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 |
|---|---|---|---|---|
def lowerCAmelCase_ ( _snake_case : int = 1000 ) -> int:
'''simple docstring'''
__magic_name__ : Dict = 2**power
__magic_name__ : Optional[Any] = str(_snake_case )
__magic_name__ : List[Any] = list(_snake_case )
__magic_name__ : Optional[int] = 0
for i in list_num:
sum_of_num += int(_snake_case )
return sum_of_num
if __name__ == "__main__":
snake_case : List[str] = int(input("Enter the power of 2: ").strip())
print("2 ^ ", power, " = ", 2**power)
snake_case : str = solution(power)
print("Sum of the digits is: ", result)
| 281 |
import unittest
import numpy as np
from transformers import RobertaPreLayerNormConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
import jax.numpy as jnp
from transformers.models.roberta_prelayernorm.modeling_flax_roberta_prelayernorm import (
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormModel,
)
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=True , _a=True , _a=True , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=16 , _a=2 , _a=0.02 , _a=4 , ):
__magic_name__ : List[Any] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Dict = seq_length
__magic_name__ : Union[str, Any] = is_training
__magic_name__ : Optional[Any] = use_attention_mask
__magic_name__ : Optional[Any] = use_token_type_ids
__magic_name__ : int = use_labels
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : Any = intermediate_size
__magic_name__ : List[Any] = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[Any] = max_position_embeddings
__magic_name__ : Tuple = type_vocab_size
__magic_name__ : List[str] = type_sequence_label_size
__magic_name__ : Dict = initializer_range
__magic_name__ : List[Any] = num_choices
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = None
if self.use_attention_mask:
__magic_name__ : List[str] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : str = None
if self.use_token_type_ids:
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__magic_name__ : List[str] = RobertaPreLayerNormConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=_a , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[Any] = config_and_inputs
__magic_name__ : List[str] = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": attention_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : Tuple = True
__magic_name__ : int = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
token_type_ids,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
# Copied from tests.models.roberta.test_modelling_flax_roberta.FlaxRobertaPreLayerNormModelTest with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta-base->andreasmadsen/efficient_mlm_m0.40
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = True
UpperCamelCase__ = (
(
FlaxRobertaPreLayerNormModel,
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
)
if is_flax_available()
else ()
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxRobertaPreLayerNormModelTester(self )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Optional[Any] = model_class_name.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(_a )
@require_flax
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxRobertaPreLayerNormForMaskedLM.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Union[str, Any] = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : List[str] = model(_a )[0]
__magic_name__ : str = [1, 11, 50_265]
self.assertEqual(list(output.shape ) , _a )
# compare the actual values for a slice.
__magic_name__ : List[str] = np.array(
[[[40.48_80, 18.01_99, -5.23_67], [-1.88_77, -4.08_85, 10.70_85], [-2.26_13, -5.61_10, 7.26_65]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = FlaxRobertaPreLayerNormModel.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Tuple = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : Tuple = model(_a )[0]
# compare the actual values for a slice.
__magic_name__ : Dict = np.array(
[[[0.02_08, -0.03_56, 0.02_37], [-0.15_69, -0.04_11, -0.26_26], [0.18_79, 0.01_25, -0.00_89]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
| 281 | 1 |
import torch
from transformers import CamembertForMaskedLM, CamembertTokenizer
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any] , _snake_case : Union[str, Any] , _snake_case : str=5 ) -> Any:
'''simple docstring'''
assert masked_input.count("<mask>" ) == 1
__magic_name__ : Union[str, Any] = torch.tensor(tokenizer.encode(_snake_case , add_special_tokens=_snake_case ) ).unsqueeze(0 ) # Batch size 1
__magic_name__ : int = model(_snake_case )[0] # The last hidden-state is the first element of the output tuple
__magic_name__ : Optional[int] = (input_ids.squeeze() == tokenizer.mask_token_id).nonzero().item()
__magic_name__ : int = logits[0, masked_index, :]
__magic_name__ : str = logits.softmax(dim=0 )
__magic_name__ , __magic_name__ : Tuple = prob.topk(k=_snake_case , dim=0 )
__magic_name__ : List[Any] = " ".join(
[tokenizer.convert_ids_to_tokens(indices[i].item() ) for i in range(len(_snake_case ) )] )
__magic_name__ : Dict = tokenizer.mask_token
__magic_name__ : List[str] = []
for index, predicted_token_bpe in enumerate(topk_predicted_token_bpe.split(" " ) ):
__magic_name__ : Any = predicted_token_bpe.replace("\u2581" , " " )
if " {0}".format(_snake_case ) in masked_input:
topk_filled_outputs.append(
(
masked_input.replace(" {0}".format(_snake_case ) , _snake_case ),
values[index].item(),
predicted_token,
) )
else:
topk_filled_outputs.append(
(
masked_input.replace(_snake_case , _snake_case ),
values[index].item(),
predicted_token,
) )
return topk_filled_outputs
snake_case : Dict = CamembertTokenizer.from_pretrained("camembert-base")
snake_case : Optional[int] = CamembertForMaskedLM.from_pretrained("camembert-base")
model.eval()
snake_case : str = "Le camembert est <mask> :)"
print(fill_mask(masked_input, model, tokenizer, topk=3))
| 281 |
def lowerCAmelCase_ ( _snake_case : list[list[int | float]] ) -> int:
'''simple docstring'''
__magic_name__ : Any = len(_snake_case )
__magic_name__ : Optional[Any] = len(matrix[0] )
__magic_name__ : Union[str, Any] = min(_snake_case , _snake_case )
for row in range(_snake_case ):
# Check if diagonal element is not zero
if matrix[row][row] != 0:
# Eliminate all the elements below the diagonal
for col in range(row + 1 , _snake_case ):
__magic_name__ : Optional[Any] = matrix[col][row] / matrix[row][row]
for i in range(_snake_case , _snake_case ):
matrix[col][i] -= multiplier * matrix[row][i]
else:
# Find a non-zero diagonal element to swap rows
__magic_name__ : str = True
for i in range(row + 1 , _snake_case ):
if matrix[i][row] != 0:
__magic_name__ , __magic_name__ : List[str] = matrix[i], matrix[row]
__magic_name__ : Union[str, Any] = False
break
if reduce:
rank -= 1
for i in range(_snake_case ):
__magic_name__ : Any = matrix[i][rank]
# Reduce the row pointer by one to stay on the same row
row -= 1
return rank
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
from collections import OrderedDict
from typing import TYPE_CHECKING, Any, Mapping, Optional, Union
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
if TYPE_CHECKING:
from ... import FeatureExtractionMixin, PreTrainedTokenizerBase, TensorType
snake_case : Optional[int] = logging.get_logger(__name__)
snake_case : List[Any] = {
"microsoft/deberta-v2-xlarge": "https://huggingface.co/microsoft/deberta-v2-xlarge/resolve/main/config.json",
"microsoft/deberta-v2-xxlarge": "https://huggingface.co/microsoft/deberta-v2-xxlarge/resolve/main/config.json",
"microsoft/deberta-v2-xlarge-mnli": (
"https://huggingface.co/microsoft/deberta-v2-xlarge-mnli/resolve/main/config.json"
),
"microsoft/deberta-v2-xxlarge-mnli": (
"https://huggingface.co/microsoft/deberta-v2-xxlarge-mnli/resolve/main/config.json"
),
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'deberta-v2'
def __init__( self , _a=128_100 , _a=1_536 , _a=24 , _a=24 , _a=6_144 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=0 , _a=0.02 , _a=1e-7 , _a=False , _a=-1 , _a=0 , _a=True , _a=None , _a=0 , _a="gelu" , **_a , ):
super().__init__(**_a )
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : List[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : List[Any] = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : Union[str, Any] = hidden_dropout_prob
__magic_name__ : str = attention_probs_dropout_prob
__magic_name__ : Union[str, Any] = max_position_embeddings
__magic_name__ : Optional[Any] = type_vocab_size
__magic_name__ : Dict = initializer_range
__magic_name__ : Optional[int] = relative_attention
__magic_name__ : Any = max_relative_positions
__magic_name__ : Dict = pad_token_id
__magic_name__ : str = position_biased_input
# Backwards compatibility
if type(_a ) == str:
__magic_name__ : Dict = [x.strip() for x in pos_att_type.lower().split("|" )]
__magic_name__ : Optional[int] = pos_att_type
__magic_name__ : Union[str, Any] = vocab_size
__magic_name__ : Any = layer_norm_eps
__magic_name__ : Optional[int] = kwargs.get("pooler_hidden_size" , _a )
__magic_name__ : List[Any] = pooler_dropout
__magic_name__ : Union[str, Any] = pooler_hidden_act
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : Union[str, Any] = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Optional[Any] = {0: "batch", 1: "sequence"}
if self._config.type_vocab_size > 0:
return OrderedDict(
[("input_ids", dynamic_axis), ("attention_mask", dynamic_axis), ("token_type_ids", dynamic_axis)] )
else:
return OrderedDict([("input_ids", dynamic_axis), ("attention_mask", dynamic_axis)] )
@property
def SCREAMING_SNAKE_CASE ( self ):
return 12
def SCREAMING_SNAKE_CASE ( self , _a , _a = -1 , _a = -1 , _a = -1 , _a = False , _a = None , _a = 3 , _a = 40 , _a = 40 , _a = None , ):
__magic_name__ : Dict = super().generate_dummy_inputs(preprocessor=_a , framework=_a )
if self._config.type_vocab_size == 0 and "token_type_ids" in dummy_inputs:
del dummy_inputs["token_type_ids"]
return dummy_inputs
| 281 |
import argparse
import collections
import json
import os
import re
import string
import sys
import numpy as np
snake_case : Dict = re.compile(R"\b(a|an|the)\b", re.UNICODE)
snake_case : Optional[int] = None
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser("Official evaluation script for SQuAD version 2.0." )
parser.add_argument("data_file" , metavar="data.json" , help="Input data JSON file." )
parser.add_argument("pred_file" , metavar="pred.json" , help="Model predictions." )
parser.add_argument(
"--out-file" , "-o" , metavar="eval.json" , help="Write accuracy metrics to file (default is stdout)." )
parser.add_argument(
"--na-prob-file" , "-n" , metavar="na_prob.json" , help="Model estimates of probability of no answer." )
parser.add_argument(
"--na-prob-thresh" , "-t" , type=_snake_case , default=1.0 , help="Predict \"\" if no-answer probability exceeds this (default = 1.0)." , )
parser.add_argument(
"--out-image-dir" , "-p" , metavar="out_images" , default=_snake_case , help="Save precision-recall curves to directory." )
parser.add_argument("--verbose" , "-v" , action="store_true" )
if len(sys.argv ) == 1:
parser.print_help()
sys.exit(1 )
return parser.parse_args()
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Optional[int] = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : str = bool(qa["answers"]["text"] )
return qid_to_has_ans
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Optional[Any]:
'''simple docstring'''
def remove_articles(_snake_case : List[str] ):
return ARTICLES_REGEX.sub(" " , _snake_case )
def white_space_fix(_snake_case : Optional[int] ):
return " ".join(text.split() )
def remove_punc(_snake_case : Optional[int] ):
__magic_name__ : Dict = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(_snake_case : str ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(_snake_case ) ) ) )
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
if not s:
return []
return normalize_answer(_snake_case ).split()
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Dict ) -> Tuple:
'''simple docstring'''
return int(normalize_answer(_snake_case ) == normalize_answer(_snake_case ) )
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : int ) -> str:
'''simple docstring'''
__magic_name__ : Any = get_tokens(_snake_case )
__magic_name__ : Optional[int] = get_tokens(_snake_case )
__magic_name__ : Tuple = collections.Counter(_snake_case ) & collections.Counter(_snake_case )
__magic_name__ : Tuple = sum(common.values() )
if len(_snake_case ) == 0 or len(_snake_case ) == 0:
# If either is no-answer, then F1 is 1 if they agree, 0 otherwise
return int(gold_toks == pred_toks )
if num_same == 0:
return 0
__magic_name__ : Dict = 1.0 * num_same / len(_snake_case )
__magic_name__ : Optional[Any] = 1.0 * num_same / len(_snake_case )
__magic_name__ : List[Any] = (2 * precision * recall) / (precision + recall)
return fa
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = {}
__magic_name__ : int = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : Union[str, Any] = qa["id"]
__magic_name__ : Any = [t for t in qa["answers"]["text"] if normalize_answer(_snake_case )]
if not gold_answers:
# For unanswerable questions, only correct answer is empty string
__magic_name__ : Tuple = [""]
if qid not in preds:
print(F'''Missing prediction for {qid}''' )
continue
__magic_name__ : Any = preds[qid]
# Take max over all gold answers
__magic_name__ : List[Any] = max(compute_exact(_snake_case , _snake_case ) for a in gold_answers )
__magic_name__ : int = max(compute_fa(_snake_case , _snake_case ) for a in gold_answers )
return exact_scores, fa_scores
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] , _snake_case : Optional[int] , _snake_case : Dict ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : str = {}
for qid, s in scores.items():
__magic_name__ : Dict = na_probs[qid] > na_prob_thresh
if pred_na:
__magic_name__ : str = float(not qid_to_has_ans[qid] )
else:
__magic_name__ : Optional[int] = s
return new_scores
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : List[str] , _snake_case : Tuple=None ) -> Tuple:
'''simple docstring'''
if not qid_list:
__magic_name__ : Any = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores.values() ) / total),
("f1", 100.0 * sum(fa_scores.values() ) / total),
("total", total),
] )
else:
__magic_name__ : Tuple = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores[k] for k in qid_list ) / total),
("f1", 100.0 * sum(fa_scores[k] for k in qid_list ) / total),
("total", total),
] )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : str , _snake_case : str ) -> Dict:
'''simple docstring'''
for k in new_eval:
__magic_name__ : int = new_eval[k]
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : Optional[Any] , _snake_case : Union[str, Any] ) -> str:
'''simple docstring'''
plt.step(_snake_case , _snake_case , color="b" , alpha=0.2 , where="post" )
plt.fill_between(_snake_case , _snake_case , step="post" , alpha=0.2 , color="b" )
plt.xlabel("Recall" )
plt.ylabel("Precision" )
plt.xlim([0.0, 1.05] )
plt.ylim([0.0, 1.05] )
plt.title(_snake_case )
plt.savefig(_snake_case )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any , _snake_case : Optional[int] , _snake_case : List[Any] , _snake_case : Optional[int]=None , _snake_case : int=None ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
__magic_name__ : Optional[int] = 0.0
__magic_name__ : str = 1.0
__magic_name__ : str = 0.0
__magic_name__ : List[str] = [1.0]
__magic_name__ : str = [0.0]
__magic_name__ : Optional[Any] = 0.0
for i, qid in enumerate(_snake_case ):
if qid_to_has_ans[qid]:
true_pos += scores[qid]
__magic_name__ : List[str] = true_pos / float(i + 1 )
__magic_name__ : Any = true_pos / float(_snake_case )
if i == len(_snake_case ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]:
# i.e., if we can put a threshold after this point
avg_prec += cur_p * (cur_r - recalls[-1])
precisions.append(_snake_case )
recalls.append(_snake_case )
if out_image:
plot_pr_curve(_snake_case , _snake_case , _snake_case , _snake_case )
return {"ap": 100.0 * avg_prec}
def lowerCAmelCase_ ( _snake_case : Tuple , _snake_case : Optional[Any] , _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : Any , _snake_case : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
if out_image_dir and not os.path.exists(_snake_case ):
os.makedirs(_snake_case )
__magic_name__ : Any = sum(1 for v in qid_to_has_ans.values() if v )
if num_true_pos == 0:
return
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_exact.png" ) , title="Precision-Recall curve for Exact Match score" , )
__magic_name__ : Union[str, Any] = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_f1.png" ) , title="Precision-Recall curve for F1 score" , )
__magic_name__ : str = {k: float(_snake_case ) for k, v in qid_to_has_ans.items()}
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_oracle.png" ) , title="Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)" , )
merge_eval(_snake_case , _snake_case , "pr_exact" )
merge_eval(_snake_case , _snake_case , "pr_f1" )
merge_eval(_snake_case , _snake_case , "pr_oracle" )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : Optional[Any] , _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
if not qid_list:
return
__magic_name__ : Dict = [na_probs[k] for k in qid_list]
__magic_name__ : str = np.ones_like(_snake_case ) / float(len(_snake_case ) )
plt.hist(_snake_case , weights=_snake_case , bins=20 , range=(0.0, 1.0) )
plt.xlabel("Model probability of no-answer" )
plt.ylabel("Proportion of dataset" )
plt.title(F'''Histogram of no-answer probability: {name}''' )
plt.savefig(os.path.join(_snake_case , F'''na_prob_hist_{name}.png''' ) )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : List[str] , _snake_case : Dict ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] )
__magic_name__ : List[str] = num_no_ans
__magic_name__ : Dict = cur_score
__magic_name__ : Dict = 0.0
__magic_name__ : Any = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
for i, qid in enumerate(_snake_case ):
if qid not in scores:
continue
if qid_to_has_ans[qid]:
__magic_name__ : Union[str, Any] = scores[qid]
else:
if preds[qid]:
__magic_name__ : List[Any] = -1
else:
__magic_name__ : Optional[int] = 0
cur_score += diff
if cur_score > best_score:
__magic_name__ : Optional[int] = cur_score
__magic_name__ : List[Any] = na_probs[qid]
return 100.0 * best_score / len(_snake_case ), best_thresh
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str , _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Dict ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ , __magic_name__ : List[str] = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ , __magic_name__ : int = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ : Optional[int] = best_exact
__magic_name__ : List[Any] = exact_thresh
__magic_name__ : Dict = best_fa
__magic_name__ : Any = fa_thresh
def lowerCAmelCase_ ( ) -> int:
'''simple docstring'''
with open(OPTS.data_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
__magic_name__ : List[Any] = dataset_json["data"]
with open(OPTS.pred_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
if OPTS.na_prob_file:
with open(OPTS.na_prob_file ) as f:
__magic_name__ : Any = json.load(_snake_case )
else:
__magic_name__ : Any = {k: 0.0 for k in preds}
__magic_name__ : str = make_qid_to_has_ans(_snake_case ) # maps qid to True/False
__magic_name__ : Tuple = [k for k, v in qid_to_has_ans.items() if v]
__magic_name__ : Optional[Any] = [k for k, v in qid_to_has_ans.items() if not v]
__magic_name__ , __magic_name__ : Union[str, Any] = get_raw_scores(_snake_case , _snake_case )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case )
if has_ans_qids:
__magic_name__ : int = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "HasAns" )
if no_ans_qids:
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "NoAns" )
if OPTS.na_prob_file:
find_all_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
if OPTS.na_prob_file and OPTS.out_image_dir:
run_precision_recall_analysis(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , OPTS.out_image_dir )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "hasAns" )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "noAns" )
if OPTS.out_file:
with open(OPTS.out_file , "w" ) as f:
json.dump(_snake_case , _snake_case )
else:
print(json.dumps(_snake_case , indent=2 ) )
if __name__ == "__main__":
snake_case : int = parse_args()
if OPTS.out_image_dir:
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
main()
| 281 | 1 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
snake_case : Optional[Any] = {"configuration_fnet": ["FNET_PRETRAINED_CONFIG_ARCHIVE_MAP", "FNetConfig"]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : str = ["FNetTokenizer"]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : List[Any] = ["FNetTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : Union[str, Any] = [
"FNET_PRETRAINED_MODEL_ARCHIVE_LIST",
"FNetForMaskedLM",
"FNetForMultipleChoice",
"FNetForNextSentencePrediction",
"FNetForPreTraining",
"FNetForQuestionAnswering",
"FNetForSequenceClassification",
"FNetForTokenClassification",
"FNetLayer",
"FNetModel",
"FNetPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_fnet import FNET_PRETRAINED_CONFIG_ARCHIVE_MAP, FNetConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_fnet import FNetTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_fnet_fast import FNetTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_fnet import (
FNET_PRETRAINED_MODEL_ARCHIVE_LIST,
FNetForMaskedLM,
FNetForMultipleChoice,
FNetForNextSentencePrediction,
FNetForPreTraining,
FNetForQuestionAnswering,
FNetForSequenceClassification,
FNetForTokenClassification,
FNetLayer,
FNetModel,
FNetPreTrainedModel,
)
else:
import sys
snake_case : Optional[Any] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 281 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
snake_case : str = "▁"
snake_case : List[Any] = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = BigBirdTokenizer
UpperCamelCase__ = BigBirdTokenizerFast
UpperCamelCase__ = True
UpperCamelCase__ = True
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
__magic_name__ : Optional[Any] = self.tokenizer_class(_a , keep_accents=_a )
tokenizer.save_pretrained(self.tmpdirname )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = "<s>"
__magic_name__ : Dict = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_a ) , _a )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_a ) , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : 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] , "[MASK]" )
self.assertEqual(len(_a ) , 1_004 )
def SCREAMING_SNAKE_CASE ( self ):
self.assertEqual(self.get_tokenizer().vocab_size , 1_000 )
def SCREAMING_SNAKE_CASE ( self ):
if not self.test_rust_tokenizer:
return
__magic_name__ : Dict = self.get_tokenizer()
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Any = "I was born in 92000, and this is falsé."
__magic_name__ : Dict = tokenizer.tokenize(_a )
__magic_name__ : Any = rust_tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : List[Any] = tokenizer.encode(_a , add_special_tokens=_a )
__magic_name__ : List[str] = rust_tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Dict = tokenizer.encode(_a )
__magic_name__ : Optional[int] = rust_tokenizer.encode(_a )
self.assertListEqual(_a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = BigBirdTokenizer(_a , keep_accents=_a )
__magic_name__ : str = tokenizer.tokenize("This is a test" )
self.assertListEqual(_a , ["▁This", "▁is", "▁a", "▁t", "est"] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_a ) , [285, 46, 10, 170, 382] , )
__magic_name__ : Dict = tokenizer.tokenize("I was born in 92000, and this is falsé." )
self.assertListEqual(
_a , [
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",
".",
] , )
__magic_name__ : Union[str, Any] = tokenizer.convert_tokens_to_ids(_a )
self.assertListEqual(
_a , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
__magic_name__ : int = tokenizer.convert_ids_to_tokens(_a )
self.assertListEqual(
_a , [
SPIECE_UNDERLINE + "I",
SPIECE_UNDERLINE + "was",
SPIECE_UNDERLINE + "b",
"or",
"n",
SPIECE_UNDERLINE + "in",
SPIECE_UNDERLINE + "",
"<unk>",
"2",
"0",
"0",
"0",
",",
SPIECE_UNDERLINE + "and",
SPIECE_UNDERLINE + "this",
SPIECE_UNDERLINE + "is",
SPIECE_UNDERLINE + "f",
"al",
"s",
"<unk>",
".",
] , )
@cached_property
def SCREAMING_SNAKE_CASE ( self ):
return BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "Hello World!"
__magic_name__ : Dict = [65, 18_536, 2_260, 101, 66]
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = (
"This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"
" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"
)
# fmt: off
__magic_name__ : List[str] = [65, 871, 419, 358, 946, 991, 2_521, 452, 358, 1_357, 387, 7_751, 3_536, 112, 985, 456, 126, 865, 938, 5_400, 5_734, 458, 1_368, 467, 786, 2_462, 5_246, 1_159, 633, 865, 4_519, 457, 582, 852, 2_557, 427, 916, 508, 405, 34_324, 497, 391, 408, 11_342, 1_244, 385, 100, 938, 985, 456, 574, 362, 12_597, 3_200, 3_129, 1_172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
__magic_name__ : Optional[Any] = list(self.big_tokenizer.get_vocab().keys() )[:10]
__magic_name__ : List[Any] = " ".join(_a )
__magic_name__ : Any = self.big_tokenizer.encode_plus(_a , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : Union[str, Any] = self.big_tokenizer.batch_encode_plus(
[sequence + " " + sequence] , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : List[str] = BigBirdConfig(attention_type="original_full" )
__magic_name__ : Optional[int] = BigBirdModel(_a )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_a )
model(**_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
__magic_name__ : int = tokenizer.decode(tokenizer("Paris is the [MASK]." ).input_ids )
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# fmt: off
__magic_name__ : Optional[Any] = {"input_ids": [[65, 39_286, 458, 36_335, 2_001, 456, 13_073, 13_266, 455, 113, 7_746, 1_741, 11_157, 391, 13_073, 13_266, 455, 113, 3_967, 35_412, 113, 4_936, 109, 3_870, 2_377, 113, 30_084, 45_720, 458, 134, 17_496, 112, 503, 11_672, 113, 118, 112, 5_665, 13_347, 38_687, 112, 1_496, 31_389, 112, 3_268, 47_264, 134, 962, 112, 16_377, 8_035, 23_130, 430, 12_169, 15_518, 28_592, 458, 146, 41_697, 109, 391, 12_169, 15_518, 16_689, 458, 146, 41_358, 109, 452, 726, 4_034, 111, 763, 35_412, 5_082, 388, 1_903, 111, 9_051, 391, 2_870, 48_918, 1_900, 1_123, 550, 998, 112, 9_586, 15_985, 455, 391, 410, 22_955, 37_636, 114, 66], [65, 448, 17_496, 419, 3_663, 385, 763, 113, 27_533, 2_870, 3_283, 13_043, 1_639, 24_713, 523, 656, 24_013, 18_550, 2_521, 517, 27_014, 21_244, 420, 1_212, 1_465, 391, 927, 4_833, 388, 578, 11_786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2_169, 7_687, 21_932, 18_146, 726, 363, 17_032, 3_391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], "attention_mask": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_a , model_name="google/bigbird-roberta-base" , revision="215c99f1600e06f83acce68422f2035b2b5c3510" , )
| 281 | 1 |
snake_case : Dict = "Input must be a string of 8 numbers plus letter"
snake_case : Any = "TRWAGMYFPDXBNJZSQVHLCKE"
def lowerCAmelCase_ ( _snake_case : str ) -> bool:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ):
__magic_name__ : List[str] = F'''Expected string as input, found {type(_snake_case ).__name__}'''
raise TypeError(_snake_case )
__magic_name__ : int = spanish_id.replace("-" , "" ).upper()
if len(_snake_case ) != 9:
raise ValueError(_snake_case )
try:
__magic_name__ : Optional[int] = int(spanish_id_clean[0:8] )
__magic_name__ : Optional[int] = spanish_id_clean[8]
except ValueError as ex:
raise ValueError(_snake_case ) from ex
if letter.isdigit():
raise ValueError(_snake_case )
return letter == LOOKUP_LETTERS[number % 23]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
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
snake_case : int = logging.get_logger(__name__)
snake_case : List[str] = {"vocab_file": "spiece.model"}
snake_case : List[str] = {
"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",
}
}
snake_case : Tuple = {
"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,
}
snake_case : List[str] = "▁"
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a=True , _a=True , _a=False , _a="[CLS]" , _a="[SEP]" , _a="<unk>" , _a="[SEP]" , _a="<pad>" , _a="[CLS]" , _a="[MASK]" , _a = None , **_a , ):
# 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.
__magic_name__ : str = (
AddedToken(_a , lstrip=_a , rstrip=_a , normalized=_a )
if isinstance(_a , _a )
else mask_token
)
__magic_name__ : Union[str, Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_a , remove_space=_a , keep_accents=_a , bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , pad_token=_a , cls_token=_a , mask_token=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Dict = do_lower_case
__magic_name__ : Tuple = remove_space
__magic_name__ : Union[str, Any] = keep_accents
__magic_name__ : Tuple = vocab_file
__magic_name__ : int = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_a )
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.sp_model )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ):
__magic_name__ : List[str] = self.__dict__.copy()
__magic_name__ : Any = None
return state
def __setstate__( self , _a ):
__magic_name__ : Union[str, Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : str = {}
__magic_name__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.remove_space:
__magic_name__ : List[Any] = " ".join(inputs.strip().split() )
else:
__magic_name__ : str = inputs
__magic_name__ : int = outputs.replace("``" , "\"" ).replace("''" , "\"" )
if not self.keep_accents:
__magic_name__ : str = unicodedata.normalize("NFKD" , _a )
__magic_name__ : Tuple = "".join([c for c in outputs if not unicodedata.combining(_a )] )
if self.do_lower_case:
__magic_name__ : int = outputs.lower()
return outputs
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = self.preprocess_text(_a )
__magic_name__ : Dict = self.sp_model.encode(_a , out_type=_a )
__magic_name__ : Any = []
for piece in pieces:
if len(_a ) > 1 and piece[-1] == str("," ) and piece[-2].isdigit():
__magic_name__ : Optional[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(_a , "" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
__magic_name__ : List[str] = cur_pieces[1:]
else:
__magic_name__ : Optional[int] = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(_a )
else:
new_pieces.append(_a )
return new_pieces
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.PieceToId(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.IdToPiece(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Any = []
__magic_name__ : Union[str, Any] = ""
__magic_name__ : int = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_a ) + token
__magic_name__ : List[Any] = True
__magic_name__ : Optional[int] = []
else:
current_sub_tokens.append(_a )
__magic_name__ : Optional[Any] = False
out_string += self.sp_model.decode(_a )
return out_string.strip()
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : List[str] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is not None:
return [1] + ([0] * len(_a )) + [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[int] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : List[str] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _a )
elif not os.path.isfile(self.vocab_file ):
with open(_a , "wb" ) as fi:
__magic_name__ : Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(_a )
return (out_vocab_file,)
| 281 | 1 |
from ...utils import (
OptionalDependencyNotAvailable,
is_flax_available,
is_torch_available,
is_transformers_available,
)
try:
if not (is_transformers_available() and is_torch_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ...utils.dummy_torch_and_transformers_objects import * # noqa F403
else:
from .multicontrolnet import MultiControlNetModel
from .pipeline_controlnet import StableDiffusionControlNetPipeline
from .pipeline_controlnet_imgaimg import StableDiffusionControlNetImgaImgPipeline
from .pipeline_controlnet_inpaint import StableDiffusionControlNetInpaintPipeline
if is_transformers_available() and is_flax_available():
from .pipeline_flax_controlnet import FlaxStableDiffusionControlNetPipeline
| 281 |
import unicodedata
from dataclasses import dataclass
from typing import Optional, Union
import numpy as np
from transformers.data.data_collator import DataCollatorMixin
from transformers.file_utils import PaddingStrategy
from transformers.tokenization_utils_base import PreTrainedTokenizerBase
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Union[str, Any] = np.full((len(_snake_case ), sequence_length, 2) , _snake_case )
else:
__magic_name__ : List[Any] = np.full((len(_snake_case ), sequence_length) , _snake_case )
for i, tensor in enumerate(_snake_case ):
if padding_side == "right":
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Optional[Any] = tensor[:sequence_length]
else:
__magic_name__ : Union[str, Any] = tensor[:sequence_length]
else:
if isinstance(_snake_case , _snake_case ):
__magic_name__ : List[Any] = tensor[:sequence_length]
else:
__magic_name__ : Optional[Any] = tensor[:sequence_length]
return out_tensor.tolist()
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = ord(_snake_case )
if (cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126):
return True
__magic_name__ : Any = unicodedata.category(_snake_case )
if cat.startswith("P" ):
return True
return False
@dataclass
class _snake_case ( snake_case ):
UpperCamelCase__ = 42
UpperCamelCase__ = True
UpperCamelCase__ = None
UpperCamelCase__ = None
UpperCamelCase__ = -100
UpperCamelCase__ = "pt"
def SCREAMING_SNAKE_CASE ( self , _a ):
import torch
__magic_name__ : List[str] = "label" if "label" in features[0].keys() else "labels"
__magic_name__ : Union[str, Any] = [feature[label_name] for feature in features] if label_name in features[0].keys() else None
__magic_name__ : Optional[int] = self.tokenizer.pad(
_a , padding=self.padding , max_length=self.max_length , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors="pt" if labels is None else None , )
if labels is None:
return batch
__magic_name__ : Dict = torch.tensor(batch["entity_ids"] ).shape[1]
__magic_name__ : List[Any] = self.tokenizer.padding_side
if padding_side == "right":
__magic_name__ : str = [
list(_a ) + [self.label_pad_token_id] * (sequence_length - len(_a )) for label in labels
]
else:
__magic_name__ : int = [
[self.label_pad_token_id] * (sequence_length - len(_a )) + list(_a ) for label in labels
]
__magic_name__ : Dict = [feature["ner_tags"] for feature in features]
__magic_name__ : List[Any] = padding_tensor(_a , -1 , _a , _a )
__magic_name__ : Any = [feature["original_entity_spans"] for feature in features]
__magic_name__ : Any = padding_tensor(_a , (-1, -1) , _a , _a )
__magic_name__ : List[Any] = {k: torch.tensor(_a , dtype=torch.intaa ) for k, v in batch.items()}
return batch
| 281 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
snake_case : Any = logging.get_logger(__name__)
snake_case : Dict = {
"facebook/dpr-ctx_encoder-single-nq-base": (
"https://huggingface.co/facebook/dpr-ctx_encoder-single-nq-base/resolve/main/config.json"
),
"facebook/dpr-question_encoder-single-nq-base": (
"https://huggingface.co/facebook/dpr-question_encoder-single-nq-base/resolve/main/config.json"
),
"facebook/dpr-reader-single-nq-base": (
"https://huggingface.co/facebook/dpr-reader-single-nq-base/resolve/main/config.json"
),
"facebook/dpr-ctx_encoder-multiset-base": (
"https://huggingface.co/facebook/dpr-ctx_encoder-multiset-base/resolve/main/config.json"
),
"facebook/dpr-question_encoder-multiset-base": (
"https://huggingface.co/facebook/dpr-question_encoder-multiset-base/resolve/main/config.json"
),
"facebook/dpr-reader-multiset-base": (
"https://huggingface.co/facebook/dpr-reader-multiset-base/resolve/main/config.json"
),
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'dpr'
def __init__( self , _a=30_522 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=0 , _a="absolute" , _a = 0 , **_a , ):
super().__init__(pad_token_id=_a , **_a )
__magic_name__ : Optional[int] = vocab_size
__magic_name__ : List[str] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : Dict = num_attention_heads
__magic_name__ : Optional[Any] = hidden_act
__magic_name__ : Union[str, Any] = intermediate_size
__magic_name__ : Optional[Any] = hidden_dropout_prob
__magic_name__ : Union[str, Any] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : int = type_vocab_size
__magic_name__ : Tuple = initializer_range
__magic_name__ : Optional[int] = layer_norm_eps
__magic_name__ : Tuple = projection_dim
__magic_name__ : Dict = position_embedding_type
| 281 |
import math
def lowerCAmelCase_ ( _snake_case : float , _snake_case : float ) -> float:
'''simple docstring'''
return math.pow(_snake_case , 2 ) - a
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
return 2 * x
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
__magic_name__ : Optional[int] = 2.0
while start <= a:
__magic_name__ : str = math.pow(_snake_case , 2 )
return start
def lowerCAmelCase_ ( _snake_case : float , _snake_case : int = 9999 , _snake_case : float = 0.00_000_000_000_001 ) -> float:
'''simple docstring'''
if a < 0:
raise ValueError("math domain error" )
__magic_name__ : Optional[int] = get_initial_point(_snake_case )
for _ in range(_snake_case ):
__magic_name__ : int = value
__magic_name__ : str = value - fx(_snake_case , _snake_case ) / fx_derivative(_snake_case )
if abs(prev_value - value ) < tolerance:
return value
return value
if __name__ == "__main__":
from doctest import testmod
testmod()
| 281 | 1 |
import json
import os
import unittest
from transformers import MgpstrTokenizer
from transformers.models.mgp_str.tokenization_mgp_str import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = MgpstrTokenizer
UpperCamelCase__ = False
UpperCamelCase__ = {}
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
# fmt: off
__magic_name__ : List[Any] = ["[GO]", "[s]", "0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z"]
# fmt: on
__magic_name__ : Optional[int] = dict(zip(_a , range(len(_a ) ) ) )
__magic_name__ : Optional[int] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["vocab_file"] )
with open(self.vocab_file , "w" , encoding="utf-8" ) as fp:
fp.write(json.dumps(_a ) + "\n" )
def SCREAMING_SNAKE_CASE ( self , **_a ):
return MgpstrTokenizer.from_pretrained(self.tmpdirname , **_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[Any] = "tester"
__magic_name__ : int = "tester"
return input_text, output_text
@unittest.skip("MGP-STR always lower cases letters." )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = self.get_tokenizers(do_lower_case=_a )
for tokenizer in tokenizers:
with self.subTest(f'''{tokenizer.__class__.__name__}''' ):
__magic_name__ : Tuple = "[SPECIAL_TOKEN]"
tokenizer.add_special_tokens({"cls_token": special_token} )
__magic_name__ : Any = tokenizer.encode([special_token] , add_special_tokens=_a )
self.assertEqual(len(_a ) , 1 )
__magic_name__ : str = tokenizer.decode(_a , skip_special_tokens=_a )
self.assertTrue(special_token not in decoded )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = self.get_tokenizers()
for tokenizer in tokenizers:
with self.subTest(f'''{tokenizer.__class__.__name__}''' ):
__magic_name__ , __magic_name__ : Dict = self.get_input_output_texts(_a )
__magic_name__ : Tuple = tokenizer.tokenize(_a )
__magic_name__ : Optional[int] = tokenizer.convert_tokens_to_ids(_a )
__magic_name__ : List[Any] = tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : Any = tokenizer.convert_ids_to_tokens(_a )
self.assertNotEqual(len(_a ) , 0 )
__magic_name__ : Dict = tokenizer.decode(_a )
self.assertIsInstance(_a , _a )
self.assertEqual(text_a.replace(" " , "" ) , _a )
@unittest.skip("MGP-STR tokenizer only handles one sequence." )
def SCREAMING_SNAKE_CASE ( self ):
pass
@unittest.skip("inputs cannot be pretokenized in MgpstrTokenizer" )
def SCREAMING_SNAKE_CASE ( self ):
pass
| 281 |
from __future__ import annotations
import unittest
from transformers import LEDConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFLEDForConditionalGeneration, TFLEDModel
@require_tf
class _snake_case :
UpperCamelCase__ = LEDConfig
UpperCamelCase__ = {}
UpperCamelCase__ = 'gelu'
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=False , _a=99 , _a=32 , _a=2 , _a=4 , _a=37 , _a=0.1 , _a=0.1 , _a=20 , _a=2 , _a=1 , _a=0 , _a=4 , ):
__magic_name__ : int = parent
__magic_name__ : Optional[int] = batch_size
__magic_name__ : Tuple = seq_length
__magic_name__ : List[Any] = is_training
__magic_name__ : Dict = use_labels
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : int = hidden_size
__magic_name__ : Optional[int] = num_hidden_layers
__magic_name__ : Optional[int] = num_attention_heads
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Any = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Any = eos_token_id
__magic_name__ : str = pad_token_id
__magic_name__ : int = bos_token_id
__magic_name__ : Optional[int] = attention_window
# `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size
# [num_attention_heads, encoder_seq_length, encoder_key_length], but TFLongformerSelfAttention
# returns attention of shape [num_attention_heads, encoder_seq_length, self.attention_window + 1]
# because its local attention only attends to `self.attention_window` and one before and one after
__magic_name__ : Tuple = self.attention_window + 2
# because of padding `encoder_seq_length`, is different from `seq_length`. Relevant for
# the `test_attention_outputs` and `test_hidden_states_output` tests
__magic_name__ : Tuple = (
self.seq_length + (self.attention_window - self.seq_length % self.attention_window) % self.attention_window
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
__magic_name__ : int = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
__magic_name__ : int = tf.concat([input_ids, eos_tensor] , axis=1 )
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : Dict = self.config_cls(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_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_ids=[2] , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.pad_token_id , attention_window=self.attention_window , **self.config_updates , )
__magic_name__ : List[str] = prepare_led_inputs_dict(_a , _a , _a )
__magic_name__ : Union[str, Any] = tf.concat(
[tf.zeros_like(_a )[:, :-1], tf.ones_like(_a )[:, -1:]] , axis=-1 , )
__magic_name__ : List[Any] = global_attention_mask
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Dict = TFLEDModel(config=_a ).get_decoder()
__magic_name__ : Optional[int] = inputs_dict["input_ids"]
__magic_name__ : Union[str, Any] = input_ids[:1, :]
__magic_name__ : str = inputs_dict["attention_mask"][:1, :]
__magic_name__ : int = 1
# first forward pass
__magic_name__ : Tuple = model(_a , attention_mask=_a , use_cache=_a )
__magic_name__ , __magic_name__ : str = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
__magic_name__ : Optional[int] = ids_tensor((self.batch_size, 3) , config.vocab_size )
__magic_name__ : Any = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
__magic_name__ : Optional[Any] = tf.concat([input_ids, next_tokens] , axis=-1 )
__magic_name__ : List[Any] = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
__magic_name__ : List[str] = model(_a , attention_mask=_a )[0]
__magic_name__ : Dict = model(_a , attention_mask=_a , past_key_values=_a )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
__magic_name__ : List[Any] = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
__magic_name__ : Union[str, Any] = output_from_no_past[:, -3:, random_slice_idx]
__magic_name__ : List[str] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(_a , _a , rtol=1e-3 )
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : List[Any] , _snake_case : Any , _snake_case : str=None , _snake_case : List[str]=None , _snake_case : int=None , _snake_case : Any=None , ) -> int:
'''simple docstring'''
if attention_mask is None:
__magic_name__ : str = tf.cast(tf.math.not_equal(_snake_case , config.pad_token_id ) , tf.inta )
if decoder_attention_mask is None:
__magic_name__ : List[Any] = tf.concat(
[
tf.ones(decoder_input_ids[:, :1].shape , dtype=tf.inta ),
tf.cast(tf.math.not_equal(decoder_input_ids[:, 1:] , config.pad_token_id ) , tf.inta ),
] , axis=-1 , )
if head_mask is None:
__magic_name__ : Dict = tf.ones((config.encoder_layers, config.encoder_attention_heads) )
if decoder_head_mask is None:
__magic_name__ : Any = tf.ones((config.decoder_layers, config.decoder_attention_heads) )
return {
"input_ids": input_ids,
"attention_mask": attention_mask,
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
}
@require_tf
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (TFLEDForConditionalGeneration, TFLEDModel) if is_tf_available() else ()
UpperCamelCase__ = (TFLEDForConditionalGeneration,) if is_tf_available() else ()
UpperCamelCase__ = (
{
'conversational': TFLEDForConditionalGeneration,
'feature-extraction': TFLEDModel,
'summarization': TFLEDForConditionalGeneration,
'text2text-generation': TFLEDForConditionalGeneration,
'translation': TFLEDForConditionalGeneration,
}
if is_tf_available()
else {}
)
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = TFLEDModelTester(self )
__magic_name__ : List[Any] = ConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : int = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[str] = tf.zeros_like(inputs_dict["attention_mask"] )
__magic_name__ : Optional[Any] = 2
__magic_name__ : Tuple = tf.where(
tf.range(self.model_tester.seq_length )[None, :] < num_global_attn_indices , 1 , inputs_dict["global_attention_mask"] , )
__magic_name__ : Any = True
__magic_name__ : str = self.model_tester.seq_length
__magic_name__ : Dict = self.model_tester.encoder_seq_length
def check_decoder_attentions_output(_a ):
__magic_name__ : str = outputs.decoder_attentions
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(decoder_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
def check_encoder_attentions_output(_a ):
__magic_name__ : Any = [t.numpy() for t in outputs.encoder_attentions]
__magic_name__ : Tuple = [t.numpy() for t in outputs.encoder_global_attentions]
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
self.assertListEqual(
list(global_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, encoder_seq_length, num_global_attn_indices] , )
for model_class in self.all_model_classes:
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = False
__magic_name__ : Tuple = False
__magic_name__ : Optional[int] = model_class(_a )
__magic_name__ : str = model(self._prepare_for_class(_a , _a ) )
__magic_name__ : Any = len(_a )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
if self.is_encoder_decoder:
__magic_name__ : Tuple = model_class(_a )
__magic_name__ : Optional[Any] = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_decoder_attentions_output(_a )
# Check that output attentions can also be changed via the config
del inputs_dict["output_attentions"]
__magic_name__ : Dict = True
__magic_name__ : str = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
# Check attention is always last and order is fine
__magic_name__ : Union[str, Any] = True
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(out_len + (2 if self.is_encoder_decoder else 1) , len(_a ) )
self.assertEqual(model.config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
@unittest.skip("LED keeps using potentially symbolic tensors in conditionals and breaks tracing." )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
# TODO: Head-masking not yet implement
pass
def lowerCAmelCase_ ( _snake_case : int ) -> Optional[int]:
'''simple docstring'''
return tf.constant(_snake_case , dtype=tf.intaa )
snake_case : Optional[int] = 1E-4
@slow
@require_tf
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" ).led
# change to intended input here
__magic_name__ : Optional[int] = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : str = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Any = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : List[Any] = model(**_a )[0]
__magic_name__ : List[str] = (1, 1_024, 768)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : int = tf.convert_to_tensor(
[[2.30_50, 2.82_79, 0.65_31], [-1.84_57, -0.14_55, -3.56_61], [-1.01_86, 0.45_86, -2.20_43]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" )
# change to intended input here
__magic_name__ : int = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Tuple = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Optional[Any] = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : Union[str, Any] = model(**_a )[0]
__magic_name__ : Optional[int] = (1, 1_024, model.config.vocab_size)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : str = tf.convert_to_tensor(
[[33.65_07, 6.45_72, 16.80_89], [5.87_39, -2.42_38, 11.29_02], [-3.21_39, -4.31_49, 4.27_83]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 , rtol=1e-3 )
| 281 | 1 |
import argparse
import os
import shutil
from pathlib import Path
import onnx
import torch
from packaging import version
from torch.onnx import export
from diffusers import OnnxRuntimeModel, OnnxStableDiffusionPipeline, StableDiffusionPipeline
snake_case : Tuple = version.parse(version.parse(torch.__version__).base_version) < version.parse("1.11")
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : tuple , _snake_case : Path , _snake_case : Optional[Any] , _snake_case : Any , _snake_case : Dict , _snake_case : Optional[Any] , _snake_case : List[str]=False , ) -> Optional[Any]:
'''simple docstring'''
output_path.parent.mkdir(parents=_snake_case , exist_ok=_snake_case )
# PyTorch deprecated the `enable_onnx_checker` and `use_external_data_format` arguments in v1.11,
# so we check the torch version for backwards compatibility
if is_torch_less_than_1_11:
export(
_snake_case , _snake_case , f=output_path.as_posix() , input_names=_snake_case , output_names=_snake_case , dynamic_axes=_snake_case , do_constant_folding=_snake_case , use_external_data_format=_snake_case , enable_onnx_checker=_snake_case , opset_version=_snake_case , )
else:
export(
_snake_case , _snake_case , f=output_path.as_posix() , input_names=_snake_case , output_names=_snake_case , dynamic_axes=_snake_case , do_constant_folding=_snake_case , opset_version=_snake_case , )
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str , _snake_case : int , _snake_case : bool = False ) -> str:
'''simple docstring'''
__magic_name__ : Tuple = torch.floataa if fpaa else torch.floataa
if fpaa and torch.cuda.is_available():
__magic_name__ : List[str] = "cuda"
elif fpaa and not torch.cuda.is_available():
raise ValueError("`float16` model export is only supported on GPUs with CUDA" )
else:
__magic_name__ : Tuple = "cpu"
__magic_name__ : Optional[Any] = StableDiffusionPipeline.from_pretrained(_snake_case , torch_dtype=_snake_case ).to(_snake_case )
__magic_name__ : List[str] = Path(_snake_case )
# TEXT ENCODER
__magic_name__ : Tuple = pipeline.text_encoder.config.max_position_embeddings
__magic_name__ : List[str] = pipeline.text_encoder.config.hidden_size
__magic_name__ : List[Any] = pipeline.tokenizer(
"A sample prompt" , padding="max_length" , max_length=pipeline.tokenizer.model_max_length , truncation=_snake_case , return_tensors="pt" , )
onnx_export(
pipeline.text_encoder , model_args=(text_input.input_ids.to(device=_snake_case , dtype=torch.intaa )) , output_path=output_path / "text_encoder" / "model.onnx" , ordered_input_names=["input_ids"] , output_names=["last_hidden_state", "pooler_output"] , dynamic_axes={
"input_ids": {0: "batch", 1: "sequence"},
} , opset=_snake_case , )
del pipeline.text_encoder
# UNET
__magic_name__ : Any = pipeline.unet.config.in_channels
__magic_name__ : Tuple = pipeline.unet.config.sample_size
__magic_name__ : int = output_path / "unet" / "model.onnx"
onnx_export(
pipeline.unet , model_args=(
torch.randn(2 , _snake_case , _snake_case , _snake_case ).to(device=_snake_case , dtype=_snake_case ),
torch.randn(2 ).to(device=_snake_case , dtype=_snake_case ),
torch.randn(2 , _snake_case , _snake_case ).to(device=_snake_case , dtype=_snake_case ),
False,
) , output_path=_snake_case , ordered_input_names=["sample", "timestep", "encoder_hidden_states", "return_dict"] , output_names=["out_sample"] , dynamic_axes={
"sample": {0: "batch", 1: "channels", 2: "height", 3: "width"},
"timestep": {0: "batch"},
"encoder_hidden_states": {0: "batch", 1: "sequence"},
} , opset=_snake_case , use_external_data_format=_snake_case , )
__magic_name__ : Any = str(unet_path.absolute().as_posix() )
__magic_name__ : int = os.path.dirname(_snake_case )
__magic_name__ : Tuple = onnx.load(_snake_case )
# clean up existing tensor files
shutil.rmtree(_snake_case )
os.mkdir(_snake_case )
# collate external tensor files into one
onnx.save_model(
_snake_case , _snake_case , save_as_external_data=_snake_case , all_tensors_to_one_file=_snake_case , location="weights.pb" , convert_attribute=_snake_case , )
del pipeline.unet
# VAE ENCODER
__magic_name__ : int = pipeline.vae
__magic_name__ : Tuple = vae_encoder.config.in_channels
__magic_name__ : List[Any] = vae_encoder.config.sample_size
# need to get the raw tensor output (sample) from the encoder
__magic_name__ : Union[str, Any] = lambda _snake_case , _snake_case : vae_encoder.encode(_snake_case , _snake_case )[0].sample()
onnx_export(
_snake_case , model_args=(
torch.randn(1 , _snake_case , _snake_case , _snake_case ).to(device=_snake_case , dtype=_snake_case ),
False,
) , output_path=output_path / "vae_encoder" / "model.onnx" , ordered_input_names=["sample", "return_dict"] , output_names=["latent_sample"] , dynamic_axes={
"sample": {0: "batch", 1: "channels", 2: "height", 3: "width"},
} , opset=_snake_case , )
# VAE DECODER
__magic_name__ : Dict = pipeline.vae
__magic_name__ : Tuple = vae_decoder.config.latent_channels
__magic_name__ : Optional[Any] = vae_decoder.config.out_channels
# forward only through the decoder part
__magic_name__ : str = vae_encoder.decode
onnx_export(
_snake_case , model_args=(
torch.randn(1 , _snake_case , _snake_case , _snake_case ).to(device=_snake_case , dtype=_snake_case ),
False,
) , output_path=output_path / "vae_decoder" / "model.onnx" , ordered_input_names=["latent_sample", "return_dict"] , output_names=["sample"] , dynamic_axes={
"latent_sample": {0: "batch", 1: "channels", 2: "height", 3: "width"},
} , opset=_snake_case , )
del pipeline.vae
# SAFETY CHECKER
if pipeline.safety_checker is not None:
__magic_name__ : List[Any] = pipeline.safety_checker
__magic_name__ : Optional[int] = safety_checker.config.vision_config.num_channels
__magic_name__ : str = safety_checker.config.vision_config.image_size
__magic_name__ : List[Any] = safety_checker.forward_onnx
onnx_export(
pipeline.safety_checker , model_args=(
torch.randn(
1 , _snake_case , _snake_case , _snake_case , ).to(device=_snake_case , dtype=_snake_case ),
torch.randn(1 , _snake_case , _snake_case , _snake_case ).to(device=_snake_case , dtype=_snake_case ),
) , output_path=output_path / "safety_checker" / "model.onnx" , ordered_input_names=["clip_input", "images"] , output_names=["out_images", "has_nsfw_concepts"] , dynamic_axes={
"clip_input": {0: "batch", 1: "channels", 2: "height", 3: "width"},
"images": {0: "batch", 1: "height", 2: "width", 3: "channels"},
} , opset=_snake_case , )
del pipeline.safety_checker
__magic_name__ : Optional[int] = OnnxRuntimeModel.from_pretrained(output_path / "safety_checker" )
__magic_name__ : Tuple = pipeline.feature_extractor
else:
__magic_name__ : Optional[Any] = None
__magic_name__ : Dict = None
__magic_name__ : Optional[int] = OnnxStableDiffusionPipeline(
vae_encoder=OnnxRuntimeModel.from_pretrained(output_path / "vae_encoder" ) , vae_decoder=OnnxRuntimeModel.from_pretrained(output_path / "vae_decoder" ) , text_encoder=OnnxRuntimeModel.from_pretrained(output_path / "text_encoder" ) , tokenizer=pipeline.tokenizer , unet=OnnxRuntimeModel.from_pretrained(output_path / "unet" ) , scheduler=pipeline.scheduler , safety_checker=_snake_case , feature_extractor=_snake_case , requires_safety_checker=safety_checker is not None , )
onnx_pipeline.save_pretrained(_snake_case )
print("ONNX pipeline saved to" , _snake_case )
del pipeline
del onnx_pipeline
__magic_name__ : int = OnnxStableDiffusionPipeline.from_pretrained(_snake_case , provider="CPUExecutionProvider" )
print("ONNX pipeline is loadable" )
if __name__ == "__main__":
snake_case : List[str] = argparse.ArgumentParser()
parser.add_argument(
"--model_path",
type=str,
required=True,
help="Path to the `diffusers` checkpoint to convert (either a local directory or on the Hub).",
)
parser.add_argument("--output_path", type=str, required=True, help="Path to the output model.")
