe(
image=__UpperCAmelCase ,generator=__UpperCAmelCase ,guidance_scale=7.5 ,num_inference_steps=50 ,output_type="""numpy""" ,).images
lowerCAmelCase__ : List[Any] = image[0, 253:256, 253:256, -1]
assert image.shape == (1, 512, 512, 3)
lowerCAmelCase__ : Tuple = np.array([0.0_4_4_1, 0.0_4_6_9, 0.0_5_0_7, 0.0_5_7_5, 0.0_6_3_2, 0.0_6_5_0, 0.0_8_6_5, 0.0_9_0_9, 0.0_9_4_5] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
| 37
|
'''simple docstring'''
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , )
@pytest.mark.usefixtures('''sm_env''' )
@parameterized_class(
[
{
'''framework''': '''pytorch''',
'''script''': '''run_glue.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_5_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''pytorch''',
'''script''': '''run_ddp.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''tensorflow''',
'''script''': '''run_tf_dist.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.6, '''eval_loss''': 0.7},
},
] )
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.framework == "pytorch":
subprocess.run(
F"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() ,encoding="""utf-8""" ,check=__UpperCAmelCase ,)
assert hasattr(self ,"""env""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = F"""{self.env.base_job_name}-{instance_count}-{'ddp' if 'ddp' in self.script else 'smd'}"""
# distributed data settings
lowerCAmelCase__ : Any = {"""smdistributed""": {"""dataparallel""": {"""enabled""": True}}} if self.script != """run_ddp.py""" else None
# creates estimator
return HuggingFace(
entry_point=self.script ,source_dir=self.env.test_path ,role=self.env.role ,image_uri=self.env.image_uri ,base_job_name=__UpperCAmelCase ,instance_count=__UpperCAmelCase ,instance_type=self.instance_type ,debugger_hook_config=__UpperCAmelCase ,hyperparameters={**self.env.distributed_hyperparameters, """model_name_or_path""": self.model_name_or_path} ,metric_definitions=self.env.metric_definitions ,distribution=__UpperCAmelCase ,py_version="""py36""" ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
TrainingJobAnalytics(__UpperCAmelCase ).export_csv(F"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(2,)] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
# create estimator
lowerCAmelCase__ : List[Any] = self.create_estimator(__UpperCAmelCase )
# run training
estimator.fit()
# result dataframe
lowerCAmelCase__ : Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
lowerCAmelCase__ : int = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
lowerCAmelCase__ : List[Any] = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
lowerCAmelCase__ : List[str] = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" ,99_9999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(F"""{estimator.latest_training_job.name}.json""" ,"""w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} ,__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
from typing import Dict, List, Optional
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''nielsr/canine-s''': 2048,
}
# Unicode defines 1,114,112 total “codepoints”
_lowerCAmelCase = 111_4112
# Below: Constants defining canonical codepoints for special, pseudo-characters.
# Copied from https://github.com/google-research/language/blob/master/language/canine/special_codepoints.py
_lowerCAmelCase = 0
_lowerCAmelCase = 0xe_000
_lowerCAmelCase = 0xe_001
_lowerCAmelCase = 0xe_002
_lowerCAmelCase = 0xe_003
_lowerCAmelCase = 0xe_004
# Maps special codepoints to human-readable names.
_lowerCAmelCase = {
# Special symbols are represented using codepoints values that are valid,
# but designated as "Private Use", meaning that they will never be assigned
# characters by the Unicode Consortium, and are thus safe for use here.
#
# NOTE: Do *NOT* add any sort of [UNK_CHAR] here. They are explicitly
# excluded and should fail with a hard error.
CLS: "[CLS]",
SEP: "[SEP]",
BOS: "[BOS]",
MASK: "[MASK]",
PAD: "[PAD]",
RESERVED: "[RESERVED]",
}
# Maps special codepoint human-readable names to their codepoint values.
_lowerCAmelCase = {name: codepoint for codepoint, name in SPECIAL_CODEPOINTS.items()}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,__UpperCAmelCase=chr(__UpperCAmelCase ) ,__UpperCAmelCase=chr(__UpperCAmelCase ) ,__UpperCAmelCase=chr(__UpperCAmelCase ) ,__UpperCAmelCase=chr(__UpperCAmelCase ) ,__UpperCAmelCase=chr(__UpperCAmelCase ) ,__UpperCAmelCase=chr(__UpperCAmelCase ) ,__UpperCAmelCase=False ,__UpperCAmelCase=2048 ,**__UpperCAmelCase ,) -> Optional[int]:
lowerCAmelCase__ : Optional[int] = AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else bos_token
lowerCAmelCase__ : Dict = AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else eos_token
lowerCAmelCase__ : Union[str, Any] = AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else sep_token
lowerCAmelCase__ : int = AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else cls_token
lowerCAmelCase__ : int = AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else pad_token
# Mask token behave like a normal word, i.e. include the space before it
lowerCAmelCase__ : str = AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ) if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else mask_token
super().__init__(
bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,model_max_length=__UpperCAmelCase ,**__UpperCAmelCase ,)
# Creates a mapping for looking up the IDs of special symbols.
lowerCAmelCase__ : Dict[str, int] = {}
for codepoint, name in SPECIAL_CODEPOINTS.items():
lowerCAmelCase__ : Tuple = codepoint
# Creates a mapping for looking up the string forms of special symbol IDs.
lowerCAmelCase__ : Dict[int, str] = {
codepoint: name for name, codepoint in self._special_codepoints.items()
}
lowerCAmelCase__ : Optional[Any] = UNICODE_VOCAB_SIZE
lowerCAmelCase__ : Any = len(self._special_codepoints )
@property
def UpperCAmelCase_ ( self ) -> int:
return self._unicode_vocab_size
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return list(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
try:
return ord(__UpperCAmelCase )
except TypeError:
raise ValueError(F"""invalid token: '{token}'""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
try:
if index in SPECIAL_CODEPOINTS:
return SPECIAL_CODEPOINTS[index]
return chr(__UpperCAmelCase )
except TypeError:
raise ValueError(F"""invalid id: {index}""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return "".join(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : str = [self.sep_token_id]
lowerCAmelCase__ : Optional[int] = [self.cls_token_id]
lowerCAmelCase__ : str = cls + token_ids_a + sep
if token_ids_a is not None:
result += token_ids_a + sep
return result
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=__UpperCAmelCase ,token_ids_a=__UpperCAmelCase ,already_has_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = [1] + ([0] * len(__UpperCAmelCase )) + [1]
if token_ids_a is not None:
result += ([0] * len(__UpperCAmelCase )) + [1]
return result
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : Optional[int] = [self.sep_token_id]
lowerCAmelCase__ : List[str] = [self.cls_token_id]
lowerCAmelCase__ : Optional[int] = len(cls + token_ids_a + sep ) * [0]
if token_ids_a is not None:
result += len(token_ids_a + sep ) * [1]
return result
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> int:
return ()
| 37
|
'''simple docstring'''
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''albert-base-v1''': '''https://huggingface.co/albert-base-v1/resolve/main/spiece.model''',
'''albert-large-v1''': '''https://huggingface.co/albert-large-v1/resolve/main/spiece.model''',
'''albert-xlarge-v1''': '''https://huggingface.co/albert-xlarge-v1/resolve/main/spiece.model''',
'''albert-xxlarge-v1''': '''https://huggingface.co/albert-xxlarge-v1/resolve/main/spiece.model''',
'''albert-base-v2''': '''https://huggingface.co/albert-base-v2/resolve/main/spiece.model''',
'''albert-large-v2''': '''https://huggingface.co/albert-large-v2/resolve/main/spiece.model''',
'''albert-xlarge-v2''': '''https://huggingface.co/albert-xlarge-v2/resolve/main/spiece.model''',
'''albert-xxlarge-v2''': '''https://huggingface.co/albert-xxlarge-v2/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''albert-base-v1''': 512,
'''albert-large-v1''': 512,
'''albert-xlarge-v1''': 512,
'''albert-xxlarge-v1''': 512,
'''albert-base-v2''': 512,
'''albert-large-v2''': 512,
'''albert-xlarge-v2''': 512,
'''albert-xxlarge-v2''': 512,
}
_lowerCAmelCase = '''▁'''
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Any = VOCAB_FILES_NAMES
__lowercase : Optional[Any] = PRETRAINED_VOCAB_FILES_MAP
__lowercase : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=True ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[MASK]" ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
# Mask token behave like a normal word, i.e. include the space before it and
# is included in the raw text, there should be a match in a non-normalized sentence.
lowerCAmelCase__ : Tuple = (
AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ,normalized=__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase )
else mask_token
)
lowerCAmelCase__ : Any = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : str = do_lower_case
lowerCAmelCase__ : int = remove_space
lowerCAmelCase__ : Tuple = keep_accents
lowerCAmelCase__ : Any = vocab_file
lowerCAmelCase__ : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> Optional[int]:
return len(self.sp_model )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Union[str, Any] = {self.convert_ids_to_tokens(__UpperCAmelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = self.__dict__.copy()
lowerCAmelCase__ : Optional[Any] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Union[str, Any] = {}
lowerCAmelCase__ : Any = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[int]:
if self.remove_space:
lowerCAmelCase__ : int = """ """.join(inputs.strip().split() )
else:
lowerCAmelCase__ : str = inputs
lowerCAmelCase__ : Tuple = outputs.replace("""``""" ,"""\"""" ).replace("""''""" ,"""\"""" )
if not self.keep_accents:
lowerCAmelCase__ : Any = unicodedata.normalize("""NFKD""" ,__UpperCAmelCase )
lowerCAmelCase__ : Dict = """""".join([c for c in outputs if not unicodedata.combining(__UpperCAmelCase )] )
if self.do_lower_case:
lowerCAmelCase__ : Tuple = outputs.lower()
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
lowerCAmelCase__ : str = []
for piece in pieces:
if len(__UpperCAmelCase ) > 1 and piece[-1] == str(""",""" ) and piece[-2].isdigit():
lowerCAmelCase__ : List[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(__UpperCAmelCase ,"""""" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
lowerCAmelCase__ : str = cur_pieces[1:]
else:
lowerCAmelCase__ : int = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(__UpperCAmelCase )
else:
new_pieces.append(__UpperCAmelCase )
return new_pieces
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
return self.sp_model.IdToPiece(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : Tuple = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : List[str] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string.strip()
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : int = [self.sep_token_id]
lowerCAmelCase__ : Dict = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
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 not None:
return [1] + ([0] * len(__UpperCAmelCase )) + [1] + ([0] * len(__UpperCAmelCase )) + [1]
return [1] + ([0] * len(__UpperCAmelCase )) + [1]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : List[str] = [self.sep_token_id]
lowerCAmelCase__ : Tuple = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : int = 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:
lowerCAmelCase__ : List[Any] = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
| 37
| 1
|
'''simple docstring'''
import argparse
import json
from pathlib import Path
import requests
import torch
from huggingface_hub import cached_download, hf_hub_download, hf_hub_url
from PIL import Image
from transformers import DetaConfig, DetaForObjectDetection, DetaImageProcessor, SwinConfig
from transformers.utils import logging
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = SwinConfig(
embed_dim=192 , depths=(2, 2, 18, 2) , num_heads=(6, 12, 24, 48) , window_size=12 , out_features=["""stage2""", """stage3""", """stage4"""] , )
lowerCAmelCase__ : Union[str, Any] = DetaConfig(
backbone_config=UpperCamelCase , num_queries=900 , encoder_ffn_dim=2048 , decoder_ffn_dim=2048 , num_feature_levels=5 , assign_first_stage=UpperCamelCase , with_box_refine=UpperCamelCase , two_stage=UpperCamelCase , )
# set labels
lowerCAmelCase__ : int = """huggingface/label-files"""
if "o365" in model_name:
lowerCAmelCase__ : Optional[int] = 366
lowerCAmelCase__ : int = """object365-id2label.json"""
else:
lowerCAmelCase__ : List[str] = 91
lowerCAmelCase__ : int = """coco-detection-id2label.json"""
lowerCAmelCase__ : Union[str, Any] = num_labels
lowerCAmelCase__ : Tuple = json.load(open(cached_download(hf_hub_url(UpperCamelCase , UpperCamelCase , repo_type="""dataset""" ) ) , """r""" ) )
lowerCAmelCase__ : List[str] = {int(UpperCamelCase ): v for k, v in idalabel.items()}
lowerCAmelCase__ : Union[str, Any] = idalabel
lowerCAmelCase__ : Optional[Any] = {v: k for k, v in idalabel.items()}
return config
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = []
# stem
# fmt: off
rename_keys.append(("""backbone.0.body.patch_embed.proj.weight""", """model.backbone.model.embeddings.patch_embeddings.projection.weight""") )
rename_keys.append(("""backbone.0.body.patch_embed.proj.bias""", """model.backbone.model.embeddings.patch_embeddings.projection.bias""") )
rename_keys.append(("""backbone.0.body.patch_embed.norm.weight""", """model.backbone.model.embeddings.norm.weight""") )
rename_keys.append(("""backbone.0.body.patch_embed.norm.bias""", """model.backbone.model.embeddings.norm.bias""") )
# stages
for i in range(len(config.backbone_config.depths ) ):
for j in range(config.backbone_config.depths[i] ):
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.norm1.weight""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_before.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.norm1.bias""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_before.bias""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.attn.relative_position_bias_table""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_bias_table""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.attn.relative_position_index""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_index""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.attn.proj.weight""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.output.dense.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.attn.proj.bias""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.attention.output.dense.bias""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.norm2.weight""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_after.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.norm2.bias""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.layernorm_after.bias""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.mlp.fc1.weight""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.intermediate.dense.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.mlp.fc1.bias""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.intermediate.dense.bias""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.mlp.fc2.weight""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.output.dense.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.blocks.{j}.mlp.fc2.bias""", f"""model.backbone.model.encoder.layers.{i}.blocks.{j}.output.dense.bias""") )
if i < 3:
rename_keys.append((f"""backbone.0.body.layers.{i}.downsample.reduction.weight""", f"""model.backbone.model.encoder.layers.{i}.downsample.reduction.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.downsample.norm.weight""", f"""model.backbone.model.encoder.layers.{i}.downsample.norm.weight""") )
rename_keys.append((f"""backbone.0.body.layers.{i}.downsample.norm.bias""", f"""model.backbone.model.encoder.layers.{i}.downsample.norm.bias""") )
rename_keys.append(("""backbone.0.body.norm1.weight""", """model.backbone.model.hidden_states_norms.stage2.weight""") )
rename_keys.append(("""backbone.0.body.norm1.bias""", """model.backbone.model.hidden_states_norms.stage2.bias""") )
rename_keys.append(("""backbone.0.body.norm2.weight""", """model.backbone.model.hidden_states_norms.stage3.weight""") )
rename_keys.append(("""backbone.0.body.norm2.bias""", """model.backbone.model.hidden_states_norms.stage3.bias""") )
rename_keys.append(("""backbone.0.body.norm3.weight""", """model.backbone.model.hidden_states_norms.stage4.weight""") )
rename_keys.append(("""backbone.0.body.norm3.bias""", """model.backbone.model.hidden_states_norms.stage4.bias""") )
# transformer encoder
for i in range(config.encoder_layers ):
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.sampling_offsets.weight""", f"""model.encoder.layers.{i}.self_attn.sampling_offsets.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.sampling_offsets.bias""", f"""model.encoder.layers.{i}.self_attn.sampling_offsets.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.attention_weights.weight""", f"""model.encoder.layers.{i}.self_attn.attention_weights.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.attention_weights.bias""", f"""model.encoder.layers.{i}.self_attn.attention_weights.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.value_proj.weight""", f"""model.encoder.layers.{i}.self_attn.value_proj.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.value_proj.bias""", f"""model.encoder.layers.{i}.self_attn.value_proj.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.output_proj.weight""", f"""model.encoder.layers.{i}.self_attn.output_proj.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.self_attn.output_proj.bias""", f"""model.encoder.layers.{i}.self_attn.output_proj.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.norm1.weight""", f"""model.encoder.layers.{i}.self_attn_layer_norm.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.norm1.bias""", f"""model.encoder.layers.{i}.self_attn_layer_norm.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.linear1.weight""", f"""model.encoder.layers.{i}.fc1.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.linear1.bias""", f"""model.encoder.layers.{i}.fc1.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.linear2.weight""", f"""model.encoder.layers.{i}.fc2.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.linear2.bias""", f"""model.encoder.layers.{i}.fc2.bias""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.norm2.weight""", f"""model.encoder.layers.{i}.final_layer_norm.weight""") )
rename_keys.append((f"""transformer.encoder.layers.{i}.norm2.bias""", f"""model.encoder.layers.{i}.final_layer_norm.bias""") )
# transformer decoder
for i in range(config.decoder_layers ):
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.sampling_offsets.weight""", f"""model.decoder.layers.{i}.encoder_attn.sampling_offsets.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.sampling_offsets.bias""", f"""model.decoder.layers.{i}.encoder_attn.sampling_offsets.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.attention_weights.weight""", f"""model.decoder.layers.{i}.encoder_attn.attention_weights.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.attention_weights.bias""", f"""model.decoder.layers.{i}.encoder_attn.attention_weights.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.value_proj.weight""", f"""model.decoder.layers.{i}.encoder_attn.value_proj.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.value_proj.bias""", f"""model.decoder.layers.{i}.encoder_attn.value_proj.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.output_proj.weight""", f"""model.decoder.layers.{i}.encoder_attn.output_proj.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.cross_attn.output_proj.bias""", f"""model.decoder.layers.{i}.encoder_attn.output_proj.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.norm1.weight""", f"""model.decoder.layers.{i}.encoder_attn_layer_norm.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.norm1.bias""", f"""model.decoder.layers.{i}.encoder_attn_layer_norm.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.self_attn.out_proj.weight""", f"""model.decoder.layers.{i}.self_attn.out_proj.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.self_attn.out_proj.bias""", f"""model.decoder.layers.{i}.self_attn.out_proj.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.norm2.weight""", f"""model.decoder.layers.{i}.self_attn_layer_norm.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.norm2.bias""", f"""model.decoder.layers.{i}.self_attn_layer_norm.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.linear1.weight""", f"""model.decoder.layers.{i}.fc1.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.linear1.bias""", f"""model.decoder.layers.{i}.fc1.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.linear2.weight""", f"""model.decoder.layers.{i}.fc2.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.linear2.bias""", f"""model.decoder.layers.{i}.fc2.bias""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.norm3.weight""", f"""model.decoder.layers.{i}.final_layer_norm.weight""") )
rename_keys.append((f"""transformer.decoder.layers.{i}.norm3.bias""", f"""model.decoder.layers.{i}.final_layer_norm.bias""") )
# fmt: on
return rename_keys
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = dct.pop(UpperCamelCase )
lowerCAmelCase__ : Dict = val
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = [int(backbone_config.embed_dim * 2**i ) for i in range(len(backbone_config.depths ) )]
for i in range(len(backbone_config.depths ) ):
lowerCAmelCase__ : Any = num_features[i]
for j in range(backbone_config.depths[i] ):
# fmt: off
# read in weights + bias of input projection layer (in original implementation, this is a single matrix + bias)
lowerCAmelCase__ : int = state_dict.pop(f"""backbone.0.body.layers.{i}.blocks.{j}.attn.qkv.weight""" )
lowerCAmelCase__ : Dict = state_dict.pop(f"""backbone.0.body.layers.{i}.blocks.{j}.attn.qkv.bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Dict = in_proj_weight[:dim, :]
lowerCAmelCase__ : str = in_proj_bias[: dim]
lowerCAmelCase__ : int = in_proj_weight[
dim : dim * 2, :
]
lowerCAmelCase__ : Union[str, Any] = in_proj_bias[
dim : dim * 2
]
lowerCAmelCase__ : Tuple = in_proj_weight[
-dim :, :
]
lowerCAmelCase__ : Any = in_proj_bias[-dim :]
# fmt: on
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = config.d_model
for i in range(config.decoder_layers ):
# read in weights + bias of input projection layer of self-attention
lowerCAmelCase__ : List[str] = state_dict.pop(f"""transformer.decoder.layers.{i}.self_attn.in_proj_weight""" )
lowerCAmelCase__ : Union[str, Any] = state_dict.pop(f"""transformer.decoder.layers.{i}.self_attn.in_proj_bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Union[str, Any] = in_proj_weight[:hidden_size, :]
lowerCAmelCase__ : Dict = in_proj_bias[:hidden_size]
lowerCAmelCase__ : Tuple = in_proj_weight[
hidden_size : hidden_size * 2, :
]
lowerCAmelCase__ : Optional[Any] = in_proj_bias[hidden_size : hidden_size * 2]
lowerCAmelCase__ : Tuple = in_proj_weight[-hidden_size:, :]
lowerCAmelCase__ : str = in_proj_bias[-hidden_size:]
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : str = """http://images.cocodataset.org/val2017/000000039769.jpg"""
lowerCAmelCase__ : Optional[Any] = Image.open(requests.get(UpperCamelCase , stream=UpperCamelCase ).raw )
return im
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = get_deta_config(UpperCamelCase )
# load original state dict
if model_name == "deta-swin-large":
lowerCAmelCase__ : Union[str, Any] = hf_hub_download(repo_id="""nielsr/deta-checkpoints""" , filename="""adet_swin_ft.pth""" )
elif model_name == "deta-swin-large-o365":
lowerCAmelCase__ : List[Any] = hf_hub_download(repo_id="""jozhang97/deta-swin-l-o365""" , filename="""deta_swin_pt_o365.pth""" )
else:
raise ValueError(f"""Model name {model_name} not supported""" )
lowerCAmelCase__ : str = torch.load(UpperCamelCase , map_location="""cpu""" )["""model"""]
# original state dict
for name, param in state_dict.items():
print(UpperCamelCase , param.shape )
# rename keys
lowerCAmelCase__ : Dict = create_rename_keys(UpperCamelCase )
for src, dest in rename_keys:
rename_key(UpperCamelCase , UpperCamelCase , UpperCamelCase )
read_in_swin_q_k_v(UpperCamelCase , config.backbone_config )
read_in_decoder_q_k_v(UpperCamelCase , UpperCamelCase )
# fix some prefixes
for key in state_dict.copy().keys():
if "transformer.decoder.class_embed" in key or "transformer.decoder.bbox_embed" in key:
lowerCAmelCase__ : Dict = state_dict.pop(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = val
if "input_proj" in key:
lowerCAmelCase__ : str = state_dict.pop(UpperCamelCase )
lowerCAmelCase__ : Any = val
if "level_embed" in key or "pos_trans" in key or "pix_trans" in key or "enc_output" in key:
lowerCAmelCase__ : Any = state_dict.pop(UpperCamelCase )
lowerCAmelCase__ : Tuple = val
# finally, create HuggingFace model and load state dict
lowerCAmelCase__ : Dict = DetaForObjectDetection(UpperCamelCase )
model.load_state_dict(UpperCamelCase )
model.eval()
lowerCAmelCase__ : Optional[Any] = """cuda""" if torch.cuda.is_available() else """cpu"""
model.to(UpperCamelCase )
# load image processor
lowerCAmelCase__ : List[Any] = DetaImageProcessor(format="""coco_detection""" )
# verify our conversion on image
lowerCAmelCase__ : List[Any] = prepare_img()
lowerCAmelCase__ : Optional[Any] = processor(images=UpperCamelCase , return_tensors="""pt""" )
lowerCAmelCase__ : int = encoding["""pixel_values"""]
lowerCAmelCase__ : List[Any] = model(pixel_values.to(UpperCamelCase ) )
# verify logits
print("""Logits:""" , outputs.logits[0, :3, :3] )
print("""Boxes:""" , outputs.pred_boxes[0, :3, :3] )
if model_name == "deta-swin-large":
lowerCAmelCase__ : Any = torch.tensor(
[[-7.6308, -2.8485, -5.3737], [-7.2037, -4.5505, -4.8027], [-7.2943, -4.2611, -4.6617]] )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0.4987, 0.4969, 0.9999], [0.2549, 0.5498, 0.4805], [0.5498, 0.2757, 0.0569]] )
elif model_name == "deta-swin-large-o365":
lowerCAmelCase__ : int = torch.tensor(
[[-8.0122, -3.5720, -4.9717], [-8.1547, -3.6886, -4.6389], [-7.6610, -3.6194, -5.0134]] )
lowerCAmelCase__ : Union[str, Any] = torch.tensor([[0.2523, 0.5549, 0.4881], [0.7715, 0.4149, 0.4601], [0.5503, 0.2753, 0.0575]] )
assert torch.allclose(outputs.logits[0, :3, :3] , expected_logits.to(UpperCamelCase ) , atol=1e-4 )
assert torch.allclose(outputs.pred_boxes[0, :3, :3] , expected_boxes.to(UpperCamelCase ) , atol=1e-4 )
print("""Everything ok!""" )
if pytorch_dump_folder_path:
# Save model and processor
logger.info(f"""Saving PyTorch model and processor to {pytorch_dump_folder_path}...""" )
Path(UpperCamelCase ).mkdir(exist_ok=UpperCamelCase )
model.save_pretrained(UpperCamelCase )
processor.save_pretrained(UpperCamelCase )
# Push to hub
if push_to_hub:
print("""Pushing model and processor to hub...""" )
model.push_to_hub(f"""jozhang97/{model_name}""" )
processor.push_to_hub(f"""jozhang97/{model_name}""" )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--model_name''',
type=str,
default='''deta-swin-large''',
choices=['''deta-swin-large''', '''deta-swin-large-o365'''],
help='''Name of the model you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''',
default=None,
type=str,
help='''Path to the folder to output PyTorch model.''',
)
parser.add_argument(
'''--push_to_hub''', action='''store_true''', help='''Whether or not to push the converted model to the 🤗 hub.'''
)
_lowerCAmelCase = parser.parse_args()
convert_deta_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 37
|
'''simple docstring'''
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''merges_file''': '''merges.txt''',
}
_lowerCAmelCase = {
'''vocab_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json'''},
'''merges_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt'''},
}
_lowerCAmelCase = {
'''ctrl''': 256,
}
_lowerCAmelCase = {
'''Pregnancy''': 16_8629,
'''Christianity''': 7675,
'''Explain''': 10_6423,
'''Fitness''': 6_3440,
'''Saving''': 6_3163,
'''Ask''': 2_7171,
'''Ass''': 9_5985,
'''Joke''': 16_3509,
'''Questions''': 4_5622,
'''Thoughts''': 4_9605,
'''Retail''': 5_2342,
'''Feminism''': 16_4338,
'''Writing''': 1_1992,
'''Atheism''': 19_2263,
'''Netflix''': 4_8616,
'''Computing''': 3_9639,
'''Opinion''': 4_3213,
'''Alone''': 4_4967,
'''Funny''': 5_8917,
'''Gaming''': 4_0358,
'''Human''': 4088,
'''India''': 1331,
'''Joker''': 7_7138,
'''Diet''': 3_6206,
'''Legal''': 1_1859,
'''Norman''': 4939,
'''Tip''': 7_2689,
'''Weight''': 5_2343,
'''Movies''': 4_6273,
'''Running''': 2_3425,
'''Science''': 2090,
'''Horror''': 3_7793,
'''Confession''': 6_0572,
'''Finance''': 1_2250,
'''Politics''': 1_6360,
'''Scary''': 19_1985,
'''Support''': 1_2654,
'''Technologies''': 3_2516,
'''Teenage''': 6_6160,
'''Event''': 3_2769,
'''Learned''': 6_7460,
'''Notion''': 18_2770,
'''Wikipedia''': 3_7583,
'''Books''': 6665,
'''Extract''': 7_6050,
'''Confessions''': 10_2701,
'''Conspiracy''': 7_5932,
'''Links''': 6_3674,
'''Narcissus''': 15_0425,
'''Relationship''': 5_4766,
'''Relationships''': 13_4796,
'''Reviews''': 4_1671,
'''News''': 4256,
'''Translation''': 2_6820,
'''multilingual''': 12_8406,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = set()
lowerCAmelCase__ : str = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCAmelCase__ : List[Any] = char
lowerCAmelCase__ : Optional[Any] = set(UpperCamelCase )
return pairs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Any = CONTROL_CODES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<unk>" ,**__UpperCAmelCase ) -> Optional[int]:
super().__init__(unk_token=__UpperCAmelCase ,**__UpperCAmelCase )
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as vocab_handle:
lowerCAmelCase__ : int = json.load(__UpperCAmelCase )
lowerCAmelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as merges_handle:
lowerCAmelCase__ : Union[str, Any] = merges_handle.read().split("""\n""" )[1:-1]
lowerCAmelCase__ : Optional[Any] = [tuple(merge.split() ) for merge in merges]
lowerCAmelCase__ : int = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return len(self.encoder )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return dict(self.encoder ,**self.added_tokens_encoder )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if token in self.cache:
return self.cache[token]
lowerCAmelCase__ : Optional[Any] = tuple(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tuple(list(word[:-1] ) + [word[-1] + """</w>"""] )
lowerCAmelCase__ : Any = get_pairs(__UpperCAmelCase )
if not pairs:
return token
while True:
lowerCAmelCase__ : List[str] = min(__UpperCAmelCase ,key=lambda __UpperCAmelCase : self.bpe_ranks.get(__UpperCAmelCase ,float("""inf""" ) ) )
if bigram not in self.bpe_ranks:
break
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = bigram
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : Dict = 0
while i < len(__UpperCAmelCase ):
try:
lowerCAmelCase__ : Optional[Any] = word.index(__UpperCAmelCase ,__UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCAmelCase__ : Dict = 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
lowerCAmelCase__ : Tuple = tuple(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = new_word
if len(__UpperCAmelCase ) == 1:
break
else:
lowerCAmelCase__ : Dict = get_pairs(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """@@ """.join(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = word[:-4]
lowerCAmelCase__ : Optional[Any] = word
return word
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : Any = re.findall(R"""\S+\n?""" ,__UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(__UpperCAmelCase ).split(""" """ ) ) )
return split_tokens
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
return self.encoder.get(__UpperCAmelCase ,self.encoder.get(self.unk_token ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = """ """.join(__UpperCAmelCase ).replace("""@@ """ ,"""""" ).strip()
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : List[Any] = os.path.join(
__UpperCAmelCase ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : Optional[int] = 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""" )
lowerCAmelCase__ : int = 0
with open(__UpperCAmelCase ,"""w""" ,encoding="""utf-8""" ) as writer:
writer.write("""#version: 0.2\n""" )
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 {merge_file}: BPE merge indices are not consecutive."""
""" Please check that the tokenizer is not corrupted!""" )
lowerCAmelCase__ : Dict = token_index
writer.write(""" """.join(__UpperCAmelCase ) + """\n""" )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 37
| 1
|
'''simple docstring'''
import argparse
from collections import OrderedDict
from pathlib import Path
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from torchvision.transforms import functional as F
from transformers import DetrImageProcessor, TableTransformerConfig, TableTransformerForObjectDetection
from transformers.utils import logging
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger(__name__)
# here we list all keys to be renamed (original name on the left, our name on the right)
_lowerCAmelCase = []
for i in range(6):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append(
(F"""transformer.encoder.layers.{i}.self_attn.out_proj.weight""", F"""encoder.layers.{i}.self_attn.out_proj.weight""")
)
rename_keys.append(
(F"""transformer.encoder.layers.{i}.self_attn.out_proj.bias""", F"""encoder.layers.{i}.self_attn.out_proj.bias""")
)
rename_keys.append((F"""transformer.encoder.layers.{i}.linear1.weight""", F"""encoder.layers.{i}.fc1.weight"""))
rename_keys.append((F"""transformer.encoder.layers.{i}.linear1.bias""", F"""encoder.layers.{i}.fc1.bias"""))
rename_keys.append((F"""transformer.encoder.layers.{i}.linear2.weight""", F"""encoder.layers.{i}.fc2.weight"""))
rename_keys.append((F"""transformer.encoder.layers.{i}.linear2.bias""", F"""encoder.layers.{i}.fc2.bias"""))
rename_keys.append(
(F"""transformer.encoder.layers.{i}.norm1.weight""", F"""encoder.layers.{i}.self_attn_layer_norm.weight""")
)
rename_keys.append((F"""transformer.encoder.layers.{i}.norm1.bias""", F"""encoder.layers.{i}.self_attn_layer_norm.bias"""))
rename_keys.append((F"""transformer.encoder.layers.{i}.norm2.weight""", F"""encoder.layers.{i}.final_layer_norm.weight"""))
rename_keys.append((F"""transformer.encoder.layers.{i}.norm2.bias""", F"""encoder.layers.{i}.final_layer_norm.bias"""))
# decoder layers: 2 times output projection, 2 feedforward neural networks and 3 layernorms
rename_keys.append(
(F"""transformer.decoder.layers.{i}.self_attn.out_proj.weight""", F"""decoder.layers.{i}.self_attn.out_proj.weight""")
)
rename_keys.append(
(F"""transformer.decoder.layers.{i}.self_attn.out_proj.bias""", F"""decoder.layers.{i}.self_attn.out_proj.bias""")
)
rename_keys.append(
(
F"""transformer.decoder.layers.{i}.multihead_attn.out_proj.weight""",
F"""decoder.layers.{i}.encoder_attn.out_proj.weight""",
)
)
rename_keys.append(
(
F"""transformer.decoder.layers.{i}.multihead_attn.out_proj.bias""",
F"""decoder.layers.{i}.encoder_attn.out_proj.bias""",
)
)
rename_keys.append((F"""transformer.decoder.layers.{i}.linear1.weight""", F"""decoder.layers.{i}.fc1.weight"""))
rename_keys.append((F"""transformer.decoder.layers.{i}.linear1.bias""", F"""decoder.layers.{i}.fc1.bias"""))
rename_keys.append((F"""transformer.decoder.layers.{i}.linear2.weight""", F"""decoder.layers.{i}.fc2.weight"""))
rename_keys.append((F"""transformer.decoder.layers.{i}.linear2.bias""", F"""decoder.layers.{i}.fc2.bias"""))
rename_keys.append(
(F"""transformer.decoder.layers.{i}.norm1.weight""", F"""decoder.layers.{i}.self_attn_layer_norm.weight""")
)
rename_keys.append((F"""transformer.decoder.layers.{i}.norm1.bias""", F"""decoder.layers.{i}.self_attn_layer_norm.bias"""))
rename_keys.append(
(F"""transformer.decoder.layers.{i}.norm2.weight""", F"""decoder.layers.{i}.encoder_attn_layer_norm.weight""")
)
rename_keys.append(
(F"""transformer.decoder.layers.{i}.norm2.bias""", F"""decoder.layers.{i}.encoder_attn_layer_norm.bias""")
)
rename_keys.append((F"""transformer.decoder.layers.{i}.norm3.weight""", F"""decoder.layers.{i}.final_layer_norm.weight"""))
rename_keys.append((F"""transformer.decoder.layers.{i}.norm3.bias""", F"""decoder.layers.{i}.final_layer_norm.bias"""))
# convolutional projection + query embeddings + layernorm of encoder + layernorm of decoder + class and bounding box heads
rename_keys.extend(
[
('''input_proj.weight''', '''input_projection.weight'''),
('''input_proj.bias''', '''input_projection.bias'''),
('''query_embed.weight''', '''query_position_embeddings.weight'''),
('''transformer.encoder.norm.weight''', '''encoder.layernorm.weight'''),
('''transformer.encoder.norm.bias''', '''encoder.layernorm.bias'''),
('''transformer.decoder.norm.weight''', '''decoder.layernorm.weight'''),
('''transformer.decoder.norm.bias''', '''decoder.layernorm.bias'''),
('''class_embed.weight''', '''class_labels_classifier.weight'''),
('''class_embed.bias''', '''class_labels_classifier.bias'''),
('''bbox_embed.layers.0.weight''', '''bbox_predictor.layers.0.weight'''),
('''bbox_embed.layers.0.bias''', '''bbox_predictor.layers.0.bias'''),
('''bbox_embed.layers.1.weight''', '''bbox_predictor.layers.1.weight'''),
('''bbox_embed.layers.1.bias''', '''bbox_predictor.layers.1.bias'''),
('''bbox_embed.layers.2.weight''', '''bbox_predictor.layers.2.weight'''),
('''bbox_embed.layers.2.bias''', '''bbox_predictor.layers.2.bias'''),
]
)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = state_dict.pop(UpperCamelCase )
lowerCAmelCase__ : Tuple = val
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = OrderedDict()
for key, value in state_dict.items():
if "backbone.0.body" in key:
lowerCAmelCase__ : List[str] = key.replace("""backbone.0.body""" , """backbone.conv_encoder.model""" )
lowerCAmelCase__ : List[Any] = value
else:
lowerCAmelCase__ : Optional[int] = value
return new_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = """"""
# first: transformer encoder
for i in range(6 ):
# read in weights + bias of input projection layer (in PyTorch's MultiHeadAttention, this is a single matrix + bias)
lowerCAmelCase__ : Union[str, Any] = state_dict.pop(f"""{prefix}transformer.encoder.layers.{i}.self_attn.in_proj_weight""" )
lowerCAmelCase__ : Any = state_dict.pop(f"""{prefix}transformer.encoder.layers.{i}.self_attn.in_proj_bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Union[str, Any] = in_proj_weight[:256, :]
lowerCAmelCase__ : List[Any] = in_proj_bias[:256]
lowerCAmelCase__ : str = in_proj_weight[256:512, :]
lowerCAmelCase__ : Optional[int] = in_proj_bias[256:512]
lowerCAmelCase__ : Union[str, Any] = in_proj_weight[-256:, :]
lowerCAmelCase__ : str = in_proj_bias[-256:]
# next: transformer decoder (which is a bit more complex because it also includes cross-attention)
for i in range(6 ):
# read in weights + bias of input projection layer of self-attention
lowerCAmelCase__ : List[str] = state_dict.pop(f"""{prefix}transformer.decoder.layers.{i}.self_attn.in_proj_weight""" )
lowerCAmelCase__ : Optional[Any] = state_dict.pop(f"""{prefix}transformer.decoder.layers.{i}.self_attn.in_proj_bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Optional[int] = in_proj_weight[:256, :]
lowerCAmelCase__ : Dict = in_proj_bias[:256]
lowerCAmelCase__ : Any = in_proj_weight[256:512, :]
lowerCAmelCase__ : Dict = in_proj_bias[256:512]
lowerCAmelCase__ : Tuple = in_proj_weight[-256:, :]
lowerCAmelCase__ : Union[str, Any] = in_proj_bias[-256:]
# read in weights + bias of input projection layer of cross-attention
lowerCAmelCase__ : List[Any] = state_dict.pop(
f"""{prefix}transformer.decoder.layers.{i}.multihead_attn.in_proj_weight""" )
lowerCAmelCase__ : List[str] = state_dict.pop(f"""{prefix}transformer.decoder.layers.{i}.multihead_attn.in_proj_bias""" )
# next, add query, keys and values (in that order) of cross-attention to the state dict
lowerCAmelCase__ : str = in_proj_weight_cross_attn[:256, :]
lowerCAmelCase__ : List[Any] = in_proj_bias_cross_attn[:256]
lowerCAmelCase__ : Optional[Any] = in_proj_weight_cross_attn[256:512, :]
lowerCAmelCase__ : str = in_proj_bias_cross_attn[256:512]
lowerCAmelCase__ : Dict = in_proj_weight_cross_attn[-256:, :]
lowerCAmelCase__ : str = in_proj_bias_cross_attn[-256:]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : str = image.size
lowerCAmelCase__ : Tuple = max(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : str = 800 if """detection""" in checkpoint_url else 1000
lowerCAmelCase__ : Union[str, Any] = target_max_size / current_max_size
lowerCAmelCase__ : str = image.resize((int(round(scale * width ) ), int(round(scale * height ) )) )
return resized_image
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = F.to_tensor(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = F.normalize(UpperCamelCase , mean=[0.485, 0.456, 0.406] , std=[0.229, 0.224, 0.225] )
return image
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
logger.info("""Converting model...""" )
# load original state dict
lowerCAmelCase__ : str = torch.hub.load_state_dict_from_url(UpperCamelCase , map_location="""cpu""" )
# rename keys
for src, dest in rename_keys:
rename_key(UpperCamelCase , UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = rename_backbone_keys(UpperCamelCase )
# query, key and value matrices need special treatment
read_in_q_k_v(UpperCamelCase )
# important: we need to prepend a prefix to each of the base model keys as the head models use different attributes for them
lowerCAmelCase__ : List[str] = """model."""
for key in state_dict.copy().keys():
if not key.startswith("""class_labels_classifier""" ) and not key.startswith("""bbox_predictor""" ):
lowerCAmelCase__ : Optional[Any] = state_dict.pop(UpperCamelCase )
lowerCAmelCase__ : int = val
# create HuggingFace model and load state dict
lowerCAmelCase__ : List[str] = TableTransformerConfig(
backbone="""resnet18""" , mask_loss_coefficient=1 , dice_loss_coefficient=1 , ce_loss_coefficient=1 , bbox_loss_coefficient=5 , giou_loss_coefficient=2 , eos_coefficient=0.4 , class_cost=1 , bbox_cost=5 , giou_cost=2 , )
if "detection" in checkpoint_url:
lowerCAmelCase__ : int = 15
lowerCAmelCase__ : Tuple = 2
lowerCAmelCase__ : Optional[Any] = {0: """table""", 1: """table rotated"""}
lowerCAmelCase__ : Union[str, Any] = idalabel
lowerCAmelCase__ : Tuple = {v: k for k, v in idalabel.items()}
else:
lowerCAmelCase__ : Optional[Any] = 125
lowerCAmelCase__ : int = 6
lowerCAmelCase__ : Union[str, Any] = {
0: """table""",
1: """table column""",
2: """table row""",
3: """table column header""",
4: """table projected row header""",
5: """table spanning cell""",
}
lowerCAmelCase__ : Any = idalabel
lowerCAmelCase__ : Any = {v: k for k, v in idalabel.items()}
lowerCAmelCase__ : List[Any] = DetrImageProcessor(
format="""coco_detection""" , max_size=800 if """detection""" in checkpoint_url else 1000 )
lowerCAmelCase__ : Optional[int] = TableTransformerForObjectDetection(UpperCamelCase )
model.load_state_dict(UpperCamelCase )
model.eval()
# verify our conversion
lowerCAmelCase__ : List[Any] = """example_pdf.png""" if """detection""" in checkpoint_url else """example_table.png"""
lowerCAmelCase__ : int = hf_hub_download(repo_id="""nielsr/example-pdf""" , repo_type="""dataset""" , filename=UpperCamelCase )
lowerCAmelCase__ : List[str] = Image.open(UpperCamelCase ).convert("""RGB""" )
lowerCAmelCase__ : Union[str, Any] = normalize(resize(UpperCamelCase , UpperCamelCase ) ).unsqueeze(0 )
lowerCAmelCase__ : Tuple = model(UpperCamelCase )
if "detection" in checkpoint_url:
lowerCAmelCase__ : List[str] = (1, 15, 3)
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[-6.7897, -16.9985, 6.7937], [-8.0186, -22.2192, 6.9677], [-7.3117, -21.0708, 7.4055]] )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0.4867, 0.1767, 0.6732], [0.6718, 0.4479, 0.3830], [0.4716, 0.1760, 0.6364]] )
else:
lowerCAmelCase__ : Any = (1, 125, 7)
lowerCAmelCase__ : str = torch.tensor(
[[-18.1430, -8.3214, 4.8274], [-18.4685, -7.1361, -4.2667], [-26.3693, -9.3429, -4.9962]] )
lowerCAmelCase__ : List[str] = torch.tensor([[0.4983, 0.5595, 0.9440], [0.4916, 0.6315, 0.5954], [0.6108, 0.8637, 0.1135]] )
assert outputs.logits.shape == expected_shape
assert torch.allclose(outputs.logits[0, :3, :3] , UpperCamelCase , atol=1e-4 )
assert torch.allclose(outputs.pred_boxes[0, :3, :3] , UpperCamelCase , atol=1e-4 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
# Save model and image processor
logger.info(f"""Saving PyTorch model and image processor to {pytorch_dump_folder_path}...""" )
Path(UpperCamelCase ).mkdir(exist_ok=UpperCamelCase )
model.save_pretrained(UpperCamelCase )
image_processor.save_pretrained(UpperCamelCase )
if push_to_hub:
# Push model to HF hub
logger.info("""Pushing model to the hub...""" )
lowerCAmelCase__ : Optional[Any] = (
"""microsoft/table-transformer-detection"""
if """detection""" in checkpoint_url
else """microsoft/table-transformer-structure-recognition"""
)
model.push_to_hub(UpperCamelCase )
image_processor.push_to_hub(UpperCamelCase )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_url''',
default='''https://pubtables1m.blob.core.windows.net/model/pubtables1m_detection_detr_r18.pth''',
type=str,
choices=[
'''https://pubtables1m.blob.core.windows.net/model/pubtables1m_detection_detr_r18.pth''',
'''https://pubtables1m.blob.core.windows.net/model/pubtables1m_structure_detr_r18.pth''',
],
help='''URL of the Table Transformer checkpoint you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the folder to output PyTorch model.'''
)
parser.add_argument(
'''--push_to_hub''', action='''store_true''', help='''Whether or not to push the converted model to the 🤗 hub.'''
)
_lowerCAmelCase = parser.parse_args()
convert_table_transformer_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub)
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : str = set()
# Replace all the whitespace in our sentence
lowerCAmelCase__ : Tuple = input_str.replace(""" """ , """""" )
for alpha in input_str:
if "a" <= alpha.lower() <= "z":
frequency.add(alpha.lower() )
return len(UpperCamelCase ) == 26
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : Any = [False] * 26
for char in input_str:
if char.islower():
lowerCAmelCase__ : Optional[Any] = True
elif char.isupper():
lowerCAmelCase__ : Any = True
return all(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
return len({char for char in input_str.lower() if char.isalpha()} ) == 26
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from timeit import timeit
lowerCAmelCase__ : Union[str, Any] = """from __main__ import is_pangram, is_pangram_faster, is_pangram_fastest"""
print(timeit("""is_pangram()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_faster()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_fastest()""" , setup=UpperCamelCase ) )
# 5.348480500048026, 2.6477354579837993, 1.8470395830227062
# 5.036091582966037, 2.644472333951853, 1.8869528750656173
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import csv
import tweepy
# Twitter API credentials
_lowerCAmelCase = ''''''
_lowerCAmelCase = ''''''
_lowerCAmelCase = ''''''
_lowerCAmelCase = ''''''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = tweepy.OAuthHandler(UpperCamelCase , UpperCamelCase )
auth.set_access_token(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = tweepy.API(UpperCamelCase )
# initialize a list to hold all the tweepy Tweets
lowerCAmelCase__ : Optional[int] = []
# make initial request for most recent tweets (200 is the maximum allowed count)
lowerCAmelCase__ : Any = api.user_timeline(screen_name=UpperCamelCase , count=200 )
# save most recent tweets
alltweets.extend(UpperCamelCase )
# save the id of the oldest tweet less one
lowerCAmelCase__ : int = alltweets[-1].id - 1
# keep grabbing tweets until there are no tweets left to grab
while len(UpperCamelCase ) > 0:
print(f"""getting tweets before {oldest}""" )
# all subsequent requests use the max_id param to prevent duplicates
lowerCAmelCase__ : Optional[Any] = api.user_timeline(
screen_name=UpperCamelCase , count=200 , max_id=UpperCamelCase )
# save most recent tweets
alltweets.extend(UpperCamelCase )
# update the id of the oldest tweet less one
lowerCAmelCase__ : Any = alltweets[-1].id - 1
print(f"""...{len(UpperCamelCase )} tweets downloaded so far""" )
# transform the tweepy tweets into a 2D array that will populate the csv
lowerCAmelCase__ : str = [[tweet.id_str, tweet.created_at, tweet.text] for tweet in alltweets]
# write the csv
with open(f"""new_{screen_name}_tweets.csv""" , """w""" ) as f:
lowerCAmelCase__ : List[str] = csv.writer(UpperCamelCase )
writer.writerow(["""id""", """created_at""", """text"""] )
writer.writerows(UpperCamelCase )
if __name__ == "__main__":
# pass in the username of the account you want to download
get_all_tweets('''FirePing32''')
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = abs(UpperCamelCase )
lowerCAmelCase__ : List[Any] = 0
while n > 0:
res += n % 10
n //= 10
return res
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = abs(UpperCamelCase )
return n if n < 10 else n % 10 + sum_of_digits(n // 10 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return sum(int(UpperCamelCase ) for c in str(abs(UpperCamelCase ) ) )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(UpperCamelCase , UpperCamelCase ) -> None:
lowerCAmelCase__ : str = f"""{func.__name__}({value})"""
lowerCAmelCase__ : str = timeit(f"""__main__.{call}""" , setup="""import __main__""" )
print(f"""{call:56} = {func(UpperCamelCase )} -- {timing:.4f} seconds""" )
for value in (262144, 1125899906842624, 1267650600228229401496703205376):
for func in (sum_of_digits, sum_of_digits_recursion, sum_of_digits_compact):
benchmark_a_function(UpperCamelCase , UpperCamelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xmod-base''': '''https://huggingface.co/facebook/xmod-base/resolve/main/config.json''',
'''facebook/xmod-large-prenorm''': '''https://huggingface.co/facebook/xmod-large-prenorm/resolve/main/config.json''',
'''facebook/xmod-base-13-125k''': '''https://huggingface.co/facebook/xmod-base-13-125k/resolve/main/config.json''',
'''facebook/xmod-base-30-125k''': '''https://huggingface.co/facebook/xmod-base-30-125k/resolve/main/config.json''',
'''facebook/xmod-base-30-195k''': '''https://huggingface.co/facebook/xmod-base-30-195k/resolve/main/config.json''',
'''facebook/xmod-base-60-125k''': '''https://huggingface.co/facebook/xmod-base-60-125k/resolve/main/config.json''',
'''facebook/xmod-base-60-265k''': '''https://huggingface.co/facebook/xmod-base-60-265k/resolve/main/config.json''',
'''facebook/xmod-base-75-125k''': '''https://huggingface.co/facebook/xmod-base-75-125k/resolve/main/config.json''',
'''facebook/xmod-base-75-269k''': '''https://huggingface.co/facebook/xmod-base-75-269k/resolve/main/config.json''',
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[int] = '''xmod'''
def __init__( self ,__UpperCAmelCase=3_0522 ,__UpperCAmelCase=768 ,__UpperCAmelCase=12 ,__UpperCAmelCase=12 ,__UpperCAmelCase=3072 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=512 ,__UpperCAmelCase=2 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-12 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,__UpperCAmelCase=False ,__UpperCAmelCase=2 ,__UpperCAmelCase=False ,__UpperCAmelCase=True ,__UpperCAmelCase=True ,__UpperCAmelCase=("en_XX",) ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> List[Any]:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = vocab_size
lowerCAmelCase__ : Optional[int] = hidden_size
lowerCAmelCase__ : Union[str, Any] = num_hidden_layers
lowerCAmelCase__ : List[Any] = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : str = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : Union[str, Any] = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[Any] = max_position_embeddings
lowerCAmelCase__ : Tuple = type_vocab_size
lowerCAmelCase__ : Optional[int] = initializer_range
lowerCAmelCase__ : Dict = layer_norm_eps
lowerCAmelCase__ : List[Any] = position_embedding_type
lowerCAmelCase__ : Dict = use_cache
lowerCAmelCase__ : Tuple = classifier_dropout
lowerCAmelCase__ : Union[str, Any] = pre_norm
lowerCAmelCase__ : Union[str, Any] = adapter_reduction_factor
lowerCAmelCase__ : List[str] = adapter_layer_norm
lowerCAmelCase__ : Optional[Any] = adapter_reuse_layer_norm
lowerCAmelCase__ : int = ln_before_adapter
lowerCAmelCase__ : Any = list(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = default_language
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : List[str] = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Optional[Any] = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
|
'''simple docstring'''
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team.
#
# 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.
# this script dumps information about the environment
import os
import sys
import transformers
_lowerCAmelCase = '''3'''
print('''Python version:''', sys.version)
print('''transformers version:''', transformers.__version__)
try:
import torch
print('''Torch version:''', torch.__version__)
print('''Cuda available:''', torch.cuda.is_available())
print('''Cuda version:''', torch.version.cuda)
print('''CuDNN version:''', torch.backends.cudnn.version())
print('''Number of GPUs available:''', torch.cuda.device_count())
print('''NCCL version:''', torch.cuda.nccl.version())
except ImportError:
print('''Torch version:''', None)
try:
import deepspeed
print('''DeepSpeed version:''', deepspeed.__version__)
except ImportError:
print('''DeepSpeed version:''', None)
try:
import tensorflow as tf
print('''TensorFlow version:''', tf.__version__)
print('''TF GPUs available:''', bool(tf.config.list_physical_devices('''GPU''')))
print('''Number of TF GPUs available:''', len(tf.config.list_physical_devices('''GPU''')))
except ImportError:
print('''TensorFlow version:''', None)
| 37
| 1
|
'''simple docstring'''
import json
import os
import tempfile
import transformers
import datasets
from utils import generate_example_dataset, get_duration
_lowerCAmelCase = 50_0000
_lowerCAmelCase , _lowerCAmelCase = os.path.split(__file__)
_lowerCAmelCase = os.path.join(RESULTS_BASEPATH, '''results''', RESULTS_FILENAME.replace('''.py''', '''.json'''))
@get_duration
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , **UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = dataset.map(**UpperCamelCase )
@get_duration
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , **UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = dataset.filter(**UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = {"""num examples""": SPEED_TEST_N_EXAMPLES}
with tempfile.TemporaryDirectory() as tmp_dir:
lowerCAmelCase__ : List[str] = datasets.Features({"""text""": datasets.Value("""string""" ), """numbers""": datasets.Value("""float32""" )} )
lowerCAmelCase__ : Union[str, Any] = generate_example_dataset(
os.path.join(UpperCamelCase , """dataset.arrow""" ) , UpperCamelCase , num_examples=UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = transformers.AutoTokenizer.from_pretrained("""bert-base-cased""" , use_fast=UpperCamelCase )
def tokenize(UpperCamelCase ):
return tokenizer(examples["""text"""] )
lowerCAmelCase__ : int = map(UpperCamelCase )
lowerCAmelCase__ : List[Any] = map(UpperCamelCase , batched=UpperCamelCase )
lowerCAmelCase__ : int = map(UpperCamelCase , function=lambda UpperCamelCase : None , batched=UpperCamelCase )
with dataset.formatted_as(type="""numpy""" ):
lowerCAmelCase__ : Optional[Any] = map(UpperCamelCase , function=lambda UpperCamelCase : None , batched=UpperCamelCase )
with dataset.formatted_as(type="""pandas""" ):
lowerCAmelCase__ : Dict = map(UpperCamelCase , function=lambda UpperCamelCase : None , batched=UpperCamelCase )
with dataset.formatted_as(type="""torch""" , columns="""numbers""" ):
lowerCAmelCase__ : Any = map(UpperCamelCase , function=lambda UpperCamelCase : None , batched=UpperCamelCase )
with dataset.formatted_as(type="""tensorflow""" , columns="""numbers""" ):
lowerCAmelCase__ : Dict = map(UpperCamelCase , function=lambda UpperCamelCase : None , batched=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = map(UpperCamelCase , function=UpperCamelCase , batched=UpperCamelCase )
lowerCAmelCase__ : Any = filter(UpperCamelCase )
# Activate later when tokenizer support batched inputs
# with dataset.formatted_as(type='numpy'):
# times[func.__name__ + " fast-tokenizer batched numpy"] = func(dataset, function=tokenize, batched=True)
with open(UpperCamelCase , """wb""" ) as f:
f.write(json.dumps(UpperCamelCase ).encode("""utf-8""" ) )
if __name__ == "__main__": # useful to run the profiler
benchmark_map_filter()
| 37
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xlm-roberta-xl''': '''https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json''',
'''facebook/xlm-roberta-xxl''': '''https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json''',
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''xlm-roberta-xl'''
def __init__( self ,__UpperCAmelCase=25_0880 ,__UpperCAmelCase=2560 ,__UpperCAmelCase=36 ,__UpperCAmelCase=32 ,__UpperCAmelCase=1_0240 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=514 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-05 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = vocab_size
lowerCAmelCase__ : int = hidden_size
lowerCAmelCase__ : int = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : str = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : List[str] = type_vocab_size
lowerCAmelCase__ : List[Any] = initializer_range
lowerCAmelCase__ : Tuple = layer_norm_eps
lowerCAmelCase__ : int = position_embedding_type
lowerCAmelCase__ : Optional[Any] = use_cache
lowerCAmelCase__ : Optional[Any] = classifier_dropout
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Dict = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Any = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = n ** (1 / 3)
return (val * val * val) == n
if __name__ == "__main__":
print(perfect_cube(27))
print(perfect_cube(4))
| 37
|
'''simple docstring'''
from __future__ import annotations
import math
from collections import Counter
from string import ascii_lowercase
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = analyze_text(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = list(""" """ + ascii_lowercase )
# what is our total sum of probabilities.
lowerCAmelCase__ : List[Any] = sum(single_char_strings.values() )
# one length string
lowerCAmelCase__ : Optional[int] = 0
# for each alpha we go in our dict and if it is in it we calculate entropy
for ch in my_alphas:
if ch in single_char_strings:
lowerCAmelCase__ : List[Any] = single_char_strings[ch]
lowerCAmelCase__ : List[Any] = my_str / all_sum
my_fir_sum += prob * math.loga(UpperCamelCase ) # entropy formula.
# print entropy
print(f"""{round(-1 * my_fir_sum ):.1f}""" )
# two len string
lowerCAmelCase__ : Dict = sum(two_char_strings.values() )
lowerCAmelCase__ : int = 0
# for each alpha (two in size) calculate entropy.
for cha in my_alphas:
for cha in my_alphas:
lowerCAmelCase__ : Union[str, Any] = cha + cha
if sequence in two_char_strings:
lowerCAmelCase__ : Dict = two_char_strings[sequence]
lowerCAmelCase__ : Tuple = int(UpperCamelCase ) / all_sum
my_sec_sum += prob * math.loga(UpperCamelCase )
# print second entropy
print(f"""{round(-1 * my_sec_sum ):.1f}""" )
# print the difference between them
print(f"""{round((-1 * my_sec_sum) - (-1 * my_fir_sum) ):.1f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = Counter() # type: ignore
lowerCAmelCase__ : Tuple = Counter() # type: ignore
single_char_strings[text[-1]] += 1
# first case when we have space at start.
two_char_strings[" " + text[0]] += 1
for i in range(0 , len(UpperCamelCase ) - 1 ):
single_char_strings[text[i]] += 1
two_char_strings[text[i : i + 2]] += 1
return single_char_strings, two_char_strings
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
import doctest
doctest.testmod()
# text = (
# "Had repulsive dashwoods suspicion sincerity but advantage now him. Remark "
# "easily garret nor nay. Civil those mrs enjoy shy fat merry. You greatest "
# "jointure saw horrible. He private he on be imagine suppose. Fertile "
# "beloved evident through no service elderly is. Blind there if every no so "
# "at. Own neglected you preferred way sincerity delivered his attempted. To "
# "of message cottage windows do besides against uncivil. Delightful "
# "unreserved impossible few estimating men favourable see entreaties. She "
# "propriety immediate was improving. He or entrance humoured likewise "
# "moderate. Much nor game son say feel. Fat make met can must form into "
# "gate. Me we offending prevailed discovery. "
# )
# calculate_prob(text)
if __name__ == "__main__":
main()
| 37
| 1
|
'''simple docstring'''
import os
from bleurt import score # From: git+https://github.com/google-research/bleurt.git
import datasets
_lowerCAmelCase = datasets.logging.get_logger(__name__)
_lowerCAmelCase = '''\
@inproceedings{bleurt,
title={BLEURT: Learning Robust Metrics for Text Generation},
author={Thibault Sellam and Dipanjan Das and Ankur P. Parikh},
booktitle={ACL},
year={2020},
url={https://arxiv.org/abs/2004.04696}
}
'''
_lowerCAmelCase = '''\
BLEURT a learnt evaluation metric for Natural Language Generation. It is built using multiple phases of transfer learning starting from a pretrained BERT model (Devlin et al. 2018)
and then employing another pre-training phrase using synthetic data. Finally it is trained on WMT human annotations. You may run BLEURT out-of-the-box or fine-tune
it for your specific application (the latter is expected to perform better).
See the project\'s README at https://github.com/google-research/bleurt#readme for more information.
'''
_lowerCAmelCase = '''
BLEURT score.
Args:
`predictions` (list of str): prediction/candidate sentences
`references` (list of str): reference sentences
`checkpoint` BLEURT checkpoint. Will default to BLEURT-tiny if None.
Returns:
\'scores\': List of scores.
Examples:
>>> predictions = ["hello there", "general kenobi"]
>>> references = ["hello there", "general kenobi"]
>>> bleurt = datasets.load_metric("bleurt")
>>> results = bleurt.compute(predictions=predictions, references=references)
>>> print([round(v, 2) for v in results["scores"]])
[1.03, 1.04]
'''
_lowerCAmelCase = {
'''bleurt-tiny-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-128.zip''',
'''bleurt-tiny-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-512.zip''',
'''bleurt-base-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-128.zip''',
'''bleurt-base-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-512.zip''',
'''bleurt-large-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-128.zip''',
'''bleurt-large-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-512.zip''',
'''BLEURT-20-D3''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D3.zip''',
'''BLEURT-20-D6''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D6.zip''',
'''BLEURT-20-D12''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D12.zip''',
'''BLEURT-20''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20.zip''',
}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,homepage="""https://github.com/google-research/bleurt""" ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(
{
"""predictions""": datasets.Value("""string""" ,id="""sequence""" ),
"""references""": datasets.Value("""string""" ,id="""sequence""" ),
} ) ,codebase_urls=["""https://github.com/google-research/bleurt"""] ,reference_urls=["""https://github.com/google-research/bleurt""", """https://arxiv.org/abs/2004.04696"""] ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
# check that config name specifies a valid BLEURT model
if self.config_name == "default":
logger.warning(
"""Using default BLEURT-Base checkpoint for sequence maximum length 128. """
"""You can use a bigger model for better results with e.g.: datasets.load_metric('bleurt', 'bleurt-large-512').""" )
lowerCAmelCase__ : str = """bleurt-base-128"""
if self.config_name.lower() in CHECKPOINT_URLS:
lowerCAmelCase__ : Union[str, Any] = self.config_name.lower()
elif self.config_name.upper() in CHECKPOINT_URLS:
lowerCAmelCase__ : List[Any] = self.config_name.upper()
else:
raise KeyError(
F"""{self.config_name} model not found. You should supply the name of a model checkpoint for bleurt in {CHECKPOINT_URLS.keys()}""" )
# download the model checkpoint specified by self.config_name and set up the scorer
lowerCAmelCase__ : int = dl_manager.download_and_extract(CHECKPOINT_URLS[checkpoint_name] )
lowerCAmelCase__ : Dict = score.BleurtScorer(os.path.join(__UpperCAmelCase ,__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Union[str, Any] = self.scorer.score(references=__UpperCAmelCase ,candidates=__UpperCAmelCase )
return {"scores": scores}
| 37
|
'''simple docstring'''
import argparse
import json
import torch
from diffusers import DDPMScheduler, LDMPipeline, UNetaDModel, VQModel
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=1 ):
"""simple docstring"""
if n_shave_prefix_segments >= 0:
return ".".join(path.split(""".""" )[n_shave_prefix_segments:] )
else:
return ".".join(path.split(""".""" )[:n_shave_prefix_segments] )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = []
for old_item in old_list:
lowerCAmelCase__ : Optional[Any] = old_item.replace("""in_layers.0""" , """norm1""" )
lowerCAmelCase__ : Optional[int] = new_item.replace("""in_layers.2""" , """conv1""" )
lowerCAmelCase__ : Dict = new_item.replace("""out_layers.0""" , """norm2""" )
lowerCAmelCase__ : str = new_item.replace("""out_layers.3""" , """conv2""" )
lowerCAmelCase__ : str = new_item.replace("""emb_layers.1""" , """time_emb_proj""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""skip_connection""" , """conv_shortcut""" )
lowerCAmelCase__ : Union[str, Any] = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : int = []
for old_item in old_list:
lowerCAmelCase__ : List[str] = old_item
lowerCAmelCase__ : int = new_item.replace("""norm.weight""" , """group_norm.weight""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""norm.bias""" , """group_norm.bias""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""proj_out.weight""" , """proj_attn.weight""" )
lowerCAmelCase__ : int = new_item.replace("""proj_out.bias""" , """proj_attn.bias""" )
lowerCAmelCase__ : str = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=None ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "Paths should be a list of dicts containing 'old' and 'new' keys."
# Splits the attention layers into three variables.
if attention_paths_to_split is not None:
for path, path_map in attention_paths_to_split.items():
lowerCAmelCase__ : Any = old_checkpoint[path]
lowerCAmelCase__ : int = old_tensor.shape[0] // 3
lowerCAmelCase__ : int = (-1, channels) if len(old_tensor.shape ) == 3 else (-1)
lowerCAmelCase__ : Tuple = old_tensor.shape[0] // config["""num_head_channels"""] // 3
lowerCAmelCase__ : List[Any] = old_tensor.reshape((num_heads, 3 * channels // num_heads) + old_tensor.shape[1:] )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : str = old_tensor.split(channels // num_heads , dim=1 )
lowerCAmelCase__ : int = query.reshape(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = key.reshape(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = value.reshape(UpperCamelCase )
for path in paths:
lowerCAmelCase__ : Any = path["""new"""]
# These have already been assigned
if attention_paths_to_split is not None and new_path in attention_paths_to_split:
continue
# Global renaming happens here
lowerCAmelCase__ : Any = new_path.replace("""middle_block.0""" , """mid_block.resnets.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.1""" , """mid_block.attentions.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.2""" , """mid_block.resnets.1""" )
if additional_replacements is not None:
for replacement in additional_replacements:
lowerCAmelCase__ : Any = new_path.replace(replacement["""old"""] , replacement["""new"""] )
# proj_attn.weight has to be converted from conv 1D to linear
if "proj_attn.weight" in new_path:
lowerCAmelCase__ : List[Any] = old_checkpoint[path["""old"""]][:, :, 0]
else:
lowerCAmelCase__ : Dict = old_checkpoint[path["""old"""]]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = {}
lowerCAmelCase__ : str = checkpoint["""time_embed.0.weight"""]
lowerCAmelCase__ : List[Any] = checkpoint["""time_embed.0.bias"""]
lowerCAmelCase__ : int = checkpoint["""time_embed.2.weight"""]
lowerCAmelCase__ : List[str] = checkpoint["""time_embed.2.bias"""]
lowerCAmelCase__ : str = checkpoint["""input_blocks.0.0.weight"""]
lowerCAmelCase__ : Any = checkpoint["""input_blocks.0.0.bias"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.weight"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.bias"""]
lowerCAmelCase__ : str = checkpoint["""out.2.weight"""]
lowerCAmelCase__ : Tuple = checkpoint["""out.2.bias"""]
# Retrieves the keys for the input blocks only
lowerCAmelCase__ : Optional[Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """input_blocks""" in layer} )
lowerCAmelCase__ : Optional[Any] = {
layer_id: [key for key in checkpoint if f"""input_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the middle blocks only
lowerCAmelCase__ : Union[str, Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """middle_block""" in layer} )
lowerCAmelCase__ : Union[str, Any] = {
layer_id: [key for key in checkpoint if f"""middle_block.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the output blocks only
lowerCAmelCase__ : List[str] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """output_blocks""" in layer} )
lowerCAmelCase__ : List[Any] = {
layer_id: [key for key in checkpoint if f"""output_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
for i in range(1 , UpperCamelCase ):
lowerCAmelCase__ : Dict = (i - 1) // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Tuple = (i - 1) % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Optional[int] = [key for key in input_blocks[i] if f"""input_blocks.{i}.0""" in key]
lowerCAmelCase__ : Optional[Any] = [key for key in input_blocks[i] if f"""input_blocks.{i}.1""" in key]
if f"""input_blocks.{i}.0.op.weight""" in checkpoint:
lowerCAmelCase__ : Optional[int] = checkpoint[
f"""input_blocks.{i}.0.op.weight"""
]
lowerCAmelCase__ : Tuple = checkpoint[
f"""input_blocks.{i}.0.op.bias"""
]
continue
lowerCAmelCase__ : Optional[Any] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Dict = {"""old""": f"""input_blocks.{i}.0""", """new""": f"""down_blocks.{block_id}.resnets.{layer_in_block_id}"""}
lowerCAmelCase__ : Optional[Any] = {"""old""": """resnets.2.op""", """new""": """downsamplers.0.op"""}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path, resnet_op] , config=UpperCamelCase )
if len(UpperCamelCase ):
lowerCAmelCase__ : Optional[Any] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""input_blocks.{i}.1""",
"""new""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : List[str] = {
f"""input_blocks.{i}.1.qkv.bias""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""input_blocks.{i}.1.qkv.weight""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase , )
lowerCAmelCase__ : Dict = middle_blocks[0]
lowerCAmelCase__ : Union[str, Any] = middle_blocks[1]
lowerCAmelCase__ : Dict = middle_blocks[2]
lowerCAmelCase__ : Any = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Dict = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {
"""middle_block.1.qkv.bias""": {
"""key""": """mid_block.attentions.0.key.bias""",
"""query""": """mid_block.attentions.0.query.bias""",
"""value""": """mid_block.attentions.0.value.bias""",
},
"""middle_block.1.qkv.weight""": {
"""key""": """mid_block.attentions.0.key.weight""",
"""query""": """mid_block.attentions.0.query.weight""",
"""value""": """mid_block.attentions.0.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase )
for i in range(UpperCamelCase ):
lowerCAmelCase__ : Tuple = i // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : List[str] = i % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : int = [shave_segments(UpperCamelCase , 2 ) for name in output_blocks[i]]
lowerCAmelCase__ : Union[str, Any] = {}
for layer in output_block_layers:
lowerCAmelCase__ , lowerCAmelCase__ : Any = layer.split(""".""" )[0], shave_segments(UpperCamelCase , 1 )
if layer_id in output_block_list:
output_block_list[layer_id].append(UpperCamelCase )
else:
lowerCAmelCase__ : str = [layer_name]
if len(UpperCamelCase ) > 1:
lowerCAmelCase__ : str = [key for key in output_blocks[i] if f"""output_blocks.{i}.0""" in key]
lowerCAmelCase__ : Dict = [key for key in output_blocks[i] if f"""output_blocks.{i}.1""" in key]
lowerCAmelCase__ : Optional[int] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {"""old""": f"""output_blocks.{i}.0""", """new""": f"""up_blocks.{block_id}.resnets.{layer_in_block_id}"""}
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , config=UpperCamelCase )
if ["conv.weight", "conv.bias"] in output_block_list.values():
lowerCAmelCase__ : List[Any] = list(output_block_list.values() ).index(["""conv.weight""", """conv.bias"""] )
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.weight"""
]
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.bias"""
]
# Clear attentions as they have been attributed above.
if len(UpperCamelCase ) == 2:
lowerCAmelCase__ : Tuple = []
if len(UpperCamelCase ):
lowerCAmelCase__ : Dict = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""output_blocks.{i}.1""",
"""new""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : Tuple = {
f"""output_blocks.{i}.1.qkv.bias""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""output_blocks.{i}.1.qkv.weight""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=to_split if any("""qkv""" in key for key in attentions ) else None , config=UpperCamelCase , )
else:
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase , n_shave_prefix_segments=1 )
for path in resnet_0_paths:
lowerCAmelCase__ : Tuple = """.""".join(["""output_blocks""", str(UpperCamelCase ), path["""old"""]] )
lowerCAmelCase__ : List[Any] = """.""".join(["""up_blocks""", str(UpperCamelCase ), """resnets""", str(UpperCamelCase ), path["""new"""]] )
lowerCAmelCase__ : Union[str, Any] = checkpoint[old_path]
return new_checkpoint
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the checkpoint to convert.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help='''The config json file corresponding to the architecture.''',
)
parser.add_argument('''--dump_path''', default=None, type=str, required=True, help='''Path to the output model.''')
_lowerCAmelCase = parser.parse_args()
_lowerCAmelCase = torch.load(args.checkpoint_path)
with open(args.config_file) as f:
_lowerCAmelCase = json.loads(f.read())
_lowerCAmelCase = convert_ldm_checkpoint(checkpoint, config)
if "ldm" in config:
del config["ldm"]
_lowerCAmelCase = UNetaDModel(**config)
model.load_state_dict(converted_checkpoint)
try:
_lowerCAmelCase = DDPMScheduler.from_config('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = VQModel.from_pretrained('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = LDMPipeline(unet=model, scheduler=scheduler, vae=vqvae)
pipe.save_pretrained(args.dump_path)
except: # noqa: E722
model.save_pretrained(args.dump_path)
| 37
| 1
|
'''simple docstring'''
# flake8: noqa
# Lint as: python3
from typing import Dict, List, Optional, Type
from .. import config
from ..utils import logging
from .formatting import (
ArrowFormatter,
CustomFormatter,
Formatter,
PandasFormatter,
PythonFormatter,
TensorFormatter,
format_table,
query_table,
)
from .np_formatter import NumpyFormatter
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {}
_lowerCAmelCase = {}
_lowerCAmelCase = {}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase = None , ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = aliases if aliases is not None else []
if format_type in _FORMAT_TYPES:
logger.warning(
f"""Overwriting format type '{format_type}' ({_FORMAT_TYPES[format_type].__name__} -> {formatter_cls.__name__})""" )
lowerCAmelCase__ : Union[str, Any] = formatter_cls
for alias in set(aliases + [format_type] ):
if alias in _FORMAT_TYPES_ALIASES:
logger.warning(
f"""Overwriting format type alias '{alias}' ({_FORMAT_TYPES_ALIASES[alias]} -> {format_type})""" )
lowerCAmelCase__ : Tuple = format_type
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase = None ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = aliases if aliases is not None else []
for alias in set(aliases + [format_type] ):
lowerCAmelCase__ : Any = unavailable_error
# Here we define all the available formatting functions that can be used by `Dataset.set_format`
_register_formatter(PythonFormatter, None, aliases=['''python'''])
_register_formatter(ArrowFormatter, '''arrow''', aliases=['''pa''', '''pyarrow'''])
_register_formatter(NumpyFormatter, '''numpy''', aliases=['''np'''])
_register_formatter(PandasFormatter, '''pandas''', aliases=['''pd'''])
_register_formatter(CustomFormatter, '''custom''')
if config.TORCH_AVAILABLE:
from .torch_formatter import TorchFormatter
_register_formatter(TorchFormatter, '''torch''', aliases=['''pt''', '''pytorch'''])
else:
_lowerCAmelCase = ValueError('''PyTorch needs to be installed to be able to return PyTorch tensors.''')
_register_unavailable_formatter(_torch_error, '''torch''', aliases=['''pt''', '''pytorch'''])
if config.TF_AVAILABLE:
from .tf_formatter import TFFormatter
_register_formatter(TFFormatter, '''tensorflow''', aliases=['''tf'''])
else:
_lowerCAmelCase = ValueError('''Tensorflow needs to be installed to be able to return Tensorflow tensors.''')
_register_unavailable_formatter(_tf_error, '''tensorflow''', aliases=['''tf'''])
if config.JAX_AVAILABLE:
from .jax_formatter import JaxFormatter
_register_formatter(JaxFormatter, '''jax''', aliases=[])
else:
_lowerCAmelCase = ValueError('''JAX needs to be installed to be able to return JAX arrays.''')
_register_unavailable_formatter(_jax_error, '''jax''', aliases=[])
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if format_type in _FORMAT_TYPES_ALIASES:
return _FORMAT_TYPES_ALIASES[format_type]
else:
return format_type
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , **UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : int = get_format_type_from_alias(UpperCamelCase )
if format_type in _FORMAT_TYPES:
return _FORMAT_TYPES[format_type](**UpperCamelCase )
if format_type in _FORMAT_TYPES_ALIASES_UNAVAILABLE:
raise _FORMAT_TYPES_ALIASES_UNAVAILABLE[format_type]
else:
raise ValueError(
f"""Return type should be None or selected in {list(type for type in _FORMAT_TYPES.keys() if type != None )}, but got '{format_type}'""" )
| 37
|
'''simple docstring'''
from math import sqrt
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' must been an int and positive"
lowerCAmelCase__ : int = True
# 0 and 1 are none primes.
if number <= 1:
lowerCAmelCase__ : Optional[Any] = False
for divisor in range(2 , int(round(sqrt(UpperCamelCase ) ) ) + 1 ):
# if 'number' divisible by 'divisor' then sets 'status'
# of false and break up the loop.
if number % divisor == 0:
lowerCAmelCase__ : Any = False
break
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'status' must been from type bool"
return status
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
# beginList: contains all natural numbers from 2 up to N
lowerCAmelCase__ : List[str] = list(range(2 , n + 1 ) )
lowerCAmelCase__ : str = [] # this list will be returns.
# actual sieve of erathostenes
for i in range(len(UpperCamelCase ) ):
for j in range(i + 1 , len(UpperCamelCase ) ):
if (begin_list[i] != 0) and (begin_list[j] % begin_list[i] == 0):
lowerCAmelCase__ : List[Any] = 0
# filters actual prime numbers.
lowerCAmelCase__ : List[Any] = [x for x in begin_list if x != 0]
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
lowerCAmelCase__ : List[str] = []
# iterates over all numbers between 2 up to N+1
# if a number is prime then appends to list 'ans'
for number in range(2 , n + 1 ):
if is_prime(UpperCamelCase ):
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and number >= 0, "'number' must been an int and >= 0"
lowerCAmelCase__ : Optional[Any] = [] # this list will be returns of the function.
# potential prime number factors.
lowerCAmelCase__ : Dict = 2
lowerCAmelCase__ : Dict = number
if number == 0 or number == 1:
ans.append(UpperCamelCase )
# if 'number' not prime then builds the prime factorization of 'number'
elif not is_prime(UpperCamelCase ):
while quotient != 1:
if is_prime(UpperCamelCase ) and (quotient % factor == 0):
ans.append(UpperCamelCase )
quotient /= factor
else:
factor += 1
else:
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : Optional[int] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = max(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = min(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 == 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 == 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 != 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 != 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase ) and (number > 2) and is_even(UpperCamelCase )
), "'number' must been an int, even and > 2"
lowerCAmelCase__ : Dict = [] # this list will returned
# creates a list of prime numbers between 2 up to 'number'
lowerCAmelCase__ : Dict = get_prime_numbers(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = len(UpperCamelCase )
# run variable for while-loops.
lowerCAmelCase__ : List[str] = 0
lowerCAmelCase__ : List[Any] = None
# exit variable. for break up the loops
lowerCAmelCase__ : Any = True
while i < len_pn and loop:
lowerCAmelCase__ : List[Any] = i + 1
while j < len_pn and loop:
if prime_numbers[i] + prime_numbers[j] == number:
lowerCAmelCase__ : Optional[Any] = False
ans.append(prime_numbers[i] )
ans.append(prime_numbers[j] )
j += 1
i += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (len(UpperCamelCase ) == 2)
and (ans[0] + ans[1] == number)
and is_prime(ans[0] )
and is_prime(ans[1] )
), "'ans' must contains two primes. And sum of elements must been eq 'number'"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 0)
and (numbera >= 0)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 0
while numbera != 0:
lowerCAmelCase__ : Any = numbera % numbera
lowerCAmelCase__ : str = numbera
lowerCAmelCase__ : List[str] = rest
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
numbera >= 0
), "'number' must been from type int and positive"
return numbera
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 1)
and (numbera >= 1)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 1 # actual answer that will be return.
# for kgV (x,1)
if numbera > 1 and numbera > 1:
# builds the prime factorization of 'number1' and 'number2'
lowerCAmelCase__ : int = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = prime_factorization(UpperCamelCase )
elif numbera == 1 or numbera == 1:
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : Dict = []
lowerCAmelCase__ : List[str] = max(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = 0
lowerCAmelCase__ : str = 0
lowerCAmelCase__ : List[Any] = [] # captured numbers int both 'primeFac1' and 'primeFac2'
# iterates through primeFac1
for n in prime_fac_a:
if n not in done:
if n in prime_fac_a:
lowerCAmelCase__ : int = prime_fac_a.count(UpperCamelCase )
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(max(UpperCamelCase , UpperCamelCase ) ):
ans *= n
else:
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# iterates through primeFac2
for n in prime_fac_a:
if n not in done:
lowerCAmelCase__ : Optional[int] = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
ans >= 0
), "'ans' must been from type int and positive"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'number' must been a positive int"
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Tuple = 2 # this variable holds the answer
while index < n:
index += 1
ans += 1 # counts to the next number
# if ans not prime then
# runs to the next prime number.
while not is_prime(UpperCamelCase ):
ans += 1
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and is_prime(
UpperCamelCase ), "'ans' must been a prime number and from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
is_prime(UpperCamelCase ) and is_prime(UpperCamelCase ) and (p_number_a < p_number_a)
), "The arguments must been prime numbers and 'pNumber1' < 'pNumber2'"
lowerCAmelCase__ : Dict = p_number_a + 1 # jump to the next number
lowerCAmelCase__ : List[Any] = [] # this list will be returns.
# if number is not prime then
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
while number < p_number_a:
ans.append(UpperCamelCase )
number += 1
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and ans[0] != p_number_a
and ans[len(UpperCamelCase ) - 1] != p_number_a
), "'ans' must been a list without the arguments"
# 'ans' contains not 'pNumber1' and 'pNumber2' !
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 1), "'n' must been int and >= 1"
lowerCAmelCase__ : List[Any] = [] # will be returned.
for divisor in range(1 , n + 1 ):
if n % divisor == 0:
ans.append(UpperCamelCase )
# precondition
assert ans[0] == 1 and ans[len(UpperCamelCase ) - 1] == n, "Error in function getDivisiors(...)"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number > 1
), "'number' must been an int and >= 1"
lowerCAmelCase__ : Optional[int] = get_divisors(UpperCamelCase )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (divisors[0] == 1)
and (divisors[len(UpperCamelCase ) - 1] == number)
), "Error in help-function getDivisiors(...)"
# summed all divisors up to 'number' (exclusive), hence [:-1]
return sum(divisors[:-1] ) == number
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (denominator != 0)
), "The arguments must been from type int and 'denominator' != 0"
# build the greatest common divisor of numerator and denominator.
lowerCAmelCase__ : int = gcd(abs(UpperCamelCase ) , abs(UpperCamelCase ) )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (numerator % gcd_of_fraction == 0)
and (denominator % gcd_of_fraction == 0)
), "Error in function gcd(...,...)"
return (numerator // gcd_of_fraction, denominator // gcd_of_fraction)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been a int and >= 0"
lowerCAmelCase__ : str = 1 # this will be return.
for factor in range(1 , n + 1 ):
ans *= factor
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
lowerCAmelCase__ : Any = 1
lowerCAmelCase__ : Optional[Any] = 1 # this will be return
for _ in range(n - 1 ):
lowerCAmelCase__ : Dict = ans
ans += fiba
lowerCAmelCase__ : str = tmp
return ans
| 37
| 1
|
'''simple docstring'''
import os
# Precomputes a list of the 100 first triangular numbers
_lowerCAmelCase = [int(0.5 * n * (n + 1)) for n in range(1, 101)]
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : str = os.path.dirname(os.path.realpath(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = os.path.join(UpperCamelCase , """words.txt""" )
lowerCAmelCase__ : str = """"""
with open(UpperCamelCase ) as f:
lowerCAmelCase__ : str = f.readline()
lowerCAmelCase__ : Union[str, Any] = [word.strip("""\"""" ) for word in words.strip("""\r\n""" ).split(""",""" )]
lowerCAmelCase__ : Optional[Any] = [
word
for word in [sum(ord(UpperCamelCase ) - 64 for x in word ) for word in words]
if word in TRIANGULAR_NUMBERS
]
return len(UpperCamelCase )
if __name__ == "__main__":
print(solution())
| 37
|
'''simple docstring'''
from sklearn.metrics import fa_score, matthews_corrcoef
import datasets
from .record_evaluation import evaluate as evaluate_record
_lowerCAmelCase = '''\
@article{wang2019superglue,
title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
journal={arXiv preprint arXiv:1905.00537},
year={2019}
}
'''
_lowerCAmelCase = '''\
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
GLUE with a new set of more difficult language understanding tasks, improved
resources, and a new public leaderboard.
'''
_lowerCAmelCase = '''
Compute SuperGLUE evaluation metric associated to each SuperGLUE dataset.