parser.add_argument(
"--opset",
default=14,
type=int,
help="The version of the ONNX operator set to use.",
)
parser.add_argument("--fp16", action="store_true", default=False, help="Export the models in `float16` mode")
snake_case : List[Any] = parser.parse_args()
convert_models(args.model_path, args.output_path, args.opset, args.fpaa)
| 281 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import (
BitConfig,
ViTHybridConfig,
ViTHybridForImageClassification,
ViTHybridImageProcessor,
ViTHybridModel,
)
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
snake_case : Optional[Any] = logging.get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any]=False ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
# fmt: off
# stem:
rename_keys.append(("cls_token", "vit.embeddings.cls_token") )
rename_keys.append(("pos_embed", "vit.embeddings.position_embeddings") )
rename_keys.append(("patch_embed.proj.weight", "vit.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("patch_embed.proj.bias", "vit.embeddings.patch_embeddings.projection.bias") )
# backbone
rename_keys.append(("patch_embed.backbone.stem.conv.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.convolution.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.bias", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.bias") )
for stage_idx in range(len(config.backbone_config.depths ) ):
for layer_idx in range(config.backbone_config.depths[stage_idx] ):
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.conv.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.conv.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.bias''') )
# transformer encoder
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''vit.encoder.layer.{i}.output.dense.bias''') )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("norm.weight", "layernorm.weight"),
("norm.bias", "layernorm.bias"),
("pre_logits.fc.weight", "pooler.dense.weight"),
("pre_logits.fc.bias", "pooler.dense.bias"),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
__magic_name__ : int = [(pair[0], pair[1][4:]) if pair[1].startswith("vit" ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
("norm.weight", "vit.layernorm.weight"),
("norm.bias", "vit.layernorm.bias"),
("head.weight", "classifier.weight"),
("head.bias", "classifier.bias"),
] )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Any , _snake_case : Dict=False ) -> int:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
__magic_name__ : int = ""
else:
__magic_name__ : Union[str, Any] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__magic_name__ : Optional[Any] = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
__magic_name__ : int = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__magic_name__ : Dict = in_proj_weight[
: config.hidden_size, :
]
__magic_name__ : List[str] = in_proj_bias[: config.hidden_size]
__magic_name__ : List[str] = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__magic_name__ : Optional[Any] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__magic_name__ : Optional[Any] = in_proj_weight[
-config.hidden_size :, :
]
__magic_name__ : int = in_proj_bias[-config.hidden_size :]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
__magic_name__ : List[str] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : int , _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : int = dct.pop(_snake_case )
__magic_name__ : List[Any] = val
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
__magic_name__ : List[str] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Any , _snake_case : int=False ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=_snake_case , )
__magic_name__ : List[str] = ViTHybridConfig(backbone_config=_snake_case , image_size=384 , num_labels=1000 )
__magic_name__ : str = False
# load original model from timm
__magic_name__ : Union[str, Any] = timm.create_model(_snake_case , pretrained=_snake_case )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__magic_name__ : List[Any] = timm_model.state_dict()
if base_model:
remove_classification_head_(_snake_case )
__magic_name__ : Tuple = create_rename_keys(_snake_case , _snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
read_in_q_k_v(_snake_case , _snake_case , _snake_case )
__magic_name__ : List[str] = "huggingface/label-files"
__magic_name__ : int = "imagenet-1k-id2label.json"
__magic_name__ : Optional[int] = json.load(open(hf_hub_download(_snake_case , _snake_case , repo_type="dataset" ) , "r" ) )
__magic_name__ : int = {int(_snake_case ): v for k, v in idalabel.items()}
__magic_name__ : List[str] = idalabel
__magic_name__ : List[str] = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
__magic_name__ : List[str] = ViTHybridModel(_snake_case ).eval()
else:
__magic_name__ : str = ViTHybridForImageClassification(_snake_case ).eval()
model.load_state_dict(_snake_case )
# create image processor
__magic_name__ : List[Any] = create_transform(**resolve_data_config({} , model=_snake_case ) )
__magic_name__ : int = transform.transforms
__magic_name__ : List[str] = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
__magic_name__ : int = ViTHybridImageProcessor(
do_resize=_snake_case , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=_snake_case , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=_snake_case , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
__magic_name__ : List[Any] = prepare_img()
__magic_name__ : Any = transform(_snake_case ).unsqueeze(0 )
__magic_name__ : Tuple = processor(_snake_case , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(_snake_case , _snake_case )
# verify logits
with torch.no_grad():
__magic_name__ : Optional[int] = model(_snake_case )
__magic_name__ : List[str] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
__magic_name__ : List[str] = timm_model.forward_features(_snake_case )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(_snake_case , outputs.pooler_output , atol=1E-3 )
else:
__magic_name__ : Any = timm_model(_snake_case )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_snake_case , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(_snake_case ).mkdir(exist_ok=_snake_case )
print(F'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(_snake_case )
print(F'''Saving processor to {pytorch_dump_folder_path}''' )
processor.save_pretrained(_snake_case )
if push_to_hub:
print(F'''Pushing model and processor to the hub {vit_name}''' )
model.push_to_hub(F'''ybelkada/{vit_name}''' )
processor.push_to_hub(F'''ybelkada/{vit_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_r50_s16_384",
type=str,
help="Name of the hybrid ViT timm model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether to upload the model to the HuggingFace hub."
)
snake_case : List[Any] = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
import warnings
from ...utils import logging
from .image_processing_yolos import YolosImageProcessor
snake_case : Optional[Any] = logging.get_logger(__name__)
class _snake_case ( snake_case ):
def __init__( self , *_a , **_a ):
warnings.warn(
"The class YolosFeatureExtractor is deprecated and will be removed in version 5 of Transformers. Please"
" use YolosImageProcessor instead." , _a , )
super().__init__(*_a , **_a )
| 281 |
# This script creates a super tiny model that is useful inside tests, when we just want to test that
# the machinery works, without needing to the check the quality of the outcomes.
#
# This version creates a tiny model through reduction of a normal pre-trained model, but keeping the
# full vocab, merges file, and thus also resulting in a larger model due to a large vocab size.
# This gives ~3MB in total for all files.
#
# If you want a 50 times smaller than this see `fsmt-make-super-tiny-model.py`, which is slightly more complicated
#
#
# It will be used then as "stas/tiny-wmt19-en-de"
# Build
from transformers import FSMTTokenizer, FSMTConfig, FSMTForConditionalGeneration
snake_case : List[str] = "facebook/wmt19-en-de"
snake_case : Dict = FSMTTokenizer.from_pretrained(mname)
# get the correct vocab sizes, etc. from the master model
snake_case : List[str] = FSMTConfig.from_pretrained(mname)
config.update(
dict(
d_model=4,
encoder_layers=1,
decoder_layers=1,
encoder_ffn_dim=4,
decoder_ffn_dim=4,
encoder_attention_heads=1,
decoder_attention_heads=1,
)
)
snake_case : int = FSMTForConditionalGeneration(config)
print(F"num of params {tiny_model.num_parameters()}")
# Test
snake_case : Optional[Any] = tokenizer(["Making tiny model"], return_tensors="pt")
snake_case : List[str] = tiny_model(**batch)
print("test output:", len(outputs.logits[0]))
# Save
snake_case : Dict = "tiny-wmt19-en-de"
tiny_model.half() # makes it smaller
tiny_model.save_pretrained(mname_tiny)
tokenizer.save_pretrained(mname_tiny)
print(F"Generated {mname_tiny}")
# Upload
# transformers-cli upload tiny-wmt19-en-de
| 281 | 1 |
import gc
import threading
import time
import psutil
import torch
class _snake_case :
def __init__( self ):
__magic_name__ : List[Any] = psutil.Process()
__magic_name__ : Union[str, Any] = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = -1
while True:
__magic_name__ : Tuple = max(self.process.memory_info().rss , self.cpu_memory_peak )
# can't sleep or will not catch the peak right (this comment is here on purpose)
if not self.peak_monitoring:
break
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = True
__magic_name__ : Optional[int] = threading.Thread(target=self.peak_monitor )
__magic_name__ : List[Any] = True
self.thread.start()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = False
self.thread.join()
return self.cpu_memory_peak
snake_case : Union[str, Any] = PeakCPUMemory()
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Tuple = {"time": time.time()}
gc.collect()
torch.cuda.empty_cache()
# CPU mem
__magic_name__ : Optional[int] = psutil.Process().memory_info().rss
cpu_peak_tracker.start()
# GPU mem
for i in range(torch.cuda.device_count() ):
__magic_name__ : Dict = torch.cuda.memory_allocated(_snake_case )
torch.cuda.reset_peak_memory_stats()
return measures
def lowerCAmelCase_ ( _snake_case : str ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : int = {"time": time.time() - start_measures["time"]}
gc.collect()
torch.cuda.empty_cache()
# CPU mem
__magic_name__ : Optional[Any] = (psutil.Process().memory_info().rss - start_measures["cpu"]) / 2**20
__magic_name__ : str = (cpu_peak_tracker.stop() - start_measures["cpu"]) / 2**20
# GPU mem
for i in range(torch.cuda.device_count() ):
__magic_name__ : Optional[int] = (torch.cuda.memory_allocated(_snake_case ) - start_measures[str(_snake_case )]) / 2**20
__magic_name__ : Optional[Any] = (torch.cuda.max_memory_allocated(_snake_case ) - start_measures[str(_snake_case )]) / 2**20
return measures
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Dict ) -> str:
'''simple docstring'''
print(F'''{description}:''' )
print(F'''- Time: {measures["time"]:.2f}s''' )
for i in range(torch.cuda.device_count() ):
print(F'''- GPU {i} allocated: {measures[str(_snake_case )]:.2f}MiB''' )
__magic_name__ : List[Any] = measures[F'''{i}-peak''']
print(F'''- GPU {i} peak: {peak:.2f}MiB''' )
print(F'''- CPU RAM allocated: {measures["cpu"]:.2f}MiB''' )
print(F'''- CPU RAM peak: {measures["cpu-peak"]:.2f}MiB''' )
| 281 |
import argparse
import csv
import logging
import os
import random
import numpy as np
import torch
from torch.utils.data import DataLoader, RandomSampler, SequentialSampler, TensorDataset
from tqdm import tqdm, trange
from transformers import (
CONFIG_NAME,
WEIGHTS_NAME,
AdamW,
OpenAIGPTDoubleHeadsModel,
OpenAIGPTTokenizer,
get_linear_schedule_with_warmup,
)
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s", datefmt="%m/%d/%Y %H:%M:%S", level=logging.INFO
)
snake_case : Optional[int] = logging.getLogger(__name__)
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : List[str] = np.argmax(_snake_case , axis=1 )
return np.sum(outputs == labels )
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
with open(_snake_case , encoding="utf_8" ) as f:
__magic_name__ : List[str] = csv.reader(_snake_case )
__magic_name__ : List[Any] = []
next(_snake_case ) # skip the first line
for line in tqdm(_snake_case ):
output.append((" ".join(line[1:5] ), line[5], line[6], int(line[-1] ) - 1) )
return output
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Tuple , _snake_case : Union[str, Any] , _snake_case : List[Any] , _snake_case : Tuple , _snake_case : Optional[int] ) -> int:
'''simple docstring'''
__magic_name__ : Optional[int] = []
for dataset in encoded_datasets:
__magic_name__ : Union[str, Any] = len(_snake_case )
__magic_name__ : Dict = np.zeros((n_batch, 2, input_len) , dtype=np.intaa )
__magic_name__ : List[str] = np.zeros((n_batch, 2) , dtype=np.intaa )
__magic_name__ : Optional[int] = np.full((n_batch, 2, input_len) , fill_value=-100 , dtype=np.intaa )
__magic_name__ : int = np.zeros((n_batch,) , dtype=np.intaa )
for (
i,
(story, conta, conta, mc_label),
) in enumerate(_snake_case ):
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : str = with_conta
__magic_name__ : Tuple = with_conta
__magic_name__ : Union[str, Any] = len(_snake_case ) - 1
__magic_name__ : int = len(_snake_case ) - 1
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[int] = mc_label
__magic_name__ : str = (input_ids, mc_token_ids, lm_labels, mc_labels)
tensor_datasets.append(tuple(torch.tensor(_snake_case ) for t in all_inputs ) )
return tensor_datasets
def lowerCAmelCase_ ( ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser()
parser.add_argument("--model_name" , type=_snake_case , default="openai-gpt" , help="pretrained model name" )
parser.add_argument("--do_train" , action="store_true" , help="Whether to run training." )
parser.add_argument("--do_eval" , action="store_true" , help="Whether to run eval on the dev set." )
parser.add_argument(
"--output_dir" , default=_snake_case , type=_snake_case , required=_snake_case , help="The output directory where the model predictions and checkpoints will be written." , )
parser.add_argument("--train_dataset" , type=_snake_case , default="" )
parser.add_argument("--eval_dataset" , type=_snake_case , default="" )
parser.add_argument("--seed" , type=_snake_case , default=42 )
parser.add_argument("--num_train_epochs" , type=_snake_case , default=3 )
parser.add_argument("--train_batch_size" , type=_snake_case , default=8 )
parser.add_argument("--eval_batch_size" , type=_snake_case , default=16 )
parser.add_argument("--adam_epsilon" , default=1E-8 , type=_snake_case , help="Epsilon for Adam optimizer." )
parser.add_argument("--max_grad_norm" , type=_snake_case , default=1 )
parser.add_argument(
"--max_steps" , default=-1 , type=_snake_case , help=(
"If > 0: set total number of training steps to perform. Override num_train_epochs."
) , )
parser.add_argument(
"--gradient_accumulation_steps" , type=_snake_case , default=1 , help="Number of updates steps to accumulate before performing a backward/update pass." , )
parser.add_argument("--learning_rate" , type=_snake_case , default=6.25E-5 )
parser.add_argument("--warmup_steps" , default=0 , type=_snake_case , help="Linear warmup over warmup_steps." )
parser.add_argument("--lr_schedule" , type=_snake_case , default="warmup_linear" )
parser.add_argument("--weight_decay" , type=_snake_case , default=0.01 )
parser.add_argument("--lm_coef" , type=_snake_case , default=0.9 )
parser.add_argument("--n_valid" , type=_snake_case , default=374 )
parser.add_argument("--server_ip" , type=_snake_case , default="" , help="Can be used for distant debugging." )
parser.add_argument("--server_port" , type=_snake_case , default="" , help="Can be used for distant debugging." )
__magic_name__ : List[Any] = parser.parse_args()
print(_snake_case )
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach" )
ptvsd.enable_attach(address=(args.server_ip, args.server_port) , redirect_output=_snake_case )
ptvsd.wait_for_attach()
random.seed(args.seed )
np.random.seed(args.seed )
torch.manual_seed(args.seed )
torch.cuda.manual_seed_all(args.seed )
__magic_name__ : Dict = torch.device("cuda" if torch.cuda.is_available() else "cpu" )
__magic_name__ : Optional[int] = torch.cuda.device_count()
logger.info("device: {}, n_gpu {}".format(_snake_case , _snake_case ) )
if not args.do_train and not args.do_eval:
raise ValueError("At least one of `do_train` or `do_eval` must be True." )
if not os.path.exists(args.output_dir ):
os.makedirs(args.output_dir )
# Load tokenizer and model
# This loading functions also add new tokens and embeddings called `special tokens`
# These new embeddings will be fine-tuned on the RocStories dataset
__magic_name__ : List[Any] = ["_start_", "_delimiter_", "_classify_"]
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.model_name )
tokenizer.add_tokens(_snake_case )
__magic_name__ : Optional[Any] = tokenizer.convert_tokens_to_ids(_snake_case )
__magic_name__ : List[str] = OpenAIGPTDoubleHeadsModel.from_pretrained(args.model_name )
model.resize_token_embeddings(len(_snake_case ) )
model.to(_snake_case )
# Load and encode the datasets
def tokenize_and_encode(_snake_case : str ):
if isinstance(_snake_case , _snake_case ):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(_snake_case ) )
elif isinstance(_snake_case , _snake_case ):
return obj
return [tokenize_and_encode(_snake_case ) for o in obj]
logger.info("Encoding dataset..." )
__magic_name__ : Optional[int] = load_rocstories_dataset(args.train_dataset )
__magic_name__ : str = load_rocstories_dataset(args.eval_dataset )
__magic_name__ : int = (train_dataset, eval_dataset)
__magic_name__ : List[str] = tokenize_and_encode(_snake_case )
# Compute the max input length for the Transformer
__magic_name__ : Optional[Any] = model.config.n_positions // 2 - 2
__magic_name__ : Optional[int] = max(
len(story[:max_length] ) + max(len(conta[:max_length] ) , len(conta[:max_length] ) ) + 3
for dataset in encoded_datasets
for story, conta, conta, _ in dataset )
__magic_name__ : List[str] = min(_snake_case , model.config.n_positions ) # Max size of input for the pre-trained model
# Prepare inputs tensors and dataloaders
__magic_name__ : List[Any] = pre_process_datasets(_snake_case , _snake_case , _snake_case , *_snake_case )
__magic_name__ , __magic_name__ : Optional[int] = tensor_datasets[0], tensor_datasets[1]
__magic_name__ : Tuple = TensorDataset(*_snake_case )
__magic_name__ : Union[str, Any] = RandomSampler(_snake_case )
__magic_name__ : Dict = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.train_batch_size )
__magic_name__ : Any = TensorDataset(*_snake_case )
__magic_name__ : Optional[Any] = SequentialSampler(_snake_case )
__magic_name__ : int = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.eval_batch_size )
# Prepare optimizer
if args.do_train:
if args.max_steps > 0:
__magic_name__ : Tuple = args.max_steps
__magic_name__ : List[str] = args.max_steps // (len(_snake_case ) // args.gradient_accumulation_steps) + 1
else:
__magic_name__ : List[str] = len(_snake_case ) // args.gradient_accumulation_steps * args.num_train_epochs
__magic_name__ : str = list(model.named_parameters() )
__magic_name__ : Dict = ["bias", "LayerNorm.bias", "LayerNorm.weight"]
__magic_name__ : str = [
{
"params": [p for n, p in param_optimizer if not any(nd in n for nd in no_decay )],
"weight_decay": args.weight_decay,
},
{"params": [p for n, p in param_optimizer if any(nd in n for nd in no_decay )], "weight_decay": 0.0},
]
__magic_name__ : str = AdamW(_snake_case , lr=args.learning_rate , eps=args.adam_epsilon )
__magic_name__ : List[str] = get_linear_schedule_with_warmup(
_snake_case , num_warmup_steps=args.warmup_steps , num_training_steps=_snake_case )
if args.do_train:
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = 0, 0, None
model.train()
for _ in trange(int(args.num_train_epochs ) , desc="Epoch" ):
__magic_name__ : List[str] = 0
__magic_name__ : Tuple = 0
__magic_name__ : Dict = tqdm(_snake_case , desc="Training" )
for step, batch in enumerate(_snake_case ):
__magic_name__ : Optional[Any] = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = batch
__magic_name__ : Optional[Any] = model(_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Optional[Any] = args.lm_coef * losses[0] + losses[1]
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
tr_loss += loss.item()
__magic_name__ : List[str] = (
loss.item() if exp_average_loss is None else 0.7 * exp_average_loss + 0.3 * loss.item()
)
nb_tr_steps += 1
__magic_name__ : int = "Training loss: {:.2e} lr: {:.2e}".format(_snake_case , scheduler.get_lr()[0] )
# Save a trained model
if args.do_train:
# Save a trained model, configuration and tokenizer
__magic_name__ : Dict = model.module if hasattr(_snake_case , "module" ) else model # Only save the model itself
# If we save using the predefined names, we can load using `from_pretrained`
__magic_name__ : List[Any] = os.path.join(args.output_dir , _snake_case )
__magic_name__ : Dict = os.path.join(args.output_dir , _snake_case )
torch.save(model_to_save.state_dict() , _snake_case )
model_to_save.config.to_json_file(_snake_case )
tokenizer.save_vocabulary(args.output_dir )
# Load a trained model and vocabulary that you have fine-tuned
__magic_name__ : Dict = OpenAIGPTDoubleHeadsModel.from_pretrained(args.output_dir )
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.output_dir )
model.to(_snake_case )
if args.do_eval:
model.eval()
__magic_name__ , __magic_name__ : Any = 0, 0
__magic_name__ , __magic_name__ : Union[str, Any] = 0, 0
for batch in tqdm(_snake_case , desc="Evaluating" ):
__magic_name__ : int = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = batch
with torch.no_grad():
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = model(
_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Tuple = mc_logits.detach().cpu().numpy()
__magic_name__ : Any = mc_labels.to("cpu" ).numpy()
__magic_name__ : str = accuracy(_snake_case , _snake_case )
eval_loss += mc_loss.mean().item()
eval_accuracy += tmp_eval_accuracy
nb_eval_examples += input_ids.size(0 )
nb_eval_steps += 1
__magic_name__ : Tuple = eval_loss / nb_eval_steps
__magic_name__ : List[Any] = eval_accuracy / nb_eval_examples
__magic_name__ : int = tr_loss / nb_tr_steps if args.do_train else None
__magic_name__ : Any = {"eval_loss": eval_loss, "eval_accuracy": eval_accuracy, "train_loss": train_loss}
__magic_name__ : int = os.path.join(args.output_dir , "eval_results.txt" )
with open(_snake_case , "w" ) as writer:
logger.info("***** Eval results *****" )
for key in sorted(result.keys() ):
logger.info(" %s = %s" , _snake_case , str(result[key] ) )
writer.write("%s = %s\n" % (key, str(result[key] )) )
if __name__ == "__main__":
main()
| 281 | 1 |
import os
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import torch
from torch import nn
from ...models.controlnet import ControlNetModel, ControlNetOutput
from ...models.modeling_utils import ModelMixin
from ...utils import logging
snake_case : Tuple = logging.get_logger(__name__)
class _snake_case ( snake_case ):
def __init__( self , _a ):
super().__init__()
__magic_name__ : str = nn.ModuleList(_a )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a = None , _a = None , _a = None , _a = None , _a = False , _a = True , ):
for i, (image, scale, controlnet) in enumerate(zip(_a , _a , self.nets ) ):
__magic_name__ , __magic_name__ : Optional[Any] = controlnet(
_a , _a , _a , _a , _a , _a , _a , _a , _a , _a , _a , )
# merge samples
if i == 0:
__magic_name__ , __magic_name__ : Union[str, Any] = down_samples, mid_sample
else:
__magic_name__ : Dict = [
samples_prev + samples_curr
for samples_prev, samples_curr in zip(_a , _a )
]
mid_block_res_sample += mid_sample
return down_block_res_samples, mid_block_res_sample
def SCREAMING_SNAKE_CASE ( self , _a , _a = True , _a = None , _a = False , _a = None , ):
__magic_name__ : str = 0
__magic_name__ : Tuple = save_directory
for controlnet in self.nets:
controlnet.save_pretrained(
_a , is_main_process=_a , save_function=_a , safe_serialization=_a , variant=_a , )
idx += 1
__magic_name__ : List[str] = model_path_to_save + f'''_{idx}'''
@classmethod
def SCREAMING_SNAKE_CASE ( cls , _a , **_a ):
__magic_name__ : int = 0
__magic_name__ : Union[str, Any] = []
# load controlnet and append to list until no controlnet directory exists anymore
# first controlnet has to be saved under `./mydirectory/controlnet` to be compliant with `DiffusionPipeline.from_prertained`
# second, third, ... controlnets have to be saved under `./mydirectory/controlnet_1`, `./mydirectory/controlnet_2`, ...
__magic_name__ : Dict = pretrained_model_path
while os.path.isdir(_a ):
__magic_name__ : Tuple = ControlNetModel.from_pretrained(_a , **_a )
controlnets.append(_a )
idx += 1
__magic_name__ : Dict = pretrained_model_path + f'''_{idx}'''
logger.info(f'''{len(_a )} controlnets loaded from {pretrained_model_path}.''' )
if len(_a ) == 0:
raise ValueError(
f'''No ControlNets found under {os.path.dirname(_a )}. Expected at least {pretrained_model_path + "_0"}.''' )
return cls(_a )
| 281 |
from . import __version__
# Backward compatibility imports, to make sure all those objects can be found in file_utils
from .utils import (
CLOUDFRONT_DISTRIB_PREFIX,
CONFIG_NAME,
DISABLE_TELEMETRY,
DUMMY_INPUTS,
DUMMY_MASK,
ENV_VARS_TRUE_AND_AUTO_VALUES,
ENV_VARS_TRUE_VALUES,
FEATURE_EXTRACTOR_NAME,
FLAX_WEIGHTS_NAME,
HF_MODULES_CACHE,
HUGGINGFACE_CO_PREFIX,
HUGGINGFACE_CO_RESOLVE_ENDPOINT,
MODEL_CARD_NAME,
MULTIPLE_CHOICE_DUMMY_INPUTS,
PYTORCH_PRETRAINED_BERT_CACHE,
PYTORCH_TRANSFORMERS_CACHE,
S3_BUCKET_PREFIX,
SENTENCEPIECE_UNDERLINE,
SPIECE_UNDERLINE,
TF2_WEIGHTS_NAME,
TF_WEIGHTS_NAME,
TORCH_FX_REQUIRED_VERSION,
TRANSFORMERS_CACHE,
TRANSFORMERS_DYNAMIC_MODULE_NAME,
USE_JAX,
USE_TF,
USE_TORCH,
WEIGHTS_INDEX_NAME,
WEIGHTS_NAME,
ContextManagers,
DummyObject,
EntryNotFoundError,
ExplicitEnum,
ModelOutput,
PaddingStrategy,
PushToHubMixin,
RepositoryNotFoundError,
RevisionNotFoundError,
TensorType,
_LazyModule,
add_code_sample_docstrings,
add_end_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
cached_property,
copy_func,
default_cache_path,
define_sagemaker_information,
get_cached_models,
get_file_from_repo,
get_full_repo_name,
get_torch_version,
has_file,
http_user_agent,
is_apex_available,
is_bsa_available,
is_coloredlogs_available,
is_datasets_available,
is_detectrona_available,
is_faiss_available,
is_flax_available,
is_ftfy_available,
is_in_notebook,
is_ipex_available,
is_librosa_available,
is_offline_mode,
is_onnx_available,
is_pandas_available,
is_phonemizer_available,
is_protobuf_available,
is_psutil_available,
is_pyanvml_available,
is_pyctcdecode_available,
is_pytesseract_available,
is_pytorch_quantization_available,
is_rjieba_available,
is_sagemaker_dp_enabled,
is_sagemaker_mp_enabled,
is_scipy_available,
is_sentencepiece_available,
is_seqio_available,
is_sklearn_available,
is_soundfile_availble,
is_spacy_available,
is_speech_available,
is_tensor,
is_tensorflow_probability_available,
is_tfaonnx_available,
is_tf_available,
is_timm_available,
is_tokenizers_available,
is_torch_available,
is_torch_bfaa_available,
is_torch_cuda_available,
is_torch_fx_available,
is_torch_fx_proxy,
is_torch_mps_available,
is_torch_tfaa_available,
is_torch_tpu_available,
is_torchaudio_available,
is_training_run_on_sagemaker,
is_vision_available,
replace_return_docstrings,
requires_backends,
to_numpy,
to_py_obj,
torch_only_method,
)
| 281 | 1 |
import os
from collections.abc import Iterator
def lowerCAmelCase_ ( _snake_case : str = "." ) -> Iterator[str]:
'''simple docstring'''
for dir_path, dir_names, filenames in os.walk(_snake_case ):
__magic_name__ : Tuple = [d for d in dir_names if d != "scripts" and d[0] not in "._"]
for filename in filenames:
if filename == "__init__.py":
continue
if os.path.splitext(_snake_case )[1] in (".py", ".ipynb"):
yield os.path.join(_snake_case , _snake_case ).lstrip("./" )
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> List[str]:
'''simple docstring'''
return F'''{i * " "}*''' if i else "\n##"
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = old_path.split(os.sep )
for i, new_part in enumerate(new_path.split(os.sep ) ):
if (i + 1 > len(_snake_case ) or old_parts[i] != new_part) and new_part:
print(F'''{md_prefix(_snake_case )} {new_part.replace("_" , " " ).title()}''' )
return new_path
def lowerCAmelCase_ ( _snake_case : str = "." ) -> None:
'''simple docstring'''
__magic_name__ : Optional[int] = ""
for filepath in sorted(good_file_paths(_snake_case ) ):
__magic_name__ , __magic_name__ : int = os.path.split(_snake_case )
if filepath != old_path:
__magic_name__ : Dict = print_path(_snake_case , _snake_case )
__magic_name__ : str = (filepath.count(os.sep ) + 1) if filepath else 0
__magic_name__ : List[str] = F'''{filepath}/{filename}'''.replace(" " , "%20" )
__magic_name__ : Dict = os.path.splitext(filename.replace("_" , " " ).title() )[0]
print(F'''{md_prefix(_snake_case )} [{filename}]({url})''' )
if __name__ == "__main__":
print_directory_md(".")
| 281 |
import importlib
import os
import fsspec
import pytest
from fsspec import register_implementation
from fsspec.registry import _registry as _fsspec_registry
from datasets.filesystems import COMPRESSION_FILESYSTEMS, HfFileSystem, extract_path_from_uri, is_remote_filesystem
from .utils import require_lza, require_zstandard
def lowerCAmelCase_ ( _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
assert "mock" in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
assert "mock" not in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Dict = "mock-s3-bucket"
__magic_name__ : Any = F'''s3://{mock_bucket}'''
__magic_name__ : str = extract_path_from_uri(_snake_case )
assert dataset_path.startswith("s3://" ) is False
__magic_name__ : Tuple = "./local/path"
__magic_name__ : Optional[Any] = extract_path_from_uri(_snake_case )
assert dataset_path == new_dataset_path
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : str = is_remote_filesystem(_snake_case )
assert is_remote is True
__magic_name__ : Optional[int] = fsspec.filesystem("file" )
__magic_name__ : int = is_remote_filesystem(_snake_case )
assert is_remote is False
@pytest.mark.parametrize("compression_fs_class" , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : int , _snake_case : Tuple , _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Any ) -> int:
'''simple docstring'''
__magic_name__ : Any = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_file, "bz2": bza_file, "lz4": lza_file}
__magic_name__ : str = input_paths[compression_fs_class.protocol]
if input_path is None:
__magic_name__ : Dict = F'''for \'{compression_fs_class.protocol}\' compression protocol, '''
if compression_fs_class.protocol == "lz4":
reason += require_lza.kwargs["reason"]
elif compression_fs_class.protocol == "zstd":
reason += require_zstandard.kwargs["reason"]
pytest.skip(_snake_case )
__magic_name__ : str = fsspec.filesystem(compression_fs_class.protocol , fo=_snake_case )
assert isinstance(_snake_case , _snake_case )
__magic_name__ : int = os.path.basename(_snake_case )
__magic_name__ : Optional[int] = expected_filename[: expected_filename.rindex("." )]
assert fs.glob("*" ) == [expected_filename]
with fs.open(_snake_case , "r" , encoding="utf-8" ) as f, open(_snake_case , encoding="utf-8" ) as expected_file:
assert f.read() == expected_file.read()
@pytest.mark.parametrize("protocol" , ["zip", "gzip"] )
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : Optional[Any] , _snake_case : Optional[Any] ) -> str:
'''simple docstring'''
__magic_name__ : int = {"zip": zip_jsonl_path, "gzip": jsonl_gz_path}
__magic_name__ : int = compressed_file_paths[protocol]
__magic_name__ : Tuple = "dataset.jsonl"
__magic_name__ : List[str] = F'''{protocol}://{member_file_path}::{compressed_file_path}'''
__magic_name__ , *__magic_name__ : Optional[Any] = fsspec.get_fs_token_paths(_snake_case )
assert fs.isfile(_snake_case )
assert not fs.isfile("non_existing_" + member_file_path )
@pytest.mark.integration
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : List[str] , _snake_case : Tuple ) -> str:
'''simple docstring'''
__magic_name__ : int = hf_api.dataset_info(_snake_case , token=_snake_case )
__magic_name__ : Optional[Any] = HfFileSystem(repo_info=_snake_case , token=_snake_case )
assert sorted(hffs.glob("*" ) ) == [".gitattributes", "data"]
assert hffs.isdir("data" )
assert hffs.isfile(".gitattributes" ) and hffs.isfile("data/text_data.txt" )
with open(_snake_case ) as f:
assert hffs.open("data/text_data.txt" , "r" ).read() == f.read()
def lowerCAmelCase_ ( ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Optional[Any] = "bz2"
# Import module
import datasets.filesystems
# Overwrite protocol and reload
register_implementation(_snake_case , _snake_case , clobber=_snake_case )
with pytest.warns(_snake_case ) as warning_info:
importlib.reload(datasets.filesystems )
assert len(_snake_case ) == 1
assert (
str(warning_info[0].message )
== F'''A filesystem protocol was already set for {protocol} and will be overwritten.'''
)
| 281 | 1 |
snake_case : Any = frozenset(
[
"prompt",
"height",
"width",
"guidance_scale",
"negative_prompt",
"prompt_embeds",
"negative_prompt_embeds",
"cross_attention_kwargs",
]
)
snake_case : Tuple = frozenset(["prompt", "negative_prompt"])
snake_case : int = frozenset([])
snake_case : int = frozenset(["image"])
snake_case : int = frozenset(
[
"image",
"height",
"width",
"guidance_scale",
]
)
snake_case : Union[str, Any] = frozenset(["image"])
snake_case : Optional[Any] = frozenset(
[
"prompt",
"image",
"height",
"width",
"guidance_scale",
"negative_prompt",
"prompt_embeds",
"negative_prompt_embeds",
]
)
snake_case : Optional[Any] = frozenset(["prompt", "image", "negative_prompt"])
snake_case : Optional[int] = frozenset(
[
# Text guided image variation with an image mask
"prompt",
"image",
"mask_image",
"height",
"width",
"guidance_scale",
"negative_prompt",
"prompt_embeds",
"negative_prompt_embeds",
]
)
snake_case : List[str] = frozenset(["prompt", "image", "mask_image", "negative_prompt"])
snake_case : str = frozenset(
[
# image variation with an image mask
"image",
"mask_image",
"height",
"width",
"guidance_scale",
]
)
snake_case : Tuple = frozenset(["image", "mask_image"])
snake_case : Tuple = frozenset(
[
"example_image",
"image",
"mask_image",
"height",
"width",
"guidance_scale",
]
)
snake_case : Any = frozenset(["example_image", "image", "mask_image"])
snake_case : Optional[int] = frozenset(["class_labels"])
snake_case : Dict = frozenset(["class_labels"])
snake_case : Tuple = frozenset(["batch_size"])
snake_case : Dict = frozenset([])
snake_case : List[Any] = frozenset(["batch_size"])
snake_case : Union[str, Any] = frozenset([])
snake_case : int = frozenset(
[
"prompt",
"audio_length_in_s",
"guidance_scale",
"negative_prompt",
"prompt_embeds",
"negative_prompt_embeds",
"cross_attention_kwargs",
]
)
snake_case : Optional[Any] = frozenset(["prompt", "negative_prompt"])
snake_case : str = frozenset(["input_tokens"])
snake_case : List[Any] = frozenset(["input_tokens"])
| 281 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : List[Any] = {
"YituTech/conv-bert-base": "https://huggingface.co/YituTech/conv-bert-base/resolve/main/config.json",
"YituTech/conv-bert-medium-small": (
"https://huggingface.co/YituTech/conv-bert-medium-small/resolve/main/config.json"
),
"YituTech/conv-bert-small": "https://huggingface.co/YituTech/conv-bert-small/resolve/main/config.json",
# See all ConvBERT models at https://huggingface.co/models?filter=convbert
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'convbert'
def __init__( self , _a=30_522 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=1 , _a=0 , _a=2 , _a=768 , _a=2 , _a=9 , _a=1 , _a=None , **_a , ):
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a , )
__magic_name__ : Tuple = vocab_size
__magic_name__ : List[Any] = hidden_size
__magic_name__ : Union[str, Any] = num_hidden_layers
__magic_name__ : List[Any] = num_attention_heads
__magic_name__ : str = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : Tuple = max_position_embeddings
__magic_name__ : str = type_vocab_size
__magic_name__ : List[str] = initializer_range
__magic_name__ : Tuple = layer_norm_eps
__magic_name__ : List[Any] = embedding_size
__magic_name__ : List[Any] = head_ratio
__magic_name__ : str = conv_kernel_size
__magic_name__ : Dict = num_groups
__magic_name__ : str = classifier_dropout
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : Dict = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
("token_type_ids", dynamic_axis),
] )
| 281 | 1 |
import numpy as np
from transformers import Pipeline
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : str = np.max(_snake_case , axis=-1 , keepdims=_snake_case )
__magic_name__ : List[Any] = np.exp(outputs - maxes )
return shifted_exp / shifted_exp.sum(axis=-1 , keepdims=_snake_case )
class _snake_case ( snake_case ):
def SCREAMING_SNAKE_CASE ( self , **_a ):
__magic_name__ : Optional[int] = {}
if "second_text" in kwargs:
__magic_name__ : Any = kwargs["second_text"]
return preprocess_kwargs, {}, {}
def SCREAMING_SNAKE_CASE ( self , _a , _a=None ):
return self.tokenizer(_a , text_pair=_a , return_tensors=self.framework )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.model(**_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : str = model_outputs.logits[0].numpy()
__magic_name__ : Any = softmax(_a )
__magic_name__ : List[Any] = np.argmax(_a )
__magic_name__ : List[Any] = self.model.config.idalabel[best_class]
__magic_name__ : str = probabilities[best_class].item()
__magic_name__ : Tuple = logits.tolist()
return {"label": label, "score": score, "logits": logits}
| 281 |
import argparse
import requests
import torch
# pip3 install salesforce-lavis
# I'm actually installing a slightly modified version: pip3 install git+https://github.com/nielsrogge/LAVIS.git@fix_lavis
from lavis.models import load_model_and_preprocess
from PIL import Image
from transformers import (
AutoTokenizer,
BlipaConfig,
BlipaForConditionalGeneration,
BlipaProcessor,
BlipaVisionConfig,
BlipImageProcessor,
OPTConfig,
TaConfig,
)
from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD
def lowerCAmelCase_ ( ) -> str:
'''simple docstring'''
__magic_name__ : int = "https://storage.googleapis.com/sfr-vision-language-research/LAVIS/assets/merlion.png"
__magic_name__ : Union[str, Any] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw ).convert("RGB" )
return image
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = []
# fmt: off
# vision encoder
rename_keys.append(("visual_encoder.cls_token", "vision_model.embeddings.class_embedding") )
rename_keys.append(("visual_encoder.pos_embed", "vision_model.embeddings.position_embedding") )
rename_keys.append(("visual_encoder.patch_embed.proj.weight", "vision_model.embeddings.patch_embedding.weight") )
rename_keys.append(("visual_encoder.patch_embed.proj.bias", "vision_model.embeddings.patch_embedding.bias") )
rename_keys.append(("ln_vision.weight", "vision_model.post_layernorm.weight") )
rename_keys.append(("ln_vision.bias", "vision_model.post_layernorm.bias") )
for i in range(config.vision_config.num_hidden_layers ):
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.weight''', F'''vision_model.encoder.layers.{i}.layer_norm1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.bias''', F'''vision_model.encoder.layers.{i}.layer_norm1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.weight''', F'''vision_model.encoder.layers.{i}.layer_norm2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.bias''', F'''vision_model.encoder.layers.{i}.layer_norm2.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.qkv.weight''', F'''vision_model.encoder.layers.{i}.self_attn.qkv.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.weight''', F'''vision_model.encoder.layers.{i}.self_attn.projection.weight''',) )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.bias''', F'''vision_model.encoder.layers.{i}.self_attn.projection.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc2.bias''') )
# QFormer
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.weight", "qformer.layernorm.weight") )
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.bias", "qformer.layernorm.bias") )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Optional[Any] ) -> int:
'''simple docstring'''
__magic_name__ : Tuple = dct.pop(_snake_case )
__magic_name__ : int = val
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
for i in range(config.vision_config.num_hidden_layers ):
# read in original q and v biases
__magic_name__ : List[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.q_bias''' )
__magic_name__ : Optional[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.v_bias''' )
# next, set bias in the state dict
__magic_name__ : Optional[int] = torch.cat((q_bias, torch.zeros_like(_snake_case , requires_grad=_snake_case ), v_bias) )
__magic_name__ : Union[str, Any] = qkv_bias
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : str ) -> int:
'''simple docstring'''
__magic_name__ : List[Any] = 364 if "coco" in model_name else 224
__magic_name__ : Union[str, Any] = BlipaVisionConfig(image_size=_snake_case ).to_dict()
# make sure the models have proper bos_token_id and eos_token_id set (important for generation)
# seems like flan-T5 models don't have bos_token_id properly set?
if "opt-2.7b" in model_name:
__magic_name__ : List[str] = OPTConfig.from_pretrained("facebook/opt-2.7b" , eos_token_id=_snake_case ).to_dict()
elif "opt-6.7b" in model_name:
__magic_name__ : Any = OPTConfig.from_pretrained("facebook/opt-6.7b" , eos_token_id=_snake_case ).to_dict()
elif "t5-xl" in model_name:
__magic_name__ : Dict = TaConfig.from_pretrained("google/flan-t5-xl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
elif "t5-xxl" in model_name:
__magic_name__ : int = TaConfig.from_pretrained("google/flan-t5-xxl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
__magic_name__ : List[Any] = BlipaConfig(vision_config=_snake_case , text_config=_snake_case )
return config, image_size
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : str=None , _snake_case : Dict=False ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Optional[int] = (
AutoTokenizer.from_pretrained("facebook/opt-2.7b" )
if "opt" in model_name
else AutoTokenizer.from_pretrained("google/flan-t5-xl" )
)
__magic_name__ : List[Any] = tokenizer("\n" , add_special_tokens=_snake_case ).input_ids[0]
__magic_name__ , __magic_name__ : Tuple = get_blipa_config(_snake_case , eos_token_id=_snake_case )
__magic_name__ : Union[str, Any] = BlipaForConditionalGeneration(_snake_case ).eval()
__magic_name__ : Any = {
"blip2-opt-2.7b": ("blip2_opt", "pretrain_opt2.7b"),
"blip2-opt-6.7b": ("blip2_opt", "pretrain_opt6.7b"),
"blip2-opt-2.7b-coco": ("blip2_opt", "caption_coco_opt2.7b"),
"blip2-opt-6.7b-coco": ("blip2_opt", "caption_coco_opt6.7b"),
"blip2-flan-t5-xl": ("blip2_t5", "pretrain_flant5xl"),
"blip2-flan-t5-xl-coco": ("blip2_t5", "caption_coco_flant5xl"),
"blip2-flan-t5-xxl": ("blip2_t5", "pretrain_flant5xxl"),
}
__magic_name__ , __magic_name__ : Union[str, Any] = model_name_to_original[model_name]
# load original model
print("Loading original model..." )
__magic_name__ : Union[str, Any] = "cuda" if torch.cuda.is_available() else "cpu"
__magic_name__ , __magic_name__ , __magic_name__ : Optional[Any] = load_model_and_preprocess(
name=_snake_case , model_type=_snake_case , is_eval=_snake_case , device=_snake_case )
original_model.eval()
print("Done!" )
# update state dict keys
__magic_name__ : Dict = original_model.state_dict()
__magic_name__ : str = create_rename_keys(_snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
# some keys can be renamed efficiently
for key, val in state_dict.copy().items():
__magic_name__ : Any = state_dict.pop(_snake_case )
if key.startswith("Qformer.bert" ):
__magic_name__ : Optional[int] = key.replace("Qformer.bert" , "qformer" )
if "attention.self" in key:
__magic_name__ : Any = key.replace("self" , "attention" )
if "opt_proj" in key:
__magic_name__ : Union[str, Any] = key.replace("opt_proj" , "language_projection" )
if "t5_proj" in key:
__magic_name__ : Optional[int] = key.replace("t5_proj" , "language_projection" )
if key.startswith("opt" ):
__magic_name__ : List[str] = key.replace("opt" , "language" )
if key.startswith("t5" ):
__magic_name__ : Tuple = key.replace("t5" , "language" )
__magic_name__ : Dict = val
# read in qv biases
read_in_q_v_bias(_snake_case , _snake_case )
__magic_name__ , __magic_name__ : Tuple = hf_model.load_state_dict(_snake_case , strict=_snake_case )
assert len(_snake_case ) == 0
assert unexpected_keys == ["qformer.embeddings.position_ids"]
__magic_name__ : List[Any] = load_demo_image()
__magic_name__ : Tuple = vis_processors["eval"](_snake_case ).unsqueeze(0 ).to(_snake_case )
__magic_name__ : Dict = tokenizer(["\n"] , return_tensors="pt" ).input_ids.to(_snake_case )
# create processor
__magic_name__ : Optional[Any] = BlipImageProcessor(
size={"height": image_size, "width": image_size} , image_mean=_snake_case , image_std=_snake_case )
__magic_name__ : Dict = BlipaProcessor(image_processor=_snake_case , tokenizer=_snake_case )
__magic_name__ : Union[str, Any] = processor(images=_snake_case , return_tensors="pt" ).pixel_values.to(_snake_case )
# make sure processor creates exact same pixel values
assert torch.allclose(_snake_case , _snake_case )
original_model.to(_snake_case )
hf_model.to(_snake_case )
with torch.no_grad():
if "opt" in model_name:
__magic_name__ : List[Any] = original_model({"image": original_pixel_values, "text_input": [""]} ).logits
__magic_name__ : Optional[int] = hf_model(_snake_case , _snake_case ).logits
else:
__magic_name__ : int = original_model(
{"image": original_pixel_values, "text_input": ["\n"], "text_output": ["\n"]} ).logits
__magic_name__ : Tuple = input_ids.masked_fill(input_ids == tokenizer.pad_token_id , -100 )
__magic_name__ : List[str] = hf_model(_snake_case , _snake_case , labels=_snake_case ).logits
assert original_logits.shape == logits.shape
print("First values of original logits:" , original_logits[0, :3, :3] )
print("First values of HF logits:" , logits[0, :3, :3] )
# assert values
if model_name == "blip2-flan-t5-xl":
__magic_name__ : List[str] = torch.tensor(
[[-41.5_850, -4.4_440, -8.9_922], [-47.4_322, -5.9_143, -1.7_340]] , device=_snake_case )
assert torch.allclose(logits[0, :3, :3] , _snake_case , atol=1E-4 )
elif model_name == "blip2-flan-t5-xl-coco":
__magic_name__ : Tuple = torch.tensor(
[[-57.0_109, -9.8_967, -12.6_280], [-68.6_578, -12.7_191, -10.5_065]] , device=_snake_case )
else:
# cast to same type
__magic_name__ : str = logits.dtype
assert torch.allclose(original_logits.to(_snake_case ) , _snake_case , atol=1E-2 )
print("Looks ok!" )
print("Generating a caption..." )
__magic_name__ : Optional[int] = ""
__magic_name__ : Dict = tokenizer(_snake_case , return_tensors="pt" ).input_ids.to(_snake_case )
__magic_name__ : int = original_model.generate({"image": original_pixel_values} )
__magic_name__ : Optional[Any] = hf_model.generate(
_snake_case , _snake_case , do_sample=_snake_case , num_beams=5 , max_length=30 , min_length=1 , top_p=0.9 , repetition_penalty=1.0 , length_penalty=1.0 , temperature=1 , )
print("Original generation:" , _snake_case )
__magic_name__ : Tuple = input_ids.shape[1]
__magic_name__ : int = processor.batch_decode(outputs[:, prompt_length:] , skip_special_tokens=_snake_case )
__magic_name__ : Union[str, Any] = [text.strip() for text in output_text]
print("HF generation:" , _snake_case )
if pytorch_dump_folder_path is not None:
processor.save_pretrained(_snake_case )
hf_model.save_pretrained(_snake_case )
if push_to_hub:
processor.push_to_hub(F'''nielsr/{model_name}''' )
hf_model.push_to_hub(F'''nielsr/{model_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
snake_case : Union[str, Any] = [
"blip2-opt-2.7b",
"blip2-opt-6.7b",
"blip2-opt-2.7b-coco",
"blip2-opt-6.7b-coco",
"blip2-flan-t5-xl",
"blip2-flan-t5-xl-coco",
"blip2-flan-t5-xxl",
]
parser.add_argument(
"--model_name",
default="blip2-opt-2.7b",
choices=choices,
type=str,
help="Path to hf config.json of model to convert",
)
parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.")
parser.add_argument(
"--push_to_hub",
action="store_true",
help="Whether to push the model and processor to the hub after converting",
)
snake_case : int = parser.parse_args()
convert_blipa_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> str:
'''simple docstring'''
if not (isinstance(_snake_case , _snake_case ) and isinstance(_snake_case , _snake_case )):
raise ValueError("longest_common_substring() takes two strings for inputs" )
__magic_name__ : List[str] = len(_snake_case )
__magic_name__ : str = len(_snake_case )
__magic_name__ : List[str] = [[0] * (texta_length + 1) for _ in range(texta_length + 1 )]
__magic_name__ : int = 0
__magic_name__ : Dict = 0
for i in range(1 , texta_length + 1 ):
for j in range(1 , texta_length + 1 ):
if texta[i - 1] == texta[j - 1]:
__magic_name__ : List[str] = 1 + dp[i - 1][j - 1]
if dp[i][j] > ans_length:
__magic_name__ : Union[str, Any] = i
__magic_name__ : Any = dp[i][j]
return texta[ans_index - ans_length : ans_index]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
import os
import re
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : Union[str, Any] = {
"vocab_file": "vocab.txt",
"merges_file": "bpe.codes",
}
snake_case : Dict = {
"vocab_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/vocab.txt",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/vocab.txt",
},
"merges_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/bpe.codes",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/bpe.codes",
},
}
snake_case : Union[str, Any] = {
"vinai/phobert-base": 256,
"vinai/phobert-large": 256,
}
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = set()
__magic_name__ : Any = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
__magic_name__ : int = char
__magic_name__ : List[str] = set(_snake_case )
return pairs
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a , _a="<s>" , _a="</s>" , _a="</s>" , _a="<s>" , _a="<unk>" , _a="<pad>" , _a="<mask>" , **_a , ):
super().__init__(
bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , cls_token=_a , pad_token=_a , mask_token=_a , **_a , )
__magic_name__ : Dict = vocab_file
__magic_name__ : Tuple = merges_file
__magic_name__ : List[Any] = {}
__magic_name__ : List[Any] = 0
__magic_name__ : Tuple = 1
__magic_name__ : int = 2
__magic_name__ : Union[str, Any] = 3
self.add_from_file(_a )
__magic_name__ : Optional[int] = {v: k for k, v in self.encoder.items()}
with open(_a , encoding="utf-8" ) as merges_handle:
__magic_name__ : List[str] = merges_handle.read().split("\n" )[:-1]
__magic_name__ : Union[str, Any] = [tuple(merge.split()[:-1] ) for merge in merges]
__magic_name__ : Union[str, Any] = dict(zip(_a , range(len(_a ) ) ) )
__magic_name__ : Optional[int] = {}
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
__magic_name__ : Optional[Any] = [self.cls_token_id]
__magic_name__ : Union[str, Any] = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is None:
return [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[Any] = [self.sep_token_id]
__magic_name__ : 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 + sep + token_ids_a + sep ) * [0]
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.encoder )
def SCREAMING_SNAKE_CASE ( self ):
return dict(self.encoder , **self.added_tokens_encoder )
def SCREAMING_SNAKE_CASE ( self , _a ):
if token in self.cache:
return self.cache[token]
__magic_name__ : List[Any] = tuple(_a )
__magic_name__ : List[Any] = tuple(list(word[:-1] ) + [word[-1] + "</w>"] )
__magic_name__ : Any = get_pairs(_a )
if not pairs:
return token
while True:
__magic_name__ : str = min(_a , key=lambda _a : self.bpe_ranks.get(_a , float("inf" ) ) )
if bigram not in self.bpe_ranks:
break
__magic_name__ , __magic_name__ : List[str] = bigram
__magic_name__ : List[str] = []
__magic_name__ : List[str] = 0
while i < len(_a ):
try:
__magic_name__ : Any = word.index(_a , _a )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
__magic_name__ : Tuple = j
if word[i] == first and i < len(_a ) - 1 and word[i + 1] == second:
new_word.append(first + second )
i += 2
else:
new_word.append(word[i] )
i += 1
__magic_name__ : Union[str, Any] = tuple(_a )
__magic_name__ : Optional[int] = new_word
if len(_a ) == 1:
break
else:
__magic_name__ : List[Any] = get_pairs(_a )
__magic_name__ : Optional[int] = "@@ ".join(_a )
__magic_name__ : Tuple = word[:-4]
__magic_name__ : str = word
return word
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = []
__magic_name__ : Dict = re.findall(r"\S+\n?" , _a )
for token in words:
split_tokens.extend(list(self.bpe(_a ).split(" " ) ) )
return split_tokens
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.encoder.get(_a , self.encoder.get(self.unk_token ) )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.decoder.get(_a , self.unk_token )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Tuple = " ".join(_a ).replace("@@ " , "" ).strip()
return out_string
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : Optional[int] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
__magic_name__ : Union[str, Any] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["merges_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ):
copyfile(self.vocab_file , _a )
if os.path.abspath(self.merges_file ) != os.path.abspath(_a ):
copyfile(self.merges_file , _a )
return out_vocab_file, out_merge_file
def SCREAMING_SNAKE_CASE ( self , _a ):
if isinstance(_a , _a ):
try:
with open(_a , "r" , encoding="utf-8" ) as fd:
self.add_from_file(_a )
except FileNotFoundError as fnfe:
raise fnfe
except UnicodeError:
raise Exception(f'''Incorrect encoding detected in {f}, please rebuild the dataset''' )
return
__magic_name__ : List[Any] = f.readlines()
for lineTmp in lines:
__magic_name__ : Optional[Any] = lineTmp.strip()
__magic_name__ : Union[str, Any] = line.rfind(" " )
if idx == -1:
raise ValueError("Incorrect dictionary format, expected '<token> <cnt>'" )
__magic_name__ : Optional[int] = line[:idx]
__magic_name__ : Dict = len(self.encoder )
| 281 | 1 |
import unittest
from transformers import PegasusConfig, PegasusTokenizer, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_flax_common import FlaxModelTesterMixin, ids_tensor
if is_flax_available():
import os
# The slow tests are often failing with OOM error on GPU
# This makes JAX allocate exactly what is needed on demand, and deallocate memory that is no longer needed
# but will be slower as stated here https://jax.readthedocs.io/en/latest/gpu_memory_allocation.html
snake_case : str = "platform"
import jax
import jax.numpy as jnp
import numpy as np
from transformers import FlaxPegasusForConditionalGeneration, FlaxPegasusModel
@require_flax
class _snake_case :
UpperCamelCase__ = PegasusConfig
UpperCamelCase__ = {}
UpperCamelCase__ = 'gelu'
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=False , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a=0.1 , _a=0.1 , _a=20 , _a=2 , _a=1 , _a=0 , ):
__magic_name__ : List[str] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Tuple = seq_length
__magic_name__ : Any = is_training
__magic_name__ : Optional[int] = use_labels
__magic_name__ : Any = vocab_size
__magic_name__ : Tuple = hidden_size
__magic_name__ : List[Any] = num_hidden_layers
__magic_name__ : Tuple = num_attention_heads
__magic_name__ : Union[str, Any] = intermediate_size
__magic_name__ : Union[str, Any] = hidden_dropout_prob
__magic_name__ : Any = attention_probs_dropout_prob
__magic_name__ : Tuple = max_position_embeddings
__magic_name__ : Tuple = eos_token_id
__magic_name__ : Any = pad_token_id
__magic_name__ : Union[str, Any] = bos_token_id
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size ).clip(3 , self.vocab_size )
__magic_name__ : Optional[int] = np.expand_dims(np.array([self.eos_token_id] * self.batch_size ) , 1 )
__magic_name__ : str = np.concatenate([input_ids, eos_tensor] , axis=1 )
__magic_name__ : Dict = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = self.config_cls(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_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_ids=[2] , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.pad_token_id , **self.config_updates , )
__magic_name__ : Any = prepare_pegasus_inputs_dict(_a , _a , _a )
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Optional[int] = 20
__magic_name__ : Optional[int] = model_class_name(_a )
__magic_name__ : Optional[int] = model.encode(inputs_dict["input_ids"] )
__magic_name__ , __magic_name__ : Any = (
inputs_dict["decoder_input_ids"],
inputs_dict["decoder_attention_mask"],
)
__magic_name__ : Tuple = model.init_cache(decoder_input_ids.shape[0] , _a , _a )
__magic_name__ : Optional[Any] = jnp.ones((decoder_input_ids.shape[0], max_decoder_length) , dtype="i4" )
__magic_name__ : int = jnp.broadcast_to(
jnp.arange(decoder_input_ids.shape[-1] - 1 )[None, :] , (decoder_input_ids.shape[0], decoder_input_ids.shape[-1] - 1) , )
__magic_name__ : Dict = model.decode(
decoder_input_ids[:, :-1] , _a , decoder_attention_mask=_a , past_key_values=_a , decoder_position_ids=_a , )
__magic_name__ : Optional[int] = jnp.array(decoder_input_ids.shape[0] * [[decoder_input_ids.shape[-1] - 1]] , dtype="i4" )
__magic_name__ : List[Any] = model.decode(
decoder_input_ids[:, -1:] , _a , decoder_attention_mask=_a , past_key_values=outputs_cache.past_key_values , decoder_position_ids=_a , )
__magic_name__ : Union[str, Any] = model.decode(_a , _a )
__magic_name__ : Dict = np.max(np.abs((outputs_cache_next[0][:, -1, :5] - outputs[0][:, -1, :5]) ) )
self.parent.assertTrue(diff < 1e-3 , msg=f'''Max diff is {diff}''' )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : List[str] = 20
__magic_name__ : Optional[Any] = model_class_name(_a )
__magic_name__ : Optional[int] = model.encode(inputs_dict["input_ids"] )
__magic_name__ , __magic_name__ : Optional[Any] = (
inputs_dict["decoder_input_ids"],
inputs_dict["decoder_attention_mask"],
)
__magic_name__ : int = jnp.concatenate(
[
decoder_attention_mask,
jnp.zeros((decoder_attention_mask.shape[0], max_decoder_length - decoder_attention_mask.shape[1]) ),
] , axis=-1 , )
__magic_name__ : Optional[int] = model.init_cache(decoder_input_ids.shape[0] , _a , _a )
__magic_name__ : List[str] = jnp.broadcast_to(
jnp.arange(decoder_input_ids.shape[-1] - 1 )[None, :] , (decoder_input_ids.shape[0], decoder_input_ids.shape[-1] - 1) , )
__magic_name__ : str = model.decode(
decoder_input_ids[:, :-1] , _a , decoder_attention_mask=_a , past_key_values=_a , decoder_position_ids=_a , )
__magic_name__ : Tuple = jnp.array(decoder_input_ids.shape[0] * [[decoder_input_ids.shape[-1] - 1]] , dtype="i4" )
__magic_name__ : Optional[Any] = model.decode(
decoder_input_ids[:, -1:] , _a , past_key_values=outputs_cache.past_key_values , decoder_attention_mask=_a , decoder_position_ids=_a , )
__magic_name__ : List[Any] = model.decode(_a , _a , decoder_attention_mask=_a )
__magic_name__ : Optional[int] = np.max(np.abs((outputs_cache_next[0][:, -1, :5] - outputs[0][:, -1, :5]) ) )
self.parent.assertTrue(diff < 1e-3 , msg=f'''Max diff is {diff}''' )
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : int , _snake_case : str , _snake_case : Optional[Any]=None , _snake_case : Any=None , ) -> Any:
'''simple docstring'''
if attention_mask is None:
__magic_name__ : int = np.not_equal(_snake_case , config.pad_token_id ).astype(np.inta )
if decoder_attention_mask is None:
__magic_name__ : str = np.concatenate(
[
np.ones(decoder_input_ids[:, :1].shape , dtype=np.inta ),
np.not_equal(decoder_input_ids[:, 1:] , config.pad_token_id ).astype(np.inta ),
] , axis=-1 , )
return {
"input_ids": input_ids,
"decoder_input_ids": decoder_input_ids,
"attention_mask": attention_mask,
"decoder_attention_mask": decoder_attention_mask,
}
@require_flax
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = (
(
FlaxPegasusForConditionalGeneration,
FlaxPegasusModel,
)
if is_flax_available()
else ()
)
UpperCamelCase__ = (FlaxPegasusForConditionalGeneration,) if is_flax_available() else ()
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxPegasusModelTester(self )
__magic_name__ : str = ConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : Optional[int] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
self.model_tester.check_use_cache_forward(_a , _a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
self.model_tester.check_use_cache_forward_with_attn_mask(_a , _a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__ ):
__magic_name__ : Union[str, Any] = self._prepare_for_class(_a , _a )
__magic_name__ : Tuple = model_class(_a )
@jax.jit
def encode_jitted(_a , _a=None , **_a ):
return model.encode(input_ids=_a , attention_mask=_a )
with self.subTest("JIT Enabled" ):
__magic_name__ : Tuple = encode_jitted(**_a ).to_tuple()
with self.subTest("JIT Disabled" ):
with jax.disable_jit():
__magic_name__ : str = encode_jitted(**_a ).to_tuple()
self.assertEqual(len(_a ) , len(_a ) )
for jitted_output, output in zip(_a , _a ):
self.assertEqual(jitted_output.shape , output.shape )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__ ):
__magic_name__ : List[str] = model_class(_a )
__magic_name__ : Any = model.encode(inputs_dict["input_ids"] , inputs_dict["attention_mask"] )
__magic_name__ : Optional[int] = {
"decoder_input_ids": inputs_dict["decoder_input_ids"],
"decoder_attention_mask": inputs_dict["decoder_attention_mask"],
"encoder_outputs": encoder_outputs,
}
@jax.jit
def decode_jitted(_a , _a , _a ):
return model.decode(
decoder_input_ids=_a , decoder_attention_mask=_a , encoder_outputs=_a , )
with self.subTest("JIT Enabled" ):
__magic_name__ : List[str] = decode_jitted(**_a ).to_tuple()
with self.subTest("JIT Disabled" ):
with jax.disable_jit():
__magic_name__ : Dict = decode_jitted(**_a ).to_tuple()
self.assertEqual(len(_a ) , len(_a ) )
for jitted_output, output in zip(_a , _a ):
self.assertEqual(jitted_output.shape , output.shape )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Any = model_class_name.from_pretrained("google/pegasus-large" , from_pt=_a )
__magic_name__ : int = np.ones((1, 1) )
__magic_name__ : int = model(_a )
self.assertIsNotNone(_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxPegasusForConditionalGeneration.from_pretrained("google/pegasus-xsum" )
__magic_name__ : str = PegasusTokenizer.from_pretrained("google/pegasus-xsum" )
__magic_name__ : Union[str, Any] = [
" PG&E stated it scheduled the blackouts in response to forecasts for high winds amid dry conditions. The aim is to reduce the risk of wildfires. Nearly 800 thousand customers were scheduled to be affected by the shutoffs which were expected to last through at least midday tomorrow.",
" The London trio are up for best UK act and best album, as well as getting two nominations in the best song category.\"We got told like this morning 'Oh I think you're nominated'\", said Dappy.\"And I was like 'Oh yeah, which one?' And now we've got nominated for four awards. I mean, wow!\"Bandmate Fazer added: \"We thought it's best of us to come down and mingle with everyone and say hello to the cameras. And now we find we've got four nominations.\"The band have two shots at the best song prize, getting the nod for their Tynchy Stryder collaboration Number One, and single Strong Again.Their album Uncle B will also go up against records by the likes of Beyonce and Kanye West.N-Dubz picked up the best newcomer Mobo in 2007, but female member Tulisa said they wouldn't be too disappointed if they didn't win this time around.\"At the end of the day we're grateful to be where we are in our careers.\"If it don't happen then it don't happen - live to fight another day and keep on making albums and hits for the fans.\"Dappy also revealed they could be performing live several times on the night.The group will be doing Number One and also a possible rendition of the War Child single, I Got Soul.The charity song is a re-working of The Killers' All These Things That I've Done and is set to feature artists like Chipmunk, Ironik and Pixie Lott.This year's Mobos will be held outside of London for the first time, in Glasgow on 30 September.N-Dubz said they were looking forward to performing for their Scottish fans and boasted about their recent shows north of the border.\"We just done Edinburgh the other day,\" said Dappy.\"We smashed up an N-Dubz show over there. We done Aberdeen about three or four months ago - we smashed up that show over there! Everywhere we go we smash it up!\" ",
]
__magic_name__ : Optional[Any] = [
"California's largest electricity provider has turned off power to hundreds of thousands of customers.",
"Pop group N-Dubz have revealed they were surprised to get four nominations for this year's Mobo Awards.",
]
__magic_name__ : Union[str, Any] = tokenizer(_a , return_tensors="np" , truncation=_a , max_length=512 , padding=_a )
__magic_name__ : Optional[int] = model.generate(**_a , num_beams=2 ).sequences
__magic_name__ : Tuple = tokenizer.batch_decode(_a , skip_special_tokens=_a )
assert tgt_text == decoded
| 281 |
from itertools import zip_longest
import requests
from bsa import BeautifulSoup
from pandas import DataFrame
def lowerCAmelCase_ ( _snake_case : str = "laptop" ) -> DataFrame:
'''simple docstring'''
__magic_name__ : Tuple = F'''https://www.amazon.in/laptop/s?k={product}'''
__magic_name__ : Dict = {
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36\n (KHTML, like Gecko)Chrome/44.0.2403.157 Safari/537.36",
"Accept-Language": "en-US, en;q=0.5",
}
__magic_name__ : Tuple = BeautifulSoup(requests.get(_snake_case , headers=_snake_case ).text )
# Initialize a Pandas dataframe with the column titles
__magic_name__ : int = DataFrame(
columns=[
"Product Title",
"Product Link",
"Current Price of the product",
"Product Rating",
"MRP of the product",
"Discount",
] )
# Loop through each entry and store them in the dataframe
for item, _ in zip_longest(
soup.find_all(
"div" , attrs={"class": "s-result-item", "data-component-type": "s-search-result"} , ) , soup.find_all("div" , attrs={"class": "a-row a-size-base a-color-base"} ) , ):
try:
__magic_name__ : Dict = item.ha.text
__magic_name__ : Optional[int] = "https://www.amazon.in/" + item.ha.a["href"]
__magic_name__ : Optional[Any] = item.find("span" , attrs={"class": "a-offscreen"} ).text
try:
__magic_name__ : Union[str, Any] = item.find("span" , attrs={"class": "a-icon-alt"} ).text
except AttributeError:
__magic_name__ : Dict = "Not available"
try:
__magic_name__ : Optional[int] = (
"₹"
+ item.find(
"span" , attrs={"class": "a-price a-text-price"} ).text.split("₹" )[1]
)
except AttributeError:
__magic_name__ : List[str] = ""
try:
__magic_name__ : int = float(
(
(
float(product_mrp.strip("₹" ).replace("," , "" ) )
- float(product_price.strip("₹" ).replace("," , "" ) )
)
/ float(product_mrp.strip("₹" ).replace("," , "" ) )
)
* 100 )
except ValueError:
__magic_name__ : str = float("nan" )
except AttributeError:
pass
__magic_name__ : Optional[int] = [
product_title,
product_link,
product_price,
product_rating,
product_mrp,
discount,
]
__magic_name__ : Optional[Any] = " "
__magic_name__ : str = " "
data_frame.index += 1
return data_frame
if __name__ == "__main__":
snake_case : Any = "headphones"
get_amazon_product_data(product).to_csv(F"Amazon Product Data for {product}.csv")
| 281 | 1 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
snake_case : Dict = {
"configuration_clap": [
"CLAP_PRETRAINED_MODEL_ARCHIVE_LIST",
"ClapAudioConfig",
"ClapConfig",
"ClapTextConfig",
],
"processing_clap": ["ClapProcessor"],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : Dict = [
"CLAP_PRETRAINED_MODEL_ARCHIVE_LIST",
"ClapModel",
"ClapPreTrainedModel",
"ClapTextModel",
"ClapTextModelWithProjection",
"ClapAudioModel",
"ClapAudioModelWithProjection",
]
snake_case : Tuple = ["ClapFeatureExtractor"]
if TYPE_CHECKING:
from .configuration_clap import (
CLAP_PRETRAINED_MODEL_ARCHIVE_LIST,
ClapAudioConfig,
ClapConfig,
ClapTextConfig,
)
from .processing_clap import ClapProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_clap import ClapFeatureExtractor
from .modeling_clap import (
CLAP_PRETRAINED_MODEL_ARCHIVE_LIST,
ClapAudioModel,
ClapAudioModelWithProjection,
ClapModel,
ClapPreTrainedModel,
ClapTextModel,
ClapTextModelWithProjection,
)
else:
import sys
snake_case : Dict = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 281 |
from __future__ import annotations
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Optional[Any] = data
__magic_name__ : Node | None = None
__magic_name__ : Node | None = None
def lowerCAmelCase_ ( _snake_case : Node | None ) -> None: # In Order traversal of the tree
'''simple docstring'''
if tree:
display(tree.left )
print(tree.data )
display(tree.right )
def lowerCAmelCase_ ( _snake_case : Node | None ) -> int:
'''simple docstring'''
return 1 + max(depth_of_tree(tree.left ) , depth_of_tree(tree.right ) ) if tree else 0
def lowerCAmelCase_ ( _snake_case : Node ) -> bool:
'''simple docstring'''
if not tree:
return True
if tree.left and tree.right:
return is_full_binary_tree(tree.left ) and is_full_binary_tree(tree.right )
else:
return not tree.left and not tree.right
def lowerCAmelCase_ ( ) -> None: # Main function for testing.
'''simple docstring'''
__magic_name__ : int = Node(1 )
__magic_name__ : Union[str, Any] = Node(2 )
__magic_name__ : Tuple = Node(3 )
__magic_name__ : Optional[Any] = Node(4 )
__magic_name__ : Union[str, Any] = Node(5 )
__magic_name__ : Any = Node(6 )
__magic_name__ : int = Node(7 )
__magic_name__ : List[str] = Node(8 )
__magic_name__ : Union[str, Any] = Node(9 )
print(is_full_binary_tree(_snake_case ) )
print(depth_of_tree(_snake_case ) )
print("Tree is: " )
display(_snake_case )
if __name__ == "__main__":
main()
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : list[int] ) -> list[list[int]]:
'''simple docstring'''
__magic_name__ : List[str] = []
if len(_snake_case ) == 1:
return [nums.copy()]
for _ in range(len(_snake_case ) ):
__magic_name__ : Optional[int] = nums.pop(0 )
__magic_name__ : List[str] = permute(_snake_case )
for perm in permutations:
perm.append(_snake_case )
result.extend(_snake_case )
nums.append(_snake_case )
return result
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Tuple:
'''simple docstring'''
def backtrack(_snake_case : List[Any] ):
if start == len(_snake_case ) - 1:
output.append(nums[:] )
else:
for i in range(_snake_case , len(_snake_case ) ):
__magic_name__ , __magic_name__ : Any = nums[i], nums[start]
backtrack(start + 1 )
__magic_name__ , __magic_name__ : Any = nums[i], nums[start] # backtrack
__magic_name__ : Union[str, Any] = []
backtrack(0 )
return output
if __name__ == "__main__":
import doctest
# use res to print the data in permute2 function
snake_case : List[str] = permutea([1, 2, 3])
print(res)
doctest.testmod()
| 281 |
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> bool:
'''simple docstring'''
__magic_name__ : Union[str, Any] = len(_snake_case ) + 1
__magic_name__ : List[str] = len(_snake_case ) + 1
# dp is a 2d matrix where dp[i][j] denotes whether prefix string of
# length i of input_string matches with prefix string of length j of
# given pattern.
# "dp" stands for dynamic programming.
__magic_name__ : str = [[0 for i in range(_snake_case )] for j in range(_snake_case )]
# since string of zero length match pattern of zero length
__magic_name__ : Optional[int] = 1
# since pattern of zero length will never match with string of non-zero length
for i in range(1 , _snake_case ):
__magic_name__ : Optional[int] = 0
# since string of zero length will match with pattern where there
# is at least one * alternatively
for j in range(1 , _snake_case ):
__magic_name__ : Union[str, Any] = dp[0][j - 2] if pattern[j - 1] == "*" else 0
# now using bottom-up approach to find for all remaining lengths
for i in range(1 , _snake_case ):
for j in range(1 , _snake_case ):
if input_string[i - 1] == pattern[j - 1] or pattern[j - 1] == ".":
__magic_name__ : Optional[int] = dp[i - 1][j - 1]
elif pattern[j - 1] == "*":
if dp[i][j - 2] == 1:
__magic_name__ : Optional[Any] = 1
elif pattern[j - 2] in (input_string[i - 1], "."):
__magic_name__ : List[Any] = dp[i - 1][j]
else:
__magic_name__ : Union[str, Any] = 0
else:
__magic_name__ : Dict = 0
return bool(dp[-1][-1] )
if __name__ == "__main__":
import doctest
doctest.testmod()
# inputing the strings
# input_string = input("input a string :")
# pattern = input("input a pattern :")
snake_case : Optional[Any] = "aab"
snake_case : List[str] = "c*a*b"
# using function to check whether given string matches the given pattern
if match_pattern(input_string, pattern):
print(F"{input_string} matches the given pattern {pattern}")
else:
print(F"{input_string} does not match with the given pattern {pattern}")
| 281 | 1 |
import pytest
from datasets.utils.sharding import _distribute_shards, _number_of_shards_in_gen_kwargs, _split_gen_kwargs
@pytest.mark.parametrize(
"kwargs, expected" , [
({"num_shards": 0, "max_num_jobs": 1}, []),
({"num_shards": 10, "max_num_jobs": 1}, [range(10 )]),
({"num_shards": 10, "max_num_jobs": 10}, [range(_snake_case , i + 1 ) for i in range(10 )]),
({"num_shards": 1, "max_num_jobs": 10}, [range(1 )]),
({"num_shards": 10, "max_num_jobs": 3}, [range(0 , 4 ), range(4 , 7 ), range(7 , 10 )]),
({"num_shards": 3, "max_num_jobs": 10}, [range(0 , 1 ), range(1 , 2 ), range(2 , 3 )]),
] , )
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Any ) -> Tuple:
'''simple docstring'''
__magic_name__ : List[str] = _distribute_shards(**_snake_case )
assert out == expected
@pytest.mark.parametrize(
"gen_kwargs, max_num_jobs, expected" , [
({"foo": 0}, 10, [{"foo": 0}]),
({"shards": [0, 1, 2, 3]}, 1, [{"shards": [0, 1, 2, 3]}]),
({"shards": [0, 1, 2, 3]}, 4, [{"shards": [0]}, {"shards": [1]}, {"shards": [2]}, {"shards": [3]}]),
({"shards": [0, 1]}, 4, [{"shards": [0]}, {"shards": [1]}]),
({"shards": [0, 1, 2, 3]}, 2, [{"shards": [0, 1]}, {"shards": [2, 3]}]),
] , )
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : Any , _snake_case : List[str] ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : int = _split_gen_kwargs(_snake_case , _snake_case )
assert out == expected
@pytest.mark.parametrize(
"gen_kwargs, expected" , [
({"foo": 0}, 1),
({"shards": [0]}, 1),
({"shards": [0, 1, 2, 3]}, 4),
({"shards": [0, 1, 2, 3], "foo": 0}, 4),
({"shards": [0, 1, 2, 3], "other": (0, 1)}, 4),
({"shards": [0, 1, 2, 3], "shards2": [0, 1]}, RuntimeError),
] , )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : List[Any] ) -> Optional[Any]:
'''simple docstring'''
if expected is RuntimeError:
with pytest.raises(_snake_case ):
_number_of_shards_in_gen_kwargs(_snake_case )
else:
__magic_name__ : int = _number_of_shards_in_gen_kwargs(_snake_case )
assert out == expected
| 281 |
import hashlib
import unittest
from typing import Dict
import numpy as np
from transformers import (
MODEL_FOR_MASK_GENERATION_MAPPING,
TF_MODEL_FOR_MASK_GENERATION_MAPPING,
is_vision_available,
pipeline,
)
from transformers.pipelines import MaskGenerationPipeline
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
)
if is_vision_available():
from PIL import Image
else:
class _snake_case :
@staticmethod
def SCREAMING_SNAKE_CASE ( *_a , **_a ):
pass
def lowerCAmelCase_ ( _snake_case : Image ) -> str:
'''simple docstring'''
__magic_name__ : Optional[int] = hashlib.mda(image.tobytes() )
return m.hexdigest()[:10]
def lowerCAmelCase_ ( _snake_case : Image ) -> Dict:
'''simple docstring'''
__magic_name__ : List[Any] = np.array(_snake_case )
__magic_name__ : Optional[int] = npimg.shape
return {"hash": hashimage(_snake_case ), "shape": shape}
@is_pipeline_test
@require_vision
@require_torch
class _snake_case ( unittest.TestCase ):
UpperCamelCase__ = dict(
(list(MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if MODEL_FOR_MASK_GENERATION_MAPPING else []) )
UpperCamelCase__ = dict(
(list(TF_MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if TF_MODEL_FOR_MASK_GENERATION_MAPPING else []) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Dict = MaskGenerationPipeline(model=_a , image_processor=_a )
return image_segmenter, [
"./tests/fixtures/tests_samples/COCO/000000039769.png",
"./tests/fixtures/tests_samples/COCO/000000039769.png",
]
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
pass
@require_tf
@unittest.skip("Image segmentation not implemented in TF" )
def SCREAMING_SNAKE_CASE ( self ):
pass
@slow
@require_torch
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = pipeline("mask-generation" , model="facebook/sam-vit-huge" )
__magic_name__ : str = image_segmenter("http://images.cocodataset.org/val2017/000000039769.jpg" , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Dict = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
# fmt: off
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.0_21},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
{"mask": {"hash": "e2d0b7a0b7", "shape": (480, 640)}, "scores": 0.99_67},
{"mask": {"hash": "453c7844bd", "shape": (480, 640)}, "scores": 0.9_93},
{"mask": {"hash": "3d44f2926d", "shape": (480, 640)}, "scores": 0.99_09},
{"mask": {"hash": "64033ddc3f", "shape": (480, 640)}, "scores": 0.98_79},
{"mask": {"hash": "801064ff79", "shape": (480, 640)}, "scores": 0.98_34},
{"mask": {"hash": "6172f276ef", "shape": (480, 640)}, "scores": 0.97_16},
{"mask": {"hash": "b49e60e084", "shape": (480, 640)}, "scores": 0.96_12},
{"mask": {"hash": "a811e775fd", "shape": (480, 640)}, "scores": 0.95_99},
{"mask": {"hash": "a6a8ebcf4b", "shape": (480, 640)}, "scores": 0.95_52},
{"mask": {"hash": "9d8257e080", "shape": (480, 640)}, "scores": 0.95_32},
{"mask": {"hash": "32de6454a8", "shape": (480, 640)}, "scores": 0.95_16},
{"mask": {"hash": "af3d4af2c8", "shape": (480, 640)}, "scores": 0.94_99},
{"mask": {"hash": "3c6db475fb", "shape": (480, 640)}, "scores": 0.94_83},
{"mask": {"hash": "c290813fb9", "shape": (480, 640)}, "scores": 0.94_64},
{"mask": {"hash": "b6f0b8f606", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "92ce16bfdf", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "c749b25868", "shape": (480, 640)}, "scores": 0.94_08},
{"mask": {"hash": "efb6cab859", "shape": (480, 640)}, "scores": 0.93_35},
{"mask": {"hash": "1ff2eafb30", "shape": (480, 640)}, "scores": 0.93_26},
{"mask": {"hash": "788b798e24", "shape": (480, 640)}, "scores": 0.92_62},
{"mask": {"hash": "abea804f0e", "shape": (480, 640)}, "scores": 0.89_99},
{"mask": {"hash": "7b9e8ddb73", "shape": (480, 640)}, "scores": 0.89_86},
{"mask": {"hash": "cd24047c8a", "shape": (480, 640)}, "scores": 0.89_84},
{"mask": {"hash": "6943e6bcbd", "shape": (480, 640)}, "scores": 0.88_73},
{"mask": {"hash": "b5f47c9191", "shape": (480, 640)}, "scores": 0.88_71}
] , )
# fmt: on
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = "facebook/sam-vit-huge"
__magic_name__ : str = pipeline("mask-generation" , model=_a )
__magic_name__ : Tuple = image_segmenter(
"http://images.cocodataset.org/val2017/000000039769.jpg" , pred_iou_thresh=1 , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Any = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.02_10},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
] , )
| 281 | 1 |
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel
from diffusers import DDIMScheduler, LDMPipeline, UNetaDModel, VQModel
from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device
enable_full_determinism()
class _snake_case ( unittest.TestCase ):
@property
def SCREAMING_SNAKE_CASE ( self ):
torch.manual_seed(0 )
__magic_name__ : Optional[Any] = UNetaDModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=3 , out_channels=3 , down_block_types=("DownBlock2D", "AttnDownBlock2D") , up_block_types=("AttnUpBlock2D", "UpBlock2D") , )
return model
@property
def SCREAMING_SNAKE_CASE ( self ):
torch.manual_seed(0 )
__magic_name__ : Optional[Any] = VQModel(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"] , up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"] , latent_channels=3 , )
return model
@property
def SCREAMING_SNAKE_CASE ( self ):
torch.manual_seed(0 )
__magic_name__ : Tuple = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
return CLIPTextModel(_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = self.dummy_uncond_unet
__magic_name__ : Tuple = DDIMScheduler()
__magic_name__ : List[Any] = self.dummy_vq_model
__magic_name__ : Union[str, Any] = LDMPipeline(unet=_a , vqvae=_a , scheduler=_a )
ldm.to(_a )
ldm.set_progress_bar_config(disable=_a )
__magic_name__ : str = torch.manual_seed(0 )
__magic_name__ : str = ldm(generator=_a , num_inference_steps=2 , output_type="numpy" ).images
__magic_name__ : int = torch.manual_seed(0 )
__magic_name__ : List[Any] = ldm(generator=_a , num_inference_steps=2 , output_type="numpy" , return_dict=_a )[0]
__magic_name__ : List[str] = image[0, -3:, -3:, -1]
__magic_name__ : Dict = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 64, 64, 3)
__magic_name__ : Any = np.array([0.85_12, 0.8_18, 0.64_11, 0.68_08, 0.44_65, 0.56_18, 0.46, 0.62_31, 0.51_72] )
__magic_name__ : Tuple = 1e-2 if torch_device != "mps" else 3e-2
assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < tolerance
@slow
@require_torch
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = LDMPipeline.from_pretrained("CompVis/ldm-celebahq-256" )
ldm.to(_a )
ldm.set_progress_bar_config(disable=_a )
__magic_name__ : Optional[int] = torch.manual_seed(0 )
__magic_name__ : Union[str, Any] = ldm(generator=_a , num_inference_steps=5 , output_type="numpy" ).images
__magic_name__ : Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 256, 256, 3)
__magic_name__ : List[Any] = np.array([0.43_99, 0.4_49_75, 0.4_68_25, 0.4_74, 0.43_59, 0.45_81, 0.4_50_95, 0.43_41, 0.44_47] )
__magic_name__ : int = 1e-2 if torch_device != "mps" else 3e-2
assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
| 281 |
import absl # noqa: F401 # Here to have a nice missing dependency error message early on
import nltk # noqa: F401 # Here to have a nice missing dependency error message early on
import numpy # noqa: F401 # Here to have a nice missing dependency error message early on
import six # noqa: F401 # Here to have a nice missing dependency error message early on
from rouge_score import rouge_scorer, scoring
import datasets
snake_case : List[Any] = "\\n@inproceedings{lin-2004-rouge,\n title = \"{ROUGE}: A Package for Automatic Evaluation of Summaries\",\n author = \"Lin, Chin-Yew\",\n booktitle = \"Text Summarization Branches Out\",\n month = jul,\n year = \"2004\",\n address = \"Barcelona, Spain\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W04-1013\",\n pages = \"74--81\",\n}\n"
snake_case : Any = "\\nROUGE, or Recall-Oriented Understudy for Gisting Evaluation, is a set of metrics and a software package used for\nevaluating automatic summarization and machine translation software in natural language processing.\nThe metrics compare an automatically produced summary or translation against a reference or a set of references (human-produced) summary or translation.\n\nNote that ROUGE is case insensitive, meaning that upper case letters are treated the same way as lower case letters.\n\nThis metrics is a wrapper around Google Research reimplementation of ROUGE:\nhttps://github.com/google-research/google-research/tree/master/rouge\n"
snake_case : str = "\nCalculates average rouge scores for a list of hypotheses and references\nArgs:\n predictions: list of predictions to score. Each prediction\n should be a string with tokens separated by spaces.\n references: list of reference for each prediction. Each\n reference should be a string with tokens separated by spaces.\n rouge_types: A list of rouge types to calculate.\n Valid names:\n `\"rouge{n}\"` (e.g. `\"rouge1\"`, `\"rouge2\"`) where: {n} is the n-gram based scoring,\n `\"rougeL\"`: Longest common subsequence based scoring.\n `\"rougeLSum\"`: rougeLsum splits text using `\"\n\"`.\n See details in https://github.com/huggingface/datasets/issues/617\n use_stemmer: Bool indicating whether Porter stemmer should be used to strip word suffixes.\n use_aggregator: Return aggregates if this is set to True\nReturns:\n rouge1: rouge_1 (precision, recall, f1),\n rouge2: rouge_2 (precision, recall, f1),\n rougeL: rouge_l (precision, recall, f1),\n rougeLsum: rouge_lsum (precision, recall, f1)\nExamples:\n\n >>> rouge = datasets.load_metric('rouge')\n >>> predictions = [\"hello there\", \"general kenobi\"]\n >>> references = [\"hello there\", \"general kenobi\"]\n >>> results = rouge.compute(predictions=predictions, references=references)\n >>> print(list(results.keys()))\n ['rouge1', 'rouge2', 'rougeL', 'rougeLsum']\n >>> print(results[\"rouge1\"])\n AggregateScore(low=Score(precision=1.0, recall=1.0, fmeasure=1.0), mid=Score(precision=1.0, recall=1.0, fmeasure=1.0), high=Score(precision=1.0, recall=1.0, fmeasure=1.0))\n >>> print(results[\"rouge1\"].mid.fmeasure)\n 1.0\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class _snake_case ( datasets.Metric ):
def SCREAMING_SNAKE_CASE ( self ):
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , 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/google-research/tree/master/rouge"] , reference_urls=[
"https://en.wikipedia.org/wiki/ROUGE_(metric)",
"https://github.com/google-research/google-research/tree/master/rouge",
] , )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a=None , _a=True , _a=False ):
if rouge_types is None:
__magic_name__ : str = ["rouge1", "rouge2", "rougeL", "rougeLsum"]
__magic_name__ : List[str] = rouge_scorer.RougeScorer(rouge_types=_a , use_stemmer=_a )
if use_aggregator:
__magic_name__ : Dict = scoring.BootstrapAggregator()
else:
__magic_name__ : str = []
for ref, pred in zip(_a , _a ):
__magic_name__ : Union[str, Any] = scorer.score(_a , _a )
if use_aggregator:
aggregator.add_scores(_a )
else:
scores.append(_a )
if use_aggregator:
__magic_name__ : Any = aggregator.aggregate()
else:
__magic_name__ : List[Any] = {}
for key in scores[0]:
__magic_name__ : str = [score[key] for score in scores]
return result
| 281 | 1 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : List[Any] = {
"YituTech/conv-bert-base": "https://huggingface.co/YituTech/conv-bert-base/resolve/main/config.json",
"YituTech/conv-bert-medium-small": (
"https://huggingface.co/YituTech/conv-bert-medium-small/resolve/main/config.json"
),
"YituTech/conv-bert-small": "https://huggingface.co/YituTech/conv-bert-small/resolve/main/config.json",
# See all ConvBERT models at https://huggingface.co/models?filter=convbert
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'convbert'
def __init__( self , _a=30_522 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=1 , _a=0 , _a=2 , _a=768 , _a=2 , _a=9 , _a=1 , _a=None , **_a , ):
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a , )
__magic_name__ : Tuple = vocab_size
__magic_name__ : List[Any] = hidden_size
__magic_name__ : Union[str, Any] = num_hidden_layers
__magic_name__ : List[Any] = num_attention_heads
__magic_name__ : str = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : Tuple = max_position_embeddings
__magic_name__ : str = type_vocab_size
__magic_name__ : List[str] = initializer_range
__magic_name__ : Tuple = layer_norm_eps
__magic_name__ : List[Any] = embedding_size
__magic_name__ : List[Any] = head_ratio
__magic_name__ : str = conv_kernel_size
__magic_name__ : Dict = num_groups
__magic_name__ : str = classifier_dropout
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : Dict = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
("token_type_ids", dynamic_axis),
] )
| 281 |
snake_case : Optional[int] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"
def lowerCAmelCase_ ( _snake_case : bytes ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ):
__magic_name__ : Tuple = F'''a bytes-like object is required, not \'{data.__class__.__name__}\''''
raise TypeError(_snake_case )
__magic_name__ : Optional[int] = "".join(bin(_snake_case )[2:].zfill(8 ) for byte in data )
__magic_name__ : List[Any] = len(_snake_case ) % 6 != 0
if padding_needed:
# The padding that will be added later
__magic_name__ : List[str] = B"=" * ((6 - len(_snake_case ) % 6) // 2)
# Append binary_stream with arbitrary binary digits (0's by default) to make its
# length a multiple of 6.
binary_stream += "0" * (6 - len(_snake_case ) % 6)
else:
__magic_name__ : List[str] = B""
# Encode every 6 binary digits to their corresponding Base64 character
return (
"".join(
B64_CHARSET[int(binary_stream[index : index + 6] , 2 )]
for index in range(0 , len(_snake_case ) , 6 ) ).encode()
+ padding
)
def lowerCAmelCase_ ( _snake_case : str ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ) and not isinstance(_snake_case , _snake_case ):
__magic_name__ : List[str] = (
"argument should be a bytes-like object or ASCII string, "
F'''not \'{encoded_data.__class__.__name__}\''''
)
raise TypeError(_snake_case )
# In case encoded_data is a bytes-like object, make sure it contains only
# ASCII characters so we convert it to a string object
if isinstance(_snake_case , _snake_case ):
try:
__magic_name__ : List[Any] = encoded_data.decode("utf-8" )
except UnicodeDecodeError:
raise ValueError("base64 encoded data should only contain ASCII characters" )
__magic_name__ : List[str] = encoded_data.count("=" )
# Check if the encoded string contains non base64 characters
if padding:
assert all(
char in B64_CHARSET for char in encoded_data[:-padding] ), "Invalid base64 character(s) found."
else:
assert all(
char in B64_CHARSET for char in encoded_data ), "Invalid base64 character(s) found."