Args:
predictions: list of predictions to score. Depending on the SuperGlUE subset:
- for \'record\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'prediction_text\': the predicted answer text
- for \'multirc\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question-answer pair as specified by the dataset
- \'prediction\': the predicted answer label
- otherwise: list of predicted labels
references: list of reference labels. Depending on the SuperGLUE subset:
- for \'record\': list of question-answers dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'answers\': list of possible answers
- otherwise: list of reference labels
Returns: depending on the SuperGLUE subset:
- for \'record\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1\': F1 score
- for \'multirc\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1_m\': Per-question macro-F1 score
- \'f1_a\': Average F1 score over all answers
- for \'axb\':
\'matthews_correlation\': Matthew Correlation
- for \'cb\':
- \'accuracy\': Accuracy
- \'f1\': F1 score
- for all others:
- \'accuracy\': Accuracy
Examples:
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'copa\') # any of ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'cb\')
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'record\')
>>> predictions = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'prediction_text\': \'answer\'}]
>>> references = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'answers\': [\'answer\', \'another_answer\']}]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'multirc\')
>>> predictions = [{\'idx\': {\'answer\': 0, \'paragraph\': 0, \'question\': 0}, \'prediction\': 0}, {\'idx\': {\'answer\': 1, \'paragraph\': 2, \'question\': 3}, \'prediction\': 1}]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1_m\': 1.0, \'f1_a\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'axb\')
>>> references = [0, 1]
>>> predictions = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'matthews_correlation\': 1.0}
'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return float((preds == labels).mean() )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase="binary" ):
"""simple docstring"""
lowerCAmelCase__ : Any = simple_accuracy(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = float(fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average=UpperCamelCase ) )
return {
"accuracy": acc,
"f1": fa,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
for id_pred, label in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : str = f"""{id_pred['idx']['paragraph']}-{id_pred['idx']['question']}"""
lowerCAmelCase__ : Dict = id_pred["""prediction"""]
if question_id in question_map:
question_map[question_id].append((pred, label) )
else:
lowerCAmelCase__ : Optional[int] = [(pred, label)]
lowerCAmelCase__ , lowerCAmelCase__ : int = [], []
for question, preds_labels in question_map.items():
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = zip(*UpperCamelCase )
lowerCAmelCase__ : List[Any] = fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average="""macro""" )
fas.append(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = int(sum(pred == label for pred, label in preds_labels ) == len(UpperCamelCase ) )
ems.append(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = float(sum(UpperCamelCase ) / len(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = sum(UpperCamelCase ) / len(UpperCamelCase )
lowerCAmelCase__ : Dict = float(fa_score(y_true=UpperCamelCase , y_pred=[id_pred["""prediction"""] for id_pred in ids_preds] ) )
return {"exact_match": em, "f1_m": fa_m, "f1_a": fa_a}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.config_name not in [
"boolq",
"cb",
"copa",
"multirc",
"record",
"rte",
"wic",
"wsc",
"wsc.fixed",
"axb",
"axg",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(self._get_feature_types() ) ,codebase_urls=[] ,reference_urls=[] ,format="""numpy""" if not self.config_name == """record""" and not self.config_name == """multirc""" else None ,)
def UpperCAmelCase_ ( self ) -> str:
if self.config_name == "record":
return {
"predictions": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"prediction_text": datasets.Value("""string""" ),
},
"references": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"answers": datasets.Sequence(datasets.Value("""string""" ) ),
},
}
elif self.config_name == "multirc":
return {
"predictions": {
"idx": {
"answer": datasets.Value("""int64""" ),
"paragraph": datasets.Value("""int64""" ),
"question": datasets.Value("""int64""" ),
},
"prediction": datasets.Value("""int64""" ),
},
"references": datasets.Value("""int64""" ),
}
else:
return {
"predictions": datasets.Value("""int64""" ),
"references": datasets.Value("""int64""" ),
}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
if self.config_name == "axb":
return {"matthews_correlation": matthews_corrcoef(__UpperCAmelCase ,__UpperCAmelCase )}
elif self.config_name == "cb":
return acc_and_fa(__UpperCAmelCase ,__UpperCAmelCase ,fa_avg="""macro""" )
elif self.config_name == "record":
lowerCAmelCase__ : Optional[Any] = [
{
"""qas""": [
{"""id""": ref["""idx"""]["""query"""], """answers""": [{"""text""": ans} for ans in ref["""answers"""]]}
for ref in references
]
}
]
lowerCAmelCase__ : Union[str, Any] = {pred["""idx"""]["""query"""]: pred["""prediction_text"""] for pred in predictions}
return evaluate_record(__UpperCAmelCase ,__UpperCAmelCase )[0]
elif self.config_name == "multirc":
return evaluate_multirc(__UpperCAmelCase ,__UpperCAmelCase )
elif self.config_name in ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]:
return {"accuracy": simple_accuracy(__UpperCAmelCase ,__UpperCAmelCase )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
| 37
| 1
|
'''simple docstring'''
_lowerCAmelCase = frozenset(
[
'''prompt''',
'''height''',
'''width''',
'''guidance_scale''',
'''negative_prompt''',
'''prompt_embeds''',
'''negative_prompt_embeds''',
'''cross_attention_kwargs''',
]
)
_lowerCAmelCase = frozenset(['''prompt''', '''negative_prompt'''])
_lowerCAmelCase = frozenset([])
_lowerCAmelCase = frozenset(['''image'''])
_lowerCAmelCase = frozenset(
[
'''image''',
'''height''',
'''width''',
'''guidance_scale''',
]
)
_lowerCAmelCase = frozenset(['''image'''])
_lowerCAmelCase = frozenset(
[
'''prompt''',
'''image''',
'''height''',
'''width''',
'''guidance_scale''',
'''negative_prompt''',
'''prompt_embeds''',
'''negative_prompt_embeds''',
]
)
_lowerCAmelCase = frozenset(['''prompt''', '''image''', '''negative_prompt'''])
_lowerCAmelCase = frozenset(
[
# Text guided image variation with an image mask
'''prompt''',
'''image''',
'''mask_image''',
'''height''',
'''width''',
'''guidance_scale''',
'''negative_prompt''',
'''prompt_embeds''',
'''negative_prompt_embeds''',
]
)
_lowerCAmelCase = frozenset(['''prompt''', '''image''', '''mask_image''', '''negative_prompt'''])
_lowerCAmelCase = frozenset(
[
# image variation with an image mask
'''image''',
'''mask_image''',
'''height''',
'''width''',
'''guidance_scale''',
]
)
_lowerCAmelCase = frozenset(['''image''', '''mask_image'''])
_lowerCAmelCase = frozenset(
[
'''example_image''',
'''image''',
'''mask_image''',
'''height''',
'''width''',
'''guidance_scale''',
]
)
_lowerCAmelCase = frozenset(['''example_image''', '''image''', '''mask_image'''])
_lowerCAmelCase = frozenset(['''class_labels'''])
_lowerCAmelCase = frozenset(['''class_labels'''])
_lowerCAmelCase = frozenset(['''batch_size'''])
_lowerCAmelCase = frozenset([])
_lowerCAmelCase = frozenset(['''batch_size'''])
_lowerCAmelCase = frozenset([])
_lowerCAmelCase = frozenset(
[
'''prompt''',
'''audio_length_in_s''',
'''guidance_scale''',
'''negative_prompt''',
'''prompt_embeds''',
'''negative_prompt_embeds''',
'''cross_attention_kwargs''',
]
)
_lowerCAmelCase = frozenset(['''prompt''', '''negative_prompt'''])
_lowerCAmelCase = frozenset(['''input_tokens'''])
_lowerCAmelCase = frozenset(['''input_tokens'''])
| 37
|
'''simple docstring'''
import unittest
from pathlib import Path
from tempfile import NamedTemporaryFile, TemporaryDirectory
from transformers import BertConfig, BertTokenizerFast, FeatureExtractionPipeline
from transformers.convert_graph_to_onnx import (
convert,
ensure_valid_input,
generate_identified_filename,
infer_shapes,
quantize,
)
from transformers.testing_utils import require_tf, require_tokenizers, require_torch, slow
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[Any]:
return None
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
return None
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
__lowercase : Dict = [
# (model_name, model_kwargs)
('''bert-base-cased''', {}),
('''gpt2''', {'''use_cache''': False}), # We don't support exporting GPT2 past keys anymore
]
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> int:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Any:
from transformers import BertModel
lowerCAmelCase__ : Optional[int] = ["""[UNK]""", """[SEP]""", """[CLS]""", """[PAD]""", """[MASK]""", """some""", """other""", """words"""]
with NamedTemporaryFile(mode="""w+t""" ) as vocab_file:
vocab_file.write("""\n""".join(__UpperCAmelCase ) )
vocab_file.flush()
lowerCAmelCase__ : Dict = BertTokenizerFast(vocab_file.name )
with TemporaryDirectory() as bert_save_dir:
lowerCAmelCase__ : Tuple = BertModel(BertConfig(vocab_size=len(__UpperCAmelCase ) ) )
model.save_pretrained(__UpperCAmelCase )
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,__UpperCAmelCase )
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> List[str]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Dict = self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = quantize(Path(__UpperCAmelCase ) )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> List[Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Any = self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = quantize(__UpperCAmelCase )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[Any]:
try:
# Compute path
with TemporaryDirectory() as tempdir:
lowerCAmelCase__ : Optional[int] = Path(__UpperCAmelCase ).joinpath("""model.onnx""" )
# Remove folder if exists
if path.parent.exists():
path.parent.rmdir()
# Export
convert(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,**__UpperCAmelCase )
return path
except Exception as e:
self.fail(__UpperCAmelCase )
@require_torch
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
from transformers import BertModel
lowerCAmelCase__ : List[Any] = BertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Union[str, Any] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""pt""" )
@require_tf
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Optional[int]:
from transformers import TFBertModel
lowerCAmelCase__ : int = TFBertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Optional[int] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""tf""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = FeatureExtractionPipeline(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = ["""input_ids""", """token_type_ids""", """attention_mask""", """output_0""", """output_1"""]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = infer_shapes(__UpperCAmelCase ,__UpperCAmelCase )
# Assert all variables are present
self.assertEqual(len(__UpperCAmelCase ) ,len(__UpperCAmelCase ) )
self.assertTrue(all(var_name in shapes for var_name in variable_names ) )
self.assertSequenceEqual(variable_names[:3] ,__UpperCAmelCase )
self.assertSequenceEqual(variable_names[3:] ,__UpperCAmelCase )
# Assert inputs are {0: batch, 1: sequence}
for var_name in ["input_ids", "token_type_ids", "attention_mask"]:
self.assertDictEqual(shapes[var_name] ,{0: """batch""", 1: """sequence"""} )
# Assert outputs are {0: batch, 1: sequence} and {0: batch}
self.assertDictEqual(shapes["""output_0"""] ,{0: """batch""", 1: """sequence"""} )
self.assertDictEqual(shapes["""output_1"""] ,{0: """batch"""} )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = ["""input_ids""", """attention_mask""", """token_type_ids"""]
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": [1, 2, 3, 4], """attention_mask""": [0, 0, 0, 0], """token_type_ids""": [1, 1, 1, 1]}
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = ensure_valid_input(FuncContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the same number of args (all are valid)
self.assertEqual(len(__UpperCAmelCase ) ,3 )
# Should have exactly the same input names
self.assertEqual(set(__UpperCAmelCase ) ,set(__UpperCAmelCase ) )
# Parameter should be reordered according to their respective place in the function:
# (input_ids, token_type_ids, attention_mask)
self.assertEqual(__UpperCAmelCase ,(tokens["""input_ids"""], tokens["""token_type_ids"""], tokens["""attention_mask"""]) )
# Generated args are interleaved with another args (for instance parameter "past" in GPT2)
lowerCAmelCase__ , lowerCAmelCase__ : int = ensure_valid_input(FuncNonContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the one arg (all before the one not provided "some_other_args")
self.assertEqual(len(__UpperCAmelCase ) ,1 )
self.assertEqual(len(__UpperCAmelCase ) ,1 )
# Should have only "input_ids"
self.assertEqual(inputs_args[0] ,tokens["""input_ids"""] )
self.assertEqual(ordered_input_names[0] ,"""input_ids""" )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Dict = generate_identified_filename(Path("""/home/something/my_fake_model.onnx""" ) ,"""-test""" )
self.assertEqual("""/home/something/my_fake_model-test.onnx""" ,generated.as_posix() )
| 37
| 1
|
'''simple docstring'''
import math
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = 100 ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = sum(i * i for i in range(1 , n + 1 ) )
lowerCAmelCase__ : Optional[int] = int(math.pow(sum(range(1 , n + 1 ) ) , 2 ) )
return square_of_sum - sum_of_squares
if __name__ == "__main__":
print(F"""{solution() = }""")
| 37
|
'''simple docstring'''
from maths.prime_factors import prime_factors
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if not isinstance(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : int = f"""Input value of [number={number}] must be an integer"""
raise TypeError(UpperCamelCase )
if number < 1:
raise ValueError("""Input must be a positive integer""" )
return -1 if len(prime_factors(UpperCamelCase ) ) % 2 else 1
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
| 1
|
'''simple docstring'''
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = XLMRobertaModel.from_pretrained("""xlm-roberta-base""" )
lowerCAmelCase__ : Optional[int] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : str = torch.Size((1, 12, 768) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.0_1_0_1, 0.1_2_1_8, -0.0_8_0_3, 0.0_8_0_1, 0.1_3_2_7, 0.0_7_7_6, -0.1_2_1_5, 0.2_3_8_3, 0.3_3_3_8, 0.3_1_0_6, 0.0_3_0_0, 0.0_2_5_2]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Optional[int] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
@slow
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : Union[str, Any] = XLMRobertaModel.from_pretrained("""xlm-roberta-large""" )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : Dict = torch.Size((1, 12, 1024) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[[-0.0_6_9_9, -0.0_3_1_8, 0.0_7_0_5, -0.1_2_4_1, 0.0_9_9_9, -0.0_5_2_0, 0.1_0_0_4, -0.1_8_3_8, -0.4_7_0_4, 0.1_4_3_7, 0.0_8_2_1, 0.0_1_2_6]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.large')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
| 37
|
'''simple docstring'''
import os
import re
import unicodedata
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import is_torch_available, logging
if is_torch_available():
import torch
if TYPE_CHECKING:
from transformers.pipelines.conversational import Conversation
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''AI-Sweden/gpt-sw3-126m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-126m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-350m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-350m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-1.6b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-1.6b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-6.7b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-6.7b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-20b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-20b/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''AI-Sweden/gpt-sw3-126m''': 2048,
'''AI-Sweden/gpt-sw3-350m''': 2048,
'''AI-Sweden/gpt-sw3-1.6b''': 2048,
'''AI-Sweden/gpt-sw3-6.7b''': 2048,
'''AI-Sweden/gpt-sw3-20b''': 2048,
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : str = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Optional[int] = ['''input_ids''', '''attention_mask''']
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
lowerCAmelCase__ : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
lowerCAmelCase__ : Dict = kwargs.get("""name_or_path""" )
if name_or_path is None:
logger.warning(
"""name_or_path not provided, will work for all GPTSw3 models except gpt-sw3-7b,"""
""" you are testing the model, this can safely be ignored""" )
lowerCAmelCase__ : Tuple = """None"""
# Default definitions for our 2 tokenizer versions, with None-checks to enable proper testing
lowerCAmelCase__ : Union[str, Any] = """<|endoftext|>""" if eos_token is None else eos_token
lowerCAmelCase__ : Dict = """<unk>""" if unk_token is None else unk_token
if "gpt-sw3-7b" in name_or_path:
lowerCAmelCase__ : Any = unk_token if pad_token is None else pad_token
lowerCAmelCase__ : Dict = eos_token if bos_token is None else bos_token
else:
lowerCAmelCase__ : List[str] = """<pad>""" if pad_token is None else pad_token
lowerCAmelCase__ : Optional[int] = """<s>""" if bos_token is None else bos_token
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : Optional[int] = do_lower_case
lowerCAmelCase__ : Dict = remove_space
lowerCAmelCase__ : Optional[Any] = keep_accents
lowerCAmelCase__ : int = vocab_file
lowerCAmelCase__ : str = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
# Used for whitespace normalization in input texts
# fmt : off
lowerCAmelCase__ : int = {""" """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """""", """"""}
# fmt : on
# Regular expression to remove non-printing characters (e.g. some unicode control chars) in preprocessing
lowerCAmelCase__ : List[str] = re.compile(
F"""[{''.join(map(__UpperCAmelCase ,list(range(0 ,9 ) ) + list(range(11 ,32 ) ) + list(range(127 ,160 ) ) + [160, 173, 8203] ) )}]""" )
def __getstate__( self ) -> Any:
lowerCAmelCase__ : int = self.__dict__.copy()
lowerCAmelCase__ : Optional[int] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Tuple = {}
lowerCAmelCase__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
@property
# Copied from transformers.models.albert.tokenization_albert.AlbertTokenizer.vocab_size
def UpperCAmelCase_ ( self ) -> int:
return len(self.sp_model )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : Tuple = self.non_printing_characters_re.sub("""""" ,__UpperCAmelCase )
# Normalize whitespaces
lowerCAmelCase__ : List[Any] = """""".join([char if char not in self.whitespaces else """ """ for char in text] )
# NFC Unicode normalization
lowerCAmelCase__ : List[Any] = unicodedata.normalize("""NFC""" ,__UpperCAmelCase )
return text
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[Any] = self.preprocess_text(__UpperCAmelCase )
return self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.IdToPiece(__UpperCAmelCase )
@staticmethod
def UpperCAmelCase_ ( __UpperCAmelCase ) -> str:
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : int = []
lowerCAmelCase__ : Optional[int] = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
# TODO: Check if this is needed, as it ensures that decode(encode(doc)) != doc by adding extra whitespace in the decoded document
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Optional[Any] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Any = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string
def UpperCAmelCase_ ( self ) -> Dict[str, int]:
lowerCAmelCase__ : Optional[int] = {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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : Optional[int] = 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:
lowerCAmelCase__ : str = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Union[List[int], List[List[int]], "torch.Tensor"]:
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : int = self.sp_model.encode(__UpperCAmelCase )
else:
lowerCAmelCase__ : int = [self.preprocess_text(__UpperCAmelCase ) for t in text]
lowerCAmelCase__ : Any = self.sp_model.encode(__UpperCAmelCase )
if return_tensors is True or return_tensors == "pt":
lowerCAmelCase__ : Tuple = torch.tensor(__UpperCAmelCase )
return token_ids
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.decode(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[int]:
lowerCAmelCase__ : List[Any] = [F"""User: {text}""" if is_user else F"""Bot: {text}""" for is_user, text in conversation.iter_texts()]
lowerCAmelCase__ : Any = (
F"""{self.eos_token}{self.bos_token}""" + F"""{self.bos_token}""".join(__UpperCAmelCase ) + F"""{self.bos_token}Bot:"""
)
return self.encode(text=__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
#
# This a `torch.distributed` diagnostics script that checks that all GPUs in the cluster (one or
# many nodes) can talk to each other via nccl and allocate gpu memory.
#
# To run first adjust the number of processes and nodes:
#
# python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
#
# You may need to add --master_addr $MASTER_ADDR --master_port $MASTER_PORT if using a custom addr:port
#
# You can also use the rdzv API: --rdzv_endpoint $MASTER_ADDR:$MASTER_PORT --rdzv_backend c10d
#
# use torch.distributed.launch instead of torch.distributed.run for torch < 1.9
#
# If you get a hanging in `barrier` calls you have some network issues, you may try to debug this with:
#
# NCCL_DEBUG=INFO python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
#
# which should tell you what's going on behind the scenes.
#
#
# This script can be run via `srun` in the SLURM environment as well. Here is a SLURM script that
# runs on 2 nodes of 4 gpus per node:
#
# #SBATCH --job-name=test-nodes # name
# #SBATCH --nodes=2 # nodes
# #SBATCH --ntasks-per-node=1 # crucial - only 1 task per dist per node!
# #SBATCH --cpus-per-task=10 # number of cores per tasks
# #SBATCH --gres=gpu:4 # number of gpus
# #SBATCH --time 0:05:00 # maximum execution time (HH:MM:SS)
# #SBATCH --output=%x-%j.out # output file name
#
# GPUS_PER_NODE=4
# MASTER_ADDR=$(scontrol show hostnames $SLURM_JOB_NODELIST | head -n 1)
# MASTER_PORT=6000
#
# srun --jobid $SLURM_JOBID bash -c 'python -m torch.distributed.run \
# --nproc_per_node $GPUS_PER_NODE --nnodes $SLURM_NNODES --node_rank $SLURM_PROCID \
# --master_addr $MASTER_ADDR --master_port $MASTER_PORT \
# torch-distributed-gpu-test.py'
#
import fcntl
import os
import socket
import torch
import torch.distributed as dist
def _SCREAMING_SNAKE_CASE ( *UpperCamelCase ):
"""simple docstring"""
with open(UpperCamelCase , """r""" ) as fh:
fcntl.flock(UpperCamelCase , fcntl.LOCK_EX )
try:
print(*UpperCamelCase )
finally:
fcntl.flock(UpperCamelCase , fcntl.LOCK_UN )
_lowerCAmelCase = int(os.environ['''LOCAL_RANK'''])
torch.cuda.set_device(local_rank)
_lowerCAmelCase = torch.device('''cuda''', local_rank)
_lowerCAmelCase = socket.gethostname()
_lowerCAmelCase = F"""[{hostname}-{local_rank}]"""
try:
# test distributed
dist.init_process_group('''nccl''')
dist.all_reduce(torch.ones(1).to(device), op=dist.ReduceOp.SUM)
dist.barrier()
# test cuda is available and can allocate memory
torch.cuda.is_available()
torch.ones(1).cuda(local_rank)
# global rank
_lowerCAmelCase = dist.get_rank()
_lowerCAmelCase = dist.get_world_size()
printflock(F"""{gpu} is OK (global rank: {rank}/{world_size})""")
dist.barrier()
if rank == 0:
printflock(F"""pt={torch.__version__}, cuda={torch.version.cuda}, nccl={torch.cuda.nccl.version()}""")
except Exception:
printflock(F"""{gpu} is broken""")
raise
| 37
|
'''simple docstring'''
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = XLMRobertaModel.from_pretrained("""xlm-roberta-base""" )
lowerCAmelCase__ : Optional[int] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : str = torch.Size((1, 12, 768) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.0_1_0_1, 0.1_2_1_8, -0.0_8_0_3, 0.0_8_0_1, 0.1_3_2_7, 0.0_7_7_6, -0.1_2_1_5, 0.2_3_8_3, 0.3_3_3_8, 0.3_1_0_6, 0.0_3_0_0, 0.0_2_5_2]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Optional[int] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
@slow
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : Union[str, Any] = XLMRobertaModel.from_pretrained("""xlm-roberta-large""" )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : Dict = torch.Size((1, 12, 1024) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[[-0.0_6_9_9, -0.0_3_1_8, 0.0_7_0_5, -0.1_2_4_1, 0.0_9_9_9, -0.0_5_2_0, 0.1_0_0_4, -0.1_8_3_8, -0.4_7_0_4, 0.1_4_3_7, 0.0_8_2_1, 0.0_1_2_6]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.large')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
| 37
| 1
|
'''simple docstring'''
import torch
from diffusers import DiffusionPipeline
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
super().__init__()
self.register_modules(unet=__UpperCAmelCase ,scheduler=__UpperCAmelCase )
def __call__( self ) -> str:
lowerCAmelCase__ : Tuple = torch.randn(
(1, self.unet.config.in_channels, self.unet.config.sample_size, self.unet.config.sample_size) ,)
lowerCAmelCase__ : Union[str, Any] = 1
lowerCAmelCase__ : Union[str, Any] = self.unet(__UpperCAmelCase ,__UpperCAmelCase ).sample
lowerCAmelCase__ : List[Any] = self.scheduler.step(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ).prev_sample
lowerCAmelCase__ : Optional[int] = scheduler_output - scheduler_output + torch.ones_like(__UpperCAmelCase )
return result
| 37
|
'''simple docstring'''
from pickle import UnpicklingError
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict
from ..utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
with open(UpperCamelCase , """rb""" ) as flax_state_f:
lowerCAmelCase__ : Union[str, Any] = from_bytes(UpperCamelCase , flax_state_f.read() )
except UnpicklingError as e:
try:
with open(UpperCamelCase ) as f:
if f.read().startswith("""version""" ):
raise OSError(
"""You seem to have cloned a repository without having git-lfs installed. Please"""
""" install git-lfs and run `git lfs install` followed by `git lfs pull` in the"""
""" folder you cloned.""" )
else:
raise ValueError from e
except (UnicodeDecodeError, ValueError):
raise EnvironmentError(f"""Unable to convert {model_file} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(UpperCamelCase , UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
import torch # noqa: F401
except ImportError:
logger.error(
"""Loading Flax weights in PyTorch requires both PyTorch and Flax to be installed. Please see"""
""" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"""
""" instructions.""" )
raise
# check if we have bf16 weights
lowerCAmelCase__ : str = flatten_dict(jax.tree_util.tree_map(lambda UpperCamelCase : x.dtype == jnp.bfloataa , UpperCamelCase ) ).values()
if any(UpperCamelCase ):
# convert all weights to fp32 if they are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
"""Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` """
"""before loading those in PyTorch model.""" )
lowerCAmelCase__ : Dict = jax.tree_util.tree_map(
lambda UpperCamelCase : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , UpperCamelCase )
lowerCAmelCase__ : Any = """"""
lowerCAmelCase__ : Any = flatten_dict(UpperCamelCase , sep=""".""" )
lowerCAmelCase__ : Optional[int] = pt_model.state_dict()
# keep track of unexpected & missing keys
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : int = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
lowerCAmelCase__ : Union[str, Any] = flax_key_tuple.split(""".""" )
if flax_key_tuple_array[-1] == "kernel" and flax_tensor.ndim == 4:
lowerCAmelCase__ : Optional[int] = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = jnp.transpose(UpperCamelCase , (3, 2, 0, 1) )
elif flax_key_tuple_array[-1] == "kernel":
lowerCAmelCase__ : str = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = flax_tensor.T
elif flax_key_tuple_array[-1] == "scale":
lowerCAmelCase__ : int = flax_key_tuple_array[:-1] + ["""weight"""]
if "time_embedding" not in flax_key_tuple_array:
for i, flax_key_tuple_string in enumerate(UpperCamelCase ):
lowerCAmelCase__ : List[str] = (
flax_key_tuple_string.replace("""_0""" , """.0""" )
.replace("""_1""" , """.1""" )
.replace("""_2""" , """.2""" )
.replace("""_3""" , """.3""" )
.replace("""_4""" , """.4""" )
.replace("""_5""" , """.5""" )
.replace("""_6""" , """.6""" )
.replace("""_7""" , """.7""" )
.replace("""_8""" , """.8""" )
.replace("""_9""" , """.9""" )
)
lowerCAmelCase__ : Union[str, Any] = """.""".join(UpperCamelCase )
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
f"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
f"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
lowerCAmelCase__ : int = np.asarray(UpperCamelCase ) if not isinstance(UpperCamelCase , np.ndarray ) else flax_tensor
lowerCAmelCase__ : int = torch.from_numpy(UpperCamelCase )
# remove from missing keys
missing_keys.remove(UpperCamelCase )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(UpperCamelCase )
pt_model.load_state_dict(UpperCamelCase )
# re-transform missing_keys to list
lowerCAmelCase__ : Optional[int] = list(UpperCamelCase )
if len(UpperCamelCase ) > 0:
logger.warning(
"""Some weights of the Flax model were not used when initializing the PyTorch model"""
f""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
f""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
""" (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This"""
f""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
""" to be exactly identical (e.g. initializing a BertForSequenceClassification model from a"""
""" FlaxBertForSequenceClassification model).""" )
if len(UpperCamelCase ) > 0:
logger.warning(
f"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
f""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
""" use it for predictions and inference.""" )
return pt_model
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = 50 ):
"""simple docstring"""
lowerCAmelCase__ : str = [1] * (length + 1)
for row_length in range(3 , length + 1 ):
for block_length in range(3 , row_length + 1 ):
for block_start in range(row_length - block_length ):
ways_number[row_length] += ways_number[
row_length - block_start - block_length - 1
]
ways_number[row_length] += 1
return ways_number[length]
if __name__ == "__main__":
print(F"""{solution() = }""")
| 37
|
'''simple docstring'''
import inspect
import unittest
import numpy as np
from tests.test_modeling_common import floats_tensor
from transformers import DetrConfig, MaskFormerConfig, SwinConfig, is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device
from transformers.utils import cached_property
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaskFormerForInstanceSegmentation, MaskFormerModel
if is_vision_available():
from transformers import MaskFormerImageProcessor
if is_vision_available():
from PIL import Image
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=2 ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase=10 ,__UpperCAmelCase=3 ,__UpperCAmelCase=32 * 4 ,__UpperCAmelCase=32 * 6 ,__UpperCAmelCase=4 ,__UpperCAmelCase=32 ,) -> str:
lowerCAmelCase__ : Optional[int] = parent
lowerCAmelCase__ : Optional[int] = batch_size
lowerCAmelCase__ : Optional[int] = is_training
lowerCAmelCase__ : Dict = use_auxiliary_loss
lowerCAmelCase__ : Union[str, Any] = num_queries
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : List[str] = min_size
lowerCAmelCase__ : int = max_size
lowerCAmelCase__ : Optional[Any] = num_labels
lowerCAmelCase__ : List[Any] = mask_feature_size
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : str = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to(
__UpperCAmelCase )
lowerCAmelCase__ : str = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=__UpperCAmelCase )
lowerCAmelCase__ : Any = (
torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=__UpperCAmelCase ) > 0.5
).float()
lowerCAmelCase__ : Optional[int] = (torch.rand((self.batch_size, self.num_labels) ,device=__UpperCAmelCase ) > 0.5).long()
lowerCAmelCase__ : Any = self.get_config()
return config, pixel_values, pixel_mask, mask_labels, class_labels
def UpperCAmelCase_ ( self ) -> Dict:
return MaskFormerConfig.from_backbone_and_decoder_configs(
backbone_config=SwinConfig(
depths=[1, 1, 1, 1] ,) ,decoder_config=DetrConfig(
decoder_ffn_dim=128 ,num_queries=self.num_queries ,decoder_attention_heads=2 ,d_model=self.mask_feature_size ,) ,mask_feature_size=self.mask_feature_size ,fpn_feature_size=self.mask_feature_size ,num_channels=self.num_channels ,num_labels=self.num_labels ,)
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.prepare_config_and_inputs()
lowerCAmelCase__ : List[str] = {"""pixel_values""": pixel_values, """pixel_mask""": pixel_mask}
return config, inputs_dict
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Optional[int] = output.encoder_hidden_states
lowerCAmelCase__ : Optional[int] = output.pixel_decoder_hidden_states
lowerCAmelCase__ : Dict = output.transformer_decoder_hidden_states
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,config.decoder_config.decoder_layers )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=False ) -> Optional[Any]:
with torch.no_grad():
lowerCAmelCase__ : int = MaskFormerModel(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
lowerCAmelCase__ : str = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : int = model(__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
# the correct shape of output.transformer_decoder_hidden_states ensure the correcteness of the
# encoder and pixel decoder
self.parent.assertEqual(
output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.mask_feature_size) ,)
# let's ensure the other two hidden state exists
self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(output.encoder_last_hidden_state is not None )
if output_hidden_states:
self.check_output_hidden_state(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : Dict = MaskFormerForInstanceSegmentation(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
def comm_check_on_output(__UpperCAmelCase ):
# let's still check that all the required stuff is there
self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.encoder_last_hidden_state is not None )
# okay, now we need to check the logits shape
# due to the encoder compression, masks have a //4 spatial size
self.parent.assertEqual(
result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,)
# + 1 for null class
self.parent.assertEqual(
result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : Dict = model(__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = model(
pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
self.parent.assertTrue(result.loss is not None )
self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) )
@require_torch
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Optional[Any] = (MaskFormerModel, MaskFormerForInstanceSegmentation) if is_torch_available() else ()
__lowercase : int = (
{'''feature-extraction''': MaskFormerModel, '''image-segmentation''': MaskFormerForInstanceSegmentation}
if is_torch_available()
else {}
)
__lowercase : Union[str, Any] = False
__lowercase : Dict = False
__lowercase : Tuple = False
__lowercase : List[Any] = False
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : str = MaskFormerModelTester(self )
lowerCAmelCase__ : List[Any] = ConfigTester(self ,config_class=__UpperCAmelCase ,has_text_modality=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
self.config_tester.run_common_tests()
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ , lowerCAmelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_maskformer_instance_segmentation_head_model(*__UpperCAmelCase )
@unittest.skip(reason="""MaskFormer does not use inputs_embeds""" )
def UpperCAmelCase_ ( self ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormer does not have a get_input_embeddings method""" )
def UpperCAmelCase_ ( self ) -> str:
pass
@unittest.skip(reason="""MaskFormer is not a generative model""" )
def UpperCAmelCase_ ( self ) -> Any:
pass
@unittest.skip(reason="""MaskFormer does not use token embeddings""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
@require_torch_multi_gpu
@unittest.skip(
reason="""MaskFormer has some layers using `add_module` which doesn't work well with `nn.DataParallel`""" )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
pass
@unittest.skip("""Will be fixed soon by reducing the size of the model used for common tests.""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : str = model_class(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowerCAmelCase__ : Dict = [*signature.parameters.keys()]
lowerCAmelCase__ : Tuple = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] ,__UpperCAmelCase )
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model_name in ["facebook/maskformer-swin-small-coco"]:
lowerCAmelCase__ : List[str] = MaskFormerModel.from_pretrained(__UpperCAmelCase )
self.assertIsNotNone(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (self.model_tester.min_size,) * 2
lowerCAmelCase__ : Any = {
"""pixel_values""": torch.randn((2, 3, *size) ,device=__UpperCAmelCase ),
"""mask_labels""": torch.randn((2, 10, *size) ,device=__UpperCAmelCase ),
"""class_labels""": torch.zeros(2 ,10 ,device=__UpperCAmelCase ).long(),
}
lowerCAmelCase__ : Tuple = MaskFormerForInstanceSegmentation(MaskFormerConfig() ).to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : int = model_class(__UpperCAmelCase ).to(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase ,output_attentions=__UpperCAmelCase )
self.assertTrue(outputs.attentions is not None )
def UpperCAmelCase_ ( self ) -> int:
if not self.model_tester.is_training:
return
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Dict = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : List[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : List[str] = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase ).loss
loss.backward()
def UpperCAmelCase_ ( self ) -> List[str]:
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Tuple = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Tuple = True
lowerCAmelCase__ : Optional[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : Dict = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
lowerCAmelCase__ : str = outputs.pixel_decoder_hidden_states[0]
pixel_decoder_hidden_states.retain_grad()
# we requires_grad=True in inputs_embeds (line 2152), the original implementation don't
lowerCAmelCase__ : Union[str, Any] = outputs.transformer_decoder_hidden_states[0]
transformer_decoder_hidden_states.retain_grad()
lowerCAmelCase__ : List[Any] = outputs.attentions[0]
attentions.retain_grad()
outputs.loss.backward(retain_graph=__UpperCAmelCase )
self.assertIsNotNone(encoder_hidden_states.grad )
self.assertIsNotNone(pixel_decoder_hidden_states.grad )
self.assertIsNotNone(transformer_decoder_hidden_states.grad )
self.assertIsNotNone(attentions.grad )
_lowerCAmelCase = 1e-4
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_vision
@slow
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@cached_property
def UpperCAmelCase_ ( self ) -> List[Any]:
return (
MaskFormerImageProcessor.from_pretrained("""facebook/maskformer-swin-small-coco""" )
if is_vision_available()
else None
)
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Any = MaskFormerModel.from_pretrained("""facebook/maskformer-swin-small-coco""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = self.default_image_processor
lowerCAmelCase__ : str = prepare_img()
lowerCAmelCase__ : Optional[int] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[-0.0_4_8_2, 0.9_2_2_8, 0.4_9_5_1], [-0.2_5_4_7, 0.8_0_1_7, 0.8_5_2_7], [-0.0_0_6_9, 0.3_3_8_5, -0.0_0_8_9]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.encoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.8_4_2_2, -0.8_4_3_4, -0.9_7_1_8], [-1.0_1_4_4, -0.5_5_6_5, -0.4_1_9_5], [-1.0_0_3_8, -0.4_4_8_4, -0.1_9_6_1]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Optional[int] = torch.tensor(
[[0.2_8_5_2, -0.0_1_5_9, 0.9_7_3_5], [0.6_2_5_4, 0.1_8_5_8, 0.8_5_2_9], [-0.0_6_8_0, -0.4_1_1_6, 1.8_4_1_3]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : List[str] = prepare_img()
lowerCAmelCase__ : str = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[int] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : Optional[int] = [
[-1.3_7_3_7_1_2_4, -1.7_7_2_4_9_3_7, -1.9_3_6_4_2_3_3],
[-1.5_9_7_7_2_8_1, -1.9_8_6_7_9_3_9, -2.1_5_2_3_6_9_5],
[-1.5_7_9_5_3_9_8, -1.9_2_6_9_8_3_2, -2.0_9_3_9_4_2],
]
lowerCAmelCase__ : Optional[int] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Tuple = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[
[1.65_12E00, -5.25_72E00, -3.35_19E00],
[3.61_69E-02, -5.90_25E00, -2.93_13E00],
[1.07_66E-04, -7.76_30E00, -5.12_63E00],
] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-resnet101-coco-stuff""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : int = prepare_img()
lowerCAmelCase__ : Optional[Any] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : str = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : int = [[-0.9_0_4_6, -2.6_3_6_6, -4.6_0_6_2], [-3.4_1_7_9, -5.7_8_9_0, -8.8_0_5_7], [-4.9_1_7_9, -7.6_5_6_0, -1_0.7_7_1_1]]
lowerCAmelCase__ : List[str] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Tuple = torch.tensor(
[[4.7_1_8_8, -3.2_5_8_5, -2.8_8_5_7], [6.6_8_7_1, -2.9_1_8_1, -1.2_4_8_7], [7.2_4_4_9, -2.2_7_6_4, -2.1_8_7_4]] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : str = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Dict = self.default_image_processor
lowerCAmelCase__ : Union[str, Any] = image_processor(
[np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors="""pt""" ,)
lowerCAmelCase__ : Tuple = inputs["""pixel_values"""].to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""mask_labels"""]]
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""class_labels"""]]
with torch.no_grad():
lowerCAmelCase__ : Any = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
| 37
| 1
|
'''simple docstring'''
import itertools
import json
import os
import unittest
from transformers import AddedToken, LongformerTokenizer, LongformerTokenizerFast
from transformers.models.longformer.tokenization_longformer import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers, slow
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Optional[int] = LongformerTokenizer
__lowercase : Union[str, Any] = True
__lowercase : Any = LongformerTokenizerFast
__lowercase : int = True
def UpperCAmelCase_ ( self ) -> List[str]:
super().setUp()
# Adapted from Sennrich et al. 2015 and https://github.com/rsennrich/subword-nmt
lowerCAmelCase__ : str = [
"""l""",
"""o""",
"""w""",
"""e""",
"""r""",
"""s""",
"""t""",
"""i""",
"""d""",
"""n""",
"""\u0120""",
"""\u0120l""",
"""\u0120n""",
"""\u0120lo""",
"""\u0120low""",
"""er""",
"""\u0120lowest""",
"""\u0120newer""",
"""\u0120wider""",
"""<unk>""",
]
lowerCAmelCase__ : Tuple = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : int = ["""#version: 0.2""", """\u0120 l""", """\u0120l o""", """\u0120lo w""", """e r""", """"""]
lowerCAmelCase__ : Union[str, Any] = {"""unk_token""": """<unk>"""}
lowerCAmelCase__ : List[Any] = os.path.join(self.tmpdirname ,VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : int = 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 ,**__UpperCAmelCase ) -> Union[str, Any]:
kwargs.update(self.special_tokens_map )
return self.tokenizer_class.from_pretrained(self.tmpdirname ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> List[str]:
kwargs.update(self.special_tokens_map )
return self.rust_tokenizer_class.from_pretrained(self.tmpdirname ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Union[str, Any] = """lower newer"""
lowerCAmelCase__ : Union[str, Any] = """lower newer"""
return input_text, output_text
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = self.tokenizer_class(self.vocab_file ,self.merges_file ,**self.special_tokens_map )
lowerCAmelCase__ : Optional[Any] = """lower newer"""
lowerCAmelCase__ : List[Any] = ["""l""", """o""", """w""", """er""", """\u0120""", """n""", """e""", """w""", """er"""]
lowerCAmelCase__ : Any = tokenizer.tokenize(__UpperCAmelCase ) # , add_prefix_space=True)
self.assertListEqual(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tokens + [tokenizer.unk_token]
lowerCAmelCase__ : int = [0, 1, 2, 15, 10, 9, 3, 2, 15, 19]
self.assertListEqual(tokenizer.convert_tokens_to_ids(__UpperCAmelCase ) ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[Any]:
lowerCAmelCase__ : Optional[Any] = self.get_tokenizer()
self.assertListEqual(tokenizer.encode("""Hello world!""" ,add_special_tokens=__UpperCAmelCase ) ,[0, 3_1414, 232, 328, 2] )
self.assertListEqual(
tokenizer.encode("""Hello world! cécé herlolip 418""" ,add_special_tokens=__UpperCAmelCase ) ,[0, 3_1414, 232, 328, 740, 1140, 1_2695, 69, 4_6078, 1588, 2] ,)
@slow
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Optional[Any] = self.tokenizer_class.from_pretrained("""allenai/longformer-base-4096""" )
lowerCAmelCase__ : List[str] = tokenizer.encode("""sequence builders""" ,add_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : Any = tokenizer.encode("""multi-sequence build""" ,add_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : int = tokenizer.encode(
"""sequence builders""" ,add_special_tokens=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = tokenizer.encode(
"""sequence builders""" ,"""multi-sequence build""" ,add_special_tokens=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = tokenizer.build_inputs_with_special_tokens(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = tokenizer.build_inputs_with_special_tokens(__UpperCAmelCase ,__UpperCAmelCase )
assert encoded_sentence == encoded_text_from_decode
assert encoded_pair == encoded_pair_from_decode
def UpperCAmelCase_ ( self ) -> Dict:
lowerCAmelCase__ : Union[str, Any] = self.get_tokenizer()
lowerCAmelCase__ : List[str] = """Encode this sequence."""
lowerCAmelCase__ : Tuple = tokenizer.byte_encoder[""" """.encode("""utf-8""" )[0]]
# Testing encoder arguments
lowerCAmelCase__ : str = tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = tokenizer.convert_ids_to_tokens(encoded[0] )[0]
self.assertNotEqual(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = tokenizer.convert_ids_to_tokens(encoded[0] )[0]
self.assertEqual(__UpperCAmelCase ,__UpperCAmelCase )
tokenizer.add_special_tokens({"""bos_token""": """<s>"""} )
lowerCAmelCase__ : int = tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tokenizer.convert_ids_to_tokens(encoded[1] )[0]
self.assertNotEqual(__UpperCAmelCase ,__UpperCAmelCase )
# Testing spaces after special tokens
lowerCAmelCase__ : Optional[int] = """<mask>"""
tokenizer.add_special_tokens(
{"""mask_token""": AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase )} ) # mask token has a left space
lowerCAmelCase__ : Optional[int] = tokenizer.convert_tokens_to_ids(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = """Encode <mask> sequence"""
lowerCAmelCase__ : int = """Encode <mask>sequence"""
lowerCAmelCase__ : str = tokenizer.encode(__UpperCAmelCase )
lowerCAmelCase__ : int = encoded.index(__UpperCAmelCase )
lowerCAmelCase__ : Any = tokenizer.convert_ids_to_tokens(encoded[mask_loc + 1] )[0]
self.assertEqual(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = tokenizer.encode(__UpperCAmelCase )
lowerCAmelCase__ : str = encoded.index(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tokenizer.convert_ids_to_tokens(encoded[mask_loc + 1] )[0]
self.assertNotEqual(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
def UpperCAmelCase_ ( self ) -> Any:
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(F"""{tokenizer.__class__.__name__} ({pretrained_name})""" ):
lowerCAmelCase__ : Union[str, Any] = self.rust_tokenizer_class.from_pretrained(__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.tokenizer_class.from_pretrained(__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = """A, <mask> AllenNLP sentence."""
lowerCAmelCase__ : int = tokenizer_r.encode_plus(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ,return_token_type_ids=__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = tokenizer_p.encode_plus(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ,return_token_type_ids=__UpperCAmelCase )
# token_type_ids should put 0 everywhere
self.assertEqual(sum(tokens_r["""token_type_ids"""] ) ,sum(tokens_p["""token_type_ids"""] ) )
# attention_mask should put 1 everywhere, so sum over length should be 1
self.assertEqual(
sum(tokens_r["""attention_mask"""] ) / len(tokens_r["""attention_mask"""] ) ,sum(tokens_p["""attention_mask"""] ) / len(tokens_p["""attention_mask"""] ) ,)
lowerCAmelCase__ : Union[str, Any] = tokenizer_r.convert_ids_to_tokens(tokens_r["""input_ids"""] )
lowerCAmelCase__ : List[str] = tokenizer_p.convert_ids_to_tokens(tokens_p["""input_ids"""] )
# Rust correctly handles the space before the mask while python doesnt
self.assertSequenceEqual(tokens_p["""input_ids"""] ,[0, 250, 6, 5_0264, 3823, 487, 2_1992, 3645, 4, 2] )
self.assertSequenceEqual(tokens_r["""input_ids"""] ,[0, 250, 6, 5_0264, 3823, 487, 2_1992, 3645, 4, 2] )
self.assertSequenceEqual(
__UpperCAmelCase ,["""<s>""", """A""", """,""", """<mask>""", """ĠAllen""", """N""", """LP""", """Ġsentence""", """.""", """</s>"""] )
self.assertSequenceEqual(
__UpperCAmelCase ,["""<s>""", """A""", """,""", """<mask>""", """ĠAllen""", """N""", """LP""", """Ġsentence""", """.""", """</s>"""] )
def UpperCAmelCase_ ( self ) -> int:
for trim_offsets, add_prefix_space in itertools.product([True, False] ,repeat=2 ):
lowerCAmelCase__ : Tuple = self.rust_tokenizer_class.from_pretrained(
self.tmpdirname ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : int = json.loads(tokenizer_r.backend_tokenizer.pre_tokenizer.__getstate__() )
lowerCAmelCase__ : List[str] = json.loads(tokenizer_r.backend_tokenizer.post_processor.__getstate__() )
self.assertEqual(pre_tokenizer_state["""add_prefix_space"""] ,__UpperCAmelCase )
self.assertEqual(post_processor_state["""add_prefix_space"""] ,__UpperCAmelCase )
self.assertEqual(post_processor_state["""trim_offsets"""] ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
# Test which aims to verify that the offsets are well adapted to the argument `add_prefix_space` and
# `trim_offsets`
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(F"""{tokenizer.__class__.__name__} ({pretrained_name})""" ):
lowerCAmelCase__ : Dict = """hello""" # `hello` is a token in the vocabulary of `pretrained_name`
lowerCAmelCase__ : Any = F"""{text_of_1_token} {text_of_1_token}"""
lowerCAmelCase__ : Tuple = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(0, len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(len(__UpperCAmelCase ) + 1, len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
lowerCAmelCase__ : Optional[int] = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : Any = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(0, len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(len(__UpperCAmelCase ) + 1, len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
lowerCAmelCase__ : Union[str, Any] = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(0, len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(len(__UpperCAmelCase ), len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
lowerCAmelCase__ : Tuple = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(0, len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(len(__UpperCAmelCase ), len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
lowerCAmelCase__ : int = F""" {text}"""
# tokenizer_r = self.rust_tokenizer_class.from_pretrained(
# pretrained_name, use_fast=True, add_prefix_space=True, trim_offsets=True
# )
# encoding = tokenizer_r(text, return_offsets_mapping=True, add_special_tokens=False)
# self.assertEqual(encoding.offset_mapping[0], (1, 1 + len(text_of_1_token)))
# self.assertEqual(
# encoding.offset_mapping[1],
# (1 + len(text_of_1_token) + 1, 1 + len(text_of_1_token) + 1 + len(text_of_1_token)),
# )
lowerCAmelCase__ : Tuple = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : Any = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(1, 1 + len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(1 + len(__UpperCAmelCase ) + 1, 1 + len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
lowerCAmelCase__ : Any = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : str = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(0, 1 + len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(1 + len(__UpperCAmelCase ), 1 + len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
lowerCAmelCase__ : Optional[Any] = self.rust_tokenizer_class.from_pretrained(
__UpperCAmelCase ,use_fast=__UpperCAmelCase ,add_prefix_space=__UpperCAmelCase ,trim_offsets=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = tokenizer_r(__UpperCAmelCase ,return_offsets_mapping=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
self.assertEqual(encoding.offset_mapping[0] ,(0, 1 + len(__UpperCAmelCase )) )
self.assertEqual(
encoding.offset_mapping[1] ,(1 + len(__UpperCAmelCase ), 1 + len(__UpperCAmelCase ) + 1 + len(__UpperCAmelCase )) ,)
| 37
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''microsoft/focalnet-tiny''': '''https://huggingface.co/microsoft/focalnet-tiny/resolve/main/config.json''',
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[Any] = '''focalnet'''
def __init__( self ,__UpperCAmelCase=224 ,__UpperCAmelCase=4 ,__UpperCAmelCase=3 ,__UpperCAmelCase=96 ,__UpperCAmelCase=False ,__UpperCAmelCase=[192, 384, 768, 768] ,__UpperCAmelCase=[2, 2, 6, 2] ,__UpperCAmelCase=[2, 2, 2, 2] ,__UpperCAmelCase=[3, 3, 3, 3] ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=4.0 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=False ,__UpperCAmelCase=1E-4 ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-5 ,__UpperCAmelCase=32 ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Optional[Any]:
super().__init__(**__UpperCAmelCase )
lowerCAmelCase__ : Dict = image_size
lowerCAmelCase__ : int = patch_size
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : Dict = embed_dim
lowerCAmelCase__ : List[str] = use_conv_embed
lowerCAmelCase__ : List[Any] = hidden_sizes
lowerCAmelCase__ : Dict = depths
lowerCAmelCase__ : List[str] = focal_levels
lowerCAmelCase__ : List[str] = focal_windows
lowerCAmelCase__ : Dict = hidden_act
lowerCAmelCase__ : Dict = mlp_ratio
lowerCAmelCase__ : Tuple = hidden_dropout_prob
lowerCAmelCase__ : Tuple = drop_path_rate
lowerCAmelCase__ : Dict = use_layerscale
lowerCAmelCase__ : Optional[Any] = layerscale_value
lowerCAmelCase__ : str = use_post_layernorm
lowerCAmelCase__ : Union[str, Any] = use_post_layernorm_in_modulation
lowerCAmelCase__ : int = normalize_modulator
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : List[str] = layer_norm_eps
lowerCAmelCase__ : List[Any] = encoder_stride
lowerCAmelCase__ : Dict = ["""stem"""] + [F"""stage{idx}""" for idx in range(1 ,len(self.depths ) + 1 )]
lowerCAmelCase__ , lowerCAmelCase__ : Any = get_aligned_output_features_output_indices(
out_features=__UpperCAmelCase ,out_indices=__UpperCAmelCase ,stage_names=self.stage_names )
| 37
| 1
|
'''simple docstring'''
import json
import os
import tempfile
import unittest
import numpy as np
from datasets import load_dataset
from transformers.testing_utils import require_torch, require_vision, slow
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import ImageGPTImageProcessor
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=7 ,__UpperCAmelCase=3 ,__UpperCAmelCase=18 ,__UpperCAmelCase=30 ,__UpperCAmelCase=400 ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,__UpperCAmelCase=True ,) -> str:
lowerCAmelCase__ : Dict = size if size is not None else {"""height""": 18, """width""": 18}
lowerCAmelCase__ : Union[str, Any] = parent
lowerCAmelCase__ : Any = batch_size
lowerCAmelCase__ : Union[str, Any] = num_channels
lowerCAmelCase__ : str = image_size
lowerCAmelCase__ : Tuple = min_resolution
lowerCAmelCase__ : Union[str, Any] = max_resolution
lowerCAmelCase__ : Any = do_resize
lowerCAmelCase__ : Optional[Any] = size
lowerCAmelCase__ : Any = do_normalize
def UpperCAmelCase_ ( self ) -> str:
return {
# here we create 2 clusters for the sake of simplicity
"clusters": np.asarray(
[
[0.8_8_6_6_4_4_3_6_3_4_0_3_3_2_0_3, 0.6_6_1_8_8_2_9_3_6_9_5_4_4_9_8_3, 0.3_8_9_1_7_4_6_4_0_1_7_8_6_8_0_4],
[-0.6_0_4_2_5_5_9_1_4_6_8_8_1_1_0_4, -0.0_2_2_9_5_0_0_8_8_6_0_5_2_8_4_6_9, 0.5_4_2_3_7_9_7_3_6_9_0_0_3_2_9_6],
] ),
"do_resize": self.do_resize,
"size": self.size,
"do_normalize": self.do_normalize,
}
@require_torch
@require_vision
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Optional[int] = ImageGPTImageProcessor if is_vision_available() else None
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Dict = ImageGPTImageProcessingTester(self )
@property
def UpperCAmelCase_ ( self ) -> Any:
return self.image_processor_tester.prepare_image_processor_dict()
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : int = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(__UpperCAmelCase ,"""clusters""" ) )
self.assertTrue(hasattr(__UpperCAmelCase ,"""do_resize""" ) )
self.assertTrue(hasattr(__UpperCAmelCase ,"""size""" ) )
self.assertTrue(hasattr(__UpperCAmelCase ,"""do_normalize""" ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = self.image_processing_class.from_dict(self.image_processor_dict )
self.assertEqual(image_processor.size ,{"""height""": 18, """width""": 18} )
lowerCAmelCase__ : List[str] = self.image_processing_class.from_dict(self.image_processor_dict ,size=42 )
self.assertEqual(image_processor.size ,{"""height""": 42, """width""": 42} )
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : List[str] = self.image_processing_class(**self.image_processor_dict )
lowerCAmelCase__ : List[Any] = json.loads(image_processor.to_json_string() )
for key, value in self.image_processor_dict.items():
if key == "clusters":
self.assertTrue(np.array_equal(__UpperCAmelCase ,obj[key] ) )
else:
self.assertEqual(obj[key] ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = self.image_processing_class(**self.image_processor_dict )
with tempfile.TemporaryDirectory() as tmpdirname:
lowerCAmelCase__ : Union[str, Any] = os.path.join(__UpperCAmelCase ,"""image_processor.json""" )
image_processor_first.to_json_file(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = self.image_processing_class.from_json_file(__UpperCAmelCase ).to_dict()
lowerCAmelCase__ : Dict = image_processor_first.to_dict()
for key, value in image_processor_first.items():
if key == "clusters":
self.assertTrue(np.array_equal(__UpperCAmelCase ,image_processor_second[key] ) )
else:
self.assertEqual(image_processor_first[key] ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : int = self.image_processing_class(**self.image_processor_dict )
with tempfile.TemporaryDirectory() as tmpdirname:
image_processor_first.save_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = self.image_processing_class.from_pretrained(__UpperCAmelCase ).to_dict()
lowerCAmelCase__ : Tuple = image_processor_first.to_dict()
for key, value in image_processor_first.items():
if key == "clusters":
self.assertTrue(np.array_equal(__UpperCAmelCase ,image_processor_second[key] ) )
else:
self.assertEqual(image_processor_first[key] ,__UpperCAmelCase )
@unittest.skip("""ImageGPT requires clusters at initialization""" )
def UpperCAmelCase_ ( self ) -> Any:
pass
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : str = load_dataset("""hf-internal-testing/fixtures_image_utils""" , split="""test""" )
lowerCAmelCase__ : int = Image.open(dataset[4]["""file"""] )
lowerCAmelCase__ : List[str] = Image.open(dataset[5]["""file"""] )
lowerCAmelCase__ : List[Any] = [imagea, imagea]
return images
@require_vision
@require_torch
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Tuple = ImageGPTImageProcessor.from_pretrained("""openai/imagegpt-small""" )
lowerCAmelCase__ : Union[str, Any] = prepare_images()
# test non-batched
lowerCAmelCase__ : Any = image_processing(images[0] ,return_tensors="""pt""" )
self.assertIsInstance(encoding.input_ids ,torch.LongTensor )
self.assertEqual(encoding.input_ids.shape ,(1, 1024) )
lowerCAmelCase__ : Union[str, Any] = [306, 191, 191]
self.assertEqual(encoding.input_ids[0, :3].tolist() ,__UpperCAmelCase )
# test batched
lowerCAmelCase__ : Optional[Any] = image_processing(__UpperCAmelCase ,return_tensors="""pt""" )
self.assertIsInstance(encoding.input_ids ,torch.LongTensor )
self.assertEqual(encoding.input_ids.shape ,(2, 1024) )
lowerCAmelCase__ : Union[str, Any] = [303, 13, 13]
self.assertEqual(encoding.input_ids[1, -3:].tolist() ,__UpperCAmelCase )
| 37
|
'''simple docstring'''
import argparse
import os.path as osp
import re
import torch
from safetensors.torch import load_file, save_file
# =================#
# UNet Conversion #
# =================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''time_embed.0.weight''', '''time_embedding.linear_1.weight'''),
('''time_embed.0.bias''', '''time_embedding.linear_1.bias'''),
('''time_embed.2.weight''', '''time_embedding.linear_2.weight'''),
('''time_embed.2.bias''', '''time_embedding.linear_2.bias'''),
('''input_blocks.0.0.weight''', '''conv_in.weight'''),
('''input_blocks.0.0.bias''', '''conv_in.bias'''),
('''out.0.weight''', '''conv_norm_out.weight'''),
('''out.0.bias''', '''conv_norm_out.bias'''),
('''out.2.weight''', '''conv_out.weight'''),
('''out.2.bias''', '''conv_out.bias'''),
]
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''in_layers.0''', '''norm1'''),
('''in_layers.2''', '''conv1'''),
('''out_layers.0''', '''norm2'''),
('''out_layers.3''', '''conv2'''),
('''emb_layers.1''', '''time_emb_proj'''),
('''skip_connection''', '''conv_shortcut'''),
]
_lowerCAmelCase = []
# hardcoded number of downblocks and resnets/attentions...
# would need smarter logic for other networks.
for i in range(4):
# loop over downblocks/upblocks
for j in range(2):
# loop over resnets/attentions for downblocks
_lowerCAmelCase = F"""down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.0."""
unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
if i < 3:
# no attention layers in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.1."""
unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
for j in range(3):
# loop over resnets/attentions for upblocks
_lowerCAmelCase = F"""up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.0."""
unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
if i > 0:
# no attention layers in up_blocks.0
_lowerCAmelCase = F"""up_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.1."""
unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
if i < 3:
# no downsample in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0.conv."""
_lowerCAmelCase = F"""input_blocks.{3*(i+1)}.0.op."""
unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
# no upsample in up_blocks.3
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""output_blocks.{3*i + 2}.{1 if i == 0 else 2}."""
unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
_lowerCAmelCase = '''mid_block.attentions.0.'''
_lowerCAmelCase = '''middle_block.1.'''
unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
for j in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{j}."""
_lowerCAmelCase = F"""middle_block.{2*j}."""
unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = {k: k for k in unet_state_dict.keys()}
for sd_name, hf_name in unet_conversion_map:
lowerCAmelCase__ : Optional[int] = sd_name
for k, v in mapping.items():
if "resnets" in k:
for sd_part, hf_part in unet_conversion_map_resnet:
lowerCAmelCase__ : Any = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
for k, v in mapping.items():
for sd_part, hf_part in unet_conversion_map_layer:
lowerCAmelCase__ : List[Any] = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
lowerCAmelCase__ : Tuple = {v: unet_state_dict[k] for k, v in mapping.items()}
return new_state_dict
# ================#
# VAE Conversion #
# ================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''nin_shortcut''', '''conv_shortcut'''),
('''norm_out''', '''conv_norm_out'''),
('''mid.attn_1.''', '''mid_block.attentions.0.'''),
]
for i in range(4):
# down_blocks have two resnets
for j in range(2):
_lowerCAmelCase = F"""encoder.down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""encoder.down.{i}.block.{j}."""
vae_conversion_map.append((sd_down_prefix, hf_down_prefix))
if i < 3:
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0."""
_lowerCAmelCase = F"""down.{i}.downsample."""
vae_conversion_map.append((sd_downsample_prefix, hf_downsample_prefix))
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""up.{3-i}.upsample."""
vae_conversion_map.append((sd_upsample_prefix, hf_upsample_prefix))
# up_blocks have three resnets
# also, up blocks in hf are numbered in reverse from sd
for j in range(3):
_lowerCAmelCase = F"""decoder.up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""decoder.up.{3-i}.block.{j}."""
vae_conversion_map.append((sd_up_prefix, hf_up_prefix))
# this part accounts for mid blocks in both the encoder and the decoder
for i in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{i}."""
_lowerCAmelCase = F"""mid.block_{i+1}."""
vae_conversion_map.append((sd_mid_res_prefix, hf_mid_res_prefix))
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''norm.''', '''group_norm.'''),
('''q.''', '''query.'''),
('''k.''', '''key.'''),
('''v.''', '''value.'''),
('''proj_out.''', '''proj_attn.'''),
]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return w.reshape(*w.shape , 1 , 1 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = {k: k for k in vae_state_dict.keys()}
for k, v in mapping.items():
for sd_part, hf_part in vae_conversion_map:
lowerCAmelCase__ : str = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
for k, v in mapping.items():
if "attentions" in k:
for sd_part, hf_part in vae_conversion_map_attn:
lowerCAmelCase__ : Dict = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
lowerCAmelCase__ : Union[str, Any] = {v: vae_state_dict[k] for k, v in mapping.items()}
lowerCAmelCase__ : Tuple = ["""q""", """k""", """v""", """proj_out"""]
for k, v in new_state_dict.items():
for weight_name in weights_to_convert:
if f"""mid.attn_1.{weight_name}.weight""" in k:
print(f"""Reshaping {k} for SD format""" )
lowerCAmelCase__ : Optional[int] = reshape_weight_for_sd(UpperCamelCase )
return new_state_dict
# =========================#
# Text Encoder Conversion #
# =========================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''resblocks.''', '''text_model.encoder.layers.'''),
('''ln_1''', '''layer_norm1'''),
('''ln_2''', '''layer_norm2'''),
('''.c_fc.''', '''.fc1.'''),
('''.c_proj.''', '''.fc2.'''),
('''.attn''', '''.self_attn'''),
('''ln_final.''', '''transformer.text_model.final_layer_norm.'''),
('''token_embedding.weight''', '''transformer.text_model.embeddings.token_embedding.weight'''),
('''positional_embedding''', '''transformer.text_model.embeddings.position_embedding.weight'''),
]
_lowerCAmelCase = {re.escape(x[1]): x[0] for x in textenc_conversion_lst}
_lowerCAmelCase = re.compile('''|'''.join(protected.keys()))
# Ordering is from https://github.com/pytorch/pytorch/blob/master/test/cpp/api/modules.cpp
_lowerCAmelCase = {'''q''': 0, '''k''': 1, '''v''': 2}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = {}
lowerCAmelCase__ : int = {}
lowerCAmelCase__ : List[Any] = {}
for k, v in text_enc_dict.items():
if (
k.endswith(""".self_attn.q_proj.weight""" )
or k.endswith(""".self_attn.k_proj.weight""" )
or k.endswith(""".self_attn.v_proj.weight""" )
):
lowerCAmelCase__ : Optional[int] = k[: -len(""".q_proj.weight""" )]
lowerCAmelCase__ : Tuple = k[-len("""q_proj.weight""" )]
if k_pre not in capture_qkv_weight:
lowerCAmelCase__ : List[Any] = [None, None, None]
lowerCAmelCase__ : Dict = v
continue
if (
k.endswith(""".self_attn.q_proj.bias""" )
or k.endswith(""".self_attn.k_proj.bias""" )
or k.endswith(""".self_attn.v_proj.bias""" )
):
lowerCAmelCase__ : str = k[: -len(""".q_proj.bias""" )]
lowerCAmelCase__ : List[str] = k[-len("""q_proj.bias""" )]
if k_pre not in capture_qkv_bias:
lowerCAmelCase__ : Union[str, Any] = [None, None, None]
lowerCAmelCase__ : Any = v
continue
lowerCAmelCase__ : Dict = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = v
for k_pre, tensors in capture_qkv_weight.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : Any = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Tuple = torch.cat(UpperCamelCase )
for k_pre, tensors in capture_qkv_bias.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : str = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : List[Any] = torch.cat(UpperCamelCase )
return new_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return text_enc_dict
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument('''--model_path''', default=None, type=str, required=True, help='''Path to the model to convert.''')
parser.add_argument('''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the output model.''')
parser.add_argument('''--half''', action='''store_true''', help='''Save weights in half precision.''')
parser.add_argument(
'''--use_safetensors''', action='''store_true''', help='''Save weights use safetensors, default is ckpt.'''
)
_lowerCAmelCase = parser.parse_args()
assert args.model_path is not None, "Must provide a model path!"
assert args.checkpoint_path is not None, "Must provide a checkpoint path!"
# Path for safetensors
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''model.safetensors''')
# Load models from safetensors if it exists, if it doesn't pytorch
if osp.exists(unet_path):
_lowerCAmelCase = load_file(unet_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(unet_path, map_location='''cpu''')
if osp.exists(vae_path):
_lowerCAmelCase = load_file(vae_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(vae_path, map_location='''cpu''')
if osp.exists(text_enc_path):
_lowerCAmelCase = load_file(text_enc_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''pytorch_model.bin''')
_lowerCAmelCase = torch.load(text_enc_path, map_location='''cpu''')
# Convert the UNet model
_lowerCAmelCase = convert_unet_state_dict(unet_state_dict)
_lowerCAmelCase = {'''model.diffusion_model.''' + k: v for k, v in unet_state_dict.items()}
# Convert the VAE model
_lowerCAmelCase = convert_vae_state_dict(vae_state_dict)
_lowerCAmelCase = {'''first_stage_model.''' + k: v for k, v in vae_state_dict.items()}
# Easiest way to identify v2.0 model seems to be that the text encoder (OpenCLIP) is deeper
_lowerCAmelCase = '''text_model.encoder.layers.22.layer_norm2.bias''' in text_enc_dict
if is_vaa_model:
# Need to add the tag 'transformer' in advance so we can knock it out from the final layer-norm
_lowerCAmelCase = {'''transformer.''' + k: v for k, v in text_enc_dict.items()}
_lowerCAmelCase = convert_text_enc_state_dict_vaa(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.model.''' + k: v for k, v in text_enc_dict.items()}
else:
_lowerCAmelCase = convert_text_enc_state_dict(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.transformer.''' + k: v for k, v in text_enc_dict.items()}
# Put together new checkpoint
_lowerCAmelCase = {**unet_state_dict, **vae_state_dict, **text_enc_dict}
if args.half:
_lowerCAmelCase = {k: v.half() for k, v in state_dict.items()}
if args.use_safetensors:
save_file(state_dict, args.checkpoint_path)
else:
_lowerCAmelCase = {'''state_dict''': state_dict}
torch.save(state_dict, args.checkpoint_path)
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return round(float(moles / volume ) * nfactor )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return round(float((moles * 0.0821 * temperature) / (volume) ) )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return round(float((moles * 0.0821 * temperature) / (pressure) ) )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return round(float((pressure * volume) / (0.0821 * moles) ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
import os
from bleurt import score # From: git+https://github.com/google-research/bleurt.git
import datasets
_lowerCAmelCase = datasets.logging.get_logger(__name__)
_lowerCAmelCase = '''\
@inproceedings{bleurt,
title={BLEURT: Learning Robust Metrics for Text Generation},
author={Thibault Sellam and Dipanjan Das and Ankur P. Parikh},
booktitle={ACL},
year={2020},
url={https://arxiv.org/abs/2004.04696}
}
'''
_lowerCAmelCase = '''\
BLEURT a learnt evaluation metric for Natural Language Generation. It is built using multiple phases of transfer learning starting from a pretrained BERT model (Devlin et al. 2018)
and then employing another pre-training phrase using synthetic data. Finally it is trained on WMT human annotations. You may run BLEURT out-of-the-box or fine-tune
it for your specific application (the latter is expected to perform better).
See the project\'s README at https://github.com/google-research/bleurt#readme for more information.
'''
_lowerCAmelCase = '''
BLEURT score.
Args:
`predictions` (list of str): prediction/candidate sentences
`references` (list of str): reference sentences
`checkpoint` BLEURT checkpoint. Will default to BLEURT-tiny if None.
Returns:
\'scores\': List of scores.
Examples:
>>> predictions = ["hello there", "general kenobi"]
>>> references = ["hello there", "general kenobi"]
>>> bleurt = datasets.load_metric("bleurt")
>>> results = bleurt.compute(predictions=predictions, references=references)
>>> print([round(v, 2) for v in results["scores"]])
[1.03, 1.04]
'''
_lowerCAmelCase = {
'''bleurt-tiny-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-128.zip''',
'''bleurt-tiny-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-512.zip''',
'''bleurt-base-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-128.zip''',
'''bleurt-base-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-512.zip''',
'''bleurt-large-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-128.zip''',
'''bleurt-large-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-512.zip''',
'''BLEURT-20-D3''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D3.zip''',
'''BLEURT-20-D6''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D6.zip''',
'''BLEURT-20-D12''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D12.zip''',
'''BLEURT-20''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20.zip''',
}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,homepage="""https://github.com/google-research/bleurt""" ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(
{
"""predictions""": datasets.Value("""string""" ,id="""sequence""" ),
"""references""": datasets.Value("""string""" ,id="""sequence""" ),
} ) ,codebase_urls=["""https://github.com/google-research/bleurt"""] ,reference_urls=["""https://github.com/google-research/bleurt""", """https://arxiv.org/abs/2004.04696"""] ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
# check that config name specifies a valid BLEURT model
if self.config_name == "default":
logger.warning(
"""Using default BLEURT-Base checkpoint for sequence maximum length 128. """
"""You can use a bigger model for better results with e.g.: datasets.load_metric('bleurt', 'bleurt-large-512').""" )
lowerCAmelCase__ : str = """bleurt-base-128"""
if self.config_name.lower() in CHECKPOINT_URLS:
lowerCAmelCase__ : Union[str, Any] = self.config_name.lower()
elif self.config_name.upper() in CHECKPOINT_URLS:
lowerCAmelCase__ : List[Any] = self.config_name.upper()
else:
raise KeyError(
F"""{self.config_name} model not found. You should supply the name of a model checkpoint for bleurt in {CHECKPOINT_URLS.keys()}""" )
# download the model checkpoint specified by self.config_name and set up the scorer
lowerCAmelCase__ : int = dl_manager.download_and_extract(CHECKPOINT_URLS[checkpoint_name] )
lowerCAmelCase__ : Dict = score.BleurtScorer(os.path.join(__UpperCAmelCase ,__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Union[str, Any] = self.scorer.score(references=__UpperCAmelCase ,candidates=__UpperCAmelCase )
return {"scores": scores}
| 37
| 1
|
'''simple docstring'''
from heapq import heappop, heappush
import numpy as np
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = grid.shape
lowerCAmelCase__ : str = [-1, 1, 0, 0]
lowerCAmelCase__ : int = [0, 0, -1, 1]
if allow_diagonal:
dx += [-1, -1, 1, 1]
dy += [-1, 1, -1, 1]
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = [(0, source)], set()
lowerCAmelCase__ : Any = np.full((rows, cols) , np.inf )
lowerCAmelCase__ : Dict = 0
lowerCAmelCase__ : List[Any] = np.empty((rows, cols) , dtype=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = None
while queue:
((lowerCAmelCase__) , (lowerCAmelCase__)) : Union[str, Any] = heappop(UpperCamelCase )
if (x, y) in visited:
continue
visited.add((x, y) )
if (x, y) == destination:
lowerCAmelCase__ : Optional[int] = []
while (x, y) != source:
path.append((x, y) )
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = predecessors[x, y]
path.append(UpperCamelCase ) # add the source manually
path.reverse()
return matrix[destination], path
for i in range(len(UpperCamelCase ) ):
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = x + dx[i], y + dy[i]
if 0 <= nx < rows and 0 <= ny < cols:
lowerCAmelCase__ : Dict = grid[nx][ny]
if next_node == 1 and matrix[nx, ny] > dist + 1:
heappush(UpperCamelCase , (dist + 1, (nx, ny)) )
lowerCAmelCase__ : List[Any] = dist + 1
lowerCAmelCase__ : Optional[int] = (x, y)
return np.inf, []
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
import uuid
from typing import Any, Dict, List, Optional, Union
from ..utils import add_end_docstrings, is_tf_available, is_torch_available, logging
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_tf_available():
import tensorflow as tf
if is_torch_available():
import torch
_lowerCAmelCase = logging.get_logger(__name__)
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ) -> str:
if not conversation_id:
lowerCAmelCase__ : List[str] = uuid.uuida()
if past_user_inputs is None:
lowerCAmelCase__ : List[Any] = []
if generated_responses is None:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : uuid.UUID = conversation_id
lowerCAmelCase__ : List[str] = past_user_inputs
lowerCAmelCase__ : List[str] = generated_responses
lowerCAmelCase__ : Optional[str] = text
def __eq__( self ,__UpperCAmelCase ) -> Dict:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
return False
if self.uuid == other.uuid:
return True
return (
self.new_user_input == other.new_user_input
and self.past_user_inputs == other.past_user_inputs
and self.generated_responses == other.generated_responses
)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Optional[Any]:
if self.new_user_input:
if overwrite:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" was overwritten """
F"""with: \"{text}\".""" )
lowerCAmelCase__ : Optional[int] = text
else:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" new input """
F"""ignored: \"{text}\". Set `overwrite` to True to overwrite unprocessed user input""" )
else:
lowerCAmelCase__ : Optional[Any] = text
def UpperCAmelCase_ ( self ) -> List[Any]:
if self.new_user_input:
self.past_user_inputs.append(self.new_user_input )
lowerCAmelCase__ : Union[str, Any] = None
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
self.generated_responses.append(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
for user_input, generated_response in zip(self.past_user_inputs ,self.generated_responses ):
yield True, user_input
yield False, generated_response
if self.new_user_input:
yield True, self.new_user_input
def __repr__( self ) -> Tuple:
lowerCAmelCase__ : Tuple = F"""Conversation id: {self.uuid} \n"""
for is_user, text in self.iter_texts():
lowerCAmelCase__ : Any = """user""" if is_user else """bot"""
output += F"""{name} >> {text} \n"""
return output
@add_end_docstrings(
SCREAMING_SNAKE_CASE_ , R'''
min_length_for_response (`int`, *optional*, defaults to 32):
The minimum length (in number of tokens) for a response.
minimum_tokens (`int`, *optional*, defaults to 10):
The minimum length of tokens to leave for a response.
''' , )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Tuple:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
if self.tokenizer.pad_token_id is None:
lowerCAmelCase__ : Tuple = self.tokenizer.eos_token
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : List[Any] = {}
lowerCAmelCase__ : Optional[int] = {}
lowerCAmelCase__ : List[str] = {}
if min_length_for_response is not None:
lowerCAmelCase__ : Any = min_length_for_response
if minimum_tokens is not None:
lowerCAmelCase__ : Optional[int] = minimum_tokens
if "max_length" in generate_kwargs:
lowerCAmelCase__ : Optional[Any] = generate_kwargs["""max_length"""]
# self.max_length = generate_kwargs.get("max_length", self.model.config.max_length)
if clean_up_tokenization_spaces is not None:
lowerCAmelCase__ : int = clean_up_tokenization_spaces
if generate_kwargs:
forward_params.update(__UpperCAmelCase )
return preprocess_params, forward_params, postprocess_params
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase=0 ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Optional[int] = super().__call__(__UpperCAmelCase ,num_workers=__UpperCAmelCase ,**__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) and len(__UpperCAmelCase ) == 1:
return outputs[0]
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=32 ) -> Dict[str, Any]:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
raise ValueError("""ConversationalPipeline, expects Conversation as inputs""" )
if conversation.new_user_input is None:
raise ValueError(
F"""Conversation with UUID {type(conversation.uuid )} does not contain new user input to process. """
"""Add user inputs with the conversation's `add_user_input` method""" )
if hasattr(self.tokenizer ,"""_build_conversation_input_ids""" ):
lowerCAmelCase__ : str = self.tokenizer._build_conversation_input_ids(__UpperCAmelCase )
else:
# If the tokenizer cannot handle conversations, we default to only the old version
lowerCAmelCase__ : List[Any] = self._legacy_parse_and_tokenize(__UpperCAmelCase )
if self.framework == "pt":
lowerCAmelCase__ : List[Any] = torch.LongTensor([input_ids] )
elif self.framework == "tf":
lowerCAmelCase__ : Dict = tf.constant([input_ids] )
return {"input_ids": input_ids, "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=10 ,**__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[Any] = generate_kwargs.get("""max_length""" ,self.model.config.max_length )
lowerCAmelCase__ : Optional[Any] = model_inputs["""input_ids"""].shape[1]
if max_length - minimum_tokens < n:
logger.warning(F"""Conversation input is to long ({n}), trimming it to ({max_length} - {minimum_tokens})""" )
lowerCAmelCase__ : str = max_length - minimum_tokens
lowerCAmelCase__ : Union[str, Any] = model_inputs["""input_ids"""][:, -trim:]
if "attention_mask" in model_inputs:
lowerCAmelCase__ : Tuple = model_inputs["""attention_mask"""][:, -trim:]
lowerCAmelCase__ : str = model_inputs.pop("""conversation""" )
lowerCAmelCase__ : Union[str, Any] = max_length
lowerCAmelCase__ : Any = self.model.generate(**__UpperCAmelCase ,**__UpperCAmelCase )
if self.model.config.is_encoder_decoder:
lowerCAmelCase__ : int = 1
else:
lowerCAmelCase__ : int = n
return {"output_ids": output_ids[:, start_position:], "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=True ) -> List[str]:
lowerCAmelCase__ : Optional[int] = model_outputs["""output_ids"""]
lowerCAmelCase__ : Tuple = self.tokenizer.decode(
output_ids[0] ,skip_special_tokens=__UpperCAmelCase ,clean_up_tokenization_spaces=__UpperCAmelCase ,)
lowerCAmelCase__ : Union[str, Any] = model_outputs["""conversation"""]
conversation.mark_processed()
conversation.append_response(__UpperCAmelCase )
return conversation
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Dict = self.tokenizer.eos_token_id
lowerCAmelCase__ : int = []
for is_user, text in conversation.iter_texts():
if eos_token_id is not None:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) + [eos_token_id] )
else:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) )
if len(__UpperCAmelCase ) > self.tokenizer.model_max_length:
lowerCAmelCase__ : Optional[Any] = input_ids[-self.tokenizer.model_max_length :]
return input_ids
| 37
| 1
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''junnyu/roformer_chinese_small''': '''https://huggingface.co/junnyu/roformer_chinese_small/resolve/main/config.json''',
'''junnyu/roformer_chinese_base''': '''https://huggingface.co/junnyu/roformer_chinese_base/resolve/main/config.json''',
'''junnyu/roformer_chinese_char_small''': (
'''https://huggingface.co/junnyu/roformer_chinese_char_small/resolve/main/config.json'''
),
'''junnyu/roformer_chinese_char_base''': (
'''https://huggingface.co/junnyu/roformer_chinese_char_base/resolve/main/config.json'''
),
'''junnyu/roformer_small_discriminator''': (
'''https://huggingface.co/junnyu/roformer_small_discriminator/resolve/main/config.json'''
),
'''junnyu/roformer_small_generator''': (
'''https://huggingface.co/junnyu/roformer_small_generator/resolve/main/config.json'''
),
# See all RoFormer models at https://huggingface.co/models?filter=roformer
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : List[str] = '''roformer'''
def __init__( self ,__UpperCAmelCase=5_0000 ,__UpperCAmelCase=None ,__UpperCAmelCase=768 ,__UpperCAmelCase=12 ,__UpperCAmelCase=12 ,__UpperCAmelCase=3072 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=1536 ,__UpperCAmelCase=2 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-12 ,__UpperCAmelCase=0 ,__UpperCAmelCase=False ,__UpperCAmelCase=True ,**__UpperCAmelCase ,) -> List[Any]:
super().__init__(pad_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Tuple = vocab_size
lowerCAmelCase__ : int = hidden_size if embedding_size is None else embedding_size
lowerCAmelCase__ : Optional[int] = hidden_size
lowerCAmelCase__ : List[str] = num_hidden_layers
lowerCAmelCase__ : Optional[int] = num_attention_heads
lowerCAmelCase__ : str = hidden_act
lowerCAmelCase__ : str = intermediate_size
lowerCAmelCase__ : Tuple = hidden_dropout_prob
lowerCAmelCase__ : int = attention_probs_dropout_prob
lowerCAmelCase__ : str = max_position_embeddings
lowerCAmelCase__ : Any = type_vocab_size
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : Optional[Any] = layer_norm_eps
lowerCAmelCase__ : int = rotary_value
lowerCAmelCase__ : List[str] = use_cache
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Optional[Any] = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : List[str] = {0: """batch""", 1: """sequence"""}
lowerCAmelCase__ : Tuple = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
("""token_type_ids""", dynamic_axis),
] )
| 37
|
'''simple docstring'''
from collections import UserDict
from typing import List, Union
from ..utils import (
add_end_docstrings,
is_tf_available,
is_torch_available,
is_vision_available,
logging,
requires_backends,
)
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
from ..tf_utils import stable_softmax
_lowerCAmelCase = logging.get_logger(__name__)
@add_end_docstrings(SCREAMING_SNAKE_CASE_ )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,**__UpperCAmelCase ) -> Tuple:
super().__init__(**__UpperCAmelCase )
requires_backends(self ,"""vision""" )
self.check_model_type(
TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if self.framework == """tf"""
else MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING )
def __call__( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> str:
return super().__call__(__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[Any] = {}
if "candidate_labels" in kwargs:
lowerCAmelCase__ : int = kwargs["""candidate_labels"""]
if "hypothesis_template" in kwargs:
lowerCAmelCase__ : Optional[int] = kwargs["""hypothesis_template"""]
return preprocess_params, {}, {}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase="This is a photo of {}." ) -> int:
lowerCAmelCase__ : str = load_image(__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.image_processor(images=[image] ,return_tensors=self.framework )
lowerCAmelCase__ : List[Any] = candidate_labels
lowerCAmelCase__ : List[str] = [hypothesis_template.format(__UpperCAmelCase ) for x in candidate_labels]
lowerCAmelCase__ : Optional[Any] = self.tokenizer(__UpperCAmelCase ,return_tensors=self.framework ,padding=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = [text_inputs]
return inputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = model_inputs.pop("""candidate_labels""" )
lowerCAmelCase__ : Union[str, Any] = model_inputs.pop("""text_inputs""" )
if isinstance(text_inputs[0] ,__UpperCAmelCase ):
lowerCAmelCase__ : int = text_inputs[0]
else:
# Batching case.
lowerCAmelCase__ : Dict = text_inputs[0][0]
lowerCAmelCase__ : Any = self.model(**__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = {
"""candidate_labels""": candidate_labels,
"""logits""": outputs.logits_per_image,
}
return model_outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Union[str, Any] = model_outputs.pop("""candidate_labels""" )
lowerCAmelCase__ : List[str] = model_outputs["""logits"""][0]
if self.framework == "pt":
lowerCAmelCase__ : List[str] = logits.softmax(dim=-1 ).squeeze(-1 )
lowerCAmelCase__ : Optional[Any] = probs.tolist()
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Dict = [scores]
elif self.framework == "tf":
lowerCAmelCase__ : Any = stable_softmax(__UpperCAmelCase ,axis=-1 )
lowerCAmelCase__ : List[Any] = probs.numpy().tolist()
else:
raise ValueError(F"""Unsupported framework: {self.framework}""" )
lowerCAmelCase__ : Tuple = [
{"""score""": score, """label""": candidate_label}
for score, candidate_label in sorted(zip(__UpperCAmelCase ,__UpperCAmelCase ) ,key=lambda __UpperCAmelCase : -x[0] )
]
return result
| 37
| 1
|
'''simple docstring'''
import os
from math import logaa
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "base_exp.txt" ):
"""simple docstring"""
lowerCAmelCase__ : float = 0
lowerCAmelCase__ : int = 0
for i, line in enumerate(open(os.path.join(os.path.dirname(UpperCamelCase ) , UpperCamelCase ) ) ):
lowerCAmelCase__ , lowerCAmelCase__ : Any = list(map(UpperCamelCase , line.split(""",""" ) ) )
if x * logaa(UpperCamelCase ) > largest:
lowerCAmelCase__ : List[str] = x * logaa(UpperCamelCase )
lowerCAmelCase__ : List[str] = i + 1
return result
if __name__ == "__main__":
print(solution())
| 37
|
'''simple docstring'''
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , )
@pytest.mark.usefixtures('''sm_env''' )
@parameterized_class(
[
{
'''framework''': '''pytorch''',
'''script''': '''run_glue.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_5_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''pytorch''',
'''script''': '''run_ddp.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''tensorflow''',
'''script''': '''run_tf_dist.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.6, '''eval_loss''': 0.7},
},
] )
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.framework == "pytorch":
subprocess.run(
F"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() ,encoding="""utf-8""" ,check=__UpperCAmelCase ,)
assert hasattr(self ,"""env""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = F"""{self.env.base_job_name}-{instance_count}-{'ddp' if 'ddp' in self.script else 'smd'}"""
# distributed data settings
lowerCAmelCase__ : Any = {"""smdistributed""": {"""dataparallel""": {"""enabled""": True}}} if self.script != """run_ddp.py""" else None
# creates estimator
return HuggingFace(
entry_point=self.script ,source_dir=self.env.test_path ,role=self.env.role ,image_uri=self.env.image_uri ,base_job_name=__UpperCAmelCase ,instance_count=__UpperCAmelCase ,instance_type=self.instance_type ,debugger_hook_config=__UpperCAmelCase ,hyperparameters={**self.env.distributed_hyperparameters, """model_name_or_path""": self.model_name_or_path} ,metric_definitions=self.env.metric_definitions ,distribution=__UpperCAmelCase ,py_version="""py36""" ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
TrainingJobAnalytics(__UpperCAmelCase ).export_csv(F"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(2,)] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
# create estimator
lowerCAmelCase__ : List[Any] = self.create_estimator(__UpperCAmelCase )
# run training
estimator.fit()
# result dataframe
lowerCAmelCase__ : Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
lowerCAmelCase__ : int = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
lowerCAmelCase__ : List[Any] = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
lowerCAmelCase__ : List[str] = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" ,99_9999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(F"""{estimator.latest_training_job.name}.json""" ,"""w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} ,__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
from __future__ import annotations
from collections import namedtuple
from dataclasses import dataclass
@dataclass
class lowerCAmelCase_:
'''simple docstring'''
__lowercase : int
__lowercase : TreeNode | None = None
__lowercase : TreeNode | None = None
_lowerCAmelCase = namedtuple('''CoinsDistribResult''', '''moves excess''')
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if root is None:
return 0
# Validation
def count_nodes(UpperCamelCase ) -> int:
if node is None:
return 0
return count_nodes(node.left ) + count_nodes(node.right ) + 1
def count_coins(UpperCamelCase ) -> int:
if node is None:
return 0
return count_coins(node.left ) + count_coins(node.right ) + node.data
if count_nodes(UpperCamelCase ) != count_coins(UpperCamelCase ):
raise ValueError("""The nodes number should be same as the number of coins""" )
# Main calculation
def get_distrib(UpperCamelCase ) -> CoinsDistribResult:
if node is None:
return CoinsDistribResult(0 , 1 )
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = get_distrib(node.left )
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = get_distrib(node.right )
lowerCAmelCase__ : Optional[int] = 1 - left_distrib_excess
lowerCAmelCase__ : Dict = 1 - right_distrib_excess
lowerCAmelCase__ : Union[str, Any] = (
left_distrib_moves
+ right_distrib_moves
+ abs(UpperCamelCase )
+ abs(UpperCamelCase )
)
lowerCAmelCase__ : Union[str, Any] = node.data - coins_to_left - coins_to_right
return CoinsDistribResult(UpperCamelCase , UpperCamelCase )
return get_distrib(UpperCamelCase )[0]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''albert-base-v1''': '''https://huggingface.co/albert-base-v1/resolve/main/spiece.model''',
'''albert-large-v1''': '''https://huggingface.co/albert-large-v1/resolve/main/spiece.model''',
'''albert-xlarge-v1''': '''https://huggingface.co/albert-xlarge-v1/resolve/main/spiece.model''',
'''albert-xxlarge-v1''': '''https://huggingface.co/albert-xxlarge-v1/resolve/main/spiece.model''',
'''albert-base-v2''': '''https://huggingface.co/albert-base-v2/resolve/main/spiece.model''',
'''albert-large-v2''': '''https://huggingface.co/albert-large-v2/resolve/main/spiece.model''',
'''albert-xlarge-v2''': '''https://huggingface.co/albert-xlarge-v2/resolve/main/spiece.model''',
'''albert-xxlarge-v2''': '''https://huggingface.co/albert-xxlarge-v2/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''albert-base-v1''': 512,
'''albert-large-v1''': 512,
'''albert-xlarge-v1''': 512,
'''albert-xxlarge-v1''': 512,
'''albert-base-v2''': 512,
'''albert-large-v2''': 512,
'''albert-xlarge-v2''': 512,
'''albert-xxlarge-v2''': 512,
}
_lowerCAmelCase = '''▁'''
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Any = VOCAB_FILES_NAMES
__lowercase : Optional[Any] = PRETRAINED_VOCAB_FILES_MAP
__lowercase : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=True ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[MASK]" ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
# Mask token behave like a normal word, i.e. include the space before it and
# is included in the raw text, there should be a match in a non-normalized sentence.
lowerCAmelCase__ : Tuple = (
AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ,normalized=__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase )
else mask_token
)
lowerCAmelCase__ : Any = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : str = do_lower_case
lowerCAmelCase__ : int = remove_space
lowerCAmelCase__ : Tuple = keep_accents
lowerCAmelCase__ : Any = vocab_file
lowerCAmelCase__ : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> Optional[int]:
return len(self.sp_model )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Union[str, Any] = {self.convert_ids_to_tokens(__UpperCAmelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = self.__dict__.copy()
lowerCAmelCase__ : Optional[Any] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Union[str, Any] = {}
lowerCAmelCase__ : Any = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[int]:
if self.remove_space:
lowerCAmelCase__ : int = """ """.join(inputs.strip().split() )
else:
lowerCAmelCase__ : str = inputs
lowerCAmelCase__ : Tuple = outputs.replace("""``""" ,"""\"""" ).replace("""''""" ,"""\"""" )
if not self.keep_accents:
lowerCAmelCase__ : Any = unicodedata.normalize("""NFKD""" ,__UpperCAmelCase )
lowerCAmelCase__ : Dict = """""".join([c for c in outputs if not unicodedata.combining(__UpperCAmelCase )] )
if self.do_lower_case:
lowerCAmelCase__ : Tuple = outputs.lower()
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
lowerCAmelCase__ : str = []
for piece in pieces:
if len(__UpperCAmelCase ) > 1 and piece[-1] == str(""",""" ) and piece[-2].isdigit():
lowerCAmelCase__ : List[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(__UpperCAmelCase ,"""""" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
lowerCAmelCase__ : str = cur_pieces[1:]
else:
lowerCAmelCase__ : int = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(__UpperCAmelCase )
else:
new_pieces.append(__UpperCAmelCase )
return new_pieces
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
return self.sp_model.IdToPiece(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : Tuple = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : List[str] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string.strip()
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : int = [self.sep_token_id]
lowerCAmelCase__ : Dict = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
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 not None:
return [1] + ([0] * len(__UpperCAmelCase )) + [1] + ([0] * len(__UpperCAmelCase )) + [1]
return [1] + ([0] * len(__UpperCAmelCase )) + [1]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : List[str] = [self.sep_token_id]
lowerCAmelCase__ : Tuple = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : int = 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:
lowerCAmelCase__ : List[Any] = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
| 37
| 1
|
'''simple docstring'''
from abc import ABC, abstractmethod
from typing import Optional, Union
from .. import Dataset, DatasetDict, Features, IterableDataset, IterableDatasetDict, NamedSplit
from ..utils.typing import NestedDataStructureLike, PathLike
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = False ,__UpperCAmelCase = False ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> Optional[Any]:
lowerCAmelCase__ : str = path_or_paths
lowerCAmelCase__ : Optional[Any] = split if split or isinstance(__UpperCAmelCase ,__UpperCAmelCase ) else """train"""
lowerCAmelCase__ : int = features
lowerCAmelCase__ : Optional[Any] = cache_dir
lowerCAmelCase__ : Any = keep_in_memory
lowerCAmelCase__ : Union[str, Any] = streaming
lowerCAmelCase__ : Optional[int] = num_proc
lowerCAmelCase__ : Tuple = kwargs
@abstractmethod
def UpperCAmelCase_ ( self ) -> Union[Dataset, DatasetDict, IterableDataset, IterableDatasetDict]:
pass
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = False ,__UpperCAmelCase = False ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> Dict:
lowerCAmelCase__ : str = features
lowerCAmelCase__ : str = cache_dir
lowerCAmelCase__ : int = keep_in_memory
lowerCAmelCase__ : int = streaming
lowerCAmelCase__ : int = num_proc
lowerCAmelCase__ : int = kwargs
@abstractmethod
def UpperCAmelCase_ ( self ) -> Union[Dataset, IterableDataset]:
pass
| 37
|
'''simple docstring'''
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''merges_file''': '''merges.txt''',
}
_lowerCAmelCase = {
'''vocab_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json'''},
'''merges_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt'''},
}
_lowerCAmelCase = {
'''ctrl''': 256,
}
_lowerCAmelCase = {
'''Pregnancy''': 16_8629,
'''Christianity''': 7675,
'''Explain''': 10_6423,
'''Fitness''': 6_3440,
'''Saving''': 6_3163,
'''Ask''': 2_7171,
'''Ass''': 9_5985,
'''Joke''': 16_3509,
'''Questions''': 4_5622,
'''Thoughts''': 4_9605,
'''Retail''': 5_2342,
'''Feminism''': 16_4338,
'''Writing''': 1_1992,
'''Atheism''': 19_2263,
'''Netflix''': 4_8616,
'''Computing''': 3_9639,
'''Opinion''': 4_3213,
'''Alone''': 4_4967,
'''Funny''': 5_8917,
'''Gaming''': 4_0358,
'''Human''': 4088,
'''India''': 1331,
'''Joker''': 7_7138,
'''Diet''': 3_6206,
'''Legal''': 1_1859,
'''Norman''': 4939,
'''Tip''': 7_2689,
'''Weight''': 5_2343,
'''Movies''': 4_6273,
'''Running''': 2_3425,
'''Science''': 2090,
'''Horror''': 3_7793,
'''Confession''': 6_0572,
'''Finance''': 1_2250,
'''Politics''': 1_6360,
'''Scary''': 19_1985,
'''Support''': 1_2654,
'''Technologies''': 3_2516,
'''Teenage''': 6_6160,
'''Event''': 3_2769,
'''Learned''': 6_7460,
'''Notion''': 18_2770,
'''Wikipedia''': 3_7583,
'''Books''': 6665,
'''Extract''': 7_6050,
'''Confessions''': 10_2701,
'''Conspiracy''': 7_5932,
'''Links''': 6_3674,
'''Narcissus''': 15_0425,
'''Relationship''': 5_4766,
'''Relationships''': 13_4796,
'''Reviews''': 4_1671,
'''News''': 4256,
'''Translation''': 2_6820,
'''multilingual''': 12_8406,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = set()
lowerCAmelCase__ : str = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCAmelCase__ : List[Any] = char
lowerCAmelCase__ : Optional[Any] = set(UpperCamelCase )
return pairs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Any = CONTROL_CODES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<unk>" ,**__UpperCAmelCase ) -> Optional[int]:
super().__init__(unk_token=__UpperCAmelCase ,**__UpperCAmelCase )
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as vocab_handle:
lowerCAmelCase__ : int = json.load(__UpperCAmelCase )
lowerCAmelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as merges_handle:
lowerCAmelCase__ : Union[str, Any] = merges_handle.read().split("""\n""" )[1:-1]
lowerCAmelCase__ : Optional[Any] = [tuple(merge.split() ) for merge in merges]
lowerCAmelCase__ : int = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return len(self.encoder )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return dict(self.encoder ,**self.added_tokens_encoder )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if token in self.cache:
return self.cache[token]
lowerCAmelCase__ : Optional[Any] = tuple(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tuple(list(word[:-1] ) + [word[-1] + """</w>"""] )
lowerCAmelCase__ : Any = get_pairs(__UpperCAmelCase )
if not pairs:
return token
while True:
lowerCAmelCase__ : List[str] = min(__UpperCAmelCase ,key=lambda __UpperCAmelCase : self.bpe_ranks.get(__UpperCAmelCase ,float("""inf""" ) ) )
if bigram not in self.bpe_ranks:
break
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = bigram
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : Dict = 0
while i < len(__UpperCAmelCase ):
try:
lowerCAmelCase__ : Optional[Any] = word.index(__UpperCAmelCase ,__UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCAmelCase__ : Dict = 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
lowerCAmelCase__ : Tuple = tuple(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = new_word
if len(__UpperCAmelCase ) == 1:
break
else:
lowerCAmelCase__ : Dict = get_pairs(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """@@ """.join(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = word[:-4]
lowerCAmelCase__ : Optional[Any] = word
return word
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : Any = re.findall(R"""\S+\n?""" ,__UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(__UpperCAmelCase ).split(""" """ ) ) )
return split_tokens
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
return self.encoder.get(__UpperCAmelCase ,self.encoder.get(self.unk_token ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = """ """.join(__UpperCAmelCase ).replace("""@@ """ ,"""""" ).strip()
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : List[Any] = os.path.join(
__UpperCAmelCase ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : Optional[int] = 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""" )
lowerCAmelCase__ : int = 0
with open(__UpperCAmelCase ,"""w""" ,encoding="""utf-8""" ) as writer:
writer.write("""#version: 0.2\n""" )
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 {merge_file}: BPE merge indices are not consecutive."""