# Check the padding
assert len(_snake_case ) % 4 == 0 and padding < 3, "Incorrect padding"
if padding:
# Remove padding if there is one
__magic_name__ : Optional[int] = encoded_data[:-padding]
__magic_name__ : Dict = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )[: -padding * 2]
else:
__magic_name__ : Union[str, Any] = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )
__magic_name__ : List[Any] = [
int(binary_stream[index : index + 8] , 2 )
for index in range(0 , len(_snake_case ) , 8 )
]
return bytes(_snake_case )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
import importlib.metadata
import operator
import re
import sys
from typing import Optional
from packaging import version
snake_case : str = {
"<": operator.lt,
"<=": operator.le,
"==": operator.eq,
"!=": operator.ne,
">=": operator.ge,
">": operator.gt,
}
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : int , _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : Tuple , _snake_case : List[Any] ) -> str:
'''simple docstring'''
if got_ver is None or want_ver is None:
raise ValueError(
F'''Unable to compare versions for {requirement}: need={want_ver} found={got_ver}. This is unusual. Consider'''
F''' reinstalling {pkg}.''' )
if not ops[op](version.parse(_snake_case ) , version.parse(_snake_case ) ):
raise ImportError(
F'''{requirement} is required for a normal functioning of this module, but found {pkg}=={got_ver}.{hint}''' )
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Optional[str] = None ) -> None:
'''simple docstring'''
__magic_name__ : Union[str, Any] = F'''\n{hint}''' if hint is not None else ""
# non-versioned check
if re.match(R"^[\w_\-\d]+$" , _snake_case ):
__magic_name__ , __magic_name__ , __magic_name__ : Tuple = requirement, None, None
else:
__magic_name__ : int = re.findall(R"^([^!=<>\s]+)([\s!=<>]{1,2}.+)" , _snake_case )
if not match:
raise ValueError(
"requirement needs to be in the pip package format, .e.g., package_a==1.23, or package_b>=1.23, but"
F''' got {requirement}''' )
__magic_name__ , __magic_name__ : int = match[0]
__magic_name__ : Union[str, Any] = want_full.split("," ) # there could be multiple requirements
__magic_name__ : Tuple = {}
for w in want_range:
__magic_name__ : Any = re.findall(R"^([\s!=<>]{1,2})(.+)" , _snake_case )
if not match:
raise ValueError(
"requirement needs to be in the pip package format, .e.g., package_a==1.23, or package_b>=1.23,"
F''' but got {requirement}''' )
__magic_name__ , __magic_name__ : int = match[0]
__magic_name__ : Tuple = want_ver
if op not in ops:
raise ValueError(F'''{requirement}: need one of {list(ops.keys() )}, but got {op}''' )
# special case
if pkg == "python":
__magic_name__ : Tuple = ".".join([str(_snake_case ) for x in sys.version_info[:3]] )
for op, want_ver in wanted.items():
_compare_versions(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
return
# check if any version is installed
try:
__magic_name__ : Optional[int] = importlib.metadata.version(_snake_case )
except importlib.metadata.PackageNotFoundError:
raise importlib.metadata.PackageNotFoundError(
F'''The \'{requirement}\' distribution was not found and is required by this application. {hint}''' )
# check that the right version is installed if version number or a range was provided
if want_ver is not None:
for op, want_ver in wanted.items():
_compare_versions(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : str = "Try: pip install transformers -U or pip install -e '.[dev]' if you're working with git main"
return require_version(_snake_case , _snake_case )
| 281 |
import unittest
import numpy as np
from transformers import RobertaPreLayerNormConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
import jax.numpy as jnp
from transformers.models.roberta_prelayernorm.modeling_flax_roberta_prelayernorm import (
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormModel,
)
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=True , _a=True , _a=True , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=16 , _a=2 , _a=0.02 , _a=4 , ):
__magic_name__ : List[Any] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Dict = seq_length
__magic_name__ : Union[str, Any] = is_training
__magic_name__ : Optional[Any] = use_attention_mask
__magic_name__ : Optional[Any] = use_token_type_ids
__magic_name__ : int = use_labels
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : Any = intermediate_size
__magic_name__ : List[Any] = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[Any] = max_position_embeddings
__magic_name__ : Tuple = type_vocab_size
__magic_name__ : List[str] = type_sequence_label_size
__magic_name__ : Dict = initializer_range
__magic_name__ : List[Any] = num_choices
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = None
if self.use_attention_mask:
__magic_name__ : List[str] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : str = None
if self.use_token_type_ids:
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__magic_name__ : List[str] = RobertaPreLayerNormConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=_a , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[Any] = config_and_inputs
__magic_name__ : List[str] = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": attention_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : Tuple = True
__magic_name__ : int = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
token_type_ids,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
# Copied from tests.models.roberta.test_modelling_flax_roberta.FlaxRobertaPreLayerNormModelTest with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta-base->andreasmadsen/efficient_mlm_m0.40
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = True
UpperCamelCase__ = (
(
FlaxRobertaPreLayerNormModel,
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
)
if is_flax_available()
else ()
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxRobertaPreLayerNormModelTester(self )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Optional[Any] = model_class_name.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(_a )
@require_flax
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxRobertaPreLayerNormForMaskedLM.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Union[str, Any] = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : List[str] = model(_a )[0]
__magic_name__ : str = [1, 11, 50_265]
self.assertEqual(list(output.shape ) , _a )
# compare the actual values for a slice.
__magic_name__ : List[str] = np.array(
[[[40.48_80, 18.01_99, -5.23_67], [-1.88_77, -4.08_85, 10.70_85], [-2.26_13, -5.61_10, 7.26_65]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = FlaxRobertaPreLayerNormModel.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Tuple = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : Tuple = model(_a )[0]
# compare the actual values for a slice.
__magic_name__ : Dict = np.array(
[[[0.02_08, -0.03_56, 0.02_37], [-0.15_69, -0.04_11, -0.26_26], [0.18_79, 0.01_25, -0.00_89]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
| 281 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
snake_case : Any = logging.get_logger(__name__)
snake_case : Optional[Any] = {
"studio-ousia/luke-base": "https://huggingface.co/studio-ousia/luke-base/resolve/main/config.json",
"studio-ousia/luke-large": "https://huggingface.co/studio-ousia/luke-large/resolve/main/config.json",
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'luke'
def __init__( self , _a=50_267 , _a=500_000 , _a=768 , _a=256 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=True , _a=None , _a=1 , _a=0 , _a=2 , **_a , ):
super().__init__(pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a )
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : Optional[Any] = entity_vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Any = entity_emb_size
__magic_name__ : Any = num_hidden_layers
__magic_name__ : Optional[int] = num_attention_heads
__magic_name__ : Optional[Any] = hidden_act
__magic_name__ : Union[str, Any] = intermediate_size
__magic_name__ : Dict = hidden_dropout_prob
__magic_name__ : str = attention_probs_dropout_prob
__magic_name__ : Dict = max_position_embeddings
__magic_name__ : int = type_vocab_size
__magic_name__ : Tuple = initializer_range
__magic_name__ : Any = layer_norm_eps
__magic_name__ : List[str] = use_entity_aware_attention
__magic_name__ : int = classifier_dropout
| 281 |
def lowerCAmelCase_ ( _snake_case : list[list[int | float]] ) -> int:
'''simple docstring'''
__magic_name__ : Any = len(_snake_case )
__magic_name__ : Optional[Any] = len(matrix[0] )
__magic_name__ : Union[str, Any] = min(_snake_case , _snake_case )
for row in range(_snake_case ):
# Check if diagonal element is not zero
if matrix[row][row] != 0:
# Eliminate all the elements below the diagonal
for col in range(row + 1 , _snake_case ):
__magic_name__ : Optional[Any] = matrix[col][row] / matrix[row][row]
for i in range(_snake_case , _snake_case ):
matrix[col][i] -= multiplier * matrix[row][i]
else:
# Find a non-zero diagonal element to swap rows
__magic_name__ : str = True
for i in range(row + 1 , _snake_case ):
if matrix[i][row] != 0:
__magic_name__ , __magic_name__ : List[str] = matrix[i], matrix[row]
__magic_name__ : Union[str, Any] = False
break
if reduce:
rank -= 1
for i in range(_snake_case ):
__magic_name__ : Any = matrix[i][rank]
# Reduce the row pointer by one to stay on the same row
row -= 1
return rank
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
import inspect
import unittest
from transformers import SegformerConfig, is_torch_available, is_vision_available
from transformers.models.auto import get_values
from transformers.testing_utils import require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
MODEL_MAPPING,
SegformerForImageClassification,
SegformerForSemanticSegmentation,
SegformerModel,
)
from transformers.models.segformer.modeling_segformer import SEGFORMER_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import SegformerImageProcessor
class _snake_case ( snake_case ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = self.config_class(**self.inputs_dict )
self.parent.assertTrue(hasattr(_a , "hidden_sizes" ) )
self.parent.assertTrue(hasattr(_a , "num_attention_heads" ) )
self.parent.assertTrue(hasattr(_a , "num_encoder_blocks" ) )
class _snake_case :
def __init__( self , _a , _a=13 , _a=64 , _a=3 , _a=4 , _a=[2, 2, 2, 2] , _a=[8, 4, 2, 1] , _a=[16, 32, 64, 128] , _a=[1, 4, 8, 16] , _a=[1, 2, 4, 8] , _a=True , _a=True , _a="gelu" , _a=0.1 , _a=0.1 , _a=0.02 , _a=3 , _a=None , ):
__magic_name__ : Union[str, Any] = parent
__magic_name__ : Union[str, Any] = batch_size
__magic_name__ : Dict = image_size
__magic_name__ : Dict = num_channels
__magic_name__ : Optional[int] = num_encoder_blocks
__magic_name__ : int = sr_ratios
__magic_name__ : Any = depths
__magic_name__ : List[str] = hidden_sizes
__magic_name__ : Optional[Any] = downsampling_rates
__magic_name__ : Optional[Any] = num_attention_heads
__magic_name__ : Tuple = is_training
__magic_name__ : Tuple = use_labels
__magic_name__ : str = hidden_act
__magic_name__ : Tuple = hidden_dropout_prob
__magic_name__ : Dict = attention_probs_dropout_prob
__magic_name__ : int = initializer_range
__magic_name__ : Tuple = num_labels
__magic_name__ : Optional[int] = scope
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
__magic_name__ : int = None
if self.use_labels:
__magic_name__ : Tuple = ids_tensor([self.batch_size, self.image_size, self.image_size] , self.num_labels )
__magic_name__ : List[Any] = self.get_config()
return config, pixel_values, labels
def SCREAMING_SNAKE_CASE ( self ):
return SegformerConfig(
image_size=self.image_size , num_channels=self.num_channels , num_encoder_blocks=self.num_encoder_blocks , depths=self.depths , hidden_sizes=self.hidden_sizes , num_attention_heads=self.num_attention_heads , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , initializer_range=self.initializer_range , )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Dict = SegformerModel(config=_a )
model.to(_a )
model.eval()
__magic_name__ : Dict = model(_a )
__magic_name__ : str = self.image_size // (self.downsampling_rates[-1] * 2)
self.parent.assertEqual(
result.last_hidden_state.shape , (self.batch_size, self.hidden_sizes[-1], expected_height, expected_width) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Optional[int] = self.num_labels
__magic_name__ : Optional[Any] = SegformerForSemanticSegmentation(_a )
model.to(_a )
model.eval()
__magic_name__ : int = model(_a )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_labels, self.image_size // 4, self.image_size // 4) )
__magic_name__ : Tuple = model(_a , labels=_a )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_labels, self.image_size // 4, self.image_size // 4) )
self.parent.assertGreater(result.loss , 0.0 )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : List[Any] = 1
__magic_name__ : Union[str, Any] = SegformerForSemanticSegmentation(config=_a )
model.to(_a )
model.eval()
__magic_name__ : Dict = torch.randint(0 , 1 , (self.batch_size, self.image_size, self.image_size) ).to(_a )
__magic_name__ : List[Any] = model(_a , labels=_a )
self.parent.assertGreater(result.loss , 0.0 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : List[Any] = {"pixel_values": pixel_values}
return config, inputs_dict
@require_torch
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (
(
SegformerModel,
SegformerForSemanticSegmentation,
SegformerForImageClassification,
)
if is_torch_available()
else ()
)
UpperCamelCase__ = (
{
'feature-extraction': SegformerModel,
'image-classification': SegformerForImageClassification,
'image-segmentation': SegformerForSemanticSegmentation,
}
if is_torch_available()
else {}
)
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = SegformerModelTester(self )
__magic_name__ : Tuple = SegformerConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_binary_image_segmentation(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_segmentation(*_a )
@unittest.skip("SegFormer does not use inputs_embeds" )
def SCREAMING_SNAKE_CASE ( self ):
pass
@unittest.skip("SegFormer does not have get_input_embeddings method and get_output_embeddings methods" )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__magic_name__ : int = model_class(_a )
__magic_name__ : Tuple = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
__magic_name__ : Dict = [*signature.parameters.keys()]
__magic_name__ : Union[str, Any] = ["pixel_values"]
self.assertListEqual(arg_names[:1] , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[str] = True
for model_class in self.all_model_classes:
__magic_name__ : int = True
__magic_name__ : Any = False
__magic_name__ : List[str] = True
__magic_name__ : Union[str, Any] = model_class(_a )
model.to(_a )
model.eval()
with torch.no_grad():
__magic_name__ : Union[str, Any] = model(**self._prepare_for_class(_a , _a ) )
__magic_name__ : int = outputs.attentions
__magic_name__ : List[str] = sum(self.model_tester.depths )
self.assertEqual(len(_a ) , _a )
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
__magic_name__ : Tuple = True
__magic_name__ : str = model_class(_a )
model.to(_a )
model.eval()
with torch.no_grad():
__magic_name__ : Optional[int] = model(**self._prepare_for_class(_a , _a ) )
__magic_name__ : Optional[int] = outputs.attentions
self.assertEqual(len(_a ) , _a )
# verify the first attentions (first block, first layer)
__magic_name__ : Optional[int] = (self.model_tester.image_size // 4) ** 2
__magic_name__ : Any = (self.model_tester.image_size // (4 * self.model_tester.sr_ratios[0])) ** 2
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads[0], expected_seq_len, expected_reduced_seq_len] , )
# verify the last attentions (last block, last layer)
__magic_name__ : int = (self.model_tester.image_size // 32) ** 2
__magic_name__ : List[str] = (self.model_tester.image_size // (32 * self.model_tester.sr_ratios[-1])) ** 2
self.assertListEqual(
list(attentions[-1].shape[-3:] ) , [self.model_tester.num_attention_heads[-1], expected_seq_len, expected_reduced_seq_len] , )
__magic_name__ : List[str] = len(_a )
# Check attention is always last and order is fine
__magic_name__ : Tuple = True
__magic_name__ : int = True
__magic_name__ : List[Any] = model_class(_a )
model.to(_a )
model.eval()
with torch.no_grad():
__magic_name__ : Any = model(**self._prepare_for_class(_a , _a ) )
self.assertEqual(out_len + 1 , len(_a ) )
__magic_name__ : Optional[Any] = outputs.attentions
self.assertEqual(len(_a ) , _a )
# verify the first attentions (first block, first layer)
__magic_name__ : Optional[Any] = (self.model_tester.image_size // 4) ** 2
__magic_name__ : int = (self.model_tester.image_size // (4 * self.model_tester.sr_ratios[0])) ** 2
self.assertListEqual(
list(self_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads[0], expected_seq_len, expected_reduced_seq_len] , )
def SCREAMING_SNAKE_CASE ( self ):
def check_hidden_states_output(_a , _a , _a ):
__magic_name__ : Union[str, Any] = model_class(_a )
model.to(_a )
model.eval()
with torch.no_grad():
__magic_name__ : Union[str, Any] = model(**self._prepare_for_class(_a , _a ) )
__magic_name__ : List[str] = outputs.hidden_states
__magic_name__ : Optional[int] = self.model_tester.num_encoder_blocks
self.assertEqual(len(_a ) , _a )
# verify the first hidden states (first block)
self.assertListEqual(
list(hidden_states[0].shape[-3:] ) , [
self.model_tester.hidden_sizes[0],
self.model_tester.image_size // 4,
self.model_tester.image_size // 4,
] , )
__magic_name__ , __magic_name__ : Optional[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
__magic_name__ : Any = True
check_hidden_states_output(_a , _a , _a )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
__magic_name__ : List[str] = True
check_hidden_states_output(_a , _a , _a )
def SCREAMING_SNAKE_CASE ( self ):
if not self.model_tester.is_training:
return
__magic_name__ , __magic_name__ : str = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[Any] = True
for model_class in self.all_model_classes:
if model_class in get_values(_a ):
continue
__magic_name__ : Dict = model_class(_a )
model.to(_a )
model.train()
__magic_name__ : Any = self._prepare_for_class(_a , _a , return_labels=_a )
__magic_name__ : int = model(**_a ).loss
loss.backward()
@unittest.skip("Will be fixed soon by reducing the size of the model used for common tests." )
def SCREAMING_SNAKE_CASE ( self ):
pass
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_name in SEGFORMER_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
__magic_name__ : Optional[Any] = SegformerModel.from_pretrained(_a )
self.assertIsNotNone(_a )
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
__magic_name__ : Optional[int] = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png" )
return image
@require_torch
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
# only resize + normalize
__magic_name__ : Optional[int] = SegformerImageProcessor(
image_scale=(512, 512) , keep_ratio=_a , align=_a , do_random_crop=_a )
__magic_name__ : int = SegformerForSemanticSegmentation.from_pretrained("nvidia/segformer-b0-finetuned-ade-512-512" ).to(
_a )
__magic_name__ : List[str] = prepare_img()
__magic_name__ : str = image_processor(images=_a , return_tensors="pt" )
__magic_name__ : Union[str, Any] = encoded_inputs.pixel_values.to(_a )
with torch.no_grad():
__magic_name__ : str = model(_a )
__magic_name__ : Any = torch.Size((1, model.config.num_labels, 128, 128) )
self.assertEqual(outputs.logits.shape , _a )
__magic_name__ : Any = torch.tensor(
[
[[-4.63_10, -5.52_32, -6.23_56], [-5.19_21, -6.14_44, -6.59_96], [-5.44_24, -6.27_90, -6.75_74]],
[[-12.13_91, -13.31_22, -13.95_54], [-12.87_32, -13.93_52, -14.35_63], [-12.94_38, -13.82_26, -14.25_13]],
[[-12.51_34, -13.46_86, -14.49_15], [-12.86_69, -14.43_43, -14.77_58], [-13.25_23, -14.58_19, -15.06_94]],
] ).to(_a )
self.assertTrue(torch.allclose(outputs.logits[0, :3, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# only resize + normalize
__magic_name__ : Any = SegformerImageProcessor(
image_scale=(512, 512) , keep_ratio=_a , align=_a , do_random_crop=_a )
__magic_name__ : Union[str, Any] = SegformerForSemanticSegmentation.from_pretrained(
"nvidia/segformer-b1-finetuned-cityscapes-1024-1024" ).to(_a )
__magic_name__ : Tuple = prepare_img()
__magic_name__ : List[Any] = image_processor(images=_a , return_tensors="pt" )
__magic_name__ : Optional[Any] = encoded_inputs.pixel_values.to(_a )
with torch.no_grad():
__magic_name__ : Union[str, Any] = model(_a )
__magic_name__ : List[str] = torch.Size((1, model.config.num_labels, 128, 128) )
self.assertEqual(outputs.logits.shape , _a )
__magic_name__ : int = torch.tensor(
[
[[-13.57_48, -13.91_11, -12.65_00], [-14.35_00, -15.36_83, -14.23_28], [-14.75_32, -16.04_24, -15.60_87]],
[[-17.16_51, -15.87_25, -12.96_53], [-17.25_80, -17.37_18, -14.82_23], [-16.60_58, -16.87_83, -16.74_52]],
[[-3.64_56, -3.02_09, -1.42_03], [-3.07_97, -3.19_59, -2.00_00], [-1.87_57, -1.92_17, -1.69_97]],
] ).to(_a )
self.assertTrue(torch.allclose(outputs.logits[0, :3, :3, :3] , _a , atol=1e-1 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# only resize + normalize
__magic_name__ : Tuple = SegformerImageProcessor(
image_scale=(512, 512) , keep_ratio=_a , align=_a , do_random_crop=_a )
__magic_name__ : Any = SegformerForSemanticSegmentation.from_pretrained("nvidia/segformer-b0-finetuned-ade-512-512" ).to(
_a )
__magic_name__ : Any = prepare_img()
__magic_name__ : Tuple = image_processor(images=_a , return_tensors="pt" )
__magic_name__ : List[str] = encoded_inputs.pixel_values.to(_a )
with torch.no_grad():
__magic_name__ : Optional[int] = model(_a )
__magic_name__ : Any = outputs.logits.detach().cpu()
__magic_name__ : Union[str, Any] = image_processor.post_process_semantic_segmentation(outputs=_a , target_sizes=[(500, 300)] )
__magic_name__ : Dict = torch.Size((500, 300) )
self.assertEqual(segmentation[0].shape , _a )
__magic_name__ : Optional[int] = image_processor.post_process_semantic_segmentation(outputs=_a )
__magic_name__ : List[str] = torch.Size((128, 128) )
self.assertEqual(segmentation[0].shape , _a )
| 281 |
import argparse
import collections
import json
import os
import re
import string
import sys
import numpy as np
snake_case : Dict = re.compile(R"\b(a|an|the)\b", re.UNICODE)
snake_case : Optional[int] = None
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser("Official evaluation script for SQuAD version 2.0." )
parser.add_argument("data_file" , metavar="data.json" , help="Input data JSON file." )
parser.add_argument("pred_file" , metavar="pred.json" , help="Model predictions." )
parser.add_argument(
"--out-file" , "-o" , metavar="eval.json" , help="Write accuracy metrics to file (default is stdout)." )
parser.add_argument(
"--na-prob-file" , "-n" , metavar="na_prob.json" , help="Model estimates of probability of no answer." )
parser.add_argument(
"--na-prob-thresh" , "-t" , type=_snake_case , default=1.0 , help="Predict \"\" if no-answer probability exceeds this (default = 1.0)." , )
parser.add_argument(
"--out-image-dir" , "-p" , metavar="out_images" , default=_snake_case , help="Save precision-recall curves to directory." )
parser.add_argument("--verbose" , "-v" , action="store_true" )
if len(sys.argv ) == 1:
parser.print_help()
sys.exit(1 )
return parser.parse_args()
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Optional[int] = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : str = bool(qa["answers"]["text"] )
return qid_to_has_ans
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Optional[Any]:
'''simple docstring'''
def remove_articles(_snake_case : List[str] ):
return ARTICLES_REGEX.sub(" " , _snake_case )
def white_space_fix(_snake_case : Optional[int] ):
return " ".join(text.split() )
def remove_punc(_snake_case : Optional[int] ):
__magic_name__ : Dict = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(_snake_case : str ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(_snake_case ) ) ) )
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
if not s:
return []
return normalize_answer(_snake_case ).split()
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Dict ) -> Tuple:
'''simple docstring'''
return int(normalize_answer(_snake_case ) == normalize_answer(_snake_case ) )
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : int ) -> str:
'''simple docstring'''
__magic_name__ : Any = get_tokens(_snake_case )
__magic_name__ : Optional[int] = get_tokens(_snake_case )
__magic_name__ : Tuple = collections.Counter(_snake_case ) & collections.Counter(_snake_case )
__magic_name__ : Tuple = sum(common.values() )
if len(_snake_case ) == 0 or len(_snake_case ) == 0:
# If either is no-answer, then F1 is 1 if they agree, 0 otherwise
return int(gold_toks == pred_toks )
if num_same == 0:
return 0
__magic_name__ : Dict = 1.0 * num_same / len(_snake_case )
__magic_name__ : Optional[Any] = 1.0 * num_same / len(_snake_case )
__magic_name__ : List[Any] = (2 * precision * recall) / (precision + recall)
return fa
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = {}
__magic_name__ : int = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : Union[str, Any] = qa["id"]
__magic_name__ : Any = [t for t in qa["answers"]["text"] if normalize_answer(_snake_case )]
if not gold_answers:
# For unanswerable questions, only correct answer is empty string
__magic_name__ : Tuple = [""]
if qid not in preds:
print(F'''Missing prediction for {qid}''' )
continue
__magic_name__ : Any = preds[qid]
# Take max over all gold answers
__magic_name__ : List[Any] = max(compute_exact(_snake_case , _snake_case ) for a in gold_answers )
__magic_name__ : int = max(compute_fa(_snake_case , _snake_case ) for a in gold_answers )
return exact_scores, fa_scores
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] , _snake_case : Optional[int] , _snake_case : Dict ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : str = {}
for qid, s in scores.items():
__magic_name__ : Dict = na_probs[qid] > na_prob_thresh
if pred_na:
__magic_name__ : str = float(not qid_to_has_ans[qid] )
else:
__magic_name__ : Optional[int] = s
return new_scores
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : List[str] , _snake_case : Tuple=None ) -> Tuple:
'''simple docstring'''
if not qid_list:
__magic_name__ : Any = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores.values() ) / total),
("f1", 100.0 * sum(fa_scores.values() ) / total),
("total", total),
] )
else:
__magic_name__ : Tuple = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores[k] for k in qid_list ) / total),
("f1", 100.0 * sum(fa_scores[k] for k in qid_list ) / total),
("total", total),
] )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : str , _snake_case : str ) -> Dict:
'''simple docstring'''
for k in new_eval:
__magic_name__ : int = new_eval[k]
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : Optional[Any] , _snake_case : Union[str, Any] ) -> str:
'''simple docstring'''
plt.step(_snake_case , _snake_case , color="b" , alpha=0.2 , where="post" )
plt.fill_between(_snake_case , _snake_case , step="post" , alpha=0.2 , color="b" )
plt.xlabel("Recall" )
plt.ylabel("Precision" )
plt.xlim([0.0, 1.05] )
plt.ylim([0.0, 1.05] )
plt.title(_snake_case )
plt.savefig(_snake_case )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any , _snake_case : Optional[int] , _snake_case : List[Any] , _snake_case : Optional[int]=None , _snake_case : int=None ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
__magic_name__ : Optional[int] = 0.0
__magic_name__ : str = 1.0
__magic_name__ : str = 0.0
__magic_name__ : List[str] = [1.0]
__magic_name__ : str = [0.0]
__magic_name__ : Optional[Any] = 0.0
for i, qid in enumerate(_snake_case ):
if qid_to_has_ans[qid]:
true_pos += scores[qid]
__magic_name__ : List[str] = true_pos / float(i + 1 )
__magic_name__ : Any = true_pos / float(_snake_case )
if i == len(_snake_case ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]:
# i.e., if we can put a threshold after this point
avg_prec += cur_p * (cur_r - recalls[-1])
precisions.append(_snake_case )
recalls.append(_snake_case )
if out_image:
plot_pr_curve(_snake_case , _snake_case , _snake_case , _snake_case )
return {"ap": 100.0 * avg_prec}
def lowerCAmelCase_ ( _snake_case : Tuple , _snake_case : Optional[Any] , _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : Any , _snake_case : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
if out_image_dir and not os.path.exists(_snake_case ):
os.makedirs(_snake_case )
__magic_name__ : Any = sum(1 for v in qid_to_has_ans.values() if v )
if num_true_pos == 0:
return
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_exact.png" ) , title="Precision-Recall curve for Exact Match score" , )
__magic_name__ : Union[str, Any] = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_f1.png" ) , title="Precision-Recall curve for F1 score" , )
__magic_name__ : str = {k: float(_snake_case ) for k, v in qid_to_has_ans.items()}
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_oracle.png" ) , title="Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)" , )
merge_eval(_snake_case , _snake_case , "pr_exact" )
merge_eval(_snake_case , _snake_case , "pr_f1" )
merge_eval(_snake_case , _snake_case , "pr_oracle" )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : Optional[Any] , _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
if not qid_list:
return
__magic_name__ : Dict = [na_probs[k] for k in qid_list]
__magic_name__ : str = np.ones_like(_snake_case ) / float(len(_snake_case ) )
plt.hist(_snake_case , weights=_snake_case , bins=20 , range=(0.0, 1.0) )
plt.xlabel("Model probability of no-answer" )
plt.ylabel("Proportion of dataset" )
plt.title(F'''Histogram of no-answer probability: {name}''' )
plt.savefig(os.path.join(_snake_case , F'''na_prob_hist_{name}.png''' ) )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : List[str] , _snake_case : Dict ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] )
__magic_name__ : List[str] = num_no_ans
__magic_name__ : Dict = cur_score
__magic_name__ : Dict = 0.0
__magic_name__ : Any = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
for i, qid in enumerate(_snake_case ):
if qid not in scores:
continue
if qid_to_has_ans[qid]:
__magic_name__ : Union[str, Any] = scores[qid]
else:
if preds[qid]:
__magic_name__ : List[Any] = -1
else:
__magic_name__ : Optional[int] = 0
cur_score += diff
if cur_score > best_score:
__magic_name__ : Optional[int] = cur_score
__magic_name__ : List[Any] = na_probs[qid]
return 100.0 * best_score / len(_snake_case ), best_thresh
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str , _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Dict ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ , __magic_name__ : List[str] = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ , __magic_name__ : int = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ : Optional[int] = best_exact
__magic_name__ : List[Any] = exact_thresh
__magic_name__ : Dict = best_fa
__magic_name__ : Any = fa_thresh
def lowerCAmelCase_ ( ) -> int:
'''simple docstring'''
with open(OPTS.data_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
__magic_name__ : List[Any] = dataset_json["data"]
with open(OPTS.pred_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
if OPTS.na_prob_file:
with open(OPTS.na_prob_file ) as f:
__magic_name__ : Any = json.load(_snake_case )
else:
__magic_name__ : Any = {k: 0.0 for k in preds}
__magic_name__ : str = make_qid_to_has_ans(_snake_case ) # maps qid to True/False
__magic_name__ : Tuple = [k for k, v in qid_to_has_ans.items() if v]
__magic_name__ : Optional[Any] = [k for k, v in qid_to_has_ans.items() if not v]
__magic_name__ , __magic_name__ : Union[str, Any] = get_raw_scores(_snake_case , _snake_case )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case )
if has_ans_qids:
__magic_name__ : int = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "HasAns" )
if no_ans_qids:
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "NoAns" )
if OPTS.na_prob_file:
find_all_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
if OPTS.na_prob_file and OPTS.out_image_dir:
run_precision_recall_analysis(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , OPTS.out_image_dir )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "hasAns" )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "noAns" )
if OPTS.out_file:
with open(OPTS.out_file , "w" ) as f:
json.dump(_snake_case , _snake_case )
else:
print(json.dumps(_snake_case , indent=2 ) )
if __name__ == "__main__":
snake_case : int = parse_args()
if OPTS.out_image_dir:
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
main()
| 281 | 1 |
from __future__ import annotations
def lowerCAmelCase_ ( _snake_case : list[int | float] , _snake_case : int , _snake_case : int ) -> int | float:
'''simple docstring'''
if len(_snake_case ) == 0:
raise ValueError("find_max() arg is an empty sequence" )
if (
left >= len(_snake_case )
or left < -len(_snake_case )
or right >= len(_snake_case )
or right < -len(_snake_case )
):
raise IndexError("list index out of range" )
if left == right:
return nums[left]
__magic_name__ : Optional[Any] = (left + right) >> 1 # the middle
__magic_name__ : Optional[int] = find_max(_snake_case , _snake_case , _snake_case ) # find max in range[left, mid]
__magic_name__ : Union[str, Any] = find_max(_snake_case , mid + 1 , _snake_case ) # find max in range[mid + 1, right]
return left_max if left_max >= right_max else right_max
if __name__ == "__main__":
import doctest
doctest.testmod(verbose=True)
| 281 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
snake_case : str = "▁"
snake_case : List[Any] = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = BigBirdTokenizer
UpperCamelCase__ = BigBirdTokenizerFast
UpperCamelCase__ = True
UpperCamelCase__ = True
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
__magic_name__ : Optional[Any] = self.tokenizer_class(_a , keep_accents=_a )
tokenizer.save_pretrained(self.tmpdirname )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = "<s>"
__magic_name__ : Dict = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_a ) , _a )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_a ) , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : 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] , "[MASK]" )
self.assertEqual(len(_a ) , 1_004 )
def SCREAMING_SNAKE_CASE ( self ):
self.assertEqual(self.get_tokenizer().vocab_size , 1_000 )
def SCREAMING_SNAKE_CASE ( self ):
if not self.test_rust_tokenizer:
return
__magic_name__ : Dict = self.get_tokenizer()
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Any = "I was born in 92000, and this is falsé."
__magic_name__ : Dict = tokenizer.tokenize(_a )
__magic_name__ : Any = rust_tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : List[Any] = tokenizer.encode(_a , add_special_tokens=_a )
__magic_name__ : List[str] = rust_tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Dict = tokenizer.encode(_a )
__magic_name__ : Optional[int] = rust_tokenizer.encode(_a )
self.assertListEqual(_a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = BigBirdTokenizer(_a , keep_accents=_a )
__magic_name__ : str = tokenizer.tokenize("This is a test" )
self.assertListEqual(_a , ["▁This", "▁is", "▁a", "▁t", "est"] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_a ) , [285, 46, 10, 170, 382] , )
__magic_name__ : Dict = tokenizer.tokenize("I was born in 92000, and this is falsé." )
self.assertListEqual(
_a , [
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",
".",
] , )
__magic_name__ : Union[str, Any] = tokenizer.convert_tokens_to_ids(_a )
self.assertListEqual(
_a , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
__magic_name__ : int = tokenizer.convert_ids_to_tokens(_a )
self.assertListEqual(
_a , [
SPIECE_UNDERLINE + "I",
SPIECE_UNDERLINE + "was",
SPIECE_UNDERLINE + "b",
"or",
"n",
SPIECE_UNDERLINE + "in",
SPIECE_UNDERLINE + "",
"<unk>",
"2",
"0",
"0",
"0",
",",
SPIECE_UNDERLINE + "and",
SPIECE_UNDERLINE + "this",
SPIECE_UNDERLINE + "is",
SPIECE_UNDERLINE + "f",
"al",
"s",
"<unk>",
".",
] , )
@cached_property
def SCREAMING_SNAKE_CASE ( self ):
return BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "Hello World!"
__magic_name__ : Dict = [65, 18_536, 2_260, 101, 66]
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = (
"This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"
" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"
)
# fmt: off
__magic_name__ : List[str] = [65, 871, 419, 358, 946, 991, 2_521, 452, 358, 1_357, 387, 7_751, 3_536, 112, 985, 456, 126, 865, 938, 5_400, 5_734, 458, 1_368, 467, 786, 2_462, 5_246, 1_159, 633, 865, 4_519, 457, 582, 852, 2_557, 427, 916, 508, 405, 34_324, 497, 391, 408, 11_342, 1_244, 385, 100, 938, 985, 456, 574, 362, 12_597, 3_200, 3_129, 1_172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
__magic_name__ : Optional[Any] = list(self.big_tokenizer.get_vocab().keys() )[:10]
__magic_name__ : List[Any] = " ".join(_a )
__magic_name__ : Any = self.big_tokenizer.encode_plus(_a , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : Union[str, Any] = self.big_tokenizer.batch_encode_plus(
[sequence + " " + sequence] , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : List[str] = BigBirdConfig(attention_type="original_full" )
__magic_name__ : Optional[int] = BigBirdModel(_a )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_a )
model(**_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
__magic_name__ : int = tokenizer.decode(tokenizer("Paris is the [MASK]." ).input_ids )
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# fmt: off
__magic_name__ : Optional[Any] = {"input_ids": [[65, 39_286, 458, 36_335, 2_001, 456, 13_073, 13_266, 455, 113, 7_746, 1_741, 11_157, 391, 13_073, 13_266, 455, 113, 3_967, 35_412, 113, 4_936, 109, 3_870, 2_377, 113, 30_084, 45_720, 458, 134, 17_496, 112, 503, 11_672, 113, 118, 112, 5_665, 13_347, 38_687, 112, 1_496, 31_389, 112, 3_268, 47_264, 134, 962, 112, 16_377, 8_035, 23_130, 430, 12_169, 15_518, 28_592, 458, 146, 41_697, 109, 391, 12_169, 15_518, 16_689, 458, 146, 41_358, 109, 452, 726, 4_034, 111, 763, 35_412, 5_082, 388, 1_903, 111, 9_051, 391, 2_870, 48_918, 1_900, 1_123, 550, 998, 112, 9_586, 15_985, 455, 391, 410, 22_955, 37_636, 114, 66], [65, 448, 17_496, 419, 3_663, 385, 763, 113, 27_533, 2_870, 3_283, 13_043, 1_639, 24_713, 523, 656, 24_013, 18_550, 2_521, 517, 27_014, 21_244, 420, 1_212, 1_465, 391, 927, 4_833, 388, 578, 11_786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2_169, 7_687, 21_932, 18_146, 726, 363, 17_032, 3_391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], "attention_mask": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_a , model_name="google/bigbird-roberta-base" , revision="215c99f1600e06f83acce68422f2035b2b5c3510" , )
| 281 | 1 |
from ..utils import DummyObject, requires_backends
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
class _snake_case ( metaclass=snake_case ):
UpperCamelCase__ = ['sentencepiece']
def __init__( self , *_a , **_a ):
requires_backends(self , ["sentencepiece"] )
| 281 |
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
snake_case : int = logging.get_logger(__name__)
snake_case : List[str] = {"vocab_file": "spiece.model"}
snake_case : List[str] = {
"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",
}
}
snake_case : Tuple = {
"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,
}
snake_case : List[str] = "▁"
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a=True , _a=True , _a=False , _a="[CLS]" , _a="[SEP]" , _a="<unk>" , _a="[SEP]" , _a="<pad>" , _a="[CLS]" , _a="[MASK]" , _a = None , **_a , ):
# 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.
__magic_name__ : str = (
AddedToken(_a , lstrip=_a , rstrip=_a , normalized=_a )
if isinstance(_a , _a )
else mask_token
)
__magic_name__ : Union[str, Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_a , remove_space=_a , keep_accents=_a , bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , pad_token=_a , cls_token=_a , mask_token=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Dict = do_lower_case
__magic_name__ : Tuple = remove_space
__magic_name__ : Union[str, Any] = keep_accents
__magic_name__ : Tuple = vocab_file
__magic_name__ : int = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_a )
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.sp_model )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ):
__magic_name__ : List[str] = self.__dict__.copy()
__magic_name__ : Any = None
return state
def __setstate__( self , _a ):
__magic_name__ : Union[str, Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : str = {}
__magic_name__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.remove_space:
__magic_name__ : List[Any] = " ".join(inputs.strip().split() )
else:
__magic_name__ : str = inputs
__magic_name__ : int = outputs.replace("``" , "\"" ).replace("''" , "\"" )
if not self.keep_accents:
__magic_name__ : str = unicodedata.normalize("NFKD" , _a )
__magic_name__ : Tuple = "".join([c for c in outputs if not unicodedata.combining(_a )] )
if self.do_lower_case:
__magic_name__ : int = outputs.lower()
return outputs
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = self.preprocess_text(_a )
__magic_name__ : Dict = self.sp_model.encode(_a , out_type=_a )
__magic_name__ : Any = []
for piece in pieces:
if len(_a ) > 1 and piece[-1] == str("," ) and piece[-2].isdigit():
__magic_name__ : Optional[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(_a , "" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
__magic_name__ : List[str] = cur_pieces[1:]
else:
__magic_name__ : Optional[int] = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(_a )
else:
new_pieces.append(_a )
return new_pieces
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.PieceToId(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.IdToPiece(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Any = []
__magic_name__ : Union[str, Any] = ""
__magic_name__ : int = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_a ) + token
__magic_name__ : List[Any] = True
__magic_name__ : Optional[int] = []
else:
current_sub_tokens.append(_a )
__magic_name__ : Optional[Any] = False
out_string += self.sp_model.decode(_a )
return out_string.strip()
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : List[str] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is not None:
return [1] + ([0] * len(_a )) + [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[int] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : List[str] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _a )
elif not os.path.isfile(self.vocab_file ):
with open(_a , "wb" ) as fi:
__magic_name__ : Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(_a )
return (out_vocab_file,)
| 281 | 1 |
from typing import List, Optional, Tuple, Union
import PIL
import torch
from torchvision import transforms
from diffusers.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
from diffusers.schedulers import DDIMScheduler
from diffusers.utils import randn_tensor
snake_case : List[str] = transforms.Compose(
[
transforms.Resize((256, 256)),
transforms.ToTensor(),
transforms.Normalize([0.5], [0.5]),
]
)
def lowerCAmelCase_ ( _snake_case : Any ) -> Dict:
'''simple docstring'''
if isinstance(_snake_case , torch.Tensor ):
return image
elif isinstance(_snake_case , PIL.Image.Image ):
__magic_name__ : List[Any] = [image]
__magic_name__ : Dict = [trans(img.convert("RGB" ) ) for img in image]
__magic_name__ : Dict = torch.stack(_snake_case )
return image
class _snake_case ( snake_case ):
def __init__( self , _a , _a ):
super().__init__()
# make sure scheduler can always be converted to DDIM
__magic_name__ : int = DDIMScheduler.from_config(scheduler.config )
self.register_modules(unet=_a , scheduler=_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
if strength < 0 or strength > 1:
raise ValueError(f'''The value of strength should in [0.0, 1.0] but is {strength}''' )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
# get the original timestep using init_timestep
__magic_name__ : int = min(int(num_inference_steps * strength ) , _a )
__magic_name__ : str = max(num_inference_steps - init_timestep , 0 )
__magic_name__ : Optional[int] = self.scheduler.timesteps[t_start:]
return timesteps, num_inference_steps - t_start
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a=None ):
if not isinstance(_a , (torch.Tensor, PIL.Image.Image, list) ):
raise ValueError(
f'''`image` has to be of type `torch.Tensor`, `PIL.Image.Image` or list but is {type(_a )}''' )
__magic_name__ : int = image.to(device=_a , dtype=_a )
if isinstance(_a , _a ) and len(_a ) != batch_size:
raise ValueError(
f'''You have passed a list of generators of length {len(_a )}, but requested an effective batch'''
f''' size of {batch_size}. Make sure the batch size matches the length of the generators.''' )
__magic_name__ : Tuple = init_latents.shape
__magic_name__ : Tuple = randn_tensor(_a , generator=_a , device=_a , dtype=_a )
# get latents
print("add noise to latents at timestep" , _a )
__magic_name__ : int = self.scheduler.add_noise(_a , _a , _a )
__magic_name__ : Dict = init_latents
return latents
@torch.no_grad()
def __call__( self , _a = None , _a = 0.8 , _a = 1 , _a = None , _a = 0.0 , _a = 50 , _a = None , _a = "pil" , _a = True , ):
self.check_inputs(_a )
# 2. Preprocess image
__magic_name__ : List[Any] = preprocess(_a )
# 3. set timesteps
self.scheduler.set_timesteps(_a , device=self.device )
__magic_name__ , __magic_name__ : List[str] = self.get_timesteps(_a , _a , self.device )
__magic_name__ : Optional[Any] = timesteps[:1].repeat(_a )
# 4. Prepare latent variables
__magic_name__ : Tuple = self.prepare_latents(_a , _a , _a , self.unet.dtype , self.device , _a )
__magic_name__ : Dict = latents
# 5. Denoising loop
for t in self.progress_bar(_a ):
# 1. predict noise model_output
__magic_name__ : Tuple = self.unet(_a , _a ).sample
# 2. predict previous mean of image x_t-1 and add variance depending on eta
# eta corresponds to η in paper and should be between [0, 1]
# do x_t -> x_t-1
__magic_name__ : List[str] = self.scheduler.step(
_a , _a , _a , eta=_a , use_clipped_model_output=_a , generator=_a , ).prev_sample
__magic_name__ : Any = (image / 2 + 0.5).clamp(0 , 1 )
__magic_name__ : Tuple = image.cpu().permute(0 , 2 , 3 , 1 ).numpy()
if output_type == "pil":
__magic_name__ : str = self.numpy_to_pil(_a )
if not return_dict:
return (image, latent_timestep.item())
return ImagePipelineOutput(images=_a )
| 281 |
import unicodedata
from dataclasses import dataclass
from typing import Optional, Union
import numpy as np
from transformers.data.data_collator import DataCollatorMixin
from transformers.file_utils import PaddingStrategy
from transformers.tokenization_utils_base import PreTrainedTokenizerBase
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Union[str, Any] = np.full((len(_snake_case ), sequence_length, 2) , _snake_case )
else:
__magic_name__ : List[Any] = np.full((len(_snake_case ), sequence_length) , _snake_case )
for i, tensor in enumerate(_snake_case ):
if padding_side == "right":
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Optional[Any] = tensor[:sequence_length]
else:
__magic_name__ : Union[str, Any] = tensor[:sequence_length]
else:
if isinstance(_snake_case , _snake_case ):
__magic_name__ : List[Any] = tensor[:sequence_length]
else:
__magic_name__ : Optional[Any] = tensor[:sequence_length]
return out_tensor.tolist()
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = ord(_snake_case )
if (cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126):
return True
__magic_name__ : Any = unicodedata.category(_snake_case )
if cat.startswith("P" ):
return True
return False
@dataclass
class _snake_case ( snake_case ):
UpperCamelCase__ = 42
UpperCamelCase__ = True
UpperCamelCase__ = None
UpperCamelCase__ = None
UpperCamelCase__ = -100
UpperCamelCase__ = "pt"
def SCREAMING_SNAKE_CASE ( self , _a ):
import torch
__magic_name__ : List[str] = "label" if "label" in features[0].keys() else "labels"
__magic_name__ : Union[str, Any] = [feature[label_name] for feature in features] if label_name in features[0].keys() else None
__magic_name__ : Optional[int] = self.tokenizer.pad(
_a , padding=self.padding , max_length=self.max_length , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors="pt" if labels is None else None , )
if labels is None:
return batch
__magic_name__ : Dict = torch.tensor(batch["entity_ids"] ).shape[1]
__magic_name__ : List[Any] = self.tokenizer.padding_side
if padding_side == "right":
__magic_name__ : str = [
list(_a ) + [self.label_pad_token_id] * (sequence_length - len(_a )) for label in labels
]
else:
__magic_name__ : int = [
[self.label_pad_token_id] * (sequence_length - len(_a )) + list(_a ) for label in labels
]
__magic_name__ : Dict = [feature["ner_tags"] for feature in features]
__magic_name__ : List[Any] = padding_tensor(_a , -1 , _a , _a )
__magic_name__ : Any = [feature["original_entity_spans"] for feature in features]
__magic_name__ : Any = padding_tensor(_a , (-1, -1) , _a , _a )
__magic_name__ : List[Any] = {k: torch.tensor(_a , dtype=torch.intaa ) for k, v in batch.items()}
return batch
| 281 | 1 |
import unittest
import numpy as np
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
from transformers import MobileViTImageProcessor
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=7 , _a=3 , _a=18 , _a=30 , _a=400 , _a=True , _a=None , _a=True , _a=None , _a=True , ):
__magic_name__ : Tuple = size if size is not None else {"shortest_edge": 20}
__magic_name__ : List[str] = crop_size if crop_size is not None else {"height": 18, "width": 18}
__magic_name__ : str = parent
__magic_name__ : Union[str, Any] = batch_size
__magic_name__ : Optional[Any] = num_channels
__magic_name__ : Dict = image_size
__magic_name__ : Optional[int] = min_resolution
__magic_name__ : Optional[Any] = max_resolution
__magic_name__ : Optional[int] = do_resize
__magic_name__ : List[Any] = size
__magic_name__ : List[Any] = do_center_crop
__magic_name__ : Any = crop_size
__magic_name__ : List[Any] = do_flip_channel_order
def SCREAMING_SNAKE_CASE ( self ):
return {
"do_resize": self.do_resize,
"size": self.size,
"do_center_crop": self.do_center_crop,
"crop_size": self.crop_size,
"do_flip_channel_order": self.do_flip_channel_order,
}
@require_torch
@require_vision
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = MobileViTImageProcessor if is_vision_available() else None
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = MobileViTImageProcessingTester(self )
@property
def SCREAMING_SNAKE_CASE ( self ):
return self.image_processor_tester.prepare_image_processor_dict()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(_a , "do_resize" ) )
self.assertTrue(hasattr(_a , "size" ) )
self.assertTrue(hasattr(_a , "do_center_crop" ) )
self.assertTrue(hasattr(_a , "center_crop" ) )
self.assertTrue(hasattr(_a , "do_flip_channel_order" ) )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = self.image_processing_class.from_dict(self.image_processor_dict )
self.assertEqual(image_processor.size , {"shortest_edge": 20} )
self.assertEqual(image_processor.crop_size , {"height": 18, "width": 18} )
__magic_name__ : Union[str, Any] = self.image_processing_class.from_dict(self.image_processor_dict , size=42 , crop_size=84 )
self.assertEqual(image_processor.size , {"shortest_edge": 42} )
self.assertEqual(image_processor.crop_size , {"height": 84, "width": 84} )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processing
__magic_name__ : Any = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__magic_name__ : Optional[int] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a )
for image in image_inputs:
self.assertIsInstance(_a , Image.Image )
# Test not batched input
__magic_name__ : str = image_processing(image_inputs[0] , return_tensors="pt" ).pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
) , )
# Test batched
__magic_name__ : List[Any] = image_processing(_a , return_tensors="pt" ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
) , )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processing
__magic_name__ : List[Any] = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
__magic_name__ : Optional[Any] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a , numpify=_a )
for image in image_inputs:
self.assertIsInstance(_a , np.ndarray )
# Test not batched input
__magic_name__ : int = image_processing(image_inputs[0] , return_tensors="pt" ).pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
) , )
# Test batched
__magic_name__ : Dict = image_processing(_a , return_tensors="pt" ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
) , )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processing
__magic_name__ : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
__magic_name__ : List[str] = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a , torchify=_a )
for image in image_inputs:
self.assertIsInstance(_a , torch.Tensor )
# Test not batched input
__magic_name__ : Union[str, Any] = image_processing(image_inputs[0] , return_tensors="pt" ).pixel_values
self.assertEqual(
encoded_images.shape , (
1,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
) , )
# Test batched
__magic_name__ : Optional[Any] = image_processing(_a , return_tensors="pt" ).pixel_values
self.assertEqual(
encoded_images.shape , (
self.image_processor_tester.batch_size,
self.image_processor_tester.num_channels,
self.image_processor_tester.crop_size["height"],
self.image_processor_tester.crop_size["width"],
) , )
| 281 |
import math
def lowerCAmelCase_ ( _snake_case : float , _snake_case : float ) -> float:
'''simple docstring'''
return math.pow(_snake_case , 2 ) - a
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
return 2 * x
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
__magic_name__ : Optional[int] = 2.0
while start <= a:
__magic_name__ : str = math.pow(_snake_case , 2 )
return start
def lowerCAmelCase_ ( _snake_case : float , _snake_case : int = 9999 , _snake_case : float = 0.00_000_000_000_001 ) -> float:
'''simple docstring'''
if a < 0:
raise ValueError("math domain error" )
__magic_name__ : Optional[int] = get_initial_point(_snake_case )
for _ in range(_snake_case ):
__magic_name__ : int = value
__magic_name__ : str = value - fx(_snake_case , _snake_case ) / fx_derivative(_snake_case )
if abs(prev_value - value ) < tolerance:
return value
return value
if __name__ == "__main__":
from doctest import testmod
testmod()
| 281 | 1 |
import os
import unittest
from transformers import LxmertTokenizer, LxmertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = LxmertTokenizer
UpperCamelCase__ = LxmertTokenizerFast
UpperCamelCase__ = True
UpperCamelCase__ = True
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
__magic_name__ : Optional[int] = [
"[UNK]",
"[CLS]",
"[SEP]",
"want",
"##want",
"##ed",
"wa",
"un",
"runn",
"##ing",
",",
"low",
"lowest",
]
__magic_name__ : str = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["vocab_file"] )
with open(self.vocab_file , "w" , encoding="utf-8" ) as vocab_writer:
vocab_writer.write("".join([x + "\n" for x in vocab_tokens] ) )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[Any] = "UNwant\u00E9d,running"
__magic_name__ : Dict = "unwanted, running"
return input_text, output_text
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.tokenizer_class(self.vocab_file )
__magic_name__ : Union[str, Any] = tokenizer.tokenize("UNwant\u00E9d,running" )
self.assertListEqual(_a , ["un", "##want", "##ed", ",", "runn", "##ing"] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(_a ) , [7, 4, 5, 10, 8, 9] )
def SCREAMING_SNAKE_CASE ( self ):
if not self.test_rust_tokenizer:
return
__magic_name__ : Union[str, Any] = self.get_tokenizer()
__magic_name__ : Tuple = self.get_rust_tokenizer()
__magic_name__ : Optional[int] = "I was born in 92000, and this is falsé."
__magic_name__ : List[str] = tokenizer.tokenize(_a )
__magic_name__ : Dict = rust_tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : Tuple = tokenizer.encode(_a , add_special_tokens=_a )
__magic_name__ : str = rust_tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Optional[int] = tokenizer.encode(_a )
__magic_name__ : Union[str, Any] = rust_tokenizer.encode(_a )
self.assertListEqual(_a , _a )
| 281 |
from __future__ import annotations
import unittest
from transformers import LEDConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFLEDForConditionalGeneration, TFLEDModel
@require_tf
class _snake_case :
UpperCamelCase__ = LEDConfig
UpperCamelCase__ = {}
UpperCamelCase__ = 'gelu'
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=False , _a=99 , _a=32 , _a=2 , _a=4 , _a=37 , _a=0.1 , _a=0.1 , _a=20 , _a=2 , _a=1 , _a=0 , _a=4 , ):
__magic_name__ : int = parent
__magic_name__ : Optional[int] = batch_size
__magic_name__ : Tuple = seq_length
__magic_name__ : List[Any] = is_training
__magic_name__ : Dict = use_labels
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : int = hidden_size
__magic_name__ : Optional[int] = num_hidden_layers
__magic_name__ : Optional[int] = num_attention_heads
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Any = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Any = eos_token_id
__magic_name__ : str = pad_token_id
__magic_name__ : int = bos_token_id
__magic_name__ : Optional[int] = attention_window
# `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size
# [num_attention_heads, encoder_seq_length, encoder_key_length], but TFLongformerSelfAttention
# returns attention of shape [num_attention_heads, encoder_seq_length, self.attention_window + 1]
# because its local attention only attends to `self.attention_window` and one before and one after
__magic_name__ : Tuple = self.attention_window + 2
# because of padding `encoder_seq_length`, is different from `seq_length`. Relevant for
# the `test_attention_outputs` and `test_hidden_states_output` tests
__magic_name__ : Tuple = (
self.seq_length + (self.attention_window - self.seq_length % self.attention_window) % self.attention_window
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
__magic_name__ : int = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
__magic_name__ : int = tf.concat([input_ids, eos_tensor] , axis=1 )
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : Dict = self.config_cls(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_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_ids=[2] , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.pad_token_id , attention_window=self.attention_window , **self.config_updates , )
__magic_name__ : List[str] = prepare_led_inputs_dict(_a , _a , _a )
__magic_name__ : Union[str, Any] = tf.concat(
[tf.zeros_like(_a )[:, :-1], tf.ones_like(_a )[:, -1:]] , axis=-1 , )
__magic_name__ : List[Any] = global_attention_mask
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Dict = TFLEDModel(config=_a ).get_decoder()
__magic_name__ : Optional[int] = inputs_dict["input_ids"]
__magic_name__ : Union[str, Any] = input_ids[:1, :]
__magic_name__ : str = inputs_dict["attention_mask"][:1, :]
__magic_name__ : int = 1
# first forward pass
__magic_name__ : Tuple = model(_a , attention_mask=_a , use_cache=_a )
__magic_name__ , __magic_name__ : str = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
__magic_name__ : Optional[int] = ids_tensor((self.batch_size, 3) , config.vocab_size )
__magic_name__ : Any = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
__magic_name__ : Optional[Any] = tf.concat([input_ids, next_tokens] , axis=-1 )
__magic_name__ : List[Any] = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
__magic_name__ : List[str] = model(_a , attention_mask=_a )[0]
__magic_name__ : Dict = model(_a , attention_mask=_a , past_key_values=_a )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
__magic_name__ : List[Any] = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
__magic_name__ : Union[str, Any] = output_from_no_past[:, -3:, random_slice_idx]
__magic_name__ : List[str] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(_a , _a , rtol=1e-3 )
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : List[Any] , _snake_case : Any , _snake_case : str=None , _snake_case : List[str]=None , _snake_case : int=None , _snake_case : Any=None , ) -> int:
'''simple docstring'''
if attention_mask is None:
__magic_name__ : str = tf.cast(tf.math.not_equal(_snake_case , config.pad_token_id ) , tf.inta )
if decoder_attention_mask is None:
__magic_name__ : List[Any] = tf.concat(
[
tf.ones(decoder_input_ids[:, :1].shape , dtype=tf.inta ),
tf.cast(tf.math.not_equal(decoder_input_ids[:, 1:] , config.pad_token_id ) , tf.inta ),
] , axis=-1 , )
if head_mask is None:
__magic_name__ : Dict = tf.ones((config.encoder_layers, config.encoder_attention_heads) )
if decoder_head_mask is None:
__magic_name__ : Any = tf.ones((config.decoder_layers, config.decoder_attention_heads) )
return {
"input_ids": input_ids,
"attention_mask": attention_mask,
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
}
@require_tf
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (TFLEDForConditionalGeneration, TFLEDModel) if is_tf_available() else ()
UpperCamelCase__ = (TFLEDForConditionalGeneration,) if is_tf_available() else ()
UpperCamelCase__ = (
{
'conversational': TFLEDForConditionalGeneration,
'feature-extraction': TFLEDModel,
'summarization': TFLEDForConditionalGeneration,
'text2text-generation': TFLEDForConditionalGeneration,
'translation': TFLEDForConditionalGeneration,
}
if is_tf_available()
else {}
)
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = TFLEDModelTester(self )
__magic_name__ : List[Any] = ConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : int = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[str] = tf.zeros_like(inputs_dict["attention_mask"] )
__magic_name__ : Optional[Any] = 2
__magic_name__ : Tuple = tf.where(
tf.range(self.model_tester.seq_length )[None, :] < num_global_attn_indices , 1 , inputs_dict["global_attention_mask"] , )
__magic_name__ : Any = True
__magic_name__ : str = self.model_tester.seq_length
__magic_name__ : Dict = self.model_tester.encoder_seq_length
def check_decoder_attentions_output(_a ):
__magic_name__ : str = outputs.decoder_attentions
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(decoder_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
def check_encoder_attentions_output(_a ):
__magic_name__ : Any = [t.numpy() for t in outputs.encoder_attentions]
__magic_name__ : Tuple = [t.numpy() for t in outputs.encoder_global_attentions]
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
self.assertListEqual(
list(global_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, encoder_seq_length, num_global_attn_indices] , )
for model_class in self.all_model_classes:
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = False
__magic_name__ : Tuple = False
__magic_name__ : Optional[int] = model_class(_a )
__magic_name__ : str = model(self._prepare_for_class(_a , _a ) )
__magic_name__ : Any = len(_a )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
if self.is_encoder_decoder:
__magic_name__ : Tuple = model_class(_a )
__magic_name__ : Optional[Any] = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_decoder_attentions_output(_a )
# Check that output attentions can also be changed via the config
del inputs_dict["output_attentions"]
__magic_name__ : Dict = True
__magic_name__ : str = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
# Check attention is always last and order is fine
__magic_name__ : Union[str, Any] = True
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(out_len + (2 if self.is_encoder_decoder else 1) , len(_a ) )
self.assertEqual(model.config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
@unittest.skip("LED keeps using potentially symbolic tensors in conditionals and breaks tracing." )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
# TODO: Head-masking not yet implement
pass
def lowerCAmelCase_ ( _snake_case : int ) -> Optional[int]:
'''simple docstring'''
return tf.constant(_snake_case , dtype=tf.intaa )
snake_case : Optional[int] = 1E-4
@slow
@require_tf
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" ).led
# change to intended input here
__magic_name__ : Optional[int] = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : str = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Any = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : List[Any] = model(**_a )[0]
__magic_name__ : List[str] = (1, 1_024, 768)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : int = tf.convert_to_tensor(
[[2.30_50, 2.82_79, 0.65_31], [-1.84_57, -0.14_55, -3.56_61], [-1.01_86, 0.45_86, -2.20_43]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" )
# change to intended input here
__magic_name__ : int = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Tuple = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Optional[Any] = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : Union[str, Any] = model(**_a )[0]
__magic_name__ : Optional[int] = (1, 1_024, model.config.vocab_size)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : str = tf.convert_to_tensor(
[[33.65_07, 6.45_72, 16.80_89], [5.87_39, -2.42_38, 11.29_02], [-3.21_39, -4.31_49, 4.27_83]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 , rtol=1e-3 )
| 281 | 1 |
import inspect
import warnings
from typing import Any, Dict, Optional, Union
from packaging import version
def lowerCAmelCase_ ( *_snake_case : List[str] , _snake_case : Optional[Union[Dict, Any]] = None , _snake_case : Dict=True , _snake_case : Tuple=2 ) -> Any:
'''simple docstring'''
from .. import __version__
__magic_name__ : List[str] = take_from
__magic_name__ : int = ()
if not isinstance(args[0] , _snake_case ):
__magic_name__ : List[Any] = (args,)
for attribute, version_name, message in args:
if version.parse(version.parse(_snake_case ).base_version ) >= version.parse(_snake_case ):
raise ValueError(
F'''The deprecation tuple {(attribute, version_name, message)} should be removed since diffusers\''''
F''' version {__version__} is >= {version_name}''' )
__magic_name__ : Tuple = None
if isinstance(_snake_case , _snake_case ) and attribute in deprecated_kwargs:
values += (deprecated_kwargs.pop(_snake_case ),)
__magic_name__ : Union[str, Any] = F'''The `{attribute}` argument is deprecated and will be removed in version {version_name}.'''
elif hasattr(_snake_case , _snake_case ):
values += (getattr(_snake_case , _snake_case ),)
__magic_name__ : int = F'''The `{attribute}` attribute is deprecated and will be removed in version {version_name}.'''
elif deprecated_kwargs is None:
__magic_name__ : List[Any] = F'''`{attribute}` is deprecated and will be removed in version {version_name}.'''
if warning is not None:
__magic_name__ : Dict = warning + " " if standard_warn else ""
warnings.warn(warning + message , _snake_case , stacklevel=_snake_case )
if isinstance(_snake_case , _snake_case ) and len(_snake_case ) > 0:
__magic_name__ : str = inspect.getouterframes(inspect.currentframe() )[1]
__magic_name__ : Tuple = call_frame.filename
__magic_name__ : int = call_frame.lineno
__magic_name__ : int = call_frame.function
__magic_name__ , __magic_name__ : List[Any] = next(iter(deprecated_kwargs.items() ) )
raise TypeError(F'''{function} in {filename} line {line_number-1} got an unexpected keyword argument `{key}`''' )
if len(_snake_case ) == 0:
return
elif len(_snake_case ) == 1:
return values[0]
return values
| 281 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import (
BitConfig,
ViTHybridConfig,
ViTHybridForImageClassification,
ViTHybridImageProcessor,
ViTHybridModel,
)
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
snake_case : Optional[Any] = logging.get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any]=False ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
# fmt: off
# stem:
rename_keys.append(("cls_token", "vit.embeddings.cls_token") )
rename_keys.append(("pos_embed", "vit.embeddings.position_embeddings") )
rename_keys.append(("patch_embed.proj.weight", "vit.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("patch_embed.proj.bias", "vit.embeddings.patch_embeddings.projection.bias") )
# backbone
rename_keys.append(("patch_embed.backbone.stem.conv.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.convolution.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.bias", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.bias") )
for stage_idx in range(len(config.backbone_config.depths ) ):
for layer_idx in range(config.backbone_config.depths[stage_idx] ):
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.conv.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.conv.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.bias''') )
# transformer encoder
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''vit.encoder.layer.{i}.output.dense.bias''') )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("norm.weight", "layernorm.weight"),
("norm.bias", "layernorm.bias"),
("pre_logits.fc.weight", "pooler.dense.weight"),
("pre_logits.fc.bias", "pooler.dense.bias"),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
__magic_name__ : int = [(pair[0], pair[1][4:]) if pair[1].startswith("vit" ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
("norm.weight", "vit.layernorm.weight"),
("norm.bias", "vit.layernorm.bias"),
("head.weight", "classifier.weight"),
("head.bias", "classifier.bias"),
] )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Any , _snake_case : Dict=False ) -> int:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
__magic_name__ : int = ""
else:
__magic_name__ : Union[str, Any] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__magic_name__ : Optional[Any] = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
__magic_name__ : int = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__magic_name__ : Dict = in_proj_weight[
: config.hidden_size, :
]
__magic_name__ : List[str] = in_proj_bias[: config.hidden_size]
__magic_name__ : List[str] = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__magic_name__ : Optional[Any] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__magic_name__ : Optional[Any] = in_proj_weight[
-config.hidden_size :, :
]
__magic_name__ : int = in_proj_bias[-config.hidden_size :]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
__magic_name__ : List[str] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : int , _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : int = dct.pop(_snake_case )
__magic_name__ : List[Any] = val
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
__magic_name__ : List[str] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Any , _snake_case : int=False ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=_snake_case , )
__magic_name__ : List[str] = ViTHybridConfig(backbone_config=_snake_case , image_size=384 , num_labels=1000 )
__magic_name__ : str = False
# load original model from timm
__magic_name__ : Union[str, Any] = timm.create_model(_snake_case , pretrained=_snake_case )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__magic_name__ : List[Any] = timm_model.state_dict()
if base_model:
remove_classification_head_(_snake_case )
__magic_name__ : Tuple = create_rename_keys(_snake_case , _snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
read_in_q_k_v(_snake_case , _snake_case , _snake_case )
__magic_name__ : List[str] = "huggingface/label-files"
__magic_name__ : int = "imagenet-1k-id2label.json"
__magic_name__ : Optional[int] = json.load(open(hf_hub_download(_snake_case , _snake_case , repo_type="dataset" ) , "r" ) )
__magic_name__ : int = {int(_snake_case ): v for k, v in idalabel.items()}
__magic_name__ : List[str] = idalabel
__magic_name__ : List[str] = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
__magic_name__ : List[str] = ViTHybridModel(_snake_case ).eval()
else:
__magic_name__ : str = ViTHybridForImageClassification(_snake_case ).eval()
model.load_state_dict(_snake_case )
# create image processor
__magic_name__ : List[Any] = create_transform(**resolve_data_config({} , model=_snake_case ) )
__magic_name__ : int = transform.transforms
__magic_name__ : List[str] = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
__magic_name__ : int = ViTHybridImageProcessor(
do_resize=_snake_case , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=_snake_case , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=_snake_case , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
__magic_name__ : List[Any] = prepare_img()
__magic_name__ : Any = transform(_snake_case ).unsqueeze(0 )
__magic_name__ : Tuple = processor(_snake_case , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(_snake_case , _snake_case )
# verify logits
with torch.no_grad():
__magic_name__ : Optional[int] = model(_snake_case )
__magic_name__ : List[str] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
__magic_name__ : List[str] = timm_model.forward_features(_snake_case )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(_snake_case , outputs.pooler_output , atol=1E-3 )
else:
__magic_name__ : Any = timm_model(_snake_case )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_snake_case , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(_snake_case ).mkdir(exist_ok=_snake_case )
print(F'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(_snake_case )
print(F'''Saving processor to {pytorch_dump_folder_path}''' )
processor.save_pretrained(_snake_case )
if push_to_hub:
print(F'''Pushing model and processor to the hub {vit_name}''' )
model.push_to_hub(F'''ybelkada/{vit_name}''' )
processor.push_to_hub(F'''ybelkada/{vit_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_r50_s16_384",
type=str,
help="Name of the hybrid ViT timm model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether to upload the model to the HuggingFace hub."
)
snake_case : List[Any] = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : int , _snake_case : int ) -> int:
'''simple docstring'''
return int(input_a == input_a == 0 )
def lowerCAmelCase_ ( ) -> None:
'''simple docstring'''
print("Truth Table of NOR Gate:" )
print("| Input 1 | Input 2 | Output |" )
print(F'''| 0 | 0 | {nor_gate(0 , 0 )} |''' )
print(F'''| 0 | 1 | {nor_gate(0 , 1 )} |''' )
print(F'''| 1 | 0 | {nor_gate(1 , 0 )} |''' )
print(F'''| 1 | 1 | {nor_gate(1 , 1 )} |''' )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 281 |
# This script creates a super tiny model that is useful inside tests, when we just want to test that
# the machinery works, without needing to the check the quality of the outcomes.
#
# This version creates a tiny model through reduction of a normal pre-trained model, but keeping the
# full vocab, merges file, and thus also resulting in a larger model due to a large vocab size.
# This gives ~3MB in total for all files.
#
# If you want a 50 times smaller than this see `fsmt-make-super-tiny-model.py`, which is slightly more complicated
#
#
# It will be used then as "stas/tiny-wmt19-en-de"
# Build
from transformers import FSMTTokenizer, FSMTConfig, FSMTForConditionalGeneration
snake_case : List[str] = "facebook/wmt19-en-de"
snake_case : Dict = FSMTTokenizer.from_pretrained(mname)
# get the correct vocab sizes, etc. from the master model
snake_case : List[str] = FSMTConfig.from_pretrained(mname)
config.update(
dict(
d_model=4,
encoder_layers=1,
decoder_layers=1,
encoder_ffn_dim=4,
decoder_ffn_dim=4,
encoder_attention_heads=1,
decoder_attention_heads=1,
)
)
snake_case : int = FSMTForConditionalGeneration(config)
print(F"num of params {tiny_model.num_parameters()}")
# Test
snake_case : Optional[Any] = tokenizer(["Making tiny model"], return_tensors="pt")
snake_case : List[str] = tiny_model(**batch)
print("test output:", len(outputs.logits[0]))
# Save
snake_case : Dict = "tiny-wmt19-en-de"
tiny_model.half() # makes it smaller
tiny_model.save_pretrained(mname_tiny)
tokenizer.save_pretrained(mname_tiny)
print(F"Generated {mname_tiny}")
# Upload
# transformers-cli upload tiny-wmt19-en-de
| 281 | 1 |
import argparse
import json
import logging
import os
import shutil
import sys
import tempfile
import unittest
from unittest import mock
import torch
from accelerate.utils import write_basic_config
from transformers.testing_utils import TestCasePlus, get_gpu_count, run_command, slow, torch_device
from transformers.utils import is_apex_available
logging.basicConfig(level=logging.DEBUG)
snake_case : Union[str, Any] = logging.getLogger()
def lowerCAmelCase_ ( ) -> str:
'''simple docstring'''
__magic_name__ : Tuple = argparse.ArgumentParser()
parser.add_argument("-f" )
__magic_name__ : Any = parser.parse_args()
return args.f
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : str = {}
__magic_name__ : List[Any] = os.path.join(_snake_case , "all_results.json" )
if os.path.exists(_snake_case ):
with open(_snake_case , "r" ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
else:
raise ValueError(F'''can\'t find {path}''' )
return results
def lowerCAmelCase_ ( ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : List[str] = torch.cuda.is_available() and torch_device == "cuda"
return is_using_cuda and is_apex_available()
snake_case : Dict = logging.StreamHandler(sys.stdout)
logger.addHandler(stream_handler)
class _snake_case ( snake_case ):
@classmethod
def SCREAMING_SNAKE_CASE ( cls ):
# Write Accelerate config, will pick up on CPU, GPU, and multi-GPU
__magic_name__ : Any = tempfile.mkdtemp()
__magic_name__ : List[Any] = os.path.join(cls.tmpdir , "default_config.yml" )
write_basic_config(save_location=cls.configPath )
__magic_name__ : Any = ["accelerate", "launch", "--config_file", cls.configPath]
@classmethod
def SCREAMING_SNAKE_CASE ( cls ):
shutil.rmtree(cls.tmpdir )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = self.get_auto_remove_tmp_dir()
__magic_name__ : str = f'''
{self.examples_dir}/pytorch/text-classification/run_glue_no_trainer.py
--model_name_or_path distilbert-base-uncased
--output_dir {tmp_dir}
--train_file ./tests/fixtures/tests_samples/MRPC/train.csv
--validation_file ./tests/fixtures/tests_samples/MRPC/dev.csv
--per_device_train_batch_size=2
--per_device_eval_batch_size=1
--learning_rate=1e-4
--seed=42
--checkpointing_steps epoch
--with_tracking
'''.split()
if is_cuda_and_apex_available():
testargs.append("--fp16" )
run_command(self._launch_args + testargs )
__magic_name__ : Optional[Any] = get_results(_a )
self.assertGreaterEqual(result["eval_accuracy"] , 0.75 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "glue_no_trainer" ) ) )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.get_auto_remove_tmp_dir()
__magic_name__ : List[str] = f'''
{self.examples_dir}/pytorch/language-modeling/run_clm_no_trainer.py
--model_name_or_path distilgpt2
--train_file ./tests/fixtures/sample_text.txt
--validation_file ./tests/fixtures/sample_text.txt
--block_size 128
--per_device_train_batch_size 5
--per_device_eval_batch_size 5
--num_train_epochs 2
--output_dir {tmp_dir}
--checkpointing_steps epoch
--with_tracking
'''.split()
if torch.cuda.device_count() > 1:
# Skipping because there are not enough batches to train the model + would need a drop_last to work.
return
run_command(self._launch_args + testargs )
__magic_name__ : Union[str, Any] = get_results(_a )
self.assertLess(result["perplexity"] , 100 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "clm_no_trainer" ) ) )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = self.get_auto_remove_tmp_dir()
__magic_name__ : Dict = f'''
{self.examples_dir}/pytorch/language-modeling/run_mlm_no_trainer.py
--model_name_or_path distilroberta-base
--train_file ./tests/fixtures/sample_text.txt
--validation_file ./tests/fixtures/sample_text.txt
--output_dir {tmp_dir}
--num_train_epochs=1
--checkpointing_steps epoch
--with_tracking
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : Tuple = get_results(_a )
self.assertLess(result["perplexity"] , 42 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "mlm_no_trainer" ) ) )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
# with so little data distributed training needs more epochs to get the score on par with 0/1 gpu
__magic_name__ : Union[str, Any] = 7 if get_gpu_count() > 1 else 2
__magic_name__ : Optional[Any] = self.get_auto_remove_tmp_dir()
__magic_name__ : Union[str, Any] = f'''
{self.examples_dir}/pytorch/token-classification/run_ner_no_trainer.py
--model_name_or_path bert-base-uncased
--train_file tests/fixtures/tests_samples/conll/sample.json
--validation_file tests/fixtures/tests_samples/conll/sample.json
--output_dir {tmp_dir}
--learning_rate=2e-4
--per_device_train_batch_size=2
--per_device_eval_batch_size=2
--num_train_epochs={epochs}
--seed 7
--checkpointing_steps epoch
--with_tracking
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : List[Any] = get_results(_a )
self.assertGreaterEqual(result["eval_accuracy"] , 0.75 )
self.assertLess(result["train_loss"] , 0.5 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "ner_no_trainer" ) ) )
@unittest.skip(reason="Fix me @muellerzr" )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.get_auto_remove_tmp_dir()
__magic_name__ : List[Any] = f'''
{self.examples_dir}/pytorch/question-answering/run_qa_no_trainer.py
--model_name_or_path bert-base-uncased
--version_2_with_negative
--train_file tests/fixtures/tests_samples/SQUAD/sample.json
--validation_file tests/fixtures/tests_samples/SQUAD/sample.json
--output_dir {tmp_dir}
--seed=42
--max_train_steps=10
--num_warmup_steps=2
--learning_rate=2e-4
--per_device_train_batch_size=2
--per_device_eval_batch_size=1
--checkpointing_steps epoch
--with_tracking
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : Optional[Any] = get_results(_a )
# Because we use --version_2_with_negative the testing script uses SQuAD v2 metrics.
self.assertGreaterEqual(result["eval_f1"] , 28 )
self.assertGreaterEqual(result["eval_exact"] , 28 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "qa_no_trainer" ) ) )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = self.get_auto_remove_tmp_dir()
__magic_name__ : str = f'''
{self.examples_dir}/pytorch/multiple-choice/run_swag_no_trainer.py
--model_name_or_path bert-base-uncased
--train_file tests/fixtures/tests_samples/swag/sample.json
--validation_file tests/fixtures/tests_samples/swag/sample.json
--output_dir {tmp_dir}
--max_train_steps=20
--num_warmup_steps=2
--learning_rate=2e-4
--per_device_train_batch_size=2
--per_device_eval_batch_size=1
--with_tracking
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : Tuple = get_results(_a )
self.assertGreaterEqual(result["eval_accuracy"] , 0.8 )
self.assertTrue(os.path.exists(os.path.join(_a , "swag_no_trainer" ) ) )
@slow
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = self.get_auto_remove_tmp_dir()
__magic_name__ : Tuple = f'''
{self.examples_dir}/pytorch/summarization/run_summarization_no_trainer.py
--model_name_or_path t5-small
--train_file tests/fixtures/tests_samples/xsum/sample.json
--validation_file tests/fixtures/tests_samples/xsum/sample.json
--output_dir {tmp_dir}
--max_train_steps=50
--num_warmup_steps=8
--learning_rate=2e-4
--per_device_train_batch_size=2
--per_device_eval_batch_size=1
--checkpointing_steps epoch
--with_tracking
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : List[Any] = get_results(_a )
self.assertGreaterEqual(result["eval_rouge1"] , 10 )
self.assertGreaterEqual(result["eval_rouge2"] , 2 )
self.assertGreaterEqual(result["eval_rougeL"] , 7 )
self.assertGreaterEqual(result["eval_rougeLsum"] , 7 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "summarization_no_trainer" ) ) )
@slow
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.get_auto_remove_tmp_dir()
__magic_name__ : List[str] = f'''
{self.examples_dir}/pytorch/translation/run_translation_no_trainer.py
--model_name_or_path sshleifer/student_marian_en_ro_6_1
--source_lang en
--target_lang ro
--train_file tests/fixtures/tests_samples/wmt16/sample.json
--validation_file tests/fixtures/tests_samples/wmt16/sample.json
--output_dir {tmp_dir}
--max_train_steps=50
--num_warmup_steps=8
--num_beams=6
--learning_rate=3e-3
--per_device_train_batch_size=2
--per_device_eval_batch_size=1
--source_lang en_XX
--target_lang ro_RO
--checkpointing_steps epoch
--with_tracking
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : Dict = get_results(_a )
self.assertGreaterEqual(result["eval_bleu"] , 30 )
self.assertTrue(os.path.exists(os.path.join(_a , "epoch_0" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "translation_no_trainer" ) ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = logging.StreamHandler(sys.stdout )
logger.addHandler(_a )
__magic_name__ : str = self.get_auto_remove_tmp_dir()
__magic_name__ : int = f'''
{self.examples_dir}/pytorch/semantic-segmentation/run_semantic_segmentation_no_trainer.py
--dataset_name huggingface/semantic-segmentation-test-sample
--output_dir {tmp_dir}
--max_train_steps=10
--num_warmup_steps=2
--learning_rate=2e-4
--per_device_train_batch_size=2
--per_device_eval_batch_size=1
--checkpointing_steps epoch
'''.split()
run_command(self._launch_args + testargs )
__magic_name__ : Union[str, Any] = get_results(_a )
self.assertGreaterEqual(result["eval_overall_accuracy"] , 0.10 )
@mock.patch.dict(os.environ , {"WANDB_MODE": "offline"} )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = self.get_auto_remove_tmp_dir()
__magic_name__ : Tuple = f'''
{self.examples_dir}/pytorch/image-classification/run_image_classification_no_trainer.py
--model_name_or_path google/vit-base-patch16-224-in21k
--dataset_name hf-internal-testing/cats_vs_dogs_sample
--learning_rate 1e-4
--per_device_train_batch_size 2
--per_device_eval_batch_size 1
--max_train_steps 2
--train_val_split 0.1
--seed 42
--output_dir {tmp_dir}
--with_tracking
--checkpointing_steps 1
'''.split()
if is_cuda_and_apex_available():
testargs.append("--fp16" )
run_command(self._launch_args + testargs )
__magic_name__ : List[str] = get_results(_a )
# The base model scores a 25%
self.assertGreaterEqual(result["eval_accuracy"] , 0.6 )
self.assertTrue(os.path.exists(os.path.join(_a , "step_1" ) ) )
self.assertTrue(os.path.exists(os.path.join(_a , "image_classification_no_trainer" ) ) )
| 281 |
import argparse
import csv
import logging
import os
import random
import numpy as np
import torch
from torch.utils.data import DataLoader, RandomSampler, SequentialSampler, TensorDataset
from tqdm import tqdm, trange
from transformers import (
CONFIG_NAME,
WEIGHTS_NAME,
AdamW,
OpenAIGPTDoubleHeadsModel,
OpenAIGPTTokenizer,
get_linear_schedule_with_warmup,
)
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s", datefmt="%m/%d/%Y %H:%M:%S", level=logging.INFO
)
snake_case : Optional[int] = logging.getLogger(__name__)
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : List[str] = np.argmax(_snake_case , axis=1 )
return np.sum(outputs == labels )
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
with open(_snake_case , encoding="utf_8" ) as f:
__magic_name__ : List[str] = csv.reader(_snake_case )
__magic_name__ : List[Any] = []
next(_snake_case ) # skip the first line
for line in tqdm(_snake_case ):
output.append((" ".join(line[1:5] ), line[5], line[6], int(line[-1] ) - 1) )
return output
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Tuple , _snake_case : Union[str, Any] , _snake_case : List[Any] , _snake_case : Tuple , _snake_case : Optional[int] ) -> int:
'''simple docstring'''
__magic_name__ : Optional[int] = []
for dataset in encoded_datasets:
__magic_name__ : Union[str, Any] = len(_snake_case )
__magic_name__ : Dict = np.zeros((n_batch, 2, input_len) , dtype=np.intaa )
__magic_name__ : List[str] = np.zeros((n_batch, 2) , dtype=np.intaa )
__magic_name__ : Optional[int] = np.full((n_batch, 2, input_len) , fill_value=-100 , dtype=np.intaa )
__magic_name__ : int = np.zeros((n_batch,) , dtype=np.intaa )
for (
i,
(story, conta, conta, mc_label),
) in enumerate(_snake_case ):
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : str = with_conta
__magic_name__ : Tuple = with_conta
__magic_name__ : Union[str, Any] = len(_snake_case ) - 1
__magic_name__ : int = len(_snake_case ) - 1
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[int] = mc_label
__magic_name__ : str = (input_ids, mc_token_ids, lm_labels, mc_labels)
tensor_datasets.append(tuple(torch.tensor(_snake_case ) for t in all_inputs ) )
return tensor_datasets
def lowerCAmelCase_ ( ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser()
parser.add_argument("--model_name" , type=_snake_case , default="openai-gpt" , help="pretrained model name" )
parser.add_argument("--do_train" , action="store_true" , help="Whether to run training." )
parser.add_argument("--do_eval" , action="store_true" , help="Whether to run eval on the dev set." )
parser.add_argument(
"--output_dir" , default=_snake_case , type=_snake_case , required=_snake_case , help="The output directory where the model predictions and checkpoints will be written." , )
parser.add_argument("--train_dataset" , type=_snake_case , default="" )
parser.add_argument("--eval_dataset" , type=_snake_case , default="" )
parser.add_argument("--seed" , type=_snake_case , default=42 )
parser.add_argument("--num_train_epochs" , type=_snake_case , default=3 )
parser.add_argument("--train_batch_size" , type=_snake_case , default=8 )
parser.add_argument("--eval_batch_size" , type=_snake_case , default=16 )
parser.add_argument("--adam_epsilon" , default=1E-8 , type=_snake_case , help="Epsilon for Adam optimizer." )
parser.add_argument("--max_grad_norm" , type=_snake_case , default=1 )
parser.add_argument(
"--max_steps" , default=-1 , type=_snake_case , help=(
"If > 0: set total number of training steps to perform. Override num_train_epochs."
) , )
parser.add_argument(
"--gradient_accumulation_steps" , type=_snake_case , default=1 , help="Number of updates steps to accumulate before performing a backward/update pass." , )
parser.add_argument("--learning_rate" , type=_snake_case , default=6.25E-5 )
parser.add_argument("--warmup_steps" , default=0 , type=_snake_case , help="Linear warmup over warmup_steps." )
parser.add_argument("--lr_schedule" , type=_snake_case , default="warmup_linear" )
parser.add_argument("--weight_decay" , type=_snake_case , default=0.01 )
parser.add_argument("--lm_coef" , type=_snake_case , default=0.9 )
parser.add_argument("--n_valid" , type=_snake_case , default=374 )
parser.add_argument("--server_ip" , type=_snake_case , default="" , help="Can be used for distant debugging." )
parser.add_argument("--server_port" , type=_snake_case , default="" , help="Can be used for distant debugging." )
__magic_name__ : List[Any] = parser.parse_args()
print(_snake_case )
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach" )
ptvsd.enable_attach(address=(args.server_ip, args.server_port) , redirect_output=_snake_case )
ptvsd.wait_for_attach()
random.seed(args.seed )
np.random.seed(args.seed )
torch.manual_seed(args.seed )
torch.cuda.manual_seed_all(args.seed )
__magic_name__ : Dict = torch.device("cuda" if torch.cuda.is_available() else "cpu" )
__magic_name__ : Optional[int] = torch.cuda.device_count()
logger.info("device: {}, n_gpu {}".format(_snake_case , _snake_case ) )
if not args.do_train and not args.do_eval:
raise ValueError("At least one of `do_train` or `do_eval` must be True." )
if not os.path.exists(args.output_dir ):
os.makedirs(args.output_dir )
# Load tokenizer and model
# This loading functions also add new tokens and embeddings called `special tokens`
# These new embeddings will be fine-tuned on the RocStories dataset
__magic_name__ : List[Any] = ["_start_", "_delimiter_", "_classify_"]
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.model_name )
tokenizer.add_tokens(_snake_case )
__magic_name__ : Optional[Any] = tokenizer.convert_tokens_to_ids(_snake_case )
__magic_name__ : List[str] = OpenAIGPTDoubleHeadsModel.from_pretrained(args.model_name )
model.resize_token_embeddings(len(_snake_case ) )
model.to(_snake_case )
# Load and encode the datasets
def tokenize_and_encode(_snake_case : str ):
if isinstance(_snake_case , _snake_case ):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(_snake_case ) )
elif isinstance(_snake_case , _snake_case ):
return obj
return [tokenize_and_encode(_snake_case ) for o in obj]
logger.info("Encoding dataset..." )
__magic_name__ : Optional[int] = load_rocstories_dataset(args.train_dataset )
__magic_name__ : str = load_rocstories_dataset(args.eval_dataset )
__magic_name__ : int = (train_dataset, eval_dataset)
__magic_name__ : List[str] = tokenize_and_encode(_snake_case )
# Compute the max input length for the Transformer
__magic_name__ : Optional[Any] = model.config.n_positions // 2 - 2
__magic_name__ : Optional[int] = max(
len(story[:max_length] ) + max(len(conta[:max_length] ) , len(conta[:max_length] ) ) + 3
for dataset in encoded_datasets
for story, conta, conta, _ in dataset )
__magic_name__ : List[str] = min(_snake_case , model.config.n_positions ) # Max size of input for the pre-trained model
# Prepare inputs tensors and dataloaders
__magic_name__ : List[Any] = pre_process_datasets(_snake_case , _snake_case , _snake_case , *_snake_case )
__magic_name__ , __magic_name__ : Optional[int] = tensor_datasets[0], tensor_datasets[1]
__magic_name__ : Tuple = TensorDataset(*_snake_case )
__magic_name__ : Union[str, Any] = RandomSampler(_snake_case )
__magic_name__ : Dict = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.train_batch_size )
__magic_name__ : Any = TensorDataset(*_snake_case )
__magic_name__ : Optional[Any] = SequentialSampler(_snake_case )
__magic_name__ : int = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.eval_batch_size )
# Prepare optimizer
if args.do_train:
if args.max_steps > 0:
__magic_name__ : Tuple = args.max_steps
__magic_name__ : List[str] = args.max_steps // (len(_snake_case ) // args.gradient_accumulation_steps) + 1
else:
__magic_name__ : List[str] = len(_snake_case ) // args.gradient_accumulation_steps * args.num_train_epochs
__magic_name__ : str = list(model.named_parameters() )
__magic_name__ : Dict = ["bias", "LayerNorm.bias", "LayerNorm.weight"]
__magic_name__ : str = [
{
"params": [p for n, p in param_optimizer if not any(nd in n for nd in no_decay )],
"weight_decay": args.weight_decay,
},
{"params": [p for n, p in param_optimizer if any(nd in n for nd in no_decay )], "weight_decay": 0.0},
]
__magic_name__ : str = AdamW(_snake_case , lr=args.learning_rate , eps=args.adam_epsilon )
__magic_name__ : List[str] = get_linear_schedule_with_warmup(
_snake_case , num_warmup_steps=args.warmup_steps , num_training_steps=_snake_case )
if args.do_train:
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = 0, 0, None
model.train()
for _ in trange(int(args.num_train_epochs ) , desc="Epoch" ):
__magic_name__ : List[str] = 0
__magic_name__ : Tuple = 0
__magic_name__ : Dict = tqdm(_snake_case , desc="Training" )
for step, batch in enumerate(_snake_case ):
__magic_name__ : Optional[Any] = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = batch
__magic_name__ : Optional[Any] = model(_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Optional[Any] = args.lm_coef * losses[0] + losses[1]
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
tr_loss += loss.item()
__magic_name__ : List[str] = (
loss.item() if exp_average_loss is None else 0.7 * exp_average_loss + 0.3 * loss.item()
)
nb_tr_steps += 1
__magic_name__ : int = "Training loss: {:.2e} lr: {:.2e}".format(_snake_case , scheduler.get_lr()[0] )
# Save a trained model
if args.do_train:
# Save a trained model, configuration and tokenizer
__magic_name__ : Dict = model.module if hasattr(_snake_case , "module" ) else model # Only save the model itself
# If we save using the predefined names, we can load using `from_pretrained`
__magic_name__ : List[Any] = os.path.join(args.output_dir , _snake_case )
__magic_name__ : Dict = os.path.join(args.output_dir , _snake_case )
torch.save(model_to_save.state_dict() , _snake_case )
model_to_save.config.to_json_file(_snake_case )
tokenizer.save_vocabulary(args.output_dir )
# Load a trained model and vocabulary that you have fine-tuned
__magic_name__ : Dict = OpenAIGPTDoubleHeadsModel.from_pretrained(args.output_dir )
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.output_dir )
model.to(_snake_case )
if args.do_eval:
model.eval()
__magic_name__ , __magic_name__ : Any = 0, 0
__magic_name__ , __magic_name__ : Union[str, Any] = 0, 0
for batch in tqdm(_snake_case , desc="Evaluating" ):
__magic_name__ : int = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = batch
with torch.no_grad():
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = model(
_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Tuple = mc_logits.detach().cpu().numpy()
__magic_name__ : Any = mc_labels.to("cpu" ).numpy()
__magic_name__ : str = accuracy(_snake_case , _snake_case )
eval_loss += mc_loss.mean().item()
eval_accuracy += tmp_eval_accuracy
nb_eval_examples += input_ids.size(0 )
nb_eval_steps += 1
__magic_name__ : Tuple = eval_loss / nb_eval_steps
__magic_name__ : List[Any] = eval_accuracy / nb_eval_examples
__magic_name__ : int = tr_loss / nb_tr_steps if args.do_train else None
__magic_name__ : Any = {"eval_loss": eval_loss, "eval_accuracy": eval_accuracy, "train_loss": train_loss}
__magic_name__ : int = os.path.join(args.output_dir , "eval_results.txt" )
with open(_snake_case , "w" ) as writer:
logger.info("***** Eval results *****" )
for key in sorted(result.keys() ):
logger.info(" %s = %s" , _snake_case , str(result[key] ) )
writer.write("%s = %s\n" % (key, str(result[key] )) )
if __name__ == "__main__":
main()
| 281 | 1 |
import sys
from collections import defaultdict
class _snake_case :
def __init__( self ):
__magic_name__ : Union[str, Any] = []
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.node_position[vertex]
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : str = pos
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a ):
if start > size // 2 - 1:
return
else:
if 2 * start + 2 >= size:
__magic_name__ : Any = 2 * start + 1
else:
if heap[2 * start + 1] < heap[2 * start + 2]:
__magic_name__ : Optional[Any] = 2 * start + 1
else:
__magic_name__ : Dict = 2 * start + 2
if heap[smallest_child] < heap[start]:
__magic_name__ , __magic_name__ : Any = heap[smallest_child], positions[smallest_child]
__magic_name__ , __magic_name__ : Dict = (
heap[start],
positions[start],
)
__magic_name__ , __magic_name__ : Union[str, Any] = temp, tempa
__magic_name__ : Optional[Any] = self.get_position(positions[smallest_child] )
self.set_position(
positions[smallest_child] , self.get_position(positions[start] ) )
self.set_position(positions[start] , _a )
self.top_to_bottom(_a , _a , _a , _a )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a ):
__magic_name__ : int = position[index]
while index != 0:
__magic_name__ : Tuple = int((index - 2) / 2 ) if index % 2 == 0 else int((index - 1) / 2 )
if val < heap[parent]:
__magic_name__ : Optional[int] = heap[parent]
__magic_name__ : List[str] = position[parent]
self.set_position(position[parent] , _a )
else:
__magic_name__ : Dict = val
__magic_name__ : Optional[Any] = temp
self.set_position(_a , _a )
break
__magic_name__ : List[Any] = parent
else:
__magic_name__ : Union[str, Any] = val
__magic_name__ : Optional[int] = temp
self.set_position(_a , 0 )
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Tuple = len(_a ) // 2 - 1
for i in range(_a , -1 , -1 ):
self.top_to_bottom(_a , _a , len(_a ) , _a )
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : List[Any] = positions[0]
__magic_name__ : List[str] = sys.maxsize
self.top_to_bottom(_a , 0 , len(_a ) , _a )
return temp
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = Heap()
__magic_name__ : str = [0] * len(_snake_case )
__magic_name__ : Optional[Any] = [-1] * len(_snake_case ) # Neighboring Tree Vertex of selected vertex
# Minimum Distance of explored vertex with neighboring vertex of partial tree
# formed in graph
__magic_name__ : Optional[Any] = [] # Heap of Distance of vertices from their neighboring vertex
__magic_name__ : List[Any] = []
for vertex in range(len(_snake_case ) ):
distance_tv.append(sys.maxsize )
positions.append(_snake_case )
heap.node_position.append(_snake_case )
__magic_name__ : Any = []
__magic_name__ : int = 1
__magic_name__ : Tuple = sys.maxsize
for neighbor, distance in adjacency_list[0]:
__magic_name__ : List[str] = 0
__magic_name__ : Union[str, Any] = distance
heap.heapify(_snake_case , _snake_case )
for _ in range(1 , len(_snake_case ) ):
__magic_name__ : str = heap.delete_minimum(_snake_case , _snake_case )
if visited[vertex] == 0:
tree_edges.append((nbr_tv[vertex], vertex) )
__magic_name__ : int = 1
for neighbor, distance in adjacency_list[vertex]:
if (
visited[neighbor] == 0
and distance < distance_tv[heap.get_position(_snake_case )]
):
__magic_name__ : str = distance
heap.bottom_to_top(
_snake_case , heap.get_position(_snake_case ) , _snake_case , _snake_case )
__magic_name__ : List[Any] = vertex
return tree_edges
if __name__ == "__main__": # pragma: no cover
# < --------- Prims Algorithm --------- >
snake_case : Any = int(input("Enter number of edges: ").strip())
snake_case : List[Any] = defaultdict(list)
for _ in range(edges_number):
snake_case : List[str] = [int(x) for x in input().strip().split()]
adjacency_list[edge[0]].append([edge[1], edge[2]])
adjacency_list[edge[1]].append([edge[0], edge[2]])
print(prisms_algorithm(adjacency_list))
| 281 |
from . import __version__
# Backward compatibility imports, to make sure all those objects can be found in file_utils
from .utils import (
CLOUDFRONT_DISTRIB_PREFIX,
CONFIG_NAME,
DISABLE_TELEMETRY,
DUMMY_INPUTS,
DUMMY_MASK,
ENV_VARS_TRUE_AND_AUTO_VALUES,
ENV_VARS_TRUE_VALUES,
FEATURE_EXTRACTOR_NAME,
FLAX_WEIGHTS_NAME,
HF_MODULES_CACHE,
HUGGINGFACE_CO_PREFIX,
HUGGINGFACE_CO_RESOLVE_ENDPOINT,
MODEL_CARD_NAME,
MULTIPLE_CHOICE_DUMMY_INPUTS,
PYTORCH_PRETRAINED_BERT_CACHE,
PYTORCH_TRANSFORMERS_CACHE,
S3_BUCKET_PREFIX,
SENTENCEPIECE_UNDERLINE,
SPIECE_UNDERLINE,
TF2_WEIGHTS_NAME,
TF_WEIGHTS_NAME,
TORCH_FX_REQUIRED_VERSION,
TRANSFORMERS_CACHE,
TRANSFORMERS_DYNAMIC_MODULE_NAME,
USE_JAX,
USE_TF,
USE_TORCH,
WEIGHTS_INDEX_NAME,
WEIGHTS_NAME,
ContextManagers,
DummyObject,
EntryNotFoundError,
ExplicitEnum,
ModelOutput,
PaddingStrategy,
PushToHubMixin,
RepositoryNotFoundError,
RevisionNotFoundError,
TensorType,
_LazyModule,
add_code_sample_docstrings,
add_end_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
cached_property,
copy_func,
default_cache_path,
define_sagemaker_information,
get_cached_models,
get_file_from_repo,
get_full_repo_name,
get_torch_version,
has_file,
http_user_agent,
is_apex_available,
is_bsa_available,
is_coloredlogs_available,
is_datasets_available,
is_detectrona_available,
is_faiss_available,
is_flax_available,
is_ftfy_available,
is_in_notebook,
is_ipex_available,
is_librosa_available,
is_offline_mode,
is_onnx_available,
is_pandas_available,
is_phonemizer_available,
is_protobuf_available,
is_psutil_available,
is_pyanvml_available,
is_pyctcdecode_available,
is_pytesseract_available,
is_pytorch_quantization_available,
is_rjieba_available,
is_sagemaker_dp_enabled,
is_sagemaker_mp_enabled,
is_scipy_available,
is_sentencepiece_available,
is_seqio_available,
is_sklearn_available,
is_soundfile_availble,
is_spacy_available,
is_speech_available,
is_tensor,
is_tensorflow_probability_available,
is_tfaonnx_available,
is_tf_available,
is_timm_available,
is_tokenizers_available,
is_torch_available,
is_torch_bfaa_available,
is_torch_cuda_available,
is_torch_fx_available,
is_torch_fx_proxy,
is_torch_mps_available,
is_torch_tfaa_available,
is_torch_tpu_available,
is_torchaudio_available,
is_training_run_on_sagemaker,
is_vision_available,
replace_return_docstrings,
requires_backends,
to_numpy,
to_py_obj,
torch_only_method,
)
| 281 | 1 |
import unittest
import numpy as np
import requests
from transformers.testing_utils import require_torch, require_vision
from transformers.utils import is_torch_available, is_vision_available
from ...test_image_processing_common import ImageProcessingSavingTestMixin, prepare_image_inputs
if is_torch_available():
import torch
from transformers.pytorch_utils import is_torch_greater_or_equal_than_1_11
else:
snake_case : List[str] = False
if is_vision_available():
from PIL import Image
from transformers import PixaStructImageProcessor
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=7 , _a=3 , _a=18 , _a=30 , _a=400 , _a=None , _a=True , _a=True , _a=None , ):
__magic_name__ : List[Any] = size if size is not None else {"height": 20, "width": 20}
__magic_name__ : List[str] = parent
__magic_name__ : Any = batch_size
__magic_name__ : Any = num_channels
__magic_name__ : Tuple = image_size
__magic_name__ : Any = min_resolution
__magic_name__ : Union[str, Any] = max_resolution
__magic_name__ : Tuple = size
__magic_name__ : Any = do_normalize
__magic_name__ : List[Any] = do_convert_rgb
__magic_name__ : Dict = [512, 1_024, 2_048, 4_096]
__magic_name__ : List[Any] = patch_size if patch_size is not None else {"height": 16, "width": 16}
def SCREAMING_SNAKE_CASE ( self ):
return {"do_normalize": self.do_normalize, "do_convert_rgb": self.do_convert_rgb}
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/australia.jpg"
__magic_name__ : Optional[Any] = Image.open(requests.get(_a , stream=_a ).raw ).convert("RGB" )
return raw_image
@unittest.skipIf(
not is_torch_greater_or_equal_than_1_11 , reason='`Pix2StructImageProcessor` requires `torch>=1.11.0`.' , )
@require_torch
@require_vision
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = PixaStructImageProcessor if is_vision_available() else None
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = PixaStructImageProcessingTester(self )
@property
def SCREAMING_SNAKE_CASE ( self ):
return self.image_processor_tester.prepare_image_processor_dict()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(_a , "do_normalize" ) )
self.assertTrue(hasattr(_a , "do_convert_rgb" ) )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = self.image_processor_tester.prepare_dummy_image()
__magic_name__ : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
__magic_name__ : List[Any] = 2_048
__magic_name__ : Union[str, Any] = image_processor(_a , return_tensors="pt" , max_patches=_a )
self.assertTrue(torch.allclose(inputs.flattened_patches.mean() , torch.tensor(0.06_06 ) , atol=1e-3 , rtol=1e-3 ) )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processor
__magic_name__ : Union[str, Any] = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__magic_name__ : str = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a )
for image in image_inputs:
self.assertIsInstance(_a , Image.Image )
# Test not batched input
__magic_name__ : Any = (
(self.image_processor_tester.patch_size["height"] * self.image_processor_tester.patch_size["width"])
* self.image_processor_tester.num_channels
) + 2
for max_patch in self.image_processor_tester.max_patches:
# Test not batched input
__magic_name__ : Dict = image_processor(
image_inputs[0] , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (1, max_patch, expected_hidden_dim) , )
# Test batched
__magic_name__ : int = image_processor(
_a , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processor
__magic_name__ : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__magic_name__ : int = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a )
for image in image_inputs:
self.assertIsInstance(_a , Image.Image )
# Test not batched input
__magic_name__ : List[Any] = (
(self.image_processor_tester.patch_size["height"] * self.image_processor_tester.patch_size["width"])
* self.image_processor_tester.num_channels
) + 2
__magic_name__ : Any = True
for max_patch in self.image_processor_tester.max_patches:
# Test not batched input
with self.assertRaises(_a ):
__magic_name__ : str = image_processor(
image_inputs[0] , return_tensors="pt" , max_patches=_a ).flattened_patches
__magic_name__ : List[str] = "Hello"
__magic_name__ : Tuple = image_processor(
image_inputs[0] , return_tensors="pt" , max_patches=_a , header_text=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (1, max_patch, expected_hidden_dim) , )
# Test batched
__magic_name__ : List[str] = image_processor(
_a , return_tensors="pt" , max_patches=_a , header_text=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processor
__magic_name__ : Dict = self.image_processing_class(**self.image_processor_dict )
# create random numpy tensors
__magic_name__ : int = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a , numpify=_a )
for image in image_inputs:
self.assertIsInstance(_a , np.ndarray )
__magic_name__ : Any = (
(self.image_processor_tester.patch_size["height"] * self.image_processor_tester.patch_size["width"])
* self.image_processor_tester.num_channels
) + 2
for max_patch in self.image_processor_tester.max_patches:
# Test not batched input
__magic_name__ : Optional[Any] = image_processor(
image_inputs[0] , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (1, max_patch, expected_hidden_dim) , )
# Test batched
__magic_name__ : Union[str, Any] = image_processor(
_a , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processor
__magic_name__ : Tuple = self.image_processing_class(**self.image_processor_dict )
# create random PyTorch tensors
__magic_name__ : Tuple = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a , torchify=_a )
for image in image_inputs:
self.assertIsInstance(_a , torch.Tensor )
# Test not batched input
__magic_name__ : Tuple = (
(self.image_processor_tester.patch_size["height"] * self.image_processor_tester.patch_size["width"])
* self.image_processor_tester.num_channels
) + 2
for max_patch in self.image_processor_tester.max_patches:
# Test not batched input
__magic_name__ : Dict = image_processor(
image_inputs[0] , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (1, max_patch, expected_hidden_dim) , )
# Test batched
__magic_name__ : int = image_processor(
_a , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
@unittest.skipIf(
not is_torch_greater_or_equal_than_1_11 , reason='`Pix2StructImageProcessor` requires `torch>=1.11.0`.' , )
@require_torch
@require_vision
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = PixaStructImageProcessor if is_vision_available() else None
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = PixaStructImageProcessingTester(self , num_channels=4 )
__magic_name__ : Union[str, Any] = 3
@property
def SCREAMING_SNAKE_CASE ( self ):
return self.image_processor_tester.prepare_image_processor_dict()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = self.image_processing_class(**self.image_processor_dict )
self.assertTrue(hasattr(_a , "do_normalize" ) )
self.assertTrue(hasattr(_a , "do_convert_rgb" ) )
def SCREAMING_SNAKE_CASE ( self ):
# Initialize image_processor
__magic_name__ : Dict = self.image_processing_class(**self.image_processor_dict )
# create random PIL images
__magic_name__ : int = prepare_image_inputs(self.image_processor_tester , equal_resolution=_a )
for image in image_inputs:
self.assertIsInstance(_a , Image.Image )
# Test not batched input
__magic_name__ : Tuple = (
(self.image_processor_tester.patch_size["height"] * self.image_processor_tester.patch_size["width"])
* (self.image_processor_tester.num_channels - 1)
) + 2
for max_patch in self.image_processor_tester.max_patches:
# Test not batched input
__magic_name__ : Optional[Any] = image_processor(
image_inputs[0] , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (1, max_patch, expected_hidden_dim) , )
# Test batched
__magic_name__ : Union[str, Any] = image_processor(
_a , return_tensors="pt" , max_patches=_a ).flattened_patches
self.assertEqual(
encoded_images.shape , (self.image_processor_tester.batch_size, max_patch, expected_hidden_dim) , )
| 281 |
import importlib
import os
import fsspec
import pytest
from fsspec import register_implementation
from fsspec.registry import _registry as _fsspec_registry
from datasets.filesystems import COMPRESSION_FILESYSTEMS, HfFileSystem, extract_path_from_uri, is_remote_filesystem
from .utils import require_lza, require_zstandard
def lowerCAmelCase_ ( _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
assert "mock" in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
assert "mock" not in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Dict = "mock-s3-bucket"
__magic_name__ : Any = F'''s3://{mock_bucket}'''
__magic_name__ : str = extract_path_from_uri(_snake_case )
assert dataset_path.startswith("s3://" ) is False
__magic_name__ : Tuple = "./local/path"
__magic_name__ : Optional[Any] = extract_path_from_uri(_snake_case )
assert dataset_path == new_dataset_path
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : str = is_remote_filesystem(_snake_case )
assert is_remote is True
__magic_name__ : Optional[int] = fsspec.filesystem("file" )
__magic_name__ : int = is_remote_filesystem(_snake_case )
assert is_remote is False
@pytest.mark.parametrize("compression_fs_class" , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : int , _snake_case : Tuple , _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Any ) -> int:
'''simple docstring'''
__magic_name__ : Any = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_file, "bz2": bza_file, "lz4": lza_file}
__magic_name__ : str = input_paths[compression_fs_class.protocol]
if input_path is None:
__magic_name__ : Dict = F'''for \'{compression_fs_class.protocol}\' compression protocol, '''
if compression_fs_class.protocol == "lz4":
reason += require_lza.kwargs["reason"]
elif compression_fs_class.protocol == "zstd":
reason += require_zstandard.kwargs["reason"]
pytest.skip(_snake_case )
__magic_name__ : str = fsspec.filesystem(compression_fs_class.protocol , fo=_snake_case )
assert isinstance(_snake_case , _snake_case )
__magic_name__ : int = os.path.basename(_snake_case )
__magic_name__ : Optional[int] = expected_filename[: expected_filename.rindex("." )]
assert fs.glob("*" ) == [expected_filename]
with fs.open(_snake_case , "r" , encoding="utf-8" ) as f, open(_snake_case , encoding="utf-8" ) as expected_file:
assert f.read() == expected_file.read()
@pytest.mark.parametrize("protocol" , ["zip", "gzip"] )
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : Optional[Any] , _snake_case : Optional[Any] ) -> str:
'''simple docstring'''
__magic_name__ : int = {"zip": zip_jsonl_path, "gzip": jsonl_gz_path}
__magic_name__ : int = compressed_file_paths[protocol]
__magic_name__ : Tuple = "dataset.jsonl"
__magic_name__ : List[str] = F'''{protocol}://{member_file_path}::{compressed_file_path}'''
__magic_name__ , *__magic_name__ : Optional[Any] = fsspec.get_fs_token_paths(_snake_case )
assert fs.isfile(_snake_case )
assert not fs.isfile("non_existing_" + member_file_path )
@pytest.mark.integration
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : List[str] , _snake_case : Tuple ) -> str:
'''simple docstring'''
__magic_name__ : int = hf_api.dataset_info(_snake_case , token=_snake_case )
__magic_name__ : Optional[Any] = HfFileSystem(repo_info=_snake_case , token=_snake_case )
assert sorted(hffs.glob("*" ) ) == [".gitattributes", "data"]
assert hffs.isdir("data" )
assert hffs.isfile(".gitattributes" ) and hffs.isfile("data/text_data.txt" )
with open(_snake_case ) as f:
assert hffs.open("data/text_data.txt" , "r" ).read() == f.read()
def lowerCAmelCase_ ( ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Optional[Any] = "bz2"
# Import module
import datasets.filesystems
# Overwrite protocol and reload
register_implementation(_snake_case , _snake_case , clobber=_snake_case )
with pytest.warns(_snake_case ) as warning_info:
importlib.reload(datasets.filesystems )
assert len(_snake_case ) == 1
assert (
str(warning_info[0].message )
== F'''A filesystem protocol was already set for {protocol} and will be overwritten.'''