""" Please check that the tokenizer is not corrupted!""" )
lowerCAmelCase__ : Dict = token_index
writer.write(""" """.join(__UpperCAmelCase ) + """\n""" )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 37
| 1
|
'''simple docstring'''
import unittest
from pathlib import Path
from tempfile import NamedTemporaryFile, TemporaryDirectory
from transformers import BertConfig, BertTokenizerFast, FeatureExtractionPipeline
from transformers.convert_graph_to_onnx import (
convert,
ensure_valid_input,
generate_identified_filename,
infer_shapes,
quantize,
)
from transformers.testing_utils import require_tf, require_tokenizers, require_torch, slow
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[Any]:
return None
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
return None
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
__lowercase : Dict = [
# (model_name, model_kwargs)
('''bert-base-cased''', {}),
('''gpt2''', {'''use_cache''': False}), # We don't support exporting GPT2 past keys anymore
]
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> int:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Any:
from transformers import BertModel
lowerCAmelCase__ : Optional[int] = ["""[UNK]""", """[SEP]""", """[CLS]""", """[PAD]""", """[MASK]""", """some""", """other""", """words"""]
with NamedTemporaryFile(mode="""w+t""" ) as vocab_file:
vocab_file.write("""\n""".join(__UpperCAmelCase ) )
vocab_file.flush()
lowerCAmelCase__ : Dict = BertTokenizerFast(vocab_file.name )
with TemporaryDirectory() as bert_save_dir:
lowerCAmelCase__ : Tuple = BertModel(BertConfig(vocab_size=len(__UpperCAmelCase ) ) )
model.save_pretrained(__UpperCAmelCase )
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,__UpperCAmelCase )
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> List[str]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Dict = self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = quantize(Path(__UpperCAmelCase ) )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> List[Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Any = self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = quantize(__UpperCAmelCase )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[Any]:
try:
# Compute path
with TemporaryDirectory() as tempdir:
lowerCAmelCase__ : Optional[int] = Path(__UpperCAmelCase ).joinpath("""model.onnx""" )
# Remove folder if exists
if path.parent.exists():
path.parent.rmdir()
# Export
convert(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,**__UpperCAmelCase )
return path
except Exception as e:
self.fail(__UpperCAmelCase )
@require_torch
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
from transformers import BertModel
lowerCAmelCase__ : List[Any] = BertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Union[str, Any] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""pt""" )
@require_tf
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Optional[int]:
from transformers import TFBertModel
lowerCAmelCase__ : int = TFBertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Optional[int] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""tf""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = FeatureExtractionPipeline(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = ["""input_ids""", """token_type_ids""", """attention_mask""", """output_0""", """output_1"""]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = infer_shapes(__UpperCAmelCase ,__UpperCAmelCase )
# Assert all variables are present
self.assertEqual(len(__UpperCAmelCase ) ,len(__UpperCAmelCase ) )
self.assertTrue(all(var_name in shapes for var_name in variable_names ) )
self.assertSequenceEqual(variable_names[:3] ,__UpperCAmelCase )
self.assertSequenceEqual(variable_names[3:] ,__UpperCAmelCase )
# Assert inputs are {0: batch, 1: sequence}
for var_name in ["input_ids", "token_type_ids", "attention_mask"]:
self.assertDictEqual(shapes[var_name] ,{0: """batch""", 1: """sequence"""} )
# Assert outputs are {0: batch, 1: sequence} and {0: batch}
self.assertDictEqual(shapes["""output_0"""] ,{0: """batch""", 1: """sequence"""} )
self.assertDictEqual(shapes["""output_1"""] ,{0: """batch"""} )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = ["""input_ids""", """attention_mask""", """token_type_ids"""]
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": [1, 2, 3, 4], """attention_mask""": [0, 0, 0, 0], """token_type_ids""": [1, 1, 1, 1]}
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = ensure_valid_input(FuncContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the same number of args (all are valid)
self.assertEqual(len(__UpperCAmelCase ) ,3 )
# Should have exactly the same input names
self.assertEqual(set(__UpperCAmelCase ) ,set(__UpperCAmelCase ) )
# Parameter should be reordered according to their respective place in the function:
# (input_ids, token_type_ids, attention_mask)
self.assertEqual(__UpperCAmelCase ,(tokens["""input_ids"""], tokens["""token_type_ids"""], tokens["""attention_mask"""]) )
# Generated args are interleaved with another args (for instance parameter "past" in GPT2)
lowerCAmelCase__ , lowerCAmelCase__ : int = ensure_valid_input(FuncNonContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the one arg (all before the one not provided "some_other_args")
self.assertEqual(len(__UpperCAmelCase ) ,1 )
self.assertEqual(len(__UpperCAmelCase ) ,1 )
# Should have only "input_ids"
self.assertEqual(inputs_args[0] ,tokens["""input_ids"""] )
self.assertEqual(ordered_input_names[0] ,"""input_ids""" )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Dict = generate_identified_filename(Path("""/home/something/my_fake_model.onnx""" ) ,"""-test""" )
self.assertEqual("""/home/something/my_fake_model-test.onnx""" ,generated.as_posix() )
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : str = set()
# Replace all the whitespace in our sentence
lowerCAmelCase__ : Tuple = input_str.replace(""" """ , """""" )
for alpha in input_str:
if "a" <= alpha.lower() <= "z":
frequency.add(alpha.lower() )
return len(UpperCamelCase ) == 26
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : Any = [False] * 26
for char in input_str:
if char.islower():
lowerCAmelCase__ : Optional[Any] = True
elif char.isupper():
lowerCAmelCase__ : Any = True
return all(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
return len({char for char in input_str.lower() if char.isalpha()} ) == 26
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from timeit import timeit
lowerCAmelCase__ : Union[str, Any] = """from __main__ import is_pangram, is_pangram_faster, is_pangram_fastest"""
print(timeit("""is_pangram()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_faster()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_fastest()""" , setup=UpperCamelCase ) )
# 5.348480500048026, 2.6477354579837993, 1.8470395830227062
# 5.036091582966037, 2.644472333951853, 1.8869528750656173
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/convnextv2-tiny-1k-224''': '''https://huggingface.co/facebook/convnextv2-tiny-1k-224/resolve/main/config.json''',
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Tuple = '''convnextv2'''
def __init__( self ,__UpperCAmelCase=3 ,__UpperCAmelCase=4 ,__UpperCAmelCase=4 ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-12 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=224 ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Optional[Any]:
super().__init__(**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = num_channels
lowerCAmelCase__ : Union[str, Any] = patch_size
lowerCAmelCase__ : List[Any] = num_stages
lowerCAmelCase__ : Optional[Any] = [96, 192, 384, 768] if hidden_sizes is None else hidden_sizes
lowerCAmelCase__ : str = [3, 3, 9, 3] if depths is None else depths
lowerCAmelCase__ : Union[str, Any] = hidden_act
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : Dict = layer_norm_eps
lowerCAmelCase__ : List[Any] = drop_path_rate
lowerCAmelCase__ : int = image_size
lowerCAmelCase__ : Any = ["""stem"""] + [F"""stage{idx}""" for idx in range(1 ,len(self.depths ) + 1 )]
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = get_aligned_output_features_output_indices(
out_features=__UpperCAmelCase ,out_indices=__UpperCAmelCase ,stage_names=self.stage_names )
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = abs(UpperCamelCase )
lowerCAmelCase__ : List[Any] = 0
while n > 0:
res += n % 10
n //= 10
return res
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = abs(UpperCamelCase )
return n if n < 10 else n % 10 + sum_of_digits(n // 10 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return sum(int(UpperCamelCase ) for c in str(abs(UpperCamelCase ) ) )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(UpperCamelCase , UpperCamelCase ) -> None:
lowerCAmelCase__ : str = f"""{func.__name__}({value})"""
lowerCAmelCase__ : str = timeit(f"""__main__.{call}""" , setup="""import __main__""" )
print(f"""{call:56} = {func(UpperCamelCase )} -- {timing:.4f} seconds""" )
for value in (262144, 1125899906842624, 1267650600228229401496703205376):
for func in (sum_of_digits, sum_of_digits_recursion, sum_of_digits_compact):
benchmark_a_function(UpperCamelCase , UpperCamelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import argparse
from collections import OrderedDict
from pathlib import Path
import requests
import torch
from PIL import Image
from transformers import GLPNConfig, GLPNForDepthEstimation, GLPNImageProcessor
from transformers.utils import logging
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = OrderedDict()
for key, value in state_dict.items():
if key.startswith("""module.encoder""" ):
lowerCAmelCase__ : Optional[Any] = key.replace("""module.encoder""" , """glpn.encoder""" )
if key.startswith("""module.decoder""" ):
lowerCAmelCase__ : int = key.replace("""module.decoder""" , """decoder.stages""" )
if "patch_embed" in key:
# replace for example patch_embed1 by patch_embeddings.0
lowerCAmelCase__ : Dict = key[key.find("""patch_embed""" ) + len("""patch_embed""" )]
lowerCAmelCase__ : Optional[Any] = key.replace(f"""patch_embed{idx}""" , f"""patch_embeddings.{int(UpperCamelCase )-1}""" )
if "norm" in key:
lowerCAmelCase__ : str = key.replace("""norm""" , """layer_norm""" )
if "glpn.encoder.layer_norm" in key:
# replace for example layer_norm1 by layer_norm.0
lowerCAmelCase__ : Any = key[key.find("""glpn.encoder.layer_norm""" ) + len("""glpn.encoder.layer_norm""" )]
lowerCAmelCase__ : Tuple = key.replace(f"""layer_norm{idx}""" , f"""layer_norm.{int(UpperCamelCase )-1}""" )
if "layer_norm1" in key:
lowerCAmelCase__ : Optional[Any] = key.replace("""layer_norm1""" , """layer_norm_1""" )
if "layer_norm2" in key:
lowerCAmelCase__ : str = key.replace("""layer_norm2""" , """layer_norm_2""" )
if "block" in key:
# replace for example block1 by block.0
lowerCAmelCase__ : List[str] = key[key.find("""block""" ) + len("""block""" )]
lowerCAmelCase__ : List[str] = key.replace(f"""block{idx}""" , f"""block.{int(UpperCamelCase )-1}""" )
if "attn.q" in key:
lowerCAmelCase__ : Optional[Any] = key.replace("""attn.q""" , """attention.self.query""" )
if "attn.proj" in key:
lowerCAmelCase__ : Union[str, Any] = key.replace("""attn.proj""" , """attention.output.dense""" )
if "attn" in key:
lowerCAmelCase__ : Optional[Any] = key.replace("""attn""" , """attention.self""" )
if "fc1" in key:
lowerCAmelCase__ : Dict = key.replace("""fc1""" , """dense1""" )
if "fc2" in key:
lowerCAmelCase__ : Any = key.replace("""fc2""" , """dense2""" )
if "linear_pred" in key:
lowerCAmelCase__ : Dict = key.replace("""linear_pred""" , """classifier""" )
if "linear_fuse" in key:
lowerCAmelCase__ : Any = key.replace("""linear_fuse.conv""" , """linear_fuse""" )
lowerCAmelCase__ : Any = key.replace("""linear_fuse.bn""" , """batch_norm""" )
if "linear_c" in key:
# replace for example linear_c4 by linear_c.3
lowerCAmelCase__ : Any = key[key.find("""linear_c""" ) + len("""linear_c""" )]
lowerCAmelCase__ : Optional[Any] = key.replace(f"""linear_c{idx}""" , f"""linear_c.{int(UpperCamelCase )-1}""" )
if "bot_conv" in key:
lowerCAmelCase__ : Optional[int] = key.replace("""bot_conv""" , """0.convolution""" )
if "skip_conv1" in key:
lowerCAmelCase__ : Optional[int] = key.replace("""skip_conv1""" , """1.convolution""" )
if "skip_conv2" in key:
lowerCAmelCase__ : Any = key.replace("""skip_conv2""" , """2.convolution""" )
if "fusion1" in key:
lowerCAmelCase__ : List[str] = key.replace("""fusion1""" , """1.fusion""" )
if "fusion2" in key:
lowerCAmelCase__ : List[str] = key.replace("""fusion2""" , """2.fusion""" )
if "fusion3" in key:
lowerCAmelCase__ : List[Any] = key.replace("""fusion3""" , """3.fusion""" )
if "fusion" in key and "conv" in key:
lowerCAmelCase__ : str = key.replace("""conv""" , """convolutional_layer""" )
if key.startswith("""module.last_layer_depth""" ):
lowerCAmelCase__ : Union[str, Any] = key.replace("""module.last_layer_depth""" , """head.head""" )
lowerCAmelCase__ : List[Any] = value
return new_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for i in range(config.num_encoder_blocks ):
for j in range(config.depths[i] ):
# read in weights + bias of keys and values (which is a single matrix in the original implementation)
lowerCAmelCase__ : Optional[int] = state_dict.pop(f"""glpn.encoder.block.{i}.{j}.attention.self.kv.weight""" )
lowerCAmelCase__ : List[Any] = state_dict.pop(f"""glpn.encoder.block.{i}.{j}.attention.self.kv.bias""" )
# next, add keys and values (in that order) to the state dict
lowerCAmelCase__ : List[Any] = kv_weight[
: config.hidden_sizes[i], :
]
lowerCAmelCase__ : Optional[Any] = kv_bias[: config.hidden_sizes[i]]
lowerCAmelCase__ : List[str] = kv_weight[
config.hidden_sizes[i] :, :
]
lowerCAmelCase__ : Union[str, Any] = kv_bias[config.hidden_sizes[i] :]
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = """http://images.cocodataset.org/val2017/000000039769.jpg"""
lowerCAmelCase__ : str = Image.open(requests.get(UpperCamelCase , stream=UpperCamelCase ).raw )
return image
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=False , UpperCamelCase=None ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = GLPNConfig(hidden_sizes=[64, 128, 320, 512] , decoder_hidden_size=64 , depths=[3, 8, 27, 3] )
# load image processor (only resize + rescale)
lowerCAmelCase__ : str = GLPNImageProcessor()
# prepare image
lowerCAmelCase__ : Dict = prepare_img()
lowerCAmelCase__ : Optional[Any] = image_processor(images=UpperCamelCase , return_tensors="""pt""" ).pixel_values
logger.info("""Converting model...""" )
# load original state dict
lowerCAmelCase__ : List[Any] = torch.load(UpperCamelCase , map_location=torch.device("""cpu""" ) )
# rename keys
lowerCAmelCase__ : Any = rename_keys(UpperCamelCase )
# key and value matrices need special treatment
read_in_k_v(UpperCamelCase , UpperCamelCase )
# create HuggingFace model and load state dict
lowerCAmelCase__ : List[Any] = GLPNForDepthEstimation(UpperCamelCase )
model.load_state_dict(UpperCamelCase )
model.eval()
# forward pass
lowerCAmelCase__ : int = model(UpperCamelCase )
lowerCAmelCase__ : List[Any] = outputs.predicted_depth
# verify output
if model_name is not None:
if "nyu" in model_name:
lowerCAmelCase__ : str = torch.tensor(
[[4.4147, 4.0873, 4.0673], [3.7890, 3.2881, 3.1525], [3.7674, 3.5423, 3.4913]] )
elif "kitti" in model_name:
lowerCAmelCase__ : Optional[int] = torch.tensor(
[[3.4291, 2.7865, 2.5151], [3.2841, 2.7021, 2.3502], [3.1147, 2.4625, 2.2481]] )
else:
raise ValueError(f"""Unknown model name: {model_name}""" )
lowerCAmelCase__ : Optional[int] = torch.Size([1, 480, 640] )
assert predicted_depth.shape == expected_shape
assert torch.allclose(predicted_depth[0, :3, :3] , UpperCamelCase , atol=1e-4 )
print("""Looks ok!""" )
# finally, push to hub if required
if push_to_hub:
logger.info("""Pushing model and image processor to the hub...""" )
model.push_to_hub(
repo_path_or_name=Path(UpperCamelCase , UpperCamelCase ) , organization="""nielsr""" , commit_message="""Add model""" , use_temp_dir=UpperCamelCase , )
image_processor.push_to_hub(
repo_path_or_name=Path(UpperCamelCase , UpperCamelCase ) , organization="""nielsr""" , commit_message="""Add image processor""" , use_temp_dir=UpperCamelCase , )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_path''',
default=None,
type=str,
help='''Path to the original PyTorch checkpoint (.pth file).''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the folder to output PyTorch model.'''
)
parser.add_argument(
'''--push_to_hub''', action='''store_true''', help='''Whether to upload the model to the HuggingFace hub.'''
)
parser.add_argument(
'''--model_name''',
default='''glpn-kitti''',
type=str,
help='''Name of the model in case you\'re pushing to the hub.''',
)
_lowerCAmelCase = parser.parse_args()
convert_glpn_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub, args.model_name)
| 37
|
'''simple docstring'''
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team.
#
# 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.
# this script dumps information about the environment
import os
import sys
import transformers
_lowerCAmelCase = '''3'''
print('''Python version:''', sys.version)
print('''transformers version:''', transformers.__version__)
try:
import torch
print('''Torch version:''', torch.__version__)
print('''Cuda available:''', torch.cuda.is_available())
print('''Cuda version:''', torch.version.cuda)
print('''CuDNN version:''', torch.backends.cudnn.version())
print('''Number of GPUs available:''', torch.cuda.device_count())
print('''NCCL version:''', torch.cuda.nccl.version())
except ImportError:
print('''Torch version:''', None)
try:
import deepspeed
print('''DeepSpeed version:''', deepspeed.__version__)
except ImportError:
print('''DeepSpeed version:''', None)
try:
import tensorflow as tf
print('''TensorFlow version:''', tf.__version__)
print('''TF GPUs available:''', bool(tf.config.list_physical_devices('''GPU''')))
print('''Number of TF GPUs available:''', len(tf.config.list_physical_devices('''GPU''')))
except ImportError:
print('''TensorFlow version:''', None)
| 37
| 1
|
'''simple docstring'''
from ..utils import DummyObject, requires_backends
class lowerCAmelCase_( metaclass=SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Tuple = ['''keras_nlp''']
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Union[str, Any]:
requires_backends(self ,["""keras_nlp"""] )
| 37
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xlm-roberta-xl''': '''https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json''',
'''facebook/xlm-roberta-xxl''': '''https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json''',
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''xlm-roberta-xl'''
def __init__( self ,__UpperCAmelCase=25_0880 ,__UpperCAmelCase=2560 ,__UpperCAmelCase=36 ,__UpperCAmelCase=32 ,__UpperCAmelCase=1_0240 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=514 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-05 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = vocab_size
lowerCAmelCase__ : int = hidden_size
lowerCAmelCase__ : int = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : str = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : List[str] = type_vocab_size
lowerCAmelCase__ : List[Any] = initializer_range
lowerCAmelCase__ : Tuple = layer_norm_eps
lowerCAmelCase__ : int = position_embedding_type
lowerCAmelCase__ : Optional[Any] = use_cache
lowerCAmelCase__ : Optional[Any] = classifier_dropout
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Dict = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Any = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
| 1
|
'''simple docstring'''
from pickle import UnpicklingError
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict
from ..utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
with open(UpperCamelCase , """rb""" ) as flax_state_f:
lowerCAmelCase__ : Union[str, Any] = from_bytes(UpperCamelCase , flax_state_f.read() )
except UnpicklingError as e:
try:
with open(UpperCamelCase ) as f:
if f.read().startswith("""version""" ):
raise OSError(
"""You seem to have cloned a repository without having git-lfs installed. Please"""
""" install git-lfs and run `git lfs install` followed by `git lfs pull` in the"""
""" folder you cloned.""" )
else:
raise ValueError from e
except (UnicodeDecodeError, ValueError):
raise EnvironmentError(f"""Unable to convert {model_file} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(UpperCamelCase , UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
import torch # noqa: F401
except ImportError:
logger.error(
"""Loading Flax weights in PyTorch requires both PyTorch and Flax to be installed. Please see"""
""" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"""
""" instructions.""" )
raise
# check if we have bf16 weights
lowerCAmelCase__ : str = flatten_dict(jax.tree_util.tree_map(lambda UpperCamelCase : x.dtype == jnp.bfloataa , UpperCamelCase ) ).values()
if any(UpperCamelCase ):
# convert all weights to fp32 if they are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
"""Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` """
"""before loading those in PyTorch model.""" )
lowerCAmelCase__ : Dict = jax.tree_util.tree_map(
lambda UpperCamelCase : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , UpperCamelCase )
lowerCAmelCase__ : Any = """"""
lowerCAmelCase__ : Any = flatten_dict(UpperCamelCase , sep=""".""" )
lowerCAmelCase__ : Optional[int] = pt_model.state_dict()
# keep track of unexpected & missing keys
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : int = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
lowerCAmelCase__ : Union[str, Any] = flax_key_tuple.split(""".""" )
if flax_key_tuple_array[-1] == "kernel" and flax_tensor.ndim == 4:
lowerCAmelCase__ : Optional[int] = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = jnp.transpose(UpperCamelCase , (3, 2, 0, 1) )
elif flax_key_tuple_array[-1] == "kernel":
lowerCAmelCase__ : str = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = flax_tensor.T
elif flax_key_tuple_array[-1] == "scale":
lowerCAmelCase__ : int = flax_key_tuple_array[:-1] + ["""weight"""]
if "time_embedding" not in flax_key_tuple_array:
for i, flax_key_tuple_string in enumerate(UpperCamelCase ):
lowerCAmelCase__ : List[str] = (
flax_key_tuple_string.replace("""_0""" , """.0""" )
.replace("""_1""" , """.1""" )
.replace("""_2""" , """.2""" )
.replace("""_3""" , """.3""" )
.replace("""_4""" , """.4""" )
.replace("""_5""" , """.5""" )
.replace("""_6""" , """.6""" )
.replace("""_7""" , """.7""" )
.replace("""_8""" , """.8""" )
.replace("""_9""" , """.9""" )
)
lowerCAmelCase__ : Union[str, Any] = """.""".join(UpperCamelCase )
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
f"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
f"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
lowerCAmelCase__ : int = np.asarray(UpperCamelCase ) if not isinstance(UpperCamelCase , np.ndarray ) else flax_tensor
lowerCAmelCase__ : int = torch.from_numpy(UpperCamelCase )
# remove from missing keys
missing_keys.remove(UpperCamelCase )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(UpperCamelCase )
pt_model.load_state_dict(UpperCamelCase )
# re-transform missing_keys to list
lowerCAmelCase__ : Optional[int] = list(UpperCamelCase )
if len(UpperCamelCase ) > 0:
logger.warning(
"""Some weights of the Flax model were not used when initializing the PyTorch model"""
f""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
f""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
""" (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This"""
f""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
""" to be exactly identical (e.g. initializing a BertForSequenceClassification model from a"""
""" FlaxBertForSequenceClassification model).""" )
if len(UpperCamelCase ) > 0:
logger.warning(
f"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
f""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
""" use it for predictions and inference.""" )
return pt_model
| 37
|
'''simple docstring'''
from __future__ import annotations
import math
from collections import Counter
from string import ascii_lowercase
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = analyze_text(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = list(""" """ + ascii_lowercase )
# what is our total sum of probabilities.
lowerCAmelCase__ : List[Any] = sum(single_char_strings.values() )
# one length string
lowerCAmelCase__ : Optional[int] = 0
# for each alpha we go in our dict and if it is in it we calculate entropy
for ch in my_alphas:
if ch in single_char_strings:
lowerCAmelCase__ : List[Any] = single_char_strings[ch]
lowerCAmelCase__ : List[Any] = my_str / all_sum
my_fir_sum += prob * math.loga(UpperCamelCase ) # entropy formula.
# print entropy
print(f"""{round(-1 * my_fir_sum ):.1f}""" )
# two len string
lowerCAmelCase__ : Dict = sum(two_char_strings.values() )
lowerCAmelCase__ : int = 0
# for each alpha (two in size) calculate entropy.
for cha in my_alphas:
for cha in my_alphas:
lowerCAmelCase__ : Union[str, Any] = cha + cha
if sequence in two_char_strings:
lowerCAmelCase__ : Dict = two_char_strings[sequence]
lowerCAmelCase__ : Tuple = int(UpperCamelCase ) / all_sum
my_sec_sum += prob * math.loga(UpperCamelCase )
# print second entropy
print(f"""{round(-1 * my_sec_sum ):.1f}""" )
# print the difference between them
print(f"""{round((-1 * my_sec_sum) - (-1 * my_fir_sum) ):.1f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = Counter() # type: ignore
lowerCAmelCase__ : Tuple = Counter() # type: ignore
single_char_strings[text[-1]] += 1
# first case when we have space at start.
two_char_strings[" " + text[0]] += 1
for i in range(0 , len(UpperCamelCase ) - 1 ):
single_char_strings[text[i]] += 1
two_char_strings[text[i : i + 2]] += 1
return single_char_strings, two_char_strings
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
import doctest
doctest.testmod()
# text = (
# "Had repulsive dashwoods suspicion sincerity but advantage now him. Remark "
# "easily garret nor nay. Civil those mrs enjoy shy fat merry. You greatest "
# "jointure saw horrible. He private he on be imagine suppose. Fertile "
# "beloved evident through no service elderly is. Blind there if every no so "
# "at. Own neglected you preferred way sincerity delivered his attempted. To "
# "of message cottage windows do besides against uncivil. Delightful "
# "unreserved impossible few estimating men favourable see entreaties. She "
# "propriety immediate was improving. He or entrance humoured likewise "
# "moderate. Much nor game son say feel. Fat make met can must form into "
# "gate. Me we offending prevailed discovery. "
# )
# calculate_prob(text)
if __name__ == "__main__":
main()
| 37
| 1
|
'''simple docstring'''
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 lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : List[Any] = ['''image_processor''', '''tokenizer''']
__lowercase : List[Any] = '''BridgeTowerImageProcessor'''
__lowercase : Dict = ('''RobertaTokenizer''', '''RobertaTokenizerFast''')
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> List[Any]:
super().__init__(__UpperCAmelCase ,__UpperCAmelCase )
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = True ,__UpperCAmelCase = False ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = 0 ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase = False ,__UpperCAmelCase = False ,__UpperCAmelCase = False ,__UpperCAmelCase = False ,__UpperCAmelCase = True ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> BatchEncoding:
lowerCAmelCase__ : int = 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
lowerCAmelCase__ : Optional[int] = self.image_processor(
__UpperCAmelCase ,return_tensors=__UpperCAmelCase ,do_normalize=__UpperCAmelCase ,do_center_crop=__UpperCAmelCase ,**__UpperCAmelCase )
encoding.update(__UpperCAmelCase )
return encoding
def UpperCAmelCase_ ( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> str:
return self.tokenizer.batch_decode(*__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Optional[Any]:
return self.tokenizer.decode(*__UpperCAmelCase ,**__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : List[str] = self.tokenizer.model_input_names
lowerCAmelCase__ : Tuple = self.image_processor.model_input_names
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
| 37
|
'''simple docstring'''
import argparse
import json
import torch
from diffusers import DDPMScheduler, LDMPipeline, UNetaDModel, VQModel
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=1 ):
"""simple docstring"""
if n_shave_prefix_segments >= 0:
return ".".join(path.split(""".""" )[n_shave_prefix_segments:] )
else:
return ".".join(path.split(""".""" )[:n_shave_prefix_segments] )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = []
for old_item in old_list:
lowerCAmelCase__ : Optional[Any] = old_item.replace("""in_layers.0""" , """norm1""" )
lowerCAmelCase__ : Optional[int] = new_item.replace("""in_layers.2""" , """conv1""" )
lowerCAmelCase__ : Dict = new_item.replace("""out_layers.0""" , """norm2""" )
lowerCAmelCase__ : str = new_item.replace("""out_layers.3""" , """conv2""" )
lowerCAmelCase__ : str = new_item.replace("""emb_layers.1""" , """time_emb_proj""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""skip_connection""" , """conv_shortcut""" )
lowerCAmelCase__ : Union[str, Any] = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : int = []
for old_item in old_list:
lowerCAmelCase__ : List[str] = old_item
lowerCAmelCase__ : int = new_item.replace("""norm.weight""" , """group_norm.weight""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""norm.bias""" , """group_norm.bias""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""proj_out.weight""" , """proj_attn.weight""" )
lowerCAmelCase__ : int = new_item.replace("""proj_out.bias""" , """proj_attn.bias""" )
lowerCAmelCase__ : str = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=None ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "Paths should be a list of dicts containing 'old' and 'new' keys."
# Splits the attention layers into three variables.
if attention_paths_to_split is not None:
for path, path_map in attention_paths_to_split.items():
lowerCAmelCase__ : Any = old_checkpoint[path]
lowerCAmelCase__ : int = old_tensor.shape[0] // 3
lowerCAmelCase__ : int = (-1, channels) if len(old_tensor.shape ) == 3 else (-1)
lowerCAmelCase__ : Tuple = old_tensor.shape[0] // config["""num_head_channels"""] // 3
lowerCAmelCase__ : List[Any] = old_tensor.reshape((num_heads, 3 * channels // num_heads) + old_tensor.shape[1:] )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : str = old_tensor.split(channels // num_heads , dim=1 )
lowerCAmelCase__ : int = query.reshape(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = key.reshape(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = value.reshape(UpperCamelCase )
for path in paths:
lowerCAmelCase__ : Any = path["""new"""]
# These have already been assigned
if attention_paths_to_split is not None and new_path in attention_paths_to_split:
continue
# Global renaming happens here
lowerCAmelCase__ : Any = new_path.replace("""middle_block.0""" , """mid_block.resnets.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.1""" , """mid_block.attentions.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.2""" , """mid_block.resnets.1""" )
if additional_replacements is not None:
for replacement in additional_replacements:
lowerCAmelCase__ : Any = new_path.replace(replacement["""old"""] , replacement["""new"""] )
# proj_attn.weight has to be converted from conv 1D to linear
if "proj_attn.weight" in new_path:
lowerCAmelCase__ : List[Any] = old_checkpoint[path["""old"""]][:, :, 0]
else:
lowerCAmelCase__ : Dict = old_checkpoint[path["""old"""]]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = {}
lowerCAmelCase__ : str = checkpoint["""time_embed.0.weight"""]
lowerCAmelCase__ : List[Any] = checkpoint["""time_embed.0.bias"""]
lowerCAmelCase__ : int = checkpoint["""time_embed.2.weight"""]
lowerCAmelCase__ : List[str] = checkpoint["""time_embed.2.bias"""]
lowerCAmelCase__ : str = checkpoint["""input_blocks.0.0.weight"""]
lowerCAmelCase__ : Any = checkpoint["""input_blocks.0.0.bias"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.weight"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.bias"""]
lowerCAmelCase__ : str = checkpoint["""out.2.weight"""]
lowerCAmelCase__ : Tuple = checkpoint["""out.2.bias"""]
# Retrieves the keys for the input blocks only
lowerCAmelCase__ : Optional[Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """input_blocks""" in layer} )
lowerCAmelCase__ : Optional[Any] = {
layer_id: [key for key in checkpoint if f"""input_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the middle blocks only
lowerCAmelCase__ : Union[str, Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """middle_block""" in layer} )
lowerCAmelCase__ : Union[str, Any] = {
layer_id: [key for key in checkpoint if f"""middle_block.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the output blocks only
lowerCAmelCase__ : List[str] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """output_blocks""" in layer} )
lowerCAmelCase__ : List[Any] = {
layer_id: [key for key in checkpoint if f"""output_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
for i in range(1 , UpperCamelCase ):
lowerCAmelCase__ : Dict = (i - 1) // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Tuple = (i - 1) % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Optional[int] = [key for key in input_blocks[i] if f"""input_blocks.{i}.0""" in key]
lowerCAmelCase__ : Optional[Any] = [key for key in input_blocks[i] if f"""input_blocks.{i}.1""" in key]
if f"""input_blocks.{i}.0.op.weight""" in checkpoint:
lowerCAmelCase__ : Optional[int] = checkpoint[
f"""input_blocks.{i}.0.op.weight"""
]
lowerCAmelCase__ : Tuple = checkpoint[
f"""input_blocks.{i}.0.op.bias"""
]
continue
lowerCAmelCase__ : Optional[Any] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Dict = {"""old""": f"""input_blocks.{i}.0""", """new""": f"""down_blocks.{block_id}.resnets.{layer_in_block_id}"""}
lowerCAmelCase__ : Optional[Any] = {"""old""": """resnets.2.op""", """new""": """downsamplers.0.op"""}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path, resnet_op] , config=UpperCamelCase )
if len(UpperCamelCase ):
lowerCAmelCase__ : Optional[Any] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""input_blocks.{i}.1""",
"""new""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : List[str] = {
f"""input_blocks.{i}.1.qkv.bias""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""input_blocks.{i}.1.qkv.weight""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase , )
lowerCAmelCase__ : Dict = middle_blocks[0]
lowerCAmelCase__ : Union[str, Any] = middle_blocks[1]
lowerCAmelCase__ : Dict = middle_blocks[2]
lowerCAmelCase__ : Any = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Dict = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {
"""middle_block.1.qkv.bias""": {
"""key""": """mid_block.attentions.0.key.bias""",
"""query""": """mid_block.attentions.0.query.bias""",
"""value""": """mid_block.attentions.0.value.bias""",
},
"""middle_block.1.qkv.weight""": {
"""key""": """mid_block.attentions.0.key.weight""",
"""query""": """mid_block.attentions.0.query.weight""",
"""value""": """mid_block.attentions.0.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase )
for i in range(UpperCamelCase ):
lowerCAmelCase__ : Tuple = i // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : List[str] = i % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : int = [shave_segments(UpperCamelCase , 2 ) for name in output_blocks[i]]
lowerCAmelCase__ : Union[str, Any] = {}
for layer in output_block_layers:
lowerCAmelCase__ , lowerCAmelCase__ : Any = layer.split(""".""" )[0], shave_segments(UpperCamelCase , 1 )
if layer_id in output_block_list:
output_block_list[layer_id].append(UpperCamelCase )
else:
lowerCAmelCase__ : str = [layer_name]
if len(UpperCamelCase ) > 1:
lowerCAmelCase__ : str = [key for key in output_blocks[i] if f"""output_blocks.{i}.0""" in key]
lowerCAmelCase__ : Dict = [key for key in output_blocks[i] if f"""output_blocks.{i}.1""" in key]
lowerCAmelCase__ : Optional[int] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {"""old""": f"""output_blocks.{i}.0""", """new""": f"""up_blocks.{block_id}.resnets.{layer_in_block_id}"""}
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , config=UpperCamelCase )
if ["conv.weight", "conv.bias"] in output_block_list.values():
lowerCAmelCase__ : List[Any] = list(output_block_list.values() ).index(["""conv.weight""", """conv.bias"""] )
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.weight"""
]
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.bias"""
]
# Clear attentions as they have been attributed above.
if len(UpperCamelCase ) == 2:
lowerCAmelCase__ : Tuple = []
if len(UpperCamelCase ):
lowerCAmelCase__ : Dict = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""output_blocks.{i}.1""",
"""new""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : Tuple = {
f"""output_blocks.{i}.1.qkv.bias""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""output_blocks.{i}.1.qkv.weight""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=to_split if any("""qkv""" in key for key in attentions ) else None , config=UpperCamelCase , )
else:
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase , n_shave_prefix_segments=1 )
for path in resnet_0_paths:
lowerCAmelCase__ : Tuple = """.""".join(["""output_blocks""", str(UpperCamelCase ), path["""old"""]] )
lowerCAmelCase__ : List[Any] = """.""".join(["""up_blocks""", str(UpperCamelCase ), """resnets""", str(UpperCamelCase ), path["""new"""]] )
lowerCAmelCase__ : Union[str, Any] = checkpoint[old_path]
return new_checkpoint
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the checkpoint to convert.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help='''The config json file corresponding to the architecture.''',
)
parser.add_argument('''--dump_path''', default=None, type=str, required=True, help='''Path to the output model.''')
_lowerCAmelCase = parser.parse_args()
_lowerCAmelCase = torch.load(args.checkpoint_path)
with open(args.config_file) as f:
_lowerCAmelCase = json.loads(f.read())
_lowerCAmelCase = convert_ldm_checkpoint(checkpoint, config)
if "ldm" in config:
del config["ldm"]
_lowerCAmelCase = UNetaDModel(**config)
model.load_state_dict(converted_checkpoint)
try:
_lowerCAmelCase = DDPMScheduler.from_config('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = VQModel.from_pretrained('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = LDMPipeline(unet=model, scheduler=scheduler, vae=vqvae)
pipe.save_pretrained(args.dump_path)
except: # noqa: E722
model.save_pretrained(args.dump_path)
| 37
| 1
|
'''simple docstring'''
from dataclasses import asdict, dataclass
from typing import Optional
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
# TODO Update this
_lowerCAmelCase = {
'''facebook/esm-1b''': '''https://huggingface.co/facebook/esm-1b/resolve/main/config.json''',
# See all ESM models at https://huggingface.co/models?filter=esm
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''esm'''
def __init__( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=768 ,__UpperCAmelCase=12 ,__UpperCAmelCase=12 ,__UpperCAmelCase=3072 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=1026 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-12 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Optional[int]:
super().__init__(pad_token_id=__UpperCAmelCase ,mask_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = vocab_size
lowerCAmelCase__ : List[Any] = hidden_size
lowerCAmelCase__ : Tuple = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : Optional[Any] = intermediate_size
lowerCAmelCase__ : Optional[Any] = hidden_dropout_prob
lowerCAmelCase__ : Union[str, Any] = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : Optional[int] = initializer_range
lowerCAmelCase__ : int = layer_norm_eps
lowerCAmelCase__ : str = position_embedding_type
lowerCAmelCase__ : int = use_cache
lowerCAmelCase__ : Tuple = emb_layer_norm_before
lowerCAmelCase__ : int = token_dropout
lowerCAmelCase__ : Any = is_folding_model
if is_folding_model:
if esmfold_config is None:
logger.info("""No esmfold_config supplied for folding model, using default values.""" )
lowerCAmelCase__ : List[Any] = EsmFoldConfig()
elif isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : int = EsmFoldConfig(**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = esmfold_config
if vocab_list is None:
logger.warning("""No vocab_list supplied for folding model, assuming the ESM-2 vocabulary!""" )
lowerCAmelCase__ : Tuple = get_default_vocab_list()
else:
lowerCAmelCase__ : Optional[Any] = vocab_list
else:
lowerCAmelCase__ : Optional[int] = None
lowerCAmelCase__ : Tuple = None
if self.esmfold_config is not None and getattr(self.esmfold_config ,"""use_esm_attn_map""" ,__UpperCAmelCase ):
raise ValueError("""The HuggingFace port of ESMFold does not support use_esm_attn_map at this time!""" )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : str = super().to_dict()
if isinstance(self.esmfold_config ,__UpperCAmelCase ):
lowerCAmelCase__ : List[str] = self.esmfold_config.to_dict()
return output
@dataclass
class lowerCAmelCase_:
'''simple docstring'''
__lowercase : str = None
__lowercase : bool = True
__lowercase : bool = False
__lowercase : bool = False
__lowercase : bool = False
__lowercase : float = 0
__lowercase : bool = True
__lowercase : bool = False
__lowercase : int = 1_2_8
__lowercase : "TrunkConfig" = None
def UpperCAmelCase_ ( self ) -> Any:
if self.trunk is None:
lowerCAmelCase__ : List[str] = TrunkConfig()
elif isinstance(self.trunk ,__UpperCAmelCase ):
lowerCAmelCase__ : Optional[int] = TrunkConfig(**self.trunk )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : str = asdict(self )
lowerCAmelCase__ : List[str] = self.trunk.to_dict()
return output
@dataclass
class lowerCAmelCase_:
'''simple docstring'''
__lowercase : int = 4_8
__lowercase : int = 1_0_2_4
__lowercase : int = 1_2_8
__lowercase : int = 3_2
__lowercase : int = 3_2
__lowercase : int = 3_2
__lowercase : float = 0
__lowercase : float = 0
__lowercase : bool = False
__lowercase : int = 4
__lowercase : Optional[int] = 1_2_8
__lowercase : "StructureModuleConfig" = None
def UpperCAmelCase_ ( self ) -> Dict:
if self.structure_module is None:
lowerCAmelCase__ : Union[str, Any] = StructureModuleConfig()
elif isinstance(self.structure_module ,__UpperCAmelCase ):
lowerCAmelCase__ : str = StructureModuleConfig(**self.structure_module )
if self.max_recycles <= 0:
raise ValueError(F"""`max_recycles` should be positive, got {self.max_recycles}.""" )
if self.sequence_state_dim % self.sequence_state_dim != 0:
raise ValueError(
"""`sequence_state_dim` should be a round multiple of `sequence_state_dim`, got"""
F""" {self.sequence_state_dim} and {self.sequence_state_dim}.""" )
if self.pairwise_state_dim % self.pairwise_state_dim != 0:
raise ValueError(
"""`pairwise_state_dim` should be a round multiple of `pairwise_state_dim`, got"""
F""" {self.pairwise_state_dim} and {self.pairwise_state_dim}.""" )
lowerCAmelCase__ : Union[str, Any] = self.sequence_state_dim // self.sequence_head_width
lowerCAmelCase__ : int = self.pairwise_state_dim // self.pairwise_head_width
if self.sequence_state_dim != sequence_num_heads * self.sequence_head_width:
raise ValueError(
"""`sequence_state_dim` should be equal to `sequence_num_heads * sequence_head_width, got"""
F""" {self.sequence_state_dim} != {sequence_num_heads} * {self.sequence_head_width}.""" )
if self.pairwise_state_dim != pairwise_num_heads * self.pairwise_head_width:
raise ValueError(
"""`pairwise_state_dim` should be equal to `pairwise_num_heads * pairwise_head_width, got"""
F""" {self.pairwise_state_dim} != {pairwise_num_heads} * {self.pairwise_head_width}.""" )
if self.pairwise_state_dim % 2 != 0:
raise ValueError(F"""`pairwise_state_dim` should be even, got {self.pairwise_state_dim}.""" )
if self.dropout >= 0.4:
raise ValueError(F"""`dropout` should not be greater than 0.4, got {self.dropout}.""" )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Any = asdict(self )
lowerCAmelCase__ : Union[str, Any] = self.structure_module.to_dict()
return output
@dataclass
class lowerCAmelCase_:
'''simple docstring'''
__lowercase : int = 3_8_4
__lowercase : int = 1_2_8
__lowercase : int = 1_6
__lowercase : int = 1_2_8
__lowercase : int = 1_2
__lowercase : int = 4
__lowercase : int = 8
__lowercase : float = 0.1
__lowercase : int = 8
__lowercase : int = 1
__lowercase : int = 2
__lowercase : int = 7
__lowercase : int = 1_0
__lowercase : float = 1e-8
__lowercase : float = 1e5
def UpperCAmelCase_ ( self ) -> Tuple:
return asdict(self )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
return (
"<cls>",
"<pad>",
"<eos>",
"<unk>",
"L",
"A",
"G",
"V",
"S",
"E",
"R",
"T",
"I",
"D",
"P",
"K",
"Q",
"N",
"F",
"Y",
"M",
"H",
"W",
"C",
"X",
"B",
"U",
"Z",
"O",
".",
"-",
"<null_1>",
"<mask>",
)
| 37
|
'''simple docstring'''
from math import sqrt
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' must been an int and positive"
lowerCAmelCase__ : int = True
# 0 and 1 are none primes.
if number <= 1:
lowerCAmelCase__ : Optional[Any] = False
for divisor in range(2 , int(round(sqrt(UpperCamelCase ) ) ) + 1 ):
# if 'number' divisible by 'divisor' then sets 'status'
# of false and break up the loop.
if number % divisor == 0:
lowerCAmelCase__ : Any = False
break
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'status' must been from type bool"
return status
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
# beginList: contains all natural numbers from 2 up to N
lowerCAmelCase__ : List[str] = list(range(2 , n + 1 ) )
lowerCAmelCase__ : str = [] # this list will be returns.
# actual sieve of erathostenes
for i in range(len(UpperCamelCase ) ):
for j in range(i + 1 , len(UpperCamelCase ) ):
if (begin_list[i] != 0) and (begin_list[j] % begin_list[i] == 0):
lowerCAmelCase__ : List[Any] = 0
# filters actual prime numbers.
lowerCAmelCase__ : List[Any] = [x for x in begin_list if x != 0]
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
lowerCAmelCase__ : List[str] = []
# iterates over all numbers between 2 up to N+1
# if a number is prime then appends to list 'ans'
for number in range(2 , n + 1 ):
if is_prime(UpperCamelCase ):
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and number >= 0, "'number' must been an int and >= 0"
lowerCAmelCase__ : Optional[Any] = [] # this list will be returns of the function.
# potential prime number factors.
lowerCAmelCase__ : Dict = 2
lowerCAmelCase__ : Dict = number
if number == 0 or number == 1:
ans.append(UpperCamelCase )
# if 'number' not prime then builds the prime factorization of 'number'
elif not is_prime(UpperCamelCase ):
while quotient != 1:
if is_prime(UpperCamelCase ) and (quotient % factor == 0):
ans.append(UpperCamelCase )
quotient /= factor
else:
factor += 1
else:
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : Optional[int] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = max(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = min(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 == 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 == 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 != 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 != 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase ) and (number > 2) and is_even(UpperCamelCase )
), "'number' must been an int, even and > 2"
lowerCAmelCase__ : Dict = [] # this list will returned
# creates a list of prime numbers between 2 up to 'number'
lowerCAmelCase__ : Dict = get_prime_numbers(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = len(UpperCamelCase )
# run variable for while-loops.
lowerCAmelCase__ : List[str] = 0
lowerCAmelCase__ : List[Any] = None
# exit variable. for break up the loops
lowerCAmelCase__ : Any = True
while i < len_pn and loop:
lowerCAmelCase__ : List[Any] = i + 1
while j < len_pn and loop:
if prime_numbers[i] + prime_numbers[j] == number:
lowerCAmelCase__ : Optional[Any] = False
ans.append(prime_numbers[i] )
ans.append(prime_numbers[j] )
j += 1
i += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (len(UpperCamelCase ) == 2)
and (ans[0] + ans[1] == number)
and is_prime(ans[0] )
and is_prime(ans[1] )
), "'ans' must contains two primes. And sum of elements must been eq 'number'"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 0)
and (numbera >= 0)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 0
while numbera != 0:
lowerCAmelCase__ : Any = numbera % numbera
lowerCAmelCase__ : str = numbera
lowerCAmelCase__ : List[str] = rest
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
numbera >= 0
), "'number' must been from type int and positive"
return numbera
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 1)
and (numbera >= 1)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 1 # actual answer that will be return.
# for kgV (x,1)
if numbera > 1 and numbera > 1:
# builds the prime factorization of 'number1' and 'number2'
lowerCAmelCase__ : int = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = prime_factorization(UpperCamelCase )
elif numbera == 1 or numbera == 1:
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : Dict = []
lowerCAmelCase__ : List[str] = max(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = 0
lowerCAmelCase__ : str = 0
lowerCAmelCase__ : List[Any] = [] # captured numbers int both 'primeFac1' and 'primeFac2'
# iterates through primeFac1
for n in prime_fac_a:
if n not in done:
if n in prime_fac_a:
lowerCAmelCase__ : int = prime_fac_a.count(UpperCamelCase )
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(max(UpperCamelCase , UpperCamelCase ) ):
ans *= n
else:
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# iterates through primeFac2
for n in prime_fac_a:
if n not in done:
lowerCAmelCase__ : Optional[int] = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
ans >= 0
), "'ans' must been from type int and positive"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'number' must been a positive int"
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Tuple = 2 # this variable holds the answer
while index < n:
index += 1
ans += 1 # counts to the next number
# if ans not prime then
# runs to the next prime number.
while not is_prime(UpperCamelCase ):
ans += 1
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and is_prime(
UpperCamelCase ), "'ans' must been a prime number and from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
is_prime(UpperCamelCase ) and is_prime(UpperCamelCase ) and (p_number_a < p_number_a)
), "The arguments must been prime numbers and 'pNumber1' < 'pNumber2'"
lowerCAmelCase__ : Dict = p_number_a + 1 # jump to the next number
lowerCAmelCase__ : List[Any] = [] # this list will be returns.
# if number is not prime then
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
while number < p_number_a:
ans.append(UpperCamelCase )
number += 1
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and ans[0] != p_number_a
and ans[len(UpperCamelCase ) - 1] != p_number_a
), "'ans' must been a list without the arguments"
# 'ans' contains not 'pNumber1' and 'pNumber2' !
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 1), "'n' must been int and >= 1"
lowerCAmelCase__ : List[Any] = [] # will be returned.
for divisor in range(1 , n + 1 ):
if n % divisor == 0:
ans.append(UpperCamelCase )
# precondition
assert ans[0] == 1 and ans[len(UpperCamelCase ) - 1] == n, "Error in function getDivisiors(...)"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number > 1
), "'number' must been an int and >= 1"
lowerCAmelCase__ : Optional[int] = get_divisors(UpperCamelCase )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (divisors[0] == 1)
and (divisors[len(UpperCamelCase ) - 1] == number)
), "Error in help-function getDivisiors(...)"
# summed all divisors up to 'number' (exclusive), hence [:-1]
return sum(divisors[:-1] ) == number
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (denominator != 0)
), "The arguments must been from type int and 'denominator' != 0"
# build the greatest common divisor of numerator and denominator.
lowerCAmelCase__ : int = gcd(abs(UpperCamelCase ) , abs(UpperCamelCase ) )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (numerator % gcd_of_fraction == 0)
and (denominator % gcd_of_fraction == 0)
), "Error in function gcd(...,...)"
return (numerator // gcd_of_fraction, denominator // gcd_of_fraction)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been a int and >= 0"
lowerCAmelCase__ : str = 1 # this will be return.
for factor in range(1 , n + 1 ):
ans *= factor
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
lowerCAmelCase__ : Any = 1
lowerCAmelCase__ : Optional[Any] = 1 # this will be return
for _ in range(n - 1 ):
lowerCAmelCase__ : Dict = ans
ans += fiba
lowerCAmelCase__ : str = tmp
return ans
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if p < 2:
raise ValueError("""p should not be less than 2!""" )
elif p == 2:
return True
lowerCAmelCase__ : List[str] = 4
lowerCAmelCase__ : Optional[int] = (1 << p) - 1
for _ in range(p - 2 ):
lowerCAmelCase__ : Optional[Any] = ((s * s) - 2) % m
return s == 0
if __name__ == "__main__":
print(lucas_lehmer_test(7))
print(lucas_lehmer_test(11))
| 37
|
'''simple docstring'''
from sklearn.metrics import fa_score, matthews_corrcoef
import datasets
from .record_evaluation import evaluate as evaluate_record
_lowerCAmelCase = '''\
@article{wang2019superglue,
title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
journal={arXiv preprint arXiv:1905.00537},
year={2019}
}
'''
_lowerCAmelCase = '''\
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
GLUE with a new set of more difficult language understanding tasks, improved
resources, and a new public leaderboard.
'''
_lowerCAmelCase = '''
Compute SuperGLUE evaluation metric associated to each SuperGLUE dataset.
Args:
predictions: list of predictions to score. Depending on the SuperGlUE subset:
- for \'record\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'prediction_text\': the predicted answer text
- for \'multirc\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question-answer pair as specified by the dataset
- \'prediction\': the predicted answer label
- otherwise: list of predicted labels
references: list of reference labels. Depending on the SuperGLUE subset:
- for \'record\': list of question-answers dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'answers\': list of possible answers
- otherwise: list of reference labels
Returns: depending on the SuperGLUE subset:
- for \'record\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1\': F1 score
- for \'multirc\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1_m\': Per-question macro-F1 score
- \'f1_a\': Average F1 score over all answers
- for \'axb\':
\'matthews_correlation\': Matthew Correlation
- for \'cb\':
- \'accuracy\': Accuracy
- \'f1\': F1 score
- for all others:
- \'accuracy\': Accuracy
Examples:
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'copa\') # any of ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'cb\')
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'record\')
>>> predictions = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'prediction_text\': \'answer\'}]
>>> references = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'answers\': [\'answer\', \'another_answer\']}]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'multirc\')
>>> predictions = [{\'idx\': {\'answer\': 0, \'paragraph\': 0, \'question\': 0}, \'prediction\': 0}, {\'idx\': {\'answer\': 1, \'paragraph\': 2, \'question\': 3}, \'prediction\': 1}]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1_m\': 1.0, \'f1_a\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'axb\')
>>> references = [0, 1]
>>> predictions = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'matthews_correlation\': 1.0}
'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return float((preds == labels).mean() )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase="binary" ):
"""simple docstring"""
lowerCAmelCase__ : Any = simple_accuracy(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = float(fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average=UpperCamelCase ) )
return {
"accuracy": acc,
"f1": fa,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
for id_pred, label in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : str = f"""{id_pred['idx']['paragraph']}-{id_pred['idx']['question']}"""
lowerCAmelCase__ : Dict = id_pred["""prediction"""]
if question_id in question_map:
question_map[question_id].append((pred, label) )
else:
lowerCAmelCase__ : Optional[int] = [(pred, label)]
lowerCAmelCase__ , lowerCAmelCase__ : int = [], []
for question, preds_labels in question_map.items():
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = zip(*UpperCamelCase )
lowerCAmelCase__ : List[Any] = fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average="""macro""" )
fas.append(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = int(sum(pred == label for pred, label in preds_labels ) == len(UpperCamelCase ) )
ems.append(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = float(sum(UpperCamelCase ) / len(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = sum(UpperCamelCase ) / len(UpperCamelCase )
lowerCAmelCase__ : Dict = float(fa_score(y_true=UpperCamelCase , y_pred=[id_pred["""prediction"""] for id_pred in ids_preds] ) )
return {"exact_match": em, "f1_m": fa_m, "f1_a": fa_a}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.config_name not in [
"boolq",
"cb",
"copa",
"multirc",
"record",
"rte",
"wic",
"wsc",
"wsc.fixed",
"axb",
"axg",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(self._get_feature_types() ) ,codebase_urls=[] ,reference_urls=[] ,format="""numpy""" if not self.config_name == """record""" and not self.config_name == """multirc""" else None ,)
def UpperCAmelCase_ ( self ) -> str:
if self.config_name == "record":
return {
"predictions": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"prediction_text": datasets.Value("""string""" ),
},
"references": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"answers": datasets.Sequence(datasets.Value("""string""" ) ),
},
}
elif self.config_name == "multirc":
return {
"predictions": {
"idx": {
"answer": datasets.Value("""int64""" ),
"paragraph": datasets.Value("""int64""" ),
"question": datasets.Value("""int64""" ),
},
"prediction": datasets.Value("""int64""" ),
},
"references": datasets.Value("""int64""" ),
}
else:
return {
"predictions": datasets.Value("""int64""" ),
"references": datasets.Value("""int64""" ),
}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
if self.config_name == "axb":
return {"matthews_correlation": matthews_corrcoef(__UpperCAmelCase ,__UpperCAmelCase )}
elif self.config_name == "cb":
return acc_and_fa(__UpperCAmelCase ,__UpperCAmelCase ,fa_avg="""macro""" )
elif self.config_name == "record":
lowerCAmelCase__ : Optional[Any] = [
{
"""qas""": [
{"""id""": ref["""idx"""]["""query"""], """answers""": [{"""text""": ans} for ans in ref["""answers"""]]}
for ref in references
]
}
]
lowerCAmelCase__ : Union[str, Any] = {pred["""idx"""]["""query"""]: pred["""prediction_text"""] for pred in predictions}
return evaluate_record(__UpperCAmelCase ,__UpperCAmelCase )[0]
elif self.config_name == "multirc":
return evaluate_multirc(__UpperCAmelCase ,__UpperCAmelCase )
elif self.config_name in ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]:
return {"accuracy": simple_accuracy(__UpperCAmelCase ,__UpperCAmelCase )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
| 37
| 1
|
'''simple docstring'''
from multiprocessing import Lock, Pipe, Process
# lock used to ensure that two processes do not access a pipe at the same time
_lowerCAmelCase = Lock()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
global process_lock
# we perform n swaps since after n swaps we know we are sorted
# we *could* stop early if we are sorted already, but it takes as long to
# find out we are sorted as it does to sort the list with this algorithm
for i in range(0 , 10 ):
if (i + position) % 2 == 0 and r_send is not None:
# send your value to your right neighbor
process_lock.acquire()
r_send[1].send(UpperCamelCase )
process_lock.release()
# receive your right neighbor's value
process_lock.acquire()
lowerCAmelCase__ : Any = rr_cv[0].recv()
process_lock.release()
# take the lower value since you are on the left
lowerCAmelCase__ : str = min(UpperCamelCase , UpperCamelCase )
elif (i + position) % 2 != 0 and l_send is not None:
# send your value to your left neighbor
process_lock.acquire()
l_send[1].send(UpperCamelCase )
process_lock.release()
# receive your left neighbor's value
process_lock.acquire()
lowerCAmelCase__ : Tuple = lr_cv[0].recv()
process_lock.release()
# take the higher value since you are on the right
lowerCAmelCase__ : Tuple = max(UpperCamelCase , UpperCamelCase )
# after all swaps are performed, send the values back to main
result_pipe[1].send(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = []
lowerCAmelCase__ : str = []
# initialize the list of pipes where the values will be retrieved
for _ in arr:
result_pipe.append(Pipe() )
# creates the processes
# the first and last process only have one neighbor so they are made outside
# of the loop
lowerCAmelCase__ : List[str] = Pipe()
lowerCAmelCase__ : Dict = Pipe()
process_array_.append(
Process(
target=UpperCamelCase , args=(0, arr[0], None, temp_rs, None, temp_rr, result_pipe[0]) , ) )
lowerCAmelCase__ : Union[str, Any] = temp_rs
lowerCAmelCase__ : Dict = temp_rr
for i in range(1 , len(UpperCamelCase ) - 1 ):
lowerCAmelCase__ : List[Any] = Pipe()
lowerCAmelCase__ : Dict = Pipe()
process_array_.append(
Process(
target=UpperCamelCase , args=(i, arr[i], temp_ls, temp_rs, temp_lr, temp_rr, result_pipe[i]) , ) )
lowerCAmelCase__ : Dict = temp_rs
lowerCAmelCase__ : Optional[Any] = temp_rr
process_array_.append(
Process(
target=UpperCamelCase , args=(
len(UpperCamelCase ) - 1,
arr[len(UpperCamelCase ) - 1],
temp_ls,
None,
temp_lr,
None,
result_pipe[len(UpperCamelCase ) - 1],
) , ) )
# start the processes
for p in process_array_:
p.start()
# wait for the processes to end and write their values to the list
for p in range(0 , len(UpperCamelCase ) ):
lowerCAmelCase__ : List[Any] = result_pipe[p][0].recv()
process_array_[p].join()
return arr
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = list(range(10 , 0 , -1 ) )
print("""Initial List""" )
print(*UpperCamelCase )
lowerCAmelCase__ : str = odd_even_transposition(UpperCamelCase )
print("""Sorted List\n""" )
print(*UpperCamelCase )
if __name__ == "__main__":
main()
| 37
|
'''simple docstring'''
import unittest
from pathlib import Path
from tempfile import NamedTemporaryFile, TemporaryDirectory
from transformers import BertConfig, BertTokenizerFast, FeatureExtractionPipeline
from transformers.convert_graph_to_onnx import (
convert,
ensure_valid_input,
generate_identified_filename,
infer_shapes,
quantize,
)
from transformers.testing_utils import require_tf, require_tokenizers, require_torch, slow
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[Any]:
return None
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
return None
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
__lowercase : Dict = [
# (model_name, model_kwargs)
('''bert-base-cased''', {}),
('''gpt2''', {'''use_cache''': False}), # We don't support exporting GPT2 past keys anymore
]
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> int:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Any:
from transformers import BertModel
lowerCAmelCase__ : Optional[int] = ["""[UNK]""", """[SEP]""", """[CLS]""", """[PAD]""", """[MASK]""", """some""", """other""", """words"""]
with NamedTemporaryFile(mode="""w+t""" ) as vocab_file:
vocab_file.write("""\n""".join(__UpperCAmelCase ) )
vocab_file.flush()
lowerCAmelCase__ : Dict = BertTokenizerFast(vocab_file.name )
with TemporaryDirectory() as bert_save_dir:
lowerCAmelCase__ : Tuple = BertModel(BertConfig(vocab_size=len(__UpperCAmelCase ) ) )
model.save_pretrained(__UpperCAmelCase )
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,__UpperCAmelCase )
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> List[str]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Dict = self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = quantize(Path(__UpperCAmelCase ) )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> List[Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Any = self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = quantize(__UpperCAmelCase )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[Any]:
try:
# Compute path
with TemporaryDirectory() as tempdir:
lowerCAmelCase__ : Optional[int] = Path(__UpperCAmelCase ).joinpath("""model.onnx""" )
# Remove folder if exists
if path.parent.exists():
path.parent.rmdir()
# Export
convert(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,**__UpperCAmelCase )
return path
except Exception as e:
self.fail(__UpperCAmelCase )
@require_torch
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
from transformers import BertModel
lowerCAmelCase__ : List[Any] = BertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Union[str, Any] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""pt""" )
@require_tf
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Optional[int]:
from transformers import TFBertModel
lowerCAmelCase__ : int = TFBertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Optional[int] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""tf""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = FeatureExtractionPipeline(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = ["""input_ids""", """token_type_ids""", """attention_mask""", """output_0""", """output_1"""]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = infer_shapes(__UpperCAmelCase ,__UpperCAmelCase )
# Assert all variables are present
self.assertEqual(len(__UpperCAmelCase ) ,len(__UpperCAmelCase ) )
self.assertTrue(all(var_name in shapes for var_name in variable_names ) )
self.assertSequenceEqual(variable_names[:3] ,__UpperCAmelCase )
self.assertSequenceEqual(variable_names[3:] ,__UpperCAmelCase )
# Assert inputs are {0: batch, 1: sequence}
for var_name in ["input_ids", "token_type_ids", "attention_mask"]:
self.assertDictEqual(shapes[var_name] ,{0: """batch""", 1: """sequence"""} )
# Assert outputs are {0: batch, 1: sequence} and {0: batch}
self.assertDictEqual(shapes["""output_0"""] ,{0: """batch""", 1: """sequence"""} )
self.assertDictEqual(shapes["""output_1"""] ,{0: """batch"""} )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = ["""input_ids""", """attention_mask""", """token_type_ids"""]
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": [1, 2, 3, 4], """attention_mask""": [0, 0, 0, 0], """token_type_ids""": [1, 1, 1, 1]}
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = ensure_valid_input(FuncContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the same number of args (all are valid)
self.assertEqual(len(__UpperCAmelCase ) ,3 )
# Should have exactly the same input names
self.assertEqual(set(__UpperCAmelCase ) ,set(__UpperCAmelCase ) )
# Parameter should be reordered according to their respective place in the function:
# (input_ids, token_type_ids, attention_mask)
self.assertEqual(__UpperCAmelCase ,(tokens["""input_ids"""], tokens["""token_type_ids"""], tokens["""attention_mask"""]) )
# Generated args are interleaved with another args (for instance parameter "past" in GPT2)
lowerCAmelCase__ , lowerCAmelCase__ : int = ensure_valid_input(FuncNonContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the one arg (all before the one not provided "some_other_args")
self.assertEqual(len(__UpperCAmelCase ) ,1 )
self.assertEqual(len(__UpperCAmelCase ) ,1 )
# Should have only "input_ids"
self.assertEqual(inputs_args[0] ,tokens["""input_ids"""] )
self.assertEqual(ordered_input_names[0] ,"""input_ids""" )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Dict = generate_identified_filename(Path("""/home/something/my_fake_model.onnx""" ) ,"""-test""" )
self.assertEqual("""/home/something/my_fake_model-test.onnx""" ,generated.as_posix() )
| 37
| 1
|
'''simple docstring'''
import tempfile
import unittest
from pathlib import Path
from shutil import copyfile
from transformers import BatchEncoding, MarianTokenizer
from transformers.testing_utils import get_tests_dir, require_sentencepiece, slow
from transformers.utils import is_sentencepiece_available, is_tf_available, is_torch_available
if is_sentencepiece_available():
from transformers.models.marian.tokenization_marian import VOCAB_FILES_NAMES, save_json
from ...test_tokenization_common import TokenizerTesterMixin
_lowerCAmelCase = get_tests_dir('''fixtures/test_sentencepiece.model''')
_lowerCAmelCase = {'''target_lang''': '''fi''', '''source_lang''': '''en'''}
_lowerCAmelCase = '''>>zh<<'''
_lowerCAmelCase = '''Helsinki-NLP/'''
if is_torch_available():
_lowerCAmelCase = '''pt'''
elif is_tf_available():
_lowerCAmelCase = '''tf'''
else:
_lowerCAmelCase = '''jax'''
@require_sentencepiece
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Any = MarianTokenizer
__lowercase : Optional[int] = False
__lowercase : Any = True
def UpperCAmelCase_ ( self ) -> str:
super().setUp()
lowerCAmelCase__ : Dict = ["""</s>""", """<unk>""", """▁This""", """▁is""", """▁a""", """▁t""", """est""", """\u0120""", """<pad>"""]
lowerCAmelCase__ : Optional[int] = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : Tuple = Path(self.tmpdirname )
save_json(__UpperCAmelCase ,save_dir / VOCAB_FILES_NAMES["""vocab"""] )
save_json(__UpperCAmelCase ,save_dir / VOCAB_FILES_NAMES["""tokenizer_config_file"""] )
if not (save_dir / VOCAB_FILES_NAMES["source_spm"]).exists():
copyfile(__UpperCAmelCase ,save_dir / VOCAB_FILES_NAMES["""source_spm"""] )
copyfile(__UpperCAmelCase ,save_dir / VOCAB_FILES_NAMES["""target_spm"""] )
lowerCAmelCase__ : List[Any] = MarianTokenizer.from_pretrained(self.tmpdirname )
tokenizer.save_pretrained(self.tmpdirname )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> MarianTokenizer:
return MarianTokenizer.from_pretrained(self.tmpdirname ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return (
"This is a test",
"This is a test",
)
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Optional[int] = """</s>"""
lowerCAmelCase__ : Optional[Any] = 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 ) -> List[Any]:
lowerCAmelCase__ : Union[str, Any] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] ,"""</s>""" )
self.assertEqual(vocab_keys[1] ,"""<unk>""" )
self.assertEqual(vocab_keys[-1] ,"""<pad>""" )
self.assertEqual(len(__UpperCAmelCase ) ,9 )
def UpperCAmelCase_ ( self ) -> Tuple:
self.assertEqual(self.get_tokenizer().vocab_size ,9 )
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Optional[int] = MarianTokenizer.from_pretrained(F"""{ORG_NAME}opus-mt-en-de""" )
lowerCAmelCase__ : Optional[Any] = en_de_tokenizer(["""I am a small frog"""] ,return_tensors=__UpperCAmelCase )
self.assertIsInstance(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : int = [38, 121, 14, 697, 3_8848, 0]
self.assertListEqual(__UpperCAmelCase ,batch.input_ids[0] )
lowerCAmelCase__ : List[str] = tempfile.mkdtemp()
en_de_tokenizer.save_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = [x.name for x in Path(__UpperCAmelCase ).glob("""*""" )]
self.assertIn("""source.spm""" ,__UpperCAmelCase )
MarianTokenizer.from_pretrained(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : Union[str, Any] = self.get_tokenizer()
lowerCAmelCase__ : Union[str, Any] = tok(
["""I am a small frog""" * 1000, """I am a small frog"""] ,padding=__UpperCAmelCase ,truncation=__UpperCAmelCase ,return_tensors=__UpperCAmelCase )
self.assertIsInstance(__UpperCAmelCase ,__UpperCAmelCase )
self.assertEqual(batch.input_ids.shape ,(2, 512) )
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Any = self.get_tokenizer()
lowerCAmelCase__ : Dict = tok(["""I am a tiny frog""", """I am a small frog"""] ,padding=__UpperCAmelCase ,return_tensors=__UpperCAmelCase )
self.assertIsInstance(__UpperCAmelCase ,__UpperCAmelCase )
self.assertEqual(batch_smaller.input_ids.shape ,(2, 10) )
@slow
def UpperCAmelCase_ ( self ) -> Tuple:
# fmt: off
lowerCAmelCase__ : List[str] = {"""input_ids""": [[4_3495, 462, 20, 4_2164, 1369, 52, 464, 132, 1703, 492, 13, 7491, 3_8999, 6, 8, 464, 132, 1703, 492, 13, 4669, 3_7867, 13, 7525, 27, 1593, 988, 13, 3_3972, 7029, 6, 20, 8251, 383, 2, 270, 5866, 3788, 2, 2353, 8251, 1_2338, 2, 1_3958, 387, 2, 3629, 6953, 188, 2900, 2, 1_3958, 8011, 1_1501, 23, 8460, 4073, 3_4009, 20, 435, 1_1439, 27, 8, 8460, 4073, 6004, 20, 9988, 375, 27, 33, 266, 1945, 1076, 1350, 3_7867, 3288, 5, 577, 1076, 4374, 8, 5082, 5, 2_6453, 257, 556, 403, 2, 242, 132, 383, 316, 492, 8, 1_0767, 6, 316, 304, 4239, 3, 0], [148, 1_5722, 19, 1839, 12, 1350, 13, 2_2327, 5082, 5418, 4_7567, 3_5938, 59, 318, 1_9552, 108, 2183, 54, 1_4976, 4835, 32, 547, 1114, 8, 315, 2417, 5, 92, 1_9088, 3, 0, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100], [36, 6395, 1_2570, 3_9147, 1_1597, 6, 266, 4, 4_5405, 7296, 3, 0, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100, 5_8100]], """attention_mask""": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 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], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=__UpperCAmelCase ,model_name="""Helsinki-NLP/opus-mt-en-de""" ,revision="""1a8c2263da11e68e50938f97e10cd57820bd504c""" ,decode_kwargs={"""use_source_tokenizer""": True} ,)
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : Optional[int] = MarianTokenizer.from_pretrained("""hf-internal-testing/test-marian-two-vocabs""" )
lowerCAmelCase__ : Tuple = """Tämä on testi"""
lowerCAmelCase__ : str = """This is a test"""
lowerCAmelCase__ : Tuple = [76, 7, 2047, 2]
lowerCAmelCase__ : Any = [69, 12, 11, 940, 2]
lowerCAmelCase__ : Union[str, Any] = tokenizer(__UpperCAmelCase ).input_ids
self.assertListEqual(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = tokenizer(text_target=__UpperCAmelCase ).input_ids
self.assertListEqual(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tokenizer.decode(__UpperCAmelCase ,skip_special_tokens=__UpperCAmelCase )
self.assertEqual(__UpperCAmelCase ,__UpperCAmelCase )
| 37
|
'''simple docstring'''
from maths.prime_factors import prime_factors
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if not isinstance(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : int = f"""Input value of [number={number}] must be an integer"""
raise TypeError(UpperCamelCase )
if number < 1:
raise ValueError("""Input must be a positive integer""" )
return -1 if len(prime_factors(UpperCamelCase ) ) % 2 else 1
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
| 1
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_torch_available,
is_vision_available,
)
_lowerCAmelCase = {
'''configuration_mobilevit''': ['''MOBILEVIT_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''MobileViTConfig''', '''MobileViTOnnxConfig'''],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = ['''MobileViTFeatureExtractor''']
_lowerCAmelCase = ['''MobileViTImageProcessor''']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = [
'''MOBILEVIT_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''MobileViTForImageClassification''',
'''MobileViTForSemanticSegmentation''',
'''MobileViTModel''',
'''MobileViTPreTrainedModel''',
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = [
'''TF_MOBILEVIT_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''TFMobileViTForImageClassification''',
'''TFMobileViTForSemanticSegmentation''',
'''TFMobileViTModel''',
'''TFMobileViTPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_mobilevit import MOBILEVIT_PRETRAINED_CONFIG_ARCHIVE_MAP, MobileViTConfig, MobileViTOnnxConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_mobilevit import MobileViTFeatureExtractor
from .image_processing_mobilevit import MobileViTImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mobilevit import (
MOBILEVIT_PRETRAINED_MODEL_ARCHIVE_LIST,
MobileViTForImageClassification,
MobileViTForSemanticSegmentation,
MobileViTModel,
MobileViTPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mobilevit import (
TF_MOBILEVIT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFMobileViTForImageClassification,
TFMobileViTForSemanticSegmentation,
TFMobileViTModel,
TFMobileViTPreTrainedModel,
)
else:
import sys
_lowerCAmelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure, module_spec=__spec__)
| 37
|
'''simple docstring'''
import os
import re
import unicodedata
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import is_torch_available, logging
if is_torch_available():
import torch
if TYPE_CHECKING:
from transformers.pipelines.conversational import Conversation
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''AI-Sweden/gpt-sw3-126m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-126m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-350m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-350m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-1.6b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-1.6b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-6.7b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-6.7b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-20b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-20b/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''AI-Sweden/gpt-sw3-126m''': 2048,
'''AI-Sweden/gpt-sw3-350m''': 2048,
'''AI-Sweden/gpt-sw3-1.6b''': 2048,
'''AI-Sweden/gpt-sw3-6.7b''': 2048,
'''AI-Sweden/gpt-sw3-20b''': 2048,
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : str = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Optional[int] = ['''input_ids''', '''attention_mask''']
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
lowerCAmelCase__ : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
lowerCAmelCase__ : Dict = kwargs.get("""name_or_path""" )
if name_or_path is None:
logger.warning(
"""name_or_path not provided, will work for all GPTSw3 models except gpt-sw3-7b,"""
""" you are testing the model, this can safely be ignored""" )
lowerCAmelCase__ : Tuple = """None"""
# Default definitions for our 2 tokenizer versions, with None-checks to enable proper testing
lowerCAmelCase__ : Union[str, Any] = """<|endoftext|>""" if eos_token is None else eos_token
lowerCAmelCase__ : Dict = """<unk>""" if unk_token is None else unk_token
if "gpt-sw3-7b" in name_or_path:
lowerCAmelCase__ : Any = unk_token if pad_token is None else pad_token
lowerCAmelCase__ : Dict = eos_token if bos_token is None else bos_token
else:
lowerCAmelCase__ : List[str] = """<pad>""" if pad_token is None else pad_token
lowerCAmelCase__ : Optional[int] = """<s>""" if bos_token is None else bos_token
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : Optional[int] = do_lower_case
lowerCAmelCase__ : Dict = remove_space
lowerCAmelCase__ : Optional[Any] = keep_accents
lowerCAmelCase__ : int = vocab_file
lowerCAmelCase__ : str = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
# Used for whitespace normalization in input texts
# fmt : off
lowerCAmelCase__ : int = {""" """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """""", """"""}
# fmt : on
# Regular expression to remove non-printing characters (e.g. some unicode control chars) in preprocessing
lowerCAmelCase__ : List[str] = re.compile(
F"""[{''.join(map(__UpperCAmelCase ,list(range(0 ,9 ) ) + list(range(11 ,32 ) ) + list(range(127 ,160 ) ) + [160, 173, 8203] ) )}]""" )
def __getstate__( self ) -> Any:
lowerCAmelCase__ : int = self.__dict__.copy()
lowerCAmelCase__ : Optional[int] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Tuple = {}
lowerCAmelCase__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
@property
# Copied from transformers.models.albert.tokenization_albert.AlbertTokenizer.vocab_size
def UpperCAmelCase_ ( self ) -> int:
return len(self.sp_model )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : Tuple = self.non_printing_characters_re.sub("""""" ,__UpperCAmelCase )
# Normalize whitespaces
lowerCAmelCase__ : List[Any] = """""".join([char if char not in self.whitespaces else """ """ for char in text] )
# NFC Unicode normalization
lowerCAmelCase__ : List[Any] = unicodedata.normalize("""NFC""" ,__UpperCAmelCase )
return text
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[Any] = self.preprocess_text(__UpperCAmelCase )
return self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.IdToPiece(__UpperCAmelCase )
@staticmethod
def UpperCAmelCase_ ( __UpperCAmelCase ) -> str:
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : int = []
lowerCAmelCase__ : Optional[int] = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
# TODO: Check if this is needed, as it ensures that decode(encode(doc)) != doc by adding extra whitespace in the decoded document
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Optional[Any] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Any = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string
def UpperCAmelCase_ ( self ) -> Dict[str, int]:
lowerCAmelCase__ : Optional[int] = {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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : Optional[int] = 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:
lowerCAmelCase__ : str = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Union[List[int], List[List[int]], "torch.Tensor"]:
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : int = self.sp_model.encode(__UpperCAmelCase )
else:
lowerCAmelCase__ : int = [self.preprocess_text(__UpperCAmelCase ) for t in text]
lowerCAmelCase__ : Any = self.sp_model.encode(__UpperCAmelCase )
if return_tensors is True or return_tensors == "pt":
lowerCAmelCase__ : Tuple = torch.tensor(__UpperCAmelCase )
return token_ids
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.decode(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[int]:
lowerCAmelCase__ : List[Any] = [F"""User: {text}""" if is_user else F"""Bot: {text}""" for is_user, text in conversation.iter_texts()]
lowerCAmelCase__ : Any = (
F"""{self.eos_token}{self.bos_token}""" + F"""{self.bos_token}""".join(__UpperCAmelCase ) + F"""{self.bos_token}Bot:"""
)
return self.encode(text=__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xlm-roberta-xl''': '''https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json''',
'''facebook/xlm-roberta-xxl''': '''https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json''',
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''xlm-roberta-xl'''
def __init__( self ,__UpperCAmelCase=25_0880 ,__UpperCAmelCase=2560 ,__UpperCAmelCase=36 ,__UpperCAmelCase=32 ,__UpperCAmelCase=1_0240 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=514 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-05 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = vocab_size
lowerCAmelCase__ : int = hidden_size
lowerCAmelCase__ : int = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : str = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : List[str] = type_vocab_size
lowerCAmelCase__ : List[Any] = initializer_range
lowerCAmelCase__ : Tuple = layer_norm_eps
lowerCAmelCase__ : int = position_embedding_type
lowerCAmelCase__ : Optional[Any] = use_cache
lowerCAmelCase__ : Optional[Any] = classifier_dropout
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Dict = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Any = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
|
'''simple docstring'''
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = XLMRobertaModel.from_pretrained("""xlm-roberta-base""" )
lowerCAmelCase__ : Optional[int] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : str = torch.Size((1, 12, 768) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.0_1_0_1, 0.1_2_1_8, -0.0_8_0_3, 0.0_8_0_1, 0.1_3_2_7, 0.0_7_7_6, -0.1_2_1_5, 0.2_3_8_3, 0.3_3_3_8, 0.3_1_0_6, 0.0_3_0_0, 0.0_2_5_2]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Optional[int] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
@slow
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : Union[str, Any] = XLMRobertaModel.from_pretrained("""xlm-roberta-large""" )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : Dict = torch.Size((1, 12, 1024) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[[-0.0_6_9_9, -0.0_3_1_8, 0.0_7_0_5, -0.1_2_4_1, 0.0_9_9_9, -0.0_5_2_0, 0.1_0_0_4, -0.1_8_3_8, -0.4_7_0_4, 0.1_4_3_7, 0.0_8_2_1, 0.0_1_2_6]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.large')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
| 37
| 1
|
'''simple docstring'''
import torch
from diffusers import DDPMScheduler
from .test_schedulers import SchedulerCommonTest
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Tuple = (DDPMScheduler,)
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Dict = {
"""num_train_timesteps""": 1000,
"""beta_start""": 0.0_0_0_1,
"""beta_end""": 0.0_2,
"""beta_schedule""": """linear""",
"""variance_type""": """fixed_small""",
"""clip_sample""": True,
}
config.update(**__UpperCAmelCase )
return config
def UpperCAmelCase_ ( self ) -> str:
for timesteps in [1, 5, 100, 1000]:
self.check_over_configs(num_train_timesteps=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[Any]:
for beta_start, beta_end in zip([0.0_0_0_1, 0.0_0_1, 0.0_1, 0.1] ,[0.0_0_2, 0.0_2, 0.2, 2] ):
self.check_over_configs(beta_start=__UpperCAmelCase ,beta_end=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> int:
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> int:
for variance in ["fixed_small", "fixed_large", "other"]:
self.check_over_configs(variance_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
self.check_over_configs(thresholding=__UpperCAmelCase )
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample", "v_prediction"]:
self.check_over_configs(
thresholding=__UpperCAmelCase ,prediction_type=__UpperCAmelCase ,sample_max_value=__UpperCAmelCase ,)
def UpperCAmelCase_ ( self ) -> int:
for prediction_type in ["epsilon", "sample", "v_prediction"]:
self.check_over_configs(prediction_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
for t in [0, 500, 999]:
self.check_over_forward(time_step=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : List[Any] = self.scheduler_classes[0]
lowerCAmelCase__ : Optional[Any] = self.get_scheduler_config()
lowerCAmelCase__ : Dict = scheduler_class(**__UpperCAmelCase )
assert torch.sum(torch.abs(scheduler._get_variance(0 ) - 0.0 ) ) < 1E-5
assert torch.sum(torch.abs(scheduler._get_variance(487 ) - 0.0_0_9_7_9 ) ) < 1E-5
assert torch.sum(torch.abs(scheduler._get_variance(999 ) - 0.0_2 ) ) < 1E-5
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : List[str] = self.scheduler_classes[0]
lowerCAmelCase__ : int = self.get_scheduler_config()
lowerCAmelCase__ : List[Any] = scheduler_class(**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = len(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = self.dummy_model()
lowerCAmelCase__ : str = self.dummy_sample_deter
lowerCAmelCase__ : List[str] = torch.manual_seed(0 )
for t in reversed(range(__UpperCAmelCase ) ):
# 1. predict noise residual
lowerCAmelCase__ : int = model(__UpperCAmelCase ,__UpperCAmelCase )
# 2. predict previous mean of sample x_t-1
lowerCAmelCase__ : Union[str, Any] = scheduler.step(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,generator=__UpperCAmelCase ).prev_sample
# if t > 0:
# noise = self.dummy_sample_deter
# variance = scheduler.get_variance(t) ** (0.5) * noise
#
# sample = pred_prev_sample + variance
lowerCAmelCase__ : List[Any] = pred_prev_sample
lowerCAmelCase__ : List[Any] = torch.sum(torch.abs(__UpperCAmelCase ) )
lowerCAmelCase__ : List[Any] = torch.mean(torch.abs(__UpperCAmelCase ) )
assert abs(result_sum.item() - 2_5_8.9_6_0_6 ) < 1E-2
assert abs(result_mean.item() - 0.3_3_7_2 ) < 1E-3
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : int = self.scheduler_classes[0]
lowerCAmelCase__ : List[str] = self.get_scheduler_config(prediction_type="""v_prediction""" )
lowerCAmelCase__ : List[str] = scheduler_class(**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = len(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = self.dummy_model()
lowerCAmelCase__ : Tuple = self.dummy_sample_deter
lowerCAmelCase__ : Union[str, Any] = torch.manual_seed(0 )
for t in reversed(range(__UpperCAmelCase ) ):
# 1. predict noise residual
lowerCAmelCase__ : List[Any] = model(__UpperCAmelCase ,__UpperCAmelCase )
# 2. predict previous mean of sample x_t-1
lowerCAmelCase__ : List[str] = scheduler.step(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,generator=__UpperCAmelCase ).prev_sample
# if t > 0:
# noise = self.dummy_sample_deter
# variance = scheduler.get_variance(t) ** (0.5) * noise
#
# sample = pred_prev_sample + variance
lowerCAmelCase__ : str = pred_prev_sample
lowerCAmelCase__ : List[Any] = torch.sum(torch.abs(__UpperCAmelCase ) )
lowerCAmelCase__ : Any = torch.mean(torch.abs(__UpperCAmelCase ) )
assert abs(result_sum.item() - 2_0_2.0_2_9_6 ) < 1E-2
assert abs(result_mean.item() - 0.2_6_3_1 ) < 1E-3
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = self.scheduler_classes[0]
lowerCAmelCase__ : List[Any] = self.get_scheduler_config()
lowerCAmelCase__ : Optional[Any] = scheduler_class(**__UpperCAmelCase )
lowerCAmelCase__ : str = [100, 87, 50, 1, 0]
scheduler.set_timesteps(timesteps=__UpperCAmelCase )
lowerCAmelCase__ : str = scheduler.timesteps
for i, timestep in enumerate(__UpperCAmelCase ):
if i == len(__UpperCAmelCase ) - 1:
lowerCAmelCase__ : Union[str, Any] = -1
else:
lowerCAmelCase__ : List[str] = timesteps[i + 1]
lowerCAmelCase__ : str = scheduler.previous_timestep(__UpperCAmelCase )
lowerCAmelCase__ : Dict = prev_t.item()
self.assertEqual(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Optional[int] = self.scheduler_classes[0]
lowerCAmelCase__ : Optional[int] = self.get_scheduler_config()
lowerCAmelCase__ : Any = scheduler_class(**__UpperCAmelCase )
lowerCAmelCase__ : Tuple = [100, 87, 50, 51, 0]
with self.assertRaises(__UpperCAmelCase ,msg="""`custom_timesteps` must be in descending order.""" ):
scheduler.set_timesteps(timesteps=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Dict = self.scheduler_classes[0]
lowerCAmelCase__ : List[str] = self.get_scheduler_config()
lowerCAmelCase__ : Tuple = scheduler_class(**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = [100, 87, 50, 1, 0]
lowerCAmelCase__ : Union[str, Any] = len(__UpperCAmelCase )
with self.assertRaises(__UpperCAmelCase ,msg="""Can only pass one of `num_inference_steps` or `custom_timesteps`.""" ):
scheduler.set_timesteps(num_inference_steps=__UpperCAmelCase ,timesteps=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[Any]:
lowerCAmelCase__ : List[Any] = self.scheduler_classes[0]
lowerCAmelCase__ : int = self.get_scheduler_config()
lowerCAmelCase__ : Tuple = scheduler_class(**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = [scheduler.config.num_train_timesteps]
with self.assertRaises(
__UpperCAmelCase ,msg="""`timesteps` must start before `self.config.train_timesteps`: {scheduler.config.num_train_timesteps}}""" ,):
scheduler.set_timesteps(timesteps=__UpperCAmelCase )
| 37
|
'''simple docstring'''
from pickle import UnpicklingError
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict
from ..utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
with open(UpperCamelCase , """rb""" ) as flax_state_f:
lowerCAmelCase__ : Union[str, Any] = from_bytes(UpperCamelCase , flax_state_f.read() )
except UnpicklingError as e:
try:
with open(UpperCamelCase ) as f:
if f.read().startswith("""version""" ):
raise OSError(
"""You seem to have cloned a repository without having git-lfs installed. Please"""
""" install git-lfs and run `git lfs install` followed by `git lfs pull` in the"""
""" folder you cloned.""" )
else:
raise ValueError from e
except (UnicodeDecodeError, ValueError):
raise EnvironmentError(f"""Unable to convert {model_file} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(UpperCamelCase , UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
import torch # noqa: F401
except ImportError:
logger.error(
"""Loading Flax weights in PyTorch requires both PyTorch and Flax to be installed. Please see"""
""" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"""
""" instructions.""" )
raise
# check if we have bf16 weights
lowerCAmelCase__ : str = flatten_dict(jax.tree_util.tree_map(lambda UpperCamelCase : x.dtype == jnp.bfloataa , UpperCamelCase ) ).values()
if any(UpperCamelCase ):
# convert all weights to fp32 if they are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
"""Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` """
"""before loading those in PyTorch model.""" )
lowerCAmelCase__ : Dict = jax.tree_util.tree_map(
lambda UpperCamelCase : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , UpperCamelCase )
lowerCAmelCase__ : Any = """"""
lowerCAmelCase__ : Any = flatten_dict(UpperCamelCase , sep=""".""" )
lowerCAmelCase__ : Optional[int] = pt_model.state_dict()
# keep track of unexpected & missing keys
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : int = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
lowerCAmelCase__ : Union[str, Any] = flax_key_tuple.split(""".""" )
if flax_key_tuple_array[-1] == "kernel" and flax_tensor.ndim == 4:
lowerCAmelCase__ : Optional[int] = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = jnp.transpose(UpperCamelCase , (3, 2, 0, 1) )
elif flax_key_tuple_array[-1] == "kernel":
lowerCAmelCase__ : str = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = flax_tensor.T
elif flax_key_tuple_array[-1] == "scale":
lowerCAmelCase__ : int = flax_key_tuple_array[:-1] + ["""weight"""]
if "time_embedding" not in flax_key_tuple_array:
for i, flax_key_tuple_string in enumerate(UpperCamelCase ):
lowerCAmelCase__ : List[str] = (
flax_key_tuple_string.replace("""_0""" , """.0""" )
.replace("""_1""" , """.1""" )
.replace("""_2""" , """.2""" )
.replace("""_3""" , """.3""" )
.replace("""_4""" , """.4""" )
.replace("""_5""" , """.5""" )
.replace("""_6""" , """.6""" )
.replace("""_7""" , """.7""" )
.replace("""_8""" , """.8""" )
.replace("""_9""" , """.9""" )
)
lowerCAmelCase__ : Union[str, Any] = """.""".join(UpperCamelCase )
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
f"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
f"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
lowerCAmelCase__ : int = np.asarray(UpperCamelCase ) if not isinstance(UpperCamelCase , np.ndarray ) else flax_tensor
lowerCAmelCase__ : int = torch.from_numpy(UpperCamelCase )
# remove from missing keys
missing_keys.remove(UpperCamelCase )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(UpperCamelCase )
pt_model.load_state_dict(UpperCamelCase )
# re-transform missing_keys to list
lowerCAmelCase__ : Optional[int] = list(UpperCamelCase )
if len(UpperCamelCase ) > 0:
logger.warning(
"""Some weights of the Flax model were not used when initializing the PyTorch model"""
f""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
f""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
""" (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This"""
f""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
""" to be exactly identical (e.g. initializing a BertForSequenceClassification model from a"""
""" FlaxBertForSequenceClassification model).""" )
if len(UpperCamelCase ) > 0:
logger.warning(
f"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
f""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
""" use it for predictions and inference.""" )
return pt_model
| 37
| 1
|
'''simple docstring'''
import copy
import json
import os
import tempfile
from transformers import is_torch_available
from .test_configuration_utils import config_common_kwargs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : List[str] = parent
lowerCAmelCase__ : Optional[Any] = config_class
lowerCAmelCase__ : List[str] = has_text_modality
lowerCAmelCase__ : Union[str, Any] = kwargs
lowerCAmelCase__ : str = common_properties
def UpperCAmelCase_ ( self ) -> List[Any]:
lowerCAmelCase__ : int = self.config_class(**self.inputs_dict )
lowerCAmelCase__ : Dict = (
["""hidden_size""", """num_attention_heads""", """num_hidden_layers"""]
if self.common_properties is None
else self.common_properties
)
# Add common fields for text models
if self.has_text_modality:
common_properties.extend(["""vocab_size"""] )
# Test that config has the common properties as getters
for prop in common_properties:
self.parent.assertTrue(hasattr(__UpperCAmelCase ,__UpperCAmelCase ) ,msg=F"""`{prop}` does not exist""" )
# Test that config has the common properties as setter
for idx, name in enumerate(__UpperCAmelCase ):
try:
setattr(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase )
self.parent.assertEqual(
getattr(__UpperCAmelCase ,__UpperCAmelCase ) ,__UpperCAmelCase ,msg=F"""`{name} value {idx} expected, but was {getattr(__UpperCAmelCase ,__UpperCAmelCase )}""" )
except NotImplementedError:
# Some models might not be able to implement setters for common_properties
# In that case, a NotImplementedError is raised
pass
# Test if config class can be called with Config(prop_name=..)
for idx, name in enumerate(__UpperCAmelCase ):
try:
lowerCAmelCase__ : str = self.config_class(**{name: idx} )
self.parent.assertEqual(
getattr(__UpperCAmelCase ,__UpperCAmelCase ) ,__UpperCAmelCase ,msg=F"""`{name} value {idx} expected, but was {getattr(__UpperCAmelCase ,__UpperCAmelCase )}""" )
except NotImplementedError:
# Some models might not be able to implement setters for common_properties
# In that case, a NotImplementedError is raised
pass
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : str = self.config_class(**self.inputs_dict )
lowerCAmelCase__ : Union[str, Any] = json.loads(config.to_json_string() )
for key, value in self.inputs_dict.items():
self.parent.assertEqual(obj[key] ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : int = self.config_class(**self.inputs_dict )
with tempfile.TemporaryDirectory() as tmpdirname:
lowerCAmelCase__ : Union[str, Any] = os.path.join(__UpperCAmelCase ,"""config.json""" )
config_first.to_json_file(__UpperCAmelCase )
lowerCAmelCase__ : str = self.config_class.from_json_file(__UpperCAmelCase )
self.parent.assertEqual(config_second.to_dict() ,config_first.to_dict() )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Optional[Any] = self.config_class(**self.inputs_dict )
with tempfile.TemporaryDirectory() as tmpdirname:
config_first.save_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = self.config_class.from_pretrained(__UpperCAmelCase )
self.parent.assertEqual(config_second.to_dict() ,config_first.to_dict() )
def UpperCAmelCase_ ( self ) -> List[Any]:
lowerCAmelCase__ : List[str] = self.config_class(**self.inputs_dict )
lowerCAmelCase__ : Union[str, Any] = """test"""
with tempfile.TemporaryDirectory() as tmpdirname:
lowerCAmelCase__ : List[Any] = os.path.join(__UpperCAmelCase ,__UpperCAmelCase )
config_first.save_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.config_class.from_pretrained(__UpperCAmelCase ,subfolder=__UpperCAmelCase )
self.parent.assertEqual(config_second.to_dict() ,config_first.to_dict() )
def UpperCAmelCase_ ( self ) -> Dict:
lowerCAmelCase__ : Dict = self.config_class(**self.inputs_dict ,num_labels=5 )
self.parent.assertEqual(len(config.idalabel ) ,5 )
self.parent.assertEqual(len(config.labelaid ) ,5 )
lowerCAmelCase__ : List[str] = 3
self.parent.assertEqual(len(config.idalabel ) ,3 )
self.parent.assertEqual(len(config.labelaid ) ,3 )
def UpperCAmelCase_ ( self ) -> str:
if self.config_class.is_composition:
return
lowerCAmelCase__ : Tuple = self.config_class()
self.parent.assertIsNotNone(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : Optional[Any] = copy.deepcopy(__UpperCAmelCase )
lowerCAmelCase__ : int = self.config_class(**__UpperCAmelCase )
lowerCAmelCase__ : Dict = []
for key, value in config_common_kwargs.items():
if key == "torch_dtype":
if not is_torch_available():
continue
else:
import torch
if config.torch_dtype != torch.floataa:
wrong_values.append(("""torch_dtype""", config.torch_dtype, torch.floataa) )
elif getattr(__UpperCAmelCase ,__UpperCAmelCase ) != value:
wrong_values.append((key, getattr(__UpperCAmelCase ,__UpperCAmelCase ), value) )
if len(__UpperCAmelCase ) > 0:
lowerCAmelCase__ : Any = """\n""".join([F"""- {v[0]}: got {v[1]} instead of {v[2]}""" for v in wrong_values] )
raise ValueError(F"""The following keys were not properly set in the config:\n{errors}""" )
def UpperCAmelCase_ ( self ) -> Tuple:
self.create_and_test_config_common_properties()
self.create_and_test_config_to_json_string()
self.create_and_test_config_to_json_file()
self.create_and_test_config_from_and_save_pretrained()
self.create_and_test_config_from_and_save_pretrained_subfolder()
self.create_and_test_config_with_num_labels()
self.check_config_can_be_init_without_params()
self.check_config_arguments_init()
| 37
|
'''simple docstring'''
import inspect
import unittest
import numpy as np
from tests.test_modeling_common import floats_tensor
from transformers import DetrConfig, MaskFormerConfig, SwinConfig, is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device
from transformers.utils import cached_property
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaskFormerForInstanceSegmentation, MaskFormerModel
if is_vision_available():
from transformers import MaskFormerImageProcessor
if is_vision_available():
from PIL import Image
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=2 ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase=10 ,__UpperCAmelCase=3 ,__UpperCAmelCase=32 * 4 ,__UpperCAmelCase=32 * 6 ,__UpperCAmelCase=4 ,__UpperCAmelCase=32 ,) -> str:
lowerCAmelCase__ : Optional[int] = parent
lowerCAmelCase__ : Optional[int] = batch_size
lowerCAmelCase__ : Optional[int] = is_training
lowerCAmelCase__ : Dict = use_auxiliary_loss
lowerCAmelCase__ : Union[str, Any] = num_queries
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : List[str] = min_size
lowerCAmelCase__ : int = max_size
lowerCAmelCase__ : Optional[Any] = num_labels
lowerCAmelCase__ : List[Any] = mask_feature_size
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : str = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to(
__UpperCAmelCase )
lowerCAmelCase__ : str = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=__UpperCAmelCase )
lowerCAmelCase__ : Any = (
torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=__UpperCAmelCase ) > 0.5
).float()
lowerCAmelCase__ : Optional[int] = (torch.rand((self.batch_size, self.num_labels) ,device=__UpperCAmelCase ) > 0.5).long()
lowerCAmelCase__ : Any = self.get_config()
return config, pixel_values, pixel_mask, mask_labels, class_labels
def UpperCAmelCase_ ( self ) -> Dict:
return MaskFormerConfig.from_backbone_and_decoder_configs(
backbone_config=SwinConfig(
depths=[1, 1, 1, 1] ,) ,decoder_config=DetrConfig(
decoder_ffn_dim=128 ,num_queries=self.num_queries ,decoder_attention_heads=2 ,d_model=self.mask_feature_size ,) ,mask_feature_size=self.mask_feature_size ,fpn_feature_size=self.mask_feature_size ,num_channels=self.num_channels ,num_labels=self.num_labels ,)
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.prepare_config_and_inputs()
lowerCAmelCase__ : List[str] = {"""pixel_values""": pixel_values, """pixel_mask""": pixel_mask}
return config, inputs_dict
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Optional[int] = output.encoder_hidden_states
lowerCAmelCase__ : Optional[int] = output.pixel_decoder_hidden_states
lowerCAmelCase__ : Dict = output.transformer_decoder_hidden_states
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,config.decoder_config.decoder_layers )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=False ) -> Optional[Any]:
with torch.no_grad():
lowerCAmelCase__ : int = MaskFormerModel(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
lowerCAmelCase__ : str = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : int = model(__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
# the correct shape of output.transformer_decoder_hidden_states ensure the correcteness of the
# encoder and pixel decoder
self.parent.assertEqual(
output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.mask_feature_size) ,)
# let's ensure the other two hidden state exists
self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(output.encoder_last_hidden_state is not None )
if output_hidden_states:
self.check_output_hidden_state(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : Dict = MaskFormerForInstanceSegmentation(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
def comm_check_on_output(__UpperCAmelCase ):
# let's still check that all the required stuff is there
self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.encoder_last_hidden_state is not None )
# okay, now we need to check the logits shape
# due to the encoder compression, masks have a //4 spatial size
self.parent.assertEqual(
result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,)
# + 1 for null class
self.parent.assertEqual(
result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : Dict = model(__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = model(
pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
self.parent.assertTrue(result.loss is not None )
self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) )
@require_torch
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Optional[Any] = (MaskFormerModel, MaskFormerForInstanceSegmentation) if is_torch_available() else ()
__lowercase : int = (
{'''feature-extraction''': MaskFormerModel, '''image-segmentation''': MaskFormerForInstanceSegmentation}
if is_torch_available()
else {}
)
__lowercase : Union[str, Any] = False
__lowercase : Dict = False
__lowercase : Tuple = False
__lowercase : List[Any] = False
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : str = MaskFormerModelTester(self )
lowerCAmelCase__ : List[Any] = ConfigTester(self ,config_class=__UpperCAmelCase ,has_text_modality=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
self.config_tester.run_common_tests()
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ , lowerCAmelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_maskformer_instance_segmentation_head_model(*__UpperCAmelCase )
@unittest.skip(reason="""MaskFormer does not use inputs_embeds""" )
def UpperCAmelCase_ ( self ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormer does not have a get_input_embeddings method""" )
def UpperCAmelCase_ ( self ) -> str:
pass
@unittest.skip(reason="""MaskFormer is not a generative model""" )
def UpperCAmelCase_ ( self ) -> Any:
pass
@unittest.skip(reason="""MaskFormer does not use token embeddings""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
@require_torch_multi_gpu
@unittest.skip(
reason="""MaskFormer has some layers using `add_module` which doesn't work well with `nn.DataParallel`""" )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
pass
@unittest.skip("""Will be fixed soon by reducing the size of the model used for common tests.""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : str = model_class(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowerCAmelCase__ : Dict = [*signature.parameters.keys()]
lowerCAmelCase__ : Tuple = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] ,__UpperCAmelCase )
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model_name in ["facebook/maskformer-swin-small-coco"]:
lowerCAmelCase__ : List[str] = MaskFormerModel.from_pretrained(__UpperCAmelCase )
self.assertIsNotNone(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (self.model_tester.min_size,) * 2
lowerCAmelCase__ : Any = {
"""pixel_values""": torch.randn((2, 3, *size) ,device=__UpperCAmelCase ),
"""mask_labels""": torch.randn((2, 10, *size) ,device=__UpperCAmelCase ),
"""class_labels""": torch.zeros(2 ,10 ,device=__UpperCAmelCase ).long(),
}
lowerCAmelCase__ : Tuple = MaskFormerForInstanceSegmentation(MaskFormerConfig() ).to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : int = model_class(__UpperCAmelCase ).to(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase ,output_attentions=__UpperCAmelCase )
self.assertTrue(outputs.attentions is not None )
def UpperCAmelCase_ ( self ) -> int:
if not self.model_tester.is_training:
return
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Dict = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : List[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : List[str] = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase ).loss
loss.backward()
def UpperCAmelCase_ ( self ) -> List[str]:
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Tuple = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Tuple = True
lowerCAmelCase__ : Optional[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : Dict = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
lowerCAmelCase__ : str = outputs.pixel_decoder_hidden_states[0]
pixel_decoder_hidden_states.retain_grad()
# we requires_grad=True in inputs_embeds (line 2152), the original implementation don't
lowerCAmelCase__ : Union[str, Any] = outputs.transformer_decoder_hidden_states[0]
transformer_decoder_hidden_states.retain_grad()
lowerCAmelCase__ : List[Any] = outputs.attentions[0]
attentions.retain_grad()
outputs.loss.backward(retain_graph=__UpperCAmelCase )
self.assertIsNotNone(encoder_hidden_states.grad )
self.assertIsNotNone(pixel_decoder_hidden_states.grad )
self.assertIsNotNone(transformer_decoder_hidden_states.grad )
self.assertIsNotNone(attentions.grad )
_lowerCAmelCase = 1e-4
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_vision
@slow
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@cached_property
def UpperCAmelCase_ ( self ) -> List[Any]:
return (
MaskFormerImageProcessor.from_pretrained("""facebook/maskformer-swin-small-coco""" )
if is_vision_available()
else None
)
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Any = MaskFormerModel.from_pretrained("""facebook/maskformer-swin-small-coco""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = self.default_image_processor
lowerCAmelCase__ : str = prepare_img()
lowerCAmelCase__ : Optional[int] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[-0.0_4_8_2, 0.9_2_2_8, 0.4_9_5_1], [-0.2_5_4_7, 0.8_0_1_7, 0.8_5_2_7], [-0.0_0_6_9, 0.3_3_8_5, -0.0_0_8_9]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.encoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.8_4_2_2, -0.8_4_3_4, -0.9_7_1_8], [-1.0_1_4_4, -0.5_5_6_5, -0.4_1_9_5], [-1.0_0_3_8, -0.4_4_8_4, -0.1_9_6_1]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Optional[int] = torch.tensor(
[[0.2_8_5_2, -0.0_1_5_9, 0.9_7_3_5], [0.6_2_5_4, 0.1_8_5_8, 0.8_5_2_9], [-0.0_6_8_0, -0.4_1_1_6, 1.8_4_1_3]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : List[str] = prepare_img()
lowerCAmelCase__ : str = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[int] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : Optional[int] = [
[-1.3_7_3_7_1_2_4, -1.7_7_2_4_9_3_7, -1.9_3_6_4_2_3_3],
[-1.5_9_7_7_2_8_1, -1.9_8_6_7_9_3_9, -2.1_5_2_3_6_9_5],
[-1.5_7_9_5_3_9_8, -1.9_2_6_9_8_3_2, -2.0_9_3_9_4_2],
]
lowerCAmelCase__ : Optional[int] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Tuple = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[
[1.65_12E00, -5.25_72E00, -3.35_19E00],
[3.61_69E-02, -5.90_25E00, -2.93_13E00],
[1.07_66E-04, -7.76_30E00, -5.12_63E00],
] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-resnet101-coco-stuff""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : int = prepare_img()
lowerCAmelCase__ : Optional[Any] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : str = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : int = [[-0.9_0_4_6, -2.6_3_6_6, -4.6_0_6_2], [-3.4_1_7_9, -5.7_8_9_0, -8.8_0_5_7], [-4.9_1_7_9, -7.6_5_6_0, -1_0.7_7_1_1]]
lowerCAmelCase__ : List[str] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Tuple = torch.tensor(
[[4.7_1_8_8, -3.2_5_8_5, -2.8_8_5_7], [6.6_8_7_1, -2.9_1_8_1, -1.2_4_8_7], [7.2_4_4_9, -2.2_7_6_4, -2.1_8_7_4]] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : str = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Dict = self.default_image_processor
lowerCAmelCase__ : Union[str, Any] = image_processor(
[np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors="""pt""" ,)
lowerCAmelCase__ : Tuple = inputs["""pixel_values"""].to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""mask_labels"""]]
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""class_labels"""]]
with torch.no_grad():
lowerCAmelCase__ : Any = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
| 37
| 1
|
'''simple docstring'''
import argparse
import json
import os
import fairseq
import torch
from fairseq.data import Dictionary
from transformers import (
WavaVecaConformerConfig,
WavaVecaConformerForCTC,
WavaVecaConformerForPreTraining,
WavaVecaCTCTokenizer,
WavaVecaFeatureExtractor,
WavaVecaProcessor,
logging,
)
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''post_extract_proj''': '''feature_projection.projection''',
'''encoder.pos_conv.0''': '''encoder.pos_conv_embed.conv''',
'''self_attn.linear_k''': '''encoder.layers.*.self_attn.linear_k''',
'''self_attn.linear_v''': '''encoder.layers.*.self_attn.linear_v''',
'''self_attn.linear_q''': '''encoder.layers.*.self_attn.linear_q''',
'''self_attn.pos_bias_u''': '''encoder.layers.*.self_attn.pos_bias_u''',
'''self_attn.pos_bias_v''': '''encoder.layers.*.self_attn.pos_bias_v''',
'''self_attn.linear_out''': '''encoder.layers.*.self_attn.linear_out''',
'''self_attn.linear_pos''': '''encoder.layers.*.self_attn.linear_pos''',
'''self_attn.rotary_emb''': '''encoder.embed_positions''',
'''self_attn_layer_norm''': '''encoder.layers.*.self_attn_layer_norm''',
'''conv_module.pointwise_conv1''': '''encoder.layers.*.conv_module.pointwise_conv1''',
'''conv_module.pointwise_conv2''': '''encoder.layers.*.conv_module.pointwise_conv2''',
'''conv_module.depthwise_conv''': '''encoder.layers.*.conv_module.depthwise_conv''',
'''conv_module.batch_norm''': '''encoder.layers.*.conv_module.batch_norm''',
'''conv_module.layer_norm''': '''encoder.layers.*.conv_module.layer_norm''',
'''ffn1.w_1''': '''encoder.layers.*.ffn1.intermediate_dense''',
'''ffn1.w_2''': '''encoder.layers.*.ffn1.output_dense''',
'''ffn1.layer_norm''': '''encoder.layers.*.ffn1_layer_norm''',
'''ffn2.w_1''': '''encoder.layers.*.ffn2.intermediate_dense''',
'''ffn2.w_2''': '''encoder.layers.*.ffn2.output_dense''',
'''ffn2.layer_norm''': '''encoder.layers.*.ffn2_layer_norm''',
'''final_layer_norm''': '''encoder.layers.*.final_layer_norm''',
'''encoder.layer_norm''': '''encoder.layer_norm''',
'''w2v_model.layer_norm''': '''feature_projection.layer_norm''',
'''quantizer.weight_proj''': '''quantizer.weight_proj''',
'''quantizer.vars''': '''quantizer.codevectors''',
'''project_q''': '''project_q''',
'''final_proj''': '''project_hid''',
'''w2v_encoder.proj''': '''lm_head''',
'''mask_emb''': '''masked_spec_embed''',
}
_lowerCAmelCase = [
'''lm_head''',
'''quantizer.weight_proj''',
'''quantizer.codevectors''',
'''project_q''',
'''project_hid''',
]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for attribute in key.split(""".""" ):
lowerCAmelCase__ : Optional[Any] = getattr(UpperCamelCase , UpperCamelCase )
if weight_type is not None:
lowerCAmelCase__ : int = getattr(UpperCamelCase , UpperCamelCase ).shape
else:
lowerCAmelCase__ : int = 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":
lowerCAmelCase__ : Union[str, Any] = value
elif weight_type == "weight_g":
lowerCAmelCase__ : Any = value
elif weight_type == "weight_v":
lowerCAmelCase__ : Optional[int] = value
elif weight_type == "bias":
lowerCAmelCase__ : List[str] = value
elif weight_type == "running_mean":
lowerCAmelCase__ : List[str] = value
elif weight_type == "running_var":
lowerCAmelCase__ : int = value
elif weight_type == "num_batches_tracked":
lowerCAmelCase__ : Tuple = value
elif weight_type == "inv_freq":
lowerCAmelCase__ : str = value
else:
lowerCAmelCase__ : str = value
logger.info(f"""{key + '.' + weight_type if weight_type is not None else ''} was initialized from {full_name}.""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : List[str] = fairseq_model.state_dict()
lowerCAmelCase__ : Any = hf_model.wavaveca_conformer.feature_extractor
for name, value in fairseq_dict.items():
lowerCAmelCase__ : int = False
if "conv_layers" in name:
load_conv_layer(
UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , hf_model.config.feat_extract_norm == """group""" , )
lowerCAmelCase__ : Optional[int] = True
else:
for key, mapped_key in MAPPING.items():
lowerCAmelCase__ : List[Any] = """wav2vec2_conformer.""" + 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]:
lowerCAmelCase__ : List[Any] = True
if "*" in mapped_key:
lowerCAmelCase__ : str = name.split(UpperCamelCase )[0].split(""".""" )[-2]
lowerCAmelCase__ : List[str] = mapped_key.replace("""*""" , UpperCamelCase )
if "pos_bias_u" in name:
lowerCAmelCase__ : List[str] = None
elif "pos_bias_v" in name:
lowerCAmelCase__ : List[str] = None
elif "weight_g" in name:
lowerCAmelCase__ : Optional[int] = """weight_g"""
elif "weight_v" in name:
lowerCAmelCase__ : Union[str, Any] = """weight_v"""
elif "bias" in name:
lowerCAmelCase__ : Union[str, Any] = """bias"""
elif "weight" in name:
# TODO: don't match quantizer.weight_proj
lowerCAmelCase__ : Union[str, Any] = """weight"""
elif "running_mean" in name:
lowerCAmelCase__ : List[Any] = """running_mean"""
elif "inv_freq" in name:
lowerCAmelCase__ : int = """inv_freq"""
elif "running_var" in name:
lowerCAmelCase__ : Any = """running_var"""
elif "num_batches_tracked" in name:
lowerCAmelCase__ : Union[str, Any] = """num_batches_tracked"""
else:
lowerCAmelCase__ : str = None
set_recursively(UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase )
continue
if not is_used:
unused_weights.append(UpperCamelCase )
logger.warning(f"""Unused weights: {unused_weights}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = full_name.split("""conv_layers.""" )[-1]
lowerCAmelCase__ : Any = name.split(""".""" )
lowerCAmelCase__ : Optional[int] = int(items[0] )
lowerCAmelCase__ : Tuple = 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.""" )
lowerCAmelCase__ : List[str] = 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.""" )
lowerCAmelCase__ : Any = 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.conv_layers[layer_id].layer_norm.bias.data.shape} was found.""" )
lowerCAmelCase__ : int = 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.conv_layers[layer_id].layer_norm.weight.data.shape} was found.""" )
lowerCAmelCase__ : Union[str, Any] = value
logger.info(f"""Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.""" )
else:
unused_weights.append(UpperCamelCase )
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=True ):
"""simple docstring"""
if config_path is not None:
lowerCAmelCase__ : Tuple = WavaVecaConformerConfig.from_pretrained(UpperCamelCase , hidden_act="""swish""" )
else:
lowerCAmelCase__ : Union[str, Any] = WavaVecaConformerConfig()
if "rope" in checkpoint_path:
lowerCAmelCase__ : Tuple = """rotary"""
if is_finetuned:
if dict_path:
lowerCAmelCase__ : int = Dictionary.load(UpperCamelCase )
# important change bos & pad token id since CTC symbol is <pad> and
# not <s> as in fairseq
lowerCAmelCase__ : Tuple = target_dict.pad_index
lowerCAmelCase__ : Union[str, Any] = target_dict.bos_index
lowerCAmelCase__ : List[str] = target_dict.eos_index
lowerCAmelCase__ : Dict = len(target_dict.symbols )
lowerCAmelCase__ : Tuple = os.path.join(UpperCamelCase , """vocab.json""" )
if not os.path.isdir(UpperCamelCase ):
logger.error("""--pytorch_dump_folder_path ({}) should be a directory""".format(UpperCamelCase ) )
return
os.makedirs(UpperCamelCase , exist_ok=UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = target_dict.indices
# fairseq has the <pad> and <s> switched
lowerCAmelCase__ : Optional[int] = 0
lowerCAmelCase__ : int = 1
with open(UpperCamelCase , """w""" , encoding="""utf-8""" ) as vocab_handle:
json.dump(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = WavaVecaCTCTokenizer(
UpperCamelCase , unk_token=target_dict.unk_word , pad_token=target_dict.pad_word , bos_token=target_dict.bos_word , eos_token=target_dict.eos_word , word_delimiter_token="""|""" , do_lower_case=UpperCamelCase , )
lowerCAmelCase__ : int = True if config.feat_extract_norm == """layer""" else False
lowerCAmelCase__ : Union[str, Any] = WavaVecaFeatureExtractor(
feature_size=1 , sampling_rate=16000 , padding_value=0 , do_normalize=UpperCamelCase , return_attention_mask=UpperCamelCase , )
lowerCAmelCase__ : List[Any] = WavaVecaProcessor(feature_extractor=UpperCamelCase , tokenizer=UpperCamelCase )
processor.save_pretrained(UpperCamelCase )
lowerCAmelCase__ : List[str] = WavaVecaConformerForCTC(UpperCamelCase )
else:
lowerCAmelCase__ : Tuple = WavaVecaConformerForPreTraining(UpperCamelCase )
if is_finetuned:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = fairseq.checkpoint_utils.load_model_ensemble_and_task(
[checkpoint_path] , arg_overrides={"""data""": """/""".join(dict_path.split("""/""" )[:-1] )} )
else:
lowerCAmelCase__ : List[Any] = argparse.Namespace(task="""audio_pretraining""" )
lowerCAmelCase__ : List[str] = fairseq.tasks.setup_task(UpperCamelCase )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : str = fairseq.checkpoint_utils.load_model_ensemble_and_task([checkpoint_path] , task=UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = model[0].eval()
recursively_load_weights(UpperCamelCase , UpperCamelCase , not is_finetuned )
hf_wavavec.save_pretrained(UpperCamelCase )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument('''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model.''')
parser.add_argument('''--checkpoint_path''', default=None, type=str, help='''Path to fairseq checkpoint''')
parser.add_argument('''--dict_path''', default=None, type=str, help='''Path to dict of fine-tuned model''')
parser.add_argument('''--config_path''', default=None, type=str, help='''Path to hf config.json of model to convert''')
parser.add_argument(
'''--not_finetuned''', action='''store_true''', help='''Whether the model to convert is a fine-tuned model or not'''
)
_lowerCAmelCase = parser.parse_args()
convert_wavaveca_conformer_checkpoint(
args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path, args.dict_path, not args.not_finetuned
)
| 37
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''microsoft/focalnet-tiny''': '''https://huggingface.co/microsoft/focalnet-tiny/resolve/main/config.json''',
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[Any] = '''focalnet'''
def __init__( self ,__UpperCAmelCase=224 ,__UpperCAmelCase=4 ,__UpperCAmelCase=3 ,__UpperCAmelCase=96 ,__UpperCAmelCase=False ,__UpperCAmelCase=[192, 384, 768, 768] ,__UpperCAmelCase=[2, 2, 6, 2] ,__UpperCAmelCase=[2, 2, 2, 2] ,__UpperCAmelCase=[3, 3, 3, 3] ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=4.0 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=False ,__UpperCAmelCase=1E-4 ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-5 ,__UpperCAmelCase=32 ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Optional[Any]:
super().__init__(**__UpperCAmelCase )
lowerCAmelCase__ : Dict = image_size
lowerCAmelCase__ : int = patch_size
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : Dict = embed_dim
lowerCAmelCase__ : List[str] = use_conv_embed
lowerCAmelCase__ : List[Any] = hidden_sizes
lowerCAmelCase__ : Dict = depths
lowerCAmelCase__ : List[str] = focal_levels
lowerCAmelCase__ : List[str] = focal_windows
lowerCAmelCase__ : Dict = hidden_act
lowerCAmelCase__ : Dict = mlp_ratio
lowerCAmelCase__ : Tuple = hidden_dropout_prob
lowerCAmelCase__ : Tuple = drop_path_rate
lowerCAmelCase__ : Dict = use_layerscale
lowerCAmelCase__ : Optional[Any] = layerscale_value
lowerCAmelCase__ : str = use_post_layernorm
lowerCAmelCase__ : Union[str, Any] = use_post_layernorm_in_modulation
lowerCAmelCase__ : int = normalize_modulator
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : List[str] = layer_norm_eps
lowerCAmelCase__ : List[Any] = encoder_stride
lowerCAmelCase__ : Dict = ["""stem"""] + [F"""stage{idx}""" for idx in range(1 ,len(self.depths ) + 1 )]
lowerCAmelCase__ , lowerCAmelCase__ : Any = get_aligned_output_features_output_indices(
out_features=__UpperCAmelCase ,out_indices=__UpperCAmelCase ,stage_names=self.stage_names )
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = 1000 ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : Dict = 1, 1
lowerCAmelCase__ : int = []
for i in range(1 , n + 1 ):
lowerCAmelCase__ : List[str] = prev_numerator + 2 * prev_denominator
lowerCAmelCase__ : List[Any] = prev_numerator + prev_denominator
if len(str(UpperCamelCase ) ) > len(str(UpperCamelCase ) ):
result.append(UpperCamelCase )
lowerCAmelCase__ : str = numerator
lowerCAmelCase__ : Union[str, Any] = denominator
return len(UpperCamelCase )
if __name__ == "__main__":
print(F"""{solution() = }""")
| 37
|
'''simple docstring'''
import argparse
import os.path as osp
import re
import torch
from safetensors.torch import load_file, save_file
# =================#
# UNet Conversion #
# =================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''time_embed.0.weight''', '''time_embedding.linear_1.weight'''),
('''time_embed.0.bias''', '''time_embedding.linear_1.bias'''),
('''time_embed.2.weight''', '''time_embedding.linear_2.weight'''),
('''time_embed.2.bias''', '''time_embedding.linear_2.bias'''),
('''input_blocks.0.0.weight''', '''conv_in.weight'''),
('''input_blocks.0.0.bias''', '''conv_in.bias'''),
('''out.0.weight''', '''conv_norm_out.weight'''),
('''out.0.bias''', '''conv_norm_out.bias'''),
('''out.2.weight''', '''conv_out.weight'''),
('''out.2.bias''', '''conv_out.bias'''),
]
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''in_layers.0''', '''norm1'''),
('''in_layers.2''', '''conv1'''),
('''out_layers.0''', '''norm2'''),
('''out_layers.3''', '''conv2'''),
('''emb_layers.1''', '''time_emb_proj'''),
('''skip_connection''', '''conv_shortcut'''),
]
_lowerCAmelCase = []
# hardcoded number of downblocks and resnets/attentions...
# would need smarter logic for other networks.
for i in range(4):
# loop over downblocks/upblocks
for j in range(2):
# loop over resnets/attentions for downblocks
_lowerCAmelCase = F"""down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.0."""
unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
if i < 3:
# no attention layers in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.1."""
unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
for j in range(3):
# loop over resnets/attentions for upblocks
_lowerCAmelCase = F"""up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.0."""
unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
if i > 0:
# no attention layers in up_blocks.0
_lowerCAmelCase = F"""up_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.1."""
unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
if i < 3:
# no downsample in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0.conv."""
_lowerCAmelCase = F"""input_blocks.{3*(i+1)}.0.op."""
unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
# no upsample in up_blocks.3
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""output_blocks.{3*i + 2}.{1 if i == 0 else 2}."""
unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
_lowerCAmelCase = '''mid_block.attentions.0.'''
_lowerCAmelCase = '''middle_block.1.'''
unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
for j in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{j}."""
_lowerCAmelCase = F"""middle_block.{2*j}."""
unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = {k: k for k in unet_state_dict.keys()}
for sd_name, hf_name in unet_conversion_map:
lowerCAmelCase__ : Optional[int] = sd_name
for k, v in mapping.items():
if "resnets" in k:
for sd_part, hf_part in unet_conversion_map_resnet:
lowerCAmelCase__ : Any = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
for k, v in mapping.items():
for sd_part, hf_part in unet_conversion_map_layer:
lowerCAmelCase__ : List[Any] = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
lowerCAmelCase__ : Tuple = {v: unet_state_dict[k] for k, v in mapping.items()}
return new_state_dict
# ================#
# VAE Conversion #
# ================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''nin_shortcut''', '''conv_shortcut'''),
('''norm_out''', '''conv_norm_out'''),
('''mid.attn_1.''', '''mid_block.attentions.0.'''),
]
for i in range(4):
# down_blocks have two resnets
for j in range(2):
_lowerCAmelCase = F"""encoder.down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""encoder.down.{i}.block.{j}."""
vae_conversion_map.append((sd_down_prefix, hf_down_prefix))
if i < 3:
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0."""
_lowerCAmelCase = F"""down.{i}.downsample."""
vae_conversion_map.append((sd_downsample_prefix, hf_downsample_prefix))
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""up.{3-i}.upsample."""
vae_conversion_map.append((sd_upsample_prefix, hf_upsample_prefix))
# up_blocks have three resnets
# also, up blocks in hf are numbered in reverse from sd
for j in range(3):
_lowerCAmelCase = F"""decoder.up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""decoder.up.{3-i}.block.{j}."""
vae_conversion_map.append((sd_up_prefix, hf_up_prefix))
# this part accounts for mid blocks in both the encoder and the decoder
for i in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{i}."""
_lowerCAmelCase = F"""mid.block_{i+1}."""
vae_conversion_map.append((sd_mid_res_prefix, hf_mid_res_prefix))
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''norm.''', '''group_norm.'''),
('''q.''', '''query.'''),
('''k.''', '''key.'''),
('''v.''', '''value.'''),
('''proj_out.''', '''proj_attn.'''),
]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return w.reshape(*w.shape , 1 , 1 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = {k: k for k in vae_state_dict.keys()}
for k, v in mapping.items():
for sd_part, hf_part in vae_conversion_map:
lowerCAmelCase__ : str = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
for k, v in mapping.items():
if "attentions" in k:
for sd_part, hf_part in vae_conversion_map_attn:
lowerCAmelCase__ : Dict = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
lowerCAmelCase__ : Union[str, Any] = {v: vae_state_dict[k] for k, v in mapping.items()}
lowerCAmelCase__ : Tuple = ["""q""", """k""", """v""", """proj_out"""]
for k, v in new_state_dict.items():
for weight_name in weights_to_convert:
if f"""mid.attn_1.{weight_name}.weight""" in k:
print(f"""Reshaping {k} for SD format""" )
lowerCAmelCase__ : Optional[int] = reshape_weight_for_sd(UpperCamelCase )
return new_state_dict
# =========================#
# Text Encoder Conversion #
# =========================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''resblocks.''', '''text_model.encoder.layers.'''),
('''ln_1''', '''layer_norm1'''),
('''ln_2''', '''layer_norm2'''),
('''.c_fc.''', '''.fc1.'''),
('''.c_proj.''', '''.fc2.'''),
('''.attn''', '''.self_attn'''),
('''ln_final.''', '''transformer.text_model.final_layer_norm.'''),
('''token_embedding.weight''', '''transformer.text_model.embeddings.token_embedding.weight'''),
('''positional_embedding''', '''transformer.text_model.embeddings.position_embedding.weight'''),
]
_lowerCAmelCase = {re.escape(x[1]): x[0] for x in textenc_conversion_lst}
_lowerCAmelCase = re.compile('''|'''.join(protected.keys()))
# Ordering is from https://github.com/pytorch/pytorch/blob/master/test/cpp/api/modules.cpp
_lowerCAmelCase = {'''q''': 0, '''k''': 1, '''v''': 2}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = {}
lowerCAmelCase__ : int = {}
lowerCAmelCase__ : List[Any] = {}
for k, v in text_enc_dict.items():
if (
k.endswith(""".self_attn.q_proj.weight""" )
or k.endswith(""".self_attn.k_proj.weight""" )
or k.endswith(""".self_attn.v_proj.weight""" )
):
lowerCAmelCase__ : Optional[int] = k[: -len(""".q_proj.weight""" )]
lowerCAmelCase__ : Tuple = k[-len("""q_proj.weight""" )]
if k_pre not in capture_qkv_weight:
lowerCAmelCase__ : List[Any] = [None, None, None]
lowerCAmelCase__ : Dict = v
continue
if (
k.endswith(""".self_attn.q_proj.bias""" )
or k.endswith(""".self_attn.k_proj.bias""" )
or k.endswith(""".self_attn.v_proj.bias""" )
):
lowerCAmelCase__ : str = k[: -len(""".q_proj.bias""" )]
lowerCAmelCase__ : List[str] = k[-len("""q_proj.bias""" )]
if k_pre not in capture_qkv_bias:
lowerCAmelCase__ : Union[str, Any] = [None, None, None]
lowerCAmelCase__ : Any = v
continue
lowerCAmelCase__ : Dict = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = v
for k_pre, tensors in capture_qkv_weight.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : Any = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Tuple = torch.cat(UpperCamelCase )
for k_pre, tensors in capture_qkv_bias.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : str = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : List[Any] = torch.cat(UpperCamelCase )
return new_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return text_enc_dict
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument('''--model_path''', default=None, type=str, required=True, help='''Path to the model to convert.''')
parser.add_argument('''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the output model.''')
parser.add_argument('''--half''', action='''store_true''', help='''Save weights in half precision.''')
parser.add_argument(
'''--use_safetensors''', action='''store_true''', help='''Save weights use safetensors, default is ckpt.'''
)
_lowerCAmelCase = parser.parse_args()
assert args.model_path is not None, "Must provide a model path!"
assert args.checkpoint_path is not None, "Must provide a checkpoint path!"
# Path for safetensors
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''model.safetensors''')
# Load models from safetensors if it exists, if it doesn't pytorch
if osp.exists(unet_path):
_lowerCAmelCase = load_file(unet_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(unet_path, map_location='''cpu''')
if osp.exists(vae_path):
_lowerCAmelCase = load_file(vae_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(vae_path, map_location='''cpu''')
if osp.exists(text_enc_path):
_lowerCAmelCase = load_file(text_enc_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''pytorch_model.bin''')
_lowerCAmelCase = torch.load(text_enc_path, map_location='''cpu''')
# Convert the UNet model
_lowerCAmelCase = convert_unet_state_dict(unet_state_dict)
_lowerCAmelCase = {'''model.diffusion_model.''' + k: v for k, v in unet_state_dict.items()}
# Convert the VAE model
_lowerCAmelCase = convert_vae_state_dict(vae_state_dict)
_lowerCAmelCase = {'''first_stage_model.''' + k: v for k, v in vae_state_dict.items()}
# Easiest way to identify v2.0 model seems to be that the text encoder (OpenCLIP) is deeper
_lowerCAmelCase = '''text_model.encoder.layers.22.layer_norm2.bias''' in text_enc_dict
if is_vaa_model:
# Need to add the tag 'transformer' in advance so we can knock it out from the final layer-norm
_lowerCAmelCase = {'''transformer.''' + k: v for k, v in text_enc_dict.items()}
_lowerCAmelCase = convert_text_enc_state_dict_vaa(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.model.''' + k: v for k, v in text_enc_dict.items()}
else:
_lowerCAmelCase = convert_text_enc_state_dict(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.transformer.''' + k: v for k, v in text_enc_dict.items()}
# Put together new checkpoint
_lowerCAmelCase = {**unet_state_dict, **vae_state_dict, **text_enc_dict}
if args.half:
_lowerCAmelCase = {k: v.half() for k, v in state_dict.items()}
if args.use_safetensors:
save_file(state_dict, args.checkpoint_path)
else:
_lowerCAmelCase = {'''state_dict''': state_dict}
torch.save(state_dict, args.checkpoint_path)
| 37
| 1
|
'''simple docstring'''
from collections.abc import Sequence
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase = False ):
"""simple docstring"""
if not arr:
return 0
lowerCAmelCase__ : Optional[Any] = 0 if allow_empty_subarrays else float("""-inf""" )
lowerCAmelCase__ : Any = 0.0
for num in arr:
lowerCAmelCase__ : Tuple = max(0 if allow_empty_subarrays else num , curr_sum + num )
lowerCAmelCase__ : Optional[Any] = max(UpperCamelCase , UpperCamelCase )
return max_sum
if __name__ == "__main__":
from doctest import testmod
testmod()
_lowerCAmelCase = [-2, 1, -3, 4, -1, 2, 1, -5, 4]
print(F"""{max_subarray_sum(nums) = }""")
| 37
|
'''simple docstring'''
import os
from bleurt import score # From: git+https://github.com/google-research/bleurt.git
import datasets
_lowerCAmelCase = datasets.logging.get_logger(__name__)
_lowerCAmelCase = '''\
@inproceedings{bleurt,
title={BLEURT: Learning Robust Metrics for Text Generation},
author={Thibault Sellam and Dipanjan Das and Ankur P. Parikh},
booktitle={ACL},
year={2020},
url={https://arxiv.org/abs/2004.04696}
}
'''
_lowerCAmelCase = '''\
BLEURT a learnt evaluation metric for Natural Language Generation. It is built using multiple phases of transfer learning starting from a pretrained BERT model (Devlin et al. 2018)
and then employing another pre-training phrase using synthetic data. Finally it is trained on WMT human annotations. You may run BLEURT out-of-the-box or fine-tune
it for your specific application (the latter is expected to perform better).
See the project\'s README at https://github.com/google-research/bleurt#readme for more information.
'''
_lowerCAmelCase = '''
BLEURT score.
Args:
`predictions` (list of str): prediction/candidate sentences
`references` (list of str): reference sentences
`checkpoint` BLEURT checkpoint. Will default to BLEURT-tiny if None.
Returns:
\'scores\': List of scores.
Examples:
>>> predictions = ["hello there", "general kenobi"]
>>> references = ["hello there", "general kenobi"]
>>> bleurt = datasets.load_metric("bleurt")
>>> results = bleurt.compute(predictions=predictions, references=references)
>>> print([round(v, 2) for v in results["scores"]])
[1.03, 1.04]
'''
_lowerCAmelCase = {
'''bleurt-tiny-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-128.zip''',
'''bleurt-tiny-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-512.zip''',
'''bleurt-base-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-128.zip''',
'''bleurt-base-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-512.zip''',
'''bleurt-large-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-128.zip''',
'''bleurt-large-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-512.zip''',
'''BLEURT-20-D3''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D3.zip''',
'''BLEURT-20-D6''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D6.zip''',
'''BLEURT-20-D12''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D12.zip''',
'''BLEURT-20''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20.zip''',
}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,homepage="""https://github.com/google-research/bleurt""" ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(
{
"""predictions""": datasets.Value("""string""" ,id="""sequence""" ),
"""references""": datasets.Value("""string""" ,id="""sequence""" ),
} ) ,codebase_urls=["""https://github.com/google-research/bleurt"""] ,reference_urls=["""https://github.com/google-research/bleurt""", """https://arxiv.org/abs/2004.04696"""] ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
# check that config name specifies a valid BLEURT model
if self.config_name == "default":
logger.warning(
"""Using default BLEURT-Base checkpoint for sequence maximum length 128. """
"""You can use a bigger model for better results with e.g.: datasets.load_metric('bleurt', 'bleurt-large-512').""" )
lowerCAmelCase__ : str = """bleurt-base-128"""
if self.config_name.lower() in CHECKPOINT_URLS:
lowerCAmelCase__ : Union[str, Any] = self.config_name.lower()
elif self.config_name.upper() in CHECKPOINT_URLS:
lowerCAmelCase__ : List[Any] = self.config_name.upper()
else:
raise KeyError(
F"""{self.config_name} model not found. You should supply the name of a model checkpoint for bleurt in {CHECKPOINT_URLS.keys()}""" )
# download the model checkpoint specified by self.config_name and set up the scorer
lowerCAmelCase__ : int = dl_manager.download_and_extract(CHECKPOINT_URLS[checkpoint_name] )
lowerCAmelCase__ : Dict = score.BleurtScorer(os.path.join(__UpperCAmelCase ,__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Union[str, Any] = self.scorer.score(references=__UpperCAmelCase ,candidates=__UpperCAmelCase )
return {"scores": scores}
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if bit_count < 0:
raise ValueError("""The given input must be positive""" )
# get the generated string sequence
lowerCAmelCase__ : List[str] = gray_code_sequence_string(UpperCamelCase )
#
# convert them to integers
for i in range(len(UpperCamelCase ) ):
lowerCAmelCase__ : int = int(sequence[i] , 2 )
return sequence
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if bit_count == 0:
return ["0"]
if bit_count == 1:
return ["0", "1"]
lowerCAmelCase__ : Optional[Any] = 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
lowerCAmelCase__ : Union[str, Any] = gray_code_sequence_string(bit_count - 1 )
lowerCAmelCase__ : List[Any] = []
# append 0 to first half of the smaller sequence generated
for i in range(seq_len // 2 ):
lowerCAmelCase__ : Optional[Any] = """0""" + smaller_sequence[i]
sequence.append(UpperCamelCase )
# append 1 to second half ... start from the end of the list
for i in reversed(range(seq_len // 2 ) ):
lowerCAmelCase__ : int = """1""" + smaller_sequence[i]
sequence.append(UpperCamelCase )
return sequence
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
import uuid
from typing import Any, Dict, List, Optional, Union
from ..utils import add_end_docstrings, is_tf_available, is_torch_available, logging
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_tf_available():
import tensorflow as tf
if is_torch_available():
import torch
_lowerCAmelCase = logging.get_logger(__name__)
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ) -> str:
if not conversation_id:
lowerCAmelCase__ : List[str] = uuid.uuida()
if past_user_inputs is None:
lowerCAmelCase__ : List[Any] = []
if generated_responses is None:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : uuid.UUID = conversation_id
lowerCAmelCase__ : List[str] = past_user_inputs
lowerCAmelCase__ : List[str] = generated_responses
lowerCAmelCase__ : Optional[str] = text
def __eq__( self ,__UpperCAmelCase ) -> Dict:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
return False
if self.uuid == other.uuid:
return True
return (
self.new_user_input == other.new_user_input
and self.past_user_inputs == other.past_user_inputs
and self.generated_responses == other.generated_responses
)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Optional[Any]:
if self.new_user_input:
if overwrite:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" was overwritten """
F"""with: \"{text}\".""" )
lowerCAmelCase__ : Optional[int] = text
else:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" new input """
F"""ignored: \"{text}\". Set `overwrite` to True to overwrite unprocessed user input""" )
else:
lowerCAmelCase__ : Optional[Any] = text
def UpperCAmelCase_ ( self ) -> List[Any]:
if self.new_user_input:
self.past_user_inputs.append(self.new_user_input )
lowerCAmelCase__ : Union[str, Any] = None
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
self.generated_responses.append(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
for user_input, generated_response in zip(self.past_user_inputs ,self.generated_responses ):
yield True, user_input
yield False, generated_response
if self.new_user_input:
yield True, self.new_user_input
def __repr__( self ) -> Tuple:
lowerCAmelCase__ : Tuple = F"""Conversation id: {self.uuid} \n"""
for is_user, text in self.iter_texts():
lowerCAmelCase__ : Any = """user""" if is_user else """bot"""
output += F"""{name} >> {text} \n"""
return output
@add_end_docstrings(
SCREAMING_SNAKE_CASE_ , R'''
min_length_for_response (`int`, *optional*, defaults to 32):
The minimum length (in number of tokens) for a response.
minimum_tokens (`int`, *optional*, defaults to 10):
The minimum length of tokens to leave for a response.
''' , )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Tuple:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
if self.tokenizer.pad_token_id is None:
lowerCAmelCase__ : Tuple = self.tokenizer.eos_token
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : List[Any] = {}
lowerCAmelCase__ : Optional[int] = {}
lowerCAmelCase__ : List[str] = {}
if min_length_for_response is not None:
lowerCAmelCase__ : Any = min_length_for_response
if minimum_tokens is not None:
lowerCAmelCase__ : Optional[int] = minimum_tokens
if "max_length" in generate_kwargs:
lowerCAmelCase__ : Optional[Any] = generate_kwargs["""max_length"""]
# self.max_length = generate_kwargs.get("max_length", self.model.config.max_length)
if clean_up_tokenization_spaces is not None:
lowerCAmelCase__ : int = clean_up_tokenization_spaces
if generate_kwargs:
forward_params.update(__UpperCAmelCase )
return preprocess_params, forward_params, postprocess_params
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase=0 ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Optional[int] = super().__call__(__UpperCAmelCase ,num_workers=__UpperCAmelCase ,**__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) and len(__UpperCAmelCase ) == 1:
return outputs[0]
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=32 ) -> Dict[str, Any]:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
raise ValueError("""ConversationalPipeline, expects Conversation as inputs""" )
if conversation.new_user_input is None:
raise ValueError(
F"""Conversation with UUID {type(conversation.uuid )} does not contain new user input to process. """
"""Add user inputs with the conversation's `add_user_input` method""" )
if hasattr(self.tokenizer ,"""_build_conversation_input_ids""" ):
lowerCAmelCase__ : str = self.tokenizer._build_conversation_input_ids(__UpperCAmelCase )
else:
# If the tokenizer cannot handle conversations, we default to only the old version
lowerCAmelCase__ : List[Any] = self._legacy_parse_and_tokenize(__UpperCAmelCase )
if self.framework == "pt":
lowerCAmelCase__ : List[Any] = torch.LongTensor([input_ids] )
elif self.framework == "tf":
lowerCAmelCase__ : Dict = tf.constant([input_ids] )
return {"input_ids": input_ids, "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=10 ,**__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[Any] = generate_kwargs.get("""max_length""" ,self.model.config.max_length )
lowerCAmelCase__ : Optional[Any] = model_inputs["""input_ids"""].shape[1]
if max_length - minimum_tokens < n:
logger.warning(F"""Conversation input is to long ({n}), trimming it to ({max_length} - {minimum_tokens})""" )
lowerCAmelCase__ : str = max_length - minimum_tokens
lowerCAmelCase__ : Union[str, Any] = model_inputs["""input_ids"""][:, -trim:]
if "attention_mask" in model_inputs:
lowerCAmelCase__ : Tuple = model_inputs["""attention_mask"""][:, -trim:]
lowerCAmelCase__ : str = model_inputs.pop("""conversation""" )
lowerCAmelCase__ : Union[str, Any] = max_length
lowerCAmelCase__ : Any = self.model.generate(**__UpperCAmelCase ,**__UpperCAmelCase )
if self.model.config.is_encoder_decoder:
lowerCAmelCase__ : int = 1
else:
lowerCAmelCase__ : int = n
return {"output_ids": output_ids[:, start_position:], "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=True ) -> List[str]:
lowerCAmelCase__ : Optional[int] = model_outputs["""output_ids"""]
lowerCAmelCase__ : Tuple = self.tokenizer.decode(
output_ids[0] ,skip_special_tokens=__UpperCAmelCase ,clean_up_tokenization_spaces=__UpperCAmelCase ,)
lowerCAmelCase__ : Union[str, Any] = model_outputs["""conversation"""]
conversation.mark_processed()
conversation.append_response(__UpperCAmelCase )
return conversation
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Dict = self.tokenizer.eos_token_id
lowerCAmelCase__ : int = []
for is_user, text in conversation.iter_texts():
if eos_token_id is not None:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) + [eos_token_id] )
else:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) )
if len(__UpperCAmelCase ) > self.tokenizer.model_max_length:
lowerCAmelCase__ : Optional[Any] = input_ids[-self.tokenizer.model_max_length :]
return input_ids
| 37
| 1
|
'''simple docstring'''
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils import AddedToken
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_barthez import BarthezTokenizer
else:
_lowerCAmelCase = None
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''sentencepiece.bpe.model''', '''tokenizer_file''': '''tokenizer.json'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''moussaKam/mbarthez''': '''https://huggingface.co/moussaKam/mbarthez/resolve/main/sentencepiece.bpe.model''',
'''moussaKam/barthez''': '''https://huggingface.co/moussaKam/barthez/resolve/main/sentencepiece.bpe.model''',
'''moussaKam/barthez-orangesum-title''': (
'''https://huggingface.co/moussaKam/barthez-orangesum-title/resolve/main/sentencepiece.bpe.model'''
),
},
'''tokenizer_file''': {
'''moussaKam/mbarthez''': '''https://huggingface.co/moussaKam/mbarthez/resolve/main/tokenizer.json''',
'''moussaKam/barthez''': '''https://huggingface.co/moussaKam/barthez/resolve/main/tokenizer.json''',
'''moussaKam/barthez-orangesum-title''': (
'''https://huggingface.co/moussaKam/barthez-orangesum-title/resolve/main/tokenizer.json'''
),
},
}
_lowerCAmelCase = {
'''moussaKam/mbarthez''': 1024,
'''moussaKam/barthez''': 1024,
'''moussaKam/barthez-orangesum-title''': 1024,
}
_lowerCAmelCase = '''▁'''
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Tuple = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Tuple = ['''input_ids''', '''attention_mask''']
__lowercase : List[str] = BarthezTokenizer
def __init__( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase="<s>" ,__UpperCAmelCase="</s>" ,__UpperCAmelCase="</s>" ,__UpperCAmelCase="<s>" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="<mask>" ,**__UpperCAmelCase ,) -> Dict:
# Mask token behave like a normal word, i.e. include the space before it
lowerCAmelCase__ : Union[str, Any] = 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 ,unk_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,**__UpperCAmelCase ,)
lowerCAmelCase__ : List[Any] = vocab_file
lowerCAmelCase__ : Union[str, Any] = False if not self.vocab_file else True
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
lowerCAmelCase__ : Tuple = [self.cls_token_id]
lowerCAmelCase__ : Union[str, Any] = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : Tuple = [self.sep_token_id]
lowerCAmelCase__ : str = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep + sep + token_ids_a + sep ) * [0]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
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
lowerCAmelCase__ : Tuple = 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,)
| 37
|
'''simple docstring'''
from collections import UserDict
from typing import List, Union
from ..utils import (
add_end_docstrings,
is_tf_available,
is_torch_available,
is_vision_available,
logging,
requires_backends,
)
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
from ..tf_utils import stable_softmax
_lowerCAmelCase = logging.get_logger(__name__)
@add_end_docstrings(SCREAMING_SNAKE_CASE_ )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,**__UpperCAmelCase ) -> Tuple:
super().__init__(**__UpperCAmelCase )
requires_backends(self ,"""vision""" )
self.check_model_type(
TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if self.framework == """tf"""
else MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING )
def __call__( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> str:
return super().__call__(__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[Any] = {}
if "candidate_labels" in kwargs:
lowerCAmelCase__ : int = kwargs["""candidate_labels"""]
if "hypothesis_template" in kwargs:
lowerCAmelCase__ : Optional[int] = kwargs["""hypothesis_template"""]
return preprocess_params, {}, {}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase="This is a photo of {}." ) -> int:
lowerCAmelCase__ : str = load_image(__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.image_processor(images=[image] ,return_tensors=self.framework )
lowerCAmelCase__ : List[Any] = candidate_labels
lowerCAmelCase__ : List[str] = [hypothesis_template.format(__UpperCAmelCase ) for x in candidate_labels]
lowerCAmelCase__ : Optional[Any] = self.tokenizer(__UpperCAmelCase ,return_tensors=self.framework ,padding=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = [text_inputs]
return inputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = model_inputs.pop("""candidate_labels""" )
lowerCAmelCase__ : Union[str, Any] = model_inputs.pop("""text_inputs""" )
if isinstance(text_inputs[0] ,__UpperCAmelCase ):
lowerCAmelCase__ : int = text_inputs[0]
else:
# Batching case.
lowerCAmelCase__ : Dict = text_inputs[0][0]
lowerCAmelCase__ : Any = self.model(**__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = {
"""candidate_labels""": candidate_labels,
"""logits""": outputs.logits_per_image,
}
return model_outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Union[str, Any] = model_outputs.pop("""candidate_labels""" )
lowerCAmelCase__ : List[str] = model_outputs["""logits"""][0]
if self.framework == "pt":
lowerCAmelCase__ : List[str] = logits.softmax(dim=-1 ).squeeze(-1 )
lowerCAmelCase__ : Optional[Any] = probs.tolist()
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Dict = [scores]
elif self.framework == "tf":
lowerCAmelCase__ : Any = stable_softmax(__UpperCAmelCase ,axis=-1 )
lowerCAmelCase__ : List[Any] = probs.numpy().tolist()
else:
raise ValueError(F"""Unsupported framework: {self.framework}""" )
lowerCAmelCase__ : Tuple = [
{"""score""": score, """label""": candidate_label}
for score, candidate_label in sorted(zip(__UpperCAmelCase ,__UpperCAmelCase ) ,key=lambda __UpperCAmelCase : -x[0] )
]
return result
| 37
| 1
|
'''simple docstring'''
import tempfile
import unittest
import numpy as np
from diffusers import (
DDIMScheduler,
DPMSolverMultistepScheduler,
EulerAncestralDiscreteScheduler,
EulerDiscreteScheduler,
LMSDiscreteScheduler,
OnnxStableDiffusionPipeline,
PNDMScheduler,
)
from diffusers.utils.testing_utils import is_onnx_available, nightly, require_onnxruntime, require_torch_gpu
from ..test_pipelines_onnx_common import OnnxPipelineTesterMixin
if is_onnx_available():
import onnxruntime as ort
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Tuple = '''hf-internal-testing/tiny-random-OnnxStableDiffusionPipeline'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase=0 ) -> int:
lowerCAmelCase__ : Any = np.random.RandomState(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = {
"""prompt""": """A painting of a squirrel eating a burger""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 7.5,
"""output_type""": """numpy""",
}
return inputs
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Optional[int] = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : str = self.get_dummy_inputs()
lowerCAmelCase__ : Optional[Any] = pipe(**__UpperCAmelCase ).images
lowerCAmelCase__ : Any = image[0, -3:, -3:, -1]
assert image.shape == (1, 128, 128, 3)
lowerCAmelCase__ : Optional[int] = np.array([0.6_5_0_7_2, 0.5_8_4_9_2, 0.4_8_2_1_9, 0.5_5_5_2_1, 0.5_3_1_8_0, 0.5_5_9_3_9, 0.5_0_6_9_7, 0.3_9_8_0_0, 0.4_6_4_5_5] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : List[str] = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
lowerCAmelCase__ : int = PNDMScheduler.from_config(pipe.scheduler.config ,skip_prk_steps=__UpperCAmelCase )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = self.get_dummy_inputs()
lowerCAmelCase__ : Optional[int] = pipe(**__UpperCAmelCase ).images
lowerCAmelCase__ : List[str] = image[0, -3:, -3:, -1]
assert image.shape == (1, 128, 128, 3)
lowerCAmelCase__ : str = np.array([0.6_5_8_6_3, 0.5_9_4_2_5, 0.4_9_3_2_6, 0.5_6_3_1_3, 0.5_3_8_7_5, 0.5_6_6_2_7, 0.5_1_0_6_5, 0.3_9_7_7_7, 0.4_6_3_3_0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[str] = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
lowerCAmelCase__ : Optional[Any] = LMSDiscreteScheduler.from_config(pipe.scheduler.config )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.get_dummy_inputs()
lowerCAmelCase__ : Union[str, Any] = pipe(**__UpperCAmelCase ).images
lowerCAmelCase__ : Optional[int] = image[0, -3:, -3:, -1]
assert image.shape == (1, 128, 128, 3)
lowerCAmelCase__ : int = np.array([0.5_3_7_5_5, 0.6_0_7_8_6, 0.4_7_4_0_2, 0.4_9_4_8_8, 0.5_1_8_6_9, 0.4_9_8_1_9, 0.4_7_9_8_5, 0.3_8_9_5_7, 0.4_4_2_7_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Tuple = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
lowerCAmelCase__ : List[Any] = EulerDiscreteScheduler.from_config(pipe.scheduler.config )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = self.get_dummy_inputs()
lowerCAmelCase__ : Union[str, Any] = pipe(**__UpperCAmelCase ).images
lowerCAmelCase__ : str = image[0, -3:, -3:, -1]
assert image.shape == (1, 128, 128, 3)
lowerCAmelCase__ : Dict = np.array([0.5_3_7_5_5, 0.6_0_7_8_6, 0.4_7_4_0_2, 0.4_9_4_8_8, 0.5_1_8_6_9, 0.4_9_8_1_9, 0.4_7_9_8_5, 0.3_8_9_5_7, 0.4_4_2_7_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : str = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
lowerCAmelCase__ : Tuple = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.get_dummy_inputs()
lowerCAmelCase__ : Union[str, Any] = pipe(**__UpperCAmelCase ).images
lowerCAmelCase__ : Dict = image[0, -3:, -3:, -1]
assert image.shape == (1, 128, 128, 3)
lowerCAmelCase__ : Optional[Any] = np.array([0.5_3_8_1_7, 0.6_0_8_1_2, 0.4_7_3_8_4, 0.4_9_5_3_0, 0.5_1_8_9_4, 0.4_9_8_1_4, 0.4_7_9_8_4, 0.3_8_9_5_8, 0.4_4_2_7_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : List[str] = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
lowerCAmelCase__ : Any = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = self.get_dummy_inputs()
lowerCAmelCase__ : Optional[Any] = pipe(**__UpperCAmelCase ).images
lowerCAmelCase__ : Any = image[0, -3:, -3:, -1]
assert image.shape == (1, 128, 128, 3)
lowerCAmelCase__ : str = np.array([0.5_3_8_9_5, 0.6_0_8_0_8, 0.4_7_9_3_3, 0.4_9_6_0_8, 0.5_1_8_8_6, 0.4_9_9_5_0, 0.4_8_0_5_3, 0.3_8_9_5_7, 0.4_4_2_0_0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Tuple = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.get_dummy_inputs()
lowerCAmelCase__ : str = 3 * [inputs["""prompt"""]]
# forward
lowerCAmelCase__ : Union[str, Any] = pipe(**__UpperCAmelCase )
lowerCAmelCase__ : Dict = output.images[0, -3:, -3:, -1]
lowerCAmelCase__ : Union[str, Any] = self.get_dummy_inputs()
lowerCAmelCase__ : Dict = 3 * [inputs.pop("""prompt""" )]
lowerCAmelCase__ : str = pipe.tokenizer(
__UpperCAmelCase ,padding="""max_length""" ,max_length=pipe.tokenizer.model_max_length ,truncation=__UpperCAmelCase ,return_tensors="""np""" ,)
lowerCAmelCase__ : Union[str, Any] = text_inputs["""input_ids"""]
lowerCAmelCase__ : Any = pipe.text_encoder(input_ids=text_inputs.astype(np.intaa ) )[0]
lowerCAmelCase__ : Optional[int] = prompt_embeds
# forward
lowerCAmelCase__ : Optional[Any] = pipe(**__UpperCAmelCase )
lowerCAmelCase__ : int = output.images[0, -3:, -3:, -1]
assert np.abs(image_slice_a.flatten() - image_slice_a.flatten() ).max() < 1E-4
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : List[Any] = OnnxStableDiffusionPipeline.from_pretrained(self.hub_checkpoint ,provider="""CPUExecutionProvider""" )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.get_dummy_inputs()
lowerCAmelCase__ : Dict = 3 * ["""this is a negative prompt"""]
lowerCAmelCase__ : List[Any] = negative_prompt
lowerCAmelCase__ : Tuple = 3 * [inputs["""prompt"""]]
# forward
lowerCAmelCase__ : Optional[int] = pipe(**__UpperCAmelCase )
lowerCAmelCase__ : Any = output.images[0, -3:, -3:, -1]
lowerCAmelCase__ : Union[str, Any] = self.get_dummy_inputs()
lowerCAmelCase__ : List[str] = 3 * [inputs.pop("""prompt""" )]
lowerCAmelCase__ : Tuple = []
for p in [prompt, negative_prompt]:
lowerCAmelCase__ : Any = pipe.tokenizer(
__UpperCAmelCase ,padding="""max_length""" ,max_length=pipe.tokenizer.model_max_length ,truncation=__UpperCAmelCase ,return_tensors="""np""" ,)
lowerCAmelCase__ : Any = text_inputs["""input_ids"""]
embeds.append(pipe.text_encoder(input_ids=text_inputs.astype(np.intaa ) )[0] )
lowerCAmelCase__ , lowerCAmelCase__ : Any = embeds
# forward
lowerCAmelCase__ : Any = pipe(**__UpperCAmelCase )
lowerCAmelCase__ : Any = output.images[0, -3:, -3:, -1]
assert np.abs(image_slice_a.flatten() - image_slice_a.flatten() ).max() < 1E-4
@nightly
@require_onnxruntime
@require_torch_gpu
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Tuple:
return (
"CUDAExecutionProvider",
{
"gpu_mem_limit": "15000000000", # 15GB
"arena_extend_strategy": "kSameAsRequested",
},
)
@property
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Dict = ort.SessionOptions()
lowerCAmelCase__ : Dict = False
return options
def UpperCAmelCase_ ( self ) -> str:
# using the PNDM scheduler by default
lowerCAmelCase__ : Union[str, Any] = OnnxStableDiffusionPipeline.from_pretrained(
"""CompVis/stable-diffusion-v1-4""" ,revision="""onnx""" ,safety_checker=__UpperCAmelCase ,feature_extractor=__UpperCAmelCase ,provider=self.gpu_provider ,sess_options=self.gpu_options ,)
sd_pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = """A painting of a squirrel eating a burger"""
np.random.seed(0 )
lowerCAmelCase__ : List[Any] = sd_pipe([prompt] ,guidance_scale=6.0 ,num_inference_steps=10 ,output_type="""np""" )
lowerCAmelCase__ : Optional[int] = output.images
lowerCAmelCase__ : Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
lowerCAmelCase__ : int = np.array([0.0_4_5_2, 0.0_3_9_0, 0.0_0_8_7, 0.0_3_5_0, 0.0_6_1_7, 0.0_3_6_4, 0.0_5_4_4, 0.0_5_2_3, 0.0_7_2_0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-3
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : Dict = DDIMScheduler.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" ,subfolder="""scheduler""" ,revision="""onnx""" )
lowerCAmelCase__ : List[Any] = OnnxStableDiffusionPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" ,revision="""onnx""" ,scheduler=__UpperCAmelCase ,safety_checker=__UpperCAmelCase ,feature_extractor=__UpperCAmelCase ,provider=self.gpu_provider ,sess_options=self.gpu_options ,)
sd_pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = """open neural network exchange"""
lowerCAmelCase__ : Optional[int] = np.random.RandomState(0 )
lowerCAmelCase__ : str = sd_pipe([prompt] ,guidance_scale=7.5 ,num_inference_steps=10 ,generator=__UpperCAmelCase ,output_type="""np""" )
lowerCAmelCase__ : Optional[int] = output.images
lowerCAmelCase__ : Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
lowerCAmelCase__ : int = np.array([0.2_8_6_7, 0.1_9_7_4, 0.1_4_8_1, 0.7_2_9_4, 0.7_2_5_1, 0.6_6_6_7, 0.4_1_9_4, 0.5_6_4_2, 0.6_4_8_6] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-3
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : str = LMSDiscreteScheduler.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" ,subfolder="""scheduler""" ,revision="""onnx""" )
lowerCAmelCase__ : Tuple = OnnxStableDiffusionPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" ,revision="""onnx""" ,scheduler=__UpperCAmelCase ,safety_checker=__UpperCAmelCase ,feature_extractor=__UpperCAmelCase ,provider=self.gpu_provider ,sess_options=self.gpu_options ,)
sd_pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """open neural network exchange"""
lowerCAmelCase__ : List[Any] = np.random.RandomState(0 )
lowerCAmelCase__ : Optional[int] = sd_pipe([prompt] ,guidance_scale=7.5 ,num_inference_steps=10 ,generator=__UpperCAmelCase ,output_type="""np""" )
lowerCAmelCase__ : str = output.images
lowerCAmelCase__ : int = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
lowerCAmelCase__ : List[Any] = np.array([0.2_3_0_6, 0.1_9_5_9, 0.1_5_9_3, 0.6_5_4_9, 0.6_3_9_4, 0.5_4_0_8, 0.5_0_6_5, 0.6_0_1_0, 0.6_1_6_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-3
def UpperCAmelCase_ ( self ) -> List[Any]:
lowerCAmelCase__ : int = 0
def test_callback_fn(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> None:
lowerCAmelCase__ : int = True
nonlocal number_of_steps
number_of_steps += 1
if step == 0:
assert latents.shape == (1, 4, 64, 64)
lowerCAmelCase__ : List[str] = latents[0, -3:, -3:, -1]
lowerCAmelCase__ : int = np.array(
[-0.6_7_7_2, -0.3_8_3_5, -1.2_4_5_6, 0.1_9_0_5, -1.0_9_7_4, 0.6_9_6_7, -1.9_3_5_3, 0.0_1_7_8, 1.0_1_6_7] )
assert np.abs(latents_slice.flatten() - expected_slice ).max() < 1E-3
elif step == 5:
assert latents.shape == (1, 4, 64, 64)
lowerCAmelCase__ : List[Any] = latents[0, -3:, -3:, -1]
lowerCAmelCase__ : Optional[Any] = np.array(
[-0.3_3_5_1, 0.2_2_4_1, -0.1_8_3_7, -0.2_3_2_5, -0.6_5_7_7, 0.3_3_9_3, -0.0_2_4_1, 0.5_8_9_9, 1.3_8_7_5] )
assert np.abs(latents_slice.flatten() - expected_slice ).max() < 1E-3
lowerCAmelCase__ : int = False
lowerCAmelCase__ : Any = OnnxStableDiffusionPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" ,revision="""onnx""" ,safety_checker=__UpperCAmelCase ,feature_extractor=__UpperCAmelCase ,provider=self.gpu_provider ,sess_options=self.gpu_options ,)
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : Dict = """Andromeda galaxy in a bottle"""
lowerCAmelCase__ : List[str] = np.random.RandomState(0 )
pipe(
prompt=__UpperCAmelCase ,num_inference_steps=5 ,guidance_scale=7.5 ,generator=__UpperCAmelCase ,callback=__UpperCAmelCase ,callback_steps=1 ,)
assert test_callback_fn.has_been_called
assert number_of_steps == 6
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Any = OnnxStableDiffusionPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" ,revision="""onnx""" ,safety_checker=__UpperCAmelCase ,feature_extractor=__UpperCAmelCase ,provider=self.gpu_provider ,sess_options=self.gpu_options ,)
assert isinstance(__UpperCAmelCase ,__UpperCAmelCase )
assert pipe.safety_checker is None
lowerCAmelCase__ : Dict = pipe("""example prompt""" ,num_inference_steps=2 ).images[0]
assert image is not None
# check that there's no error when saving a pipeline with one of the models being None
with tempfile.TemporaryDirectory() as tmpdirname:
pipe.save_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : int = OnnxStableDiffusionPipeline.from_pretrained(__UpperCAmelCase )
# sanity check that the pipeline still works
assert pipe.safety_checker is None
lowerCAmelCase__ : Union[str, Any] = pipe("""example prompt""" ,num_inference_steps=2 ).images[0]
assert image is not None
| 37
|
'''simple docstring'''
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , )
@pytest.mark.usefixtures('''sm_env''' )
@parameterized_class(
[
{
'''framework''': '''pytorch''',
'''script''': '''run_glue.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_5_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''pytorch''',
'''script''': '''run_ddp.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''tensorflow''',
'''script''': '''run_tf_dist.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.6, '''eval_loss''': 0.7},
},
] )
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.framework == "pytorch":
subprocess.run(
F"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() ,encoding="""utf-8""" ,check=__UpperCAmelCase ,)
assert hasattr(self ,"""env""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = F"""{self.env.base_job_name}-{instance_count}-{'ddp' if 'ddp' in self.script else 'smd'}"""
# distributed data settings
lowerCAmelCase__ : Any = {"""smdistributed""": {"""dataparallel""": {"""enabled""": True}}} if self.script != """run_ddp.py""" else None
# creates estimator
return HuggingFace(
entry_point=self.script ,source_dir=self.env.test_path ,role=self.env.role ,image_uri=self.env.image_uri ,base_job_name=__UpperCAmelCase ,instance_count=__UpperCAmelCase ,instance_type=self.instance_type ,debugger_hook_config=__UpperCAmelCase ,hyperparameters={**self.env.distributed_hyperparameters, """model_name_or_path""": self.model_name_or_path} ,metric_definitions=self.env.metric_definitions ,distribution=__UpperCAmelCase ,py_version="""py36""" ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
TrainingJobAnalytics(__UpperCAmelCase ).export_csv(F"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(2,)] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
# create estimator
lowerCAmelCase__ : List[Any] = self.create_estimator(__UpperCAmelCase )
# run training
estimator.fit()
# result dataframe
lowerCAmelCase__ : Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
lowerCAmelCase__ : int = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
lowerCAmelCase__ : List[Any] = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
lowerCAmelCase__ : List[str] = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" ,99_9999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(F"""{estimator.latest_training_job.name}.json""" ,"""w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} ,__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
import os
from distutils.util import strtobool
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for e in env_keys:
lowerCAmelCase__ : List[Any] = int(os.environ.get(UpperCamelCase , -1 ) )
if val >= 0:
return val
return default
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=False ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = os.environ.get(UpperCamelCase , str(UpperCamelCase ) )
return strtobool(UpperCamelCase ) == 1 # As its name indicates `strtobool` actually returns an int...