)
| 281 | 1 |
import logging
from dataclasses import dataclass, field
from typing import Optional
from seqaseq_trainer import arg_to_scheduler
from transformers import TrainingArguments
snake_case : List[str] = logging.getLogger(__name__)
@dataclass
class _snake_case ( snake_case ):
UpperCamelCase__ = field(
default=0.0 , metadata={'help': 'The label smoothing epsilon to apply (if not zero).'} )
UpperCamelCase__ = field(default=snake_case , metadata={'help': 'Whether to SortishSamler or not.'} )
UpperCamelCase__ = field(
default=snake_case , metadata={'help': 'Whether to use generate to calculate generative metrics (ROUGE, BLEU).'} )
UpperCamelCase__ = field(default=snake_case , metadata={'help': 'whether to use adafactor'} )
UpperCamelCase__ = field(
default=snake_case , metadata={'help': 'Encoder layer dropout probability. Goes into model.config.'} )
UpperCamelCase__ = field(
default=snake_case , metadata={'help': 'Decoder layer dropout probability. Goes into model.config.'} )
UpperCamelCase__ = field(default=snake_case , metadata={'help': 'Dropout probability. Goes into model.config.'} )
UpperCamelCase__ = field(
default=snake_case , metadata={'help': 'Attention dropout probability. Goes into model.config.'} )
UpperCamelCase__ = field(
default='linear' , metadata={'help': F'Which lr scheduler to use. Selected in {sorted(arg_to_scheduler.keys() )}'} , )
| 281 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : List[Any] = {
"YituTech/conv-bert-base": "https://huggingface.co/YituTech/conv-bert-base/resolve/main/config.json",
"YituTech/conv-bert-medium-small": (
"https://huggingface.co/YituTech/conv-bert-medium-small/resolve/main/config.json"
),
"YituTech/conv-bert-small": "https://huggingface.co/YituTech/conv-bert-small/resolve/main/config.json",
# See all ConvBERT models at https://huggingface.co/models?filter=convbert
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'convbert'
def __init__( self , _a=30_522 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=1 , _a=0 , _a=2 , _a=768 , _a=2 , _a=9 , _a=1 , _a=None , **_a , ):
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a , )
__magic_name__ : Tuple = vocab_size
__magic_name__ : List[Any] = hidden_size
__magic_name__ : Union[str, Any] = num_hidden_layers
__magic_name__ : List[Any] = num_attention_heads
__magic_name__ : str = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : Tuple = max_position_embeddings
__magic_name__ : str = type_vocab_size
__magic_name__ : List[str] = initializer_range
__magic_name__ : Tuple = layer_norm_eps
__magic_name__ : List[Any] = embedding_size
__magic_name__ : List[Any] = head_ratio
__magic_name__ : str = conv_kernel_size
__magic_name__ : Dict = num_groups
__magic_name__ : str = classifier_dropout
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : Dict = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
("token_type_ids", dynamic_axis),
] )
| 281 | 1 |
import os
import pytest
from datasets import (
get_dataset_config_info,
get_dataset_config_names,
get_dataset_infos,
get_dataset_split_names,
inspect_dataset,
inspect_metric,
)
snake_case : int = pytest.mark.integration
@pytest.mark.parametrize("path" , ["paws", "csv"] )
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : str ) -> Dict:
'''simple docstring'''
inspect_dataset(_snake_case , _snake_case )
__magic_name__ : Any = path + ".py"
assert script_name in os.listdir(_snake_case )
assert "__pycache__" not in os.listdir(_snake_case )
@pytest.mark.filterwarnings("ignore:inspect_metric is deprecated:FutureWarning" )
@pytest.mark.filterwarnings("ignore:metric_module_factory is deprecated:FutureWarning" )
@pytest.mark.parametrize("path" , ["accuracy"] )
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any ) -> Optional[int]:
'''simple docstring'''
inspect_metric(_snake_case , _snake_case )
__magic_name__ : int = path + ".py"
assert script_name in os.listdir(_snake_case )
assert "__pycache__" not in os.listdir(_snake_case )
@pytest.mark.parametrize(
"path, config_name, expected_splits" , [
("squad", "plain_text", ["train", "validation"]),
("dalle-mini/wit", "dalle-mini--wit", ["train"]),
("paws", "labeled_final", ["train", "test", "validation"]),
] , )
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : int , _snake_case : str ) -> Any:
'''simple docstring'''
__magic_name__ : Optional[int] = get_dataset_config_info(_snake_case , config_name=_snake_case )
assert info.config_name == config_name
assert list(info.splits.keys() ) == expected_splits
@pytest.mark.parametrize(
"path, config_name, expected_exception" , [
("paws", None, ValueError),
] , )
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : Union[str, Any] , _snake_case : int ) -> int:
'''simple docstring'''
with pytest.raises(_snake_case ):
get_dataset_config_info(_snake_case , config_name=_snake_case )
@pytest.mark.parametrize(
"path, expected" , [
("squad", "plain_text"),
("acronym_identification", "default"),
("lhoestq/squad", "plain_text"),
("lhoestq/test", "default"),
("lhoestq/demo1", "lhoestq--demo1"),
("dalle-mini/wit", "dalle-mini--wit"),
] , )
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : List[str] = get_dataset_config_names(_snake_case )
assert expected in config_names
@pytest.mark.parametrize(
"path, expected_configs, expected_splits_in_first_config" , [
("squad", ["plain_text"], ["train", "validation"]),
("dalle-mini/wit", ["dalle-mini--wit"], ["train"]),
("paws", ["labeled_final", "labeled_swap", "unlabeled_final"], ["train", "test", "validation"]),
] , )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : List[str] , _snake_case : int ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = get_dataset_infos(_snake_case )
assert list(infos.keys() ) == expected_configs
__magic_name__ : Any = expected_configs[0]
assert expected_config in infos
__magic_name__ : List[str] = infos[expected_config]
assert info.config_name == expected_config
assert list(info.splits.keys() ) == expected_splits_in_first_config
@pytest.mark.parametrize(
"path, expected_config, expected_splits" , [
("squad", "plain_text", ["train", "validation"]),
("dalle-mini/wit", "dalle-mini--wit", ["train"]),
("paws", "labeled_final", ["train", "test", "validation"]),
] , )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : Dict , _snake_case : int ) -> List[str]:
'''simple docstring'''
__magic_name__ : Tuple = get_dataset_infos(_snake_case )
assert expected_config in infos
__magic_name__ : str = infos[expected_config]
assert info.config_name == expected_config
assert list(info.splits.keys() ) == expected_splits
@pytest.mark.parametrize(
"path, config_name, expected_exception" , [
("paws", None, ValueError),
] , )
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : Tuple , _snake_case : Any ) -> Tuple:
'''simple docstring'''
with pytest.raises(_snake_case ):
get_dataset_split_names(_snake_case , config_name=_snake_case )
| 281 |
import argparse
import requests
import torch
# pip3 install salesforce-lavis
# I'm actually installing a slightly modified version: pip3 install git+https://github.com/nielsrogge/LAVIS.git@fix_lavis
from lavis.models import load_model_and_preprocess
from PIL import Image
from transformers import (
AutoTokenizer,
BlipaConfig,
BlipaForConditionalGeneration,
BlipaProcessor,
BlipaVisionConfig,
BlipImageProcessor,
OPTConfig,
TaConfig,
)
from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD
def lowerCAmelCase_ ( ) -> str:
'''simple docstring'''
__magic_name__ : int = "https://storage.googleapis.com/sfr-vision-language-research/LAVIS/assets/merlion.png"
__magic_name__ : Union[str, Any] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw ).convert("RGB" )
return image
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = []
# fmt: off
# vision encoder
rename_keys.append(("visual_encoder.cls_token", "vision_model.embeddings.class_embedding") )
rename_keys.append(("visual_encoder.pos_embed", "vision_model.embeddings.position_embedding") )
rename_keys.append(("visual_encoder.patch_embed.proj.weight", "vision_model.embeddings.patch_embedding.weight") )
rename_keys.append(("visual_encoder.patch_embed.proj.bias", "vision_model.embeddings.patch_embedding.bias") )
rename_keys.append(("ln_vision.weight", "vision_model.post_layernorm.weight") )
rename_keys.append(("ln_vision.bias", "vision_model.post_layernorm.bias") )
for i in range(config.vision_config.num_hidden_layers ):
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.weight''', F'''vision_model.encoder.layers.{i}.layer_norm1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.bias''', F'''vision_model.encoder.layers.{i}.layer_norm1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.weight''', F'''vision_model.encoder.layers.{i}.layer_norm2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.bias''', F'''vision_model.encoder.layers.{i}.layer_norm2.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.qkv.weight''', F'''vision_model.encoder.layers.{i}.self_attn.qkv.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.weight''', F'''vision_model.encoder.layers.{i}.self_attn.projection.weight''',) )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.bias''', F'''vision_model.encoder.layers.{i}.self_attn.projection.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc2.bias''') )
# QFormer
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.weight", "qformer.layernorm.weight") )
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.bias", "qformer.layernorm.bias") )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Optional[Any] ) -> int:
'''simple docstring'''
__magic_name__ : Tuple = dct.pop(_snake_case )
__magic_name__ : int = val
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
for i in range(config.vision_config.num_hidden_layers ):
# read in original q and v biases
__magic_name__ : List[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.q_bias''' )
__magic_name__ : Optional[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.v_bias''' )
# next, set bias in the state dict
__magic_name__ : Optional[int] = torch.cat((q_bias, torch.zeros_like(_snake_case , requires_grad=_snake_case ), v_bias) )
__magic_name__ : Union[str, Any] = qkv_bias
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : str ) -> int:
'''simple docstring'''
__magic_name__ : List[Any] = 364 if "coco" in model_name else 224
__magic_name__ : Union[str, Any] = BlipaVisionConfig(image_size=_snake_case ).to_dict()
# make sure the models have proper bos_token_id and eos_token_id set (important for generation)
# seems like flan-T5 models don't have bos_token_id properly set?
if "opt-2.7b" in model_name:
__magic_name__ : List[str] = OPTConfig.from_pretrained("facebook/opt-2.7b" , eos_token_id=_snake_case ).to_dict()
elif "opt-6.7b" in model_name:
__magic_name__ : Any = OPTConfig.from_pretrained("facebook/opt-6.7b" , eos_token_id=_snake_case ).to_dict()
elif "t5-xl" in model_name:
__magic_name__ : Dict = TaConfig.from_pretrained("google/flan-t5-xl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
elif "t5-xxl" in model_name:
__magic_name__ : int = TaConfig.from_pretrained("google/flan-t5-xxl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
__magic_name__ : List[Any] = BlipaConfig(vision_config=_snake_case , text_config=_snake_case )
return config, image_size
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : str=None , _snake_case : Dict=False ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Optional[int] = (
AutoTokenizer.from_pretrained("facebook/opt-2.7b" )
if "opt" in model_name
else AutoTokenizer.from_pretrained("google/flan-t5-xl" )
)
__magic_name__ : List[Any] = tokenizer("\n" , add_special_tokens=_snake_case ).input_ids[0]
__magic_name__ , __magic_name__ : Tuple = get_blipa_config(_snake_case , eos_token_id=_snake_case )
__magic_name__ : Union[str, Any] = BlipaForConditionalGeneration(_snake_case ).eval()
__magic_name__ : Any = {
"blip2-opt-2.7b": ("blip2_opt", "pretrain_opt2.7b"),
"blip2-opt-6.7b": ("blip2_opt", "pretrain_opt6.7b"),
"blip2-opt-2.7b-coco": ("blip2_opt", "caption_coco_opt2.7b"),
"blip2-opt-6.7b-coco": ("blip2_opt", "caption_coco_opt6.7b"),
"blip2-flan-t5-xl": ("blip2_t5", "pretrain_flant5xl"),
"blip2-flan-t5-xl-coco": ("blip2_t5", "caption_coco_flant5xl"),
"blip2-flan-t5-xxl": ("blip2_t5", "pretrain_flant5xxl"),
}
__magic_name__ , __magic_name__ : Union[str, Any] = model_name_to_original[model_name]
# load original model
print("Loading original model..." )
__magic_name__ : Union[str, Any] = "cuda" if torch.cuda.is_available() else "cpu"
__magic_name__ , __magic_name__ , __magic_name__ : Optional[Any] = load_model_and_preprocess(
name=_snake_case , model_type=_snake_case , is_eval=_snake_case , device=_snake_case )
original_model.eval()
print("Done!" )
# update state dict keys
__magic_name__ : Dict = original_model.state_dict()
__magic_name__ : str = create_rename_keys(_snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
# some keys can be renamed efficiently
for key, val in state_dict.copy().items():
__magic_name__ : Any = state_dict.pop(_snake_case )
if key.startswith("Qformer.bert" ):
__magic_name__ : Optional[int] = key.replace("Qformer.bert" , "qformer" )
if "attention.self" in key:
__magic_name__ : Any = key.replace("self" , "attention" )
if "opt_proj" in key:
__magic_name__ : Union[str, Any] = key.replace("opt_proj" , "language_projection" )
if "t5_proj" in key:
__magic_name__ : Optional[int] = key.replace("t5_proj" , "language_projection" )
if key.startswith("opt" ):
__magic_name__ : List[str] = key.replace("opt" , "language" )
if key.startswith("t5" ):
__magic_name__ : Tuple = key.replace("t5" , "language" )
__magic_name__ : Dict = val
# read in qv biases
read_in_q_v_bias(_snake_case , _snake_case )
__magic_name__ , __magic_name__ : Tuple = hf_model.load_state_dict(_snake_case , strict=_snake_case )
assert len(_snake_case ) == 0
assert unexpected_keys == ["qformer.embeddings.position_ids"]
__magic_name__ : List[Any] = load_demo_image()
__magic_name__ : Tuple = vis_processors["eval"](_snake_case ).unsqueeze(0 ).to(_snake_case )
__magic_name__ : Dict = tokenizer(["\n"] , return_tensors="pt" ).input_ids.to(_snake_case )
# create processor
__magic_name__ : Optional[Any] = BlipImageProcessor(
size={"height": image_size, "width": image_size} , image_mean=_snake_case , image_std=_snake_case )
__magic_name__ : Dict = BlipaProcessor(image_processor=_snake_case , tokenizer=_snake_case )
__magic_name__ : Union[str, Any] = processor(images=_snake_case , return_tensors="pt" ).pixel_values.to(_snake_case )
# make sure processor creates exact same pixel values
assert torch.allclose(_snake_case , _snake_case )
original_model.to(_snake_case )
hf_model.to(_snake_case )
with torch.no_grad():
if "opt" in model_name:
__magic_name__ : List[Any] = original_model({"image": original_pixel_values, "text_input": [""]} ).logits
__magic_name__ : Optional[int] = hf_model(_snake_case , _snake_case ).logits
else:
__magic_name__ : int = original_model(
{"image": original_pixel_values, "text_input": ["\n"], "text_output": ["\n"]} ).logits
__magic_name__ : Tuple = input_ids.masked_fill(input_ids == tokenizer.pad_token_id , -100 )
__magic_name__ : List[str] = hf_model(_snake_case , _snake_case , labels=_snake_case ).logits
assert original_logits.shape == logits.shape
print("First values of original logits:" , original_logits[0, :3, :3] )
print("First values of HF logits:" , logits[0, :3, :3] )
# assert values
if model_name == "blip2-flan-t5-xl":
__magic_name__ : List[str] = torch.tensor(
[[-41.5_850, -4.4_440, -8.9_922], [-47.4_322, -5.9_143, -1.7_340]] , device=_snake_case )
assert torch.allclose(logits[0, :3, :3] , _snake_case , atol=1E-4 )
elif model_name == "blip2-flan-t5-xl-coco":
__magic_name__ : Tuple = torch.tensor(
[[-57.0_109, -9.8_967, -12.6_280], [-68.6_578, -12.7_191, -10.5_065]] , device=_snake_case )
else:
# cast to same type
__magic_name__ : str = logits.dtype
assert torch.allclose(original_logits.to(_snake_case ) , _snake_case , atol=1E-2 )
print("Looks ok!" )
print("Generating a caption..." )
__magic_name__ : Optional[int] = ""
__magic_name__ : Dict = tokenizer(_snake_case , return_tensors="pt" ).input_ids.to(_snake_case )
__magic_name__ : int = original_model.generate({"image": original_pixel_values} )
__magic_name__ : Optional[Any] = hf_model.generate(
_snake_case , _snake_case , do_sample=_snake_case , num_beams=5 , max_length=30 , min_length=1 , top_p=0.9 , repetition_penalty=1.0 , length_penalty=1.0 , temperature=1 , )
print("Original generation:" , _snake_case )
__magic_name__ : Tuple = input_ids.shape[1]
__magic_name__ : int = processor.batch_decode(outputs[:, prompt_length:] , skip_special_tokens=_snake_case )
__magic_name__ : Union[str, Any] = [text.strip() for text in output_text]
print("HF generation:" , _snake_case )
if pytorch_dump_folder_path is not None:
processor.save_pretrained(_snake_case )
hf_model.save_pretrained(_snake_case )
if push_to_hub:
processor.push_to_hub(F'''nielsr/{model_name}''' )
hf_model.push_to_hub(F'''nielsr/{model_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
snake_case : Union[str, Any] = [
"blip2-opt-2.7b",
"blip2-opt-6.7b",
"blip2-opt-2.7b-coco",
"blip2-opt-6.7b-coco",
"blip2-flan-t5-xl",
"blip2-flan-t5-xl-coco",
"blip2-flan-t5-xxl",
]
parser.add_argument(
"--model_name",
default="blip2-opt-2.7b",
choices=choices,
type=str,
help="Path to hf config.json of model to convert",
)
parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.")
parser.add_argument(
"--push_to_hub",
action="store_true",
help="Whether to push the model and processor to the hub after converting",
)
snake_case : int = parser.parse_args()
convert_blipa_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
from __future__ import annotations
from typing import Any
def lowerCAmelCase_ ( _snake_case : list[Any] ) -> None:
'''simple docstring'''
create_state_space_tree(_snake_case , [] , 0 )
def lowerCAmelCase_ ( _snake_case : list[Any] , _snake_case : list[Any] , _snake_case : int ) -> None:
'''simple docstring'''
if index == len(_snake_case ):
print(_snake_case )
return
create_state_space_tree(_snake_case , _snake_case , index + 1 )
current_subsequence.append(sequence[index] )
create_state_space_tree(_snake_case , _snake_case , index + 1 )
current_subsequence.pop()
if __name__ == "__main__":
snake_case : list[Any] = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["A", "B", "C"])
generate_all_subsequences(seq)
| 281 |
import os
import re
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : Union[str, Any] = {
"vocab_file": "vocab.txt",
"merges_file": "bpe.codes",
}
snake_case : Dict = {
"vocab_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/vocab.txt",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/vocab.txt",
},
"merges_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/bpe.codes",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/bpe.codes",
},
}
snake_case : Union[str, Any] = {
"vinai/phobert-base": 256,
"vinai/phobert-large": 256,
}
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = set()
__magic_name__ : Any = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
__magic_name__ : int = char
__magic_name__ : List[str] = set(_snake_case )
return pairs
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a , _a="<s>" , _a="</s>" , _a="</s>" , _a="<s>" , _a="<unk>" , _a="<pad>" , _a="<mask>" , **_a , ):
super().__init__(
bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , cls_token=_a , pad_token=_a , mask_token=_a , **_a , )
__magic_name__ : Dict = vocab_file
__magic_name__ : Tuple = merges_file
__magic_name__ : List[Any] = {}
__magic_name__ : List[Any] = 0
__magic_name__ : Tuple = 1
__magic_name__ : int = 2
__magic_name__ : Union[str, Any] = 3
self.add_from_file(_a )
__magic_name__ : Optional[int] = {v: k for k, v in self.encoder.items()}
with open(_a , encoding="utf-8" ) as merges_handle:
__magic_name__ : List[str] = merges_handle.read().split("\n" )[:-1]
__magic_name__ : Union[str, Any] = [tuple(merge.split()[:-1] ) for merge in merges]
__magic_name__ : Union[str, Any] = dict(zip(_a , range(len(_a ) ) ) )
__magic_name__ : Optional[int] = {}
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
__magic_name__ : Optional[Any] = [self.cls_token_id]
__magic_name__ : Union[str, Any] = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is None:
return [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[Any] = [self.sep_token_id]
__magic_name__ : 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 + sep + token_ids_a + sep ) * [0]
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.encoder )
def SCREAMING_SNAKE_CASE ( self ):
return dict(self.encoder , **self.added_tokens_encoder )
def SCREAMING_SNAKE_CASE ( self , _a ):
if token in self.cache:
return self.cache[token]
__magic_name__ : List[Any] = tuple(_a )
__magic_name__ : List[Any] = tuple(list(word[:-1] ) + [word[-1] + "</w>"] )
__magic_name__ : Any = get_pairs(_a )
if not pairs:
return token
while True:
__magic_name__ : str = min(_a , key=lambda _a : self.bpe_ranks.get(_a , float("inf" ) ) )
if bigram not in self.bpe_ranks:
break
__magic_name__ , __magic_name__ : List[str] = bigram
__magic_name__ : List[str] = []
__magic_name__ : List[str] = 0
while i < len(_a ):
try:
__magic_name__ : Any = word.index(_a , _a )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
__magic_name__ : Tuple = j
if word[i] == first and i < len(_a ) - 1 and word[i + 1] == second:
new_word.append(first + second )
i += 2
else:
new_word.append(word[i] )
i += 1
__magic_name__ : Union[str, Any] = tuple(_a )
__magic_name__ : Optional[int] = new_word
if len(_a ) == 1:
break
else:
__magic_name__ : List[Any] = get_pairs(_a )
__magic_name__ : Optional[int] = "@@ ".join(_a )
__magic_name__ : Tuple = word[:-4]
__magic_name__ : str = word
return word
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = []
__magic_name__ : Dict = re.findall(r"\S+\n?" , _a )
for token in words:
split_tokens.extend(list(self.bpe(_a ).split(" " ) ) )
return split_tokens
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.encoder.get(_a , self.encoder.get(self.unk_token ) )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.decoder.get(_a , self.unk_token )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Tuple = " ".join(_a ).replace("@@ " , "" ).strip()
return out_string
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : Optional[int] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
__magic_name__ : Union[str, Any] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["merges_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ):
copyfile(self.vocab_file , _a )
if os.path.abspath(self.merges_file ) != os.path.abspath(_a ):
copyfile(self.merges_file , _a )
return out_vocab_file, out_merge_file
def SCREAMING_SNAKE_CASE ( self , _a ):
if isinstance(_a , _a ):
try:
with open(_a , "r" , encoding="utf-8" ) as fd:
self.add_from_file(_a )
except FileNotFoundError as fnfe:
raise fnfe
except UnicodeError:
raise Exception(f'''Incorrect encoding detected in {f}, please rebuild the dataset''' )
return
__magic_name__ : List[Any] = f.readlines()
for lineTmp in lines:
__magic_name__ : Optional[Any] = lineTmp.strip()
__magic_name__ : Union[str, Any] = line.rfind(" " )
if idx == -1:
raise ValueError("Incorrect dictionary format, expected '<token> <cnt>'" )
__magic_name__ : Optional[int] = line[:idx]
__magic_name__ : Dict = len(self.encoder )
| 281 | 1 |
from typing import TYPE_CHECKING
from ....utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
snake_case : List[Any] = {"configuration_van": ["VAN_PRETRAINED_CONFIG_ARCHIVE_MAP", "VanConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : List[Any] = [
"VAN_PRETRAINED_MODEL_ARCHIVE_LIST",
"VanForImageClassification",
"VanModel",
"VanPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_van import VAN_PRETRAINED_CONFIG_ARCHIVE_MAP, VanConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_van import (
VAN_PRETRAINED_MODEL_ARCHIVE_LIST,
VanForImageClassification,
VanModel,
VanPreTrainedModel,
)
else:
import sys
snake_case : List[Any] = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 281 |
from itertools import zip_longest
import requests
from bsa import BeautifulSoup
from pandas import DataFrame
def lowerCAmelCase_ ( _snake_case : str = "laptop" ) -> DataFrame:
'''simple docstring'''
__magic_name__ : Tuple = F'''https://www.amazon.in/laptop/s?k={product}'''
__magic_name__ : Dict = {
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36\n (KHTML, like Gecko)Chrome/44.0.2403.157 Safari/537.36",
"Accept-Language": "en-US, en;q=0.5",
}
__magic_name__ : Tuple = BeautifulSoup(requests.get(_snake_case , headers=_snake_case ).text )
# Initialize a Pandas dataframe with the column titles
__magic_name__ : int = DataFrame(
columns=[
"Product Title",
"Product Link",
"Current Price of the product",
"Product Rating",
"MRP of the product",
"Discount",
] )
# Loop through each entry and store them in the dataframe
for item, _ in zip_longest(
soup.find_all(
"div" , attrs={"class": "s-result-item", "data-component-type": "s-search-result"} , ) , soup.find_all("div" , attrs={"class": "a-row a-size-base a-color-base"} ) , ):
try:
__magic_name__ : Dict = item.ha.text
__magic_name__ : Optional[int] = "https://www.amazon.in/" + item.ha.a["href"]
__magic_name__ : Optional[Any] = item.find("span" , attrs={"class": "a-offscreen"} ).text
try:
__magic_name__ : Union[str, Any] = item.find("span" , attrs={"class": "a-icon-alt"} ).text
except AttributeError:
__magic_name__ : Dict = "Not available"
try:
__magic_name__ : Optional[int] = (
"₹"
+ item.find(
"span" , attrs={"class": "a-price a-text-price"} ).text.split("₹" )[1]
)
except AttributeError:
__magic_name__ : List[str] = ""
try:
__magic_name__ : int = float(
(
(
float(product_mrp.strip("₹" ).replace("," , "" ) )
- float(product_price.strip("₹" ).replace("," , "" ) )
)
/ float(product_mrp.strip("₹" ).replace("," , "" ) )
)
* 100 )
except ValueError:
__magic_name__ : str = float("nan" )
except AttributeError:
pass
__magic_name__ : Optional[int] = [
product_title,
product_link,
product_price,
product_rating,
product_mrp,
discount,
]
__magic_name__ : Optional[Any] = " "
__magic_name__ : str = " "
data_frame.index += 1
return data_frame
if __name__ == "__main__":
snake_case : Any = "headphones"
get_amazon_product_data(product).to_csv(F"Amazon Product Data for {product}.csv")
| 281 | 1 |
import math
from dataclasses import dataclass
from typing import Optional, Tuple, Union
import numpy as np
import torch
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import BaseOutput, randn_tensor
from .scheduling_utils import SchedulerMixin
@dataclass
# Copied from diffusers.schedulers.scheduling_ddpm.DDPMSchedulerOutput with DDPM->UnCLIP
class _snake_case ( snake_case ):
UpperCamelCase__ = 42
UpperCamelCase__ = None
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : List[str]=0.999 , _snake_case : Optional[Any]="cosine" , ) -> Union[str, Any]:
'''simple docstring'''
if alpha_transform_type == "cosine":
def alpha_bar_fn(_snake_case : Optional[int] ):
return math.cos((t + 0.008) / 1.008 * math.pi / 2 ) ** 2
elif alpha_transform_type == "exp":
def alpha_bar_fn(_snake_case : Any ):
return math.exp(t * -12.0 )
else:
raise ValueError(F'''Unsupported alpha_tranform_type: {alpha_transform_type}''' )
__magic_name__ : Tuple = []
for i in range(_snake_case ):
__magic_name__ : List[str] = i / num_diffusion_timesteps
__magic_name__ : Optional[int] = (i + 1) / num_diffusion_timesteps
betas.append(min(1 - alpha_bar_fn(_snake_case ) / alpha_bar_fn(_snake_case ) , _snake_case ) )
return torch.tensor(_snake_case , dtype=torch.floataa )
class _snake_case ( snake_case , snake_case ):
@register_to_config
def __init__( self , _a = 1_000 , _a = "fixed_small_log" , _a = True , _a = 1.0 , _a = "epsilon" , _a = "squaredcos_cap_v2" , ):
if beta_schedule != "squaredcos_cap_v2":
raise ValueError("UnCLIPScheduler only supports `beta_schedule`: 'squaredcos_cap_v2'" )
__magic_name__ : Dict = betas_for_alpha_bar(_a )
__magic_name__ : int = 1.0 - self.betas
__magic_name__ : List[str] = torch.cumprod(self.alphas , dim=0 )
__magic_name__ : str = torch.tensor(1.0 )
# standard deviation of the initial noise distribution
__magic_name__ : Any = 1.0
# setable values
__magic_name__ : int = None
__magic_name__ : Union[str, Any] = torch.from_numpy(np.arange(0 , _a )[::-1].copy() )
__magic_name__ : Any = variance_type
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
return sample
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Any = num_inference_steps
__magic_name__ : str = (self.config.num_train_timesteps - 1) / (self.num_inference_steps - 1)
__magic_name__ : Any = (np.arange(0 , _a ) * step_ratio).round()[::-1].copy().astype(np.intaa )
__magic_name__ : Optional[Any] = torch.from_numpy(_a ).to(_a )
def SCREAMING_SNAKE_CASE ( self , _a , _a=None , _a=None , _a=None ):
if prev_timestep is None:
__magic_name__ : Union[str, Any] = t - 1
__magic_name__ : Optional[int] = self.alphas_cumprod[t]
__magic_name__ : str = self.alphas_cumprod[prev_timestep] if prev_timestep >= 0 else self.one
__magic_name__ : Union[str, Any] = 1 - alpha_prod_t
__magic_name__ : int = 1 - alpha_prod_t_prev
if prev_timestep == t - 1:
__magic_name__ : Dict = self.betas[t]
else:
__magic_name__ : Dict = 1 - alpha_prod_t / alpha_prod_t_prev
# For t > 0, compute predicted variance βt (see formula (6) and (7) from https://arxiv.org/pdf/2006.11239.pdf)
# and sample from it to get previous sample
# x_{t-1} ~ N(pred_prev_sample, variance) == add variance to pred_sample
__magic_name__ : List[str] = beta_prod_t_prev / beta_prod_t * beta
if variance_type is None:
__magic_name__ : Any = self.config.variance_type
# hacks - were probably added for training stability
if variance_type == "fixed_small_log":
__magic_name__ : Any = torch.log(torch.clamp(_a , min=1e-20 ) )
__magic_name__ : Optional[int] = torch.exp(0.5 * variance )
elif variance_type == "learned_range":
# NOTE difference with DDPM scheduler
__magic_name__ : Any = variance.log()
__magic_name__ : str = beta.log()
__magic_name__ : List[str] = (predicted_variance + 1) / 2
__magic_name__ : str = frac * max_log + (1 - frac) * min_log
return variance
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a = None , _a=None , _a = True , ):
__magic_name__ : Tuple = timestep
if model_output.shape[1] == sample.shape[1] * 2 and self.variance_type == "learned_range":
__magic_name__ , __magic_name__ : Optional[int] = torch.split(_a , sample.shape[1] , dim=1 )
else:
__magic_name__ : Tuple = None
# 1. compute alphas, betas
if prev_timestep is None:
__magic_name__ : Optional[Any] = t - 1
__magic_name__ : Tuple = self.alphas_cumprod[t]
__magic_name__ : List[str] = self.alphas_cumprod[prev_timestep] if prev_timestep >= 0 else self.one
__magic_name__ : List[Any] = 1 - alpha_prod_t
__magic_name__ : int = 1 - alpha_prod_t_prev
if prev_timestep == t - 1:
__magic_name__ : List[str] = self.betas[t]
__magic_name__ : List[str] = self.alphas[t]
else:
__magic_name__ : Optional[int] = 1 - alpha_prod_t / alpha_prod_t_prev
__magic_name__ : List[Any] = 1 - beta
# 2. compute predicted original sample from predicted noise also called
# "predicted x_0" of formula (15) from https://arxiv.org/pdf/2006.11239.pdf
if self.config.prediction_type == "epsilon":
__magic_name__ : Dict = (sample - beta_prod_t ** 0.5 * model_output) / alpha_prod_t ** 0.5
elif self.config.prediction_type == "sample":
__magic_name__ : Dict = model_output
else:
raise ValueError(
f'''prediction_type given as {self.config.prediction_type} must be one of `epsilon` or `sample`'''
" for the UnCLIPScheduler." )
# 3. Clip "predicted x_0"
if self.config.clip_sample:
__magic_name__ : Any = torch.clamp(
_a , -self.config.clip_sample_range , self.config.clip_sample_range )
# 4. Compute coefficients for pred_original_sample x_0 and current sample x_t
# See formula (7) from https://arxiv.org/pdf/2006.11239.pdf
__magic_name__ : Optional[int] = (alpha_prod_t_prev ** 0.5 * beta) / beta_prod_t
__magic_name__ : str = alpha ** 0.5 * beta_prod_t_prev / beta_prod_t
# 5. Compute predicted previous sample µ_t
# See formula (7) from https://arxiv.org/pdf/2006.11239.pdf
__magic_name__ : Dict = pred_original_sample_coeff * pred_original_sample + current_sample_coeff * sample
# 6. Add noise
__magic_name__ : str = 0
if t > 0:
__magic_name__ : Union[str, Any] = randn_tensor(
model_output.shape , dtype=model_output.dtype , generator=_a , device=model_output.device )
__magic_name__ : Tuple = self._get_variance(
_a , predicted_variance=_a , prev_timestep=_a , )
if self.variance_type == "fixed_small_log":
__magic_name__ : Tuple = variance
elif self.variance_type == "learned_range":
__magic_name__ : Tuple = (0.5 * variance).exp()
else:
raise ValueError(
f'''variance_type given as {self.variance_type} must be one of `fixed_small_log` or `learned_range`'''
" for the UnCLIPScheduler." )
__magic_name__ : str = variance * variance_noise
__magic_name__ : int = pred_prev_sample + variance
if not return_dict:
return (pred_prev_sample,)
return UnCLIPSchedulerOutput(prev_sample=_a , pred_original_sample=_a )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , ):
# Make sure alphas_cumprod and timestep have same device and dtype as original_samples
__magic_name__ : str = self.alphas_cumprod.to(device=original_samples.device , dtype=original_samples.dtype )
__magic_name__ : List[Any] = timesteps.to(original_samples.device )
__magic_name__ : int = alphas_cumprod[timesteps] ** 0.5
__magic_name__ : List[str] = sqrt_alpha_prod.flatten()
while len(sqrt_alpha_prod.shape ) < len(original_samples.shape ):
__magic_name__ : Optional[int] = sqrt_alpha_prod.unsqueeze(-1 )
__magic_name__ : int = (1 - alphas_cumprod[timesteps]) ** 0.5
__magic_name__ : str = sqrt_one_minus_alpha_prod.flatten()
while len(sqrt_one_minus_alpha_prod.shape ) < len(original_samples.shape ):
__magic_name__ : int = sqrt_one_minus_alpha_prod.unsqueeze(-1 )
__magic_name__ : List[str] = sqrt_alpha_prod * original_samples + sqrt_one_minus_alpha_prod * noise
return noisy_samples
| 281 |
from __future__ import annotations
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Optional[Any] = data
__magic_name__ : Node | None = None
__magic_name__ : Node | None = None
def lowerCAmelCase_ ( _snake_case : Node | None ) -> None: # In Order traversal of the tree
'''simple docstring'''
if tree:
display(tree.left )
print(tree.data )
display(tree.right )
def lowerCAmelCase_ ( _snake_case : Node | None ) -> int:
'''simple docstring'''
return 1 + max(depth_of_tree(tree.left ) , depth_of_tree(tree.right ) ) if tree else 0
def lowerCAmelCase_ ( _snake_case : Node ) -> bool:
'''simple docstring'''
if not tree:
return True
if tree.left and tree.right:
return is_full_binary_tree(tree.left ) and is_full_binary_tree(tree.right )
else:
return not tree.left and not tree.right
def lowerCAmelCase_ ( ) -> None: # Main function for testing.