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase="no" ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = os.environ.get(UpperCamelCase , str(UpperCamelCase ) )
return value
| 37
|
'''simple docstring'''
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''albert-base-v1''': '''https://huggingface.co/albert-base-v1/resolve/main/spiece.model''',
'''albert-large-v1''': '''https://huggingface.co/albert-large-v1/resolve/main/spiece.model''',
'''albert-xlarge-v1''': '''https://huggingface.co/albert-xlarge-v1/resolve/main/spiece.model''',
'''albert-xxlarge-v1''': '''https://huggingface.co/albert-xxlarge-v1/resolve/main/spiece.model''',
'''albert-base-v2''': '''https://huggingface.co/albert-base-v2/resolve/main/spiece.model''',
'''albert-large-v2''': '''https://huggingface.co/albert-large-v2/resolve/main/spiece.model''',
'''albert-xlarge-v2''': '''https://huggingface.co/albert-xlarge-v2/resolve/main/spiece.model''',
'''albert-xxlarge-v2''': '''https://huggingface.co/albert-xxlarge-v2/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''albert-base-v1''': 512,
'''albert-large-v1''': 512,
'''albert-xlarge-v1''': 512,
'''albert-xxlarge-v1''': 512,
'''albert-base-v2''': 512,
'''albert-large-v2''': 512,
'''albert-xlarge-v2''': 512,
'''albert-xxlarge-v2''': 512,
}
_lowerCAmelCase = '''▁'''
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Any = VOCAB_FILES_NAMES
__lowercase : Optional[Any] = PRETRAINED_VOCAB_FILES_MAP
__lowercase : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=True ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[MASK]" ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
# Mask token behave like a normal word, i.e. include the space before it and
# is included in the raw text, there should be a match in a non-normalized sentence.
lowerCAmelCase__ : Tuple = (
AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ,normalized=__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase )
else mask_token
)
lowerCAmelCase__ : Any = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : str = do_lower_case
lowerCAmelCase__ : int = remove_space
lowerCAmelCase__ : Tuple = keep_accents
lowerCAmelCase__ : Any = vocab_file
lowerCAmelCase__ : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> Optional[int]:
return len(self.sp_model )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Union[str, Any] = {self.convert_ids_to_tokens(__UpperCAmelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = self.__dict__.copy()
lowerCAmelCase__ : Optional[Any] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Union[str, Any] = {}
lowerCAmelCase__ : Any = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[int]:
if self.remove_space:
lowerCAmelCase__ : int = """ """.join(inputs.strip().split() )
else:
lowerCAmelCase__ : str = inputs
lowerCAmelCase__ : Tuple = outputs.replace("""``""" ,"""\"""" ).replace("""''""" ,"""\"""" )
if not self.keep_accents:
lowerCAmelCase__ : Any = unicodedata.normalize("""NFKD""" ,__UpperCAmelCase )
lowerCAmelCase__ : Dict = """""".join([c for c in outputs if not unicodedata.combining(__UpperCAmelCase )] )
if self.do_lower_case:
lowerCAmelCase__ : Tuple = outputs.lower()
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
lowerCAmelCase__ : str = []
for piece in pieces:
if len(__UpperCAmelCase ) > 1 and piece[-1] == str(""",""" ) and piece[-2].isdigit():
lowerCAmelCase__ : List[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(__UpperCAmelCase ,"""""" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
lowerCAmelCase__ : str = cur_pieces[1:]
else:
lowerCAmelCase__ : int = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(__UpperCAmelCase )
else:
new_pieces.append(__UpperCAmelCase )
return new_pieces
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
return self.sp_model.IdToPiece(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : Tuple = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : List[str] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string.strip()
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : int = [self.sep_token_id]
lowerCAmelCase__ : Dict = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
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 not None:
return [1] + ([0] * len(__UpperCAmelCase )) + [1] + ([0] * len(__UpperCAmelCase )) + [1]
return [1] + ([0] * len(__UpperCAmelCase )) + [1]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : List[str] = [self.sep_token_id]
lowerCAmelCase__ : Tuple = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : int = 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:
lowerCAmelCase__ : List[Any] = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
| 37
| 1
|
'''simple docstring'''
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''merges_file''': '''merges.txt''',
}
_lowerCAmelCase = {
'''vocab_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json'''},
'''merges_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt'''},
}
_lowerCAmelCase = {
'''ctrl''': 256,
}
_lowerCAmelCase = {
'''Pregnancy''': 16_8629,
'''Christianity''': 7675,
'''Explain''': 10_6423,
'''Fitness''': 6_3440,
'''Saving''': 6_3163,
'''Ask''': 2_7171,
'''Ass''': 9_5985,
'''Joke''': 16_3509,
'''Questions''': 4_5622,
'''Thoughts''': 4_9605,
'''Retail''': 5_2342,
'''Feminism''': 16_4338,
'''Writing''': 1_1992,
'''Atheism''': 19_2263,
'''Netflix''': 4_8616,
'''Computing''': 3_9639,
'''Opinion''': 4_3213,
'''Alone''': 4_4967,
'''Funny''': 5_8917,
'''Gaming''': 4_0358,
'''Human''': 4088,
'''India''': 1331,
'''Joker''': 7_7138,
'''Diet''': 3_6206,
'''Legal''': 1_1859,
'''Norman''': 4939,
'''Tip''': 7_2689,
'''Weight''': 5_2343,
'''Movies''': 4_6273,
'''Running''': 2_3425,
'''Science''': 2090,
'''Horror''': 3_7793,
'''Confession''': 6_0572,
'''Finance''': 1_2250,
'''Politics''': 1_6360,
'''Scary''': 19_1985,
'''Support''': 1_2654,
'''Technologies''': 3_2516,
'''Teenage''': 6_6160,
'''Event''': 3_2769,
'''Learned''': 6_7460,
'''Notion''': 18_2770,
'''Wikipedia''': 3_7583,
'''Books''': 6665,
'''Extract''': 7_6050,
'''Confessions''': 10_2701,
'''Conspiracy''': 7_5932,
'''Links''': 6_3674,
'''Narcissus''': 15_0425,
'''Relationship''': 5_4766,
'''Relationships''': 13_4796,
'''Reviews''': 4_1671,
'''News''': 4256,
'''Translation''': 2_6820,
'''multilingual''': 12_8406,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = set()
lowerCAmelCase__ : str = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCAmelCase__ : List[Any] = char
lowerCAmelCase__ : Optional[Any] = set(UpperCamelCase )
return pairs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Any = CONTROL_CODES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<unk>" ,**__UpperCAmelCase ) -> Optional[int]:
super().__init__(unk_token=__UpperCAmelCase ,**__UpperCAmelCase )
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as vocab_handle:
lowerCAmelCase__ : int = json.load(__UpperCAmelCase )
lowerCAmelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as merges_handle:
lowerCAmelCase__ : Union[str, Any] = merges_handle.read().split("""\n""" )[1:-1]
lowerCAmelCase__ : Optional[Any] = [tuple(merge.split() ) for merge in merges]
lowerCAmelCase__ : int = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return len(self.encoder )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return dict(self.encoder ,**self.added_tokens_encoder )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if token in self.cache:
return self.cache[token]
lowerCAmelCase__ : Optional[Any] = tuple(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tuple(list(word[:-1] ) + [word[-1] + """</w>"""] )
lowerCAmelCase__ : Any = get_pairs(__UpperCAmelCase )
if not pairs:
return token
while True:
lowerCAmelCase__ : List[str] = min(__UpperCAmelCase ,key=lambda __UpperCAmelCase : self.bpe_ranks.get(__UpperCAmelCase ,float("""inf""" ) ) )
if bigram not in self.bpe_ranks:
break
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = bigram
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : Dict = 0
while i < len(__UpperCAmelCase ):
try:
lowerCAmelCase__ : Optional[Any] = word.index(__UpperCAmelCase ,__UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCAmelCase__ : Dict = 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
lowerCAmelCase__ : Tuple = tuple(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = new_word
if len(__UpperCAmelCase ) == 1:
break
else:
lowerCAmelCase__ : Dict = get_pairs(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """@@ """.join(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = word[:-4]
lowerCAmelCase__ : Optional[Any] = word
return word
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : Any = re.findall(R"""\S+\n?""" ,__UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(__UpperCAmelCase ).split(""" """ ) ) )
return split_tokens
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
return self.encoder.get(__UpperCAmelCase ,self.encoder.get(self.unk_token ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = """ """.join(__UpperCAmelCase ).replace("""@@ """ ,"""""" ).strip()
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : List[Any] = os.path.join(
__UpperCAmelCase ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : Optional[int] = 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""" )
lowerCAmelCase__ : int = 0
with open(__UpperCAmelCase ,"""w""" ,encoding="""utf-8""" ) as writer:
writer.write("""#version: 0.2\n""" )
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 {merge_file}: BPE merge indices are not consecutive."""
""" Please check that the tokenizer is not corrupted!""" )
lowerCAmelCase__ : Dict = token_index
writer.write(""" """.join(__UpperCAmelCase ) + """\n""" )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 37
|
'''simple docstring'''
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''merges_file''': '''merges.txt''',
}
_lowerCAmelCase = {
'''vocab_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json'''},
'''merges_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt'''},
}
_lowerCAmelCase = {
'''ctrl''': 256,
}
_lowerCAmelCase = {
'''Pregnancy''': 16_8629,
'''Christianity''': 7675,
'''Explain''': 10_6423,
'''Fitness''': 6_3440,
'''Saving''': 6_3163,
'''Ask''': 2_7171,
'''Ass''': 9_5985,
'''Joke''': 16_3509,
'''Questions''': 4_5622,
'''Thoughts''': 4_9605,
'''Retail''': 5_2342,
'''Feminism''': 16_4338,
'''Writing''': 1_1992,
'''Atheism''': 19_2263,
'''Netflix''': 4_8616,
'''Computing''': 3_9639,
'''Opinion''': 4_3213,
'''Alone''': 4_4967,
'''Funny''': 5_8917,
'''Gaming''': 4_0358,
'''Human''': 4088,
'''India''': 1331,
'''Joker''': 7_7138,
'''Diet''': 3_6206,
'''Legal''': 1_1859,
'''Norman''': 4939,
'''Tip''': 7_2689,
'''Weight''': 5_2343,
'''Movies''': 4_6273,
'''Running''': 2_3425,
'''Science''': 2090,
'''Horror''': 3_7793,
'''Confession''': 6_0572,
'''Finance''': 1_2250,
'''Politics''': 1_6360,
'''Scary''': 19_1985,
'''Support''': 1_2654,
'''Technologies''': 3_2516,
'''Teenage''': 6_6160,
'''Event''': 3_2769,
'''Learned''': 6_7460,
'''Notion''': 18_2770,
'''Wikipedia''': 3_7583,
'''Books''': 6665,
'''Extract''': 7_6050,
'''Confessions''': 10_2701,
'''Conspiracy''': 7_5932,
'''Links''': 6_3674,
'''Narcissus''': 15_0425,
'''Relationship''': 5_4766,
'''Relationships''': 13_4796,
'''Reviews''': 4_1671,
'''News''': 4256,
'''Translation''': 2_6820,
'''multilingual''': 12_8406,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = set()
lowerCAmelCase__ : str = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCAmelCase__ : List[Any] = char
lowerCAmelCase__ : Optional[Any] = set(UpperCamelCase )
return pairs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Any = CONTROL_CODES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<unk>" ,**__UpperCAmelCase ) -> Optional[int]:
super().__init__(unk_token=__UpperCAmelCase ,**__UpperCAmelCase )
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as vocab_handle:
lowerCAmelCase__ : int = json.load(__UpperCAmelCase )
lowerCAmelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as merges_handle:
lowerCAmelCase__ : Union[str, Any] = merges_handle.read().split("""\n""" )[1:-1]
lowerCAmelCase__ : Optional[Any] = [tuple(merge.split() ) for merge in merges]
lowerCAmelCase__ : int = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return len(self.encoder )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return dict(self.encoder ,**self.added_tokens_encoder )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if token in self.cache:
return self.cache[token]
lowerCAmelCase__ : Optional[Any] = tuple(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tuple(list(word[:-1] ) + [word[-1] + """</w>"""] )
lowerCAmelCase__ : Any = get_pairs(__UpperCAmelCase )
if not pairs:
return token
while True:
lowerCAmelCase__ : List[str] = min(__UpperCAmelCase ,key=lambda __UpperCAmelCase : self.bpe_ranks.get(__UpperCAmelCase ,float("""inf""" ) ) )
if bigram not in self.bpe_ranks:
break
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = bigram
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : Dict = 0
while i < len(__UpperCAmelCase ):
try:
lowerCAmelCase__ : Optional[Any] = word.index(__UpperCAmelCase ,__UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCAmelCase__ : Dict = 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
lowerCAmelCase__ : Tuple = tuple(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = new_word
if len(__UpperCAmelCase ) == 1:
break
else:
lowerCAmelCase__ : Dict = get_pairs(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """@@ """.join(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = word[:-4]
lowerCAmelCase__ : Optional[Any] = word
return word
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : Any = re.findall(R"""\S+\n?""" ,__UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(__UpperCAmelCase ).split(""" """ ) ) )
return split_tokens
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
return self.encoder.get(__UpperCAmelCase ,self.encoder.get(self.unk_token ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = """ """.join(__UpperCAmelCase ).replace("""@@ """ ,"""""" ).strip()
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : List[Any] = os.path.join(
__UpperCAmelCase ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : Optional[int] = 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""" )
lowerCAmelCase__ : int = 0
with open(__UpperCAmelCase ,"""w""" ,encoding="""utf-8""" ) as writer:
writer.write("""#version: 0.2\n""" )
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 {merge_file}: BPE merge indices are not consecutive."""
""" Please check that the tokenizer is not corrupted!""" )
lowerCAmelCase__ : Dict = token_index
writer.write(""" """.join(__UpperCAmelCase ) + """\n""" )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 37
| 1
|
'''simple docstring'''
import math
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> int:
lowerCAmelCase__ : int = 0.0
lowerCAmelCase__ : Tuple = 0.0
for i in range(len(__UpperCAmelCase ) ):
da += math.pow((sample[i] - weights[0][i]) ,2 )
da += math.pow((sample[i] - weights[1][i]) ,2 )
return 0 if da > da else 1
return 0
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> list[list[int | float]]:
for i in range(len(__UpperCAmelCase ) ):
weights[j][i] += alpha * (sample[i] - weights[j][i])
return weights
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Dict = [[1, 1, 0, 0], [0, 0, 0, 1], [1, 0, 0, 0], [0, 0, 1, 1]]
# weight initialization ( n, C )
lowerCAmelCase__ : int = [[0.2, 0.6, 0.5, 0.9], [0.8, 0.4, 0.7, 0.3]]
# training
lowerCAmelCase__ : List[Any] = SelfOrganizingMap()
lowerCAmelCase__ : Dict = 3
lowerCAmelCase__ : Optional[Any] = 0.5
for _ in range(UpperCamelCase ):
for j in range(len(UpperCamelCase ) ):
# training sample
lowerCAmelCase__ : Optional[int] = training_samples[j]
# Compute the winning vector
lowerCAmelCase__ : Dict = self_organizing_map.get_winner(UpperCamelCase , UpperCamelCase )
# Update the winning vector
lowerCAmelCase__ : Optional[int] = self_organizing_map.update(UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase )
# classify test sample
lowerCAmelCase__ : Union[str, Any] = [0, 0, 0, 1]
lowerCAmelCase__ : str = self_organizing_map.get_winner(UpperCamelCase , UpperCamelCase )
# results
print(f"""Clusters that the test sample belongs to : {winner}""" )
print(f"""Weights that have been trained : {weights}""" )
# running the main() function
if __name__ == "__main__":
main()
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : str = set()
# Replace all the whitespace in our sentence
lowerCAmelCase__ : Tuple = input_str.replace(""" """ , """""" )
for alpha in input_str:
if "a" <= alpha.lower() <= "z":
frequency.add(alpha.lower() )
return len(UpperCamelCase ) == 26
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : Any = [False] * 26
for char in input_str:
if char.islower():
lowerCAmelCase__ : Optional[Any] = True
elif char.isupper():
lowerCAmelCase__ : Any = True
return all(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
return len({char for char in input_str.lower() if char.isalpha()} ) == 26
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from timeit import timeit
lowerCAmelCase__ : Union[str, Any] = """from __main__ import is_pangram, is_pangram_faster, is_pangram_fastest"""
print(timeit("""is_pangram()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_faster()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_fastest()""" , setup=UpperCamelCase ) )
# 5.348480500048026, 2.6477354579837993, 1.8470395830227062
# 5.036091582966037, 2.644472333951853, 1.8869528750656173
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import argparse
import ast
import logging
import os
import sys
import pandas as pd
import torch
from tqdm import tqdm
from transformers import BartForConditionalGeneration, RagRetriever, RagSequenceForGeneration, RagTokenForGeneration
from transformers import logging as transformers_logging
sys.path.append(os.path.join(os.getcwd())) # noqa: E402 # isort:skip
from utils_rag import exact_match_score, fa_score # noqa: E402 # isort:skip
_lowerCAmelCase = logging.getLogger(__name__)
logging.basicConfig(level=logging.INFO)
transformers_logging.set_verbosity_info()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if "token" in model_name_or_path:
return "rag_token"
if "sequence" in model_name_or_path:
return "rag_sequence"
if "bart" in model_name_or_path:
return "bart"
return None
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return max(metric_fn(UpperCamelCase , UpperCamelCase ) for gt in ground_truths )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : int = [line.strip() for line in open(UpperCamelCase , """r""" ).readlines()]
lowerCAmelCase__ : Union[str, Any] = []
if args.gold_data_mode == "qa":
lowerCAmelCase__ : Tuple = pd.read_csv(UpperCamelCase , sep="""\t""" , header=UpperCamelCase )
for answer_list in data[1]:
lowerCAmelCase__ : Optional[Any] = ast.literal_eval(UpperCamelCase )
answers.append(UpperCamelCase )
else:
lowerCAmelCase__ : str = [line.strip() for line in open(UpperCamelCase , """r""" ).readlines()]
lowerCAmelCase__ : Tuple = [[reference] for reference in references]
lowerCAmelCase__ : Optional[Any] = 0
for prediction, ground_truths in zip(UpperCamelCase , UpperCamelCase ):
total += 1
em += metric_max_over_ground_truths(UpperCamelCase , UpperCamelCase , UpperCamelCase )
fa += metric_max_over_ground_truths(UpperCamelCase , UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : str = 100.0 * em / total
lowerCAmelCase__ : Optional[int] = 100.0 * fa / total
logger.info(f"""F1: {fa:.2f}""" )
logger.info(f"""EM: {em:.2f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = args.k
lowerCAmelCase__ : Optional[int] = [line.strip() for line in open(UpperCamelCase , """r""" ).readlines()]
lowerCAmelCase__ : Optional[int] = [line.strip() for line in open(UpperCamelCase , """r""" ).readlines()]
lowerCAmelCase__ : str = 0
for hypo, reference in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : Union[str, Any] = set(hypo.split("""\t""" )[:k] )
lowerCAmelCase__ : List[Any] = set(reference.split("""\t""" ) )
total += 1
em += len(hypo_provenance & ref_provenance ) / k
lowerCAmelCase__ : Optional[int] = 100.0 * em / total
logger.info(f"""Precision@{k}: {em: .2f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
def strip_title(UpperCamelCase ):
if title.startswith("""\"""" ):
lowerCAmelCase__ : Optional[Any] = title[1:]
if title.endswith("""\"""" ):
lowerCAmelCase__ : int = title[:-1]
return title
lowerCAmelCase__ : Tuple = rag_model.retriever.question_encoder_tokenizer.batch_encode_plus(
UpperCamelCase , return_tensors="""pt""" , padding=UpperCamelCase , truncation=UpperCamelCase , )["""input_ids"""].to(args.device )
lowerCAmelCase__ : Optional[int] = rag_model.rag.question_encoder(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = question_enc_outputs[0]
lowerCAmelCase__ : List[str] = rag_model.retriever(
UpperCamelCase , question_enc_pool_output.cpu().detach().to(torch.floataa ).numpy() , prefix=rag_model.rag.generator.config.prefix , n_docs=rag_model.config.n_docs , return_tensors="""pt""" , )
lowerCAmelCase__ : List[str] = rag_model.retriever.index.get_doc_dicts(result.doc_ids )
lowerCAmelCase__ : List[Any] = []
for docs in all_docs:
lowerCAmelCase__ : Tuple = [strip_title(UpperCamelCase ) for title in docs["""title"""]]
provenance_strings.append("""\t""".join(UpperCamelCase ) )
return provenance_strings
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
with torch.no_grad():
lowerCAmelCase__ : str = rag_model.retriever.question_encoder_tokenizer.batch_encode_plus(
UpperCamelCase , return_tensors="""pt""" , padding=UpperCamelCase , truncation=UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = inputs_dict.input_ids.to(args.device )
lowerCAmelCase__ : List[Any] = inputs_dict.attention_mask.to(args.device )
lowerCAmelCase__ : Optional[Any] = rag_model.generate( # rag_model overwrites generate
UpperCamelCase , attention_mask=UpperCamelCase , num_beams=args.num_beams , min_length=args.min_length , max_length=args.max_length , early_stopping=UpperCamelCase , num_return_sequences=1 , bad_words_ids=[[0, 0]] , )
lowerCAmelCase__ : Optional[Any] = rag_model.retriever.generator_tokenizer.batch_decode(UpperCamelCase , skip_special_tokens=UpperCamelCase )
if args.print_predictions:
for q, a in zip(UpperCamelCase , UpperCamelCase ):
logger.info("""Q: {} - A: {}""".format(UpperCamelCase , UpperCamelCase ) )
return answers
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = argparse.ArgumentParser()
parser.add_argument(
"""--model_type""" , choices=["""rag_sequence""", """rag_token""", """bart"""] , type=UpperCamelCase , help=(
"""RAG model type: rag_sequence, rag_token or bart, if none specified, the type is inferred from the"""
""" model_name_or_path"""
) , )
parser.add_argument(
"""--index_name""" , default=UpperCamelCase , choices=["""exact""", """compressed""", """legacy"""] , type=UpperCamelCase , help="""RAG model retriever type""" , )
parser.add_argument(
"""--index_path""" , default=UpperCamelCase , type=UpperCamelCase , help="""Path to the retrieval index""" , )
parser.add_argument("""--n_docs""" , default=5 , type=UpperCamelCase , help="""Number of retrieved docs""" )
parser.add_argument(
"""--model_name_or_path""" , default=UpperCamelCase , type=UpperCamelCase , required=UpperCamelCase , help="""Path to pretrained checkpoints or model identifier from huggingface.co/models""" , )
parser.add_argument(
"""--eval_mode""" , choices=["""e2e""", """retrieval"""] , default="""e2e""" , type=UpperCamelCase , help=(
"""Evaluation mode, e2e calculates exact match and F1 of the downstream task, retrieval calculates"""
""" precision@k."""
) , )
parser.add_argument("""--k""" , default=1 , type=UpperCamelCase , help="""k for the precision@k calculation""" )
parser.add_argument(
"""--evaluation_set""" , default=UpperCamelCase , type=UpperCamelCase , required=UpperCamelCase , help="""Path to a file containing evaluation samples""" , )
parser.add_argument(
"""--gold_data_path""" , default=UpperCamelCase , type=UpperCamelCase , required=UpperCamelCase , help="""Path to a tab-separated file with gold samples""" , )
parser.add_argument(
"""--gold_data_mode""" , default="""qa""" , type=UpperCamelCase , choices=["""qa""", """ans"""] , help=(
"""Format of the gold data file"""
"""qa - a single line in the following format: question [tab] answer_list"""
"""ans - a single line of the gold file contains the expected answer string"""
) , )
parser.add_argument(
"""--predictions_path""" , type=UpperCamelCase , default="""predictions.txt""" , help="""Name of the predictions file, to be stored in the checkpoints directory""" , )
parser.add_argument(
"""--eval_all_checkpoints""" , action="""store_true""" , help="""Evaluate all checkpoints starting with the same prefix as model_name ending and ending with step number""" , )
parser.add_argument(
"""--eval_batch_size""" , default=8 , type=UpperCamelCase , help="""Batch size per GPU/CPU for evaluation.""" , )
parser.add_argument(
"""--recalculate""" , help="""Recalculate predictions even if the prediction file exists""" , action="""store_true""" , )
parser.add_argument(
"""--num_beams""" , default=4 , type=UpperCamelCase , help="""Number of beams to be used when generating answers""" , )
parser.add_argument("""--min_length""" , default=1 , type=UpperCamelCase , help="""Min length of the generated answers""" )
parser.add_argument("""--max_length""" , default=50 , type=UpperCamelCase , help="""Max length of the generated answers""" )
parser.add_argument(
"""--print_predictions""" , action="""store_true""" , help="""If True, prints predictions while evaluating.""" , )
parser.add_argument(
"""--print_docs""" , action="""store_true""" , help="""If True, prints docs retried while generating.""" , )
lowerCAmelCase__ : Dict = parser.parse_args()
lowerCAmelCase__ : Tuple = torch.device("""cuda""" if torch.cuda.is_available() else """cpu""" )
return args
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
if args.model_type is None:
lowerCAmelCase__ : Any = infer_model_type(args.model_name_or_path )
assert args.model_type is not None
if args.model_type.startswith("""rag""" ):
lowerCAmelCase__ : str = RagTokenForGeneration if args.model_type == """rag_token""" else RagSequenceForGeneration
lowerCAmelCase__ : Any = args.n_docs
if args.index_name is not None:
lowerCAmelCase__ : Any = args.index_name
if args.index_path is not None:
lowerCAmelCase__ : Dict = args.index_path
else:
lowerCAmelCase__ : Tuple = BartForConditionalGeneration
lowerCAmelCase__ : List[Any] = (
[f.path for f in os.scandir(args.model_name_or_path ) if f.is_dir()]
if args.eval_all_checkpoints
else [args.model_name_or_path]
)
logger.info("""Evaluate the following checkpoints: %s""" , UpperCamelCase )
lowerCAmelCase__ : Optional[int] = get_scores if args.eval_mode == """e2e""" else get_precision_at_k
lowerCAmelCase__ : Tuple = evaluate_batch_eae if args.eval_mode == """e2e""" else evaluate_batch_retrieval
for checkpoint in checkpoints:
if os.path.exists(args.predictions_path ) and (not args.recalculate):
logger.info("""Calculating metrics based on an existing predictions file: {}""".format(args.predictions_path ) )
score_fn(UpperCamelCase , args.predictions_path , args.gold_data_path )
continue
logger.info("""***** Running evaluation for {} *****""".format(UpperCamelCase ) )
logger.info(""" Batch size = %d""" , args.eval_batch_size )
logger.info(""" Predictions will be stored under {}""".format(args.predictions_path ) )
if args.model_type.startswith("""rag""" ):
lowerCAmelCase__ : str = RagRetriever.from_pretrained(UpperCamelCase , **UpperCamelCase )
lowerCAmelCase__ : List[str] = model_class.from_pretrained(UpperCamelCase , retriever=UpperCamelCase , **UpperCamelCase )
model.retriever.init_retrieval()
else:
lowerCAmelCase__ : Dict = model_class.from_pretrained(UpperCamelCase , **UpperCamelCase )
model.to(args.device )
with open(args.evaluation_set , """r""" ) as eval_file, open(args.predictions_path , """w""" ) as preds_file:
lowerCAmelCase__ : int = []
for line in tqdm(UpperCamelCase ):
questions.append(line.strip() )
if len(UpperCamelCase ) == args.eval_batch_size:
lowerCAmelCase__ : Tuple = evaluate_batch_fn(UpperCamelCase , UpperCamelCase , UpperCamelCase )
preds_file.write("""\n""".join(UpperCamelCase ) + """\n""" )
preds_file.flush()
lowerCAmelCase__ : Optional[int] = []
if len(UpperCamelCase ) > 0:
lowerCAmelCase__ : Tuple = evaluate_batch_fn(UpperCamelCase , UpperCamelCase , UpperCamelCase )
preds_file.write("""\n""".join(UpperCamelCase ) )
preds_file.flush()
score_fn(UpperCamelCase , args.predictions_path , args.gold_data_path )
if __name__ == "__main__":
_lowerCAmelCase = get_args()
main(args)
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = abs(UpperCamelCase )
lowerCAmelCase__ : List[Any] = 0
while n > 0:
res += n % 10
n //= 10
return res
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = abs(UpperCamelCase )
return n if n < 10 else n % 10 + sum_of_digits(n // 10 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return sum(int(UpperCamelCase ) for c in str(abs(UpperCamelCase ) ) )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(UpperCamelCase , UpperCamelCase ) -> None:
lowerCAmelCase__ : str = f"""{func.__name__}({value})"""
lowerCAmelCase__ : str = timeit(f"""__main__.{call}""" , setup="""import __main__""" )
print(f"""{call:56} = {func(UpperCamelCase )} -- {timing:.4f} seconds""" )
for value in (262144, 1125899906842624, 1267650600228229401496703205376):
for func in (sum_of_digits, sum_of_digits_recursion, sum_of_digits_compact):
benchmark_a_function(UpperCamelCase , UpperCamelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import gc
import random
import unittest
import numpy as np
import torch
from transformers import XLMRobertaTokenizer
from diffusers import (
AltDiffusionImgaImgPipeline,
AutoencoderKL,
PNDMScheduler,
UNetaDConditionModel,
)
from diffusers.image_processor import VaeImageProcessor
from diffusers.pipelines.alt_diffusion.modeling_roberta_series import (
RobertaSeriesConfig,
RobertaSeriesModelWithTransformation,
)
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
enable_full_determinism()
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> List[Any]:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
@property
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : Union[str, Any] = 1
lowerCAmelCase__ : Optional[Any] = 3
lowerCAmelCase__ : Optional[int] = (32, 32)
lowerCAmelCase__ : Optional[int] = floats_tensor((batch_size, num_channels) + sizes ,rng=random.Random(0 ) ).to(__UpperCAmelCase )
return image
@property
def UpperCAmelCase_ ( self ) -> str:
torch.manual_seed(0 )
lowerCAmelCase__ : Any = 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 ,)
return model
@property
def UpperCAmelCase_ ( self ) -> List[str]:
torch.manual_seed(0 )
lowerCAmelCase__ : Optional[Any] = AutoencoderKL(
block_out_channels=[32, 64] ,in_channels=3 ,out_channels=3 ,down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] ,up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] ,latent_channels=4 ,)
return model
@property
def UpperCAmelCase_ ( self ) -> str:
torch.manual_seed(0 )
lowerCAmelCase__ : List[Any] = RobertaSeriesConfig(
hidden_size=32 ,project_dim=32 ,intermediate_size=37 ,layer_norm_eps=1E-05 ,num_attention_heads=4 ,num_hidden_layers=5 ,pad_token_id=1 ,vocab_size=5006 ,)
return RobertaSeriesModelWithTransformation(__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> int:
def extract(*__UpperCAmelCase ,**__UpperCAmelCase ):
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ) -> Optional[int]:
lowerCAmelCase__ : Any = torch.ones([0] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
self.pixel_values.to(__UpperCAmelCase )
return self
return Out()
return extract
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Optional[int] = """cpu""" # ensure determinism for the device-dependent torch.Generator
lowerCAmelCase__ : List[Any] = self.dummy_cond_unet
lowerCAmelCase__ : Tuple = PNDMScheduler(skip_prk_steps=__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.dummy_vae
lowerCAmelCase__ : Tuple = self.dummy_text_encoder
lowerCAmelCase__ : Union[str, Any] = XLMRobertaTokenizer.from_pretrained("""hf-internal-testing/tiny-xlm-roberta""" )
lowerCAmelCase__ : str = 77
lowerCAmelCase__ : List[str] = self.dummy_image.to(__UpperCAmelCase )
lowerCAmelCase__ : int = init_image / 2 + 0.5
# make sure here that pndm scheduler skips prk
lowerCAmelCase__ : Dict = AltDiffusionImgaImgPipeline(
unet=__UpperCAmelCase ,scheduler=__UpperCAmelCase ,vae=__UpperCAmelCase ,text_encoder=__UpperCAmelCase ,tokenizer=__UpperCAmelCase ,safety_checker=__UpperCAmelCase ,feature_extractor=self.dummy_extractor ,)
lowerCAmelCase__ : List[Any] = VaeImageProcessor(vae_scale_factor=alt_pipe.vae_scale_factor ,do_normalize=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = alt_pipe.to(__UpperCAmelCase )
alt_pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : int = """A painting of a squirrel eating a burger"""
lowerCAmelCase__ : Optional[Any] = torch.Generator(device=__UpperCAmelCase ).manual_seed(0 )
lowerCAmelCase__ : Optional[Any] = alt_pipe(
[prompt] ,generator=__UpperCAmelCase ,guidance_scale=6.0 ,num_inference_steps=2 ,output_type="""np""" ,image=__UpperCAmelCase ,)
lowerCAmelCase__ : List[Any] = output.images
lowerCAmelCase__ : Optional[Any] = torch.Generator(device=__UpperCAmelCase ).manual_seed(0 )
lowerCAmelCase__ : Dict = alt_pipe(
[prompt] ,generator=__UpperCAmelCase ,guidance_scale=6.0 ,num_inference_steps=2 ,output_type="""np""" ,image=__UpperCAmelCase ,return_dict=__UpperCAmelCase ,)[0]
lowerCAmelCase__ : Optional[Any] = image[0, -3:, -3:, -1]
lowerCAmelCase__ : str = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 32, 32, 3)
lowerCAmelCase__ : Optional[Any] = np.array([0.4_4_2_7, 0.3_7_3_1, 0.4_2_4_9, 0.4_9_4_1, 0.4_5_4_6, 0.4_1_4_8, 0.4_1_9_3, 0.4_6_6_6, 0.4_4_9_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-3
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 5E-3
@unittest.skipIf(torch_device != """cuda""" ,"""This test requires a GPU""" )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : Any = self.dummy_cond_unet
lowerCAmelCase__ : Any = PNDMScheduler(skip_prk_steps=__UpperCAmelCase )
lowerCAmelCase__ : int = self.dummy_vae
lowerCAmelCase__ : Union[str, Any] = self.dummy_text_encoder
lowerCAmelCase__ : Union[str, Any] = XLMRobertaTokenizer.from_pretrained("""hf-internal-testing/tiny-xlm-roberta""" )
lowerCAmelCase__ : Tuple = 77
lowerCAmelCase__ : Union[str, Any] = self.dummy_image.to(__UpperCAmelCase )
# put models in fp16
lowerCAmelCase__ : Tuple = unet.half()
lowerCAmelCase__ : str = vae.half()
lowerCAmelCase__ : Union[str, Any] = bert.half()
# make sure here that pndm scheduler skips prk
lowerCAmelCase__ : Dict = AltDiffusionImgaImgPipeline(
unet=__UpperCAmelCase ,scheduler=__UpperCAmelCase ,vae=__UpperCAmelCase ,text_encoder=__UpperCAmelCase ,tokenizer=__UpperCAmelCase ,safety_checker=__UpperCAmelCase ,feature_extractor=self.dummy_extractor ,)
lowerCAmelCase__ : Optional[int] = VaeImageProcessor(vae_scale_factor=alt_pipe.vae_scale_factor ,do_normalize=__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = alt_pipe.to(__UpperCAmelCase )
alt_pipe.set_progress_bar_config(disable=__UpperCAmelCase )
lowerCAmelCase__ : str = """A painting of a squirrel eating a burger"""
lowerCAmelCase__ : int = torch.manual_seed(0 )
lowerCAmelCase__ : Optional[Any] = alt_pipe(
[prompt] ,generator=__UpperCAmelCase ,num_inference_steps=2 ,output_type="""np""" ,image=__UpperCAmelCase ,).images
assert image.shape == (1, 32, 32, 3)
@unittest.skipIf(torch_device != """cuda""" ,"""This test requires a GPU""" )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main"""
"""/img2img/sketch-mountains-input.jpg""" )
# resize to resolution that is divisible by 8 but not 16 or 32
lowerCAmelCase__ : List[Any] = init_image.resize((760, 504) )
lowerCAmelCase__ : Tuple = """BAAI/AltDiffusion"""
lowerCAmelCase__ : str = AltDiffusionImgaImgPipeline.from_pretrained(
__UpperCAmelCase ,safety_checker=__UpperCAmelCase ,)
pipe.to(__UpperCAmelCase )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
pipe.enable_attention_slicing()
lowerCAmelCase__ : Union[str, Any] = """A fantasy landscape, trending on artstation"""
lowerCAmelCase__ : Optional[Any] = torch.manual_seed(0 )
lowerCAmelCase__ : Union[str, Any] = pipe(
prompt=__UpperCAmelCase ,image=__UpperCAmelCase ,strength=0.7_5 ,guidance_scale=7.5 ,generator=__UpperCAmelCase ,output_type="""np""" ,)
lowerCAmelCase__ : List[Any] = output.images[0]
lowerCAmelCase__ : str = image[255:258, 383:386, -1]
assert image.shape == (504, 760, 3)
lowerCAmelCase__ : Any = np.array([0.9_3_5_8, 0.9_3_9_7, 0.9_5_9_9, 0.9_9_0_1, 1.0_0_0_0, 1.0_0_0_0, 0.9_8_8_2, 1.0_0_0_0, 1.0_0_0_0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
@slow
@require_torch_gpu
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : Dict = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main"""
"""/img2img/sketch-mountains-input.jpg""" )
lowerCAmelCase__ : Optional[int] = init_image.resize((768, 512) )
lowerCAmelCase__ : Dict = load_numpy(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/img2img/fantasy_landscape_alt.npy""" )
lowerCAmelCase__ : List[str] = """BAAI/AltDiffusion"""
lowerCAmelCase__ : Optional[Any] = AltDiffusionImgaImgPipeline.from_pretrained(
__UpperCAmelCase ,safety_checker=__UpperCAmelCase ,)
pipe.to(__UpperCAmelCase )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
pipe.enable_attention_slicing()
lowerCAmelCase__ : int = """A fantasy landscape, trending on artstation"""
lowerCAmelCase__ : List[str] = torch.manual_seed(0 )
lowerCAmelCase__ : List[str] = pipe(
prompt=__UpperCAmelCase ,image=__UpperCAmelCase ,strength=0.7_5 ,guidance_scale=7.5 ,generator=__UpperCAmelCase ,output_type="""np""" ,)
lowerCAmelCase__ : List[Any] = output.images[0]
assert image.shape == (512, 768, 3)
# img2img is flaky across GPUs even in fp32, so using MAE here
assert np.abs(expected_image - image ).max() < 1E-2
| 37
|
'''simple docstring'''
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team.
#
# 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.
# this script dumps information about the environment
import os
import sys
import transformers
_lowerCAmelCase = '''3'''
print('''Python version:''', sys.version)
print('''transformers version:''', transformers.__version__)
try:
import torch
print('''Torch version:''', torch.__version__)
print('''Cuda available:''', torch.cuda.is_available())
print('''Cuda version:''', torch.version.cuda)
print('''CuDNN version:''', torch.backends.cudnn.version())
print('''Number of GPUs available:''', torch.cuda.device_count())
print('''NCCL version:''', torch.cuda.nccl.version())
except ImportError:
print('''Torch version:''', None)
try:
import deepspeed
print('''DeepSpeed version:''', deepspeed.__version__)
except ImportError:
print('''DeepSpeed version:''', None)
try:
import tensorflow as tf
print('''TensorFlow version:''', tf.__version__)
print('''TF GPUs available:''', bool(tf.config.list_physical_devices('''GPU''')))
print('''Number of TF GPUs available:''', len(tf.config.list_physical_devices('''GPU''')))
except ImportError:
print('''TensorFlow version:''', None)
| 37
| 1
|
'''simple docstring'''
import argparse
import requests
import torch
from PIL import Image
from transformers import ViTMAEConfig, ViTMAEForPreTraining, ViTMAEImageProcessor
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if "cls_token" in name:
lowerCAmelCase__ : Dict = name.replace("""cls_token""" , """vit.embeddings.cls_token""" )
if "mask_token" in name:
lowerCAmelCase__ : Any = name.replace("""mask_token""" , """decoder.mask_token""" )
if "decoder_pos_embed" in name:
lowerCAmelCase__ : str = name.replace("""decoder_pos_embed""" , """decoder.decoder_pos_embed""" )
if "pos_embed" in name and "decoder" not in name:
lowerCAmelCase__ : List[str] = name.replace("""pos_embed""" , """vit.embeddings.position_embeddings""" )
if "patch_embed.proj" in name:
lowerCAmelCase__ : List[Any] = name.replace("""patch_embed.proj""" , """vit.embeddings.patch_embeddings.projection""" )
if "patch_embed.norm" in name:
lowerCAmelCase__ : Optional[Any] = name.replace("""patch_embed.norm""" , """vit.embeddings.norm""" )
if "decoder_blocks" in name:
lowerCAmelCase__ : Optional[int] = name.replace("""decoder_blocks""" , """decoder.decoder_layers""" )
if "blocks" in name:
lowerCAmelCase__ : Optional[int] = name.replace("""blocks""" , """vit.encoder.layer""" )
if "attn.proj" in name:
lowerCAmelCase__ : Optional[Any] = name.replace("""attn.proj""" , """attention.output.dense""" )
if "attn" in name:
lowerCAmelCase__ : Optional[int] = name.replace("""attn""" , """attention.self""" )
if "norm1" in name:
lowerCAmelCase__ : Dict = name.replace("""norm1""" , """layernorm_before""" )
if "norm2" in name:
lowerCAmelCase__ : Dict = name.replace("""norm2""" , """layernorm_after""" )
if "mlp.fc1" in name:
lowerCAmelCase__ : Union[str, Any] = name.replace("""mlp.fc1""" , """intermediate.dense""" )
if "mlp.fc2" in name:
lowerCAmelCase__ : List[Any] = name.replace("""mlp.fc2""" , """output.dense""" )
if "decoder_embed" in name:
lowerCAmelCase__ : Tuple = name.replace("""decoder_embed""" , """decoder.decoder_embed""" )
if "decoder_norm" in name:
lowerCAmelCase__ : int = name.replace("""decoder_norm""" , """decoder.decoder_norm""" )
if "decoder_pred" in name:
lowerCAmelCase__ : int = name.replace("""decoder_pred""" , """decoder.decoder_pred""" )
if "norm.weight" in name and "decoder" not in name:
lowerCAmelCase__ : Tuple = name.replace("""norm.weight""" , """vit.layernorm.weight""" )
if "norm.bias" in name and "decoder" not in name:
lowerCAmelCase__ : Dict = name.replace("""norm.bias""" , """vit.layernorm.bias""" )
return name
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for key in orig_state_dict.copy().keys():
lowerCAmelCase__ : Any = orig_state_dict.pop(UpperCamelCase )
if "qkv" in key:
lowerCAmelCase__ : str = key.split(""".""" )
lowerCAmelCase__ : Optional[int] = int(key_split[1] )
if "decoder_blocks" in key:
lowerCAmelCase__ : str = config.decoder_hidden_size
lowerCAmelCase__ : Any = """decoder.decoder_layers."""
if "weight" in key:
lowerCAmelCase__ : int = val[:dim, :]
lowerCAmelCase__ : Union[str, Any] = val[dim : dim * 2, :]
lowerCAmelCase__ : Optional[int] = val[-dim:, :]
elif "bias" in key:
lowerCAmelCase__ : Dict = val[:dim]
lowerCAmelCase__ : Any = val[dim : dim * 2]
lowerCAmelCase__ : int = val[-dim:]
else:
lowerCAmelCase__ : Tuple = config.hidden_size
lowerCAmelCase__ : Any = """vit.encoder.layer."""
if "weight" in key:
lowerCAmelCase__ : List[Any] = val[:dim, :]
lowerCAmelCase__ : int = val[dim : dim * 2, :]
lowerCAmelCase__ : Optional[Any] = val[-dim:, :]
elif "bias" in key:
lowerCAmelCase__ : int = val[:dim]
lowerCAmelCase__ : Optional[Any] = val[dim : dim * 2]
lowerCAmelCase__ : Any = val[-dim:]
else:
lowerCAmelCase__ : List[str] = val
return orig_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = ViTMAEConfig()
if "large" in checkpoint_url:
lowerCAmelCase__ : Optional[int] = 1024
lowerCAmelCase__ : int = 4096
lowerCAmelCase__ : Any = 24
lowerCAmelCase__ : Any = 16
elif "huge" in checkpoint_url:
lowerCAmelCase__ : List[Any] = 14
lowerCAmelCase__ : str = 1280
lowerCAmelCase__ : Dict = 5120
lowerCAmelCase__ : Any = 32
lowerCAmelCase__ : Union[str, Any] = 16
lowerCAmelCase__ : Tuple = ViTMAEForPreTraining(UpperCamelCase )
lowerCAmelCase__ : Tuple = torch.hub.load_state_dict_from_url(UpperCamelCase , map_location="""cpu""" )["""model"""]
lowerCAmelCase__ : str = ViTMAEImageProcessor(size=config.image_size )
lowerCAmelCase__ : str = convert_state_dict(UpperCamelCase , UpperCamelCase )
model.load_state_dict(UpperCamelCase )
model.eval()
lowerCAmelCase__ : List[str] = """https://user-images.githubusercontent.com/11435359/147738734-196fd92f-9260-48d5-ba7e-bf103d29364d.jpg"""
lowerCAmelCase__ : Tuple = Image.open(requests.get(UpperCamelCase , stream=UpperCamelCase ).raw )
lowerCAmelCase__ : int = ViTMAEImageProcessor(size=config.image_size )
lowerCAmelCase__ : Optional[Any] = image_processor(images=UpperCamelCase , return_tensors="""pt""" )
# forward pass
torch.manual_seed(2 )
lowerCAmelCase__ : str = model(**UpperCamelCase )
lowerCAmelCase__ : List[Any] = outputs.logits
if "large" in checkpoint_url:
lowerCAmelCase__ : List[str] = torch.tensor(
[[-0.7309, -0.7128, -1.0169], [-1.0161, -0.9058, -1.1878], [-1.0478, -0.9411, -1.1911]] )
elif "huge" in checkpoint_url:
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[[-1.1599, -0.9199, -1.2221], [-1.1952, -0.9269, -1.2307], [-1.2143, -0.9337, -1.2262]] )
else:
lowerCAmelCase__ : int = torch.tensor(
[[-0.9192, -0.8481, -1.1259], [-1.1349, -1.0034, -1.2599], [-1.1757, -1.0429, -1.2726]] )
# verify logits
assert torch.allclose(logits[0, :3, :3] , UpperCamelCase , atol=1e-4 )
print(f"""Saving model to {pytorch_dump_folder_path}""" )
model.save_pretrained(UpperCamelCase )
print(f"""Saving image processor to {pytorch_dump_folder_path}""" )
image_processor.save_pretrained(UpperCamelCase )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--checkpoint_url''',
default='''https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_base.pth''',
type=str,
help='''URL of the checkpoint you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model directory.'''
)
_lowerCAmelCase = parser.parse_args()
convert_vit_mae_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path)
| 37
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xlm-roberta-xl''': '''https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json''',
'''facebook/xlm-roberta-xxl''': '''https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json''',
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''xlm-roberta-xl'''
def __init__( self ,__UpperCAmelCase=25_0880 ,__UpperCAmelCase=2560 ,__UpperCAmelCase=36 ,__UpperCAmelCase=32 ,__UpperCAmelCase=1_0240 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=514 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-05 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = vocab_size
lowerCAmelCase__ : int = hidden_size
lowerCAmelCase__ : int = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : str = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : List[str] = type_vocab_size
lowerCAmelCase__ : List[Any] = initializer_range
lowerCAmelCase__ : Tuple = layer_norm_eps
lowerCAmelCase__ : int = position_embedding_type
lowerCAmelCase__ : Optional[Any] = use_cache
lowerCAmelCase__ : Optional[Any] = classifier_dropout
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Dict = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Any = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
| 1
|
'''simple docstring'''
import fire
from torch.utils.data import DataLoader
from tqdm import tqdm
from transformers import AutoTokenizer
from utils import SeqaSeqDataset, pickle_save
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=1024 , UpperCamelCase=1024 , UpperCamelCase=False , **UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = AutoTokenizer.from_pretrained(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = SeqaSeqDataset(UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , type_path="""train""" , **UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = tok.pad_token_id
def get_lens(UpperCamelCase ):
lowerCAmelCase__ : Optional[int] = tqdm(
DataLoader(UpperCamelCase , batch_size=512 , num_workers=8 , shuffle=UpperCamelCase , collate_fn=ds.collate_fn ) , desc=str(ds.len_file ) , )
lowerCAmelCase__ : List[str] = []
for batch in dl:
lowerCAmelCase__ : Dict = batch["""input_ids"""].ne(UpperCamelCase ).sum(1 ).tolist()
lowerCAmelCase__ : str = batch["""labels"""].ne(UpperCamelCase ).sum(1 ).tolist()
if consider_target:
for src, tgt in zip(UpperCamelCase , UpperCamelCase ):
max_lens.append(max(UpperCamelCase , UpperCamelCase ) )
else:
max_lens.extend(UpperCamelCase )
return max_lens
lowerCAmelCase__ : int = get_lens(UpperCamelCase )
lowerCAmelCase__ : int = SeqaSeqDataset(UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , type_path="""val""" , **UpperCamelCase )
lowerCAmelCase__ : str = get_lens(UpperCamelCase )
pickle_save(UpperCamelCase , train_ds.len_file )
pickle_save(UpperCamelCase , val_ds.len_file )
if __name__ == "__main__":
fire.Fire(save_len_file)
| 37
|
'''simple docstring'''
from __future__ import annotations
import math
from collections import Counter
from string import ascii_lowercase
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = analyze_text(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = list(""" """ + ascii_lowercase )
# what is our total sum of probabilities.
lowerCAmelCase__ : List[Any] = sum(single_char_strings.values() )
# one length string
lowerCAmelCase__ : Optional[int] = 0
# for each alpha we go in our dict and if it is in it we calculate entropy
for ch in my_alphas:
if ch in single_char_strings:
lowerCAmelCase__ : List[Any] = single_char_strings[ch]
lowerCAmelCase__ : List[Any] = my_str / all_sum
my_fir_sum += prob * math.loga(UpperCamelCase ) # entropy formula.
# print entropy
print(f"""{round(-1 * my_fir_sum ):.1f}""" )
# two len string
lowerCAmelCase__ : Dict = sum(two_char_strings.values() )
lowerCAmelCase__ : int = 0
# for each alpha (two in size) calculate entropy.
for cha in my_alphas:
for cha in my_alphas:
lowerCAmelCase__ : Union[str, Any] = cha + cha
if sequence in two_char_strings:
lowerCAmelCase__ : Dict = two_char_strings[sequence]
lowerCAmelCase__ : Tuple = int(UpperCamelCase ) / all_sum
my_sec_sum += prob * math.loga(UpperCamelCase )
# print second entropy
print(f"""{round(-1 * my_sec_sum ):.1f}""" )
# print the difference between them
print(f"""{round((-1 * my_sec_sum) - (-1 * my_fir_sum) ):.1f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = Counter() # type: ignore
lowerCAmelCase__ : Tuple = Counter() # type: ignore
single_char_strings[text[-1]] += 1
# first case when we have space at start.
two_char_strings[" " + text[0]] += 1
for i in range(0 , len(UpperCamelCase ) - 1 ):
single_char_strings[text[i]] += 1
two_char_strings[text[i : i + 2]] += 1
return single_char_strings, two_char_strings
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
import doctest
doctest.testmod()
# text = (
# "Had repulsive dashwoods suspicion sincerity but advantage now him. Remark "
# "easily garret nor nay. Civil those mrs enjoy shy fat merry. You greatest "
# "jointure saw horrible. He private he on be imagine suppose. Fertile "
# "beloved evident through no service elderly is. Blind there if every no so "
# "at. Own neglected you preferred way sincerity delivered his attempted. To "
# "of message cottage windows do besides against uncivil. Delightful "
# "unreserved impossible few estimating men favourable see entreaties. She "
# "propriety immediate was improving. He or entrance humoured likewise "
# "moderate. Much nor game son say feel. Fat make met can must form into "
# "gate. Me we offending prevailed discovery. "
# )
# calculate_prob(text)
if __name__ == "__main__":
main()
| 37
| 1
|
'''simple docstring'''
from dataclasses import dataclass
from typing import Optional
import numpy as np
import torch
import torch.nn as nn
from ..utils import BaseOutput, is_torch_version, randn_tensor
from .attention_processor import SpatialNorm
from .unet_ad_blocks import UNetMidBlockaD, get_down_block, get_up_block
@dataclass
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : torch.FloatTensor
class lowerCAmelCase_( nn.Module ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase=3 ,__UpperCAmelCase=3 ,__UpperCAmelCase=("DownEncoderBlock2D",) ,__UpperCAmelCase=(64,) ,__UpperCAmelCase=2 ,__UpperCAmelCase=32 ,__UpperCAmelCase="silu" ,__UpperCAmelCase=True ,) -> Optional[int]:
super().__init__()
lowerCAmelCase__ : Union[str, Any] = layers_per_block
lowerCAmelCase__ : Dict = torch.nn.Convad(
__UpperCAmelCase ,block_out_channels[0] ,kernel_size=3 ,stride=1 ,padding=1 ,)
lowerCAmelCase__ : Optional[Any] = None
lowerCAmelCase__ : Optional[Any] = nn.ModuleList([] )
# down
lowerCAmelCase__ : str = block_out_channels[0]
for i, down_block_type in enumerate(__UpperCAmelCase ):
lowerCAmelCase__ : Any = output_channel
lowerCAmelCase__ : str = block_out_channels[i]
lowerCAmelCase__ : Union[str, Any] = i == len(__UpperCAmelCase ) - 1
lowerCAmelCase__ : Tuple = get_down_block(
__UpperCAmelCase ,num_layers=self.layers_per_block ,in_channels=__UpperCAmelCase ,out_channels=__UpperCAmelCase ,add_downsample=not is_final_block ,resnet_eps=1E-6 ,downsample_padding=0 ,resnet_act_fn=__UpperCAmelCase ,resnet_groups=__UpperCAmelCase ,attention_head_dim=__UpperCAmelCase ,temb_channels=__UpperCAmelCase ,)
self.down_blocks.append(__UpperCAmelCase )
# mid
lowerCAmelCase__ : Any = UNetMidBlockaD(
in_channels=block_out_channels[-1] ,resnet_eps=1E-6 ,resnet_act_fn=__UpperCAmelCase ,output_scale_factor=1 ,resnet_time_scale_shift="""default""" ,attention_head_dim=block_out_channels[-1] ,resnet_groups=__UpperCAmelCase ,temb_channels=__UpperCAmelCase ,)
# out
lowerCAmelCase__ : List[str] = nn.GroupNorm(num_channels=block_out_channels[-1] ,num_groups=__UpperCAmelCase ,eps=1E-6 )
lowerCAmelCase__ : List[str] = nn.SiLU()
lowerCAmelCase__ : int = 2 * out_channels if double_z else out_channels
lowerCAmelCase__ : Dict = nn.Convad(block_out_channels[-1] ,__UpperCAmelCase ,3 ,padding=1 )
lowerCAmelCase__ : Optional[int] = False
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Any = x
lowerCAmelCase__ : Optional[Any] = self.conv_in(__UpperCAmelCase )
if self.training and self.gradient_checkpointing:
def create_custom_forward(__UpperCAmelCase ):
def custom_forward(*__UpperCAmelCase ):
return module(*__UpperCAmelCase )
return custom_forward
# down
if is_torch_version(""">=""" ,"""1.11.0""" ):
for down_block in self.down_blocks:
lowerCAmelCase__ : Tuple = torch.utils.checkpoint.checkpoint(
create_custom_forward(__UpperCAmelCase ) ,__UpperCAmelCase ,use_reentrant=__UpperCAmelCase )
# middle
lowerCAmelCase__ : List[Any] = torch.utils.checkpoint.checkpoint(
create_custom_forward(self.mid_block ) ,__UpperCAmelCase ,use_reentrant=__UpperCAmelCase )
else:
for down_block in self.down_blocks:
lowerCAmelCase__ : Union[str, Any] = torch.utils.checkpoint.checkpoint(create_custom_forward(__UpperCAmelCase ) ,__UpperCAmelCase )
# middle
lowerCAmelCase__ : List[str] = torch.utils.checkpoint.checkpoint(create_custom_forward(self.mid_block ) ,__UpperCAmelCase )
else:
# down
for down_block in self.down_blocks:
lowerCAmelCase__ : Union[str, Any] = down_block(__UpperCAmelCase )
# middle
lowerCAmelCase__ : Optional[int] = self.mid_block(__UpperCAmelCase )
# post-process
lowerCAmelCase__ : str = self.conv_norm_out(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.conv_act(__UpperCAmelCase )
lowerCAmelCase__ : int = self.conv_out(__UpperCAmelCase )
return sample
class lowerCAmelCase_( nn.Module ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase=3 ,__UpperCAmelCase=3 ,__UpperCAmelCase=("UpDecoderBlock2D",) ,__UpperCAmelCase=(64,) ,__UpperCAmelCase=2 ,__UpperCAmelCase=32 ,__UpperCAmelCase="silu" ,__UpperCAmelCase="group" ,) -> List[Any]:
super().__init__()
lowerCAmelCase__ : Dict = layers_per_block
lowerCAmelCase__ : Any = nn.Convad(
__UpperCAmelCase ,block_out_channels[-1] ,kernel_size=3 ,stride=1 ,padding=1 ,)
lowerCAmelCase__ : Union[str, Any] = None
lowerCAmelCase__ : Dict = nn.ModuleList([] )
lowerCAmelCase__ : Tuple = in_channels if norm_type == """spatial""" else None
# mid
lowerCAmelCase__ : Optional[Any] = UNetMidBlockaD(
in_channels=block_out_channels[-1] ,resnet_eps=1E-6 ,resnet_act_fn=__UpperCAmelCase ,output_scale_factor=1 ,resnet_time_scale_shift="""default""" if norm_type == """group""" else norm_type ,attention_head_dim=block_out_channels[-1] ,resnet_groups=__UpperCAmelCase ,temb_channels=__UpperCAmelCase ,)
# up
lowerCAmelCase__ : Optional[int] = list(reversed(__UpperCAmelCase ) )
lowerCAmelCase__ : Tuple = reversed_block_out_channels[0]
for i, up_block_type in enumerate(__UpperCAmelCase ):
lowerCAmelCase__ : Optional[Any] = output_channel
lowerCAmelCase__ : int = reversed_block_out_channels[i]
lowerCAmelCase__ : Dict = i == len(__UpperCAmelCase ) - 1
lowerCAmelCase__ : Union[str, Any] = get_up_block(
__UpperCAmelCase ,num_layers=self.layers_per_block + 1 ,in_channels=__UpperCAmelCase ,out_channels=__UpperCAmelCase ,prev_output_channel=__UpperCAmelCase ,add_upsample=not is_final_block ,resnet_eps=1E-6 ,resnet_act_fn=__UpperCAmelCase ,resnet_groups=__UpperCAmelCase ,attention_head_dim=__UpperCAmelCase ,temb_channels=__UpperCAmelCase ,resnet_time_scale_shift=__UpperCAmelCase ,)
self.up_blocks.append(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = output_channel
# out
if norm_type == "spatial":
lowerCAmelCase__ : Optional[Any] = SpatialNorm(block_out_channels[0] ,__UpperCAmelCase )
else:
lowerCAmelCase__ : List[Any] = nn.GroupNorm(num_channels=block_out_channels[0] ,num_groups=__UpperCAmelCase ,eps=1E-6 )
lowerCAmelCase__ : int = nn.SiLU()
lowerCAmelCase__ : Dict = nn.Convad(block_out_channels[0] ,__UpperCAmelCase ,3 ,padding=1 )
lowerCAmelCase__ : Optional[Any] = False
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ) -> Optional[Any]:
lowerCAmelCase__ : Optional[Any] = z
lowerCAmelCase__ : List[Any] = self.conv_in(__UpperCAmelCase )
lowerCAmelCase__ : Dict = next(iter(self.up_blocks.parameters() ) ).dtype
if self.training and self.gradient_checkpointing:
def create_custom_forward(__UpperCAmelCase ):
def custom_forward(*__UpperCAmelCase ):
return module(*__UpperCAmelCase )
return custom_forward
if is_torch_version(""">=""" ,"""1.11.0""" ):
# middle
lowerCAmelCase__ : Union[str, Any] = torch.utils.checkpoint.checkpoint(
create_custom_forward(self.mid_block ) ,__UpperCAmelCase ,__UpperCAmelCase ,use_reentrant=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = sample.to(__UpperCAmelCase )
# up
for up_block in self.up_blocks:
lowerCAmelCase__ : Optional[Any] = torch.utils.checkpoint.checkpoint(
create_custom_forward(__UpperCAmelCase ) ,__UpperCAmelCase ,__UpperCAmelCase ,use_reentrant=__UpperCAmelCase )
else:
# middle
lowerCAmelCase__ : int = torch.utils.checkpoint.checkpoint(
create_custom_forward(self.mid_block ) ,__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : str = sample.to(__UpperCAmelCase )
# up
for up_block in self.up_blocks:
lowerCAmelCase__ : List[Any] = torch.utils.checkpoint.checkpoint(create_custom_forward(__UpperCAmelCase ) ,__UpperCAmelCase ,__UpperCAmelCase )
else:
# middle
lowerCAmelCase__ : Any = self.mid_block(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : int = sample.to(__UpperCAmelCase )
# up
for up_block in self.up_blocks:
lowerCAmelCase__ : Dict = up_block(__UpperCAmelCase ,__UpperCAmelCase )
# post-process
if latent_embeds is None:
lowerCAmelCase__ : Union[str, Any] = self.conv_norm_out(__UpperCAmelCase )
else:
lowerCAmelCase__ : List[Any] = self.conv_norm_out(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.conv_act(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.conv_out(__UpperCAmelCase )
return sample
class lowerCAmelCase_( nn.Module ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase="random" ,__UpperCAmelCase=False ,__UpperCAmelCase=True ) -> List[Any]:
super().__init__()
lowerCAmelCase__ : List[str] = n_e
lowerCAmelCase__ : Dict = vq_embed_dim
lowerCAmelCase__ : List[Any] = beta
lowerCAmelCase__ : Union[str, Any] = legacy
lowerCAmelCase__ : Dict = nn.Embedding(self.n_e ,self.vq_embed_dim )
self.embedding.weight.data.uniform_(-1.0 / self.n_e ,1.0 / self.n_e )
lowerCAmelCase__ : int = remap
if self.remap is not None:
self.register_buffer("""used""" ,torch.tensor(np.load(self.remap ) ) )
lowerCAmelCase__ : List[Any] = self.used.shape[0]
lowerCAmelCase__ : Optional[Any] = unknown_index # "random" or "extra" or integer
if self.unknown_index == "extra":
lowerCAmelCase__ : Optional[int] = self.re_embed
lowerCAmelCase__ : Dict = self.re_embed + 1
print(
F"""Remapping {self.n_e} indices to {self.re_embed} indices. """
F"""Using {self.unknown_index} for unknown indices.""" )
else:
lowerCAmelCase__ : List[Any] = n_e
lowerCAmelCase__ : Tuple = sane_index_shape
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Union[str, Any] = inds.shape
assert len(__UpperCAmelCase ) > 1
lowerCAmelCase__ : Optional[int] = inds.reshape(ishape[0] ,-1 )
lowerCAmelCase__ : Any = self.used.to(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = (inds[:, :, None] == used[None, None, ...]).long()
lowerCAmelCase__ : List[Any] = match.argmax(-1 )
lowerCAmelCase__ : Optional[int] = match.sum(2 ) < 1
if self.unknown_index == "random":
lowerCAmelCase__ : int = torch.randint(0 ,self.re_embed ,size=new[unknown].shape ).to(device=new.device )
else:
lowerCAmelCase__ : Union[str, Any] = self.unknown_index
return new.reshape(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : Dict = inds.shape
assert len(__UpperCAmelCase ) > 1
lowerCAmelCase__ : List[str] = inds.reshape(ishape[0] ,-1 )
lowerCAmelCase__ : int = self.used.to(__UpperCAmelCase )
if self.re_embed > self.used.shape[0]: # extra token
lowerCAmelCase__ : Any = 0 # simply set to zero
lowerCAmelCase__ : Union[str, Any] = torch.gather(used[None, :][inds.shape[0] * [0], :] ,1 ,__UpperCAmelCase )
return back.reshape(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Union[str, Any]:
# reshape z -> (batch, height, width, channel) and flatten
lowerCAmelCase__ : Optional[int] = z.permute(0 ,2 ,3 ,1 ).contiguous()
lowerCAmelCase__ : Tuple = z.view(-1 ,self.vq_embed_dim )
# distances from z to embeddings e_j (z - e)^2 = z^2 + e^2 - 2 e * z
lowerCAmelCase__ : Optional[Any] = torch.argmin(torch.cdist(__UpperCAmelCase ,self.embedding.weight ) ,dim=1 )
lowerCAmelCase__ : Dict = self.embedding(__UpperCAmelCase ).view(z.shape )
lowerCAmelCase__ : Tuple = None
lowerCAmelCase__ : int = None
# compute loss for embedding
if not self.legacy:
lowerCAmelCase__ : List[Any] = self.beta * torch.mean((z_q.detach() - z) ** 2 ) + torch.mean((z_q - z.detach()) ** 2 )
else:
lowerCAmelCase__ : int = torch.mean((z_q.detach() - z) ** 2 ) + self.beta * torch.mean((z_q - z.detach()) ** 2 )
# preserve gradients
lowerCAmelCase__ : Optional[Any] = z + (z_q - z).detach()
# reshape back to match original input shape
lowerCAmelCase__ : List[Any] = z_q.permute(0 ,3 ,1 ,2 ).contiguous()
if self.remap is not None:
lowerCAmelCase__ : Tuple = min_encoding_indices.reshape(z.shape[0] ,-1 ) # add batch axis
lowerCAmelCase__ : int = self.remap_to_used(__UpperCAmelCase )
lowerCAmelCase__ : Dict = min_encoding_indices.reshape(-1 ,1 ) # flatten
if self.sane_index_shape:
lowerCAmelCase__ : Union[str, Any] = min_encoding_indices.reshape(z_q.shape[0] ,z_q.shape[2] ,z_q.shape[3] )
return z_q, loss, (perplexity, min_encodings, min_encoding_indices)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
# shape specifying (batch, height, width, channel)
if self.remap is not None:
lowerCAmelCase__ : str = indices.reshape(shape[0] ,-1 ) # add batch axis
lowerCAmelCase__ : Dict = self.unmap_to_all(__UpperCAmelCase )
lowerCAmelCase__ : Dict = indices.reshape(-1 ) # flatten again
# get quantized latent vectors
lowerCAmelCase__ : Optional[int] = self.embedding(__UpperCAmelCase )
if shape is not None:
lowerCAmelCase__ : str = z_q.view(__UpperCAmelCase )
# reshape back to match original input shape
lowerCAmelCase__ : Dict = z_q.permute(0 ,3 ,1 ,2 ).contiguous()
return z_q
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=False ) -> Dict:
lowerCAmelCase__ : int = parameters
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = torch.chunk(__UpperCAmelCase ,2 ,dim=1 )
lowerCAmelCase__ : Optional[Any] = torch.clamp(self.logvar ,-3_0.0 ,2_0.0 )
lowerCAmelCase__ : List[Any] = deterministic
lowerCAmelCase__ : str = torch.exp(0.5 * self.logvar )
lowerCAmelCase__ : str = torch.exp(self.logvar )
if self.deterministic:
lowerCAmelCase__ : Optional[int] = torch.zeros_like(
self.mean ,device=self.parameters.device ,dtype=self.parameters.dtype )
def UpperCAmelCase_ ( self ,__UpperCAmelCase = None ) -> torch.FloatTensor:
# make sure sample is on the same device as the parameters and has same dtype
lowerCAmelCase__ : str = randn_tensor(
self.mean.shape ,generator=__UpperCAmelCase ,device=self.parameters.device ,dtype=self.parameters.dtype )
lowerCAmelCase__ : List[str] = self.mean + self.std * sample
return x
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ) -> str:
if self.deterministic:
return torch.Tensor([0.0] )
else:
if other is None:
return 0.5 * torch.sum(torch.pow(self.mean ,2 ) + self.var - 1.0 - self.logvar ,dim=[1, 2, 3] )
else:
return 0.5 * torch.sum(
torch.pow(self.mean - other.mean ,2 ) / other.var
+ self.var / other.var
- 1.0
- self.logvar
+ other.logvar ,dim=[1, 2, 3] ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=[1, 2, 3] ) -> Optional[Any]:
if self.deterministic:
return torch.Tensor([0.0] )
lowerCAmelCase__ : str = np.log(2.0 * np.pi )
return 0.5 * torch.sum(logtwopi + self.logvar + torch.pow(sample - self.mean ,2 ) / self.var ,dim=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
return self.mean
| 37
|
'''simple docstring'''
import argparse
import json
import torch
from diffusers import DDPMScheduler, LDMPipeline, UNetaDModel, VQModel
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=1 ):
"""simple docstring"""
if n_shave_prefix_segments >= 0:
return ".".join(path.split(""".""" )[n_shave_prefix_segments:] )
else:
return ".".join(path.split(""".""" )[:n_shave_prefix_segments] )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = []
for old_item in old_list:
lowerCAmelCase__ : Optional[Any] = old_item.replace("""in_layers.0""" , """norm1""" )
lowerCAmelCase__ : Optional[int] = new_item.replace("""in_layers.2""" , """conv1""" )
lowerCAmelCase__ : Dict = new_item.replace("""out_layers.0""" , """norm2""" )
lowerCAmelCase__ : str = new_item.replace("""out_layers.3""" , """conv2""" )
lowerCAmelCase__ : str = new_item.replace("""emb_layers.1""" , """time_emb_proj""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""skip_connection""" , """conv_shortcut""" )
lowerCAmelCase__ : Union[str, Any] = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : int = []
for old_item in old_list:
lowerCAmelCase__ : List[str] = old_item
lowerCAmelCase__ : int = new_item.replace("""norm.weight""" , """group_norm.weight""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""norm.bias""" , """group_norm.bias""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""proj_out.weight""" , """proj_attn.weight""" )
lowerCAmelCase__ : int = new_item.replace("""proj_out.bias""" , """proj_attn.bias""" )
lowerCAmelCase__ : str = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=None ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "Paths should be a list of dicts containing 'old' and 'new' keys."
# Splits the attention layers into three variables.
if attention_paths_to_split is not None:
for path, path_map in attention_paths_to_split.items():
lowerCAmelCase__ : Any = old_checkpoint[path]
lowerCAmelCase__ : int = old_tensor.shape[0] // 3
lowerCAmelCase__ : int = (-1, channels) if len(old_tensor.shape ) == 3 else (-1)
lowerCAmelCase__ : Tuple = old_tensor.shape[0] // config["""num_head_channels"""] // 3
lowerCAmelCase__ : List[Any] = old_tensor.reshape((num_heads, 3 * channels // num_heads) + old_tensor.shape[1:] )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : str = old_tensor.split(channels // num_heads , dim=1 )
lowerCAmelCase__ : int = query.reshape(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = key.reshape(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = value.reshape(UpperCamelCase )
for path in paths:
lowerCAmelCase__ : Any = path["""new"""]
# These have already been assigned
if attention_paths_to_split is not None and new_path in attention_paths_to_split:
continue
# Global renaming happens here
lowerCAmelCase__ : Any = new_path.replace("""middle_block.0""" , """mid_block.resnets.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.1""" , """mid_block.attentions.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.2""" , """mid_block.resnets.1""" )
if additional_replacements is not None:
for replacement in additional_replacements:
lowerCAmelCase__ : Any = new_path.replace(replacement["""old"""] , replacement["""new"""] )
# proj_attn.weight has to be converted from conv 1D to linear
if "proj_attn.weight" in new_path:
lowerCAmelCase__ : List[Any] = old_checkpoint[path["""old"""]][:, :, 0]
else:
lowerCAmelCase__ : Dict = old_checkpoint[path["""old"""]]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = {}
lowerCAmelCase__ : str = checkpoint["""time_embed.0.weight"""]
lowerCAmelCase__ : List[Any] = checkpoint["""time_embed.0.bias"""]
lowerCAmelCase__ : int = checkpoint["""time_embed.2.weight"""]
lowerCAmelCase__ : List[str] = checkpoint["""time_embed.2.bias"""]
lowerCAmelCase__ : str = checkpoint["""input_blocks.0.0.weight"""]
lowerCAmelCase__ : Any = checkpoint["""input_blocks.0.0.bias"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.weight"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.bias"""]
lowerCAmelCase__ : str = checkpoint["""out.2.weight"""]
lowerCAmelCase__ : Tuple = checkpoint["""out.2.bias"""]
# Retrieves the keys for the input blocks only
lowerCAmelCase__ : Optional[Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """input_blocks""" in layer} )
lowerCAmelCase__ : Optional[Any] = {
layer_id: [key for key in checkpoint if f"""input_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the middle blocks only
lowerCAmelCase__ : Union[str, Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """middle_block""" in layer} )
lowerCAmelCase__ : Union[str, Any] = {
layer_id: [key for key in checkpoint if f"""middle_block.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the output blocks only
lowerCAmelCase__ : List[str] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """output_blocks""" in layer} )
lowerCAmelCase__ : List[Any] = {
layer_id: [key for key in checkpoint if f"""output_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
for i in range(1 , UpperCamelCase ):
lowerCAmelCase__ : Dict = (i - 1) // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Tuple = (i - 1) % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Optional[int] = [key for key in input_blocks[i] if f"""input_blocks.{i}.0""" in key]
lowerCAmelCase__ : Optional[Any] = [key for key in input_blocks[i] if f"""input_blocks.{i}.1""" in key]
if f"""input_blocks.{i}.0.op.weight""" in checkpoint:
lowerCAmelCase__ : Optional[int] = checkpoint[
f"""input_blocks.{i}.0.op.weight"""
]
lowerCAmelCase__ : Tuple = checkpoint[
f"""input_blocks.{i}.0.op.bias"""
]
continue
lowerCAmelCase__ : Optional[Any] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Dict = {"""old""": f"""input_blocks.{i}.0""", """new""": f"""down_blocks.{block_id}.resnets.{layer_in_block_id}"""}
lowerCAmelCase__ : Optional[Any] = {"""old""": """resnets.2.op""", """new""": """downsamplers.0.op"""}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path, resnet_op] , config=UpperCamelCase )
if len(UpperCamelCase ):
lowerCAmelCase__ : Optional[Any] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""input_blocks.{i}.1""",
"""new""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : List[str] = {
f"""input_blocks.{i}.1.qkv.bias""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""input_blocks.{i}.1.qkv.weight""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase , )
lowerCAmelCase__ : Dict = middle_blocks[0]
lowerCAmelCase__ : Union[str, Any] = middle_blocks[1]
lowerCAmelCase__ : Dict = middle_blocks[2]
lowerCAmelCase__ : Any = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Dict = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {
"""middle_block.1.qkv.bias""": {
"""key""": """mid_block.attentions.0.key.bias""",
"""query""": """mid_block.attentions.0.query.bias""",
"""value""": """mid_block.attentions.0.value.bias""",
},
"""middle_block.1.qkv.weight""": {
"""key""": """mid_block.attentions.0.key.weight""",
"""query""": """mid_block.attentions.0.query.weight""",
"""value""": """mid_block.attentions.0.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase )
for i in range(UpperCamelCase ):
lowerCAmelCase__ : Tuple = i // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : List[str] = i % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : int = [shave_segments(UpperCamelCase , 2 ) for name in output_blocks[i]]
lowerCAmelCase__ : Union[str, Any] = {}
for layer in output_block_layers:
lowerCAmelCase__ , lowerCAmelCase__ : Any = layer.split(""".""" )[0], shave_segments(UpperCamelCase , 1 )
if layer_id in output_block_list:
output_block_list[layer_id].append(UpperCamelCase )
else:
lowerCAmelCase__ : str = [layer_name]
if len(UpperCamelCase ) > 1:
lowerCAmelCase__ : str = [key for key in output_blocks[i] if f"""output_blocks.{i}.0""" in key]
lowerCAmelCase__ : Dict = [key for key in output_blocks[i] if f"""output_blocks.{i}.1""" in key]
lowerCAmelCase__ : Optional[int] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {"""old""": f"""output_blocks.{i}.0""", """new""": f"""up_blocks.{block_id}.resnets.{layer_in_block_id}"""}
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , config=UpperCamelCase )
if ["conv.weight", "conv.bias"] in output_block_list.values():
lowerCAmelCase__ : List[Any] = list(output_block_list.values() ).index(["""conv.weight""", """conv.bias"""] )
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.weight"""
]
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.bias"""
]
# Clear attentions as they have been attributed above.
if len(UpperCamelCase ) == 2:
lowerCAmelCase__ : Tuple = []
if len(UpperCamelCase ):
lowerCAmelCase__ : Dict = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""output_blocks.{i}.1""",
"""new""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : Tuple = {
f"""output_blocks.{i}.1.qkv.bias""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""output_blocks.{i}.1.qkv.weight""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=to_split if any("""qkv""" in key for key in attentions ) else None , config=UpperCamelCase , )
else:
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase , n_shave_prefix_segments=1 )
for path in resnet_0_paths:
lowerCAmelCase__ : Tuple = """.""".join(["""output_blocks""", str(UpperCamelCase ), path["""old"""]] )
lowerCAmelCase__ : List[Any] = """.""".join(["""up_blocks""", str(UpperCamelCase ), """resnets""", str(UpperCamelCase ), path["""new"""]] )
lowerCAmelCase__ : Union[str, Any] = checkpoint[old_path]
return new_checkpoint
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the checkpoint to convert.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help='''The config json file corresponding to the architecture.''',
)
parser.add_argument('''--dump_path''', default=None, type=str, required=True, help='''Path to the output model.''')
_lowerCAmelCase = parser.parse_args()
_lowerCAmelCase = torch.load(args.checkpoint_path)
with open(args.config_file) as f:
_lowerCAmelCase = json.loads(f.read())
_lowerCAmelCase = convert_ldm_checkpoint(checkpoint, config)
if "ldm" in config:
del config["ldm"]
_lowerCAmelCase = UNetaDModel(**config)
model.load_state_dict(converted_checkpoint)
try:
_lowerCAmelCase = DDPMScheduler.from_config('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = VQModel.from_pretrained('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = LDMPipeline(unet=model, scheduler=scheduler, vae=vqvae)
pipe.save_pretrained(args.dump_path)
except: # noqa: E722
model.save_pretrained(args.dump_path)
| 37
| 1
|
'''simple docstring'''
from __future__ import annotations
from collections.abc import Generator
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : dict[int, int] = {}
lowerCAmelCase__ : List[str] = 2
while True:
lowerCAmelCase__ : int = factor_map.pop(UpperCamelCase , UpperCamelCase )
if factor:
lowerCAmelCase__ : Dict = factor + prime
while x in factor_map:
x += factor
lowerCAmelCase__ : str = factor
else:
lowerCAmelCase__ : Dict = prime
yield prime
prime += 1
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = 1e10 ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = sieve()
lowerCAmelCase__ : Union[str, Any] = 1
while True:
lowerCAmelCase__ : List[Any] = next(UpperCamelCase )
if (2 * prime * n) > limit:
return n
# Ignore the next prime as the reminder will be 2.
next(UpperCamelCase )
n += 2
if __name__ == "__main__":
print(solution())
| 37
|
'''simple docstring'''
from math import sqrt
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' must been an int and positive"
lowerCAmelCase__ : int = True
# 0 and 1 are none primes.
if number <= 1:
lowerCAmelCase__ : Optional[Any] = False
for divisor in range(2 , int(round(sqrt(UpperCamelCase ) ) ) + 1 ):
# if 'number' divisible by 'divisor' then sets 'status'
# of false and break up the loop.
if number % divisor == 0:
lowerCAmelCase__ : Any = False
break
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'status' must been from type bool"
return status
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
# beginList: contains all natural numbers from 2 up to N
lowerCAmelCase__ : List[str] = list(range(2 , n + 1 ) )
lowerCAmelCase__ : str = [] # this list will be returns.
# actual sieve of erathostenes
for i in range(len(UpperCamelCase ) ):
for j in range(i + 1 , len(UpperCamelCase ) ):
if (begin_list[i] != 0) and (begin_list[j] % begin_list[i] == 0):
lowerCAmelCase__ : List[Any] = 0
# filters actual prime numbers.
lowerCAmelCase__ : List[Any] = [x for x in begin_list if x != 0]
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
lowerCAmelCase__ : List[str] = []
# iterates over all numbers between 2 up to N+1
# if a number is prime then appends to list 'ans'
for number in range(2 , n + 1 ):
if is_prime(UpperCamelCase ):
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and number >= 0, "'number' must been an int and >= 0"
lowerCAmelCase__ : Optional[Any] = [] # this list will be returns of the function.
# potential prime number factors.
lowerCAmelCase__ : Dict = 2
lowerCAmelCase__ : Dict = number
if number == 0 or number == 1:
ans.append(UpperCamelCase )
# if 'number' not prime then builds the prime factorization of 'number'
elif not is_prime(UpperCamelCase ):
while quotient != 1:
if is_prime(UpperCamelCase ) and (quotient % factor == 0):
ans.append(UpperCamelCase )
quotient /= factor
else:
factor += 1
else:
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : Optional[int] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = max(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = min(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 == 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 == 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 != 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 != 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase ) and (number > 2) and is_even(UpperCamelCase )
), "'number' must been an int, even and > 2"
lowerCAmelCase__ : Dict = [] # this list will returned
# creates a list of prime numbers between 2 up to 'number'
lowerCAmelCase__ : Dict = get_prime_numbers(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = len(UpperCamelCase )
# run variable for while-loops.
lowerCAmelCase__ : List[str] = 0
lowerCAmelCase__ : List[Any] = None
# exit variable. for break up the loops
lowerCAmelCase__ : Any = True
while i < len_pn and loop:
lowerCAmelCase__ : List[Any] = i + 1
while j < len_pn and loop:
if prime_numbers[i] + prime_numbers[j] == number:
lowerCAmelCase__ : Optional[Any] = False
ans.append(prime_numbers[i] )
ans.append(prime_numbers[j] )
j += 1
i += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (len(UpperCamelCase ) == 2)
and (ans[0] + ans[1] == number)
and is_prime(ans[0] )
and is_prime(ans[1] )
), "'ans' must contains two primes. And sum of elements must been eq 'number'"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 0)
and (numbera >= 0)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 0
while numbera != 0:
lowerCAmelCase__ : Any = numbera % numbera
lowerCAmelCase__ : str = numbera
lowerCAmelCase__ : List[str] = rest
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
numbera >= 0
), "'number' must been from type int and positive"
return numbera
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 1)
and (numbera >= 1)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 1 # actual answer that will be return.
# for kgV (x,1)
if numbera > 1 and numbera > 1:
# builds the prime factorization of 'number1' and 'number2'
lowerCAmelCase__ : int = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = prime_factorization(UpperCamelCase )
elif numbera == 1 or numbera == 1:
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : Dict = []
lowerCAmelCase__ : List[str] = max(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = 0
lowerCAmelCase__ : str = 0
lowerCAmelCase__ : List[Any] = [] # captured numbers int both 'primeFac1' and 'primeFac2'
# iterates through primeFac1
for n in prime_fac_a:
if n not in done:
if n in prime_fac_a:
lowerCAmelCase__ : int = prime_fac_a.count(UpperCamelCase )
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(max(UpperCamelCase , UpperCamelCase ) ):
ans *= n
else:
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# iterates through primeFac2
for n in prime_fac_a:
if n not in done:
lowerCAmelCase__ : Optional[int] = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
ans >= 0
), "'ans' must been from type int and positive"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'number' must been a positive int"
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Tuple = 2 # this variable holds the answer
while index < n:
index += 1
ans += 1 # counts to the next number
# if ans not prime then
# runs to the next prime number.
while not is_prime(UpperCamelCase ):
ans += 1
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and is_prime(
UpperCamelCase ), "'ans' must been a prime number and from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
is_prime(UpperCamelCase ) and is_prime(UpperCamelCase ) and (p_number_a < p_number_a)
), "The arguments must been prime numbers and 'pNumber1' < 'pNumber2'"
lowerCAmelCase__ : Dict = p_number_a + 1 # jump to the next number
lowerCAmelCase__ : List[Any] = [] # this list will be returns.