'''simple docstring'''
__magic_name__ : int = Node(1 )
__magic_name__ : Union[str, Any] = Node(2 )
__magic_name__ : Tuple = Node(3 )
__magic_name__ : Optional[Any] = Node(4 )
__magic_name__ : Union[str, Any] = Node(5 )
__magic_name__ : Any = Node(6 )
__magic_name__ : int = Node(7 )
__magic_name__ : List[str] = Node(8 )
__magic_name__ : Union[str, Any] = Node(9 )
print(is_full_binary_tree(_snake_case ) )
print(depth_of_tree(_snake_case ) )
print("Tree is: " )
display(_snake_case )
if __name__ == "__main__":
main()
| 281 | 1 |
import json
import os
import unittest
from transformers.models.ctrl.tokenization_ctrl import VOCAB_FILES_NAMES, CTRLTokenizer
from ...test_tokenization_common import TokenizerTesterMixin
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = CTRLTokenizer
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
# Adapted from Sennrich et al. 2015 and https://github.com/rsennrich/subword-nmt
__magic_name__ : Any = ["adapt", "re@@", "a@@", "apt", "c@@", "t", "<unk>"]
__magic_name__ : Tuple = dict(zip(_a , range(len(_a ) ) ) )
__magic_name__ : int = ["#version: 0.2", "a p", "ap t</w>", "r e", "a d", "ad apt</w>", ""]
__magic_name__ : Tuple = {"unk_token": "<unk>"}
__magic_name__ : Dict = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["vocab_file"] )
__magic_name__ : Optional[Any] = 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(_a ) + "\n" )
with open(self.merges_file , "w" , encoding="utf-8" ) as fp:
fp.write("\n".join(_a ) )
def SCREAMING_SNAKE_CASE ( self , **_a ):
kwargs.update(self.special_tokens_map )
return CTRLTokenizer.from_pretrained(self.tmpdirname , **_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Dict = "adapt react readapt apt"
__magic_name__ : Dict = "adapt react readapt apt"
return input_text, output_text
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = CTRLTokenizer(self.vocab_file , self.merges_file , **self.special_tokens_map )
__magic_name__ : Tuple = "adapt react readapt apt"
__magic_name__ : Tuple = "adapt re@@ a@@ c@@ t re@@ adapt apt".split()
__magic_name__ : Dict = tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : Dict = tokens + [tokenizer.unk_token]
__magic_name__ : str = [0, 1, 2, 4, 5, 1, 0, 3, 6]
self.assertListEqual(tokenizer.convert_tokens_to_ids(_a ) , _a )
| 281 |
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> bool:
'''simple docstring'''
__magic_name__ : Union[str, Any] = len(_snake_case ) + 1
__magic_name__ : List[str] = len(_snake_case ) + 1
# dp is a 2d matrix where dp[i][j] denotes whether prefix string of
# length i of input_string matches with prefix string of length j of
# given pattern.
# "dp" stands for dynamic programming.
__magic_name__ : str = [[0 for i in range(_snake_case )] for j in range(_snake_case )]
# since string of zero length match pattern of zero length
__magic_name__ : Optional[int] = 1
# since pattern of zero length will never match with string of non-zero length
for i in range(1 , _snake_case ):
__magic_name__ : Optional[int] = 0
# since string of zero length will match with pattern where there
# is at least one * alternatively
for j in range(1 , _snake_case ):
__magic_name__ : Union[str, Any] = dp[0][j - 2] if pattern[j - 1] == "*" else 0
# now using bottom-up approach to find for all remaining lengths
for i in range(1 , _snake_case ):
for j in range(1 , _snake_case ):
if input_string[i - 1] == pattern[j - 1] or pattern[j - 1] == ".":
__magic_name__ : Optional[int] = dp[i - 1][j - 1]
elif pattern[j - 1] == "*":
if dp[i][j - 2] == 1:
__magic_name__ : Optional[Any] = 1
elif pattern[j - 2] in (input_string[i - 1], "."):
__magic_name__ : List[Any] = dp[i - 1][j]
else:
__magic_name__ : Union[str, Any] = 0
else:
__magic_name__ : Dict = 0
return bool(dp[-1][-1] )
if __name__ == "__main__":
import doctest
doctest.testmod()
# inputing the strings
# input_string = input("input a string :")
# pattern = input("input a pattern :")
snake_case : Optional[Any] = "aab"
snake_case : List[str] = "c*a*b"
# using function to check whether given string matches the given pattern
if match_pattern(input_string, pattern):
print(F"{input_string} matches the given pattern {pattern}")
else:
print(F"{input_string} does not match with the given pattern {pattern}")
| 281 | 1 |
import numpy as np
from transformers import BatchFeature
from transformers.testing_utils import require_tf, require_torch
from .test_feature_extraction_common import FeatureExtractionSavingTestMixin
class _snake_case ( snake_case ):
# to overwrite at feature extractactor specific tests
UpperCamelCase__ = None
UpperCamelCase__ = None
@property
def SCREAMING_SNAKE_CASE ( self ):
return self.feat_extract_tester.prepare_feat_extract_dict()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = self.feature_extraction_class(**self.feat_extract_dict )
self.assertTrue(hasattr(_a , "feature_size" ) )
self.assertTrue(hasattr(_a , "sampling_rate" ) )
self.assertTrue(hasattr(_a , "padding_value" ) )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = self.feat_extract_tester.prepare_inputs_for_common()
__magic_name__ : Dict = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : List[Any] = feat_extract.model_input_names[0]
__magic_name__ : Dict = BatchFeature({input_name: speech_inputs} )
self.assertTrue(all(len(_a ) == len(_a ) for x, y in zip(_a , processed_features[input_name] ) ) )
__magic_name__ : int = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a )
__magic_name__ : str = BatchFeature({input_name: speech_inputs} , tensor_type="np" )
__magic_name__ : int = processed_features[input_name]
if len(batch_features_input.shape ) < 3:
__magic_name__ : List[str] = batch_features_input[:, :, None]
self.assertTrue(
batch_features_input.shape
== (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) )
@require_torch
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a )
__magic_name__ : str = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : Tuple = feat_extract.model_input_names[0]
__magic_name__ : Optional[int] = BatchFeature({input_name: speech_inputs} , tensor_type="pt" )
__magic_name__ : str = processed_features[input_name]
if len(batch_features_input.shape ) < 3:
__magic_name__ : Optional[int] = batch_features_input[:, :, None]
self.assertTrue(
batch_features_input.shape
== (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) )
@require_tf
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = self.feat_extract_tester.prepare_inputs_for_common(equal_length=_a )
__magic_name__ : Dict = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : Dict = feat_extract.model_input_names[0]
__magic_name__ : Optional[int] = BatchFeature({input_name: speech_inputs} , tensor_type="tf" )
__magic_name__ : Union[str, Any] = processed_features[input_name]
if len(batch_features_input.shape ) < 3:
__magic_name__ : Union[str, Any] = batch_features_input[:, :, None]
self.assertTrue(
batch_features_input.shape
== (self.feat_extract_tester.batch_size, len(speech_inputs[0] ), self.feat_extract_tester.feature_size) )
def SCREAMING_SNAKE_CASE ( self , _a=False ):
def _inputs_have_equal_length(_a ):
__magic_name__ : Tuple = len(input[0] )
for input_slice in input[1:]:
if len(_a ) != length:
return False
return True
def _inputs_are_equal(_a , _a ):
if len(_a ) != len(_a ):
return False
for input_slice_a, input_slice_a in zip(_a , _a ):
if not np.allclose(np.asarray(_a ) , np.asarray(_a ) , atol=1e-3 ):
return False
return True
__magic_name__ : Dict = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : List[Any] = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a )
__magic_name__ : List[Any] = feat_extract.model_input_names[0]
__magic_name__ : Union[str, Any] = BatchFeature({input_name: speech_inputs} )
__magic_name__ : Tuple = self.feat_extract_tester.seq_length_diff
__magic_name__ : List[Any] = self.feat_extract_tester.max_seq_length + pad_diff
__magic_name__ : Any = self.feat_extract_tester.min_seq_length
__magic_name__ : Dict = self.feat_extract_tester.batch_size
__magic_name__ : Optional[Any] = self.feat_extract_tester.feature_size
# test padding for List[int] + numpy
__magic_name__ : int = feat_extract.pad(_a , padding=_a )
__magic_name__ : int = input_a[input_name]
__magic_name__ : int = feat_extract.pad(_a , padding="longest" )
__magic_name__ : int = input_a[input_name]
__magic_name__ : List[Any] = feat_extract.pad(_a , padding="max_length" , max_length=len(speech_inputs[-1] ) )
__magic_name__ : List[Any] = input_a[input_name]
__magic_name__ : str = feat_extract.pad(_a , padding="longest" , return_tensors="np" )
__magic_name__ : Any = input_a[input_name]
# max_length parameter has to be provided when setting `padding="max_length"`
with self.assertRaises(_a ):
feat_extract.pad(_a , padding="max_length" )[input_name]
__magic_name__ : List[Any] = feat_extract.pad(
_a , padding="max_length" , max_length=_a , return_tensors="np" )
__magic_name__ : Optional[int] = input_a[input_name]
self.assertFalse(_inputs_have_equal_length(_a ) )
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertTrue(_inputs_are_equal(_a , _a ) )
self.assertTrue(len(input_a[0] ) == pad_min_length )
self.assertTrue(len(input_a[1] ) == pad_min_length + pad_diff )
self.assertTrue(input_a.shape[:2] == (batch_size, len(input_a[0] )) )
self.assertTrue(input_a.shape[:2] == (batch_size, pad_max_length) )
if feature_size > 1:
self.assertTrue(input_a.shape[2] == input_a.shape[2] == feature_size )
# test padding for `pad_to_multiple_of` for List[int] + numpy
__magic_name__ : str = feat_extract.pad(_a , pad_to_multiple_of=10 )
__magic_name__ : int = input_a[input_name]
__magic_name__ : Union[str, Any] = feat_extract.pad(_a , padding="longest" , pad_to_multiple_of=10 )
__magic_name__ : Dict = input_a[input_name]
__magic_name__ : List[Any] = feat_extract.pad(
_a , padding="max_length" , pad_to_multiple_of=10 , max_length=_a )
__magic_name__ : int = input_a[input_name]
__magic_name__ : Optional[Any] = feat_extract.pad(
_a , padding="max_length" , pad_to_multiple_of=10 , max_length=_a , return_tensors="np" , )
__magic_name__ : str = input_a[input_name]
self.assertTrue(all(len(_a ) % 10 == 0 for x in input_a ) )
self.assertTrue(_inputs_are_equal(_a , _a ) )
__magic_name__ : List[str] = pad_max_length if pad_max_length % 10 == 0 else (pad_max_length // 10 + 1) * 10
self.assertTrue(all(len(_a ) == expected_mult_pad_length for x in input_a ) )
self.assertEqual(input_a.shape[:2] , (batch_size, expected_mult_pad_length) )
if feature_size > 1:
self.assertTrue(input_a.shape[2] == feature_size )
# Check padding value is correct
__magic_name__ : Optional[int] = (np.ones(self.feat_extract_tester.feature_size ) * feat_extract.padding_value).sum()
self.assertTrue(
abs(np.asarray(input_a[0] )[pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) )
< 1e-3 )
self.assertTrue(
abs(
np.asarray(input_a[1] )[pad_min_length + pad_diff :].sum()
- padding_vector_sum * (pad_max_length - pad_min_length - pad_diff) )
< 1e-3 )
self.assertTrue(
abs(
np.asarray(input_a[2] )[pad_min_length + 2 * pad_diff :].sum()
- padding_vector_sum * (pad_max_length - pad_min_length - 2 * pad_diff) )
< 1e-3 )
self.assertTrue(
abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (pad_max_length - pad_min_length) ) < 1e-3 )
self.assertTrue(
abs(input_a[0, pad_min_length:].sum() - padding_vector_sum * (expected_mult_pad_length - pad_min_length) )
< 1e-3 )
def SCREAMING_SNAKE_CASE ( self , _a=False ):
def _inputs_have_equal_length(_a ):
__magic_name__ : Optional[Any] = len(input[0] )
for input_slice in input[1:]:
if len(_a ) != length:
return False
return True
def _inputs_are_equal(_a , _a ):
if len(_a ) != len(_a ):
return False
for input_slice_a, input_slice_a in zip(_a , _a ):
if not np.allclose(np.asarray(_a ) , np.asarray(_a ) , atol=1e-3 ):
return False
return True
__magic_name__ : List[str] = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : Any = self.feat_extract_tester.prepare_inputs_for_common(numpify=_a )
__magic_name__ : Any = feat_extract.model_input_names[0]
__magic_name__ : Union[str, Any] = BatchFeature({input_name: speech_inputs} )
# truncate to smallest
__magic_name__ : Tuple = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[0] ) , truncation=_a )
__magic_name__ : Union[str, Any] = input_a[input_name]
__magic_name__ : List[Any] = feat_extract.pad(_a , padding="max_length" , max_length=len(speech_inputs[0] ) )
__magic_name__ : Union[str, Any] = input_a[input_name]
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertFalse(_inputs_have_equal_length(_a ) )
# truncate to smallest with np
__magic_name__ : Optional[int] = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[0] ) , return_tensors="np" , truncation=_a , )
__magic_name__ : int = input_a[input_name]
__magic_name__ : Optional[int] = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[0] ) , return_tensors="np" )
__magic_name__ : Optional[Any] = input_a[input_name]
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertTrue(input_a.shape[1] == len(speech_inputs[0] ) )
# since truncation forces padding to be smaller than longest input
# function can't return `np.ndarray`, but has to return list
self.assertFalse(_inputs_have_equal_length(_a ) )
# truncate to middle
__magic_name__ : Optional[int] = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[1] ) , truncation=_a , return_tensors="np" , )
__magic_name__ : List[str] = input_a[input_name]
__magic_name__ : Tuple = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[1] ) , truncation=_a )
__magic_name__ : List[Any] = input_a[input_name]
__magic_name__ : int = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[1] ) , return_tensors="np" )
__magic_name__ : Union[str, Any] = input_a[input_name]
self.assertTrue(input_a.shape[1] == len(speech_inputs[1] ) )
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertTrue(_inputs_are_equal(_a , _a ) )
# since truncation forces padding to be smaller than longest input
# function can't return `np.ndarray`, but has to return list
self.assertFalse(_inputs_have_equal_length(_a ) )
self.assertTrue(len(input_a[-1] ) == len(speech_inputs[-1] ) )
# padding has to be max_length when setting `truncation=True`
with self.assertRaises(_a ):
feat_extract.pad(_a , truncation=_a )[input_name]
# padding has to be max_length when setting `truncation=True`
with self.assertRaises(_a ):
feat_extract.pad(_a , padding="longest" , truncation=_a )[input_name]
# padding has to be max_length when setting `truncation=True`
with self.assertRaises(_a ):
feat_extract.pad(_a , padding="longest" , truncation=_a )[input_name]
# max_length parameter has to be provided when setting `truncation=True` and padding="max_length"
with self.assertRaises(_a ):
feat_extract.pad(_a , padding="max_length" , truncation=_a )[input_name]
# test truncation for `pad_to_multiple_of` for List[int] + numpy
__magic_name__ : List[str] = 12
__magic_name__ : Tuple = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[0] ) , pad_to_multiple_of=_a , truncation=_a , )
__magic_name__ : Optional[int] = input_a[input_name]
__magic_name__ : Tuple = feat_extract.pad(
_a , padding="max_length" , max_length=len(speech_inputs[0] ) , pad_to_multiple_of=_a , )
__magic_name__ : str = input_a[input_name]
# retrieve expected_length as multiple of pad_to_multiple_of
__magic_name__ : List[Any] = len(speech_inputs[0] )
if expected_length % pad_to_multiple_of != 0:
__magic_name__ : Tuple = ((len(speech_inputs[0] ) // pad_to_multiple_of) + 1) * pad_to_multiple_of
self.assertTrue(len(input_a[0] ) == expected_length )
self.assertTrue(_inputs_have_equal_length(_a ) )
self.assertFalse(_inputs_have_equal_length(_a ) )
def SCREAMING_SNAKE_CASE ( self ):
self._check_padding(numpify=_a )
def SCREAMING_SNAKE_CASE ( self ):
self._check_padding(numpify=_a )
def SCREAMING_SNAKE_CASE ( self ):
self._check_truncation(numpify=_a )
def SCREAMING_SNAKE_CASE ( self ):
self._check_truncation(numpify=_a )
@require_torch
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : str = self.feat_extract_tester.prepare_inputs_for_common()
__magic_name__ : Optional[Any] = feat_extract.model_input_names[0]
__magic_name__ : List[Any] = BatchFeature({input_name: speech_inputs} )
__magic_name__ : List[str] = feat_extract.pad(_a , padding="longest" , return_tensors="np" )[input_name]
__magic_name__ : Dict = feat_extract.pad(_a , padding="longest" , return_tensors="pt" )[input_name]
self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_pt.numpy().astype(np.floataa ).sum() ) < 1e-2 )
@require_tf
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.feature_extraction_class(**self.feat_extract_dict )
__magic_name__ : Optional[int] = self.feat_extract_tester.prepare_inputs_for_common()
__magic_name__ : Any = feat_extract.model_input_names[0]
__magic_name__ : Union[str, Any] = BatchFeature({input_name: speech_inputs} )
__magic_name__ : Tuple = feat_extract.pad(_a , padding="longest" , return_tensors="np" )[input_name]
__magic_name__ : List[str] = feat_extract.pad(_a , padding="longest" , return_tensors="tf" )[input_name]
self.assertTrue(abs(input_np.astype(np.floataa ).sum() - input_tf.numpy().astype(np.floataa ).sum() ) < 1e-2 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.feat_extract_dict
__magic_name__ : int = True
__magic_name__ : Union[str, Any] = self.feature_extraction_class(**_a )
__magic_name__ : List[Any] = self.feat_extract_tester.prepare_inputs_for_common()
__magic_name__ : List[str] = [len(_a ) for x in speech_inputs]
__magic_name__ : Dict = feat_extract.model_input_names[0]
__magic_name__ : Union[str, Any] = BatchFeature({input_name: speech_inputs} )
__magic_name__ : Optional[Any] = feat_extract.pad(_a , padding="longest" , return_tensors="np" )
self.assertIn("attention_mask" , _a )
self.assertListEqual(list(processed.attention_mask.shape ) , list(processed[input_name].shape[:2] ) )
self.assertListEqual(processed.attention_mask.sum(-1 ).tolist() , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.feat_extract_dict
__magic_name__ : Dict = True
__magic_name__ : Any = self.feature_extraction_class(**_a )
__magic_name__ : str = self.feat_extract_tester.prepare_inputs_for_common()
__magic_name__ : Union[str, Any] = [len(_a ) for x in speech_inputs]
__magic_name__ : Dict = feat_extract.model_input_names[0]
__magic_name__ : List[Any] = BatchFeature({input_name: speech_inputs} )
__magic_name__ : Dict = min(_a )
__magic_name__ : Dict = feat_extract.pad(
_a , padding="max_length" , max_length=_a , truncation=_a , return_tensors="np" )
self.assertIn("attention_mask" , _a )
self.assertListEqual(
list(processed_pad.attention_mask.shape ) , [processed_pad[input_name].shape[0], max_length] )
self.assertListEqual(
processed_pad.attention_mask[:, :max_length].sum(-1 ).tolist() , [max_length for x in speech_inputs] )
| 281 |
import hashlib
import unittest
from typing import Dict
import numpy as np
from transformers import (
MODEL_FOR_MASK_GENERATION_MAPPING,
TF_MODEL_FOR_MASK_GENERATION_MAPPING,
is_vision_available,
pipeline,
)
from transformers.pipelines import MaskGenerationPipeline
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
)
if is_vision_available():
from PIL import Image
else:
class _snake_case :
@staticmethod
def SCREAMING_SNAKE_CASE ( *_a , **_a ):
pass
def lowerCAmelCase_ ( _snake_case : Image ) -> str:
'''simple docstring'''
__magic_name__ : Optional[int] = hashlib.mda(image.tobytes() )
return m.hexdigest()[:10]
def lowerCAmelCase_ ( _snake_case : Image ) -> Dict:
'''simple docstring'''
__magic_name__ : List[Any] = np.array(_snake_case )
__magic_name__ : Optional[int] = npimg.shape
return {"hash": hashimage(_snake_case ), "shape": shape}
@is_pipeline_test
@require_vision
@require_torch
class _snake_case ( unittest.TestCase ):
UpperCamelCase__ = dict(
(list(MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if MODEL_FOR_MASK_GENERATION_MAPPING else []) )
UpperCamelCase__ = dict(
(list(TF_MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if TF_MODEL_FOR_MASK_GENERATION_MAPPING else []) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Dict = MaskGenerationPipeline(model=_a , image_processor=_a )
return image_segmenter, [
"./tests/fixtures/tests_samples/COCO/000000039769.png",
"./tests/fixtures/tests_samples/COCO/000000039769.png",
]
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
pass
@require_tf
@unittest.skip("Image segmentation not implemented in TF" )
def SCREAMING_SNAKE_CASE ( self ):
pass
@slow
@require_torch
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = pipeline("mask-generation" , model="facebook/sam-vit-huge" )
__magic_name__ : str = image_segmenter("http://images.cocodataset.org/val2017/000000039769.jpg" , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Dict = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
# fmt: off
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.0_21},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
{"mask": {"hash": "e2d0b7a0b7", "shape": (480, 640)}, "scores": 0.99_67},
{"mask": {"hash": "453c7844bd", "shape": (480, 640)}, "scores": 0.9_93},
{"mask": {"hash": "3d44f2926d", "shape": (480, 640)}, "scores": 0.99_09},
{"mask": {"hash": "64033ddc3f", "shape": (480, 640)}, "scores": 0.98_79},
{"mask": {"hash": "801064ff79", "shape": (480, 640)}, "scores": 0.98_34},
{"mask": {"hash": "6172f276ef", "shape": (480, 640)}, "scores": 0.97_16},
{"mask": {"hash": "b49e60e084", "shape": (480, 640)}, "scores": 0.96_12},
{"mask": {"hash": "a811e775fd", "shape": (480, 640)}, "scores": 0.95_99},
{"mask": {"hash": "a6a8ebcf4b", "shape": (480, 640)}, "scores": 0.95_52},
{"mask": {"hash": "9d8257e080", "shape": (480, 640)}, "scores": 0.95_32},
{"mask": {"hash": "32de6454a8", "shape": (480, 640)}, "scores": 0.95_16},
{"mask": {"hash": "af3d4af2c8", "shape": (480, 640)}, "scores": 0.94_99},
{"mask": {"hash": "3c6db475fb", "shape": (480, 640)}, "scores": 0.94_83},
{"mask": {"hash": "c290813fb9", "shape": (480, 640)}, "scores": 0.94_64},
{"mask": {"hash": "b6f0b8f606", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "92ce16bfdf", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "c749b25868", "shape": (480, 640)}, "scores": 0.94_08},
{"mask": {"hash": "efb6cab859", "shape": (480, 640)}, "scores": 0.93_35},
{"mask": {"hash": "1ff2eafb30", "shape": (480, 640)}, "scores": 0.93_26},
{"mask": {"hash": "788b798e24", "shape": (480, 640)}, "scores": 0.92_62},
{"mask": {"hash": "abea804f0e", "shape": (480, 640)}, "scores": 0.89_99},
{"mask": {"hash": "7b9e8ddb73", "shape": (480, 640)}, "scores": 0.89_86},
{"mask": {"hash": "cd24047c8a", "shape": (480, 640)}, "scores": 0.89_84},
{"mask": {"hash": "6943e6bcbd", "shape": (480, 640)}, "scores": 0.88_73},
{"mask": {"hash": "b5f47c9191", "shape": (480, 640)}, "scores": 0.88_71}
] , )
# fmt: on
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = "facebook/sam-vit-huge"
__magic_name__ : str = pipeline("mask-generation" , model=_a )
__magic_name__ : Tuple = image_segmenter(
"http://images.cocodataset.org/val2017/000000039769.jpg" , pred_iou_thresh=1 , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Any = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.02_10},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
] , )
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : int = 100 ) -> int:
'''simple docstring'''
__magic_name__ : Any = n * (n + 1) * (2 * n + 1) / 6
__magic_name__ : List[str] = (n * (n + 1) / 2) ** 2
return int(square_of_sum - sum_of_squares )
if __name__ == "__main__":
print(F"{solution() = }")
| 281 |
import absl # noqa: F401 # Here to have a nice missing dependency error message early on
import nltk # noqa: F401 # Here to have a nice missing dependency error message early on
import numpy # noqa: F401 # Here to have a nice missing dependency error message early on
import six # noqa: F401 # Here to have a nice missing dependency error message early on
from rouge_score import rouge_scorer, scoring
import datasets
snake_case : List[Any] = "\\n@inproceedings{lin-2004-rouge,\n title = \"{ROUGE}: A Package for Automatic Evaluation of Summaries\",\n author = \"Lin, Chin-Yew\",\n booktitle = \"Text Summarization Branches Out\",\n month = jul,\n year = \"2004\",\n address = \"Barcelona, Spain\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W04-1013\",\n pages = \"74--81\",\n}\n"
snake_case : Any = "\\nROUGE, or Recall-Oriented Understudy for Gisting Evaluation, is a set of metrics and a software package used for\nevaluating automatic summarization and machine translation software in natural language processing.\nThe metrics compare an automatically produced summary or translation against a reference or a set of references (human-produced) summary or translation.\n\nNote that ROUGE is case insensitive, meaning that upper case letters are treated the same way as lower case letters.\n\nThis metrics is a wrapper around Google Research reimplementation of ROUGE:\nhttps://github.com/google-research/google-research/tree/master/rouge\n"
snake_case : str = "\nCalculates average rouge scores for a list of hypotheses and references\nArgs:\n predictions: list of predictions to score. Each prediction\n should be a string with tokens separated by spaces.\n references: list of reference for each prediction. Each\n reference should be a string with tokens separated by spaces.\n rouge_types: A list of rouge types to calculate.\n Valid names:\n `\"rouge{n}\"` (e.g. `\"rouge1\"`, `\"rouge2\"`) where: {n} is the n-gram based scoring,\n `\"rougeL\"`: Longest common subsequence based scoring.\n `\"rougeLSum\"`: rougeLsum splits text using `\"\n\"`.\n See details in https://github.com/huggingface/datasets/issues/617\n use_stemmer: Bool indicating whether Porter stemmer should be used to strip word suffixes.\n use_aggregator: Return aggregates if this is set to True\nReturns:\n rouge1: rouge_1 (precision, recall, f1),\n rouge2: rouge_2 (precision, recall, f1),\n rougeL: rouge_l (precision, recall, f1),\n rougeLsum: rouge_lsum (precision, recall, f1)\nExamples:\n\n >>> rouge = datasets.load_metric('rouge')\n >>> predictions = [\"hello there\", \"general kenobi\"]\n >>> references = [\"hello there\", \"general kenobi\"]\n >>> results = rouge.compute(predictions=predictions, references=references)\n >>> print(list(results.keys()))\n ['rouge1', 'rouge2', 'rougeL', 'rougeLsum']\n >>> print(results[\"rouge1\"])\n AggregateScore(low=Score(precision=1.0, recall=1.0, fmeasure=1.0), mid=Score(precision=1.0, recall=1.0, fmeasure=1.0), high=Score(precision=1.0, recall=1.0, fmeasure=1.0))\n >>> print(results[\"rouge1\"].mid.fmeasure)\n 1.0\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class _snake_case ( datasets.Metric ):
def SCREAMING_SNAKE_CASE ( self ):
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , 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/google-research/tree/master/rouge"] , reference_urls=[
"https://en.wikipedia.org/wiki/ROUGE_(metric)",
"https://github.com/google-research/google-research/tree/master/rouge",
] , )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a=None , _a=True , _a=False ):
if rouge_types is None:
__magic_name__ : str = ["rouge1", "rouge2", "rougeL", "rougeLsum"]
__magic_name__ : List[str] = rouge_scorer.RougeScorer(rouge_types=_a , use_stemmer=_a )
if use_aggregator:
__magic_name__ : Dict = scoring.BootstrapAggregator()
else:
__magic_name__ : str = []
for ref, pred in zip(_a , _a ):
__magic_name__ : Union[str, Any] = scorer.score(_a , _a )
if use_aggregator:
aggregator.add_scores(_a )
else:
scores.append(_a )
if use_aggregator:
__magic_name__ : Any = aggregator.aggregate()
else:
__magic_name__ : List[Any] = {}
for key in scores[0]:
__magic_name__ : str = [score[key] for score in scores]
return result
| 281 | 1 |
from collections import OrderedDict
from typing import Any, Mapping, Optional, Union
from ...configuration_utils import PretrainedConfig
from ...feature_extraction_utils import FeatureExtractionMixin
from ...onnx import OnnxConfig
from ...onnx.utils import compute_effective_axis_dimension
from ...tokenization_utils_base import PreTrainedTokenizerBase
from ...utils import TensorType, logging
snake_case : Optional[int] = logging.get_logger(__name__)
snake_case : str = {
"deepmind/language-perceiver": "https://huggingface.co/deepmind/language-perceiver/resolve/main/config.json",
# See all Perceiver models at https://huggingface.co/models?filter=perceiver
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'perceiver'
def __init__( self , _a=256 , _a=1_280 , _a=768 , _a=1 , _a=26 , _a=8 , _a=8 , _a=None , _a=None , _a="kv" , _a=1 , _a=1 , _a="gelu" , _a=0.1 , _a=0.02 , _a=1e-12 , _a=True , _a=262 , _a=2_048 , _a=56 , _a=[368, 496] , _a=16 , _a=1_920 , _a=16 , _a=[1, 16, 224, 224] , **_a , ):
super().__init__(**_a )
__magic_name__ : str = num_latents
__magic_name__ : List[Any] = d_latents
__magic_name__ : Tuple = d_model
__magic_name__ : int = num_blocks
__magic_name__ : Tuple = num_self_attends_per_block
__magic_name__ : Union[str, Any] = num_self_attention_heads
__magic_name__ : List[str] = num_cross_attention_heads
__magic_name__ : Tuple = qk_channels
__magic_name__ : Dict = v_channels
__magic_name__ : Optional[Any] = cross_attention_shape_for_attention
__magic_name__ : Optional[Any] = self_attention_widening_factor
__magic_name__ : Union[str, Any] = cross_attention_widening_factor
__magic_name__ : Union[str, Any] = hidden_act
__magic_name__ : int = attention_probs_dropout_prob
__magic_name__ : int = initializer_range
__magic_name__ : str = layer_norm_eps
__magic_name__ : List[Any] = use_query_residual
# masked language modeling attributes
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : Tuple = max_position_embeddings
# image classification attributes
__magic_name__ : List[Any] = image_size
# flow attributes
__magic_name__ : Optional[Any] = train_size
# multimodal autoencoding attributes
__magic_name__ : List[str] = num_frames
__magic_name__ : Optional[Any] = audio_samples_per_frame
__magic_name__ : Optional[int] = samples_per_patch
__magic_name__ : Optional[int] = output_shape
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : List[str] = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("inputs", dynamic_axis),
("attention_mask", dynamic_axis),
] )
@property
def SCREAMING_SNAKE_CASE ( self ):
return 1e-4
def SCREAMING_SNAKE_CASE ( self , _a , _a = -1 , _a = -1 , _a = -1 , _a = False , _a = None , _a = 3 , _a = 40 , _a = 40 , ):
# copied from `transformers.onnx.config.OnnxConfig` and slightly altered/simplified
if isinstance(_a , _a ):
# If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX
__magic_name__ : int = compute_effective_axis_dimension(
_a , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 )
# If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX
__magic_name__ : Any = preprocessor.num_special_tokens_to_add(_a )
__magic_name__ : List[Any] = compute_effective_axis_dimension(
_a , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=_a )
# Generate dummy inputs according to compute batch and sequence
__magic_name__ : str = [" ".join(["a"] ) * seq_length] * batch_size
__magic_name__ : List[str] = dict(preprocessor(_a , return_tensors=_a ) )
__magic_name__ : List[str] = inputs.pop("input_ids" )
return inputs
elif isinstance(_a , _a ) and preprocessor.model_input_names[0] == "pixel_values":
# If dynamic axis (-1) we forward with a fixed dimension of 2 samples to avoid optimizations made by ONNX
__magic_name__ : Optional[Any] = compute_effective_axis_dimension(_a , fixed_dimension=OnnxConfig.default_fixed_batch )
__magic_name__ : Optional[Any] = self._generate_dummy_images(_a , _a , _a , _a )
__magic_name__ : List[Any] = dict(preprocessor(images=_a , return_tensors=_a ) )
__magic_name__ : str = inputs.pop("pixel_values" )
return inputs
else:
raise ValueError(
"Unable to generate dummy inputs for the model. Please provide a tokenizer or a preprocessor." )
| 281 |
snake_case : Optional[int] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"
def lowerCAmelCase_ ( _snake_case : bytes ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ):
__magic_name__ : Tuple = F'''a bytes-like object is required, not \'{data.__class__.__name__}\''''
raise TypeError(_snake_case )
__magic_name__ : Optional[int] = "".join(bin(_snake_case )[2:].zfill(8 ) for byte in data )
__magic_name__ : List[Any] = len(_snake_case ) % 6 != 0
if padding_needed:
# The padding that will be added later
__magic_name__ : List[str] = B"=" * ((6 - len(_snake_case ) % 6) // 2)
# Append binary_stream with arbitrary binary digits (0's by default) to make its
# length a multiple of 6.
binary_stream += "0" * (6 - len(_snake_case ) % 6)
else:
__magic_name__ : List[str] = B""
# Encode every 6 binary digits to their corresponding Base64 character
return (
"".join(
B64_CHARSET[int(binary_stream[index : index + 6] , 2 )]
for index in range(0 , len(_snake_case ) , 6 ) ).encode()
+ padding
)
def lowerCAmelCase_ ( _snake_case : str ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ) and not isinstance(_snake_case , _snake_case ):
__magic_name__ : List[str] = (
"argument should be a bytes-like object or ASCII string, "
F'''not \'{encoded_data.__class__.__name__}\''''
)
raise TypeError(_snake_case )
# In case encoded_data is a bytes-like object, make sure it contains only
# ASCII characters so we convert it to a string object
if isinstance(_snake_case , _snake_case ):
try:
__magic_name__ : List[Any] = encoded_data.decode("utf-8" )
except UnicodeDecodeError:
raise ValueError("base64 encoded data should only contain ASCII characters" )
__magic_name__ : List[str] = encoded_data.count("=" )
# Check if the encoded string contains non base64 characters
if padding:
assert all(
char in B64_CHARSET for char in encoded_data[:-padding] ), "Invalid base64 character(s) found."
else:
assert all(
char in B64_CHARSET for char in encoded_data ), "Invalid base64 character(s) found."
# Check the padding
assert len(_snake_case ) % 4 == 0 and padding < 3, "Incorrect padding"
if padding:
# Remove padding if there is one
__magic_name__ : Optional[int] = encoded_data[:-padding]
__magic_name__ : Dict = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )[: -padding * 2]
else:
__magic_name__ : Union[str, Any] = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )
__magic_name__ : List[Any] = [
int(binary_stream[index : index + 8] , 2 )
for index in range(0 , len(_snake_case ) , 8 )
]
return bytes(_snake_case )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
from abc import ABC, abstractmethod
from typing import List, Optional
class _snake_case ( snake_case ):
def __init__( self ):
# test for the above condition
self.test()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = 0
__magic_name__ : Dict = False
while not completed:
if counter == 1:
self.reset()
__magic_name__ : Dict = self.advance()
if not self.does_advance(_a ):
raise Exception(
"Custom Constraint is not defined correctly. self.does_advance(self.advance()) must be true." )
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = self.update(_a )
counter += 1
if counter > 10_000:
raise Exception("update() does not fulfill the constraint." )
if self.remaining() != 0:
raise Exception("Custom Constraint is not defined correctly." )
@abstractmethod
def SCREAMING_SNAKE_CASE ( self ):
raise NotImplementedError(
f'''{self.__class__} is an abstract class. Only classes inheriting this class can be called.''' )
@abstractmethod
def SCREAMING_SNAKE_CASE ( self , _a ):
raise NotImplementedError(
f'''{self.__class__} is an abstract class. Only classes inheriting this class can be called.''' )
@abstractmethod
def SCREAMING_SNAKE_CASE ( self , _a ):
raise NotImplementedError(
f'''{self.__class__} is an abstract class. Only classes inheriting this class can be called.''' )
@abstractmethod
def SCREAMING_SNAKE_CASE ( self ):
raise NotImplementedError(
f'''{self.__class__} is an abstract class. Only classes inheriting this class can be called.''' )
@abstractmethod
def SCREAMING_SNAKE_CASE ( self ):
raise NotImplementedError(
f'''{self.__class__} is an abstract class. Only classes inheriting this class can be called.''' )
@abstractmethod
def SCREAMING_SNAKE_CASE ( self , _a=False ):
raise NotImplementedError(
f'''{self.__class__} is an abstract class. Only classes inheriting this class can be called.''' )
class _snake_case ( snake_case ):
def __init__( self , _a ):
super(_a , self ).__init__()
if not isinstance(_a , _a ) or len(_a ) == 0:
raise ValueError(f'''`token_ids` has to be a non-empty list, but is {token_ids}.''' )
if any((not isinstance(_a , _a ) or token_id < 0) for token_id in token_ids ):
raise ValueError(f'''Each list in `token_ids` has to be a list of positive integers, but is {token_ids}.''' )
__magic_name__ : List[str] = token_ids
__magic_name__ : List[str] = len(self.token_ids )
__magic_name__ : Dict = -1 # the index of the currently fulfilled step
__magic_name__ : Optional[int] = False
def SCREAMING_SNAKE_CASE ( self ):
if self.completed:
return None
return self.token_ids[self.fulfilled_idx + 1]
def SCREAMING_SNAKE_CASE ( self , _a ):
if not isinstance(_a , _a ):
raise ValueError(f'''`token_id` has to be an `int`, but is {token_id} of type {type(_a )}''' )
if self.completed:
return False
return token_id == self.token_ids[self.fulfilled_idx + 1]
def SCREAMING_SNAKE_CASE ( self , _a ):
if not isinstance(_a , _a ):
raise ValueError(f'''`token_id` has to be an `int`, but is {token_id} of type {type(_a )}''' )
__magic_name__ : Tuple = False
__magic_name__ : List[str] = False
__magic_name__ : Any = False
if self.does_advance(_a ):
self.fulfilled_idx += 1
__magic_name__ : Optional[int] = True
if self.fulfilled_idx == (self.seqlen - 1):
__magic_name__ : Any = True
__magic_name__ : Optional[int] = completed
else:
# failed to make progress.
__magic_name__ : Any = True
self.reset()
return stepped, completed, reset
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = False
__magic_name__ : str = 0
def SCREAMING_SNAKE_CASE ( self ):
return self.seqlen - (self.fulfilled_idx + 1)
def SCREAMING_SNAKE_CASE ( self , _a=False ):
__magic_name__ : str = PhrasalConstraint(self.token_ids )
if stateful:
__magic_name__ : str = self.seqlen
__magic_name__ : Any = self.fulfilled_idx
__magic_name__ : Tuple = self.completed
return new_constraint
class _snake_case :
def __init__( self , _a , _a=True ):
__magic_name__ : Union[str, Any] = max([len(_a ) for one in nested_token_ids] )
__magic_name__ : List[Any] = {}
for token_ids in nested_token_ids:
__magic_name__ : Optional[int] = root
for tidx, token_id in enumerate(_a ):
if token_id not in level:
__magic_name__ : List[Any] = {}
__magic_name__ : Optional[int] = level[token_id]
if no_subsets and self.has_subsets(_a , _a ):
raise ValueError(
"Each list in `nested_token_ids` can't be a complete subset of another list, but is"
f''' {nested_token_ids}.''' )
__magic_name__ : int = root
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : int = self.trie
for current_token in current_seq:
__magic_name__ : Union[str, Any] = start[current_token]
__magic_name__ : int = list(start.keys() )
return next_tokens
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : int = self.next_tokens(_a )
return len(_a ) == 0
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Tuple = list(root.values() )
if len(_a ) == 0:
return 1
else:
return sum([self.count_leaves(_a ) for nn in next_nodes] )
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : List[Any] = self.count_leaves(_a )
return len(_a ) != leaf_count
class _snake_case ( snake_case ):
def __init__( self , _a ):
super(_a , self ).__init__()
if not isinstance(_a , _a ) or len(_a ) == 0:
raise ValueError(f'''`nested_token_ids` has to be a non-empty list, but is {nested_token_ids}.''' )
if any(not isinstance(_a , _a ) for token_ids in nested_token_ids ):
raise ValueError(f'''`nested_token_ids` has to be a list of lists, but is {nested_token_ids}.''' )
if any(
any((not isinstance(_a , _a ) or token_id < 0) for token_id in token_ids )
for token_ids in nested_token_ids ):
raise ValueError(
f'''Each list in `nested_token_ids` has to be a list of positive integers, but is {nested_token_ids}.''' )
__magic_name__ : Any = DisjunctiveTrie(_a )
__magic_name__ : str = nested_token_ids
__magic_name__ : List[str] = self.trie.max_height
__magic_name__ : Optional[Any] = []
__magic_name__ : Union[str, Any] = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = self.trie.next_tokens(self.current_seq )
if len(_a ) == 0:
return None
else:
return token_list
def SCREAMING_SNAKE_CASE ( self , _a ):
if not isinstance(_a , _a ):
raise ValueError(f'''`token_id` is supposed to be type `int`, but is {token_id} of type {type(_a )}''' )
__magic_name__ : List[Any] = self.trie.next_tokens(self.current_seq )
return token_id in next_tokens
def SCREAMING_SNAKE_CASE ( self , _a ):
if not isinstance(_a , _a ):
raise ValueError(f'''`token_id` is supposed to be type `int`, but is {token_id} of type {type(_a )}''' )
__magic_name__ : str = False
__magic_name__ : List[str] = False
__magic_name__ : Tuple = False
if self.does_advance(_a ):
self.current_seq.append(_a )
__magic_name__ : Optional[int] = True
else:
__magic_name__ : Optional[int] = True
self.reset()
__magic_name__ : str = self.trie.reached_leaf(self.current_seq )
__magic_name__ : int = completed
return stepped, completed, reset
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = False
__magic_name__ : Dict = []
def SCREAMING_SNAKE_CASE ( self ):
if self.completed:
# since this can be completed without reaching max height
return 0
else:
return self.seqlen - len(self.current_seq )
def SCREAMING_SNAKE_CASE ( self , _a=False ):
__magic_name__ : Union[str, Any] = DisjunctiveConstraint(self.token_ids )
if stateful:
__magic_name__ : Any = self.seqlen
__magic_name__ : Dict = self.current_seq
__magic_name__ : Optional[int] = self.completed
return new_constraint
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Optional[int] = constraints
# max # of steps required to fulfill a given constraint
__magic_name__ : int = max([c.seqlen for c in constraints] )
__magic_name__ : List[Any] = len(_a )
__magic_name__ : Tuple = False
self.init_state()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = []
__magic_name__ : Any = None
__magic_name__ : Optional[Any] = [constraint.copy(stateful=_a ) for constraint in self.constraints]
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = 0
if self.inprogress_constraint:
# extra points for having a constraint mid-fulfilled
add += self.max_seqlen - self.inprogress_constraint.remaining()
return (len(self.complete_constraints ) * self.max_seqlen) + add
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = []
if self.inprogress_constraint is None:
for constraint in self.pending_constraints: # "pending" == "unfulfilled yet"
__magic_name__ : Optional[Any] = constraint.advance()
if isinstance(_a , _a ):
token_list.append(_a )
elif isinstance(_a , _a ):
token_list.extend(_a )
else:
__magic_name__ : List[str] = self.inprogress_constraint.advance()
if isinstance(_a , _a ):
token_list.append(_a )
elif isinstance(_a , _a ):
token_list.extend(_a )
if len(_a ) == 0:
return None
else:
return token_list
def SCREAMING_SNAKE_CASE ( self , _a ):
self.init_state()
if token_ids is not None:
for token in token_ids:
# completes or steps **one** constraint
__magic_name__ , __magic_name__ : Optional[Any] = self.add(_a )
# the entire list of constraints are fulfilled
if self.completed:
break
def SCREAMING_SNAKE_CASE ( self , _a ):
if not isinstance(_a , _a ):
raise ValueError(f'''`token_id` should be an `int`, but is `{token_id}`.''' )
__magic_name__ , __magic_name__ : int = False, False
if self.completed:
__magic_name__ : str = True
__magic_name__ : List[Any] = False
return complete, stepped
if self.inprogress_constraint is not None:
# In the middle of fulfilling a constraint. If the `token_id` *does* makes an incremental progress to current
# job, simply update the state
__magic_name__ , __magic_name__ , __magic_name__ : List[str] = self.inprogress_constraint.update(_a )
if reset:
# 1. If the next token breaks the progress, then we must restart.
# e.g. constraint = "I love pies" and sequence so far is "I love" but `token_id` == "books".
# But that doesn't mean we self.init_state(), since we only reset the state for this particular
# constraint, not the full list of constraints.
self.pending_constraints.append(self.inprogress_constraint.copy(stateful=_a ) )
__magic_name__ : Optional[int] = None
if complete:
# 2. If the next token completes the constraint, move it to completed list, set
# inprogress to None. If there are no pending constraints either, then this full list of constraints
# is complete.
self.complete_constraints.append(self.inprogress_constraint )
__magic_name__ : Dict = None
if len(self.pending_constraints ) == 0:
# we're done!
__magic_name__ : Union[str, Any] = True
else:
# Not in the middle of fulfilling a constraint. So does this `token_id` helps us step towards any of our list
# of constraints?
for cidx, pending_constraint in enumerate(self.pending_constraints ):
if pending_constraint.does_advance(_a ):
__magic_name__ , __magic_name__ , __magic_name__ : str = pending_constraint.update(_a )
if not stepped:
raise Exception(
"`constraint.update(token_id)` is not yielding incremental progress, "
"even though `constraint.does_advance(token_id)` is true." )
if complete:
self.complete_constraints.append(_a )
__magic_name__ : Any = None
if not complete and stepped:
__magic_name__ : str = pending_constraint
if complete or stepped:
# If we made any progress at all, then it's at least not a "pending constraint".
__magic_name__ : Optional[int] = (
self.pending_constraints[:cidx] + self.pending_constraints[cidx + 1 :]
)
if len(self.pending_constraints ) == 0 and self.inprogress_constraint is None:
# If there's no longer any pending after this and no inprogress either, then we must be
# complete.
__magic_name__ : Optional[Any] = True
break # prevent accidentally stepping through multiple constraints with just one token.