# if number is not prime then
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
while number < p_number_a:
ans.append(UpperCamelCase )
number += 1
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and ans[0] != p_number_a
and ans[len(UpperCamelCase ) - 1] != p_number_a
), "'ans' must been a list without the arguments"
# 'ans' contains not 'pNumber1' and 'pNumber2' !
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 1), "'n' must been int and >= 1"
lowerCAmelCase__ : List[Any] = [] # will be returned.
for divisor in range(1 , n + 1 ):
if n % divisor == 0:
ans.append(UpperCamelCase )
# precondition
assert ans[0] == 1 and ans[len(UpperCamelCase ) - 1] == n, "Error in function getDivisiors(...)"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number > 1
), "'number' must been an int and >= 1"
lowerCAmelCase__ : Optional[int] = get_divisors(UpperCamelCase )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (divisors[0] == 1)
and (divisors[len(UpperCamelCase ) - 1] == number)
), "Error in help-function getDivisiors(...)"
# summed all divisors up to 'number' (exclusive), hence [:-1]
return sum(divisors[:-1] ) == number
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (denominator != 0)
), "The arguments must been from type int and 'denominator' != 0"
# build the greatest common divisor of numerator and denominator.
lowerCAmelCase__ : int = gcd(abs(UpperCamelCase ) , abs(UpperCamelCase ) )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (numerator % gcd_of_fraction == 0)
and (denominator % gcd_of_fraction == 0)
), "Error in function gcd(...,...)"
return (numerator // gcd_of_fraction, denominator // gcd_of_fraction)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been a int and >= 0"
lowerCAmelCase__ : str = 1 # this will be return.
for factor in range(1 , n + 1 ):
ans *= factor
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
lowerCAmelCase__ : Any = 1
lowerCAmelCase__ : Optional[Any] = 1 # this will be return
for _ in range(n - 1 ):
lowerCAmelCase__ : Dict = ans
ans += fiba
lowerCAmelCase__ : str = tmp
return ans
| 37
| 1
|
'''simple docstring'''
from typing import Any, Dict, List, Optional, Tuple, Union
import torch
from torch import nn
from torch.utils.data import DistributedSampler, RandomSampler
from transformers import PreTrainedModel, Trainer, logging
from transformers.integrations import is_fairscale_available
from transformers.models.fsmt.configuration_fsmt import FSMTConfig
from transformers.optimization import (
Adafactor,
AdamW,
get_constant_schedule,
get_constant_schedule_with_warmup,
get_cosine_schedule_with_warmup,
get_cosine_with_hard_restarts_schedule_with_warmup,
get_linear_schedule_with_warmup,
get_polynomial_decay_schedule_with_warmup,
)
from transformers.trainer_pt_utils import get_tpu_sampler
from transformers.training_args import ParallelMode
from transformers.utils import is_torch_tpu_available
if is_fairscale_available():
from fairscale.optim import OSS
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''linear''': get_linear_schedule_with_warmup,
'''cosine''': get_cosine_schedule_with_warmup,
'''cosine_w_restarts''': get_cosine_with_hard_restarts_schedule_with_warmup,
'''polynomial''': get_polynomial_decay_schedule_with_warmup,
'''constant''': get_constant_schedule,
'''constant_w_warmup''': get_constant_schedule_with_warmup,
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> List[Any]:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
if config is None:
assert isinstance(self.model ,__UpperCAmelCase ), (
"If no `config` is passed the model to be trained has to be of type `PreTrainedModel`, but is"
F""" {self.model.__class__}"""
)
lowerCAmelCase__ : str = self.model.config
else:
lowerCAmelCase__ : Dict = config
lowerCAmelCase__ : Dict = data_args
lowerCAmelCase__ : str = self.config.tgt_vocab_size if isinstance(self.config ,__UpperCAmelCase ) else self.config.vocab_size
if self.args.label_smoothing != 0 or (self.data_args is not None and self.data_args.ignore_pad_token_for_loss):
assert self.config.pad_token_id is not None, (
"Make sure that `config.pad_token_id` is correcly defined when ignoring `pad_token` for loss"
" calculation or doing label smoothing."
)
if self.config.pad_token_id is None and self.config.eos_token_id is not None:
logger.warning(
F"""The `config.pad_token_id` is `None`. Using `config.eos_token_id` = {self.config.eos_token_id} for"""
""" padding..""" )
if self.args.label_smoothing == 0:
lowerCAmelCase__ : Optional[int] = torch.nn.CrossEntropyLoss(ignore_index=self.config.pad_token_id )
else:
# dynamically import label_smoothed_nll_loss
from utils import label_smoothed_nll_loss
lowerCAmelCase__ : Any = label_smoothed_nll_loss
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
if self.optimizer is None:
lowerCAmelCase__ : str = ["""bias""", """LayerNorm.weight"""]
lowerCAmelCase__ : Any = [
{
"""params""": [p for n, p in self.model.named_parameters() if not any(nd in n for nd in no_decay )],
"""weight_decay""": self.args.weight_decay,
},
{
"""params""": [p for n, p in self.model.named_parameters() if any(nd in n for nd in no_decay )],
"""weight_decay""": 0.0,
},
]
lowerCAmelCase__ : Tuple = Adafactor if self.args.adafactor else AdamW
if self.args.adafactor:
lowerCAmelCase__ : Union[str, Any] = Adafactor
lowerCAmelCase__ : int = {"""scale_parameter""": False, """relative_step""": False}
else:
lowerCAmelCase__ : List[Any] = AdamW
lowerCAmelCase__ : Dict = {
"""betas""": (self.args.adam_betaa, self.args.adam_betaa),
"""eps""": self.args.adam_epsilon,
}
lowerCAmelCase__ : Tuple = self.args.learning_rate
if self.sharded_ddp:
lowerCAmelCase__ : Optional[int] = OSS(
params=__UpperCAmelCase ,optim=__UpperCAmelCase ,**__UpperCAmelCase ,)
else:
lowerCAmelCase__ : Union[str, Any] = optimizer_cls(__UpperCAmelCase ,**__UpperCAmelCase )
if self.lr_scheduler is None:
lowerCAmelCase__ : int = self._get_lr_scheduler(__UpperCAmelCase )
else: # ignoring --lr_scheduler
logger.warning("""scheduler is passed to `Seq2SeqTrainer`, `--lr_scheduler` arg is ignored.""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : int = arg_to_scheduler[self.args.lr_scheduler]
if self.args.lr_scheduler == "constant":
lowerCAmelCase__ : str = schedule_func(self.optimizer )
elif self.args.lr_scheduler == "constant_w_warmup":
lowerCAmelCase__ : Union[str, Any] = schedule_func(self.optimizer ,num_warmup_steps=self.args.warmup_steps )
else:
lowerCAmelCase__ : str = schedule_func(
self.optimizer ,num_warmup_steps=self.args.warmup_steps ,num_training_steps=__UpperCAmelCase )
return scheduler
def UpperCAmelCase_ ( self ) -> Optional[torch.utils.data.Sampler]:
if isinstance(self.train_dataset ,torch.utils.data.IterableDataset ):
return None
elif is_torch_tpu_available():
return get_tpu_sampler(self.train_dataset )
else:
if self.args.sortish_sampler:
self.train_dataset.make_sortish_sampler(
self.args.per_device_train_batch_size ,distributed=(self.args.parallel_mode == ParallelMode.DISTRIBUTED) ,)
return (
RandomSampler(self.train_dataset )
if self.args.local_rank == -1
else DistributedSampler(self.train_dataset )
)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> List[str]:
if self.args.label_smoothing == 0:
if self.data_args is not None and self.data_args.ignore_pad_token_for_loss:
# force training to ignore pad token
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase ,use_cache=__UpperCAmelCase )[0]
lowerCAmelCase__ : Dict = self.loss_fn(logits.view(-1 ,logits.shape[-1] ) ,labels.view(-1 ) )
else:
# compute usual loss via models
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase ,labels=__UpperCAmelCase ,use_cache=__UpperCAmelCase )[:2]
else:
# compute label smoothed loss
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase ,use_cache=__UpperCAmelCase )[0]
lowerCAmelCase__ : Tuple = torch.nn.functional.log_softmax(__UpperCAmelCase ,dim=-1 )
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = self.loss_fn(__UpperCAmelCase ,__UpperCAmelCase ,self.args.label_smoothing ,ignore_index=self.config.pad_token_id )
return loss, logits
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Optional[int] = inputs.pop("""labels""" )
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self._compute_loss(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase )
return loss
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase = None ,) -> Tuple[Optional[float], Optional[torch.Tensor], Optional[torch.Tensor]]:
lowerCAmelCase__ : Optional[Any] = self._prepare_inputs(__UpperCAmelCase )
lowerCAmelCase__ : Any = {
"""max_length""": self.data_args.val_max_target_length
if self.data_args is not None
else self.config.max_length,
"""num_beams""": self.data_args.eval_beams if self.data_args is not None else self.config.num_beams,
}
if self.args.predict_with_generate and not self.args.prediction_loss_only:
lowerCAmelCase__ : Tuple = self.model.generate(
inputs["""input_ids"""] ,attention_mask=inputs["""attention_mask"""] ,**__UpperCAmelCase ,)
# in case the batch is shorter than max length, the output should be padded
if generated_tokens.shape[-1] < gen_kwargs["max_length"]:
lowerCAmelCase__ : int = self._pad_tensors_to_max_len(__UpperCAmelCase ,gen_kwargs["""max_length"""] )
lowerCAmelCase__ : Union[str, Any] = inputs.pop("""labels""" )
with torch.no_grad():
# compute loss on predict data
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self._compute_loss(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : Any = loss.mean().detach()
if self.args.prediction_loss_only:
return (loss, None, None)
lowerCAmelCase__ : Union[str, Any] = generated_tokens if self.args.predict_with_generate else logits
if labels.shape[-1] < gen_kwargs["max_length"]:
lowerCAmelCase__ : Optional[int] = self._pad_tensors_to_max_len(__UpperCAmelCase ,gen_kwargs["""max_length"""] )
return (loss, logits, labels)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
# If PAD token is not defined at least EOS token has to be defined
lowerCAmelCase__ : Dict = self.config.pad_token_id if self.config.pad_token_id is not None else self.config.eos_token_id
if pad_token_id is None:
raise ValueError(
"""Make sure that either `config.pad_token_id` or `config.eos_token_id` is defined if tensor has to be"""
F""" padded to `max_length`={max_length}""" )
lowerCAmelCase__ : int = pad_token_id * torch.ones(
(tensor.shape[0], max_length) ,dtype=tensor.dtype ,device=tensor.device )
lowerCAmelCase__ : Optional[int] = tensor
return padded_tensor
| 37
|
'''simple docstring'''
from sklearn.metrics import fa_score, matthews_corrcoef
import datasets
from .record_evaluation import evaluate as evaluate_record
_lowerCAmelCase = '''\
@article{wang2019superglue,
title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
journal={arXiv preprint arXiv:1905.00537},
year={2019}
}
'''
_lowerCAmelCase = '''\
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
GLUE with a new set of more difficult language understanding tasks, improved
resources, and a new public leaderboard.
'''
_lowerCAmelCase = '''
Compute SuperGLUE evaluation metric associated to each SuperGLUE dataset.
Args:
predictions: list of predictions to score. Depending on the SuperGlUE subset:
- for \'record\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'prediction_text\': the predicted answer text
- for \'multirc\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question-answer pair as specified by the dataset
- \'prediction\': the predicted answer label
- otherwise: list of predicted labels
references: list of reference labels. Depending on the SuperGLUE subset:
- for \'record\': list of question-answers dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'answers\': list of possible answers
- otherwise: list of reference labels
Returns: depending on the SuperGLUE subset:
- for \'record\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1\': F1 score
- for \'multirc\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1_m\': Per-question macro-F1 score
- \'f1_a\': Average F1 score over all answers
- for \'axb\':
\'matthews_correlation\': Matthew Correlation
- for \'cb\':
- \'accuracy\': Accuracy
- \'f1\': F1 score
- for all others:
- \'accuracy\': Accuracy
Examples:
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'copa\') # any of ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'cb\')
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'record\')
>>> predictions = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'prediction_text\': \'answer\'}]
>>> references = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'answers\': [\'answer\', \'another_answer\']}]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'multirc\')
>>> predictions = [{\'idx\': {\'answer\': 0, \'paragraph\': 0, \'question\': 0}, \'prediction\': 0}, {\'idx\': {\'answer\': 1, \'paragraph\': 2, \'question\': 3}, \'prediction\': 1}]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1_m\': 1.0, \'f1_a\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'axb\')
>>> references = [0, 1]
>>> predictions = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'matthews_correlation\': 1.0}
'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return float((preds == labels).mean() )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase="binary" ):
"""simple docstring"""
lowerCAmelCase__ : Any = simple_accuracy(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = float(fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average=UpperCamelCase ) )
return {
"accuracy": acc,
"f1": fa,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
for id_pred, label in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : str = f"""{id_pred['idx']['paragraph']}-{id_pred['idx']['question']}"""
lowerCAmelCase__ : Dict = id_pred["""prediction"""]
if question_id in question_map:
question_map[question_id].append((pred, label) )
else:
lowerCAmelCase__ : Optional[int] = [(pred, label)]
lowerCAmelCase__ , lowerCAmelCase__ : int = [], []
for question, preds_labels in question_map.items():
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = zip(*UpperCamelCase )
lowerCAmelCase__ : List[Any] = fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average="""macro""" )
fas.append(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = int(sum(pred == label for pred, label in preds_labels ) == len(UpperCamelCase ) )
ems.append(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = float(sum(UpperCamelCase ) / len(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = sum(UpperCamelCase ) / len(UpperCamelCase )
lowerCAmelCase__ : Dict = float(fa_score(y_true=UpperCamelCase , y_pred=[id_pred["""prediction"""] for id_pred in ids_preds] ) )
return {"exact_match": em, "f1_m": fa_m, "f1_a": fa_a}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.config_name not in [
"boolq",
"cb",
"copa",
"multirc",
"record",
"rte",
"wic",
"wsc",
"wsc.fixed",
"axb",
"axg",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(self._get_feature_types() ) ,codebase_urls=[] ,reference_urls=[] ,format="""numpy""" if not self.config_name == """record""" and not self.config_name == """multirc""" else None ,)
def UpperCAmelCase_ ( self ) -> str:
if self.config_name == "record":
return {
"predictions": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"prediction_text": datasets.Value("""string""" ),
},
"references": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"answers": datasets.Sequence(datasets.Value("""string""" ) ),
},
}
elif self.config_name == "multirc":
return {
"predictions": {
"idx": {
"answer": datasets.Value("""int64""" ),
"paragraph": datasets.Value("""int64""" ),
"question": datasets.Value("""int64""" ),
},
"prediction": datasets.Value("""int64""" ),
},
"references": datasets.Value("""int64""" ),
}
else:
return {
"predictions": datasets.Value("""int64""" ),
"references": datasets.Value("""int64""" ),
}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
if self.config_name == "axb":
return {"matthews_correlation": matthews_corrcoef(__UpperCAmelCase ,__UpperCAmelCase )}
elif self.config_name == "cb":
return acc_and_fa(__UpperCAmelCase ,__UpperCAmelCase ,fa_avg="""macro""" )
elif self.config_name == "record":
lowerCAmelCase__ : Optional[Any] = [
{
"""qas""": [
{"""id""": ref["""idx"""]["""query"""], """answers""": [{"""text""": ans} for ans in ref["""answers"""]]}
for ref in references
]
}
]
lowerCAmelCase__ : Union[str, Any] = {pred["""idx"""]["""query"""]: pred["""prediction_text"""] for pred in predictions}
return evaluate_record(__UpperCAmelCase ,__UpperCAmelCase )[0]
elif self.config_name == "multirc":
return evaluate_multirc(__UpperCAmelCase ,__UpperCAmelCase )
elif self.config_name in ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]:
return {"accuracy": simple_accuracy(__UpperCAmelCase ,__UpperCAmelCase )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
| 37
| 1
|
'''simple docstring'''
import logging
from dataclasses import dataclass, field
from typing import Optional
from seqaseq_trainer import arg_to_scheduler
from transformers import TrainingArguments
_lowerCAmelCase = logging.getLogger(__name__)
@dataclass
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[float] = field(
default=0.0 , metadata={'''help''': '''The label smoothing epsilon to apply (if not zero).'''} )
__lowercase : bool = field(default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Whether to SortishSamler or not.'''} )
__lowercase : bool = field(
default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Whether to use generate to calculate generative metrics (ROUGE, BLEU).'''} )
__lowercase : bool = field(default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''whether to use adafactor'''} )
__lowercase : Optional[float] = field(
default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Encoder layer dropout probability. Goes into model.config.'''} )
__lowercase : Optional[float] = field(
default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Decoder layer dropout probability. Goes into model.config.'''} )
__lowercase : Optional[float] = field(default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Dropout probability. Goes into model.config.'''} )
__lowercase : Optional[float] = field(
default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Attention dropout probability. Goes into model.config.'''} )
__lowercase : Optional[str] = field(
default='''linear''' , metadata={'''help''': f"""Which lr scheduler to use. Selected in {sorted(arg_to_scheduler.keys() )}"""} , )
| 37
|
'''simple docstring'''
import unittest
from pathlib import Path
from tempfile import NamedTemporaryFile, TemporaryDirectory
from transformers import BertConfig, BertTokenizerFast, FeatureExtractionPipeline
from transformers.convert_graph_to_onnx import (
convert,
ensure_valid_input,
generate_identified_filename,
infer_shapes,
quantize,
)
from transformers.testing_utils import require_tf, require_tokenizers, require_torch, slow
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[Any]:
return None
class lowerCAmelCase_:
'''simple docstring'''
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
return None
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
__lowercase : Dict = [
# (model_name, model_kwargs)
('''bert-base-cased''', {}),
('''gpt2''', {'''use_cache''': False}), # We don't support exporting GPT2 past keys anymore
]
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> int:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> Any:
from transformers import BertModel
lowerCAmelCase__ : Optional[int] = ["""[UNK]""", """[SEP]""", """[CLS]""", """[PAD]""", """[MASK]""", """some""", """other""", """words"""]
with NamedTemporaryFile(mode="""w+t""" ) as vocab_file:
vocab_file.write("""\n""".join(__UpperCAmelCase ) )
vocab_file.flush()
lowerCAmelCase__ : Dict = BertTokenizerFast(vocab_file.name )
with TemporaryDirectory() as bert_save_dir:
lowerCAmelCase__ : Tuple = BertModel(BertConfig(vocab_size=len(__UpperCAmelCase ) ) )
model.save_pretrained(__UpperCAmelCase )
self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,__UpperCAmelCase )
@require_tf
@slow
def UpperCAmelCase_ ( self ) -> List[str]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Dict = self._test_export(__UpperCAmelCase ,"""tf""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = quantize(Path(__UpperCAmelCase ) )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
@require_torch
@slow
def UpperCAmelCase_ ( self ) -> List[Any]:
for model, model_kwargs in OnnxExportTestCase.MODEL_TO_TEST:
lowerCAmelCase__ : Any = self._test_export(__UpperCAmelCase ,"""pt""" ,12 ,**__UpperCAmelCase )
lowerCAmelCase__ : Dict = quantize(__UpperCAmelCase )
# Ensure the actual quantized model is not bigger than the original one
if quantized_path.stat().st_size >= Path(__UpperCAmelCase ).stat().st_size:
self.fail("""Quantized model is bigger than initial ONNX model""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[Any]:
try:
# Compute path
with TemporaryDirectory() as tempdir:
lowerCAmelCase__ : Optional[int] = Path(__UpperCAmelCase ).joinpath("""model.onnx""" )
# Remove folder if exists
if path.parent.exists():
path.parent.rmdir()
# Export
convert(__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,**__UpperCAmelCase )
return path
except Exception as e:
self.fail(__UpperCAmelCase )
@require_torch
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
from transformers import BertModel
lowerCAmelCase__ : List[Any] = BertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Union[str, Any] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""pt""" )
@require_tf
@require_tokenizers
@slow
def UpperCAmelCase_ ( self ) -> Optional[int]:
from transformers import TFBertModel
lowerCAmelCase__ : int = TFBertModel(BertConfig.from_pretrained("""lysandre/tiny-bert-random""" ) )
lowerCAmelCase__ : Optional[int] = BertTokenizerFast.from_pretrained("""lysandre/tiny-bert-random""" )
self._test_infer_dynamic_axis(__UpperCAmelCase ,__UpperCAmelCase ,"""tf""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = FeatureExtractionPipeline(__UpperCAmelCase ,__UpperCAmelCase )
lowerCAmelCase__ : List[str] = ["""input_ids""", """token_type_ids""", """attention_mask""", """output_0""", """output_1"""]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = infer_shapes(__UpperCAmelCase ,__UpperCAmelCase )
# Assert all variables are present
self.assertEqual(len(__UpperCAmelCase ) ,len(__UpperCAmelCase ) )
self.assertTrue(all(var_name in shapes for var_name in variable_names ) )
self.assertSequenceEqual(variable_names[:3] ,__UpperCAmelCase )
self.assertSequenceEqual(variable_names[3:] ,__UpperCAmelCase )
# Assert inputs are {0: batch, 1: sequence}
for var_name in ["input_ids", "token_type_ids", "attention_mask"]:
self.assertDictEqual(shapes[var_name] ,{0: """batch""", 1: """sequence"""} )
# Assert outputs are {0: batch, 1: sequence} and {0: batch}
self.assertDictEqual(shapes["""output_0"""] ,{0: """batch""", 1: """sequence"""} )
self.assertDictEqual(shapes["""output_1"""] ,{0: """batch"""} )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = ["""input_ids""", """attention_mask""", """token_type_ids"""]
lowerCAmelCase__ : Union[str, Any] = {"""input_ids""": [1, 2, 3, 4], """attention_mask""": [0, 0, 0, 0], """token_type_ids""": [1, 1, 1, 1]}
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = ensure_valid_input(FuncContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the same number of args (all are valid)
self.assertEqual(len(__UpperCAmelCase ) ,3 )
# Should have exactly the same input names
self.assertEqual(set(__UpperCAmelCase ) ,set(__UpperCAmelCase ) )
# Parameter should be reordered according to their respective place in the function:
# (input_ids, token_type_ids, attention_mask)
self.assertEqual(__UpperCAmelCase ,(tokens["""input_ids"""], tokens["""token_type_ids"""], tokens["""attention_mask"""]) )
# Generated args are interleaved with another args (for instance parameter "past" in GPT2)
lowerCAmelCase__ , lowerCAmelCase__ : int = ensure_valid_input(FuncNonContiguousArgs() ,__UpperCAmelCase ,__UpperCAmelCase )
# Should have exactly the one arg (all before the one not provided "some_other_args")
self.assertEqual(len(__UpperCAmelCase ) ,1 )
self.assertEqual(len(__UpperCAmelCase ) ,1 )
# Should have only "input_ids"
self.assertEqual(inputs_args[0] ,tokens["""input_ids"""] )
self.assertEqual(ordered_input_names[0] ,"""input_ids""" )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Dict = generate_identified_filename(Path("""/home/something/my_fake_model.onnx""" ) ,"""-test""" )
self.assertEqual("""/home/something/my_fake_model-test.onnx""" ,generated.as_posix() )
| 37
| 1
|
'''simple docstring'''
from ....configuration_utils import PretrainedConfig
from ....utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''speechbrain/m-ctc-t-large''': '''https://huggingface.co/speechbrain/m-ctc-t-large/resolve/main/config.json''',
# See all M-CTC-T models at https://huggingface.co/models?filter=mctct
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[int] = '''mctct'''
def __init__( self ,__UpperCAmelCase=8065 ,__UpperCAmelCase=1536 ,__UpperCAmelCase=36 ,__UpperCAmelCase=6144 ,__UpperCAmelCase=4 ,__UpperCAmelCase=384 ,__UpperCAmelCase=920 ,__UpperCAmelCase=1E-5 ,__UpperCAmelCase=0.3 ,__UpperCAmelCase="relu" ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=0.3 ,__UpperCAmelCase=0.3 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.3 ,__UpperCAmelCase=1 ,__UpperCAmelCase=(7,) ,__UpperCAmelCase=(3,) ,__UpperCAmelCase=80 ,__UpperCAmelCase=1 ,__UpperCAmelCase=None ,__UpperCAmelCase="sum" ,__UpperCAmelCase=False ,**__UpperCAmelCase ,) -> Tuple:
super().__init__(**__UpperCAmelCase ,pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase )
lowerCAmelCase__ : str = vocab_size
lowerCAmelCase__ : Union[str, Any] = hidden_size
lowerCAmelCase__ : Union[str, Any] = num_hidden_layers
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : str = attention_head_dim
lowerCAmelCase__ : Dict = max_position_embeddings
lowerCAmelCase__ : Union[str, Any] = layer_norm_eps
lowerCAmelCase__ : List[str] = layerdrop
lowerCAmelCase__ : str = hidden_act
lowerCAmelCase__ : Dict = initializer_range
lowerCAmelCase__ : str = hidden_dropout_prob
lowerCAmelCase__ : List[str] = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[Any] = pad_token_id
lowerCAmelCase__ : Union[str, Any] = bos_token_id
lowerCAmelCase__ : List[Any] = eos_token_id
lowerCAmelCase__ : str = conv_glu_dim
lowerCAmelCase__ : Dict = conv_dropout
lowerCAmelCase__ : Optional[int] = num_conv_layers
lowerCAmelCase__ : Dict = input_feat_per_channel
lowerCAmelCase__ : str = input_channels
lowerCAmelCase__ : str = conv_channels
lowerCAmelCase__ : List[Any] = ctc_loss_reduction
lowerCAmelCase__ : int = ctc_zero_infinity
# prevents config testing fail with exporting to json
lowerCAmelCase__ : Optional[Any] = list(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = list(__UpperCAmelCase )
if len(self.conv_kernel ) != self.num_conv_layers:
raise ValueError(
"""Configuration for convolutional module is incorrect. """
"""It is required that `len(config.conv_kernel)` == `config.num_conv_layers` """
F"""but is `len(config.conv_kernel) = {len(self.conv_kernel )}`, """
F"""`config.num_conv_layers = {self.num_conv_layers}`.""" )
| 37
|
'''simple docstring'''
from maths.prime_factors import prime_factors
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
if not isinstance(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : int = f"""Input value of [number={number}] must be an integer"""
raise TypeError(UpperCamelCase )
if number < 1:
raise ValueError("""Input must be a positive integer""" )
return -1 if len(prime_factors(UpperCamelCase ) ) % 2 else 1
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
| 1
|
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_torch_available,
is_vision_available,
)
_lowerCAmelCase = {
'''configuration_convnext''': ['''CONVNEXT_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''ConvNextConfig''', '''ConvNextOnnxConfig''']
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = ['''ConvNextFeatureExtractor''']
_lowerCAmelCase = ['''ConvNextImageProcessor''']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = [
'''CONVNEXT_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''ConvNextForImageClassification''',
'''ConvNextModel''',
'''ConvNextPreTrainedModel''',
'''ConvNextBackbone''',
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_lowerCAmelCase = [
'''TFConvNextForImageClassification''',
'''TFConvNextModel''',
'''TFConvNextPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_convnext import CONVNEXT_PRETRAINED_CONFIG_ARCHIVE_MAP, ConvNextConfig, ConvNextOnnxConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_convnext import ConvNextFeatureExtractor
from .image_processing_convnext import ConvNextImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_convnext import (
CONVNEXT_PRETRAINED_MODEL_ARCHIVE_LIST,
ConvNextBackbone,
ConvNextForImageClassification,
ConvNextModel,
ConvNextPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_convnext import TFConvNextForImageClassification, TFConvNextModel, TFConvNextPreTrainedModel
else:
import sys
_lowerCAmelCase = _LazyModule(__name__, globals()['''__file__'''], _import_structure)
| 37
|
'''simple docstring'''
import os
import re
import unicodedata
from shutil import copyfile
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import is_torch_available, logging
if is_torch_available():
import torch
if TYPE_CHECKING:
from transformers.pipelines.conversational import Conversation
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''AI-Sweden/gpt-sw3-126m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-126m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-350m''': '''https://huggingface.co/AI-Sweden/gpt-sw3-350m/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-1.6b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-1.6b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-6.7b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-6.7b/resolve/main/spiece.model''',
'''AI-Sweden/gpt-sw3-20b''': '''https://huggingface.co/AI-Sweden/gpt-sw3-20b/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''AI-Sweden/gpt-sw3-126m''': 2048,
'''AI-Sweden/gpt-sw3-350m''': 2048,
'''AI-Sweden/gpt-sw3-1.6b''': 2048,
'''AI-Sweden/gpt-sw3-6.7b''': 2048,
'''AI-Sweden/gpt-sw3-20b''': 2048,
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : str = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Optional[int] = ['''input_ids''', '''attention_mask''']
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
lowerCAmelCase__ : Optional[int] = {} if sp_model_kwargs is None else sp_model_kwargs
lowerCAmelCase__ : Dict = kwargs.get("""name_or_path""" )
if name_or_path is None:
logger.warning(
"""name_or_path not provided, will work for all GPTSw3 models except gpt-sw3-7b,"""
""" you are testing the model, this can safely be ignored""" )
lowerCAmelCase__ : Tuple = """None"""
# Default definitions for our 2 tokenizer versions, with None-checks to enable proper testing
lowerCAmelCase__ : Union[str, Any] = """<|endoftext|>""" if eos_token is None else eos_token
lowerCAmelCase__ : Dict = """<unk>""" if unk_token is None else unk_token
if "gpt-sw3-7b" in name_or_path:
lowerCAmelCase__ : Any = unk_token if pad_token is None else pad_token
lowerCAmelCase__ : Dict = eos_token if bos_token is None else bos_token
else:
lowerCAmelCase__ : List[str] = """<pad>""" if pad_token is None else pad_token
lowerCAmelCase__ : Optional[int] = """<s>""" if bos_token is None else bos_token
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : Optional[int] = do_lower_case
lowerCAmelCase__ : Dict = remove_space
lowerCAmelCase__ : Optional[Any] = keep_accents
lowerCAmelCase__ : int = vocab_file
lowerCAmelCase__ : str = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
# Used for whitespace normalization in input texts
# fmt : off
lowerCAmelCase__ : int = {""" """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """ """, """""", """"""}
# fmt : on
# Regular expression to remove non-printing characters (e.g. some unicode control chars) in preprocessing
lowerCAmelCase__ : List[str] = re.compile(
F"""[{''.join(map(__UpperCAmelCase ,list(range(0 ,9 ) ) + list(range(11 ,32 ) ) + list(range(127 ,160 ) ) + [160, 173, 8203] ) )}]""" )
def __getstate__( self ) -> Any:
lowerCAmelCase__ : int = self.__dict__.copy()
lowerCAmelCase__ : Optional[int] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Tuple = {}
lowerCAmelCase__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
@property
# Copied from transformers.models.albert.tokenization_albert.AlbertTokenizer.vocab_size
def UpperCAmelCase_ ( self ) -> int:
return len(self.sp_model )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : Tuple = self.non_printing_characters_re.sub("""""" ,__UpperCAmelCase )
# Normalize whitespaces
lowerCAmelCase__ : List[Any] = """""".join([char if char not in self.whitespaces else """ """ for char in text] )
# NFC Unicode normalization
lowerCAmelCase__ : List[Any] = unicodedata.normalize("""NFC""" ,__UpperCAmelCase )
return text
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[Any] = self.preprocess_text(__UpperCAmelCase )
return self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.IdToPiece(__UpperCAmelCase )
@staticmethod
def UpperCAmelCase_ ( __UpperCAmelCase ) -> str:
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : int = []
lowerCAmelCase__ : Optional[int] = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
# TODO: Check if this is needed, as it ensures that decode(encode(doc)) != doc by adding extra whitespace in the decoded document
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Optional[Any] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Any = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string
def UpperCAmelCase_ ( self ) -> Dict[str, int]:
lowerCAmelCase__ : Optional[int] = {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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : Optional[int] = 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:
lowerCAmelCase__ : str = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Union[List[int], List[List[int]], "torch.Tensor"]:
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Tuple = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : int = self.sp_model.encode(__UpperCAmelCase )
else:
lowerCAmelCase__ : int = [self.preprocess_text(__UpperCAmelCase ) for t in text]
lowerCAmelCase__ : Any = self.sp_model.encode(__UpperCAmelCase )
if return_tensors is True or return_tensors == "pt":
lowerCAmelCase__ : Tuple = torch.tensor(__UpperCAmelCase )
return token_ids
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
return self.sp_model.decode(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[int]:
lowerCAmelCase__ : List[Any] = [F"""User: {text}""" if is_user else F"""Bot: {text}""" for is_user, text in conversation.iter_texts()]
lowerCAmelCase__ : Any = (
F"""{self.eos_token}{self.bos_token}""" + F"""{self.bos_token}""".join(__UpperCAmelCase ) + F"""{self.bos_token}Bot:"""
)
return self.encode(text=__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = 0
while num > 0:
digit_sum += num % 10
num //= 10
return digit_sum
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = 100 ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = 1
lowerCAmelCase__ : List[str] = 2
for i in range(2 , max_n + 1 ):
lowerCAmelCase__ : Dict = pre_numerator
lowerCAmelCase__ : Optional[int] = 2 * i // 3 if i % 3 == 0 else 1
lowerCAmelCase__ : Tuple = cur_numerator
lowerCAmelCase__ : Optional[Any] = e_cont * pre_numerator + temp
return sum_digits(UpperCamelCase )
if __name__ == "__main__":
print(F"""{solution() = }""")
| 37
|
'''simple docstring'''
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = XLMRobertaModel.from_pretrained("""xlm-roberta-base""" )
lowerCAmelCase__ : Optional[int] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : str = torch.Size((1, 12, 768) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.0_1_0_1, 0.1_2_1_8, -0.0_8_0_3, 0.0_8_0_1, 0.1_3_2_7, 0.0_7_7_6, -0.1_2_1_5, 0.2_3_8_3, 0.3_3_3_8, 0.3_1_0_6, 0.0_3_0_0, 0.0_2_5_2]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.base')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Optional[int] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
@slow
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : Union[str, Any] = XLMRobertaModel.from_pretrained("""xlm-roberta-large""" )
lowerCAmelCase__ : Optional[Any] = torch.tensor([[0, 581, 1_0269, 83, 9_9942, 136, 6_0742, 23, 70, 8_0583, 1_8276, 2]] )
# The dog is cute and lives in the garden house
lowerCAmelCase__ : Dict = torch.Size((1, 12, 1024) ) # batch_size, sequence_length, embedding_vector_dim
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[[-0.0_6_9_9, -0.0_3_1_8, 0.0_7_0_5, -0.1_2_4_1, 0.0_9_9_9, -0.0_5_2_0, 0.1_0_0_4, -0.1_8_3_8, -0.4_7_0_4, 0.1_4_3_7, 0.0_8_2_1, 0.0_1_2_6]] )
# xlmr = torch.hub.load('pytorch/fairseq', 'xlmr.large')
# xlmr.eval()
# expected_output_values_last_dim = xlmr.extract_features(input_ids[0])[:, :, -1]
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(__UpperCAmelCase )["""last_hidden_state"""].detach()
self.assertEqual(output.shape ,__UpperCAmelCase )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] ,__UpperCAmelCase ,atol=1E-3 ) )
| 37
| 1
|
'''simple docstring'''
from typing import List, Union
from ..utils import (
add_end_docstrings,
is_tf_available,
is_torch_available,
is_vision_available,
logging,
requires_backends,
)
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_tf_available():
import tensorflow as tf
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING
from ..tf_utils import stable_softmax
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING
_lowerCAmelCase = logging.get_logger(__name__)
@add_end_docstrings(SCREAMING_SNAKE_CASE_ )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> int:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
requires_backends(self ,"""vision""" )
self.check_model_type(
TF_MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING
if self.framework == """tf"""
else MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING )
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ) -> int:
lowerCAmelCase__ : Dict = {}
if top_k is not None:
lowerCAmelCase__ : Optional[int] = top_k
return {}, {}, postprocess_params
def __call__( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> Union[str, Any]:
return super().__call__(__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : List[str] = load_image(__UpperCAmelCase )
lowerCAmelCase__ : Any = self.image_processor(images=__UpperCAmelCase ,return_tensors=self.framework )
return model_inputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : Any = self.model(**__UpperCAmelCase )
return model_outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=5 ) -> Any:
if top_k > self.model.config.num_labels:
lowerCAmelCase__ : Union[str, Any] = self.model.config.num_labels
if self.framework == "pt":
lowerCAmelCase__ : Union[str, Any] = model_outputs.logits.softmax(-1 )[0]
lowerCAmelCase__ , lowerCAmelCase__ : Dict = probs.topk(__UpperCAmelCase )
elif self.framework == "tf":
lowerCAmelCase__ : Any = stable_softmax(model_outputs.logits ,axis=-1 )[0]
lowerCAmelCase__ : List[Any] = tf.math.top_k(__UpperCAmelCase ,k=__UpperCAmelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = topk.values.numpy(), topk.indices.numpy()
else:
raise ValueError(F"""Unsupported framework: {self.framework}""" )
lowerCAmelCase__ : Tuple = scores.tolist()
lowerCAmelCase__ : int = ids.tolist()
return [{"score": score, "label": self.model.config.idalabel[_id]} for score, _id in zip(__UpperCAmelCase ,__UpperCAmelCase )]
| 37
|
'''simple docstring'''
from pickle import UnpicklingError
import jax
import jax.numpy as jnp
import numpy as np
from flax.serialization import from_bytes
from flax.traverse_util import flatten_dict
from ..utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
with open(UpperCamelCase , """rb""" ) as flax_state_f:
lowerCAmelCase__ : Union[str, Any] = from_bytes(UpperCamelCase , flax_state_f.read() )
except UnpicklingError as e:
try:
with open(UpperCamelCase ) as f:
if f.read().startswith("""version""" ):
raise OSError(
"""You seem to have cloned a repository without having git-lfs installed. Please"""
""" install git-lfs and run `git lfs install` followed by `git lfs pull` in the"""
""" folder you cloned.""" )
else:
raise ValueError from e
except (UnicodeDecodeError, ValueError):
raise EnvironmentError(f"""Unable to convert {model_file} to Flax deserializable object. """ )
return load_flax_weights_in_pytorch_model(UpperCamelCase , UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
try:
import torch # noqa: F401
except ImportError:
logger.error(
"""Loading Flax weights in PyTorch requires both PyTorch and Flax to be installed. Please see"""
""" https://pytorch.org/ and https://flax.readthedocs.io/en/latest/installation.html for installation"""
""" instructions.""" )
raise
# check if we have bf16 weights
lowerCAmelCase__ : str = flatten_dict(jax.tree_util.tree_map(lambda UpperCamelCase : x.dtype == jnp.bfloataa , UpperCamelCase ) ).values()
if any(UpperCamelCase ):
# convert all weights to fp32 if they are bf16 since torch.from_numpy can-not handle bf16
# and bf16 is not fully supported in PT yet.
logger.warning(
"""Found ``bfloat16`` weights in Flax model. Casting all ``bfloat16`` weights to ``float32`` """
"""before loading those in PyTorch model.""" )
lowerCAmelCase__ : Dict = jax.tree_util.tree_map(
lambda UpperCamelCase : params.astype(np.floataa ) if params.dtype == jnp.bfloataa else params , UpperCamelCase )
lowerCAmelCase__ : Any = """"""
lowerCAmelCase__ : Any = flatten_dict(UpperCamelCase , sep=""".""" )
lowerCAmelCase__ : Optional[int] = pt_model.state_dict()
# keep track of unexpected & missing keys
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : int = set(pt_model_dict.keys() )
for flax_key_tuple, flax_tensor in flax_state_dict.items():
lowerCAmelCase__ : Union[str, Any] = flax_key_tuple.split(""".""" )
if flax_key_tuple_array[-1] == "kernel" and flax_tensor.ndim == 4:
lowerCAmelCase__ : Optional[int] = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = jnp.transpose(UpperCamelCase , (3, 2, 0, 1) )
elif flax_key_tuple_array[-1] == "kernel":
lowerCAmelCase__ : str = flax_key_tuple_array[:-1] + ["""weight"""]
lowerCAmelCase__ : Any = flax_tensor.T
elif flax_key_tuple_array[-1] == "scale":
lowerCAmelCase__ : int = flax_key_tuple_array[:-1] + ["""weight"""]
if "time_embedding" not in flax_key_tuple_array:
for i, flax_key_tuple_string in enumerate(UpperCamelCase ):
lowerCAmelCase__ : List[str] = (
flax_key_tuple_string.replace("""_0""" , """.0""" )
.replace("""_1""" , """.1""" )
.replace("""_2""" , """.2""" )
.replace("""_3""" , """.3""" )
.replace("""_4""" , """.4""" )
.replace("""_5""" , """.5""" )
.replace("""_6""" , """.6""" )
.replace("""_7""" , """.7""" )
.replace("""_8""" , """.8""" )
.replace("""_9""" , """.9""" )
)
lowerCAmelCase__ : Union[str, Any] = """.""".join(UpperCamelCase )
if flax_key in pt_model_dict:
if flax_tensor.shape != pt_model_dict[flax_key].shape:
raise ValueError(
f"""Flax checkpoint seems to be incorrect. Weight {flax_key_tuple} was expected """
f"""to be of shape {pt_model_dict[flax_key].shape}, but is {flax_tensor.shape}.""" )
else:
# add weight to pytorch dict
lowerCAmelCase__ : int = np.asarray(UpperCamelCase ) if not isinstance(UpperCamelCase , np.ndarray ) else flax_tensor
lowerCAmelCase__ : int = torch.from_numpy(UpperCamelCase )
# remove from missing keys
missing_keys.remove(UpperCamelCase )
else:
# weight is not expected by PyTorch model
unexpected_keys.append(UpperCamelCase )
pt_model.load_state_dict(UpperCamelCase )
# re-transform missing_keys to list
lowerCAmelCase__ : Optional[int] = list(UpperCamelCase )
if len(UpperCamelCase ) > 0:
logger.warning(
"""Some weights of the Flax model were not used when initializing the PyTorch model"""
f""" {pt_model.__class__.__name__}: {unexpected_keys}\n- This IS expected if you are initializing"""
f""" {pt_model.__class__.__name__} from a Flax model trained on another task or with another architecture"""
""" (e.g. initializing a BertForSequenceClassification model from a FlaxBertForPreTraining model).\n- This"""
f""" IS NOT expected if you are initializing {pt_model.__class__.__name__} from a Flax model that you expect"""
""" to be exactly identical (e.g. initializing a BertForSequenceClassification model from a"""
""" FlaxBertForSequenceClassification model).""" )
if len(UpperCamelCase ) > 0:
logger.warning(
f"""Some weights of {pt_model.__class__.__name__} were not initialized from the Flax model and are newly"""
f""" initialized: {missing_keys}\nYou should probably TRAIN this model on a down-stream task to be able to"""
""" use it for predictions and inference.""" )
return pt_model
| 37
| 1
|
'''simple docstring'''
import argparse
import os
from . import (
ALBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
BART_PRETRAINED_MODEL_ARCHIVE_LIST,
BERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
CAMEMBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
CTRL_PRETRAINED_CONFIG_ARCHIVE_MAP,
DISTILBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
DPR_CONTEXT_ENCODER_PRETRAINED_MODEL_ARCHIVE_LIST,
DPR_QUESTION_ENCODER_PRETRAINED_MODEL_ARCHIVE_LIST,
DPR_READER_PRETRAINED_MODEL_ARCHIVE_LIST,
ELECTRA_PRETRAINED_CONFIG_ARCHIVE_MAP,
FLAUBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
GPT2_PRETRAINED_CONFIG_ARCHIVE_MAP,
LAYOUTLM_PRETRAINED_MODEL_ARCHIVE_LIST,
LXMERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
OPENAI_GPT_PRETRAINED_CONFIG_ARCHIVE_MAP,
ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP,
T5_PRETRAINED_CONFIG_ARCHIVE_MAP,
TRANSFO_XL_PRETRAINED_CONFIG_ARCHIVE_MAP,
WAV_2_VEC_2_PRETRAINED_CONFIG_ARCHIVE_MAP,
XLM_PRETRAINED_CONFIG_ARCHIVE_MAP,
XLM_ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP,
XLNET_PRETRAINED_CONFIG_ARCHIVE_MAP,
AlbertConfig,
BartConfig,
BertConfig,
CamembertConfig,
CTRLConfig,
DistilBertConfig,
DPRConfig,
ElectraConfig,
FlaubertConfig,
GPTaConfig,
LayoutLMConfig,
LxmertConfig,
OpenAIGPTConfig,
RobertaConfig,
TaConfig,
TFAlbertForPreTraining,
TFBartForConditionalGeneration,
TFBartForSequenceClassification,
TFBertForPreTraining,
TFBertForQuestionAnswering,
TFBertForSequenceClassification,
TFCamembertForMaskedLM,
TFCTRLLMHeadModel,
TFDistilBertForMaskedLM,
TFDistilBertForQuestionAnswering,
TFDPRContextEncoder,
TFDPRQuestionEncoder,
TFDPRReader,
TFElectraForPreTraining,
TFFlaubertWithLMHeadModel,
TFGPTaLMHeadModel,
TFLayoutLMForMaskedLM,
TFLxmertForPreTraining,
TFLxmertVisualFeatureEncoder,
TFOpenAIGPTLMHeadModel,
TFRobertaForCausalLM,
TFRobertaForMaskedLM,
TFRobertaForSequenceClassification,
TFTaForConditionalGeneration,
TFTransfoXLLMHeadModel,
TFWavaVecaModel,
TFXLMRobertaForMaskedLM,
TFXLMWithLMHeadModel,
TFXLNetLMHeadModel,
TransfoXLConfig,
WavaVecaConfig,
WavaVecaModel,
XLMConfig,
XLMRobertaConfig,
XLNetConfig,
is_torch_available,
load_pytorch_checkpoint_in_tfa_model,
)
from .utils import CONFIG_NAME, WEIGHTS_NAME, cached_file, logging
if is_torch_available():
import numpy as np
import torch
from . import (
AlbertForPreTraining,
BartForConditionalGeneration,
BertForPreTraining,
BertForQuestionAnswering,
BertForSequenceClassification,
CamembertForMaskedLM,
CTRLLMHeadModel,
DistilBertForMaskedLM,
DistilBertForQuestionAnswering,
DPRContextEncoder,
DPRQuestionEncoder,
DPRReader,
ElectraForPreTraining,
FlaubertWithLMHeadModel,
GPTaLMHeadModel,
LayoutLMForMaskedLM,
LxmertForPreTraining,
LxmertVisualFeatureEncoder,
OpenAIGPTLMHeadModel,
RobertaForMaskedLM,
RobertaForSequenceClassification,
TaForConditionalGeneration,
TransfoXLLMHeadModel,
XLMRobertaForMaskedLM,
XLMWithLMHeadModel,
XLNetLMHeadModel,
)
logging.set_verbosity_info()
_lowerCAmelCase = {
'''bart''': (
BartConfig,
TFBartForConditionalGeneration,
TFBartForSequenceClassification,
BartForConditionalGeneration,
BART_PRETRAINED_MODEL_ARCHIVE_LIST,
),
'''bert''': (
BertConfig,
TFBertForPreTraining,
BertForPreTraining,
BERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''bert-large-uncased-whole-word-masking-finetuned-squad''': (
BertConfig,
TFBertForQuestionAnswering,
BertForQuestionAnswering,
BERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''bert-large-cased-whole-word-masking-finetuned-squad''': (
BertConfig,
TFBertForQuestionAnswering,
BertForQuestionAnswering,
BERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''bert-base-cased-finetuned-mrpc''': (
BertConfig,
TFBertForSequenceClassification,
BertForSequenceClassification,
BERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''dpr''': (
DPRConfig,
TFDPRQuestionEncoder,
TFDPRContextEncoder,
TFDPRReader,
DPRQuestionEncoder,
DPRContextEncoder,
DPRReader,
DPR_CONTEXT_ENCODER_PRETRAINED_MODEL_ARCHIVE_LIST,
DPR_QUESTION_ENCODER_PRETRAINED_MODEL_ARCHIVE_LIST,
DPR_READER_PRETRAINED_MODEL_ARCHIVE_LIST,
),
'''gpt2''': (
GPTaConfig,
TFGPTaLMHeadModel,
GPTaLMHeadModel,
GPT2_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''xlnet''': (
XLNetConfig,
TFXLNetLMHeadModel,
XLNetLMHeadModel,
XLNET_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''xlm''': (
XLMConfig,
TFXLMWithLMHeadModel,
XLMWithLMHeadModel,
XLM_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''xlm-roberta''': (
XLMRobertaConfig,
TFXLMRobertaForMaskedLM,
XLMRobertaForMaskedLM,
XLM_ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''transfo-xl''': (
TransfoXLConfig,
TFTransfoXLLMHeadModel,
TransfoXLLMHeadModel,
TRANSFO_XL_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''openai-gpt''': (
OpenAIGPTConfig,
TFOpenAIGPTLMHeadModel,
OpenAIGPTLMHeadModel,
OPENAI_GPT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''roberta''': (
RobertaConfig,
TFRobertaForCausalLM,
TFRobertaForMaskedLM,
RobertaForMaskedLM,
ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''layoutlm''': (
LayoutLMConfig,
TFLayoutLMForMaskedLM,
LayoutLMForMaskedLM,
LAYOUTLM_PRETRAINED_MODEL_ARCHIVE_LIST,
),
'''roberta-large-mnli''': (
RobertaConfig,
TFRobertaForSequenceClassification,
RobertaForSequenceClassification,
ROBERTA_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''camembert''': (
CamembertConfig,
TFCamembertForMaskedLM,
CamembertForMaskedLM,
CAMEMBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''flaubert''': (
FlaubertConfig,
TFFlaubertWithLMHeadModel,
FlaubertWithLMHeadModel,
FLAUBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''distilbert''': (
DistilBertConfig,
TFDistilBertForMaskedLM,
DistilBertForMaskedLM,
DISTILBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''distilbert-base-distilled-squad''': (
DistilBertConfig,
TFDistilBertForQuestionAnswering,
DistilBertForQuestionAnswering,
DISTILBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''lxmert''': (
LxmertConfig,
TFLxmertForPreTraining,
LxmertForPreTraining,
LXMERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''lxmert-visual-feature-encoder''': (
LxmertConfig,
TFLxmertVisualFeatureEncoder,
LxmertVisualFeatureEncoder,
LXMERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''ctrl''': (
CTRLConfig,
TFCTRLLMHeadModel,
CTRLLMHeadModel,
CTRL_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''albert''': (
AlbertConfig,
TFAlbertForPreTraining,
AlbertForPreTraining,
ALBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''t5''': (
TaConfig,
TFTaForConditionalGeneration,
TaForConditionalGeneration,
T5_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''electra''': (
ElectraConfig,
TFElectraForPreTraining,
ElectraForPreTraining,
ELECTRA_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
'''wav2vec2''': (
WavaVecaConfig,
TFWavaVecaModel,
WavaVecaModel,
WAV_2_VEC_2_PRETRAINED_CONFIG_ARCHIVE_MAP,
),
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=False , UpperCamelCase=True ):
"""simple docstring"""
if model_type not in MODEL_CLASSES:
raise ValueError(f"""Unrecognized model type, should be one of {list(MODEL_CLASSES.keys() )}.""" )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = MODEL_CLASSES[model_type]
# Initialise TF model
if config_file in aws_config_map:
lowerCAmelCase__ : Tuple = cached_file(UpperCamelCase , UpperCamelCase , force_download=not use_cached_models )
lowerCAmelCase__ : List[Any] = config_class.from_json_file(UpperCamelCase )
lowerCAmelCase__ : str = True
lowerCAmelCase__ : int = True
print(f"""Building TensorFlow model from configuration: {config}""" )
lowerCAmelCase__ : Optional[int] = model_class(UpperCamelCase )
# Load weights from tf checkpoint
if pytorch_checkpoint_path in aws_config_map.keys():
lowerCAmelCase__ : int = cached_file(
UpperCamelCase , UpperCamelCase , force_download=not use_cached_models )
# Load PyTorch checkpoint in tf2 model:
lowerCAmelCase__ : Any = load_pytorch_checkpoint_in_tfa_model(UpperCamelCase , UpperCamelCase )
if compare_with_pt_model:
lowerCAmelCase__ : Optional[Any] = tf_model(tf_model.dummy_inputs , training=UpperCamelCase ) # build the network
lowerCAmelCase__ : List[Any] = torch.load(UpperCamelCase , map_location="""cpu""" )
lowerCAmelCase__ : Optional[Any] = pt_model_class.from_pretrained(
pretrained_model_name_or_path=UpperCamelCase , config=UpperCamelCase , state_dict=UpperCamelCase )
with torch.no_grad():
lowerCAmelCase__ : int = pt_model(**pt_model.dummy_inputs )
lowerCAmelCase__ : Dict = pto[0].numpy()
lowerCAmelCase__ : Union[str, Any] = tfo[0].numpy()
lowerCAmelCase__ : Any = np.amax(np.abs(np_pt - np_tf ) )
print(f"""Max absolute difference between models outputs {diff}""" )
assert diff <= 2e-2, f"""Error, model absolute difference is >2e-2: {diff}"""
# Save pytorch-model
print(f"""Save TensorFlow model to {tf_dump_path}""" )
tf_model.save_weights(UpperCamelCase , save_format="""h5""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=False , UpperCamelCase=False , UpperCamelCase=False , UpperCamelCase=False , ):
"""simple docstring"""
if args_model_type is None:
lowerCAmelCase__ : int = list(MODEL_CLASSES.keys() )
else:
lowerCAmelCase__ : List[str] = [args_model_type]
for j, model_type in enumerate(UpperCamelCase , start=1 ):
print("""=""" * 100 )
print(f""" Converting model type {j}/{len(UpperCamelCase )}: {model_type}""" )
print("""=""" * 100 )
if model_type not in MODEL_CLASSES:
raise ValueError(f"""Unrecognized model type {model_type}, should be one of {list(MODEL_CLASSES.keys() )}.""" )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = MODEL_CLASSES[model_type]
if model_shortcut_names_or_path is None:
lowerCAmelCase__ : List[str] = list(aws_model_maps.keys() )
if config_shortcut_names_or_path is None:
lowerCAmelCase__ : int = model_shortcut_names_or_path
for i, (model_shortcut_name, config_shortcut_name) in enumerate(
zip(UpperCamelCase , UpperCamelCase ) , start=1 ):
print("""-""" * 100 )
if "-squad" in model_shortcut_name or "-mrpc" in model_shortcut_name or "-mnli" in model_shortcut_name:
if not only_convert_finetuned_models:
print(f""" Skipping finetuned checkpoint {model_shortcut_name}""" )
continue
lowerCAmelCase__ : str = model_shortcut_name
elif only_convert_finetuned_models:
print(f""" Skipping not finetuned checkpoint {model_shortcut_name}""" )
continue
print(
f""" Converting checkpoint {i}/{len(UpperCamelCase )}: {model_shortcut_name} - model_type {model_type}""" )
print("""-""" * 100 )
if config_shortcut_name in aws_config_map:
lowerCAmelCase__ : List[Any] = cached_file(UpperCamelCase , UpperCamelCase , force_download=not use_cached_models )
else:
lowerCAmelCase__ : Tuple = config_shortcut_name
if model_shortcut_name in aws_model_maps:
lowerCAmelCase__ : List[Any] = cached_file(UpperCamelCase , UpperCamelCase , force_download=not use_cached_models )
else:
lowerCAmelCase__ : Optional[int] = model_shortcut_name
if os.path.isfile(UpperCamelCase ):
lowerCAmelCase__ : int = """converted_model"""
convert_pt_checkpoint_to_tf(
model_type=UpperCamelCase , pytorch_checkpoint_path=UpperCamelCase , config_file=UpperCamelCase , tf_dump_path=os.path.join(UpperCamelCase , model_shortcut_name + """-tf_model.h5""" ) , compare_with_pt_model=UpperCamelCase , )
if remove_cached_files:
os.remove(UpperCamelCase )
os.remove(UpperCamelCase )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--tf_dump_path''', default=None, type=str, required=True, help='''Path to the output Tensorflow dump file.'''
)
parser.add_argument(
'''--model_type''',
default=None,
type=str,
help=(
F"""Model type selected in the list of {list(MODEL_CLASSES.keys())}. If not given, will download and """
'''convert all the models from AWS.'''
),
)
parser.add_argument(
'''--pytorch_checkpoint_path''',
default=None,
type=str,
help=(
'''Path to the PyTorch checkpoint path or shortcut name to download from AWS. '''
'''If not given, will download and convert all the checkpoints from AWS.'''
),
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
help=(
'''The config json file corresponding to the pre-trained model. \n'''
'''This specifies the model architecture. If not given and '''
'''--pytorch_checkpoint_path is not given or is a shortcut name '''
'''use the configuration associated to the shortcut name on the AWS'''
),
)
parser.add_argument(
'''--compare_with_pt_model''', action='''store_true''', help='''Compare Tensorflow and PyTorch model predictions.'''
)
parser.add_argument(
'''--use_cached_models''',
action='''store_true''',
help='''Use cached models if possible instead of updating to latest checkpoint versions.''',
)
parser.add_argument(
'''--remove_cached_files''',
action='''store_true''',
help='''Remove pytorch models after conversion (save memory when converting in batches).''',
)
parser.add_argument('''--only_convert_finetuned_models''', action='''store_true''', help='''Only convert finetuned models.''')
_lowerCAmelCase = parser.parse_args()
# if args.pytorch_checkpoint_path is not None:
# convert_pt_checkpoint_to_tf(args.model_type.lower(),
# args.pytorch_checkpoint_path,
# args.config_file if args.config_file is not None else args.pytorch_checkpoint_path,
# args.tf_dump_path,
# compare_with_pt_model=args.compare_with_pt_model,
# use_cached_models=args.use_cached_models)
# else:
convert_all_pt_checkpoints_to_tf(
args.model_type.lower() if args.model_type is not None else None,
args.tf_dump_path,
model_shortcut_names_or_path=[args.pytorch_checkpoint_path]
if args.pytorch_checkpoint_path is not None
else None,
config_shortcut_names_or_path=[args.config_file] if args.config_file is not None else None,
compare_with_pt_model=args.compare_with_pt_model,
use_cached_models=args.use_cached_models,
remove_cached_files=args.remove_cached_files,
only_convert_finetuned_models=args.only_convert_finetuned_models,
)
| 37
|
'''simple docstring'''
import inspect
import unittest
import numpy as np
from tests.test_modeling_common import floats_tensor
from transformers import DetrConfig, MaskFormerConfig, SwinConfig, is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_torch_multi_gpu, require_vision, slow, torch_device
from transformers.utils import cached_property
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import MaskFormerForInstanceSegmentation, MaskFormerModel
if is_vision_available():
from transformers import MaskFormerImageProcessor
if is_vision_available():
from PIL import Image
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=2 ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase=10 ,__UpperCAmelCase=3 ,__UpperCAmelCase=32 * 4 ,__UpperCAmelCase=32 * 6 ,__UpperCAmelCase=4 ,__UpperCAmelCase=32 ,) -> str:
lowerCAmelCase__ : Optional[int] = parent
lowerCAmelCase__ : Optional[int] = batch_size
lowerCAmelCase__ : Optional[int] = is_training
lowerCAmelCase__ : Dict = use_auxiliary_loss
lowerCAmelCase__ : Union[str, Any] = num_queries
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : List[str] = min_size
lowerCAmelCase__ : int = max_size
lowerCAmelCase__ : Optional[Any] = num_labels
lowerCAmelCase__ : List[Any] = mask_feature_size
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : str = floats_tensor([self.batch_size, self.num_channels, self.min_size, self.max_size] ).to(
__UpperCAmelCase )
lowerCAmelCase__ : str = torch.ones([self.batch_size, self.min_size, self.max_size] ,device=__UpperCAmelCase )
lowerCAmelCase__ : Any = (
torch.rand([self.batch_size, self.num_labels, self.min_size, self.max_size] ,device=__UpperCAmelCase ) > 0.5
).float()
lowerCAmelCase__ : Optional[int] = (torch.rand((self.batch_size, self.num_labels) ,device=__UpperCAmelCase ) > 0.5).long()
lowerCAmelCase__ : Any = self.get_config()
return config, pixel_values, pixel_mask, mask_labels, class_labels
def UpperCAmelCase_ ( self ) -> Dict:
return MaskFormerConfig.from_backbone_and_decoder_configs(
backbone_config=SwinConfig(
depths=[1, 1, 1, 1] ,) ,decoder_config=DetrConfig(
decoder_ffn_dim=128 ,num_queries=self.num_queries ,decoder_attention_heads=2 ,d_model=self.mask_feature_size ,) ,mask_feature_size=self.mask_feature_size ,fpn_feature_size=self.mask_feature_size ,num_channels=self.num_channels ,num_labels=self.num_labels ,)
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.prepare_config_and_inputs()
lowerCAmelCase__ : List[str] = {"""pixel_values""": pixel_values, """pixel_mask""": pixel_mask}
return config, inputs_dict
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Optional[int] = output.encoder_hidden_states
lowerCAmelCase__ : Optional[int] = output.pixel_decoder_hidden_states
lowerCAmelCase__ : Dict = output.transformer_decoder_hidden_states
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,len(config.backbone_config.depths ) )
self.parent.assertTrue(len(__UpperCAmelCase ) ,config.decoder_config.decoder_layers )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=False ) -> Optional[Any]:
with torch.no_grad():
lowerCAmelCase__ : int = MaskFormerModel(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
lowerCAmelCase__ : str = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : int = model(__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
# the correct shape of output.transformer_decoder_hidden_states ensure the correcteness of the
# encoder and pixel decoder
self.parent.assertEqual(
output.transformer_decoder_last_hidden_state.shape ,(self.batch_size, self.num_queries, self.mask_feature_size) ,)
# let's ensure the other two hidden state exists
self.parent.assertTrue(output.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(output.encoder_last_hidden_state is not None )
if output_hidden_states:
self.check_output_hidden_state(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : Dict = MaskFormerForInstanceSegmentation(config=__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.eval()
def comm_check_on_output(__UpperCAmelCase ):
# let's still check that all the required stuff is there
self.parent.assertTrue(result.transformer_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.pixel_decoder_last_hidden_state is not None )
self.parent.assertTrue(result.encoder_last_hidden_state is not None )
# okay, now we need to check the logits shape
# due to the encoder compression, masks have a //4 spatial size
self.parent.assertEqual(
result.masks_queries_logits.shape ,(self.batch_size, self.num_queries, self.min_size // 4, self.max_size // 4) ,)
# + 1 for null class
self.parent.assertEqual(
result.class_queries_logits.shape ,(self.batch_size, self.num_queries, self.num_labels + 1) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase )
lowerCAmelCase__ : Dict = model(__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = model(
pixel_values=__UpperCAmelCase ,pixel_mask=__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
comm_check_on_output(__UpperCAmelCase )
self.parent.assertTrue(result.loss is not None )
self.parent.assertEqual(result.loss.shape ,torch.Size([1] ) )
@require_torch
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Optional[Any] = (MaskFormerModel, MaskFormerForInstanceSegmentation) if is_torch_available() else ()
__lowercase : int = (
{'''feature-extraction''': MaskFormerModel, '''image-segmentation''': MaskFormerForInstanceSegmentation}
if is_torch_available()
else {}
)
__lowercase : Union[str, Any] = False
__lowercase : Dict = False
__lowercase : Tuple = False
__lowercase : List[Any] = False
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : str = MaskFormerModelTester(self )
lowerCAmelCase__ : List[Any] = ConfigTester(self ,config_class=__UpperCAmelCase ,has_text_modality=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
self.config_tester.run_common_tests()
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ , lowerCAmelCase__ : str = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_maskformer_instance_segmentation_head_model(*__UpperCAmelCase )
@unittest.skip(reason="""MaskFormer does not use inputs_embeds""" )
def UpperCAmelCase_ ( self ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormer does not have a get_input_embeddings method""" )
def UpperCAmelCase_ ( self ) -> str:
pass
@unittest.skip(reason="""MaskFormer is not a generative model""" )
def UpperCAmelCase_ ( self ) -> Any:
pass
@unittest.skip(reason="""MaskFormer does not use token embeddings""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
@require_torch_multi_gpu
@unittest.skip(
reason="""MaskFormer has some layers using `add_module` which doesn't work well with `nn.DataParallel`""" )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
pass
@unittest.skip("""Will be fixed soon by reducing the size of the model used for common tests.""" )
def UpperCAmelCase_ ( self ) -> List[str]:
pass
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : str = model_class(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
lowerCAmelCase__ : Dict = [*signature.parameters.keys()]
lowerCAmelCase__ : Tuple = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] ,__UpperCAmelCase )
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
for model_name in ["facebook/maskformer-swin-small-coco"]:
lowerCAmelCase__ : List[str] = MaskFormerModel.from_pretrained(__UpperCAmelCase )
self.assertIsNotNone(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (self.model_tester.min_size,) * 2
lowerCAmelCase__ : Any = {
"""pixel_values""": torch.randn((2, 3, *size) ,device=__UpperCAmelCase ),
"""mask_labels""": torch.randn((2, 10, *size) ,device=__UpperCAmelCase ),
"""class_labels""": torch.zeros(2 ,10 ,device=__UpperCAmelCase ).long(),
}
lowerCAmelCase__ : Tuple = MaskFormerForInstanceSegmentation(MaskFormerConfig() ).to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.create_and_check_maskformer_model(__UpperCAmelCase ,**__UpperCAmelCase ,output_hidden_states=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ , lowerCAmelCase__ : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
lowerCAmelCase__ : int = model_class(__UpperCAmelCase ).to(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase ,output_attentions=__UpperCAmelCase )
self.assertTrue(outputs.attentions is not None )
def UpperCAmelCase_ ( self ) -> int:
if not self.model_tester.is_training:
return
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Dict = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : List[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : List[str] = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase ).loss
loss.backward()
def UpperCAmelCase_ ( self ) -> List[str]:
# only MaskFormerForInstanceSegmentation has the loss
lowerCAmelCase__ : Tuple = self.all_model_classes[1]
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : Tuple = True
lowerCAmelCase__ : Optional[Any] = model_class(__UpperCAmelCase )
model.to(__UpperCAmelCase )
model.train()
lowerCAmelCase__ : Dict = model(__UpperCAmelCase ,mask_labels=__UpperCAmelCase ,class_labels=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = outputs.encoder_hidden_states[0]
encoder_hidden_states.retain_grad()
lowerCAmelCase__ : str = outputs.pixel_decoder_hidden_states[0]
pixel_decoder_hidden_states.retain_grad()
# we requires_grad=True in inputs_embeds (line 2152), the original implementation don't
lowerCAmelCase__ : Union[str, Any] = outputs.transformer_decoder_hidden_states[0]
transformer_decoder_hidden_states.retain_grad()
lowerCAmelCase__ : List[Any] = outputs.attentions[0]
attentions.retain_grad()
outputs.loss.backward(retain_graph=__UpperCAmelCase )
self.assertIsNotNone(encoder_hidden_states.grad )
self.assertIsNotNone(pixel_decoder_hidden_states.grad )
self.assertIsNotNone(transformer_decoder_hidden_states.grad )
self.assertIsNotNone(attentions.grad )
_lowerCAmelCase = 1e-4
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
return image
@require_vision
@slow
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@cached_property
def UpperCAmelCase_ ( self ) -> List[Any]:
return (
MaskFormerImageProcessor.from_pretrained("""facebook/maskformer-swin-small-coco""" )
if is_vision_available()
else None
)
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Any = MaskFormerModel.from_pretrained("""facebook/maskformer-swin-small-coco""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = self.default_image_processor
lowerCAmelCase__ : str = prepare_img()
lowerCAmelCase__ : Optional[int] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Dict = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : Union[str, Any] = model(**__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[-0.0_4_8_2, 0.9_2_2_8, 0.4_9_5_1], [-0.2_5_4_7, 0.8_0_1_7, 0.8_5_2_7], [-0.0_0_6_9, 0.3_3_8_5, -0.0_0_8_9]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.encoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.8_4_2_2, -0.8_4_3_4, -0.9_7_1_8], [-1.0_1_4_4, -0.5_5_6_5, -0.4_1_9_5], [-1.0_0_3_8, -0.4_4_8_4, -0.1_9_6_1]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.pixel_decoder_last_hidden_state[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
lowerCAmelCase__ : Optional[int] = torch.tensor(
[[0.2_8_5_2, -0.0_1_5_9, 0.9_7_3_5], [0.6_2_5_4, 0.1_8_5_8, 0.8_5_2_9], [-0.0_6_8_0, -0.4_1_1_6, 1.8_4_1_3]] ).to(__UpperCAmelCase )
self.assertTrue(
torch.allclose(
outputs.transformer_decoder_last_hidden_state[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : List[str] = prepare_img()
lowerCAmelCase__ : str = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : List[Any] = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[int] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : Optional[int] = [
[-1.3_7_3_7_1_2_4, -1.7_7_2_4_9_3_7, -1.9_3_6_4_2_3_3],
[-1.5_9_7_7_2_8_1, -1.9_8_6_7_9_3_9, -2.1_5_2_3_6_9_5],
[-1.5_7_9_5_3_9_8, -1.9_2_6_9_8_3_2, -2.0_9_3_9_4_2],
]
lowerCAmelCase__ : Optional[int] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Tuple = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Union[str, Any] = torch.tensor(
[
[1.65_12E00, -5.25_72E00, -3.35_19E00],
[3.61_69E-02, -5.90_25E00, -2.93_13E00],
[1.07_66E-04, -7.76_30E00, -5.12_63E00],
] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : List[Any] = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-resnet101-coco-stuff""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Optional[Any] = self.default_image_processor
lowerCAmelCase__ : int = prepare_img()
lowerCAmelCase__ : Optional[Any] = image_processor(__UpperCAmelCase ,return_tensors="""pt""" ).to(__UpperCAmelCase )
lowerCAmelCase__ : str = inputs["""pixel_values"""].shape
# check size is divisible by 32
self.assertTrue((inputs_shape[-1] % 32) == 0 and (inputs_shape[-2] % 32) == 0 )
# check size
self.assertEqual(__UpperCAmelCase ,(1, 3, 800, 1088) )
with torch.no_grad():
lowerCAmelCase__ : str = model(**__UpperCAmelCase )
# masks_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.masks_queries_logits
self.assertEqual(
masks_queries_logits.shape ,(1, model.config.decoder_config.num_queries, inputs_shape[-2] // 4, inputs_shape[-1] // 4) ,)
lowerCAmelCase__ : int = [[-0.9_0_4_6, -2.6_3_6_6, -4.6_0_6_2], [-3.4_1_7_9, -5.7_8_9_0, -8.8_0_5_7], [-4.9_1_7_9, -7.6_5_6_0, -1_0.7_7_1_1]]
lowerCAmelCase__ : List[str] = torch.tensor(__UpperCAmelCase ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(masks_queries_logits[0, 0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
# class_queries_logits
lowerCAmelCase__ : Optional[Any] = outputs.class_queries_logits
self.assertEqual(
class_queries_logits.shape ,(1, model.config.decoder_config.num_queries, model.config.num_labels + 1) )
lowerCAmelCase__ : Tuple = torch.tensor(
[[4.7_1_8_8, -3.2_5_8_5, -2.8_8_5_7], [6.6_8_7_1, -2.9_1_8_1, -1.2_4_8_7], [7.2_4_4_9, -2.2_7_6_4, -2.1_8_7_4]] ).to(__UpperCAmelCase )
self.assertTrue(torch.allclose(outputs.class_queries_logits[0, :3, :3] ,__UpperCAmelCase ,atol=__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : str = (
MaskFormerForInstanceSegmentation.from_pretrained("""facebook/maskformer-swin-small-coco""" )
.to(__UpperCAmelCase )
.eval()
)
lowerCAmelCase__ : Dict = self.default_image_processor
lowerCAmelCase__ : Union[str, Any] = image_processor(
[np.zeros((3, 800, 1333) ), np.zeros((3, 800, 1333) )] ,segmentation_maps=[np.zeros((384, 384) ).astype(np.floataa ), np.zeros((384, 384) ).astype(np.floataa )] ,return_tensors="""pt""" ,)
lowerCAmelCase__ : Tuple = inputs["""pixel_values"""].to(__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""mask_labels"""]]
lowerCAmelCase__ : Union[str, Any] = [el.to(__UpperCAmelCase ) for el in inputs["""class_labels"""]]
with torch.no_grad():
lowerCAmelCase__ : Any = model(**__UpperCAmelCase )
self.assertTrue(outputs.loss is not None )
| 37
| 1
|
'''simple docstring'''
import argparse
import json
import pickle
from pathlib import Path
import requests
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import MaskFormerConfig, MaskFormerForInstanceSegmentation, MaskFormerImageProcessor, SwinConfig
from transformers.utils import logging
logging.set_verbosity_info()
_lowerCAmelCase = logging.get_logger(__name__)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Dict = SwinConfig.from_pretrained(
"""microsoft/swin-tiny-patch4-window7-224""" , out_features=["""stage1""", """stage2""", """stage3""", """stage4"""] )
lowerCAmelCase__ : Optional[int] = MaskFormerConfig(backbone_config=UpperCamelCase )
lowerCAmelCase__ : str = """huggingface/label-files"""
if "ade20k-full" in model_name:
# this should be ok
lowerCAmelCase__ : List[Any] = 847
lowerCAmelCase__ : Union[str, Any] = """maskformer-ade20k-full-id2label.json"""
elif "ade" in model_name:
# this should be ok
lowerCAmelCase__ : Optional[int] = 150
lowerCAmelCase__ : Dict = """ade20k-id2label.json"""
elif "coco-stuff" in model_name:
# this should be ok
lowerCAmelCase__ : Optional[Any] = 171
lowerCAmelCase__ : Any = """maskformer-coco-stuff-id2label.json"""
elif "coco" in model_name:
# TODO
lowerCAmelCase__ : Optional[Any] = 133
lowerCAmelCase__ : Union[str, Any] = """coco-panoptic-id2label.json"""
elif "cityscapes" in model_name:
# this should be ok
lowerCAmelCase__ : int = 19
lowerCAmelCase__ : Union[str, Any] = """cityscapes-id2label.json"""
elif "vistas" in model_name:
# this should be ok
lowerCAmelCase__ : Tuple = 65
lowerCAmelCase__ : List[Any] = """mapillary-vistas-id2label.json"""
lowerCAmelCase__ : Any = json.load(open(hf_hub_download(UpperCamelCase , UpperCamelCase , repo_type="""dataset""" ) , """r""" ) )
lowerCAmelCase__ : Union[str, Any] = {int(UpperCamelCase ): v for k, v in idalabel.items()}
return config
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = []
# stem
# fmt: off
rename_keys.append(("""backbone.patch_embed.proj.weight""", """model.pixel_level_module.encoder.model.embeddings.patch_embeddings.projection.weight""") )
rename_keys.append(("""backbone.patch_embed.proj.bias""", """model.pixel_level_module.encoder.model.embeddings.patch_embeddings.projection.bias""") )
rename_keys.append(("""backbone.patch_embed.norm.weight""", """model.pixel_level_module.encoder.model.embeddings.norm.weight""") )
rename_keys.append(("""backbone.patch_embed.norm.bias""", """model.pixel_level_module.encoder.model.embeddings.norm.bias""") )
# stages
for i in range(len(config.backbone_config.depths ) ):
for j in range(config.backbone_config.depths[i] ):
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.norm1.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.layernorm_before.weight""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.norm1.bias""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.layernorm_before.bias""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.attn.relative_position_bias_table""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_bias_table""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.attn.relative_position_index""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_index""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.attn.proj.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.attention.output.dense.weight""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.attn.proj.bias""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.attention.output.dense.bias""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.norm2.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.layernorm_after.weight""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.norm2.bias""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.layernorm_after.bias""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.mlp.fc1.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.intermediate.dense.weight""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.mlp.fc1.bias""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.intermediate.dense.bias""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.mlp.fc2.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.output.dense.weight""") )
rename_keys.append((f"""backbone.layers.{i}.blocks.{j}.mlp.fc2.bias""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.blocks.{j}.output.dense.bias""") )
if i < 3:
rename_keys.append((f"""backbone.layers.{i}.downsample.reduction.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.downsample.reduction.weight""") )
rename_keys.append((f"""backbone.layers.{i}.downsample.norm.weight""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.downsample.norm.weight""") )
rename_keys.append((f"""backbone.layers.{i}.downsample.norm.bias""", f"""model.pixel_level_module.encoder.model.encoder.layers.{i}.downsample.norm.bias""") )
rename_keys.append((f"""backbone.norm{i}.weight""", f"""model.pixel_level_module.encoder.hidden_states_norms.{i}.weight""") )
rename_keys.append((f"""backbone.norm{i}.bias""", f"""model.pixel_level_module.encoder.hidden_states_norms.{i}.bias""") )
# FPN
rename_keys.append(("""sem_seg_head.layer_4.weight""", """model.pixel_level_module.decoder.fpn.stem.0.weight""") )
rename_keys.append(("""sem_seg_head.layer_4.norm.weight""", """model.pixel_level_module.decoder.fpn.stem.1.weight""") )
rename_keys.append(("""sem_seg_head.layer_4.norm.bias""", """model.pixel_level_module.decoder.fpn.stem.1.bias""") )
for source_index, target_index in zip(range(3 , 0 , -1 ) , range(0 , 3 ) ):
rename_keys.append((f"""sem_seg_head.adapter_{source_index}.weight""", f"""model.pixel_level_module.decoder.fpn.layers.{target_index}.proj.0.weight""") )
rename_keys.append((f"""sem_seg_head.adapter_{source_index}.norm.weight""", f"""model.pixel_level_module.decoder.fpn.layers.{target_index}.proj.1.weight""") )
rename_keys.append((f"""sem_seg_head.adapter_{source_index}.norm.bias""", f"""model.pixel_level_module.decoder.fpn.layers.{target_index}.proj.1.bias""") )
rename_keys.append((f"""sem_seg_head.layer_{source_index}.weight""", f"""model.pixel_level_module.decoder.fpn.layers.{target_index}.block.0.weight""") )
rename_keys.append((f"""sem_seg_head.layer_{source_index}.norm.weight""", f"""model.pixel_level_module.decoder.fpn.layers.{target_index}.block.1.weight""") )
rename_keys.append((f"""sem_seg_head.layer_{source_index}.norm.bias""", f"""model.pixel_level_module.decoder.fpn.layers.{target_index}.block.1.bias""") )
rename_keys.append(("""sem_seg_head.mask_features.weight""", """model.pixel_level_module.decoder.mask_projection.weight""") )
rename_keys.append(("""sem_seg_head.mask_features.bias""", """model.pixel_level_module.decoder.mask_projection.bias""") )
# Transformer decoder
for idx in range(config.decoder_config.decoder_layers ):
# self-attention out projection
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.self_attn.out_proj.weight""", f"""model.transformer_module.decoder.layers.{idx}.self_attn.out_proj.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.self_attn.out_proj.bias""", f"""model.transformer_module.decoder.layers.{idx}.self_attn.out_proj.bias""") )
# cross-attention out projection
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.multihead_attn.out_proj.weight""", f"""model.transformer_module.decoder.layers.{idx}.encoder_attn.out_proj.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.multihead_attn.out_proj.bias""", f"""model.transformer_module.decoder.layers.{idx}.encoder_attn.out_proj.bias""") )
# MLP 1
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.linear1.weight""", f"""model.transformer_module.decoder.layers.{idx}.fc1.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.linear1.bias""", f"""model.transformer_module.decoder.layers.{idx}.fc1.bias""") )
# MLP 2
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.linear2.weight""", f"""model.transformer_module.decoder.layers.{idx}.fc2.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.linear2.bias""", f"""model.transformer_module.decoder.layers.{idx}.fc2.bias""") )
# layernorm 1 (self-attention layernorm)
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.norm1.weight""", f"""model.transformer_module.decoder.layers.{idx}.self_attn_layer_norm.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.norm1.bias""", f"""model.transformer_module.decoder.layers.{idx}.self_attn_layer_norm.bias""") )
# layernorm 2 (cross-attention layernorm)
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.norm2.weight""", f"""model.transformer_module.decoder.layers.{idx}.encoder_attn_layer_norm.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.norm2.bias""", f"""model.transformer_module.decoder.layers.{idx}.encoder_attn_layer_norm.bias""") )
# layernorm 3 (final layernorm)
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.norm3.weight""", f"""model.transformer_module.decoder.layers.{idx}.final_layer_norm.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.norm3.bias""", f"""model.transformer_module.decoder.layers.{idx}.final_layer_norm.bias""") )
rename_keys.append(("""sem_seg_head.predictor.transformer.decoder.norm.weight""", """model.transformer_module.decoder.layernorm.weight""") )
rename_keys.append(("""sem_seg_head.predictor.transformer.decoder.norm.bias""", """model.transformer_module.decoder.layernorm.bias""") )
# heads on top
rename_keys.append(("""sem_seg_head.predictor.query_embed.weight""", """model.transformer_module.queries_embedder.weight""") )
rename_keys.append(("""sem_seg_head.predictor.input_proj.weight""", """model.transformer_module.input_projection.weight""") )
rename_keys.append(("""sem_seg_head.predictor.input_proj.bias""", """model.transformer_module.input_projection.bias""") )
rename_keys.append(("""sem_seg_head.predictor.class_embed.weight""", """class_predictor.weight""") )
rename_keys.append(("""sem_seg_head.predictor.class_embed.bias""", """class_predictor.bias""") )
for i in range(3 ):
rename_keys.append((f"""sem_seg_head.predictor.mask_embed.layers.{i}.weight""", f"""mask_embedder.{i}.0.weight""") )
rename_keys.append((f"""sem_seg_head.predictor.mask_embed.layers.{i}.bias""", f"""mask_embedder.{i}.0.bias""") )
# fmt: on
return rename_keys
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = dct.pop(UpperCamelCase )
lowerCAmelCase__ : List[str] = val
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = [int(backbone_config.embed_dim * 2**i ) for i in range(len(backbone_config.depths ) )]
for i in range(len(backbone_config.depths ) ):
lowerCAmelCase__ : Optional[int] = num_features[i]
for j in range(backbone_config.depths[i] ):
# fmt: off
# read in weights + bias of input projection layer (in original implementation, this is a single matrix + bias)
lowerCAmelCase__ : Dict = state_dict.pop(f"""backbone.layers.{i}.blocks.{j}.attn.qkv.weight""" )
lowerCAmelCase__ : List[str] = state_dict.pop(f"""backbone.layers.{i}.blocks.{j}.attn.qkv.bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Union[str, Any] = in_proj_weight[:dim, :]
lowerCAmelCase__ : Optional[int] = in_proj_bias[: dim]
lowerCAmelCase__ : Optional[Any] = in_proj_weight[
dim : dim * 2, :
]
lowerCAmelCase__ : int = in_proj_bias[
dim : dim * 2
]
lowerCAmelCase__ : int = in_proj_weight[
-dim :, :
]
lowerCAmelCase__ : int = in_proj_bias[-dim :]
# fmt: on
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = config.decoder_config.hidden_size
for idx in range(config.decoder_config.decoder_layers ):
# read in weights + bias of self-attention input projection layer (in the original implementation, this is a single matrix + bias)
lowerCAmelCase__ : Tuple = state_dict.pop(f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.self_attn.in_proj_weight""" )
lowerCAmelCase__ : List[Any] = state_dict.pop(f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.self_attn.in_proj_bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Dict = in_proj_weight[: hidden_size, :]
lowerCAmelCase__ : Dict = in_proj_bias[:config.hidden_size]
lowerCAmelCase__ : str = in_proj_weight[hidden_size : hidden_size * 2, :]
lowerCAmelCase__ : Union[str, Any] = in_proj_bias[hidden_size : hidden_size * 2]
lowerCAmelCase__ : int = in_proj_weight[-hidden_size :, :]
lowerCAmelCase__ : List[str] = in_proj_bias[-hidden_size :]
# read in weights + bias of cross-attention input projection layer (in the original implementation, this is a single matrix + bias)
lowerCAmelCase__ : str = state_dict.pop(f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.multihead_attn.in_proj_weight""" )
lowerCAmelCase__ : int = state_dict.pop(f"""sem_seg_head.predictor.transformer.decoder.layers.{idx}.multihead_attn.in_proj_bias""" )
# next, add query, keys and values (in that order) to the state dict
lowerCAmelCase__ : Optional[Any] = in_proj_weight[: hidden_size, :]
lowerCAmelCase__ : Optional[Any] = in_proj_bias[:config.hidden_size]
lowerCAmelCase__ : List[str] = in_proj_weight[hidden_size : hidden_size * 2, :]
lowerCAmelCase__ : Any = in_proj_bias[hidden_size : hidden_size * 2]
lowerCAmelCase__ : List[str] = in_proj_weight[-hidden_size :, :]
lowerCAmelCase__ : Dict = in_proj_bias[-hidden_size :]
# fmt: on
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : int = """http://images.cocodataset.org/val2017/000000039769.jpg"""
lowerCAmelCase__ : List[str] = Image.open(requests.get(UpperCamelCase , stream=UpperCamelCase ).raw )
return im
@torch.no_grad()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase = False ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = get_maskformer_config(UpperCamelCase )
# load original state_dict
with open(UpperCamelCase , """rb""" ) as f:
lowerCAmelCase__ : List[str] = pickle.load(UpperCamelCase )
lowerCAmelCase__ : List[Any] = data["""model"""]
# for name, param in state_dict.items():
# print(name, param.shape)
# rename keys
lowerCAmelCase__ : List[Any] = create_rename_keys(UpperCamelCase )
for src, dest in rename_keys:
rename_key(UpperCamelCase , UpperCamelCase , UpperCamelCase )
read_in_swin_q_k_v(UpperCamelCase , config.backbone_config )
read_in_decoder_q_k_v(UpperCamelCase , UpperCamelCase )
# update to torch tensors
for key, value in state_dict.items():
lowerCAmelCase__ : Optional[Any] = torch.from_numpy(UpperCamelCase )
# load 🤗 model
lowerCAmelCase__ : int = MaskFormerForInstanceSegmentation(UpperCamelCase )
model.eval()
for name, param in model.named_parameters():
print(UpperCamelCase , param.shape )
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = model.load_state_dict(UpperCamelCase , strict=UpperCamelCase )
assert missing_keys == [
"model.pixel_level_module.encoder.model.layernorm.weight",
"model.pixel_level_module.encoder.model.layernorm.bias",
]
assert len(UpperCamelCase ) == 0, f"""Unexpected keys: {unexpected_keys}"""
# verify results
lowerCAmelCase__ : List[Any] = prepare_img()
if "vistas" in model_name:
lowerCAmelCase__ : Any = 65
elif "cityscapes" in model_name:
lowerCAmelCase__ : int = 65535
else:
lowerCAmelCase__ : Optional[int] = 255
lowerCAmelCase__ : Union[str, Any] = True if """ade""" in model_name else False
lowerCAmelCase__ : Optional[int] = MaskFormerImageProcessor(ignore_index=UpperCamelCase , reduce_labels=UpperCamelCase )
lowerCAmelCase__ : List[Any] = image_processor(UpperCamelCase , return_tensors="""pt""" )
lowerCAmelCase__ : List[str] = model(**UpperCamelCase )
print("""Logits:""" , outputs.class_queries_logits[0, :3, :3] )
if model_name == "maskformer-swin-tiny-ade":
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[3.6353, -4.4770, -2.6065], [0.5081, -4.2394, -3.5343], [2.1909, -5.0353, -1.9323]] )
assert torch.allclose(outputs.class_queries_logits[0, :3, :3] , UpperCamelCase , atol=1e-4 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model and image processor to {pytorch_dump_folder_path}""" )
Path(UpperCamelCase ).mkdir(exist_ok=UpperCamelCase )
model.save_pretrained(UpperCamelCase )
image_processor.save_pretrained(UpperCamelCase )
if push_to_hub:
print("""Pushing model and image processor to the hub...""" )
model.push_to_hub(f"""nielsr/{model_name}""" )
image_processor.push_to_hub(f"""nielsr/{model_name}""" )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--model_name''',
default='''maskformer-swin-tiny-ade''',
type=str,
help=('''Name of the MaskFormer model you\'d like to convert''',),
)
parser.add_argument(
'''--checkpoint_path''',
default='''/Users/nielsrogge/Documents/MaskFormer_checkpoints/MaskFormer-Swin-tiny-ADE20k/model.pkl''',
type=str,
help='''Path to the original state dict (.pth file).''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model directory.'''
)
parser.add_argument(
'''--push_to_hub''', action='''store_true''', help='''Whether or not to push the converted model to the 🤗 hub.'''
)
_lowerCAmelCase = parser.parse_args()
convert_maskformer_checkpoint(
args.model_name, args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub
)
| 37
|
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''microsoft/focalnet-tiny''': '''https://huggingface.co/microsoft/focalnet-tiny/resolve/main/config.json''',
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Optional[Any] = '''focalnet'''
def __init__( self ,__UpperCAmelCase=224 ,__UpperCAmelCase=4 ,__UpperCAmelCase=3 ,__UpperCAmelCase=96 ,__UpperCAmelCase=False ,__UpperCAmelCase=[192, 384, 768, 768] ,__UpperCAmelCase=[2, 2, 6, 2] ,__UpperCAmelCase=[2, 2, 2, 2] ,__UpperCAmelCase=[3, 3, 3, 3] ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=4.0 ,__UpperCAmelCase=0.0 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=False ,__UpperCAmelCase=1E-4 ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=False ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-5 ,__UpperCAmelCase=32 ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> Optional[Any]:
super().__init__(**__UpperCAmelCase )
lowerCAmelCase__ : Dict = image_size
lowerCAmelCase__ : int = patch_size
lowerCAmelCase__ : str = num_channels
lowerCAmelCase__ : Dict = embed_dim
lowerCAmelCase__ : List[str] = use_conv_embed
lowerCAmelCase__ : List[Any] = hidden_sizes
lowerCAmelCase__ : Dict = depths
lowerCAmelCase__ : List[str] = focal_levels
lowerCAmelCase__ : List[str] = focal_windows
lowerCAmelCase__ : Dict = hidden_act
lowerCAmelCase__ : Dict = mlp_ratio
lowerCAmelCase__ : Tuple = hidden_dropout_prob
lowerCAmelCase__ : Tuple = drop_path_rate
lowerCAmelCase__ : Dict = use_layerscale
lowerCAmelCase__ : Optional[Any] = layerscale_value
lowerCAmelCase__ : str = use_post_layernorm
lowerCAmelCase__ : Union[str, Any] = use_post_layernorm_in_modulation
lowerCAmelCase__ : int = normalize_modulator
lowerCAmelCase__ : Optional[Any] = initializer_range
lowerCAmelCase__ : List[str] = layer_norm_eps
lowerCAmelCase__ : List[Any] = encoder_stride
lowerCAmelCase__ : Dict = ["""stem"""] + [F"""stage{idx}""" for idx in range(1 ,len(self.depths ) + 1 )]
lowerCAmelCase__ , lowerCAmelCase__ : Any = get_aligned_output_features_output_indices(
out_features=__UpperCAmelCase ,out_indices=__UpperCAmelCase ,stage_names=self.stage_names )
| 37
| 1
|
'''simple docstring'''
import argparse
import requests
import torch
from PIL import Image
from torchvision.transforms import Compose, Normalize, Resize, ToTensor
from transformers import SwinaSRConfig, SwinaSRForImageSuperResolution, SwinaSRImageProcessor
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = SwinaSRConfig()
if "Swin2SR_ClassicalSR_X4_64" in checkpoint_url:
lowerCAmelCase__ : Union[str, Any] = 4
elif "Swin2SR_CompressedSR_X4_48" in checkpoint_url:
lowerCAmelCase__ : int = 4
lowerCAmelCase__ : List[str] = 48
lowerCAmelCase__ : List[Any] = """pixelshuffle_aux"""
elif "Swin2SR_Lightweight_X2_64" in checkpoint_url:
lowerCAmelCase__ : List[Any] = [6, 6, 6, 6]
lowerCAmelCase__ : List[str] = 60
lowerCAmelCase__ : Tuple = [6, 6, 6, 6]
lowerCAmelCase__ : Optional[Any] = """pixelshuffledirect"""
elif "Swin2SR_RealworldSR_X4_64_BSRGAN_PSNR" in checkpoint_url:
lowerCAmelCase__ : List[Any] = 4
lowerCAmelCase__ : int = """nearest+conv"""
elif "Swin2SR_Jpeg_dynamic" in checkpoint_url:
lowerCAmelCase__ : List[Any] = 1
lowerCAmelCase__ : int = 1
lowerCAmelCase__ : Optional[Any] = 126
lowerCAmelCase__ : str = 7
lowerCAmelCase__ : Union[str, Any] = 255.0
lowerCAmelCase__ : Optional[int] = """"""
return config
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
if "patch_embed.proj" in name and "layers" not in name:
lowerCAmelCase__ : Dict = name.replace("""patch_embed.proj""" , """embeddings.patch_embeddings.projection""" )
if "patch_embed.norm" in name:
lowerCAmelCase__ : List[Any] = name.replace("""patch_embed.norm""" , """embeddings.patch_embeddings.layernorm""" )
if "layers" in name:
lowerCAmelCase__ : Dict = name.replace("""layers""" , """encoder.stages""" )
if "residual_group.blocks" in name:
lowerCAmelCase__ : List[Any] = name.replace("""residual_group.blocks""" , """layers""" )
if "attn.proj" in name:
lowerCAmelCase__ : Any = name.replace("""attn.proj""" , """attention.output.dense""" )
if "attn" in name:
lowerCAmelCase__ : Union[str, Any] = name.replace("""attn""" , """attention.self""" )
if "norm1" in name:
lowerCAmelCase__ : Any = name.replace("""norm1""" , """layernorm_before""" )
if "norm2" in name:
lowerCAmelCase__ : List[str] = name.replace("""norm2""" , """layernorm_after""" )
if "mlp.fc1" in name:
lowerCAmelCase__ : str = name.replace("""mlp.fc1""" , """intermediate.dense""" )
if "mlp.fc2" in name:
lowerCAmelCase__ : Any = name.replace("""mlp.fc2""" , """output.dense""" )
if "q_bias" in name:
lowerCAmelCase__ : int = name.replace("""q_bias""" , """query.bias""" )
if "k_bias" in name:
lowerCAmelCase__ : Any = name.replace("""k_bias""" , """key.bias""" )
if "v_bias" in name:
lowerCAmelCase__ : List[str] = name.replace("""v_bias""" , """value.bias""" )
if "cpb_mlp" in name:
lowerCAmelCase__ : Optional[Any] = name.replace("""cpb_mlp""" , """continuous_position_bias_mlp""" )
if "patch_embed.proj" in name:
lowerCAmelCase__ : Optional[Any] = name.replace("""patch_embed.proj""" , """patch_embed.projection""" )
if name == "norm.weight":
lowerCAmelCase__ : Any = """layernorm.weight"""
if name == "norm.bias":
lowerCAmelCase__ : Optional[Any] = """layernorm.bias"""
if "conv_first" in name:
lowerCAmelCase__ : Tuple = name.replace("""conv_first""" , """first_convolution""" )
if (
"upsample" in name
or "conv_before_upsample" in name
or "conv_bicubic" in name
or "conv_up" in name
or "conv_hr" in name
or "conv_last" in name
or "aux" in name
):
# heads
if "conv_last" in name:
lowerCAmelCase__ : Any = name.replace("""conv_last""" , """final_convolution""" )
if config.upsampler in ["pixelshuffle", "pixelshuffle_aux", "nearest+conv"]:
if "conv_before_upsample.0" in name:
lowerCAmelCase__ : Union[str, Any] = name.replace("""conv_before_upsample.0""" , """conv_before_upsample""" )
if "upsample.0" in name:
lowerCAmelCase__ : Optional[int] = name.replace("""upsample.0""" , """upsample.convolution_0""" )
if "upsample.2" in name:
lowerCAmelCase__ : Optional[int] = name.replace("""upsample.2""" , """upsample.convolution_1""" )
lowerCAmelCase__ : int = """upsample.""" + name
elif config.upsampler == "pixelshuffledirect":
lowerCAmelCase__ : str = name.replace("""upsample.0.weight""" , """upsample.conv.weight""" )
lowerCAmelCase__ : List[str] = name.replace("""upsample.0.bias""" , """upsample.conv.bias""" )
else:
pass
else:
lowerCAmelCase__ : Any = """swin2sr.""" + name
return name
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
for key in orig_state_dict.copy().keys():
lowerCAmelCase__ : Optional[Any] = orig_state_dict.pop(UpperCamelCase )
if "qkv" in key:
lowerCAmelCase__ : Dict = key.split(""".""" )
lowerCAmelCase__ : Any = int(key_split[1] )
lowerCAmelCase__ : Dict = int(key_split[4] )
lowerCAmelCase__ : List[str] = config.embed_dim
if "weight" in key:
lowerCAmelCase__ : List[str] = val[:dim, :]
lowerCAmelCase__ : int = val[dim : dim * 2, :]
lowerCAmelCase__ : Dict = val[-dim:, :]
else:
lowerCAmelCase__ : Union[str, Any] = val[:dim]
lowerCAmelCase__ : Optional[Any] = val[dim : dim * 2]
lowerCAmelCase__ : List[Any] = val[-dim:]
pass
else:
lowerCAmelCase__ : Optional[int] = val
return orig_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[Any] = get_config(UpperCamelCase )
lowerCAmelCase__ : Any = SwinaSRForImageSuperResolution(UpperCamelCase )
model.eval()
lowerCAmelCase__ : Tuple = torch.hub.load_state_dict_from_url(UpperCamelCase , map_location="""cpu""" )
lowerCAmelCase__ : Optional[Any] = convert_state_dict(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ , lowerCAmelCase__ : Dict = model.load_state_dict(UpperCamelCase , strict=UpperCamelCase )
if len(UpperCamelCase ) > 0:
raise ValueError("""Missing keys when converting: {}""".format(UpperCamelCase ) )
for key in unexpected_keys:
if not ("relative_position_index" in key or "relative_coords_table" in key or "self_mask" in key):
raise ValueError(f"""Unexpected key {key} in state_dict""" )
# verify values
lowerCAmelCase__ : Union[str, Any] = """https://github.com/mv-lab/swin2sr/blob/main/testsets/real-inputs/shanghai.jpg?raw=true"""
lowerCAmelCase__ : Dict = Image.open(requests.get(UpperCamelCase , stream=UpperCamelCase ).raw ).convert("""RGB""" )
lowerCAmelCase__ : int = SwinaSRImageProcessor()
# pixel_values = processor(image, return_tensors="pt").pixel_values
lowerCAmelCase__ : str = 126 if """Jpeg""" in checkpoint_url else 256
lowerCAmelCase__ : List[Any] = Compose(
[
Resize((image_size, image_size) ),
ToTensor(),
Normalize(mean=[0.485, 0.456, 0.406] , std=[0.229, 0.224, 0.225] ),
] )
lowerCAmelCase__ : Optional[int] = transforms(UpperCamelCase ).unsqueeze(0 )
if config.num_channels == 1:
lowerCAmelCase__ : int = pixel_values[:, 0, :, :].unsqueeze(1 )
lowerCAmelCase__ : Optional[Any] = model(UpperCamelCase )
# assert values
if "Swin2SR_ClassicalSR_X2_64" in checkpoint_url:
lowerCAmelCase__ : Tuple = torch.Size([1, 3, 512, 512] )
lowerCAmelCase__ : Optional[Any] = torch.tensor(
[[-0.7087, -0.7138, -0.6721], [-0.8340, -0.8095, -0.7298], [-0.9149, -0.8414, -0.7940]] )
elif "Swin2SR_ClassicalSR_X4_64" in checkpoint_url:
lowerCAmelCase__ : Union[str, Any] = torch.Size([1, 3, 1024, 1024] )
lowerCAmelCase__ : str = torch.tensor(
[[-0.7775, -0.8105, -0.8933], [-0.7764, -0.8356, -0.9225], [-0.7976, -0.8686, -0.9579]] )
elif "Swin2SR_CompressedSR_X4_48" in checkpoint_url:
# TODO values didn't match exactly here
lowerCAmelCase__ : str = torch.Size([1, 3, 1024, 1024] )
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.8035, -0.7504, -0.7491], [-0.8538, -0.8124, -0.7782], [-0.8804, -0.8651, -0.8493]] )
elif "Swin2SR_Lightweight_X2_64" in checkpoint_url:
lowerCAmelCase__ : Tuple = torch.Size([1, 3, 512, 512] )
lowerCAmelCase__ : Dict = torch.tensor(
[[-0.7669, -0.8662, -0.8767], [-0.8810, -0.9962, -0.9820], [-0.9340, -1.0322, -1.1149]] )
elif "Swin2SR_RealworldSR_X4_64_BSRGAN_PSNR" in checkpoint_url:
lowerCAmelCase__ : Tuple = torch.Size([1, 3, 1024, 1024] )
lowerCAmelCase__ : int = torch.tensor(
[[-0.5238, -0.5557, -0.6321], [-0.6016, -0.5903, -0.6391], [-0.6244, -0.6334, -0.6889]] )
assert (
outputs.reconstruction.shape == expected_shape
), f"""Shape of reconstruction should be {expected_shape}, but is {outputs.reconstruction.shape}"""
assert torch.allclose(outputs.reconstruction[0, 0, :3, :3] , UpperCamelCase , atol=1e-3 )
print("""Looks ok!""" )
lowerCAmelCase__ : Optional[Any] = {
"""https://github.com/mv-lab/swin2sr/releases/download/v0.0.1/Swin2SR_ClassicalSR_X2_64.pth""": (
"""swin2SR-classical-sr-x2-64"""
),
"""https://github.com/mv-lab/swin2sr/releases/download/v0.0.1/Swin2SR_ClassicalSR_X4_64.pth""": (
"""swin2SR-classical-sr-x4-64"""
),
"""https://github.com/mv-lab/swin2sr/releases/download/v0.0.1/Swin2SR_CompressedSR_X4_48.pth""": (
"""swin2SR-compressed-sr-x4-48"""
),
"""https://github.com/mv-lab/swin2sr/releases/download/v0.0.1/Swin2SR_Lightweight_X2_64.pth""": (
"""swin2SR-lightweight-x2-64"""
),
"""https://github.com/mv-lab/swin2sr/releases/download/v0.0.1/Swin2SR_RealworldSR_X4_64_BSRGAN_PSNR.pth""": (
"""swin2SR-realworld-sr-x4-64-bsrgan-psnr"""
),
}
lowerCAmelCase__ : Optional[int] = url_to_name[checkpoint_url]
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(UpperCamelCase )
print(f"""Saving image processor to {pytorch_dump_folder_path}""" )
processor.save_pretrained(UpperCamelCase )
if push_to_hub:
model.push_to_hub(f"""caidas/{model_name}""" )
processor.push_to_hub(f"""caidas/{model_name}""" )
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
'''--checkpoint_url''',
default='''https://github.com/mv-lab/swin2sr/releases/download/v0.0.1/Swin2SR_ClassicalSR_X2_64.pth''',
type=str,
help='''URL of the original Swin2SR checkpoint you\'d like to convert.''',
)
parser.add_argument(
'''--pytorch_dump_folder_path''', default=None, type=str, help='''Path to the output PyTorch model directory.'''
)
parser.add_argument('''--push_to_hub''', action='''store_true''', help='''Whether to push the converted model to the hub.''')
_lowerCAmelCase = parser.parse_args()
convert_swinasr_checkpoint(args.checkpoint_url, args.pytorch_dump_folder_path, args.push_to_hub)
| 37
|
'''simple docstring'''
import argparse
import os.path as osp
import re
import torch
from safetensors.torch import load_file, save_file
# =================#
# UNet Conversion #
# =================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''time_embed.0.weight''', '''time_embedding.linear_1.weight'''),
('''time_embed.0.bias''', '''time_embedding.linear_1.bias'''),
('''time_embed.2.weight''', '''time_embedding.linear_2.weight'''),
('''time_embed.2.bias''', '''time_embedding.linear_2.bias'''),
('''input_blocks.0.0.weight''', '''conv_in.weight'''),
('''input_blocks.0.0.bias''', '''conv_in.bias'''),
('''out.0.weight''', '''conv_norm_out.weight'''),
('''out.0.bias''', '''conv_norm_out.bias'''),
('''out.2.weight''', '''conv_out.weight'''),
('''out.2.bias''', '''conv_out.bias'''),
]
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''in_layers.0''', '''norm1'''),
('''in_layers.2''', '''conv1'''),
('''out_layers.0''', '''norm2'''),
('''out_layers.3''', '''conv2'''),
('''emb_layers.1''', '''time_emb_proj'''),
('''skip_connection''', '''conv_shortcut'''),
]
_lowerCAmelCase = []
# hardcoded number of downblocks and resnets/attentions...
# would need smarter logic for other networks.
for i in range(4):
# loop over downblocks/upblocks
for j in range(2):
# loop over resnets/attentions for downblocks
_lowerCAmelCase = F"""down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.0."""
unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
if i < 3:
# no attention layers in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""input_blocks.{3*i + j + 1}.1."""
unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
for j in range(3):
# loop over resnets/attentions for upblocks
_lowerCAmelCase = F"""up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.0."""
unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
if i > 0:
# no attention layers in up_blocks.0
_lowerCAmelCase = F"""up_blocks.{i}.attentions.{j}."""
_lowerCAmelCase = F"""output_blocks.{3*i + j}.1."""
unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
if i < 3:
# no downsample in down_blocks.3
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0.conv."""
_lowerCAmelCase = F"""input_blocks.{3*(i+1)}.0.op."""
unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
# no upsample in up_blocks.3
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""output_blocks.{3*i + 2}.{1 if i == 0 else 2}."""
unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
_lowerCAmelCase = '''mid_block.attentions.0.'''
_lowerCAmelCase = '''middle_block.1.'''
unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
for j in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{j}."""
_lowerCAmelCase = F"""middle_block.{2*j}."""
unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Any = {k: k for k in unet_state_dict.keys()}
for sd_name, hf_name in unet_conversion_map:
lowerCAmelCase__ : Optional[int] = sd_name
for k, v in mapping.items():
if "resnets" in k:
for sd_part, hf_part in unet_conversion_map_resnet:
lowerCAmelCase__ : Any = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
for k, v in mapping.items():
for sd_part, hf_part in unet_conversion_map_layer:
lowerCAmelCase__ : List[Any] = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
lowerCAmelCase__ : Tuple = {v: unet_state_dict[k] for k, v in mapping.items()}
return new_state_dict
# ================#
# VAE Conversion #
# ================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''nin_shortcut''', '''conv_shortcut'''),
('''norm_out''', '''conv_norm_out'''),
('''mid.attn_1.''', '''mid_block.attentions.0.'''),
]
for i in range(4):
# down_blocks have two resnets
for j in range(2):
_lowerCAmelCase = F"""encoder.down_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""encoder.down.{i}.block.{j}."""
vae_conversion_map.append((sd_down_prefix, hf_down_prefix))
if i < 3:
_lowerCAmelCase = F"""down_blocks.{i}.downsamplers.0."""
_lowerCAmelCase = F"""down.{i}.downsample."""
vae_conversion_map.append((sd_downsample_prefix, hf_downsample_prefix))
_lowerCAmelCase = F"""up_blocks.{i}.upsamplers.0."""
_lowerCAmelCase = F"""up.{3-i}.upsample."""
vae_conversion_map.append((sd_upsample_prefix, hf_upsample_prefix))
# up_blocks have three resnets
# also, up blocks in hf are numbered in reverse from sd
for j in range(3):
_lowerCAmelCase = F"""decoder.up_blocks.{i}.resnets.{j}."""
_lowerCAmelCase = F"""decoder.up.{3-i}.block.{j}."""
vae_conversion_map.append((sd_up_prefix, hf_up_prefix))
# this part accounts for mid blocks in both the encoder and the decoder
for i in range(2):
_lowerCAmelCase = F"""mid_block.resnets.{i}."""
_lowerCAmelCase = F"""mid.block_{i+1}."""
vae_conversion_map.append((sd_mid_res_prefix, hf_mid_res_prefix))
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''norm.''', '''group_norm.'''),
('''q.''', '''query.'''),
('''k.''', '''key.'''),
('''v.''', '''value.'''),
('''proj_out.''', '''proj_attn.'''),
]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return w.reshape(*w.shape , 1 , 1 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = {k: k for k in vae_state_dict.keys()}
for k, v in mapping.items():
for sd_part, hf_part in vae_conversion_map:
lowerCAmelCase__ : str = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = v
for k, v in mapping.items():
if "attentions" in k:
for sd_part, hf_part in vae_conversion_map_attn:
lowerCAmelCase__ : Dict = v.replace(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : List[Any] = v
lowerCAmelCase__ : Union[str, Any] = {v: vae_state_dict[k] for k, v in mapping.items()}
lowerCAmelCase__ : Tuple = ["""q""", """k""", """v""", """proj_out"""]
for k, v in new_state_dict.items():
for weight_name in weights_to_convert:
if f"""mid.attn_1.{weight_name}.weight""" in k:
print(f"""Reshaping {k} for SD format""" )
lowerCAmelCase__ : Optional[int] = reshape_weight_for_sd(UpperCamelCase )
return new_state_dict
# =========================#
# Text Encoder Conversion #
# =========================#
_lowerCAmelCase = [
# (stable-diffusion, HF Diffusers)
('''resblocks.''', '''text_model.encoder.layers.'''),
('''ln_1''', '''layer_norm1'''),
('''ln_2''', '''layer_norm2'''),
('''.c_fc.''', '''.fc1.'''),
('''.c_proj.''', '''.fc2.'''),
('''.attn''', '''.self_attn'''),
('''ln_final.''', '''transformer.text_model.final_layer_norm.'''),
('''token_embedding.weight''', '''transformer.text_model.embeddings.token_embedding.weight'''),
('''positional_embedding''', '''transformer.text_model.embeddings.position_embedding.weight'''),
]
_lowerCAmelCase = {re.escape(x[1]): x[0] for x in textenc_conversion_lst}
_lowerCAmelCase = re.compile('''|'''.join(protected.keys()))
# Ordering is from https://github.com/pytorch/pytorch/blob/master/test/cpp/api/modules.cpp
_lowerCAmelCase = {'''q''': 0, '''k''': 1, '''v''': 2}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = {}
lowerCAmelCase__ : int = {}
lowerCAmelCase__ : List[Any] = {}
for k, v in text_enc_dict.items():
if (
k.endswith(""".self_attn.q_proj.weight""" )
or k.endswith(""".self_attn.k_proj.weight""" )
or k.endswith(""".self_attn.v_proj.weight""" )
):
lowerCAmelCase__ : Optional[int] = k[: -len(""".q_proj.weight""" )]
lowerCAmelCase__ : Tuple = k[-len("""q_proj.weight""" )]
if k_pre not in capture_qkv_weight:
lowerCAmelCase__ : List[Any] = [None, None, None]
lowerCAmelCase__ : Dict = v
continue
if (
k.endswith(""".self_attn.q_proj.bias""" )
or k.endswith(""".self_attn.k_proj.bias""" )
or k.endswith(""".self_attn.v_proj.bias""" )
):
lowerCAmelCase__ : str = k[: -len(""".q_proj.bias""" )]
lowerCAmelCase__ : List[str] = k[-len("""q_proj.bias""" )]
if k_pre not in capture_qkv_bias:
lowerCAmelCase__ : Union[str, Any] = [None, None, None]
lowerCAmelCase__ : Any = v
continue
lowerCAmelCase__ : Dict = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = v
for k_pre, tensors in capture_qkv_weight.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : Any = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : Tuple = torch.cat(UpperCamelCase )
for k_pre, tensors in capture_qkv_bias.items():
if None in tensors:
raise Exception("""CORRUPTED MODEL: one of the q-k-v values for the text encoder was missing""" )
lowerCAmelCase__ : str = textenc_pattern.sub(lambda UpperCamelCase : protected[re.escape(m.group(0 ) )] , UpperCamelCase )
lowerCAmelCase__ : List[Any] = torch.cat(UpperCamelCase )
return new_state_dict
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return text_enc_dict
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument('''--model_path''', default=None, type=str, required=True, help='''Path to the model to convert.''')
parser.add_argument('''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the output model.''')
parser.add_argument('''--half''', action='''store_true''', help='''Save weights in half precision.''')
parser.add_argument(
'''--use_safetensors''', action='''store_true''', help='''Save weights use safetensors, default is ckpt.'''
)
_lowerCAmelCase = parser.parse_args()
assert args.model_path is not None, "Must provide a model path!"
assert args.checkpoint_path is not None, "Must provide a checkpoint path!"
# Path for safetensors
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.safetensors''')
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''model.safetensors''')
# Load models from safetensors if it exists, if it doesn't pytorch
if osp.exists(unet_path):
_lowerCAmelCase = load_file(unet_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''unet''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(unet_path, map_location='''cpu''')
if osp.exists(vae_path):
_lowerCAmelCase = load_file(vae_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''vae''', '''diffusion_pytorch_model.bin''')
_lowerCAmelCase = torch.load(vae_path, map_location='''cpu''')
if osp.exists(text_enc_path):
_lowerCAmelCase = load_file(text_enc_path, device='''cpu''')
else:
_lowerCAmelCase = osp.join(args.model_path, '''text_encoder''', '''pytorch_model.bin''')
_lowerCAmelCase = torch.load(text_enc_path, map_location='''cpu''')
# Convert the UNet model
_lowerCAmelCase = convert_unet_state_dict(unet_state_dict)
_lowerCAmelCase = {'''model.diffusion_model.''' + k: v for k, v in unet_state_dict.items()}
# Convert the VAE model
_lowerCAmelCase = convert_vae_state_dict(vae_state_dict)
_lowerCAmelCase = {'''first_stage_model.''' + k: v for k, v in vae_state_dict.items()}
# Easiest way to identify v2.0 model seems to be that the text encoder (OpenCLIP) is deeper
_lowerCAmelCase = '''text_model.encoder.layers.22.layer_norm2.bias''' in text_enc_dict
if is_vaa_model:
# Need to add the tag 'transformer' in advance so we can knock it out from the final layer-norm
_lowerCAmelCase = {'''transformer.''' + k: v for k, v in text_enc_dict.items()}
_lowerCAmelCase = convert_text_enc_state_dict_vaa(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.model.''' + k: v for k, v in text_enc_dict.items()}
else:
_lowerCAmelCase = convert_text_enc_state_dict(text_enc_dict)
_lowerCAmelCase = {'''cond_stage_model.transformer.''' + k: v for k, v in text_enc_dict.items()}
# Put together new checkpoint
_lowerCAmelCase = {**unet_state_dict, **vae_state_dict, **text_enc_dict}
if args.half:
_lowerCAmelCase = {k: v.half() for k, v in state_dict.items()}
if args.use_safetensors:
save_file(state_dict, args.checkpoint_path)
else:
_lowerCAmelCase = {'''state_dict''': state_dict}
torch.save(state_dict, args.checkpoint_path)
| 37
| 1
|
'''simple docstring'''
# 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 lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = '''facebook/bart-large-mnli'''
__lowercase : Optional[int] = (
'''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.'''
)
__lowercase : str = '''text_classifier'''
__lowercase : Union[str, Any] = AutoTokenizer
__lowercase : int = AutoModelForSequenceClassification
__lowercase : Dict = ['''text''', ['''text''']]
__lowercase : int = ['''text''']
def UpperCAmelCase_ ( self ) -> List[Any]:
super().setup()
lowerCAmelCase__ : Optional[int] = self.model.config
lowerCAmelCase__ : str = -1
for idx, label in config.idalabel.items():
if label.lower().startswith("""entail""" ):
lowerCAmelCase__ : List[Any] = 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 ,__UpperCAmelCase ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : Optional[int] = 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 ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[str] = outputs.logits
lowerCAmelCase__ : Union[str, Any] = torch.argmax(logits[:, 2] ).item()
return self._labels[label_id]
| 37
|
'''simple docstring'''
import os
from bleurt import score # From: git+https://github.com/google-research/bleurt.git
import datasets
_lowerCAmelCase = datasets.logging.get_logger(__name__)
_lowerCAmelCase = '''\
@inproceedings{bleurt,
title={BLEURT: Learning Robust Metrics for Text Generation},
author={Thibault Sellam and Dipanjan Das and Ankur P. Parikh},
booktitle={ACL},
year={2020},
url={https://arxiv.org/abs/2004.04696}
}
'''
_lowerCAmelCase = '''\
BLEURT a learnt evaluation metric for Natural Language Generation. It is built using multiple phases of transfer learning starting from a pretrained BERT model (Devlin et al. 2018)
and then employing another pre-training phrase using synthetic data. Finally it is trained on WMT human annotations. You may run BLEURT out-of-the-box or fine-tune
it for your specific application (the latter is expected to perform better).
See the project\'s README at https://github.com/google-research/bleurt#readme for more information.
'''
_lowerCAmelCase = '''
BLEURT score.
Args:
`predictions` (list of str): prediction/candidate sentences
`references` (list of str): reference sentences
`checkpoint` BLEURT checkpoint. Will default to BLEURT-tiny if None.
Returns:
\'scores\': List of scores.
Examples:
>>> predictions = ["hello there", "general kenobi"]
>>> references = ["hello there", "general kenobi"]
>>> bleurt = datasets.load_metric("bleurt")
>>> results = bleurt.compute(predictions=predictions, references=references)
>>> print([round(v, 2) for v in results["scores"]])
[1.03, 1.04]
'''
_lowerCAmelCase = {
'''bleurt-tiny-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-128.zip''',
'''bleurt-tiny-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-tiny-512.zip''',
'''bleurt-base-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-128.zip''',
'''bleurt-base-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-base-512.zip''',
'''bleurt-large-128''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-128.zip''',
'''bleurt-large-512''': '''https://storage.googleapis.com/bleurt-oss/bleurt-large-512.zip''',
'''BLEURT-20-D3''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D3.zip''',
'''BLEURT-20-D6''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D6.zip''',
'''BLEURT-20-D12''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20-D12.zip''',
'''BLEURT-20''': '''https://storage.googleapis.com/bleurt-oss-21/BLEURT-20.zip''',
}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,homepage="""https://github.com/google-research/bleurt""" ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(
{
"""predictions""": datasets.Value("""string""" ,id="""sequence""" ),
"""references""": datasets.Value("""string""" ,id="""sequence""" ),
} ) ,codebase_urls=["""https://github.com/google-research/bleurt"""] ,reference_urls=["""https://github.com/google-research/bleurt""", """https://arxiv.org/abs/2004.04696"""] ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
# check that config name specifies a valid BLEURT model
if self.config_name == "default":
logger.warning(
"""Using default BLEURT-Base checkpoint for sequence maximum length 128. """
"""You can use a bigger model for better results with e.g.: datasets.load_metric('bleurt', 'bleurt-large-512').""" )
lowerCAmelCase__ : str = """bleurt-base-128"""
if self.config_name.lower() in CHECKPOINT_URLS:
lowerCAmelCase__ : Union[str, Any] = self.config_name.lower()
elif self.config_name.upper() in CHECKPOINT_URLS:
lowerCAmelCase__ : List[Any] = self.config_name.upper()
else:
raise KeyError(
F"""{self.config_name} model not found. You should supply the name of a model checkpoint for bleurt in {CHECKPOINT_URLS.keys()}""" )
# download the model checkpoint specified by self.config_name and set up the scorer
lowerCAmelCase__ : int = dl_manager.download_and_extract(CHECKPOINT_URLS[checkpoint_name] )
lowerCAmelCase__ : Dict = score.BleurtScorer(os.path.join(__UpperCAmelCase ,__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Union[str, Any] = self.scorer.score(references=__UpperCAmelCase ,candidates=__UpperCAmelCase )
return {"scores": scores}
| 37
| 1
|
'''simple docstring'''
from dataclasses import dataclass, field
from typing import Tuple
from ..utils import cached_property, is_torch_available, is_torch_tpu_available, logging, requires_backends
from .benchmark_args_utils import BenchmarkArguments
if is_torch_available():
import torch
if is_torch_tpu_available(check_device=False):
import torch_xla.core.xla_model as xm
_lowerCAmelCase = logging.get_logger(__name__)
@dataclass
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = [
'''no_inference''',
'''no_cuda''',
'''no_tpu''',
'''no_speed''',
'''no_memory''',
'''no_env_print''',
'''no_multi_process''',
]
def __init__( self ,**__UpperCAmelCase ) -> str:
for deprecated_arg in self.deprecated_args:
if deprecated_arg in kwargs:
lowerCAmelCase__ : Union[str, Any] = deprecated_arg[3:]
setattr(self ,__UpperCAmelCase ,not kwargs.pop(__UpperCAmelCase ) )
logger.warning(
F"""{deprecated_arg} is depreciated. Please use --no_{positive_arg} or"""
F""" {positive_arg}={kwargs[positive_arg]}""" )
lowerCAmelCase__ : Union[str, Any] = kwargs.pop("""torchscript""" ,self.torchscript )
lowerCAmelCase__ : Any = kwargs.pop("""torch_xla_tpu_print_metrics""" ,self.torch_xla_tpu_print_metrics )
lowerCAmelCase__ : Optional[Any] = kwargs.pop("""fp16_opt_level""" ,self.fpaa_opt_level )
super().__init__(**__UpperCAmelCase )
__lowercase : bool = field(default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Trace the models using torchscript'''} )
__lowercase : bool = field(default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Print Xla/PyTorch tpu metrics'''} )
__lowercase : str = field(
default='''O1''' , metadata={
'''help''': (
'''For fp16: Apex AMP optimization level selected in [\'O0\', \'O1\', \'O2\', and \'O3\']. '''
'''See details at https://nvidia.github.io/apex/amp.html'''
)
} , )
@cached_property
def UpperCAmelCase_ ( self ) -> Tuple["torch.device", int]:
requires_backends(self ,["""torch"""] )
logger.info("""PyTorch: setting up devices""" )
if not self.cuda:
lowerCAmelCase__ : str = torch.device("""cpu""" )
lowerCAmelCase__ : List[str] = 0
elif is_torch_tpu_available():
lowerCAmelCase__ : Union[str, Any] = xm.xla_device()
lowerCAmelCase__ : Dict = 0
else:
lowerCAmelCase__ : str = torch.device("""cuda""" if torch.cuda.is_available() else """cpu""" )
lowerCAmelCase__ : Union[str, Any] = torch.cuda.device_count()
return device, n_gpu
@property
def UpperCAmelCase_ ( self ) -> Optional[int]:
return is_torch_tpu_available() and self.tpu
@property
def UpperCAmelCase_ ( self ) -> int:
requires_backends(self ,["""torch"""] )
# TODO(PVP): currently only single GPU is supported
return torch.cuda.current_device()
@property
def UpperCAmelCase_ ( self ) -> "torch.device":
requires_backends(self ,["""torch"""] )
return self._setup_devices[0]
@property
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
requires_backends(self ,["""torch"""] )
return self._setup_devices[1]
@property
def UpperCAmelCase_ ( self ) -> Dict:
return self.n_gpu > 0
| 37
|
'''simple docstring'''
import uuid
from typing import Any, Dict, List, Optional, Union
from ..utils import add_end_docstrings, is_tf_available, is_torch_available, logging
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_tf_available():
import tensorflow as tf
if is_torch_available():
import torch
_lowerCAmelCase = logging.get_logger(__name__)
class lowerCAmelCase_:
'''simple docstring'''
def __init__( self ,__UpperCAmelCase = None ,__UpperCAmelCase = None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ) -> str:
if not conversation_id:
lowerCAmelCase__ : List[str] = uuid.uuida()
if past_user_inputs is None:
lowerCAmelCase__ : List[Any] = []
if generated_responses is None:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : uuid.UUID = conversation_id
lowerCAmelCase__ : List[str] = past_user_inputs
lowerCAmelCase__ : List[str] = generated_responses
lowerCAmelCase__ : Optional[str] = text
def __eq__( self ,__UpperCAmelCase ) -> Dict:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
return False
if self.uuid == other.uuid:
return True
return (
self.new_user_input == other.new_user_input
and self.past_user_inputs == other.past_user_inputs
and self.generated_responses == other.generated_responses
)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = False ) -> Optional[Any]:
if self.new_user_input:
if overwrite:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" was overwritten """
F"""with: \"{text}\".""" )
lowerCAmelCase__ : Optional[int] = text
else:
logger.warning(
F"""User input added while unprocessed input was existing: \"{self.new_user_input}\" new input """
F"""ignored: \"{text}\". Set `overwrite` to True to overwrite unprocessed user input""" )
else:
lowerCAmelCase__ : Optional[Any] = text
def UpperCAmelCase_ ( self ) -> List[Any]:
if self.new_user_input:
self.past_user_inputs.append(self.new_user_input )
lowerCAmelCase__ : Union[str, Any] = None
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
self.generated_responses.append(__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> List[str]:
for user_input, generated_response in zip(self.past_user_inputs ,self.generated_responses ):
yield True, user_input
yield False, generated_response
if self.new_user_input:
yield True, self.new_user_input
def __repr__( self ) -> Tuple:
lowerCAmelCase__ : Tuple = F"""Conversation id: {self.uuid} \n"""
for is_user, text in self.iter_texts():
lowerCAmelCase__ : Any = """user""" if is_user else """bot"""
output += F"""{name} >> {text} \n"""
return output
@add_end_docstrings(
SCREAMING_SNAKE_CASE_ , R'''
min_length_for_response (`int`, *optional*, defaults to 32):
The minimum length (in number of tokens) for a response.
minimum_tokens (`int`, *optional*, defaults to 10):
The minimum length of tokens to leave for a response.
''' , )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,*__UpperCAmelCase ,**__UpperCAmelCase ) -> Tuple:
super().__init__(*__UpperCAmelCase ,**__UpperCAmelCase )
if self.tokenizer.pad_token_id is None:
lowerCAmelCase__ : Tuple = self.tokenizer.eos_token
def UpperCAmelCase_ ( self ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,__UpperCAmelCase=None ,**__UpperCAmelCase ) -> Optional[int]:
lowerCAmelCase__ : List[Any] = {}
lowerCAmelCase__ : Optional[int] = {}
lowerCAmelCase__ : List[str] = {}
if min_length_for_response is not None:
lowerCAmelCase__ : Any = min_length_for_response
if minimum_tokens is not None:
lowerCAmelCase__ : Optional[int] = minimum_tokens
if "max_length" in generate_kwargs:
lowerCAmelCase__ : Optional[Any] = generate_kwargs["""max_length"""]
# self.max_length = generate_kwargs.get("max_length", self.model.config.max_length)
if clean_up_tokenization_spaces is not None:
lowerCAmelCase__ : int = clean_up_tokenization_spaces
if generate_kwargs:
forward_params.update(__UpperCAmelCase )
return preprocess_params, forward_params, postprocess_params
def __call__( self ,__UpperCAmelCase ,__UpperCAmelCase=0 ,**__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : Optional[int] = super().__call__(__UpperCAmelCase ,num_workers=__UpperCAmelCase ,**__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase ) and len(__UpperCAmelCase ) == 1:
return outputs[0]
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=32 ) -> Dict[str, Any]:
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
raise ValueError("""ConversationalPipeline, expects Conversation as inputs""" )
if conversation.new_user_input is None:
raise ValueError(
F"""Conversation with UUID {type(conversation.uuid )} does not contain new user input to process. """
"""Add user inputs with the conversation's `add_user_input` method""" )
if hasattr(self.tokenizer ,"""_build_conversation_input_ids""" ):
lowerCAmelCase__ : str = self.tokenizer._build_conversation_input_ids(__UpperCAmelCase )
else:
# If the tokenizer cannot handle conversations, we default to only the old version
lowerCAmelCase__ : List[Any] = self._legacy_parse_and_tokenize(__UpperCAmelCase )
if self.framework == "pt":
lowerCAmelCase__ : List[Any] = torch.LongTensor([input_ids] )
elif self.framework == "tf":
lowerCAmelCase__ : Dict = tf.constant([input_ids] )
return {"input_ids": input_ids, "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=10 ,**__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[Any] = generate_kwargs.get("""max_length""" ,self.model.config.max_length )
lowerCAmelCase__ : Optional[Any] = model_inputs["""input_ids"""].shape[1]
if max_length - minimum_tokens < n:
logger.warning(F"""Conversation input is to long ({n}), trimming it to ({max_length} - {minimum_tokens})""" )
lowerCAmelCase__ : str = max_length - minimum_tokens
lowerCAmelCase__ : Union[str, Any] = model_inputs["""input_ids"""][:, -trim:]
if "attention_mask" in model_inputs:
lowerCAmelCase__ : Tuple = model_inputs["""attention_mask"""][:, -trim:]
lowerCAmelCase__ : str = model_inputs.pop("""conversation""" )
lowerCAmelCase__ : Union[str, Any] = max_length
lowerCAmelCase__ : Any = self.model.generate(**__UpperCAmelCase ,**__UpperCAmelCase )
if self.model.config.is_encoder_decoder:
lowerCAmelCase__ : int = 1
else:
lowerCAmelCase__ : int = n
return {"output_ids": output_ids[:, start_position:], "conversation": conversation}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=True ) -> List[str]:
lowerCAmelCase__ : Optional[int] = model_outputs["""output_ids"""]
lowerCAmelCase__ : Tuple = self.tokenizer.decode(
output_ids[0] ,skip_special_tokens=__UpperCAmelCase ,clean_up_tokenization_spaces=__UpperCAmelCase ,)
lowerCAmelCase__ : Union[str, Any] = model_outputs["""conversation"""]
conversation.mark_processed()
conversation.append_response(__UpperCAmelCase )
return conversation
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Dict = self.tokenizer.eos_token_id
lowerCAmelCase__ : int = []
for is_user, text in conversation.iter_texts():
if eos_token_id is not None:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) + [eos_token_id] )
else:
input_ids.extend(self.tokenizer.encode(__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase ) )
if len(__UpperCAmelCase ) > self.tokenizer.model_max_length:
lowerCAmelCase__ : Optional[Any] = input_ids[-self.tokenizer.model_max_length :]
return input_ids
| 37
| 1
|
'''simple docstring'''
from dataclasses import dataclass, field
from typing import Tuple
from ..utils import cached_property, is_tf_available, logging, requires_backends
from .benchmark_args_utils import BenchmarkArguments
if is_tf_available():
import tensorflow as tf
_lowerCAmelCase = logging.get_logger(__name__)
@dataclass
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Any = [
'''no_inference''',
'''no_cuda''',
'''no_tpu''',
'''no_speed''',
'''no_memory''',
'''no_env_print''',
'''no_multi_process''',
]
def __init__( self ,**__UpperCAmelCase ) -> int:
for deprecated_arg in self.deprecated_args:
if deprecated_arg in kwargs:
lowerCAmelCase__ : str = deprecated_arg[3:]
lowerCAmelCase__ : Union[str, Any] = not kwargs.pop(__UpperCAmelCase )
logger.warning(
F"""{deprecated_arg} is depreciated. Please use --no-{positive_arg} or"""
F""" {positive_arg}={kwargs[positive_arg]}""" )
lowerCAmelCase__ : Union[str, Any] = kwargs.pop("""tpu_name""" ,self.tpu_name )
lowerCAmelCase__ : str = kwargs.pop("""device_idx""" ,self.device_idx )
lowerCAmelCase__ : int = kwargs.pop("""eager_mode""" ,self.eager_mode )
lowerCAmelCase__ : int = kwargs.pop("""use_xla""" ,self.use_xla )
super().__init__(**__UpperCAmelCase )
__lowercase : str = field(
default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Name of TPU'''} , )
__lowercase : int = field(
default=0 , metadata={'''help''': '''CPU / GPU device index. Defaults to 0.'''} , )
__lowercase : bool = field(default=SCREAMING_SNAKE_CASE_ , metadata={'''help''': '''Benchmark models in eager model.'''} )
__lowercase : bool = field(
default=SCREAMING_SNAKE_CASE_ , metadata={
'''help''': '''Benchmark models using XLA JIT compilation. Note that `eager_model` has to be set to `False`.'''
} , )
@cached_property
def UpperCAmelCase_ ( self ) -> Tuple["tf.distribute.cluster_resolver.TPUClusterResolver"]:
requires_backends(self ,["""tf"""] )
lowerCAmelCase__ : Any = None
if self.tpu:
try:
if self.tpu_name:
lowerCAmelCase__ : List[Any] = tf.distribute.cluster_resolver.TPUClusterResolver(self.tpu_name )
else:
lowerCAmelCase__ : Any = tf.distribute.cluster_resolver.TPUClusterResolver()
except ValueError:
lowerCAmelCase__ : Optional[Any] = None
return tpu
@cached_property
def UpperCAmelCase_ ( self ) -> Tuple["tf.distribute.Strategy", "tf.distribute.cluster_resolver.TPUClusterResolver"]:
requires_backends(self ,["""tf"""] )
if self.is_tpu:
tf.config.experimental_connect_to_cluster(self._setup_tpu )
tf.tpu.experimental.initialize_tpu_system(self._setup_tpu )
lowerCAmelCase__ : Any = tf.distribute.TPUStrategy(self._setup_tpu )
else:
# currently no multi gpu is allowed
if self.is_gpu:
# TODO: Currently only single GPU is supported
tf.config.set_visible_devices(self.gpu_list[self.device_idx] ,"""GPU""" )
lowerCAmelCase__ : str = tf.distribute.OneDeviceStrategy(device=F"""/gpu:{self.device_idx}""" )
else:
tf.config.set_visible_devices([] ,"""GPU""" ) # disable GPU
lowerCAmelCase__ : str = tf.distribute.OneDeviceStrategy(device=F"""/cpu:{self.device_idx}""" )
return strategy
@property
def UpperCAmelCase_ ( self ) -> bool:
requires_backends(self ,["""tf"""] )
return self._setup_tpu is not None
@property
def UpperCAmelCase_ ( self ) -> "tf.distribute.Strategy":
requires_backends(self ,["""tf"""] )
return self._setup_strategy
@property
def UpperCAmelCase_ ( self ) -> str:
requires_backends(self ,["""tf"""] )
return tf.config.list_physical_devices("""GPU""" )
@property
def UpperCAmelCase_ ( self ) -> int:
requires_backends(self ,["""tf"""] )
if self.cuda:
return len(self.gpu_list )
return 0
@property
def UpperCAmelCase_ ( self ) -> bool:
return self.n_gpu > 0
| 37
|
'''simple docstring'''
from collections import UserDict
from typing import List, Union
from ..utils import (
add_end_docstrings,
is_tf_available,
is_torch_available,
is_vision_available,
logging,
requires_backends,
)
from .base import PIPELINE_INIT_ARGS, Pipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
from ..tf_utils import stable_softmax
_lowerCAmelCase = logging.get_logger(__name__)
@add_end_docstrings(SCREAMING_SNAKE_CASE_ )
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
def __init__( self ,**__UpperCAmelCase ) -> Tuple:
super().__init__(**__UpperCAmelCase )
requires_backends(self ,"""vision""" )
self.check_model_type(
TF_MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING
if self.framework == """tf"""
else MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING )
def __call__( self ,__UpperCAmelCase ,**__UpperCAmelCase ) -> str:
return super().__call__(__UpperCAmelCase ,**__UpperCAmelCase )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> str:
lowerCAmelCase__ : List[Any] = {}
if "candidate_labels" in kwargs:
lowerCAmelCase__ : int = kwargs["""candidate_labels"""]
if "hypothesis_template" in kwargs:
lowerCAmelCase__ : Optional[int] = kwargs["""hypothesis_template"""]
return preprocess_params, {}, {}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=None ,__UpperCAmelCase="This is a photo of {}." ) -> int:
lowerCAmelCase__ : str = load_image(__UpperCAmelCase )
lowerCAmelCase__ : Dict = self.image_processor(images=[image] ,return_tensors=self.framework )
lowerCAmelCase__ : List[Any] = candidate_labels
lowerCAmelCase__ : List[str] = [hypothesis_template.format(__UpperCAmelCase ) for x in candidate_labels]
lowerCAmelCase__ : Optional[Any] = self.tokenizer(__UpperCAmelCase ,return_tensors=self.framework ,padding=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = [text_inputs]
return inputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = model_inputs.pop("""candidate_labels""" )
lowerCAmelCase__ : Union[str, Any] = model_inputs.pop("""text_inputs""" )
if isinstance(text_inputs[0] ,__UpperCAmelCase ):
lowerCAmelCase__ : int = text_inputs[0]
else:
# Batching case.
lowerCAmelCase__ : Dict = text_inputs[0][0]
lowerCAmelCase__ : Any = self.model(**__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = {
"""candidate_labels""": candidate_labels,
"""logits""": outputs.logits_per_image,
}
return model_outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
lowerCAmelCase__ : Union[str, Any] = model_outputs.pop("""candidate_labels""" )
lowerCAmelCase__ : List[str] = model_outputs["""logits"""][0]
if self.framework == "pt":
lowerCAmelCase__ : List[str] = logits.softmax(dim=-1 ).squeeze(-1 )
lowerCAmelCase__ : Optional[Any] = probs.tolist()
if not isinstance(__UpperCAmelCase ,__UpperCAmelCase ):
lowerCAmelCase__ : Dict = [scores]
elif self.framework == "tf":
lowerCAmelCase__ : Any = stable_softmax(__UpperCAmelCase ,axis=-1 )
lowerCAmelCase__ : List[Any] = probs.numpy().tolist()
else:
raise ValueError(F"""Unsupported framework: {self.framework}""" )
lowerCAmelCase__ : Tuple = [
{"""score""": score, """label""": candidate_label}
for score, candidate_label in sorted(zip(__UpperCAmelCase ,__UpperCAmelCase ) ,key=lambda __UpperCAmelCase : -x[0] )
]
return result
| 37
| 1
|
'''simple docstring'''
import unittest
from transformers import SPIECE_UNDERLINE, XLNetTokenizer, XLNetTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, slow
from ...test_tokenization_common import TokenizerTesterMixin
_lowerCAmelCase = get_tests_dir('''fixtures/test_sentencepiece.model''')
@require_sentencepiece
@require_tokenizers
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Union[str, Any] = XLNetTokenizer
__lowercase : List[Any] = XLNetTokenizerFast
__lowercase : List[str] = True
__lowercase : str = True
def UpperCAmelCase_ ( self ) -> int:
super().setUp()
# We have a SentencePiece fixture for testing
lowerCAmelCase__ : Optional[int] = XLNetTokenizer(__UpperCAmelCase ,keep_accents=__UpperCAmelCase )
tokenizer.sanitize_special_tokens()
tokenizer.save_pretrained(self.tmpdirname )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : List[Any] = """<s>"""
lowerCAmelCase__ : Optional[int] = 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 ) -> Dict:
lowerCAmelCase__ : Optional[int] = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] ,"""<unk>""" )
self.assertEqual(vocab_keys[1] ,"""<s>""" )
self.assertEqual(vocab_keys[-1] ,"""<eod>""" )
self.assertEqual(len(__UpperCAmelCase ) ,1006 )
def UpperCAmelCase_ ( self ) -> Dict:
self.assertEqual(self.get_tokenizer().vocab_size ,1000 )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Tuple = XLNetTokenizer(__UpperCAmelCase ,keep_accents=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = 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] )
lowerCAmelCase__ : Tuple = 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""",
""".""",
] ,)
lowerCAmelCase__ : List[Any] = 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] )
lowerCAmelCase__ : List[Any] = 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>""",
""".""",
] ,)
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Optional[Any] = XLNetTokenizer(__UpperCAmelCase ,do_lower_case=__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = 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""",
"""se""",
""".""",
] ,)
self.assertListEqual(tokenizer.tokenize("""H\u00E9llo""" ) ,["""▁he""", """ll""", """o"""] )
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Dict = XLNetTokenizer(__UpperCAmelCase ,do_lower_case=__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = 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""",
"""se""",
""".""",
] ,)
@slow
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : Tuple = XLNetTokenizer.from_pretrained("""xlnet-base-cased""" )
lowerCAmelCase__ : int = tokenizer.encode("""sequence builders""" ,add_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : Dict = tokenizer.encode("""multi-sequence build""" ,add_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : Dict = tokenizer.build_inputs_with_special_tokens(__UpperCAmelCase )
lowerCAmelCase__ : Any = tokenizer.build_inputs_with_special_tokens(__UpperCAmelCase ,__UpperCAmelCase )
assert encoded_sentence == text + [4, 3]
assert encoded_pair == text + [4] + text_a + [4, 3]
@slow
def UpperCAmelCase_ ( self ) -> int:
# fmt: off
lowerCAmelCase__ : Optional[Any] = {"""input_ids""": [[17, 2_1442, 270, 17, 10, 1_4645, 318, 34, 17, 4546, 3145, 787, 13, 7752, 2_2018, 23, 21, 17, 4546, 3145, 787, 13, 3352, 1_4431, 13, 5500, 11, 1176, 580, 13, 1_6819, 4797, 23, 17, 10, 1_7135, 658, 19, 457, 7932, 13, 184, 19, 3154, 1_7135, 6468, 19, 1404, 1_2269, 19, 4229, 5356, 1_6264, 46, 19, 17, 2_0545, 1_0395, 9, 9, 9, 11, 28, 6421, 9531, 2_0729, 17, 10, 353, 1_7022, 11, 21, 6421, 9531, 1_6949, 17, 10, 1_1509, 753, 11, 33, 95, 2421, 7385, 956, 1_4431, 2626, 25, 842, 7385, 4836, 21, 1429, 2272, 9855, 3120, 161, 2_4738, 19, 1_3203, 658, 218, 787, 21, 430, 1_8482, 847, 2637, 9, 4, 3], [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 322, 2_2178, 27, 1064, 22, 956, 13, 1_1101, 1429, 5854, 2_4313, 1_8953, 40, 422, 2_4366, 68, 1758, 37, 1_0483, 1_4257, 31, 207, 263, 21, 203, 3773, 25, 71, 9735, 9, 4, 3], [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 32, 2049, 3442, 17, 1_3894, 3380, 23, 95, 18, 1_7634, 2288, 9, 4, 3]], """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, 2], [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2], [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2]], """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], [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 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, 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, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 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, 1, 1]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=__UpperCAmelCase ,model_name="""xlnet-base-cased""" ,revision="""c841166438c31ec7ca9a106dee7bb312b73ae511""" ,)
| 37
|
'''simple docstring'''
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv('''TEST_SAGEMAKER''' , '''False''' ) ) is not True , reason='''Skipping test because should only be run when releasing minor transformers version''' , )
@pytest.mark.usefixtures('''sm_env''' )
@parameterized_class(
[
{
'''framework''': '''pytorch''',
'''script''': '''run_glue.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_5_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''pytorch''',
'''script''': '''run_ddp.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.7, '''eval_loss''': 0.6},
},
{
'''framework''': '''tensorflow''',
'''script''': '''run_tf_dist.py''',
'''model_name_or_path''': '''distilbert-base-cased''',
'''instance_type''': '''ml.p3.16xlarge''',
'''results''': {'''train_runtime''': 6_0_0, '''eval_accuracy''': 0.6, '''eval_loss''': 0.7},
},
] )
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.framework == "pytorch":
subprocess.run(
F"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() ,encoding="""utf-8""" ,check=__UpperCAmelCase ,)
assert hasattr(self ,"""env""" )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Optional[int] = F"""{self.env.base_job_name}-{instance_count}-{'ddp' if 'ddp' in self.script else 'smd'}"""
# distributed data settings
lowerCAmelCase__ : Any = {"""smdistributed""": {"""dataparallel""": {"""enabled""": True}}} if self.script != """run_ddp.py""" else None
# creates estimator
return HuggingFace(
entry_point=self.script ,source_dir=self.env.test_path ,role=self.env.role ,image_uri=self.env.image_uri ,base_job_name=__UpperCAmelCase ,instance_count=__UpperCAmelCase ,instance_type=self.instance_type ,debugger_hook_config=__UpperCAmelCase ,hyperparameters={**self.env.distributed_hyperparameters, """model_name_or_path""": self.model_name_or_path} ,metric_definitions=self.env.metric_definitions ,distribution=__UpperCAmelCase ,py_version="""py36""" ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
TrainingJobAnalytics(__UpperCAmelCase ).export_csv(F"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(2,)] )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
# create estimator
lowerCAmelCase__ : List[Any] = self.create_estimator(__UpperCAmelCase )
# run training
estimator.fit()
# result dataframe
lowerCAmelCase__ : Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
lowerCAmelCase__ : int = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
lowerCAmelCase__ : List[Any] = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
lowerCAmelCase__ : List[str] = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" ,99_9999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(F"""{estimator.latest_training_job.name}.json""" ,"""w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} ,__UpperCAmelCase )
| 37
| 1
|
'''simple docstring'''
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import torch
import torch.nn as nn
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput
from .embeddings import GaussianFourierProjection, TimestepEmbedding, Timesteps
from .modeling_utils import ModelMixin
from .unet_ad_blocks import get_down_block, get_mid_block, get_out_block, get_up_block
@dataclass
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : torch.FloatTensor
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@register_to_config
def __init__( self ,__UpperCAmelCase = 6_5536 ,__UpperCAmelCase = None ,__UpperCAmelCase = 2 ,__UpperCAmelCase = 2 ,__UpperCAmelCase = 0 ,__UpperCAmelCase = "fourier" ,__UpperCAmelCase = True ,__UpperCAmelCase = False ,__UpperCAmelCase = 0.0 ,__UpperCAmelCase = ("DownBlock1DNoSkip", "DownBlock1D", "AttnDownBlock1D") ,__UpperCAmelCase = ("AttnUpBlock1D", "UpBlock1D", "UpBlock1DNoSkip") ,__UpperCAmelCase = "UNetMidBlock1D" ,__UpperCAmelCase = None ,__UpperCAmelCase = (32, 32, 64) ,__UpperCAmelCase = None ,__UpperCAmelCase = 8 ,__UpperCAmelCase = 1 ,__UpperCAmelCase = False ,) -> str:
super().__init__()
lowerCAmelCase__ : int = sample_size
# time
if time_embedding_type == "fourier":
lowerCAmelCase__ : Dict = GaussianFourierProjection(
embedding_size=8 ,set_W_to_weight=__UpperCAmelCase ,log=__UpperCAmelCase ,flip_sin_to_cos=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = 2 * block_out_channels[0]
elif time_embedding_type == "positional":
lowerCAmelCase__ : Any = Timesteps(
block_out_channels[0] ,flip_sin_to_cos=__UpperCAmelCase ,downscale_freq_shift=__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = block_out_channels[0]
if use_timestep_embedding:
lowerCAmelCase__ : List[str] = block_out_channels[0] * 4
lowerCAmelCase__ : Union[str, Any] = TimestepEmbedding(
in_channels=__UpperCAmelCase ,time_embed_dim=__UpperCAmelCase ,act_fn=__UpperCAmelCase ,out_dim=block_out_channels[0] ,)
lowerCAmelCase__ : Dict = nn.ModuleList([] )
lowerCAmelCase__ : int = None
lowerCAmelCase__ : List[str] = nn.ModuleList([] )
lowerCAmelCase__ : Any = None
# down
lowerCAmelCase__ : Optional[int] = in_channels
for i, down_block_type in enumerate(__UpperCAmelCase ):
lowerCAmelCase__ : str = output_channel
lowerCAmelCase__ : Union[str, Any] = block_out_channels[i]
if i == 0:
input_channel += extra_in_channels
lowerCAmelCase__ : Union[str, Any] = i == len(__UpperCAmelCase ) - 1
lowerCAmelCase__ : Dict = get_down_block(
__UpperCAmelCase ,num_layers=__UpperCAmelCase ,in_channels=__UpperCAmelCase ,out_channels=__UpperCAmelCase ,temb_channels=block_out_channels[0] ,add_downsample=not is_final_block or downsample_each_block ,)
self.down_blocks.append(__UpperCAmelCase )
# mid
lowerCAmelCase__ : Any = get_mid_block(
__UpperCAmelCase ,in_channels=block_out_channels[-1] ,mid_channels=block_out_channels[-1] ,out_channels=block_out_channels[-1] ,embed_dim=block_out_channels[0] ,num_layers=__UpperCAmelCase ,add_downsample=__UpperCAmelCase ,)
# up
lowerCAmelCase__ : Union[str, Any] = list(reversed(__UpperCAmelCase ) )
lowerCAmelCase__ : Optional[Any] = reversed_block_out_channels[0]
if out_block_type is None:
lowerCAmelCase__ : List[str] = out_channels
else:
lowerCAmelCase__ : Any = block_out_channels[0]
for i, up_block_type in enumerate(__UpperCAmelCase ):
lowerCAmelCase__ : int = output_channel
lowerCAmelCase__ : List[str] = (
reversed_block_out_channels[i + 1] if i < len(__UpperCAmelCase ) - 1 else final_upsample_channels
)
lowerCAmelCase__ : Any = i == len(__UpperCAmelCase ) - 1
lowerCAmelCase__ : int = get_up_block(
__UpperCAmelCase ,num_layers=__UpperCAmelCase ,in_channels=__UpperCAmelCase ,out_channels=__UpperCAmelCase ,temb_channels=block_out_channels[0] ,add_upsample=not is_final_block ,)
self.up_blocks.append(__UpperCAmelCase )
lowerCAmelCase__ : int = output_channel
# out
lowerCAmelCase__ : str = norm_num_groups if norm_num_groups is not None else min(block_out_channels[0] // 4 ,32 )
lowerCAmelCase__ : Tuple = get_out_block(
out_block_type=__UpperCAmelCase ,num_groups_out=__UpperCAmelCase ,embed_dim=block_out_channels[0] ,out_channels=__UpperCAmelCase ,act_fn=__UpperCAmelCase ,fc_dim=block_out_channels[-1] // 4 ,)
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase = True ,) -> Union[UNetaDOutput, Tuple]:
lowerCAmelCase__ : Optional[Any] = timestep
if not torch.is_tensor(__UpperCAmelCase ):
lowerCAmelCase__ : Optional[Any] = torch.tensor([timesteps] ,dtype=torch.long ,device=sample.device )
elif torch.is_tensor(__UpperCAmelCase ) and len(timesteps.shape ) == 0:
lowerCAmelCase__ : Optional[int] = timesteps[None].to(sample.device )
lowerCAmelCase__ : Any = self.time_proj(__UpperCAmelCase )
if self.config.use_timestep_embedding:
lowerCAmelCase__ : str = self.time_mlp(__UpperCAmelCase )
else:
lowerCAmelCase__ : Union[str, Any] = timestep_embed[..., None]
lowerCAmelCase__ : Any = timestep_embed.repeat([1, 1, sample.shape[2]] ).to(sample.dtype )
lowerCAmelCase__ : Optional[Any] = timestep_embed.broadcast_to((sample.shape[:1] + timestep_embed.shape[1:]) )
# 2. down
lowerCAmelCase__ : Union[str, Any] = ()
for downsample_block in self.down_blocks:
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = downsample_block(hidden_states=__UpperCAmelCase ,temb=__UpperCAmelCase )
down_block_res_samples += res_samples
# 3. mid
if self.mid_block:
lowerCAmelCase__ : int = self.mid_block(__UpperCAmelCase ,__UpperCAmelCase )
# 4. up
for i, upsample_block in enumerate(self.up_blocks ):
lowerCAmelCase__ : Optional[Any] = down_block_res_samples[-1:]
lowerCAmelCase__ : Optional[Any] = down_block_res_samples[:-1]
lowerCAmelCase__ : Tuple = upsample_block(__UpperCAmelCase ,res_hidden_states_tuple=__UpperCAmelCase ,temb=__UpperCAmelCase )
# 5. post-process
if self.out_block:
lowerCAmelCase__ : Optional[Any] = self.out_block(__UpperCAmelCase ,__UpperCAmelCase )
if not return_dict:
return (sample,)
return UNetaDOutput(sample=__UpperCAmelCase )
| 37
|
'''simple docstring'''
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {'''vocab_file''': '''spiece.model'''}
_lowerCAmelCase = {
'''vocab_file''': {
'''albert-base-v1''': '''https://huggingface.co/albert-base-v1/resolve/main/spiece.model''',
'''albert-large-v1''': '''https://huggingface.co/albert-large-v1/resolve/main/spiece.model''',
'''albert-xlarge-v1''': '''https://huggingface.co/albert-xlarge-v1/resolve/main/spiece.model''',
'''albert-xxlarge-v1''': '''https://huggingface.co/albert-xxlarge-v1/resolve/main/spiece.model''',
'''albert-base-v2''': '''https://huggingface.co/albert-base-v2/resolve/main/spiece.model''',
'''albert-large-v2''': '''https://huggingface.co/albert-large-v2/resolve/main/spiece.model''',
'''albert-xlarge-v2''': '''https://huggingface.co/albert-xlarge-v2/resolve/main/spiece.model''',
'''albert-xxlarge-v2''': '''https://huggingface.co/albert-xxlarge-v2/resolve/main/spiece.model''',
}
}
_lowerCAmelCase = {
'''albert-base-v1''': 512,
'''albert-large-v1''': 512,
'''albert-xlarge-v1''': 512,
'''albert-xxlarge-v1''': 512,
'''albert-base-v2''': 512,
'''albert-large-v2''': 512,
'''albert-xlarge-v2''': 512,
'''albert-xxlarge-v2''': 512,
}
_lowerCAmelCase = '''▁'''
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Any = VOCAB_FILES_NAMES
__lowercase : Optional[Any] = PRETRAINED_VOCAB_FILES_MAP
__lowercase : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=True ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<unk>" ,__UpperCAmelCase="[SEP]" ,__UpperCAmelCase="<pad>" ,__UpperCAmelCase="[CLS]" ,__UpperCAmelCase="[MASK]" ,__UpperCAmelCase = None ,**__UpperCAmelCase ,) -> None:
# Mask token behave like a normal word, i.e. include the space before it and
# is included in the raw text, there should be a match in a non-normalized sentence.
lowerCAmelCase__ : Tuple = (
AddedToken(__UpperCAmelCase ,lstrip=__UpperCAmelCase ,rstrip=__UpperCAmelCase ,normalized=__UpperCAmelCase )
if isinstance(__UpperCAmelCase ,__UpperCAmelCase )
else mask_token
)
lowerCAmelCase__ : Any = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=__UpperCAmelCase ,remove_space=__UpperCAmelCase ,keep_accents=__UpperCAmelCase ,bos_token=__UpperCAmelCase ,eos_token=__UpperCAmelCase ,unk_token=__UpperCAmelCase ,sep_token=__UpperCAmelCase ,pad_token=__UpperCAmelCase ,cls_token=__UpperCAmelCase ,mask_token=__UpperCAmelCase ,sp_model_kwargs=self.sp_model_kwargs ,**__UpperCAmelCase ,)
lowerCAmelCase__ : str = do_lower_case
lowerCAmelCase__ : int = remove_space
lowerCAmelCase__ : Tuple = keep_accents
lowerCAmelCase__ : Any = vocab_file
lowerCAmelCase__ : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(__UpperCAmelCase )
@property
def UpperCAmelCase_ ( self ) -> Optional[int]:
return len(self.sp_model )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Union[str, Any] = {self.convert_ids_to_tokens(__UpperCAmelCase ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = self.__dict__.copy()
lowerCAmelCase__ : Optional[Any] = None
return state
def __setstate__( self ,__UpperCAmelCase ) -> List[Any]:
lowerCAmelCase__ : List[str] = d
# for backward compatibility
if not hasattr(self ,"""sp_model_kwargs""" ):
lowerCAmelCase__ : Union[str, Any] = {}
lowerCAmelCase__ : Any = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[int]:
if self.remove_space:
lowerCAmelCase__ : int = """ """.join(inputs.strip().split() )
else:
lowerCAmelCase__ : str = inputs
lowerCAmelCase__ : Tuple = outputs.replace("""``""" ,"""\"""" ).replace("""''""" ,"""\"""" )
if not self.keep_accents:
lowerCAmelCase__ : Any = unicodedata.normalize("""NFKD""" ,__UpperCAmelCase )
lowerCAmelCase__ : Dict = """""".join([c for c in outputs if not unicodedata.combining(__UpperCAmelCase )] )
if self.do_lower_case:
lowerCAmelCase__ : Tuple = outputs.lower()
return outputs
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
lowerCAmelCase__ : List[str] = self.preprocess_text(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = self.sp_model.encode(__UpperCAmelCase ,out_type=__UpperCAmelCase )
lowerCAmelCase__ : str = []
for piece in pieces:
if len(__UpperCAmelCase ) > 1 and piece[-1] == str(""",""" ) and piece[-2].isdigit():
lowerCAmelCase__ : List[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(__UpperCAmelCase ,"""""" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
lowerCAmelCase__ : str = cur_pieces[1:]
else:
lowerCAmelCase__ : int = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(__UpperCAmelCase )
else:
new_pieces.append(__UpperCAmelCase )
return new_pieces
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[str]:
return self.sp_model.PieceToId(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Any:
return self.sp_model.IdToPiece(__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> str:
lowerCAmelCase__ : str = []
lowerCAmelCase__ : Tuple = """"""
lowerCAmelCase__ : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(__UpperCAmelCase ) + token
lowerCAmelCase__ : Union[str, Any] = True
lowerCAmelCase__ : List[str] = []
else:
current_sub_tokens.append(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = False
out_string += self.sp_model.decode(__UpperCAmelCase )
return out_string.strip()
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : int = [self.sep_token_id]
lowerCAmelCase__ : Dict = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ,__UpperCAmelCase = False ) -> List[int]:
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 not None:
return [1] + ([0] * len(__UpperCAmelCase )) + [1] + ([0] * len(__UpperCAmelCase )) + [1]
return [1] + ([0] * len(__UpperCAmelCase )) + [1]
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> List[int]:
lowerCAmelCase__ : List[str] = [self.sep_token_id]
lowerCAmelCase__ : Tuple = [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 ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : int = 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:
lowerCAmelCase__ : List[Any] = self.sp_model.serialized_model_proto()
fi.write(__UpperCAmelCase )
return (out_vocab_file,)
| 37
| 1
|
'''simple docstring'''
import torch
from transformers import AutoModel
class lowerCAmelCase_( torch.nn.Module ):
'''simple docstring'''
def __init__( self ,__UpperCAmelCase="sayef/fsner-bert-base-uncased" ) -> str:
super(__UpperCAmelCase ,self ).__init__()
lowerCAmelCase__ : List[str] = AutoModel.from_pretrained(__UpperCAmelCase ,return_dict=__UpperCAmelCase )
lowerCAmelCase__ : str = torch.nn.CosineSimilarity(3 ,1E-08 )
lowerCAmelCase__ : List[Any] = torch.nn.Softmax(dim=1 )
def UpperCAmelCase_ ( self ,**__UpperCAmelCase ) -> int:
return self.bert(**__UpperCAmelCase ).last_hidden_state
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> List[Any]:
return token_embeddings.sum(2 ,keepdim=__UpperCAmelCase )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase=1 ) -> Dict:
return self.softmax(T * self.cos(__UpperCAmelCase ,__UpperCAmelCase ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Union[str, Any]:
lowerCAmelCase__ : List[Any] = W_supports["""sizes"""].tolist()
lowerCAmelCase__ : List[Any] = W_supports["""start_token_id"""].item()
lowerCAmelCase__ : int = W_supports["""end_token_id"""].item()
del W_supports["sizes"]
del W_supports["start_token_id"]
del W_supports["end_token_id"]
lowerCAmelCase__ : List[str] = self.BERT(**__UpperCAmelCase )
lowerCAmelCase__ : Tuple = self.BERT(**__UpperCAmelCase )
lowerCAmelCase__ : List[str] = None
lowerCAmelCase__ : int = None
lowerCAmelCase__ : str = W_supports["""input_ids"""] == start_token_id
lowerCAmelCase__ : Optional[Any] = W_supports["""input_ids"""] == end_token_id
for i, size in enumerate(__UpperCAmelCase ):
if i == 0:
lowerCAmelCase__ : Any = 0
else:
lowerCAmelCase__ : List[Any] = support_sizes[i - 1]
lowerCAmelCase__ : Optional[Any] = S[s : s + size][start_token_masks[s : s + size]]
lowerCAmelCase__ : List[str] = S[s : s + size][end_token_masks[s : s + size]]
lowerCAmelCase__ : int = torch.matmul(q[i] ,s_start.T ).sum(1 ).softmax(0 )
lowerCAmelCase__ : Optional[int] = torch.matmul(q[i] ,s_end.T ).sum(1 ).softmax(0 )
if p_starts is not None:
lowerCAmelCase__ : Optional[int] = torch.vstack((p_starts, p_start) )
lowerCAmelCase__ : Union[str, Any] = torch.vstack((p_ends, p_end) )
else:
lowerCAmelCase__ : Optional[Any] = p_start
lowerCAmelCase__ : str = p_end
return p_starts, p_ends
| 37
|
'''simple docstring'''
import json
import os
from typing import Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''vocab_file''': '''vocab.json''',
'''merges_file''': '''merges.txt''',
}
_lowerCAmelCase = {
'''vocab_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-vocab.json'''},
'''merges_file''': {'''ctrl''': '''https://raw.githubusercontent.com/salesforce/ctrl/master/ctrl-merges.txt'''},
}
_lowerCAmelCase = {
'''ctrl''': 256,
}
_lowerCAmelCase = {
'''Pregnancy''': 16_8629,
'''Christianity''': 7675,
'''Explain''': 10_6423,
'''Fitness''': 6_3440,
'''Saving''': 6_3163,
'''Ask''': 2_7171,
'''Ass''': 9_5985,
'''Joke''': 16_3509,
'''Questions''': 4_5622,
'''Thoughts''': 4_9605,
'''Retail''': 5_2342,
'''Feminism''': 16_4338,
'''Writing''': 1_1992,
'''Atheism''': 19_2263,
'''Netflix''': 4_8616,
'''Computing''': 3_9639,
'''Opinion''': 4_3213,
'''Alone''': 4_4967,
'''Funny''': 5_8917,
'''Gaming''': 4_0358,
'''Human''': 4088,
'''India''': 1331,
'''Joker''': 7_7138,
'''Diet''': 3_6206,
'''Legal''': 1_1859,
'''Norman''': 4939,
'''Tip''': 7_2689,
'''Weight''': 5_2343,
'''Movies''': 4_6273,
'''Running''': 2_3425,
'''Science''': 2090,
'''Horror''': 3_7793,
'''Confession''': 6_0572,
'''Finance''': 1_2250,
'''Politics''': 1_6360,
'''Scary''': 19_1985,
'''Support''': 1_2654,
'''Technologies''': 3_2516,
'''Teenage''': 6_6160,
'''Event''': 3_2769,
'''Learned''': 6_7460,
'''Notion''': 18_2770,
'''Wikipedia''': 3_7583,
'''Books''': 6665,
'''Extract''': 7_6050,
'''Confessions''': 10_2701,
'''Conspiracy''': 7_5932,
'''Links''': 6_3674,
'''Narcissus''': 15_0425,
'''Relationship''': 5_4766,
'''Relationships''': 13_4796,
'''Reviews''': 4_1671,
'''News''': 4256,
'''Translation''': 2_6820,
'''multilingual''': 12_8406,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[int] = set()
lowerCAmelCase__ : str = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
lowerCAmelCase__ : List[Any] = char
lowerCAmelCase__ : Optional[Any] = set(UpperCamelCase )
return pairs
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : Dict = VOCAB_FILES_NAMES
__lowercase : Dict = PRETRAINED_VOCAB_FILES_MAP
__lowercase : Optional[Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
__lowercase : Any = CONTROL_CODES
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase ,__UpperCAmelCase="<unk>" ,**__UpperCAmelCase ) -> Optional[int]:
super().__init__(unk_token=__UpperCAmelCase ,**__UpperCAmelCase )
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as vocab_handle:
lowerCAmelCase__ : int = json.load(__UpperCAmelCase )
lowerCAmelCase__ : Any = {v: k for k, v in self.encoder.items()}
with open(__UpperCAmelCase ,encoding="""utf-8""" ) as merges_handle:
lowerCAmelCase__ : Union[str, Any] = merges_handle.read().split("""\n""" )[1:-1]
lowerCAmelCase__ : Optional[Any] = [tuple(merge.split() ) for merge in merges]
lowerCAmelCase__ : int = dict(zip(__UpperCAmelCase ,range(len(__UpperCAmelCase ) ) ) )
lowerCAmelCase__ : List[Any] = {}
@property
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return len(self.encoder )
def UpperCAmelCase_ ( self ) -> Optional[Any]:
return dict(self.encoder ,**self.added_tokens_encoder )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> int:
if token in self.cache:
return self.cache[token]
lowerCAmelCase__ : Optional[Any] = tuple(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = tuple(list(word[:-1] ) + [word[-1] + """</w>"""] )
lowerCAmelCase__ : Any = get_pairs(__UpperCAmelCase )
if not pairs:
return token
while True:
lowerCAmelCase__ : List[str] = min(__UpperCAmelCase ,key=lambda __UpperCAmelCase : self.bpe_ranks.get(__UpperCAmelCase ,float("""inf""" ) ) )
if bigram not in self.bpe_ranks:
break
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = bigram
lowerCAmelCase__ : List[str] = []
lowerCAmelCase__ : Dict = 0
while i < len(__UpperCAmelCase ):
try:
lowerCAmelCase__ : Optional[Any] = word.index(__UpperCAmelCase ,__UpperCAmelCase )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
lowerCAmelCase__ : Dict = 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
lowerCAmelCase__ : Tuple = tuple(__UpperCAmelCase )
lowerCAmelCase__ : Tuple = new_word
if len(__UpperCAmelCase ) == 1:
break
else:
lowerCAmelCase__ : Dict = get_pairs(__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = """@@ """.join(__UpperCAmelCase )
lowerCAmelCase__ : Optional[Any] = word[:-4]
lowerCAmelCase__ : Optional[Any] = word
return word
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Dict:
lowerCAmelCase__ : Optional[int] = []
lowerCAmelCase__ : Any = re.findall(R"""\S+\n?""" ,__UpperCAmelCase )
for token in words:
split_tokens.extend(list(self.bpe(__UpperCAmelCase ).split(""" """ ) ) )
return split_tokens
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Optional[Any]:
return self.encoder.get(__UpperCAmelCase ,self.encoder.get(self.unk_token ) )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
return self.decoder.get(__UpperCAmelCase ,self.unk_token )
def UpperCAmelCase_ ( self ,__UpperCAmelCase ) -> Tuple:
lowerCAmelCase__ : Any = """ """.join(__UpperCAmelCase ).replace("""@@ """ ,"""""" ).strip()
return out_string
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase = None ) -> Tuple[str]:
if not os.path.isdir(__UpperCAmelCase ):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""" )
return
lowerCAmelCase__ : List[Any] = os.path.join(
__UpperCAmelCase ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
lowerCAmelCase__ : Optional[int] = 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""" )
lowerCAmelCase__ : int = 0
with open(__UpperCAmelCase ,"""w""" ,encoding="""utf-8""" ) as writer:
writer.write("""#version: 0.2\n""" )
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 {merge_file}: BPE merge indices are not consecutive."""
""" Please check that the tokenizer is not corrupted!""" )
lowerCAmelCase__ : Dict = token_index
writer.write(""" """.join(__UpperCAmelCase ) + """\n""" )
index += 1
return vocab_file, merge_file
# def decode(self, token_ids, skip_special_tokens=False, clean_up_tokenization_spaces=True):
# filtered_tokens = ' '.join(self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens))
# tokens_generated_so_far = re.sub('(@@ )', '', string=filtered_tokens)
# tokens_generated_so_far = re.sub('(@@ ?$)', '', string=tokens_generated_so_far)
# return ''.join(tokens_generated_so_far)
| 37
| 1
|
'''simple docstring'''
import datasets
import faiss
import numpy as np
import streamlit as st
import torch
from elasticsearch import Elasticsearch
from elia_utils import (
embed_questions_for_retrieval,
make_qa_sas_model,
qa_sas_generate,
query_es_index,
query_qa_dense_index,
)
import transformers
from transformers import AutoModel, AutoModelForSeqaSeqLM, AutoTokenizer
_lowerCAmelCase = '''bart'''
_lowerCAmelCase = True
@st.cache(allow_output_mutation=UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
if LOAD_DENSE_INDEX:
lowerCAmelCase__ : Union[str, Any] = AutoTokenizer.from_pretrained("""yjernite/retribert-base-uncased""" )
lowerCAmelCase__ : str = AutoModel.from_pretrained("""yjernite/retribert-base-uncased""" ).to("""cuda:0""" )
lowerCAmelCase__ : str = qar_model.eval()
else:
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = (None, None)
if MODEL_TYPE == "bart":
lowerCAmelCase__ : str = AutoTokenizer.from_pretrained("""yjernite/bart_eli5""" )
lowerCAmelCase__ : int = AutoModelForSeqaSeqLM.from_pretrained("""yjernite/bart_eli5""" ).to("""cuda:0""" )
lowerCAmelCase__ : int = torch.load("""seq2seq_models/eli5_bart_model_blm_2.pth""" )
sas_model.load_state_dict(save_dict["""model"""] )
lowerCAmelCase__ : Dict = sas_model.eval()
else:
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = make_qa_sas_model(
model_name="""t5-small""" , from_file="""seq2seq_models/eli5_t5_model_1024_4.pth""" , device="""cuda:0""" )
return (qar_tokenizer, qar_model, sas_tokenizer, sas_model)
@st.cache(allow_output_mutation=UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
if LOAD_DENSE_INDEX:
lowerCAmelCase__ : Union[str, Any] = faiss.StandardGpuResources()
lowerCAmelCase__ : List[str] = datasets.load_dataset(path="""wiki_snippets""" , name="""wiki40b_en_100_0""" )["""train"""]
lowerCAmelCase__ : Dict = np.memmap(
"""wiki40b_passages_reps_32_l-8_h-768_b-512-512.dat""" , dtype="""float32""" , mode="""r""" , shape=(wikiaab_passages.num_rows, 128) , )
lowerCAmelCase__ : int = faiss.IndexFlatIP(128 )
lowerCAmelCase__ : str = faiss.index_cpu_to_gpu(UpperCamelCase , 1 , UpperCamelCase )
wikiaab_gpu_index_flat.add(UpperCamelCase ) # TODO fix for larger GPU
else:
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = (None, None)
lowerCAmelCase__ : Optional[Any] = Elasticsearch([{"""host""": """localhost""", """port""": """9200"""}] )
return (wikiaab_passages, wikiaab_gpu_index_flat, es_client)
@st.cache(allow_output_mutation=UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = datasets.load_dataset("""eli5""" , name="""LFQA_reddit""" )
lowerCAmelCase__ : Dict = elia["""train_eli5"""]
lowerCAmelCase__ : Any = np.memmap(
"""eli5_questions_reps.dat""" , dtype="""float32""" , mode="""r""" , shape=(elia_train.num_rows, 128) )
lowerCAmelCase__ : Optional[int] = faiss.IndexFlatIP(128 )
eli5_train_q_index.add(UpperCamelCase )
return (elia_train, eli5_train_q_index)
_lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase = load_indexes()
_lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase = load_models()
_lowerCAmelCase , _lowerCAmelCase = load_train_data()
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=10 ):
"""simple docstring"""
lowerCAmelCase__ : str = embed_questions_for_retrieval([question] , UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ , lowerCAmelCase__ : int = eli5_train_q_index.search(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = [elia_train[int(UpperCamelCase )] for i in I[0]]
return nn_examples
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase="wiki40b" , UpperCamelCase="dense" , UpperCamelCase=10 ):
"""simple docstring"""
if source == "none":
lowerCAmelCase__ , lowerCAmelCase__ : Any = (""" <P> """.join(["""""" for _ in range(11 )] ).strip(), [])
else:
if method == "dense":
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = query_qa_dense_index(
UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase )
else:
lowerCAmelCase__ , lowerCAmelCase__ : int = query_es_index(
UpperCamelCase , UpperCamelCase , index_name="""english_wiki40b_snippets_100w""" , n_results=UpperCamelCase , )
lowerCAmelCase__ : int = [
(res["""article_title"""], res["""section_title"""].strip(), res["""score"""], res["""passage_text"""]) for res in hit_lst
]
lowerCAmelCase__ : Optional[Any] = """question: {} context: {}""".format(UpperCamelCase , UpperCamelCase )
return question_doc, support_list
@st.cache(
hash_funcs={
torch.Tensor: (lambda UpperCamelCase : None),
transformers.models.bart.tokenization_bart.BartTokenizer: (lambda UpperCamelCase : None),
} )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=64 , UpperCamelCase=256 , UpperCamelCase=False , UpperCamelCase=2 , UpperCamelCase=0.95 , UpperCamelCase=0.8 ):
"""simple docstring"""
with torch.no_grad():
lowerCAmelCase__ : str = qa_sas_generate(
UpperCamelCase , UpperCamelCase , UpperCamelCase , num_answers=1 , num_beams=UpperCamelCase , min_len=UpperCamelCase , max_len=UpperCamelCase , do_sample=UpperCamelCase , temp=UpperCamelCase , top_p=UpperCamelCase , top_k=UpperCamelCase , max_input_length=1024 , device="""cuda:0""" , )[0]
return (answer, support_list)
st.title('''Long Form Question Answering with ELI5''')
# Start sidebar
_lowerCAmelCase = '''<img src=\'https://huggingface.co/front/assets/huggingface_logo.svg\'>'''
_lowerCAmelCase = '''
<html>
<head>
<style>
.img-container {
padding-left: 90px;
padding-right: 90px;
padding-top: 50px;
padding-bottom: 50px;
background-color: #f0f3f9;
}
</style>
</head>
<body>
<span class="img-container"> <!-- Inline parent element -->
%s
</span>
</body>
</html>
''' % (
header_html,
)
st.sidebar.markdown(
header_full,
unsafe_allow_html=True,
)
# Long Form QA with ELI5 and Wikipedia
_lowerCAmelCase = '''
This demo presents a model trained to [provide long-form answers to open-domain questions](https://yjernite.github.io/lfqa.html).
First, a document retriever fetches a set of relevant Wikipedia passages given the question from the [Wiki40b](https://research.google/pubs/pub49029/) dataset,
a pre-processed fixed snapshot of Wikipedia.
'''
st.sidebar.markdown(description, unsafe_allow_html=True)
_lowerCAmelCase = [
'''Answer the question''',
'''View the retrieved document only''',
'''View the most similar ELI5 question and answer''',
'''Show me everything, please!''',
]
_lowerCAmelCase = st.sidebar.checkbox('''Demo options''')
if demo_options:
_lowerCAmelCase = st.sidebar.selectbox(
'''''',
action_list,
index=3,
)
_lowerCAmelCase = action_list.index(action_st)
_lowerCAmelCase = st.sidebar.selectbox(
'''''',
['''Show full text of passages''', '''Show passage section titles'''],
index=0,
)
_lowerCAmelCase = show_type == '''Show full text of passages'''
else:
_lowerCAmelCase = 3
_lowerCAmelCase = True
_lowerCAmelCase = st.sidebar.checkbox('''Retrieval options''')
if retrieval_options:
_lowerCAmelCase = '''
### Information retriever options
The **sparse** retriever uses ElasticSearch, while the **dense** retriever uses max-inner-product search between a question and passage embedding
trained using the [ELI5](https://arxiv.org/abs/1907.09190) questions-answer pairs.
The answer is then generated by sequence to sequence model which takes the question and retrieved document as input.
'''
st.sidebar.markdown(retriever_info)
_lowerCAmelCase = st.sidebar.selectbox('''Which Wikipedia format should the model use?''', ['''wiki40b''', '''none'''])
_lowerCAmelCase = st.sidebar.selectbox('''Which Wikipedia indexer should the model use?''', ['''dense''', '''sparse''', '''mixed'''])
else:
_lowerCAmelCase = '''wiki40b'''
_lowerCAmelCase = '''dense'''
_lowerCAmelCase = '''beam'''
_lowerCAmelCase = 2
_lowerCAmelCase = 64
_lowerCAmelCase = 256
_lowerCAmelCase = None
_lowerCAmelCase = None
_lowerCAmelCase = st.sidebar.checkbox('''Generation options''')
if generate_options:
_lowerCAmelCase = '''
### Answer generation options
The sequence-to-sequence model was initialized with [BART](https://huggingface.co/facebook/bart-large)
weights and fine-tuned on the ELI5 QA pairs and retrieved documents. You can use the model for greedy decoding with
**beam** search, or **sample** from the decoder\'s output probabilities.
'''
st.sidebar.markdown(generate_info)
_lowerCAmelCase = st.sidebar.selectbox('''Would you like to use beam search or sample an answer?''', ['''beam''', '''sampled'''])
_lowerCAmelCase = st.sidebar.slider(
'''Minimum generation length''', min_value=8, max_value=256, value=64, step=8, format=None, key=None
)
_lowerCAmelCase = st.sidebar.slider(
'''Maximum generation length''', min_value=64, max_value=512, value=256, step=16, format=None, key=None
)
if sampled == "beam":
_lowerCAmelCase = st.sidebar.slider('''Beam size''', min_value=1, max_value=8, value=2, step=None, format=None, key=None)
else:
_lowerCAmelCase = st.sidebar.slider(
'''Nucleus sampling p''', min_value=0.1, max_value=1.0, value=0.95, step=0.01, format=None, key=None
)
_lowerCAmelCase = st.sidebar.slider(
'''Temperature''', min_value=0.1, max_value=1.0, value=0.7, step=0.01, format=None, key=None
)
_lowerCAmelCase = None
# start main text
_lowerCAmelCase = [
'''<MY QUESTION>''',
'''How do people make chocolate?''',
'''Why do we get a fever when we are sick?''',
'''How can different animals perceive different colors?''',
'''What is natural language processing?''',
'''What\'s the best way to treat a sunburn?''',
'''What exactly are vitamins ?''',
'''How does nuclear energy provide electricity?''',
'''What\'s the difference between viruses and bacteria?''',
'''Why are flutes classified as woodwinds when most of them are made out of metal ?''',
'''Why do people like drinking coffee even though it tastes so bad?''',
'''What happens when wine ages? How does it make the wine taste better?''',
'''If an animal is an herbivore, where does it get the protein that it needs to survive if it only eats grass?''',
'''How can we set a date to the beginning or end of an artistic period? Doesn\'t the change happen gradually?''',
'''How does New Zealand have so many large bird predators?''',
]
_lowerCAmelCase = st.selectbox(
'''What would you like to ask? ---- select <MY QUESTION> to enter a new query''',
questions_list,
index=1,
)
if question_s == "<MY QUESTION>":
_lowerCAmelCase = st.text_input('''Enter your question here:''', '''''')
else:
_lowerCAmelCase = question_s
if st.button('''Show me!'''):
if action in [0, 1, 3]:
if index_type == "mixed":
_lowerCAmelCase , _lowerCAmelCase = make_support(question, source=wiki_source, method='''dense''', n_results=10)
_lowerCAmelCase , _lowerCAmelCase = make_support(question, source=wiki_source, method='''sparse''', n_results=10)
_lowerCAmelCase = []
for res_d, res_s in zip(support_list_dense, support_list_sparse):
if tuple(res_d) not in support_list:
support_list += [tuple(res_d)]
if tuple(res_s) not in support_list:
support_list += [tuple(res_s)]
_lowerCAmelCase = support_list[:10]
_lowerCAmelCase = '''<P> ''' + ''' <P> '''.join([res[-1] for res in support_list])
else:
_lowerCAmelCase , _lowerCAmelCase = make_support(question, source=wiki_source, method=index_type, n_results=10)
if action in [0, 3]:
_lowerCAmelCase , _lowerCAmelCase = answer_question(
question_doc,
sas_model,
sas_tokenizer,
min_len=min_len,
max_len=int(max_len),
sampling=(sampled == '''sampled'''),
n_beams=n_beams,
top_p=top_p,
temp=temp,
)
st.markdown('''### The model generated answer is:''')
st.write(answer)
if action in [0, 1, 3] and wiki_source != "none":
st.markdown('''--- \n ### The model is drawing information from the following Wikipedia passages:''')
for i, res in enumerate(support_list):
_lowerCAmelCase = '''https://en.wikipedia.org/wiki/{}'''.format(res[0].replace(''' ''', '''_'''))
_lowerCAmelCase = res[1].strip()
if sec_titles == "":
_lowerCAmelCase = '''[{}]({})'''.format(res[0], wiki_url)
else:
_lowerCAmelCase = sec_titles.split(''' & ''')
_lowerCAmelCase = ''' & '''.join(
['''[{}]({}#{})'''.format(sec.strip(), wiki_url, sec.strip().replace(''' ''', '''_''')) for sec in sec_list]
)
st.markdown(
'''{0:02d} - **Article**: {1:<18} <br> _Section_: {2}'''.format(i + 1, res[0], sections),
unsafe_allow_html=True,
)
if show_passages:
st.write(
'''> <span style="font-family:arial; font-size:10pt;">''' + res[-1] + '''</span>''', unsafe_allow_html=True
)
if action in [2, 3]:
_lowerCAmelCase = find_nearest_training(question)
_lowerCAmelCase = nn_train_list[0]
st.markdown(
'''--- \n ### The most similar question in the ELI5 training set was: \n\n {}'''.format(train_exple['''title'''])
)
_lowerCAmelCase = [
'''{}. {}'''.format(i + 1, ''' \n'''.join([line.strip() for line in ans.split('''\n''') if line.strip() != '''''']))
for i, (ans, sc) in enumerate(zip(train_exple['''answers''']['''text'''], train_exple['''answers''']['''score''']))
if i == 0 or sc > 2
]
st.markdown('''##### Its answers were: \n\n {}'''.format('''\n'''.join(answers_st)))
_lowerCAmelCase = '''
---
**Disclaimer**
*The intent of this app is to provide some (hopefully entertaining) insights into the behavior of a current LFQA system.
Evaluating biases of such a model and ensuring factual generations are still very much open research problems.
Therefore, until some significant progress is achieved, we caution against using the generated answers for practical purposes.*
'''
st.sidebar.markdown(disclaimer, unsafe_allow_html=True)
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : str = set()
# Replace all the whitespace in our sentence
lowerCAmelCase__ : Tuple = input_str.replace(""" """ , """""" )
for alpha in input_str:
if "a" <= alpha.lower() <= "z":
frequency.add(alpha.lower() )
return len(UpperCamelCase ) == 26
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
lowerCAmelCase__ : Any = [False] * 26
for char in input_str:
if char.islower():
lowerCAmelCase__ : Optional[Any] = True
elif char.isupper():
lowerCAmelCase__ : Any = True
return all(UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase = "The quick brown fox jumps over the lazy dog" , ):
"""simple docstring"""
return len({char for char in input_str.lower() if char.isalpha()} ) == 26
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from timeit import timeit
lowerCAmelCase__ : Union[str, Any] = """from __main__ import is_pangram, is_pangram_faster, is_pangram_fastest"""
print(timeit("""is_pangram()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_faster()""" , setup=UpperCamelCase ) )
print(timeit("""is_pangram_fastest()""" , setup=UpperCamelCase ) )
# 5.348480500048026, 2.6477354579837993, 1.8470395830227062
# 5.036091582966037, 2.644472333951853, 1.8869528750656173
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
from math import pow
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , ):
"""simple docstring"""
if current_sum == needed_sum:
# If the sum of the powers is equal to needed_sum, then we have a solution.
solutions_count += 1
return current_sum, solutions_count
lowerCAmelCase__ : Optional[int] = int(pow(UpperCamelCase , UpperCamelCase ) )
if current_sum + i_to_n <= needed_sum:
# If the sum of the powers is less than needed_sum, then continue adding powers.
current_sum += i_to_n
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = backtrack(
UpperCamelCase , UpperCamelCase , current_number + 1 , UpperCamelCase , UpperCamelCase )
current_sum -= i_to_n
if i_to_n < needed_sum:
# If the power of i is less than needed_sum, then try with the next power.
lowerCAmelCase__ , lowerCAmelCase__ : Tuple = backtrack(
UpperCamelCase , UpperCamelCase , current_number + 1 , UpperCamelCase , UpperCamelCase )
return current_sum, solutions_count
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
if not (1 <= needed_sum <= 1000 and 2 <= power <= 10):
raise ValueError(
"""Invalid input\n"""
"""needed_sum must be between 1 and 1000, power between 2 and 10.""" )
return backtrack(UpperCamelCase , UpperCamelCase , 1 , 0 , 0 )[1] # Return the solutions_count
if __name__ == "__main__":
import doctest
doctest.testmod()
| 37
|
'''simple docstring'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = abs(UpperCamelCase )
lowerCAmelCase__ : List[Any] = 0
while n > 0:
res += n % 10
n //= 10
return res
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = abs(UpperCamelCase )
return n if n < 10 else n % 10 + sum_of_digits(n // 10 )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return sum(int(UpperCamelCase ) for c in str(abs(UpperCamelCase ) ) )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(UpperCamelCase , UpperCamelCase ) -> None:
lowerCAmelCase__ : str = f"""{func.__name__}({value})"""
lowerCAmelCase__ : str = timeit(f"""__main__.{call}""" , setup="""import __main__""" )
print(f"""{call:56} = {func(UpperCamelCase )} -- {timing:.4f} seconds""" )
for value in (262144, 1125899906842624, 1267650600228229401496703205376):
for func in (sum_of_digits, sum_of_digits_recursion, sum_of_digits_compact):
benchmark_a_function(UpperCamelCase , UpperCamelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 37
| 1
|
'''simple docstring'''
import fire
from utils import calculate_rouge, save_json
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=None , **UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Tuple = [x.strip() for x in open(UpperCamelCase ).readlines()]
lowerCAmelCase__ : List[Any] = [x.strip() for x in open(UpperCamelCase ).readlines()][: len(UpperCamelCase )]
lowerCAmelCase__ : Tuple = calculate_rouge(UpperCamelCase , UpperCamelCase , **UpperCamelCase )
if save_path is not None:
save_json(UpperCamelCase , UpperCamelCase , indent=UpperCamelCase )
return metrics # these print nicely
if __name__ == "__main__":
fire.Fire(calculate_rouge_path)
| 37
|
'''simple docstring'''
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team.
#
# 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.
# this script dumps information about the environment
import os
import sys
import transformers
_lowerCAmelCase = '''3'''
print('''Python version:''', sys.version)
print('''transformers version:''', transformers.__version__)
try:
import torch
print('''Torch version:''', torch.__version__)
print('''Cuda available:''', torch.cuda.is_available())
print('''Cuda version:''', torch.version.cuda)
print('''CuDNN version:''', torch.backends.cudnn.version())
print('''Number of GPUs available:''', torch.cuda.device_count())
print('''NCCL version:''', torch.cuda.nccl.version())
except ImportError:
print('''Torch version:''', None)
try:
import deepspeed
print('''DeepSpeed version:''', deepspeed.__version__)
except ImportError:
print('''DeepSpeed version:''', None)
try:
import tensorflow as tf
print('''TensorFlow version:''', tf.__version__)
print('''TF GPUs available:''', bool(tf.config.list_physical_devices('''GPU''')))
print('''Number of TF GPUs available:''', len(tf.config.list_physical_devices('''GPU''')))
except ImportError:
print('''TensorFlow version:''', None)
| 37
| 1
|
'''simple docstring'''
from __future__ import annotations
import unittest
import numpy as np
from transformers import OPTConfig, is_tf_available
from transformers.testing_utils import require_sentencepiece, require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import GPTaTokenizer, TFOPTForCausalLM, TFOPTModel
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None ):
"""simple docstring"""
if attention_mask is None:
lowerCAmelCase__ : Any = tf.cast(tf.math.not_equal(UpperCamelCase , config.pad_token_id ) , tf.inta )
return {"input_ids": input_ids, "attention_mask": attention_mask}
@require_tf
class lowerCAmelCase_:
'''simple docstring'''
__lowercase : str = OPTConfig
__lowercase : List[Any] = {}
__lowercase : List[str] = '''gelu'''
def __init__( self ,__UpperCAmelCase ,__UpperCAmelCase=13 ,__UpperCAmelCase=7 ,__UpperCAmelCase=True ,__UpperCAmelCase=False ,__UpperCAmelCase=99 ,__UpperCAmelCase=16 ,__UpperCAmelCase=2 ,__UpperCAmelCase=4 ,__UpperCAmelCase=4 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=20 ,__UpperCAmelCase=2 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=16 ,__UpperCAmelCase=16 ,) -> List[str]:
lowerCAmelCase__ : str = parent
lowerCAmelCase__ : Dict = batch_size
lowerCAmelCase__ : str = seq_length
lowerCAmelCase__ : int = is_training
lowerCAmelCase__ : Optional[Any] = use_labels
lowerCAmelCase__ : Union[str, Any] = vocab_size
lowerCAmelCase__ : Optional[Any] = hidden_size
lowerCAmelCase__ : Dict = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : Any = intermediate_size
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : Optional[Any] = attention_probs_dropout_prob
lowerCAmelCase__ : Any = max_position_embeddings
lowerCAmelCase__ : Union[str, Any] = eos_token_id
lowerCAmelCase__ : Any = pad_token_id
lowerCAmelCase__ : List[str] = bos_token_id
lowerCAmelCase__ : List[str] = embed_dim
lowerCAmelCase__ : int = word_embed_proj_dim
lowerCAmelCase__ : Optional[int] = False
def UpperCAmelCase_ ( self ) -> Dict:
lowerCAmelCase__ : str = ids_tensor([self.batch_size, self.seq_length - 1] ,self.vocab_size )
lowerCAmelCase__ : int = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) ,1 )
lowerCAmelCase__ : str = tf.concat([input_ids, eos_tensor] ,axis=1 )
lowerCAmelCase__ : Dict = self.config_cls(
vocab_size=self.vocab_size ,hidden_size=self.hidden_size ,num_hidden_layers=self.num_hidden_layers ,num_attention_heads=self.num_attention_heads ,ffn_dim=self.intermediate_size ,dropout=self.hidden_dropout_prob ,attention_dropout=self.attention_probs_dropout_prob ,max_position_embeddings=self.max_position_embeddings ,eos_token_id=self.eos_token_id ,bos_token_id=self.bos_token_id ,pad_token_id=self.pad_token_id ,embed_dim=self.embed_dim ,word_embed_proj_dim=self.word_embed_proj_dim ,is_encoder_decoder=__UpperCAmelCase ,**self.config_updates ,)
lowerCAmelCase__ : List[str] = prepare_opt_inputs_dict(__UpperCAmelCase ,__UpperCAmelCase )
return config, inputs_dict
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Optional[Any]:
lowerCAmelCase__ : Any = TFOPTModel(config=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = inputs_dict["""input_ids"""]
lowerCAmelCase__ : List[str] = input_ids[:1, :]
lowerCAmelCase__ : str = inputs_dict["""attention_mask"""][:1, :]
lowerCAmelCase__ : int = 1
# first forward pass
lowerCAmelCase__ : int = model(__UpperCAmelCase ,attention_mask=__UpperCAmelCase ,use_cache=__UpperCAmelCase )
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
lowerCAmelCase__ : Tuple = ids_tensor((self.batch_size, 3) ,config.vocab_size )
lowerCAmelCase__ : str = tf.cast(ids_tensor((self.batch_size, 3) ,2 ) ,tf.inta )
# append to next input_ids and
lowerCAmelCase__ : Tuple = tf.concat([input_ids, next_tokens] ,axis=-1 )
lowerCAmelCase__ : Dict = tf.concat([attention_mask, next_attn_mask] ,axis=-1 )
lowerCAmelCase__ : Union[str, Any] = model(__UpperCAmelCase ,attention_mask=__UpperCAmelCase )[0]
lowerCAmelCase__ : Optional[int] = model(__UpperCAmelCase ,attention_mask=__UpperCAmelCase ,past_key_values=__UpperCAmelCase )[0]
self.parent.assertEqual(next_tokens.shape[1] ,output_from_past.shape[1] )
# select random slice
lowerCAmelCase__ : str = int(ids_tensor((1,) ,output_from_past.shape[-1] ) )
lowerCAmelCase__ : Optional[Any] = output_from_no_past[:, -3:, random_slice_idx]
lowerCAmelCase__ : int = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(__UpperCAmelCase ,__UpperCAmelCase ,rtol=1E-3 )
@require_tf
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : Tuple = (TFOPTModel, TFOPTForCausalLM) if is_tf_available() else ()
__lowercase : Dict = (TFOPTForCausalLM,) if is_tf_available() else ()
__lowercase : List[str] = (
{'''feature-extraction''': TFOPTModel, '''text-generation''': TFOPTForCausalLM} if is_tf_available() else {}
)
__lowercase : int = False
__lowercase : Any = False
__lowercase : Union[str, Any] = False
__lowercase : Union[str, Any] = 1_0
def UpperCAmelCase_ ( self ) -> Tuple:
lowerCAmelCase__ : Optional[Any] = TFOPTModelTester(self )
lowerCAmelCase__ : List[Any] = ConfigTester(self ,config_class=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
self.config_tester.run_common_tests()
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ , lowerCAmelCase__ : Optional[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def _get_word_embedding_weight(__UpperCAmelCase ,__UpperCAmelCase ):
if hasattr(__UpperCAmelCase ,"""weight""" ):
return embedding_layer.weight
else:
# Here we build the word embeddings weights if not exists.
# And then we retry to get the attribute once built.
model.build()
if hasattr(__UpperCAmelCase ,"""weight""" ):
return embedding_layer.weight
else:
return None
for model_class in self.all_model_classes:
for size in [config.vocab_size - 10, config.vocab_size + 10]:
# build the embeddings
lowerCAmelCase__ : Optional[int] = model_class(config=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = _get_word_embedding_weight(__UpperCAmelCase ,model.get_input_embeddings() )
lowerCAmelCase__ : Optional[int] = _get_word_embedding_weight(__UpperCAmelCase ,model.get_output_embeddings() )
# reshape the embeddings
model.resize_token_embeddings(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = _get_word_embedding_weight(__UpperCAmelCase ,model.get_input_embeddings() )
lowerCAmelCase__ : Optional[int] = _get_word_embedding_weight(__UpperCAmelCase ,model.get_output_embeddings() )
# check that the resized embeddings size matches the desired size.
lowerCAmelCase__ : str = size if size is not None else config.vocab_size
self.assertEqual(new_input_embeddings.shape[0] ,__UpperCAmelCase )
# check that weights remain the same after resizing
lowerCAmelCase__ : Dict = True
for pa, pa in zip(old_input_embeddings.value() ,new_input_embeddings.value() ):
if tf.math.reduce_sum(tf.math.abs(pa - pa ) ) > 0:
lowerCAmelCase__ : Any = False
self.assertTrue(__UpperCAmelCase )
if old_output_embeddings is not None and new_output_embeddings is not None:
self.assertEqual(new_output_embeddings.shape[0] ,__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = True
for pa, pa in zip(old_output_embeddings.value() ,new_output_embeddings.value() ):
if tf.math.reduce_sum(tf.math.abs(pa - pa ) ) > 0:
lowerCAmelCase__ : List[Any] = False
self.assertTrue(__UpperCAmelCase )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return tf.constant(UpperCamelCase , dtype=tf.intaa )
@require_tf
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
__lowercase : List[Any] = 9_9
def UpperCAmelCase_ ( self ) -> List[Any]:
lowerCAmelCase__ : Optional[int] = tf.ones((4, 1) ,dtype=tf.intaa ) * 2
lowerCAmelCase__ : List[Any] = tf.concat([ids_tensor((4, 6) ,self.vocab_size - 3 ) + 3, eos_column_vector] ,axis=1 )
lowerCAmelCase__ : Dict = input_ids.shape[0]
lowerCAmelCase__ : Tuple = OPTConfig(
vocab_size=self.vocab_size ,hidden_size=24 ,num_hidden_layers=2 ,num_attention_heads=2 ,ffn_dim=32 ,max_position_embeddings=48 ,eos_token_id=2 ,pad_token_id=1 ,bos_token_id=0 ,)
return config, input_ids, batch_size
@require_sentencepiece
@require_tf
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@slow
def UpperCAmelCase_ ( self ) -> Any:
lowerCAmelCase__ : Optional[Any] = TFOPTModel.from_pretrained("""facebook/opt-350m""" )
lowerCAmelCase__ : Tuple = _long_tensor([[0, 3_1414, 232, 328, 740, 1140, 1_2695, 69, 4_6078, 1588, 2]] )
lowerCAmelCase__ : Optional[Any] = tf.not_equal(__UpperCAmelCase ,model.config.pad_token_id )
with tf.GradientTape():
lowerCAmelCase__ : str = model(input_ids=__UpperCAmelCase ,attention_mask=__UpperCAmelCase ).last_hidden_state
lowerCAmelCase__ : Union[str, Any] = (1, 11, 512)
self.assertEqual(output.shape ,__UpperCAmelCase )
lowerCAmelCase__ : Tuple = tf.constant(
[[-0.2_8_7_3, -1.9_2_1_8, -0.3_0_3_3], [-1.2_7_1_0, -0.1_3_3_8, -0.1_9_0_2], [0.4_0_9_5, 0.1_2_1_4, -1.3_1_2_1]] )
self.assertTrue(np.allclose(output[:, :3, :3] ,__UpperCAmelCase ,atol=4E-3 ) )
lowerCAmelCase__ : str = tf.function(__UpperCAmelCase ,jit_compile=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = xla_generate(__UpperCAmelCase ,__UpperCAmelCase )[0]
self.assertTrue(np.allclose(output[:, :3, :3] ,__UpperCAmelCase ,atol=4E-2 ) )
@require_tf
@slow
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Tuple:
super().setUp()
lowerCAmelCase__ : Union[str, Any] = """facebook/opt-350m"""
def UpperCAmelCase_ ( self ) -> int:
lowerCAmelCase__ : str = TFOPTForCausalLM.from_pretrained(self.path_model )
lowerCAmelCase__ : Tuple = GPTaTokenizer.from_pretrained(self.path_model )
lowerCAmelCase__ : List[Any] = [
"""Today is a beautiful day and I want to""",
"""In the city of""",
"""Paris is the capital of France and""",
"""Computers and mobile phones have taken""",
]
# verify that prompt without BOS token is identical to Metaseq -> add_special_tokens=False
lowerCAmelCase__ : List[str] = tokenizer(__UpperCAmelCase ,return_tensors="""tf""" ,padding=__UpperCAmelCase ,add_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : int = tf.math.reduce_mean(model(inputs.input_ids ,attention_mask=inputs.attention_mask )[0] ,axis=-1 )
lowerCAmelCase__ : Tuple = tf.constant(
[
[1.3_8_5_1, -1_3.8_9_2_3, -1_0.5_2_2_9, -1_0.7_5_3_3, -0.2_3_0_9, -1_0.2_3_8_4, -0.5_3_6_5, -9.0_9_4_7, -5.1_6_7_0],
[-4.7_0_7_3, -1_0.6_2_7_6, -3.9_4_1_5, -2_1.5_2_4_2, -0.2_8_2_2, -0.2_8_2_2, -0.2_8_2_2, -0.2_8_2_2, -0.2_8_2_2],
[0.6_2_4_7, -3.4_2_2_9, -8.9_1_7_9, -1.4_2_9_7, -1_4.1_6_5_0, 1.4_1_4_6, -9.0_2_1_8, -0.2_7_0_3, -0.2_7_0_3],
[6.4_7_8_3, -1.9_9_1_3, -1_0.7_9_2_6, -2.3_3_3_6, 1.5_0_9_2, -0.9_9_7_4, -6.8_2_1_3, 1.3_4_7_7, 1.3_4_7_7],
] )
self.assertTrue(np.allclose(__UpperCAmelCase ,__UpperCAmelCase ,atol=1E-4 ) )
lowerCAmelCase__ : Union[str, Any] = tf.function(__UpperCAmelCase ,jit_compile=__UpperCAmelCase )
lowerCAmelCase__ : Dict = tf.math.reduce_mean(xla_generate(inputs.input_ids ,attention_mask=inputs.attention_mask )[0] ,axis=-1 )
self.assertTrue(np.allclose(__UpperCAmelCase ,__UpperCAmelCase ,atol=1E-4 ) )
@require_tf
@slow
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> str:
return [
"Today is a beautiful day and I want",
"In the city of",
"Paris is the capital of France and",
"Computers and mobile phones have taken",
]
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Optional[Any] = """facebook/opt-125m"""
lowerCAmelCase__ : int = [
"""Today is a beautiful day and I want to""",
"""In the city of New York, the city""",
"""Paris is the capital of France and the capital""",
"""Computers and mobile phones have taken over the""",
]
lowerCAmelCase__ : int = []
lowerCAmelCase__ : str = GPTaTokenizer.from_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : Optional[int] = TFOPTForCausalLM.from_pretrained(__UpperCAmelCase )
for prompt in self.prompts:
lowerCAmelCase__ : Optional[int] = tokenizer(__UpperCAmelCase ,return_tensors="""tf""" ).input_ids
lowerCAmelCase__ : int = model.generate(__UpperCAmelCase ,max_length=10 )
lowerCAmelCase__ : int = tokenizer.batch_decode(__UpperCAmelCase ,skip_special_tokens=__UpperCAmelCase )
predicted_outputs += generated_string
self.assertListEqual(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> str:
lowerCAmelCase__ : str = """facebook/opt-350m"""
lowerCAmelCase__ : List[str] = GPTaTokenizer.from_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : List[str] = TFOPTForCausalLM.from_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : Any = """left"""
# use different length sentences to test batching
lowerCAmelCase__ : Union[str, Any] = [
"""Hello, my dog is a little""",
"""Today, I""",
]
lowerCAmelCase__ : List[str] = tokenizer(__UpperCAmelCase ,return_tensors="""tf""" ,padding=__UpperCAmelCase )
lowerCAmelCase__ : Tuple = inputs["""input_ids"""]
lowerCAmelCase__ : Any = model.generate(input_ids=__UpperCAmelCase ,attention_mask=inputs["""attention_mask"""] )
lowerCAmelCase__ : List[str] = tokenizer(sentences[0] ,return_tensors="""tf""" ).input_ids
lowerCAmelCase__ : int = model.generate(input_ids=__UpperCAmelCase )
lowerCAmelCase__ : str = inputs_non_padded.shape[-1] - tf.math.reduce_sum(
tf.cast(inputs["""attention_mask"""][-1] ,tf.intaa ) )
lowerCAmelCase__ : List[str] = tokenizer(sentences[1] ,return_tensors="""tf""" ).input_ids
lowerCAmelCase__ : Dict = model.generate(input_ids=__UpperCAmelCase ,max_length=model.config.max_length - num_paddings )
lowerCAmelCase__ : List[Any] = tokenizer.batch_decode(__UpperCAmelCase ,skip_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : int = tokenizer.decode(output_non_padded[0] ,skip_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : Union[str, Any] = tokenizer.decode(output_padded[0] ,skip_special_tokens=__UpperCAmelCase )
lowerCAmelCase__ : List[str] = [
"""Hello, my dog is a little bit of a dork.\nI'm a little bit""",
"""Today, I was in the middle of a conversation with a friend about the""",
]
self.assertListEqual(__UpperCAmelCase ,__UpperCAmelCase )
self.assertListEqual(__UpperCAmelCase ,[non_padded_sentence, padded_sentence] )
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : Union[str, Any] = """facebook/opt-350m"""
lowerCAmelCase__ : Dict = [
"""Today is a beautiful day and I want to""",
"""In the city of San Francisco, the city""",
"""Paris is the capital of France and the capital""",
"""Computers and mobile phones have taken over the""",
]
lowerCAmelCase__ : Dict = []
lowerCAmelCase__ : str = GPTaTokenizer.from_pretrained(__UpperCAmelCase )
lowerCAmelCase__ : str = TFOPTForCausalLM.from_pretrained(__UpperCAmelCase )
for prompt in self.prompts:
lowerCAmelCase__ : int = tokenizer(__UpperCAmelCase ,return_tensors="""tf""" ).input_ids
lowerCAmelCase__ : str = model.generate(__UpperCAmelCase ,max_length=10 )
lowerCAmelCase__ : Optional[Any] = tokenizer.batch_decode(__UpperCAmelCase ,skip_special_tokens=__UpperCAmelCase )
predicted_outputs += generated_string
self.assertListEqual(__UpperCAmelCase ,__UpperCAmelCase )
| 37
|
'''simple docstring'''
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
_lowerCAmelCase = logging.get_logger(__name__)
_lowerCAmelCase = {
'''facebook/xlm-roberta-xl''': '''https://huggingface.co/facebook/xlm-roberta-xl/resolve/main/config.json''',
'''facebook/xlm-roberta-xxl''': '''https://huggingface.co/facebook/xlm-roberta-xxl/resolve/main/config.json''',
# See all XLM-RoBERTa-XL models at https://huggingface.co/models?filter=xlm-roberta-xl
}
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
__lowercase : int = '''xlm-roberta-xl'''
def __init__( self ,__UpperCAmelCase=25_0880 ,__UpperCAmelCase=2560 ,__UpperCAmelCase=36 ,__UpperCAmelCase=32 ,__UpperCAmelCase=1_0240 ,__UpperCAmelCase="gelu" ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=0.1 ,__UpperCAmelCase=514 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0.0_2 ,__UpperCAmelCase=1E-05 ,__UpperCAmelCase=1 ,__UpperCAmelCase=0 ,__UpperCAmelCase=2 ,__UpperCAmelCase="absolute" ,__UpperCAmelCase=True ,__UpperCAmelCase=None ,**__UpperCAmelCase ,) -> str:
super().__init__(pad_token_id=__UpperCAmelCase ,bos_token_id=__UpperCAmelCase ,eos_token_id=__UpperCAmelCase ,**__UpperCAmelCase )
lowerCAmelCase__ : List[Any] = vocab_size
lowerCAmelCase__ : int = hidden_size
lowerCAmelCase__ : int = num_hidden_layers
lowerCAmelCase__ : str = num_attention_heads
lowerCAmelCase__ : int = hidden_act
lowerCAmelCase__ : Dict = intermediate_size
lowerCAmelCase__ : List[Any] = hidden_dropout_prob
lowerCAmelCase__ : str = attention_probs_dropout_prob
lowerCAmelCase__ : Optional[int] = max_position_embeddings
lowerCAmelCase__ : List[str] = type_vocab_size
lowerCAmelCase__ : List[Any] = initializer_range
lowerCAmelCase__ : Tuple = layer_norm_eps
lowerCAmelCase__ : int = position_embedding_type
lowerCAmelCase__ : Optional[Any] = use_cache
lowerCAmelCase__ : Optional[Any] = classifier_dropout
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ ):
'''simple docstring'''
@property
def UpperCAmelCase_ ( self ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
lowerCAmelCase__ : Dict = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
lowerCAmelCase__ : Any = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 37
| 1
|
'''simple docstring'''
import gc
import unittest
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTextModelWithProjection, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
DDPMScheduler,
PriorTransformer,
StableUnCLIPPipeline,
UNetaDConditionModel,
)
from diffusers.pipelines.stable_diffusion.stable_unclip_image_normalizer import StableUnCLIPImageNormalizer
from diffusers.utils.testing_utils import enable_full_determinism, load_numpy, require_torch_gpu, slow, torch_device
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
from ..test_pipelines_common import (
PipelineKarrasSchedulerTesterMixin,
PipelineLatentTesterMixin,
PipelineTesterMixin,
assert_mean_pixel_difference,
)
enable_full_determinism()
class lowerCAmelCase_( SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ , unittest.TestCase ):
'''simple docstring'''
__lowercase : int = StableUnCLIPPipeline
__lowercase : str = TEXT_TO_IMAGE_PARAMS
__lowercase : List[Any] = TEXT_TO_IMAGE_BATCH_PARAMS
__lowercase : Optional[Any] = TEXT_TO_IMAGE_IMAGE_PARAMS
__lowercase : Optional[Any] = TEXT_TO_IMAGE_IMAGE_PARAMS
# TODO(will) Expected attn_bias.stride(1) == 0 to be true, but got false
__lowercase : str = False
def UpperCAmelCase_ ( self ) -> List[str]:
lowerCAmelCase__ : Optional[Any] = 32
lowerCAmelCase__ : Optional[int] = embedder_hidden_size
# prior components
torch.manual_seed(0 )
lowerCAmelCase__ : List[str] = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
torch.manual_seed(0 )
lowerCAmelCase__ : Any = CLIPTextModelWithProjection(
CLIPTextConfig(
bos_token_id=0 ,eos_token_id=2 ,hidden_size=__UpperCAmelCase ,projection_dim=__UpperCAmelCase ,intermediate_size=37 ,layer_norm_eps=1E-05 ,num_attention_heads=4 ,num_hidden_layers=5 ,pad_token_id=1 ,vocab_size=1000 ,) )
torch.manual_seed(0 )
lowerCAmelCase__ : List[str] = PriorTransformer(
num_attention_heads=2 ,attention_head_dim=12 ,embedding_dim=__UpperCAmelCase ,num_layers=1 ,)
torch.manual_seed(0 )
lowerCAmelCase__ : Tuple = DDPMScheduler(
variance_type="""fixed_small_log""" ,prediction_type="""sample""" ,num_train_timesteps=1000 ,clip_sample=__UpperCAmelCase ,clip_sample_range=5.0 ,beta_schedule="""squaredcos_cap_v2""" ,)
# regular denoising components
torch.manual_seed(0 )
lowerCAmelCase__ : Dict = StableUnCLIPImageNormalizer(embedding_dim=__UpperCAmelCase )
lowerCAmelCase__ : Any = DDPMScheduler(beta_schedule="""squaredcos_cap_v2""" )
torch.manual_seed(0 )
lowerCAmelCase__ : Dict = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
torch.manual_seed(0 )
lowerCAmelCase__ : int = CLIPTextModel(
CLIPTextConfig(
bos_token_id=0 ,eos_token_id=2 ,hidden_size=__UpperCAmelCase ,projection_dim=32 ,intermediate_size=37 ,layer_norm_eps=1E-05 ,num_attention_heads=4 ,num_hidden_layers=5 ,pad_token_id=1 ,vocab_size=1000 ,) )
torch.manual_seed(0 )
lowerCAmelCase__ : str = UNetaDConditionModel(
sample_size=32 ,in_channels=4 ,out_channels=4 ,down_block_types=("""CrossAttnDownBlock2D""", """DownBlock2D""") ,up_block_types=("""UpBlock2D""", """CrossAttnUpBlock2D""") ,block_out_channels=(32, 64) ,attention_head_dim=(2, 4) ,class_embed_type="""projection""" ,projection_class_embeddings_input_dim=embedder_projection_dim * 2 ,cross_attention_dim=__UpperCAmelCase ,layers_per_block=1 ,upcast_attention=__UpperCAmelCase ,use_linear_projection=__UpperCAmelCase ,)
torch.manual_seed(0 )
lowerCAmelCase__ : Any = DDIMScheduler(
beta_schedule="""scaled_linear""" ,beta_start=0.0_0_0_8_5 ,beta_end=0.0_1_2 ,prediction_type="""v_prediction""" ,set_alpha_to_one=__UpperCAmelCase ,steps_offset=1 ,)
torch.manual_seed(0 )
lowerCAmelCase__ : List[str] = AutoencoderKL()
lowerCAmelCase__ : List[str] = {
# prior components
"""prior_tokenizer""": prior_tokenizer,
"""prior_text_encoder""": prior_text_encoder,
"""prior""": prior,
"""prior_scheduler""": prior_scheduler,
# image noising components
"""image_normalizer""": image_normalizer,
"""image_noising_scheduler""": image_noising_scheduler,
# regular denoising components
"""tokenizer""": tokenizer,
"""text_encoder""": text_encoder,
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
}
return components
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase=0 ) -> Any:
if str(__UpperCAmelCase ).startswith("""mps""" ):
lowerCAmelCase__ : int = torch.manual_seed(__UpperCAmelCase )
else:
lowerCAmelCase__ : Any = torch.Generator(device=__UpperCAmelCase ).manual_seed(__UpperCAmelCase )
lowerCAmelCase__ : Any = {
"""prompt""": """A painting of a squirrel eating a burger""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""prior_num_inference_steps""": 2,
"""output_type""": """numpy""",
}
return inputs
def UpperCAmelCase_ ( self ) -> Optional[int]:
lowerCAmelCase__ : Tuple = torch_device == """cpu"""
self._test_attention_slicing_forward_pass(test_max_difference=__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Union[str, Any]:
lowerCAmelCase__ : List[str] = torch_device in ["""cpu""", """mps"""]
self._test_inference_batch_single_identical(test_max_difference=__UpperCAmelCase )
@slow
@require_torch_gpu
class lowerCAmelCase_( unittest.TestCase ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def UpperCAmelCase_ ( self ) -> Optional[Any]:
lowerCAmelCase__ : Union[str, Any] = load_numpy(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/stable_unclip_2_1_l_anime_turtle_fp16.npy""" )
lowerCAmelCase__ : Optional[Any] = StableUnCLIPPipeline.from_pretrained("""fusing/stable-unclip-2-1-l""" ,torch_dtype=torch.floataa )
pipe.to(__UpperCAmelCase )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
# stable unclip will oom when integration tests are run on a V100,
# so turn on memory savings
pipe.enable_attention_slicing()
pipe.enable_sequential_cpu_offload()
lowerCAmelCase__ : Union[str, Any] = torch.Generator(device="""cpu""" ).manual_seed(0 )
lowerCAmelCase__ : Tuple = pipe("""anime turle""" ,generator=__UpperCAmelCase ,output_type="""np""" )
lowerCAmelCase__ : str = output.images[0]
assert image.shape == (768, 768, 3)
assert_mean_pixel_difference(__UpperCAmelCase ,__UpperCAmelCase )
def UpperCAmelCase_ ( self ) -> Any:
torch.cuda.empty_cache()
torch.cuda.reset_max_memory_allocated()
torch.cuda.reset_peak_memory_stats()
lowerCAmelCase__ : int = StableUnCLIPPipeline.from_pretrained("""fusing/stable-unclip-2-1-l""" ,torch_dtype=torch.floataa )
lowerCAmelCase__ : List[str] = pipe.to(__UpperCAmelCase )
pipe.set_progress_bar_config(disable=__UpperCAmelCase )
pipe.enable_attention_slicing()
pipe.enable_sequential_cpu_offload()
lowerCAmelCase__ : Tuple = pipe(
"""anime turtle""" ,prior_num_inference_steps=2 ,num_inference_steps=2 ,output_type="""np""" ,)
lowerCAmelCase__ : Optional[Any] = torch.cuda.max_memory_allocated()
# make sure that less than 7 GB is allocated
assert mem_bytes < 7 * 10**9
| 37
|
'''simple docstring'''
from __future__ import annotations
import math
from collections import Counter
from string import ascii_lowercase
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : List[str] = analyze_text(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = list(""" """ + ascii_lowercase )
# what is our total sum of probabilities.
lowerCAmelCase__ : List[Any] = sum(single_char_strings.values() )
# one length string
lowerCAmelCase__ : Optional[int] = 0
# for each alpha we go in our dict and if it is in it we calculate entropy
for ch in my_alphas:
if ch in single_char_strings:
lowerCAmelCase__ : List[Any] = single_char_strings[ch]
lowerCAmelCase__ : List[Any] = my_str / all_sum
my_fir_sum += prob * math.loga(UpperCamelCase ) # entropy formula.
# print entropy
print(f"""{round(-1 * my_fir_sum ):.1f}""" )
# two len string
lowerCAmelCase__ : Dict = sum(two_char_strings.values() )
lowerCAmelCase__ : int = 0
# for each alpha (two in size) calculate entropy.
for cha in my_alphas:
for cha in my_alphas:
lowerCAmelCase__ : Union[str, Any] = cha + cha
if sequence in two_char_strings:
lowerCAmelCase__ : Dict = two_char_strings[sequence]
lowerCAmelCase__ : Tuple = int(UpperCamelCase ) / all_sum
my_sec_sum += prob * math.loga(UpperCamelCase )
# print second entropy
print(f"""{round(-1 * my_sec_sum ):.1f}""" )
# print the difference between them
print(f"""{round((-1 * my_sec_sum) - (-1 * my_fir_sum) ):.1f}""" )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = Counter() # type: ignore
lowerCAmelCase__ : Tuple = Counter() # type: ignore
single_char_strings[text[-1]] += 1
# first case when we have space at start.
two_char_strings[" " + text[0]] += 1
for i in range(0 , len(UpperCamelCase ) - 1 ):
single_char_strings[text[i]] += 1
two_char_strings[text[i : i + 2]] += 1
return single_char_strings, two_char_strings
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
import doctest
doctest.testmod()
# text = (
# "Had repulsive dashwoods suspicion sincerity but advantage now him. Remark "
# "easily garret nor nay. Civil those mrs enjoy shy fat merry. You greatest "
# "jointure saw horrible. He private he on be imagine suppose. Fertile "
# "beloved evident through no service elderly is. Blind there if every no so "
# "at. Own neglected you preferred way sincerity delivered his attempted. To "
# "of message cottage windows do besides against uncivil. Delightful "
# "unreserved impossible few estimating men favourable see entreaties. She "
# "propriety immediate was improving. He or entrance humoured likewise "
# "moderate. Much nor game son say feel. Fat make met can must form into "
# "gate. Me we offending prevailed discovery. "
# )
# calculate_prob(text)
if __name__ == "__main__":
main()
| 37
| 1
|
'''simple docstring'''
import argparse
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
########################################################################
# This is a fully working simple example to use Accelerate,
# specifically showcasing how to properly calculate the metrics on the
# validation dataset when in a distributed system, and builds off the
# `nlp_example.py` script.
#
# This example trains a Bert base model on GLUE MRPC
# in any of the following settings (with the same script):
# - single CPU or single GPU
# - multi GPUS (using PyTorch distributed mode)
# - (multi) TPUs
# - fp16 (mixed-precision) or fp32 (normal precision)
#
# To help focus on the differences in the code, building `DataLoaders`
# was refactored into its own function.
# New additions from the base script can be found quickly by
# looking for the # New Code # tags
#
# To run it in each of these various modes, follow the instructions
# in the readme for examples:
# https://github.com/huggingface/accelerate/tree/main/examples
#
########################################################################
_lowerCAmelCase = 16
_lowerCAmelCase = 32
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase = 16 ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = AutoTokenizer.from_pretrained("""bert-base-cased""" )
lowerCAmelCase__ : Dict = load_dataset("""glue""" , """mrpc""" )
def tokenize_function(UpperCamelCase ):
# max_length=None => use the model max length (it's actually the default)
lowerCAmelCase__ : Union[str, Any] = tokenizer(examples["""sentence1"""] , examples["""sentence2"""] , truncation=UpperCamelCase , max_length=UpperCamelCase )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
# starting with the main process first:
with accelerator.main_process_first():
lowerCAmelCase__ : Any = datasets.map(
UpperCamelCase , batched=UpperCamelCase , remove_columns=["""idx""", """sentence1""", """sentence2"""] , )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
lowerCAmelCase__ : List[str] = tokenized_datasets.rename_column("""label""" , """labels""" )
def collate_fn(UpperCamelCase ):
# On TPU it's best to pad everything to the same length or training will be very slow.
lowerCAmelCase__ : Dict = 128 if accelerator.distributed_type == DistributedType.TPU else None
# When using mixed precision we want round multiples of 8/16
if accelerator.mixed_precision == "fp8":
lowerCAmelCase__ : Dict = 16
elif accelerator.mixed_precision != "no":
lowerCAmelCase__ : Optional[Any] = 8
else:
lowerCAmelCase__ : Optional[Any] = None
return tokenizer.pad(
UpperCamelCase , padding="""longest""" , max_length=UpperCamelCase , pad_to_multiple_of=UpperCamelCase , return_tensors="""pt""" , )
# Instantiate dataloaders.
lowerCAmelCase__ : List[str] = DataLoader(
tokenized_datasets["""train"""] , shuffle=UpperCamelCase , collate_fn=UpperCamelCase , batch_size=UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = DataLoader(
tokenized_datasets["""validation"""] , shuffle=UpperCamelCase , collate_fn=UpperCamelCase , batch_size=UpperCamelCase )
return train_dataloader, eval_dataloader
# For testing only
if os.environ.get('''TESTING_MOCKED_DATALOADERS''', None) == "1":
from accelerate.test_utils.training import mocked_dataloaders
_lowerCAmelCase = mocked_dataloaders # noqa: F811
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
if os.environ.get("""TESTING_MOCKED_DATALOADERS""" , UpperCamelCase ) == "1":
lowerCAmelCase__ : Any = 2
# Initialize accelerator
lowerCAmelCase__ : List[str] = Accelerator(cpu=args.cpu , mixed_precision=args.mixed_precision )
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
lowerCAmelCase__ : Optional[int] = config["""lr"""]
lowerCAmelCase__ : Dict = int(config["""num_epochs"""] )
lowerCAmelCase__ : Optional[int] = int(config["""seed"""] )
lowerCAmelCase__ : Optional[Any] = int(config["""batch_size"""] )
lowerCAmelCase__ : List[Any] = evaluate.load("""glue""" , """mrpc""" )
# If the batch size is too big we use gradient accumulation
lowerCAmelCase__ : List[str] = 1
if batch_size > MAX_GPU_BATCH_SIZE and accelerator.distributed_type != DistributedType.TPU:
lowerCAmelCase__ : List[Any] = batch_size // MAX_GPU_BATCH_SIZE
lowerCAmelCase__ : Dict = MAX_GPU_BATCH_SIZE
set_seed(UpperCamelCase )
lowerCAmelCase__ , lowerCAmelCase__ : List[Any] = get_dataloaders(UpperCamelCase , UpperCamelCase )
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
lowerCAmelCase__ : Tuple = AutoModelForSequenceClassification.from_pretrained("""bert-base-cased""" , return_dict=UpperCamelCase )
# We could avoid this line since the accelerator is set with `device_placement=True` (default value).
# Note that if you are placing tensors on devices manually, this line absolutely needs to be before the optimizer
# creation otherwise training will not work on TPU (`accelerate` will kindly throw an error to make us aware of that).
lowerCAmelCase__ : Optional[int] = model.to(accelerator.device )
# Instantiate optimizer
lowerCAmelCase__ : List[str] = AdamW(params=model.parameters() , lr=UpperCamelCase )
# Instantiate scheduler
lowerCAmelCase__ : Tuple = get_linear_schedule_with_warmup(
optimizer=UpperCamelCase , num_warmup_steps=100 , num_training_steps=(len(UpperCamelCase ) * num_epochs) // gradient_accumulation_steps , )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = accelerator.prepare(
UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase )
# Now we train the model
for epoch in range(UpperCamelCase ):
model.train()
for step, batch in enumerate(UpperCamelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
lowerCAmelCase__ : List[str] = model(**UpperCamelCase )
lowerCAmelCase__ : int = outputs.loss
lowerCAmelCase__ : Union[str, Any] = loss / gradient_accumulation_steps
accelerator.backward(UpperCamelCase )
if step % gradient_accumulation_steps == 0:
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
model.eval()
lowerCAmelCase__ : str = 0
for step, batch in enumerate(UpperCamelCase ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
with torch.no_grad():
lowerCAmelCase__ : List[str] = model(**UpperCamelCase )
lowerCAmelCase__ : Tuple = outputs.logits.argmax(dim=-1 )
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = accelerator.gather((predictions, batch["""labels"""]) )
# New Code #
# First we check if it's a distributed system
if accelerator.use_distributed:
# Then see if we're on the last batch of our eval dataloader
if step == len(UpperCamelCase ) - 1:
# Last batch needs to be truncated on distributed systems as it contains additional samples
lowerCAmelCase__ : int = predictions[: len(eval_dataloader.dataset ) - samples_seen]
lowerCAmelCase__ : int = references[: len(eval_dataloader.dataset ) - samples_seen]
else:
# Otherwise we add the number of samples seen
samples_seen += references.shape[0]
# All of this can be avoided if you use `Accelerator.gather_for_metrics` instead of `Accelerator.gather`:
# accelerator.gather_for_metrics((predictions, batch["labels"]))
metric.add_batch(
predictions=UpperCamelCase , references=UpperCamelCase , )
lowerCAmelCase__ : Any = metric.compute()
# Use accelerator.print to print only on the main process.
accelerator.print(f"""epoch {epoch}:""" , UpperCamelCase )
def _SCREAMING_SNAKE_CASE ( ):
"""simple docstring"""
lowerCAmelCase__ : Optional[Any] = argparse.ArgumentParser(description="""Simple example of training script.""" )
parser.add_argument(
"""--mixed_precision""" , type=UpperCamelCase , default=UpperCamelCase , choices=["""no""", """fp16""", """bf16""", """fp8"""] , help="""Whether to use mixed precision. Choose"""
"""between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >= 1.10."""
"""and an Nvidia Ampere GPU.""" , )
parser.add_argument("""--cpu""" , action="""store_true""" , help="""If passed, will train on the CPU.""" )
lowerCAmelCase__ : Optional[int] = parser.parse_args()
lowerCAmelCase__ : Tuple = {"""lr""": 2e-5, """num_epochs""": 3, """seed""": 42, """batch_size""": 16}
training_function(UpperCamelCase , UpperCamelCase )
if __name__ == "__main__":
main()
| 37
|
'''simple docstring'''
import argparse
import json
import torch
from diffusers import DDPMScheduler, LDMPipeline, UNetaDModel, VQModel
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=1 ):
"""simple docstring"""
if n_shave_prefix_segments >= 0:
return ".".join(path.split(""".""" )[n_shave_prefix_segments:] )
else:
return ".".join(path.split(""".""" )[:n_shave_prefix_segments] )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : Union[str, Any] = []
for old_item in old_list:
lowerCAmelCase__ : Optional[Any] = old_item.replace("""in_layers.0""" , """norm1""" )
lowerCAmelCase__ : Optional[int] = new_item.replace("""in_layers.2""" , """conv1""" )
lowerCAmelCase__ : Dict = new_item.replace("""out_layers.0""" , """norm2""" )
lowerCAmelCase__ : str = new_item.replace("""out_layers.3""" , """conv2""" )
lowerCAmelCase__ : str = new_item.replace("""emb_layers.1""" , """time_emb_proj""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""skip_connection""" , """conv_shortcut""" )
lowerCAmelCase__ : Union[str, Any] = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase=0 ):
"""simple docstring"""
lowerCAmelCase__ : int = []
for old_item in old_list:
lowerCAmelCase__ : List[str] = old_item
lowerCAmelCase__ : int = new_item.replace("""norm.weight""" , """group_norm.weight""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""norm.bias""" , """group_norm.bias""" )
lowerCAmelCase__ : Optional[Any] = new_item.replace("""proj_out.weight""" , """proj_attn.weight""" )
lowerCAmelCase__ : int = new_item.replace("""proj_out.bias""" , """proj_attn.bias""" )
lowerCAmelCase__ : str = shave_segments(UpperCamelCase , n_shave_prefix_segments=UpperCamelCase )
mapping.append({"""old""": old_item, """new""": new_item} )
return mapping
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase , UpperCamelCase=None , UpperCamelCase=None , UpperCamelCase=None ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "Paths should be a list of dicts containing 'old' and 'new' keys."
# Splits the attention layers into three variables.
if attention_paths_to_split is not None:
for path, path_map in attention_paths_to_split.items():
lowerCAmelCase__ : Any = old_checkpoint[path]
lowerCAmelCase__ : int = old_tensor.shape[0] // 3
lowerCAmelCase__ : int = (-1, channels) if len(old_tensor.shape ) == 3 else (-1)
lowerCAmelCase__ : Tuple = old_tensor.shape[0] // config["""num_head_channels"""] // 3
lowerCAmelCase__ : List[Any] = old_tensor.reshape((num_heads, 3 * channels // num_heads) + old_tensor.shape[1:] )
lowerCAmelCase__ , lowerCAmelCase__ , lowerCAmelCase__ : str = old_tensor.split(channels // num_heads , dim=1 )
lowerCAmelCase__ : int = query.reshape(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = key.reshape(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = value.reshape(UpperCamelCase )
for path in paths:
lowerCAmelCase__ : Any = path["""new"""]
# These have already been assigned
if attention_paths_to_split is not None and new_path in attention_paths_to_split:
continue
# Global renaming happens here
lowerCAmelCase__ : Any = new_path.replace("""middle_block.0""" , """mid_block.resnets.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.1""" , """mid_block.attentions.0""" )
lowerCAmelCase__ : Any = new_path.replace("""middle_block.2""" , """mid_block.resnets.1""" )
if additional_replacements is not None:
for replacement in additional_replacements:
lowerCAmelCase__ : Any = new_path.replace(replacement["""old"""] , replacement["""new"""] )
# proj_attn.weight has to be converted from conv 1D to linear
if "proj_attn.weight" in new_path:
lowerCAmelCase__ : List[Any] = old_checkpoint[path["""old"""]][:, :, 0]
else:
lowerCAmelCase__ : Dict = old_checkpoint[path["""old"""]]
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : str = {}
lowerCAmelCase__ : str = checkpoint["""time_embed.0.weight"""]
lowerCAmelCase__ : List[Any] = checkpoint["""time_embed.0.bias"""]
lowerCAmelCase__ : int = checkpoint["""time_embed.2.weight"""]
lowerCAmelCase__ : List[str] = checkpoint["""time_embed.2.bias"""]
lowerCAmelCase__ : str = checkpoint["""input_blocks.0.0.weight"""]
lowerCAmelCase__ : Any = checkpoint["""input_blocks.0.0.bias"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.weight"""]
lowerCAmelCase__ : Union[str, Any] = checkpoint["""out.0.bias"""]
lowerCAmelCase__ : str = checkpoint["""out.2.weight"""]
lowerCAmelCase__ : Tuple = checkpoint["""out.2.bias"""]
# Retrieves the keys for the input blocks only
lowerCAmelCase__ : Optional[Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """input_blocks""" in layer} )
lowerCAmelCase__ : Optional[Any] = {
layer_id: [key for key in checkpoint if f"""input_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the middle blocks only
lowerCAmelCase__ : Union[str, Any] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """middle_block""" in layer} )
lowerCAmelCase__ : Union[str, Any] = {
layer_id: [key for key in checkpoint if f"""middle_block.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
# Retrieves the keys for the output blocks only
lowerCAmelCase__ : List[str] = len({""".""".join(layer.split(""".""" )[:2] ) for layer in checkpoint if """output_blocks""" in layer} )
lowerCAmelCase__ : List[Any] = {
layer_id: [key for key in checkpoint if f"""output_blocks.{layer_id}""" in key]
for layer_id in range(UpperCamelCase )
}
for i in range(1 , UpperCamelCase ):
lowerCAmelCase__ : Dict = (i - 1) // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Tuple = (i - 1) % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : Optional[int] = [key for key in input_blocks[i] if f"""input_blocks.{i}.0""" in key]
lowerCAmelCase__ : Optional[Any] = [key for key in input_blocks[i] if f"""input_blocks.{i}.1""" in key]
if f"""input_blocks.{i}.0.op.weight""" in checkpoint:
lowerCAmelCase__ : Optional[int] = checkpoint[
f"""input_blocks.{i}.0.op.weight"""
]
lowerCAmelCase__ : Tuple = checkpoint[
f"""input_blocks.{i}.0.op.bias"""
]
continue
lowerCAmelCase__ : Optional[Any] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Dict = {"""old""": f"""input_blocks.{i}.0""", """new""": f"""down_blocks.{block_id}.resnets.{layer_in_block_id}"""}
lowerCAmelCase__ : Optional[Any] = {"""old""": """resnets.2.op""", """new""": """downsamplers.0.op"""}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path, resnet_op] , config=UpperCamelCase )
if len(UpperCamelCase ):
lowerCAmelCase__ : Optional[Any] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""input_blocks.{i}.1""",
"""new""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : List[str] = {
f"""input_blocks.{i}.1.qkv.bias""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""input_blocks.{i}.1.qkv.weight""": {
"""key""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""down_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase , )
lowerCAmelCase__ : Dict = middle_blocks[0]
lowerCAmelCase__ : Union[str, Any] = middle_blocks[1]
lowerCAmelCase__ : Dict = middle_blocks[2]
lowerCAmelCase__ : Any = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Dict = renew_resnet_paths(UpperCamelCase )
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , config=UpperCamelCase )
lowerCAmelCase__ : Optional[int] = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {
"""middle_block.1.qkv.bias""": {
"""key""": """mid_block.attentions.0.key.bias""",
"""query""": """mid_block.attentions.0.query.bias""",
"""value""": """mid_block.attentions.0.value.bias""",
},
"""middle_block.1.qkv.weight""": {
"""key""": """mid_block.attentions.0.key.weight""",
"""query""": """mid_block.attentions.0.query.weight""",
"""value""": """mid_block.attentions.0.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , attention_paths_to_split=UpperCamelCase , config=UpperCamelCase )
for i in range(UpperCamelCase ):
lowerCAmelCase__ : Tuple = i // (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : List[str] = i % (config["""num_res_blocks"""] + 1)
lowerCAmelCase__ : int = [shave_segments(UpperCamelCase , 2 ) for name in output_blocks[i]]
lowerCAmelCase__ : Union[str, Any] = {}
for layer in output_block_layers:
lowerCAmelCase__ , lowerCAmelCase__ : Any = layer.split(""".""" )[0], shave_segments(UpperCamelCase , 1 )
if layer_id in output_block_list:
output_block_list[layer_id].append(UpperCamelCase )
else:
lowerCAmelCase__ : str = [layer_name]
if len(UpperCamelCase ) > 1:
lowerCAmelCase__ : str = [key for key in output_blocks[i] if f"""output_blocks.{i}.0""" in key]
lowerCAmelCase__ : Dict = [key for key in output_blocks[i] if f"""output_blocks.{i}.1""" in key]
lowerCAmelCase__ : Optional[int] = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = {"""old""": f"""output_blocks.{i}.0""", """new""": f"""up_blocks.{block_id}.resnets.{layer_in_block_id}"""}
assign_to_checkpoint(UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , config=UpperCamelCase )
if ["conv.weight", "conv.bias"] in output_block_list.values():
lowerCAmelCase__ : List[Any] = list(output_block_list.values() ).index(["""conv.weight""", """conv.bias"""] )
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.weight"""
]
lowerCAmelCase__ : int = checkpoint[
f"""output_blocks.{i}.{index}.conv.bias"""
]
# Clear attentions as they have been attributed above.
if len(UpperCamelCase ) == 2:
lowerCAmelCase__ : Tuple = []
if len(UpperCamelCase ):
lowerCAmelCase__ : Dict = renew_attention_paths(UpperCamelCase )
lowerCAmelCase__ : Tuple = {
"""old""": f"""output_blocks.{i}.1""",
"""new""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}""",
}
lowerCAmelCase__ : Tuple = {
f"""output_blocks.{i}.1.qkv.bias""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.bias""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.bias""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.bias""",
},
f"""output_blocks.{i}.1.qkv.weight""": {
"""key""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.key.weight""",
"""query""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.query.weight""",
"""value""": f"""up_blocks.{block_id}.attentions.{layer_in_block_id}.value.weight""",
},
}
assign_to_checkpoint(
UpperCamelCase , UpperCamelCase , UpperCamelCase , additional_replacements=[meta_path] , attention_paths_to_split=to_split if any("""qkv""" in key for key in attentions ) else None , config=UpperCamelCase , )
else:
lowerCAmelCase__ : int = renew_resnet_paths(UpperCamelCase , n_shave_prefix_segments=1 )
for path in resnet_0_paths:
lowerCAmelCase__ : Tuple = """.""".join(["""output_blocks""", str(UpperCamelCase ), path["""old"""]] )
lowerCAmelCase__ : List[Any] = """.""".join(["""up_blocks""", str(UpperCamelCase ), """resnets""", str(UpperCamelCase ), path["""new"""]] )
lowerCAmelCase__ : Union[str, Any] = checkpoint[old_path]
return new_checkpoint
if __name__ == "__main__":
_lowerCAmelCase = argparse.ArgumentParser()
parser.add_argument(
'''--checkpoint_path''', default=None, type=str, required=True, help='''Path to the checkpoint to convert.'''
)
parser.add_argument(
'''--config_file''',
default=None,
type=str,
required=True,
help='''The config json file corresponding to the architecture.''',
)
parser.add_argument('''--dump_path''', default=None, type=str, required=True, help='''Path to the output model.''')
_lowerCAmelCase = parser.parse_args()
_lowerCAmelCase = torch.load(args.checkpoint_path)
with open(args.config_file) as f:
_lowerCAmelCase = json.loads(f.read())
_lowerCAmelCase = convert_ldm_checkpoint(checkpoint, config)
if "ldm" in config:
del config["ldm"]
_lowerCAmelCase = UNetaDModel(**config)
model.load_state_dict(converted_checkpoint)
try:
_lowerCAmelCase = DDPMScheduler.from_config('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = VQModel.from_pretrained('''/'''.join(args.checkpoint_path.split('''/''')[:-1]))
_lowerCAmelCase = LDMPipeline(unet=model, scheduler=scheduler, vae=vqvae)
pipe.save_pretrained(args.dump_path)
except: # noqa: E722
model.save_pretrained(args.dump_path)
| 37
| 1
|
'''simple docstring'''
from __future__ import annotations
import collections
import pprint
from pathlib import Path
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return "".join(sorted(UpperCamelCase ) )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
return word_by_signature[signature(UpperCamelCase )]
_lowerCAmelCase = Path(__file__).parent.joinpath('''words.txt''').read_text(encoding='''utf-8''')
_lowerCAmelCase = sorted({word.strip().lower() for word in data.splitlines()})
_lowerCAmelCase = collections.defaultdict(list)
for word in word_list:
word_by_signature[signature(word)].append(word)
if __name__ == "__main__":
_lowerCAmelCase = {word: anagram(word) for word in word_list if len(anagram(word)) > 1}
with open('''anagrams.txt''', '''w''') as file:
file.write('''all_anagrams = \n ''')
file.write(pprint.pformat(all_anagrams))
| 37
|
'''simple docstring'''
from math import sqrt
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' must been an int and positive"
lowerCAmelCase__ : int = True
# 0 and 1 are none primes.
if number <= 1:
lowerCAmelCase__ : Optional[Any] = False
for divisor in range(2 , int(round(sqrt(UpperCamelCase ) ) ) + 1 ):
# if 'number' divisible by 'divisor' then sets 'status'
# of false and break up the loop.
if number % divisor == 0:
lowerCAmelCase__ : Any = False
break
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'status' must been from type bool"
return status
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
# beginList: contains all natural numbers from 2 up to N
lowerCAmelCase__ : List[str] = list(range(2 , n + 1 ) )
lowerCAmelCase__ : str = [] # this list will be returns.
# actual sieve of erathostenes
for i in range(len(UpperCamelCase ) ):
for j in range(i + 1 , len(UpperCamelCase ) ):
if (begin_list[i] != 0) and (begin_list[j] % begin_list[i] == 0):
lowerCAmelCase__ : List[Any] = 0
# filters actual prime numbers.
lowerCAmelCase__ : List[Any] = [x for x in begin_list if x != 0]
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n > 2), "'N' must been an int and > 2"
lowerCAmelCase__ : List[str] = []
# iterates over all numbers between 2 up to N+1
# if a number is prime then appends to list 'ans'
for number in range(2 , n + 1 ):
if is_prime(UpperCamelCase ):
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and number >= 0, "'number' must been an int and >= 0"
lowerCAmelCase__ : Optional[Any] = [] # this list will be returns of the function.
# potential prime number factors.
lowerCAmelCase__ : Dict = 2
lowerCAmelCase__ : Dict = number
if number == 0 or number == 1:
ans.append(UpperCamelCase )
# if 'number' not prime then builds the prime factorization of 'number'
elif not is_prime(UpperCamelCase ):
while quotient != 1:
if is_prime(UpperCamelCase ) and (quotient % factor == 0):
ans.append(UpperCamelCase )
quotient /= factor
else:
factor += 1
else:
ans.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type list"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : Optional[int] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = max(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number >= 0
), "'number' bust been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
# prime factorization of 'number'
lowerCAmelCase__ : List[str] = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Optional[int] = min(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ), "'ans' must been from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 == 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 == 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ), "'number' must been an int"
assert isinstance(number % 2 != 0 , UpperCamelCase ), "compare bust been from type bool"
return number % 2 != 0
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase ) and (number > 2) and is_even(UpperCamelCase )
), "'number' must been an int, even and > 2"
lowerCAmelCase__ : Dict = [] # this list will returned
# creates a list of prime numbers between 2 up to 'number'
lowerCAmelCase__ : Dict = get_prime_numbers(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = len(UpperCamelCase )
# run variable for while-loops.
lowerCAmelCase__ : List[str] = 0
lowerCAmelCase__ : List[Any] = None
# exit variable. for break up the loops
lowerCAmelCase__ : Any = True
while i < len_pn and loop:
lowerCAmelCase__ : List[Any] = i + 1
while j < len_pn and loop:
if prime_numbers[i] + prime_numbers[j] == number:
lowerCAmelCase__ : Optional[Any] = False
ans.append(prime_numbers[i] )
ans.append(prime_numbers[j] )
j += 1
i += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (len(UpperCamelCase ) == 2)
and (ans[0] + ans[1] == number)
and is_prime(ans[0] )
and is_prime(ans[1] )
), "'ans' must contains two primes. And sum of elements must been eq 'number'"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 0)
and (numbera >= 0)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 0
while numbera != 0:
lowerCAmelCase__ : Any = numbera % numbera
lowerCAmelCase__ : str = numbera
lowerCAmelCase__ : List[str] = rest
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
numbera >= 0
), "'number' must been from type int and positive"
return numbera
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (numbera >= 1)
and (numbera >= 1)
), "'number1' and 'number2' must been positive integer."
lowerCAmelCase__ : int = 1 # actual answer that will be return.
# for kgV (x,1)
if numbera > 1 and numbera > 1:
# builds the prime factorization of 'number1' and 'number2'
lowerCAmelCase__ : int = prime_factorization(UpperCamelCase )
lowerCAmelCase__ : Any = prime_factorization(UpperCamelCase )
elif numbera == 1 or numbera == 1:
lowerCAmelCase__ : Optional[Any] = []
lowerCAmelCase__ : Dict = []
lowerCAmelCase__ : List[str] = max(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = 0
lowerCAmelCase__ : str = 0
lowerCAmelCase__ : List[Any] = [] # captured numbers int both 'primeFac1' and 'primeFac2'
# iterates through primeFac1
for n in prime_fac_a:
if n not in done:
if n in prime_fac_a:
lowerCAmelCase__ : int = prime_fac_a.count(UpperCamelCase )
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(max(UpperCamelCase , UpperCamelCase ) ):
ans *= n
else:
lowerCAmelCase__ : Any = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# iterates through primeFac2
for n in prime_fac_a:
if n not in done:
lowerCAmelCase__ : Optional[int] = prime_fac_a.count(UpperCamelCase )
for _ in range(UpperCamelCase ):
ans *= n
done.append(UpperCamelCase )
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
ans >= 0
), "'ans' must been from type int and positive"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'number' must been a positive int"
lowerCAmelCase__ : Optional[Any] = 0
lowerCAmelCase__ : Tuple = 2 # this variable holds the answer
while index < n:
index += 1
ans += 1 # counts to the next number
# if ans not prime then
# runs to the next prime number.
while not is_prime(UpperCamelCase ):
ans += 1
# precondition
assert isinstance(UpperCamelCase , UpperCamelCase ) and is_prime(
UpperCamelCase ), "'ans' must been a prime number and from type int"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
is_prime(UpperCamelCase ) and is_prime(UpperCamelCase ) and (p_number_a < p_number_a)
), "The arguments must been prime numbers and 'pNumber1' < 'pNumber2'"
lowerCAmelCase__ : Dict = p_number_a + 1 # jump to the next number
lowerCAmelCase__ : List[Any] = [] # this list will be returns.
# if number is not prime then
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
while number < p_number_a:
ans.append(UpperCamelCase )
number += 1
# fetch the next prime number.
while not is_prime(UpperCamelCase ):
number += 1
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and ans[0] != p_number_a
and ans[len(UpperCamelCase ) - 1] != p_number_a
), "'ans' must been a list without the arguments"
# 'ans' contains not 'pNumber1' and 'pNumber2' !
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 1), "'n' must been int and >= 1"
lowerCAmelCase__ : List[Any] = [] # will be returned.
for divisor in range(1 , n + 1 ):
if n % divisor == 0:
ans.append(UpperCamelCase )
# precondition
assert ans[0] == 1 and ans[len(UpperCamelCase ) - 1] == n, "Error in function getDivisiors(...)"
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (
number > 1
), "'number' must been an int and >= 1"
lowerCAmelCase__ : Optional[int] = get_divisors(UpperCamelCase )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (divisors[0] == 1)
and (divisors[len(UpperCamelCase ) - 1] == number)
), "Error in help-function getDivisiors(...)"
# summed all divisors up to 'number' (exclusive), hence [:-1]
return sum(divisors[:-1] ) == number
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and isinstance(UpperCamelCase , UpperCamelCase )
and (denominator != 0)
), "The arguments must been from type int and 'denominator' != 0"
# build the greatest common divisor of numerator and denominator.
lowerCAmelCase__ : int = gcd(abs(UpperCamelCase ) , abs(UpperCamelCase ) )
# precondition
assert (
isinstance(UpperCamelCase , UpperCamelCase )
and (numerator % gcd_of_fraction == 0)
and (denominator % gcd_of_fraction == 0)
), "Error in function gcd(...,...)"
return (numerator // gcd_of_fraction, denominator // gcd_of_fraction)
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been a int and >= 0"
lowerCAmelCase__ : str = 1 # this will be return.
for factor in range(1 , n + 1 ):
ans *= factor
return ans
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
assert isinstance(UpperCamelCase , UpperCamelCase ) and (n >= 0), "'n' must been an int and >= 0"
lowerCAmelCase__ : List[Any] = 0
lowerCAmelCase__ : Any = 1
lowerCAmelCase__ : Optional[Any] = 1 # this will be return
for _ in range(n - 1 ):
lowerCAmelCase__ : Dict = ans
ans += fiba
lowerCAmelCase__ : str = tmp
return ans
| 37
| 1
|
'''simple docstring'''
from sklearn.metrics import fa_score, matthews_corrcoef
import datasets
from .record_evaluation import evaluate as evaluate_record
_lowerCAmelCase = '''\
@article{wang2019superglue,
title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
journal={arXiv preprint arXiv:1905.00537},
year={2019}
}
'''
_lowerCAmelCase = '''\
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
GLUE with a new set of more difficult language understanding tasks, improved
resources, and a new public leaderboard.
'''
_lowerCAmelCase = '''
Compute SuperGLUE evaluation metric associated to each SuperGLUE dataset.
Args:
predictions: list of predictions to score. Depending on the SuperGlUE subset:
- for \'record\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'prediction_text\': the predicted answer text
- for \'multirc\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question-answer pair as specified by the dataset
- \'prediction\': the predicted answer label
- otherwise: list of predicted labels
references: list of reference labels. Depending on the SuperGLUE subset:
- for \'record\': list of question-answers dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'answers\': list of possible answers
- otherwise: list of reference labels
Returns: depending on the SuperGLUE subset:
- for \'record\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1\': F1 score
- for \'multirc\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1_m\': Per-question macro-F1 score
- \'f1_a\': Average F1 score over all answers
- for \'axb\':
\'matthews_correlation\': Matthew Correlation
- for \'cb\':
- \'accuracy\': Accuracy
- \'f1\': F1 score
- for all others:
- \'accuracy\': Accuracy
Examples:
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'copa\') # any of ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'cb\')
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'record\')
>>> predictions = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'prediction_text\': \'answer\'}]
>>> references = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'answers\': [\'answer\', \'another_answer\']}]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'multirc\')
>>> predictions = [{\'idx\': {\'answer\': 0, \'paragraph\': 0, \'question\': 0}, \'prediction\': 0}, {\'idx\': {\'answer\': 1, \'paragraph\': 2, \'question\': 3}, \'prediction\': 1}]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1_m\': 1.0, \'f1_a\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'axb\')
>>> references = [0, 1]
>>> predictions = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'matthews_correlation\': 1.0}
'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return float((preds == labels).mean() )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase="binary" ):
"""simple docstring"""
lowerCAmelCase__ : Any = simple_accuracy(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = float(fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average=UpperCamelCase ) )
return {
"accuracy": acc,
"f1": fa,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
for id_pred, label in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : str = f"""{id_pred['idx']['paragraph']}-{id_pred['idx']['question']}"""
lowerCAmelCase__ : Dict = id_pred["""prediction"""]
if question_id in question_map:
question_map[question_id].append((pred, label) )
else:
lowerCAmelCase__ : Optional[int] = [(pred, label)]
lowerCAmelCase__ , lowerCAmelCase__ : int = [], []
for question, preds_labels in question_map.items():
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = zip(*UpperCamelCase )
lowerCAmelCase__ : List[Any] = fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average="""macro""" )
fas.append(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = int(sum(pred == label for pred, label in preds_labels ) == len(UpperCamelCase ) )
ems.append(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = float(sum(UpperCamelCase ) / len(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = sum(UpperCamelCase ) / len(UpperCamelCase )
lowerCAmelCase__ : Dict = float(fa_score(y_true=UpperCamelCase , y_pred=[id_pred["""prediction"""] for id_pred in ids_preds] ) )
return {"exact_match": em, "f1_m": fa_m, "f1_a": fa_a}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.config_name not in [
"boolq",
"cb",
"copa",
"multirc",
"record",
"rte",
"wic",
"wsc",
"wsc.fixed",
"axb",
"axg",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(self._get_feature_types() ) ,codebase_urls=[] ,reference_urls=[] ,format="""numpy""" if not self.config_name == """record""" and not self.config_name == """multirc""" else None ,)
def UpperCAmelCase_ ( self ) -> str:
if self.config_name == "record":
return {
"predictions": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"prediction_text": datasets.Value("""string""" ),
},
"references": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"answers": datasets.Sequence(datasets.Value("""string""" ) ),
},
}
elif self.config_name == "multirc":
return {
"predictions": {
"idx": {
"answer": datasets.Value("""int64""" ),
"paragraph": datasets.Value("""int64""" ),
"question": datasets.Value("""int64""" ),
},
"prediction": datasets.Value("""int64""" ),
},
"references": datasets.Value("""int64""" ),
}
else:
return {
"predictions": datasets.Value("""int64""" ),
"references": datasets.Value("""int64""" ),
}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
if self.config_name == "axb":
return {"matthews_correlation": matthews_corrcoef(__UpperCAmelCase ,__UpperCAmelCase )}
elif self.config_name == "cb":
return acc_and_fa(__UpperCAmelCase ,__UpperCAmelCase ,fa_avg="""macro""" )
elif self.config_name == "record":
lowerCAmelCase__ : Optional[Any] = [
{
"""qas""": [
{"""id""": ref["""idx"""]["""query"""], """answers""": [{"""text""": ans} for ans in ref["""answers"""]]}
for ref in references
]
}
]
lowerCAmelCase__ : Union[str, Any] = {pred["""idx"""]["""query"""]: pred["""prediction_text"""] for pred in predictions}
return evaluate_record(__UpperCAmelCase ,__UpperCAmelCase )[0]
elif self.config_name == "multirc":
return evaluate_multirc(__UpperCAmelCase ,__UpperCAmelCase )
elif self.config_name in ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]:
return {"accuracy": simple_accuracy(__UpperCAmelCase ,__UpperCAmelCase )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
| 37
|
'''simple docstring'''
from sklearn.metrics import fa_score, matthews_corrcoef
import datasets
from .record_evaluation import evaluate as evaluate_record
_lowerCAmelCase = '''\
@article{wang2019superglue,
title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},
author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},
journal={arXiv preprint arXiv:1905.00537},
year={2019}
}
'''
_lowerCAmelCase = '''\
SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after
GLUE with a new set of more difficult language understanding tasks, improved
resources, and a new public leaderboard.
'''
_lowerCAmelCase = '''
Compute SuperGLUE evaluation metric associated to each SuperGLUE dataset.
Args:
predictions: list of predictions to score. Depending on the SuperGlUE subset:
- for \'record\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'prediction_text\': the predicted answer text
- for \'multirc\': list of question-answer dictionaries with the following keys:
- \'idx\': index of the question-answer pair as specified by the dataset
- \'prediction\': the predicted answer label
- otherwise: list of predicted labels
references: list of reference labels. Depending on the SuperGLUE subset:
- for \'record\': list of question-answers dictionaries with the following keys:
- \'idx\': index of the question as specified by the dataset
- \'answers\': list of possible answers
- otherwise: list of reference labels
Returns: depending on the SuperGLUE subset:
- for \'record\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1\': F1 score
- for \'multirc\':
- \'exact_match\': Exact match between answer and gold answer
- \'f1_m\': Per-question macro-F1 score
- \'f1_a\': Average F1 score over all answers
- for \'axb\':
\'matthews_correlation\': Matthew Correlation
- for \'cb\':
- \'accuracy\': Accuracy
- \'f1\': F1 score
- for all others:
- \'accuracy\': Accuracy
Examples:
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'copa\') # any of ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'cb\')
>>> predictions = [0, 1]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'accuracy\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'record\')
>>> predictions = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'prediction_text\': \'answer\'}]
>>> references = [{\'idx\': {\'passage\': 0, \'query\': 0}, \'answers\': [\'answer\', \'another_answer\']}]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'multirc\')
>>> predictions = [{\'idx\': {\'answer\': 0, \'paragraph\': 0, \'question\': 0}, \'prediction\': 0}, {\'idx\': {\'answer\': 1, \'paragraph\': 2, \'question\': 3}, \'prediction\': 1}]
>>> references = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'exact_match\': 1.0, \'f1_m\': 1.0, \'f1_a\': 1.0}
>>> super_glue_metric = datasets.load_metric(\'super_glue\', \'axb\')
>>> references = [0, 1]
>>> predictions = [0, 1]
>>> results = super_glue_metric.compute(predictions=predictions, references=references)
>>> print(results)
{\'matthews_correlation\': 1.0}
'''
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
return float((preds == labels).mean() )
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase , UpperCamelCase="binary" ):
"""simple docstring"""
lowerCAmelCase__ : Any = simple_accuracy(UpperCamelCase , UpperCamelCase )
lowerCAmelCase__ : Tuple = float(fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average=UpperCamelCase ) )
return {
"accuracy": acc,
"f1": fa,
}
def _SCREAMING_SNAKE_CASE ( UpperCamelCase , UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ : List[str] = {}
for id_pred, label in zip(UpperCamelCase , UpperCamelCase ):
lowerCAmelCase__ : str = f"""{id_pred['idx']['paragraph']}-{id_pred['idx']['question']}"""
lowerCAmelCase__ : Dict = id_pred["""prediction"""]
if question_id in question_map:
question_map[question_id].append((pred, label) )
else:
lowerCAmelCase__ : Optional[int] = [(pred, label)]
lowerCAmelCase__ , lowerCAmelCase__ : int = [], []
for question, preds_labels in question_map.items():
lowerCAmelCase__ , lowerCAmelCase__ : Optional[int] = zip(*UpperCamelCase )
lowerCAmelCase__ : List[Any] = fa_score(y_true=UpperCamelCase , y_pred=UpperCamelCase , average="""macro""" )
fas.append(UpperCamelCase )
lowerCAmelCase__ : Union[str, Any] = int(sum(pred == label for pred, label in preds_labels ) == len(UpperCamelCase ) )
ems.append(UpperCamelCase )
lowerCAmelCase__ : Optional[Any] = float(sum(UpperCamelCase ) / len(UpperCamelCase ) )
lowerCAmelCase__ : List[Any] = sum(UpperCamelCase ) / len(UpperCamelCase )
lowerCAmelCase__ : Dict = float(fa_score(y_true=UpperCamelCase , y_pred=[id_pred["""prediction"""] for id_pred in ids_preds] ) )
return {"exact_match": em, "f1_m": fa_m, "f1_a": fa_a}
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCAmelCase_( datasets.Metric ):
'''simple docstring'''
def UpperCAmelCase_ ( self ) -> Optional[Any]:
if self.config_name not in [
"boolq",
"cb",
"copa",
"multirc",
"record",
"rte",
"wic",
"wsc",
"wsc.fixed",
"axb",
"axg",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
return datasets.MetricInfo(
description=_DESCRIPTION ,citation=_CITATION ,inputs_description=_KWARGS_DESCRIPTION ,features=datasets.Features(self._get_feature_types() ) ,codebase_urls=[] ,reference_urls=[] ,format="""numpy""" if not self.config_name == """record""" and not self.config_name == """multirc""" else None ,)
def UpperCAmelCase_ ( self ) -> str:
if self.config_name == "record":
return {
"predictions": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"prediction_text": datasets.Value("""string""" ),
},
"references": {
"idx": {
"passage": datasets.Value("""int64""" ),
"query": datasets.Value("""int64""" ),
},
"answers": datasets.Sequence(datasets.Value("""string""" ) ),
},
}
elif self.config_name == "multirc":
return {
"predictions": {
"idx": {
"answer": datasets.Value("""int64""" ),
"paragraph": datasets.Value("""int64""" ),
"question": datasets.Value("""int64""" ),
},
"prediction": datasets.Value("""int64""" ),
},
"references": datasets.Value("""int64""" ),
}
else:
return {
"predictions": datasets.Value("""int64""" ),
"references": datasets.Value("""int64""" ),
}
def UpperCAmelCase_ ( self ,__UpperCAmelCase ,__UpperCAmelCase ) -> Any:
if self.config_name == "axb":
return {"matthews_correlation": matthews_corrcoef(__UpperCAmelCase ,__UpperCAmelCase )}
elif self.config_name == "cb":
return acc_and_fa(__UpperCAmelCase ,__UpperCAmelCase ,fa_avg="""macro""" )
elif self.config_name == "record":
lowerCAmelCase__ : Optional[Any] = [
{
"""qas""": [
{"""id""": ref["""idx"""]["""query"""], """answers""": [{"""text""": ans} for ans in ref["""answers"""]]}
for ref in references
]
}
]
lowerCAmelCase__ : Union[str, Any] = {pred["""idx"""]["""query"""]: pred["""prediction_text"""] for pred in predictions}
return evaluate_record(__UpperCAmelCase ,__UpperCAmelCase )[0]
elif self.config_name == "multirc":
return evaluate_multirc(__UpperCAmelCase ,__UpperCAmelCase )
elif self.config_name in ["copa", "rte", "wic", "wsc", "wsc.fixed", "boolq", "axg"]:
return {"accuracy": simple_accuracy(__UpperCAmelCase ,__UpperCAmelCase )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"boolq\", \"cb\", \"copa\", \"multirc\", \"record\", \"rte\", \"wic\", \"wsc\", \"wsc.fixed\", \"axb\", \"axg\",]""" )
| 37
| 1
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.