return complete, stepped
def SCREAMING_SNAKE_CASE ( self , _a=True ):
__magic_name__ : Optional[Any] = ConstraintListState(self.constraints ) # we actually never though self.constraints objects
# throughout this process. So it's at initialization state.
if stateful:
__magic_name__ : Optional[Any] = [
constraint.copy(stateful=_a ) for constraint in self.complete_constraints
]
if self.inprogress_constraint is not None:
__magic_name__ : Optional[int] = self.inprogress_constraint.copy(stateful=_a )
__magic_name__ : str = [constraint.copy() for constraint in self.pending_constraints]
return new_state
| 281 |
import unittest
import numpy as np
from transformers import RobertaPreLayerNormConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
import jax.numpy as jnp
from transformers.models.roberta_prelayernorm.modeling_flax_roberta_prelayernorm import (
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormModel,
)
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=True , _a=True , _a=True , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=16 , _a=2 , _a=0.02 , _a=4 , ):
__magic_name__ : List[Any] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Dict = seq_length
__magic_name__ : Union[str, Any] = is_training
__magic_name__ : Optional[Any] = use_attention_mask
__magic_name__ : Optional[Any] = use_token_type_ids
__magic_name__ : int = use_labels
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : Any = intermediate_size
__magic_name__ : List[Any] = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[Any] = max_position_embeddings
__magic_name__ : Tuple = type_vocab_size
__magic_name__ : List[str] = type_sequence_label_size
__magic_name__ : Dict = initializer_range
__magic_name__ : List[Any] = num_choices
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = None
if self.use_attention_mask:
__magic_name__ : List[str] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : str = None
if self.use_token_type_ids:
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__magic_name__ : List[str] = RobertaPreLayerNormConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=_a , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[Any] = config_and_inputs
__magic_name__ : List[str] = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": attention_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : Tuple = True
__magic_name__ : int = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
token_type_ids,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
# Copied from tests.models.roberta.test_modelling_flax_roberta.FlaxRobertaPreLayerNormModelTest with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta-base->andreasmadsen/efficient_mlm_m0.40
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = True
UpperCamelCase__ = (
(
FlaxRobertaPreLayerNormModel,
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
)
if is_flax_available()
else ()
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxRobertaPreLayerNormModelTester(self )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Optional[Any] = model_class_name.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(_a )
@require_flax
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxRobertaPreLayerNormForMaskedLM.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Union[str, Any] = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : List[str] = model(_a )[0]
__magic_name__ : str = [1, 11, 50_265]
self.assertEqual(list(output.shape ) , _a )
# compare the actual values for a slice.
__magic_name__ : List[str] = np.array(
[[[40.48_80, 18.01_99, -5.23_67], [-1.88_77, -4.08_85, 10.70_85], [-2.26_13, -5.61_10, 7.26_65]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = FlaxRobertaPreLayerNormModel.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Tuple = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : Tuple = model(_a )[0]
# compare the actual values for a slice.
__magic_name__ : Dict = np.array(
[[[0.02_08, -0.03_56, 0.02_37], [-0.15_69, -0.04_11, -0.26_26], [0.18_79, 0.01_25, -0.00_89]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
| 281 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
snake_case : Union[str, Any] = logging.get_logger(__name__)
snake_case : Any = {}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'llama'
UpperCamelCase__ = ['past_key_values']
def __init__( self , _a=32_000 , _a=4_096 , _a=11_008 , _a=32 , _a=32 , _a=None , _a="silu" , _a=2_048 , _a=0.02 , _a=1e-6 , _a=True , _a=0 , _a=1 , _a=2 , _a=1 , _a=False , _a=None , **_a , ):
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Dict = hidden_size
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Any = num_hidden_layers
__magic_name__ : Union[str, Any] = num_attention_heads
# for backward compatibility
if num_key_value_heads is None:
__magic_name__ : List[str] = num_attention_heads
__magic_name__ : Union[str, Any] = num_key_value_heads
__magic_name__ : str = hidden_act
__magic_name__ : List[str] = initializer_range
__magic_name__ : List[str] = rms_norm_eps
__magic_name__ : Optional[int] = pretraining_tp
__magic_name__ : Any = use_cache
__magic_name__ : Optional[int] = rope_scaling
self._rope_scaling_validation()
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , tie_word_embeddings=_a , **_a , )
def SCREAMING_SNAKE_CASE ( self ):
if self.rope_scaling is None:
return
if not isinstance(self.rope_scaling , _a ) or len(self.rope_scaling ) != 2:
raise ValueError(
"`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, "
f'''got {self.rope_scaling}''' )
__magic_name__ : Optional[Any] = self.rope_scaling.get("type" , _a )
__magic_name__ : Dict = self.rope_scaling.get("factor" , _a )
if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
raise ValueError(
f'''`rope_scaling`\'s name field must be one of [\'linear\', \'dynamic\'], got {rope_scaling_type}''' )
if rope_scaling_factor is None or not isinstance(_a , _a ) or rope_scaling_factor <= 1.0:
raise ValueError(f'''`rope_scaling`\'s factor field must be an float > 1, got {rope_scaling_factor}''' )
| 281 |
def lowerCAmelCase_ ( _snake_case : list[list[int | float]] ) -> int:
'''simple docstring'''
__magic_name__ : Any = len(_snake_case )
__magic_name__ : Optional[Any] = len(matrix[0] )
__magic_name__ : Union[str, Any] = min(_snake_case , _snake_case )
for row in range(_snake_case ):
# Check if diagonal element is not zero
if matrix[row][row] != 0:
# Eliminate all the elements below the diagonal
for col in range(row + 1 , _snake_case ):
__magic_name__ : Optional[Any] = matrix[col][row] / matrix[row][row]
for i in range(_snake_case , _snake_case ):
matrix[col][i] -= multiplier * matrix[row][i]
else:
# Find a non-zero diagonal element to swap rows
__magic_name__ : str = True
for i in range(row + 1 , _snake_case ):
if matrix[i][row] != 0:
__magic_name__ , __magic_name__ : List[str] = matrix[i], matrix[row]
__magic_name__ : Union[str, Any] = False
break
if reduce:
rank -= 1
for i in range(_snake_case ):
__magic_name__ : Any = matrix[i][rank]
# Reduce the row pointer by one to stay on the same row
row -= 1
return rank
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
import gc
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
PNDMScheduler,
StableDiffusionLDMaDPipeline,
UNetaDConditionModel,
)
from diffusers.utils import nightly, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
enable_full_determinism()
class _snake_case ( unittest.TestCase ):
UpperCamelCase__ = StableDiffusionLDMaDPipeline
UpperCamelCase__ = TEXT_TO_IMAGE_PARAMS
UpperCamelCase__ = TEXT_TO_IMAGE_BATCH_PARAMS
UpperCamelCase__ = TEXT_TO_IMAGE_IMAGE_PARAMS
def SCREAMING_SNAKE_CASE ( self ):
torch.manual_seed(0 )
__magic_name__ : Tuple = 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 , )
__magic_name__ : int = DDIMScheduler(
beta_start=0.0_00_85 , beta_end=0.0_12 , beta_schedule="scaled_linear" , clip_sample=_a , set_alpha_to_one=_a , )
torch.manual_seed(0 )
__magic_name__ : int = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=6 , out_channels=6 , down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"] , up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"] , latent_channels=4 , )
torch.manual_seed(0 )
__magic_name__ : List[str] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
__magic_name__ : Optional[Any] = CLIPTextModel(_a )
__magic_name__ : Union[str, Any] = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip" )
__magic_name__ : List[Any] = {
"unet": unet,
"scheduler": scheduler,
"vae": vae,
"text_encoder": text_encoder,
"tokenizer": tokenizer,
"safety_checker": None,
"feature_extractor": None,
}
return components
def SCREAMING_SNAKE_CASE ( self , _a , _a=0 ):
if str(_a ).startswith("mps" ):
__magic_name__ : Tuple = torch.manual_seed(_a )
else:
__magic_name__ : Optional[int] = torch.Generator(device=_a ).manual_seed(_a )
__magic_name__ : Optional[Any] = {
"prompt": "A painting of a squirrel eating a burger",
"generator": generator,
"num_inference_steps": 2,
"guidance_scale": 6.0,
"output_type": "numpy",
}
return inputs
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = "cpu" # ensure determinism for the device-dependent torch.Generator
__magic_name__ : Optional[int] = self.get_dummy_components()
__magic_name__ : List[str] = StableDiffusionLDMaDPipeline(**_a )
__magic_name__ : List[Any] = ldmad_pipe.to(_a )
ldmad_pipe.set_progress_bar_config(disable=_a )
__magic_name__ : Optional[Any] = self.get_dummy_inputs(_a )
__magic_name__ : List[str] = ldmad_pipe(**_a )
__magic_name__ , __magic_name__ : Tuple = output.rgb, output.depth
__magic_name__ : Union[str, Any] = rgb[0, -3:, -3:, -1]
__magic_name__ : List[str] = depth[0, -3:, -1]
assert rgb.shape == (1, 64, 64, 3)
assert depth.shape == (1, 64, 64)
__magic_name__ : str = np.array(
[0.37_33_81_76, 0.7_02_47, 0.74_20_31_93, 0.51_64_36_04, 0.58_25_67_93, 0.60_93_21_36, 0.4_18_10_95, 0.48_35_58_77, 0.46_53_52_62] )
__magic_name__ : Optional[int] = np.array([1_03.4_67_27, 85.81_20_04, 87.84_92_36] )
assert np.abs(image_slice_rgb.flatten() - expected_slice_rgb ).max() < 1e-2
assert np.abs(image_slice_depth.flatten() - expected_slice_depth ).max() < 1e-2
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = self.get_dummy_components()
__magic_name__ : Tuple = StableDiffusionLDMaDPipeline(**_a )
__magic_name__ : Optional[Any] = ldmad_pipe.to(_a )
ldmad_pipe.set_progress_bar_config(disable=_a )
__magic_name__ : Any = self.get_dummy_inputs(_a )
__magic_name__ : List[str] = 3 * [inputs["prompt"]]
# forward
__magic_name__ : Union[str, Any] = ldmad_pipe(**_a )
__magic_name__ , __magic_name__ : int = output.rgb, output.depth
__magic_name__ : Union[str, Any] = rgb_slice_a[0, -3:, -3:, -1]
__magic_name__ : List[str] = depth_slice_a[0, -3:, -1]
__magic_name__ : List[str] = self.get_dummy_inputs(_a )
__magic_name__ : List[str] = 3 * [inputs.pop("prompt" )]
__magic_name__ : List[Any] = ldmad_pipe.tokenizer(
_a , padding="max_length" , max_length=ldmad_pipe.tokenizer.model_max_length , truncation=_a , return_tensors="pt" , )
__magic_name__ : Any = text_inputs["input_ids"].to(_a )
__magic_name__ : Optional[int] = ldmad_pipe.text_encoder(_a )[0]
__magic_name__ : List[Any] = prompt_embeds
# forward
__magic_name__ : Dict = ldmad_pipe(**_a )
__magic_name__ , __magic_name__ : Optional[int] = output.rgb, output.depth
__magic_name__ : str = rgb_slice_a[0, -3:, -3:, -1]
__magic_name__ : Union[str, Any] = depth_slice_a[0, -3:, -1]
assert np.abs(rgb_slice_a.flatten() - rgb_slice_a.flatten() ).max() < 1e-4
assert np.abs(depth_slice_a.flatten() - depth_slice_a.flatten() ).max() < 1e-4
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = "cpu" # ensure determinism for the device-dependent torch.Generator
__magic_name__ : List[Any] = self.get_dummy_components()
__magic_name__ : List[str] = PNDMScheduler(skip_prk_steps=_a )
__magic_name__ : Optional[Any] = StableDiffusionLDMaDPipeline(**_a )
__magic_name__ : Union[str, Any] = ldmad_pipe.to(_a )
ldmad_pipe.set_progress_bar_config(disable=_a )
__magic_name__ : str = self.get_dummy_inputs(_a )
__magic_name__ : Tuple = "french fries"
__magic_name__ : List[Any] = ldmad_pipe(**_a , negative_prompt=_a )
__magic_name__ , __magic_name__ : Union[str, Any] = output.rgb, output.depth
__magic_name__ : Dict = rgb[0, -3:, -3:, -1]
__magic_name__ : List[str] = depth[0, -3:, -1]
assert rgb.shape == (1, 64, 64, 3)
assert depth.shape == (1, 64, 64)
__magic_name__ : Dict = np.array(
[0.3_70_44, 0.71_81_15_03, 0.7_22_32_51, 0.48_60_36_75, 0.5_63_83_91, 0.6_36_49_48, 0.42_83_37_04, 0.4_90_13_15, 0.47_92_62_17] )
__magic_name__ : Dict = np.array([1_07.8_47_38, 84.6_28_02, 89.96_21_35] )
assert np.abs(rgb_slice.flatten() - expected_slice_rgb ).max() < 1e-2
assert np.abs(depth_slice.flatten() - expected_slice_depth ).max() < 1e-2
@slow
@require_torch_gpu
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def SCREAMING_SNAKE_CASE ( self , _a , _a="cpu" , _a=torch.floataa , _a=0 ):
__magic_name__ : Tuple = torch.Generator(device=_a ).manual_seed(_a )
__magic_name__ : Union[str, Any] = np.random.RandomState(_a ).standard_normal((1, 4, 64, 64) )
__magic_name__ : Tuple = torch.from_numpy(_a ).to(device=_a , dtype=_a )
__magic_name__ : Union[str, Any] = {
"prompt": "a photograph of an astronaut riding a horse",
"latents": latents,
"generator": generator,
"num_inference_steps": 3,
"guidance_scale": 7.5,
"output_type": "numpy",
}
return inputs
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = StableDiffusionLDMaDPipeline.from_pretrained("Intel/ldm3d" )
__magic_name__ : Optional[Any] = ldmad_pipe.to(_a )
ldmad_pipe.set_progress_bar_config(disable=_a )
__magic_name__ : List[Any] = self.get_inputs(_a )
__magic_name__ : List[str] = ldmad_pipe(**_a )
__magic_name__ , __magic_name__ : int = output.rgb, output.depth
__magic_name__ : Dict = rgb[0, -3:, -3:, -1].flatten()
__magic_name__ : str = rgb[0, -3:, -1].flatten()
assert rgb.shape == (1, 512, 512, 3)
assert depth.shape == (1, 512, 512)
__magic_name__ : Union[str, Any] = np.array(
[0.53_80_54_65, 0.56_70_73_05, 0.5_48_65_15, 0.57_01_22_36, 0.5_81_45_11, 0.56_25_34_87, 0.54_84_30_14, 0.55_09_22_63, 0.6_45_97_06] )
__magic_name__ : Any = np.array(
[0.9_26_37_81, 0.6_67_86_72, 0.5_48_65_15, 0.92_20_21_45, 0.67_83_11_35, 0.56_25_34_87, 0.9_24_16_94, 0.7_55_14_78, 0.6_45_97_06] )
assert np.abs(rgb_slice - expected_slice_rgb ).max() < 3e-3
assert np.abs(depth_slice - expected_slice_depth ).max() < 3e-3
@nightly
@require_torch_gpu
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def SCREAMING_SNAKE_CASE ( self , _a , _a="cpu" , _a=torch.floataa , _a=0 ):
__magic_name__ : Union[str, Any] = torch.Generator(device=_a ).manual_seed(_a )
__magic_name__ : Tuple = np.random.RandomState(_a ).standard_normal((1, 4, 64, 64) )
__magic_name__ : List[str] = torch.from_numpy(_a ).to(device=_a , dtype=_a )
__magic_name__ : Optional[Any] = {
"prompt": "a photograph of an astronaut riding a horse",
"latents": latents,
"generator": generator,
"num_inference_steps": 50,
"guidance_scale": 7.5,
"output_type": "numpy",
}
return inputs
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = StableDiffusionLDMaDPipeline.from_pretrained("Intel/ldm3d" ).to(_a )
ldmad_pipe.set_progress_bar_config(disable=_a )
__magic_name__ : Tuple = self.get_inputs(_a )
__magic_name__ : Dict = ldmad_pipe(**_a )
__magic_name__ , __magic_name__ : Tuple = output.rgb, output.depth
__magic_name__ : List[Any] = 0.49_55_86
__magic_name__ : Any = 0.33_79_55_15
__magic_name__ : str = 1_12.4_85_18
__magic_name__ : Tuple = 98.48_97_46
assert np.abs(expected_rgb_mean - rgb.mean() ) < 1e-3
assert np.abs(expected_rgb_std - rgb.std() ) < 1e-3
assert np.abs(expected_depth_mean - depth.mean() ) < 1e-3
assert np.abs(expected_depth_std - depth.std() ) < 1e-3
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = StableDiffusionLDMaDPipeline.from_pretrained("Intel/ldm3d-4c" ).to(_a )
ldmad_pipe.set_progress_bar_config(disable=_a )
__magic_name__ : str = self.get_inputs(_a )
__magic_name__ : int = ldmad_pipe(**_a )
__magic_name__ , __magic_name__ : Any = output.rgb, output.depth
__magic_name__ : Optional[int] = 0.4_19_41_27
__magic_name__ : Any = 0.35_37_55_86
__magic_name__ : Tuple = 0.5_63_85_02
__magic_name__ : Dict = 0.34_68_61_03
assert rgb.shape == (1, 512, 512, 3)
assert depth.shape == (1, 512, 512, 1)
assert np.abs(expected_rgb_mean - rgb.mean() ) < 1e-3
assert np.abs(expected_rgb_std - rgb.std() ) < 1e-3
assert np.abs(expected_depth_mean - depth.mean() ) < 1e-3
assert np.abs(expected_depth_std - depth.std() ) < 1e-3
| 281 |
import argparse
import collections
import json
import os
import re
import string
import sys
import numpy as np
snake_case : Dict = re.compile(R"\b(a|an|the)\b", re.UNICODE)
snake_case : Optional[int] = None
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser("Official evaluation script for SQuAD version 2.0." )
parser.add_argument("data_file" , metavar="data.json" , help="Input data JSON file." )
parser.add_argument("pred_file" , metavar="pred.json" , help="Model predictions." )
parser.add_argument(
"--out-file" , "-o" , metavar="eval.json" , help="Write accuracy metrics to file (default is stdout)." )
parser.add_argument(
"--na-prob-file" , "-n" , metavar="na_prob.json" , help="Model estimates of probability of no answer." )
parser.add_argument(
"--na-prob-thresh" , "-t" , type=_snake_case , default=1.0 , help="Predict \"\" if no-answer probability exceeds this (default = 1.0)." , )
parser.add_argument(
"--out-image-dir" , "-p" , metavar="out_images" , default=_snake_case , help="Save precision-recall curves to directory." )
parser.add_argument("--verbose" , "-v" , action="store_true" )
if len(sys.argv ) == 1:
parser.print_help()
sys.exit(1 )
return parser.parse_args()
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Optional[int] = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : str = bool(qa["answers"]["text"] )
return qid_to_has_ans
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Optional[Any]:
'''simple docstring'''
def remove_articles(_snake_case : List[str] ):
return ARTICLES_REGEX.sub(" " , _snake_case )
def white_space_fix(_snake_case : Optional[int] ):
return " ".join(text.split() )
def remove_punc(_snake_case : Optional[int] ):
__magic_name__ : Dict = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(_snake_case : str ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(_snake_case ) ) ) )
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
if not s:
return []
return normalize_answer(_snake_case ).split()
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Dict ) -> Tuple:
'''simple docstring'''
return int(normalize_answer(_snake_case ) == normalize_answer(_snake_case ) )
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : int ) -> str:
'''simple docstring'''
__magic_name__ : Any = get_tokens(_snake_case )
__magic_name__ : Optional[int] = get_tokens(_snake_case )
__magic_name__ : Tuple = collections.Counter(_snake_case ) & collections.Counter(_snake_case )
__magic_name__ : Tuple = sum(common.values() )
if len(_snake_case ) == 0 or len(_snake_case ) == 0:
# If either is no-answer, then F1 is 1 if they agree, 0 otherwise
return int(gold_toks == pred_toks )
if num_same == 0:
return 0
__magic_name__ : Dict = 1.0 * num_same / len(_snake_case )
__magic_name__ : Optional[Any] = 1.0 * num_same / len(_snake_case )
__magic_name__ : List[Any] = (2 * precision * recall) / (precision + recall)
return fa
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = {}
__magic_name__ : int = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : Union[str, Any] = qa["id"]
__magic_name__ : Any = [t for t in qa["answers"]["text"] if normalize_answer(_snake_case )]
if not gold_answers:
# For unanswerable questions, only correct answer is empty string
__magic_name__ : Tuple = [""]
if qid not in preds:
print(F'''Missing prediction for {qid}''' )
continue
__magic_name__ : Any = preds[qid]
# Take max over all gold answers
__magic_name__ : List[Any] = max(compute_exact(_snake_case , _snake_case ) for a in gold_answers )
__magic_name__ : int = max(compute_fa(_snake_case , _snake_case ) for a in gold_answers )
return exact_scores, fa_scores
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] , _snake_case : Optional[int] , _snake_case : Dict ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : str = {}
for qid, s in scores.items():
__magic_name__ : Dict = na_probs[qid] > na_prob_thresh
if pred_na:
__magic_name__ : str = float(not qid_to_has_ans[qid] )
else:
__magic_name__ : Optional[int] = s
return new_scores
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : List[str] , _snake_case : Tuple=None ) -> Tuple:
'''simple docstring'''
if not qid_list:
__magic_name__ : Any = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores.values() ) / total),
("f1", 100.0 * sum(fa_scores.values() ) / total),
("total", total),
] )
else:
__magic_name__ : Tuple = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores[k] for k in qid_list ) / total),
("f1", 100.0 * sum(fa_scores[k] for k in qid_list ) / total),
("total", total),
] )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : str , _snake_case : str ) -> Dict:
'''simple docstring'''
for k in new_eval:
__magic_name__ : int = new_eval[k]
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : Optional[Any] , _snake_case : Union[str, Any] ) -> str:
'''simple docstring'''
plt.step(_snake_case , _snake_case , color="b" , alpha=0.2 , where="post" )
plt.fill_between(_snake_case , _snake_case , step="post" , alpha=0.2 , color="b" )
plt.xlabel("Recall" )
plt.ylabel("Precision" )
plt.xlim([0.0, 1.05] )
plt.ylim([0.0, 1.05] )
plt.title(_snake_case )
plt.savefig(_snake_case )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any , _snake_case : Optional[int] , _snake_case : List[Any] , _snake_case : Optional[int]=None , _snake_case : int=None ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
__magic_name__ : Optional[int] = 0.0
__magic_name__ : str = 1.0
__magic_name__ : str = 0.0
__magic_name__ : List[str] = [1.0]
__magic_name__ : str = [0.0]
__magic_name__ : Optional[Any] = 0.0
for i, qid in enumerate(_snake_case ):
if qid_to_has_ans[qid]:
true_pos += scores[qid]
__magic_name__ : List[str] = true_pos / float(i + 1 )
__magic_name__ : Any = true_pos / float(_snake_case )
if i == len(_snake_case ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]:
# i.e., if we can put a threshold after this point
avg_prec += cur_p * (cur_r - recalls[-1])
precisions.append(_snake_case )
recalls.append(_snake_case )
if out_image:
plot_pr_curve(_snake_case , _snake_case , _snake_case , _snake_case )
return {"ap": 100.0 * avg_prec}
def lowerCAmelCase_ ( _snake_case : Tuple , _snake_case : Optional[Any] , _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : Any , _snake_case : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
if out_image_dir and not os.path.exists(_snake_case ):
os.makedirs(_snake_case )
__magic_name__ : Any = sum(1 for v in qid_to_has_ans.values() if v )
if num_true_pos == 0:
return
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_exact.png" ) , title="Precision-Recall curve for Exact Match score" , )
__magic_name__ : Union[str, Any] = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_f1.png" ) , title="Precision-Recall curve for F1 score" , )
__magic_name__ : str = {k: float(_snake_case ) for k, v in qid_to_has_ans.items()}
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_oracle.png" ) , title="Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)" , )
merge_eval(_snake_case , _snake_case , "pr_exact" )
merge_eval(_snake_case , _snake_case , "pr_f1" )
merge_eval(_snake_case , _snake_case , "pr_oracle" )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : Optional[Any] , _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
if not qid_list:
return
__magic_name__ : Dict = [na_probs[k] for k in qid_list]
__magic_name__ : str = np.ones_like(_snake_case ) / float(len(_snake_case ) )
plt.hist(_snake_case , weights=_snake_case , bins=20 , range=(0.0, 1.0) )
plt.xlabel("Model probability of no-answer" )
plt.ylabel("Proportion of dataset" )
plt.title(F'''Histogram of no-answer probability: {name}''' )
plt.savefig(os.path.join(_snake_case , F'''na_prob_hist_{name}.png''' ) )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : List[str] , _snake_case : Dict ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] )
__magic_name__ : List[str] = num_no_ans
__magic_name__ : Dict = cur_score
__magic_name__ : Dict = 0.0
__magic_name__ : Any = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
for i, qid in enumerate(_snake_case ):
if qid not in scores:
continue
if qid_to_has_ans[qid]:
__magic_name__ : Union[str, Any] = scores[qid]
else:
if preds[qid]:
__magic_name__ : List[Any] = -1
else:
__magic_name__ : Optional[int] = 0
cur_score += diff
if cur_score > best_score:
__magic_name__ : Optional[int] = cur_score
__magic_name__ : List[Any] = na_probs[qid]
return 100.0 * best_score / len(_snake_case ), best_thresh
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str , _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Dict ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ , __magic_name__ : List[str] = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ , __magic_name__ : int = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ : Optional[int] = best_exact
__magic_name__ : List[Any] = exact_thresh
__magic_name__ : Dict = best_fa
__magic_name__ : Any = fa_thresh
def lowerCAmelCase_ ( ) -> int:
'''simple docstring'''
with open(OPTS.data_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
__magic_name__ : List[Any] = dataset_json["data"]
with open(OPTS.pred_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
if OPTS.na_prob_file:
with open(OPTS.na_prob_file ) as f:
__magic_name__ : Any = json.load(_snake_case )
else:
__magic_name__ : Any = {k: 0.0 for k in preds}
__magic_name__ : str = make_qid_to_has_ans(_snake_case ) # maps qid to True/False
__magic_name__ : Tuple = [k for k, v in qid_to_has_ans.items() if v]
__magic_name__ : Optional[Any] = [k for k, v in qid_to_has_ans.items() if not v]
__magic_name__ , __magic_name__ : Union[str, Any] = get_raw_scores(_snake_case , _snake_case )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case )
if has_ans_qids:
__magic_name__ : int = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "HasAns" )
if no_ans_qids:
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "NoAns" )
if OPTS.na_prob_file:
find_all_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
if OPTS.na_prob_file and OPTS.out_image_dir:
run_precision_recall_analysis(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , OPTS.out_image_dir )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "hasAns" )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "noAns" )
if OPTS.out_file:
with open(OPTS.out_file , "w" ) as f:
json.dump(_snake_case , _snake_case )
else:
print(json.dumps(_snake_case , indent=2 ) )
if __name__ == "__main__":
snake_case : int = parse_args()
if OPTS.out_image_dir:
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
main()
| 281 | 1 |
from typing import List, Optional, Union
from ...image_utils import ImageInput
from ...processing_utils import ProcessorMixin
from ...tokenization_utils_base import BatchEncoding, PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
from ...utils import TensorType
class _snake_case ( snake_case ):
UpperCamelCase__ = ['image_processor', 'tokenizer']
UpperCamelCase__ = 'BlipImageProcessor'
UpperCamelCase__ = 'AutoTokenizer'
def __init__( self , _a , _a ):
__magic_name__ : Optional[int] = False
super().__init__(_a , _a )
__magic_name__ : Optional[Any] = self.image_processor
def __call__( self , _a = None , _a = None , _a = True , _a = False , _a = None , _a = None , _a = 0 , _a = None , _a = None , _a = False , _a = False , _a = False , _a = False , _a = False , _a = True , _a = None , **_a , ):
if images is None and text is None:
raise ValueError("You have to specify either images or text." )
# Get only text
if images is None:
__magic_name__ : List[Any] = self.tokenizer
__magic_name__ : Dict = self.tokenizer(
text=_a , add_special_tokens=_a , padding=_a , truncation=_a , max_length=_a , stride=_a , pad_to_multiple_of=_a , return_attention_mask=_a , return_overflowing_tokens=_a , return_special_tokens_mask=_a , return_offsets_mapping=_a , return_token_type_ids=_a , return_length=_a , verbose=_a , return_tensors=_a , **_a , )
return text_encoding
# add pixel_values
__magic_name__ : int = self.image_processor(_a , return_tensors=_a )
if text is not None:
__magic_name__ : Dict = self.tokenizer(
text=_a , add_special_tokens=_a , padding=_a , truncation=_a , max_length=_a , stride=_a , pad_to_multiple_of=_a , return_attention_mask=_a , return_overflowing_tokens=_a , return_special_tokens_mask=_a , return_offsets_mapping=_a , return_token_type_ids=_a , return_length=_a , verbose=_a , return_tensors=_a , **_a , )
else:
__magic_name__ : Any = None
if text_encoding is not None:
encoding_image_processor.update(_a )
return encoding_image_processor
def SCREAMING_SNAKE_CASE ( self , *_a , **_a ):
return self.tokenizer.batch_decode(*_a , **_a )
def SCREAMING_SNAKE_CASE ( self , *_a , **_a ):
return self.tokenizer.decode(*_a , **_a )
@property
# Copied from transformers.models.blip.processing_blip.BlipProcessor.model_input_names
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = self.tokenizer.model_input_names
__magic_name__ : Dict = self.image_processor.model_input_names
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names ) )
| 281 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
snake_case : str = "▁"
snake_case : List[Any] = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = BigBirdTokenizer
UpperCamelCase__ = BigBirdTokenizerFast
UpperCamelCase__ = True
UpperCamelCase__ = True
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
__magic_name__ : Optional[Any] = self.tokenizer_class(_a , keep_accents=_a )
tokenizer.save_pretrained(self.tmpdirname )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = "<s>"
__magic_name__ : Dict = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_a ) , _a )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_a ) , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : 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] , "[MASK]" )
self.assertEqual(len(_a ) , 1_004 )
def SCREAMING_SNAKE_CASE ( self ):
self.assertEqual(self.get_tokenizer().vocab_size , 1_000 )
def SCREAMING_SNAKE_CASE ( self ):
if not self.test_rust_tokenizer:
return
__magic_name__ : Dict = self.get_tokenizer()
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Any = "I was born in 92000, and this is falsé."
__magic_name__ : Dict = tokenizer.tokenize(_a )
__magic_name__ : Any = rust_tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : List[Any] = tokenizer.encode(_a , add_special_tokens=_a )
__magic_name__ : List[str] = rust_tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Dict = tokenizer.encode(_a )
__magic_name__ : Optional[int] = rust_tokenizer.encode(_a )
self.assertListEqual(_a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = BigBirdTokenizer(_a , keep_accents=_a )
__magic_name__ : str = tokenizer.tokenize("This is a test" )
self.assertListEqual(_a , ["▁This", "▁is", "▁a", "▁t", "est"] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_a ) , [285, 46, 10, 170, 382] , )
__magic_name__ : Dict = tokenizer.tokenize("I was born in 92000, and this is falsé." )
self.assertListEqual(
_a , [
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",
".",
] , )
__magic_name__ : Union[str, Any] = tokenizer.convert_tokens_to_ids(_a )
self.assertListEqual(
_a , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
__magic_name__ : int = tokenizer.convert_ids_to_tokens(_a )
self.assertListEqual(
_a , [
SPIECE_UNDERLINE + "I",
SPIECE_UNDERLINE + "was",
SPIECE_UNDERLINE + "b",
"or",
"n",
SPIECE_UNDERLINE + "in",
SPIECE_UNDERLINE + "",
"<unk>",
"2",
"0",
"0",
"0",
",",
SPIECE_UNDERLINE + "and",
SPIECE_UNDERLINE + "this",
SPIECE_UNDERLINE + "is",
SPIECE_UNDERLINE + "f",
"al",
"s",
"<unk>",
".",
] , )
@cached_property
def SCREAMING_SNAKE_CASE ( self ):
return BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "Hello World!"
__magic_name__ : Dict = [65, 18_536, 2_260, 101, 66]
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = (
"This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"
" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"
)
# fmt: off
__magic_name__ : List[str] = [65, 871, 419, 358, 946, 991, 2_521, 452, 358, 1_357, 387, 7_751, 3_536, 112, 985, 456, 126, 865, 938, 5_400, 5_734, 458, 1_368, 467, 786, 2_462, 5_246, 1_159, 633, 865, 4_519, 457, 582, 852, 2_557, 427, 916, 508, 405, 34_324, 497, 391, 408, 11_342, 1_244, 385, 100, 938, 985, 456, 574, 362, 12_597, 3_200, 3_129, 1_172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
__magic_name__ : Optional[Any] = list(self.big_tokenizer.get_vocab().keys() )[:10]
__magic_name__ : List[Any] = " ".join(_a )
__magic_name__ : Any = self.big_tokenizer.encode_plus(_a , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : Union[str, Any] = self.big_tokenizer.batch_encode_plus(
[sequence + " " + sequence] , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : List[str] = BigBirdConfig(attention_type="original_full" )
__magic_name__ : Optional[int] = BigBirdModel(_a )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_a )
model(**_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
__magic_name__ : int = tokenizer.decode(tokenizer("Paris is the [MASK]." ).input_ids )
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# fmt: off
__magic_name__ : Optional[Any] = {"input_ids": [[65, 39_286, 458, 36_335, 2_001, 456, 13_073, 13_266, 455, 113, 7_746, 1_741, 11_157, 391, 13_073, 13_266, 455, 113, 3_967, 35_412, 113, 4_936, 109, 3_870, 2_377, 113, 30_084, 45_720, 458, 134, 17_496, 112, 503, 11_672, 113, 118, 112, 5_665, 13_347, 38_687, 112, 1_496, 31_389, 112, 3_268, 47_264, 134, 962, 112, 16_377, 8_035, 23_130, 430, 12_169, 15_518, 28_592, 458, 146, 41_697, 109, 391, 12_169, 15_518, 16_689, 458, 146, 41_358, 109, 452, 726, 4_034, 111, 763, 35_412, 5_082, 388, 1_903, 111, 9_051, 391, 2_870, 48_918, 1_900, 1_123, 550, 998, 112, 9_586, 15_985, 455, 391, 410, 22_955, 37_636, 114, 66], [65, 448, 17_496, 419, 3_663, 385, 763, 113, 27_533, 2_870, 3_283, 13_043, 1_639, 24_713, 523, 656, 24_013, 18_550, 2_521, 517, 27_014, 21_244, 420, 1_212, 1_465, 391, 927, 4_833, 388, 578, 11_786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2_169, 7_687, 21_932, 18_146, 726, 363, 17_032, 3_391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], "attention_mask": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_a , model_name="google/bigbird-roberta-base" , revision="215c99f1600e06f83acce68422f2035b2b5c3510" , )
| 281 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
snake_case : Optional[Any] = logging.get_logger(__name__)
snake_case : Union[str, Any] = {
"google/bit-50": "https://huggingface.co/google/bit-50/resolve/main/config.json",
}
class _snake_case ( snake_case , snake_case ):
UpperCamelCase__ = 'bit'
UpperCamelCase__ = ['preactivation', 'bottleneck']
UpperCamelCase__ = ['SAME', 'VALID']
def __init__( self , _a=3 , _a=64 , _a=[256, 512, 1_024, 2_048] , _a=[3, 4, 6, 3] , _a="preactivation" , _a="relu" , _a=None , _a=32 , _a=0.0 , _a=False , _a=32 , _a=1 , _a=None , _a=None , **_a , ):
super().__init__(**_a )
if layer_type not in self.layer_types:
raise ValueError(f'''layer_type={layer_type} is not one of {",".join(self.layer_types )}''' )
if global_padding is not None:
if global_padding.upper() in self.supported_padding:
__magic_name__ : List[Any] = global_padding.upper()
else:
raise ValueError(f'''Padding strategy {global_padding} not supported''' )
__magic_name__ : List[str] = num_channels
__magic_name__ : List[Any] = embedding_size
__magic_name__ : List[str] = hidden_sizes
__magic_name__ : Optional[Any] = depths
__magic_name__ : List[str] = layer_type
__magic_name__ : str = hidden_act
__magic_name__ : int = global_padding
__magic_name__ : Optional[int] = num_groups
__magic_name__ : Optional[int] = drop_path_rate
__magic_name__ : Tuple = embedding_dynamic_padding
__magic_name__ : List[Any] = output_stride
__magic_name__ : int = width_factor
__magic_name__ : Dict = ["stem"] + [f'''stage{idx}''' for idx in range(1 , len(_a ) + 1 )]
__magic_name__ , __magic_name__ : int = get_aligned_output_features_output_indices(
out_features=_a , out_indices=_a , stage_names=self.stage_names )
| 281 |
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
snake_case : int = logging.get_logger(__name__)
snake_case : List[str] = {"vocab_file": "spiece.model"}
snake_case : List[str] = {
"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",
}
}
snake_case : Tuple = {
"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,
}
snake_case : List[str] = "▁"
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a=True , _a=True , _a=False , _a="[CLS]" , _a="[SEP]" , _a="<unk>" , _a="[SEP]" , _a="<pad>" , _a="[CLS]" , _a="[MASK]" , _a = None , **_a , ):
# 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.
__magic_name__ : str = (
AddedToken(_a , lstrip=_a , rstrip=_a , normalized=_a )
if isinstance(_a , _a )
else mask_token
)
__magic_name__ : Union[str, Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_a , remove_space=_a , keep_accents=_a , bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , pad_token=_a , cls_token=_a , mask_token=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Dict = do_lower_case
__magic_name__ : Tuple = remove_space
__magic_name__ : Union[str, Any] = keep_accents
__magic_name__ : Tuple = vocab_file
__magic_name__ : int = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_a )
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.sp_model )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ):
__magic_name__ : List[str] = self.__dict__.copy()
__magic_name__ : Any = None
return state
def __setstate__( self , _a ):
__magic_name__ : Union[str, Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : str = {}
__magic_name__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.remove_space:
__magic_name__ : List[Any] = " ".join(inputs.strip().split() )
else:
__magic_name__ : str = inputs
__magic_name__ : int = outputs.replace("``" , "\"" ).replace("''" , "\"" )
if not self.keep_accents:
__magic_name__ : str = unicodedata.normalize("NFKD" , _a )
__magic_name__ : Tuple = "".join([c for c in outputs if not unicodedata.combining(_a )] )
if self.do_lower_case:
__magic_name__ : int = outputs.lower()
return outputs
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = self.preprocess_text(_a )
__magic_name__ : Dict = self.sp_model.encode(_a , out_type=_a )
__magic_name__ : Any = []
for piece in pieces:
if len(_a ) > 1 and piece[-1] == str("," ) and piece[-2].isdigit():
__magic_name__ : Optional[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(_a , "" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
__magic_name__ : List[str] = cur_pieces[1:]
else:
__magic_name__ : Optional[int] = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(_a )
else:
new_pieces.append(_a )
return new_pieces
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.PieceToId(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.IdToPiece(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Any = []
__magic_name__ : Union[str, Any] = ""
__magic_name__ : int = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_a ) + token
__magic_name__ : List[Any] = True
__magic_name__ : Optional[int] = []
else:
current_sub_tokens.append(_a )
__magic_name__ : Optional[Any] = False
out_string += self.sp_model.decode(_a )
return out_string.strip()
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : List[str] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is not None:
return [1] + ([0] * len(_a )) + [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[int] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : List[str] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _a )
elif not os.path.isfile(self.vocab_file ):
with open(_a , "wb" ) as fi:
__magic_name__ : Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(_a )
return (out_vocab_file,)
| 281 | 1 |
from unittest import TestCase
from datasets import Dataset
from minhash_deduplication import deduplicate_dataset, make_duplicate_clusters
def lowerCAmelCase_ ( ) -> str:
'''simple docstring'''
__magic_name__ : List[str] = {
"repo_name": ["test_repo1", "test_repo2", "test_repo3"],
"path": ["test_1.py", "test_2.py", "unit_test.py"],
"content": ["a " * 20, "a " * 30, "b " * 7],
}
__magic_name__ : List[str] = Dataset.from_dict(_snake_case )
return dataset
class _snake_case ( snake_case ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = get_dataset()
__magic_name__ : Optional[Any] = make_duplicate_clusters(_a , 0.85 )
self.assertEqual(len(duplicate_clusters[0] ) , 2 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = get_dataset()
__magic_name__ , __magic_name__ : List[Any] = deduplicate_dataset(_a )
self.assertEqual(len(_a ) , 2 )
print(_a )
self.assertEqual(duplicate_clusters[0][0]["copies"] , 2 )
self.assertEqual(duplicate_clusters[0][0]["is_extreme"] , _a )
| 281 |
import unicodedata
from dataclasses import dataclass
from typing import Optional, Union
import numpy as np
from transformers.data.data_collator import DataCollatorMixin
from transformers.file_utils import PaddingStrategy
from transformers.tokenization_utils_base import PreTrainedTokenizerBase
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Union[str, Any] = np.full((len(_snake_case ), sequence_length, 2) , _snake_case )
else:
__magic_name__ : List[Any] = np.full((len(_snake_case ), sequence_length) , _snake_case )
for i, tensor in enumerate(_snake_case ):
if padding_side == "right":
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Optional[Any] = tensor[:sequence_length]
else:
__magic_name__ : Union[str, Any] = tensor[:sequence_length]
else:
if isinstance(_snake_case , _snake_case ):
__magic_name__ : List[Any] = tensor[:sequence_length]
else:
__magic_name__ : Optional[Any] = tensor[:sequence_length]
return out_tensor.tolist()
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = ord(_snake_case )
if (cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126):
return True
__magic_name__ : Any = unicodedata.category(_snake_case )
if cat.startswith("P" ):
return True
return False
@dataclass
class _snake_case ( snake_case ):
UpperCamelCase__ = 42
UpperCamelCase__ = True
UpperCamelCase__ = None
UpperCamelCase__ = None
UpperCamelCase__ = -100
UpperCamelCase__ = "pt"
def SCREAMING_SNAKE_CASE ( self , _a ):
import torch
__magic_name__ : List[str] = "label" if "label" in features[0].keys() else "labels"
__magic_name__ : Union[str, Any] = [feature[label_name] for feature in features] if label_name in features[0].keys() else None
__magic_name__ : Optional[int] = self.tokenizer.pad(
_a , padding=self.padding , max_length=self.max_length , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors="pt" if labels is None else None , )
if labels is None:
return batch
__magic_name__ : Dict = torch.tensor(batch["entity_ids"] ).shape[1]
__magic_name__ : List[Any] = self.tokenizer.padding_side
if padding_side == "right":
__magic_name__ : str = [
list(_a ) + [self.label_pad_token_id] * (sequence_length - len(_a )) for label in labels
]
else:
__magic_name__ : int = [
[self.label_pad_token_id] * (sequence_length - len(_a )) + list(_a ) for label in labels
]
__magic_name__ : Dict = [feature["ner_tags"] for feature in features]
__magic_name__ : List[Any] = padding_tensor(_a , -1 , _a , _a )
__magic_name__ : Any = [feature["original_entity_spans"] for feature in features]
__magic_name__ : Any = padding_tensor(_a , (-1, -1) , _a , _a )
__magic_name__ : List[Any] = {k: torch.tensor(_a , dtype=torch.intaa ) for k, v in batch.items()}
return batch
| 281 | 1 |
import numpy as np
def lowerCAmelCase_ ( _snake_case : np.ndarray , _snake_case : float ) -> np.ndarray:
'''simple docstring'''
return np.where(vector > 0 , _snake_case , (alpha * (np.exp(_snake_case ) - 1)) )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
import math
def lowerCAmelCase_ ( _snake_case : float , _snake_case : float ) -> float:
'''simple docstring'''
return math.pow(_snake_case , 2 ) - a
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
return 2 * x
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
__magic_name__ : Optional[int] = 2.0
while start <= a:
__magic_name__ : str = math.pow(_snake_case , 2 )
return start
def lowerCAmelCase_ ( _snake_case : float , _snake_case : int = 9999 , _snake_case : float = 0.00_000_000_000_001 ) -> float:
'''simple docstring'''
if a < 0:
raise ValueError("math domain error" )
__magic_name__ : Optional[int] = get_initial_point(_snake_case )
for _ in range(_snake_case ):
__magic_name__ : int = value
__magic_name__ : str = value - fx(_snake_case , _snake_case ) / fx_derivative(_snake_case )
if abs(prev_value - value ) < tolerance:
return value
return value
if __name__ == "__main__":
from doctest import testmod
testmod()
| 281 | 1 |
# limitations under the License.
# NOTE: This file is deprecated and will be removed in a future version.
# It only exists so that temporarely `from diffusers.pipelines import DiffusionPipeline` works
from .pipelines import DiffusionPipeline, ImagePipelineOutput # noqa: F401
from .utils import deprecate
deprecate(
"pipelines_utils",
"0.22.0",
"Importing `DiffusionPipeline` or `ImagePipelineOutput` from diffusers.pipeline_utils is deprecated. Please import from diffusers.pipelines.pipeline_utils instead.",
standard_warn=False,
stacklevel=3,
)
| 281 |
from __future__ import annotations
import unittest
from transformers import LEDConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFLEDForConditionalGeneration, TFLEDModel
@require_tf
class _snake_case :
UpperCamelCase__ = LEDConfig
UpperCamelCase__ = {}
UpperCamelCase__ = 'gelu'
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=False , _a=99 , _a=32 , _a=2 , _a=4 , _a=37 , _a=0.1 , _a=0.1 , _a=20 , _a=2 , _a=1 , _a=0 , _a=4 , ):
__magic_name__ : int = parent
__magic_name__ : Optional[int] = batch_size
__magic_name__ : Tuple = seq_length
__magic_name__ : List[Any] = is_training
__magic_name__ : Dict = use_labels
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : int = hidden_size
__magic_name__ : Optional[int] = num_hidden_layers
__magic_name__ : Optional[int] = num_attention_heads
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Any = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Any = eos_token_id
__magic_name__ : str = pad_token_id
__magic_name__ : int = bos_token_id
__magic_name__ : Optional[int] = attention_window
# `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size
# [num_attention_heads, encoder_seq_length, encoder_key_length], but TFLongformerSelfAttention
# returns attention of shape [num_attention_heads, encoder_seq_length, self.attention_window + 1]
# because its local attention only attends to `self.attention_window` and one before and one after
__magic_name__ : Tuple = self.attention_window + 2
# because of padding `encoder_seq_length`, is different from `seq_length`. Relevant for
# the `test_attention_outputs` and `test_hidden_states_output` tests
__magic_name__ : Tuple = (
self.seq_length + (self.attention_window - self.seq_length % self.attention_window) % self.attention_window
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
__magic_name__ : int = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
__magic_name__ : int = tf.concat([input_ids, eos_tensor] , axis=1 )
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : Dict = self.config_cls(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_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_ids=[2] , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.pad_token_id , attention_window=self.attention_window , **self.config_updates , )
__magic_name__ : List[str] = prepare_led_inputs_dict(_a , _a , _a )
__magic_name__ : Union[str, Any] = tf.concat(
[tf.zeros_like(_a )[:, :-1], tf.ones_like(_a )[:, -1:]] , axis=-1 , )
__magic_name__ : List[Any] = global_attention_mask
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Dict = TFLEDModel(config=_a ).get_decoder()
__magic_name__ : Optional[int] = inputs_dict["input_ids"]
__magic_name__ : Union[str, Any] = input_ids[:1, :]
__magic_name__ : str = inputs_dict["attention_mask"][:1, :]
__magic_name__ : int = 1
# first forward pass
__magic_name__ : Tuple = model(_a , attention_mask=_a , use_cache=_a )
__magic_name__ , __magic_name__ : str = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
__magic_name__ : Optional[int] = ids_tensor((self.batch_size, 3) , config.vocab_size )
__magic_name__ : Any = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
__magic_name__ : Optional[Any] = tf.concat([input_ids, next_tokens] , axis=-1 )
__magic_name__ : List[Any] = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
__magic_name__ : List[str] = model(_a , attention_mask=_a )[0]
__magic_name__ : Dict = model(_a , attention_mask=_a , past_key_values=_a )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
__magic_name__ : List[Any] = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
__magic_name__ : Union[str, Any] = output_from_no_past[:, -3:, random_slice_idx]
__magic_name__ : List[str] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(_a , _a , rtol=1e-3 )
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : List[Any] , _snake_case : Any , _snake_case : str=None , _snake_case : List[str]=None , _snake_case : int=None , _snake_case : Any=None , ) -> int:
'''simple docstring'''
if attention_mask is None:
__magic_name__ : str = tf.cast(tf.math.not_equal(_snake_case , config.pad_token_id ) , tf.inta )
if decoder_attention_mask is None:
__magic_name__ : List[Any] = tf.concat(
[
tf.ones(decoder_input_ids[:, :1].shape , dtype=tf.inta ),
tf.cast(tf.math.not_equal(decoder_input_ids[:, 1:] , config.pad_token_id ) , tf.inta ),
] , axis=-1 , )
if head_mask is None:
__magic_name__ : Dict = tf.ones((config.encoder_layers, config.encoder_attention_heads) )
if decoder_head_mask is None:
__magic_name__ : Any = tf.ones((config.decoder_layers, config.decoder_attention_heads) )
return {
"input_ids": input_ids,
"attention_mask": attention_mask,
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
}
@require_tf
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (TFLEDForConditionalGeneration, TFLEDModel) if is_tf_available() else ()
UpperCamelCase__ = (TFLEDForConditionalGeneration,) if is_tf_available() else ()
UpperCamelCase__ = (
{
'conversational': TFLEDForConditionalGeneration,
'feature-extraction': TFLEDModel,
'summarization': TFLEDForConditionalGeneration,
'text2text-generation': TFLEDForConditionalGeneration,
'translation': TFLEDForConditionalGeneration,
}
if is_tf_available()
else {}
)
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = TFLEDModelTester(self )
__magic_name__ : List[Any] = ConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : int = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[str] = tf.zeros_like(inputs_dict["attention_mask"] )
__magic_name__ : Optional[Any] = 2
__magic_name__ : Tuple = tf.where(
tf.range(self.model_tester.seq_length )[None, :] < num_global_attn_indices , 1 , inputs_dict["global_attention_mask"] , )
__magic_name__ : Any = True
__magic_name__ : str = self.model_tester.seq_length
__magic_name__ : Dict = self.model_tester.encoder_seq_length
def check_decoder_attentions_output(_a ):
__magic_name__ : str = outputs.decoder_attentions
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(decoder_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
def check_encoder_attentions_output(_a ):
__magic_name__ : Any = [t.numpy() for t in outputs.encoder_attentions]
__magic_name__ : Tuple = [t.numpy() for t in outputs.encoder_global_attentions]
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
self.assertListEqual(
list(global_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, encoder_seq_length, num_global_attn_indices] , )
for model_class in self.all_model_classes:
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = False
__magic_name__ : Tuple = False
__magic_name__ : Optional[int] = model_class(_a )
__magic_name__ : str = model(self._prepare_for_class(_a , _a ) )
__magic_name__ : Any = len(_a )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
if self.is_encoder_decoder:
__magic_name__ : Tuple = model_class(_a )
__magic_name__ : Optional[Any] = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_decoder_attentions_output(_a )
# Check that output attentions can also be changed via the config
del inputs_dict["output_attentions"]
__magic_name__ : Dict = True
__magic_name__ : str = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
# Check attention is always last and order is fine
__magic_name__ : Union[str, Any] = True
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(out_len + (2 if self.is_encoder_decoder else 1) , len(_a ) )
self.assertEqual(model.config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
@unittest.skip("LED keeps using potentially symbolic tensors in conditionals and breaks tracing." )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
# TODO: Head-masking not yet implement
pass
def lowerCAmelCase_ ( _snake_case : int ) -> Optional[int]:
'''simple docstring'''
return tf.constant(_snake_case , dtype=tf.intaa )
snake_case : Optional[int] = 1E-4
@slow
@require_tf
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" ).led
# change to intended input here
__magic_name__ : Optional[int] = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : str = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Any = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : List[Any] = model(**_a )[0]
__magic_name__ : List[str] = (1, 1_024, 768)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : int = tf.convert_to_tensor(
[[2.30_50, 2.82_79, 0.65_31], [-1.84_57, -0.14_55, -3.56_61], [-1.01_86, 0.45_86, -2.20_43]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" )
# change to intended input here
__magic_name__ : int = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Tuple = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Optional[Any] = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : Union[str, Any] = model(**_a )[0]
__magic_name__ : Optional[int] = (1, 1_024, model.config.vocab_size)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : str = tf.convert_to_tensor(
[[33.65_07, 6.45_72, 16.80_89], [5.87_39, -2.42_38, 11.29_02], [-3.21_39, -4.31_49, 4.27_83]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 , rtol=1e-3 )
| 281 | 1 |
import argparse
import logging
from collections import namedtuple
import torch
from model_bertabs import BertAbsSummarizer
from models.model_builder import AbsSummarizer # The authors' implementation
from transformers import BertTokenizer
logging.basicConfig(level=logging.INFO)
snake_case : str = logging.getLogger(__name__)
snake_case : str = "Hello world! cécé herlolip"
snake_case : Optional[Any] = namedtuple(
"BertAbsConfig",
[
"temp_dir",
"large",
"use_bert_emb",
"finetune_bert",
"encoder",
"share_emb",
"max_pos",
"enc_layers",
"enc_hidden_size",
"enc_heads",
"enc_ff_size",
"enc_dropout",
"dec_layers",
"dec_hidden_size",
"dec_heads",
"dec_ff_size",
"dec_dropout",
],
)
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[str] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = BertAbsConfig(
temp_dir="." , finetune_bert=_snake_case , large=_snake_case , share_emb=_snake_case , use_bert_emb=_snake_case , encoder="bert" , max_pos=512 , enc_layers=6 , enc_hidden_size=512 , enc_heads=8 , enc_ff_size=512 , enc_dropout=0.2 , dec_layers=6 , dec_hidden_size=768 , dec_heads=8 , dec_ff_size=2048 , dec_dropout=0.2 , )
__magic_name__ : Tuple = torch.load(_snake_case , lambda _snake_case , _snake_case : storage )
__magic_name__ : Any = AbsSummarizer(_snake_case , torch.device("cpu" ) , _snake_case )
original.eval()
__magic_name__ : Optional[int] = BertAbsSummarizer(_snake_case , torch.device("cpu" ) )
new_model.eval()
# -------------------
# Convert the weights
# -------------------
logging.info("convert the model" )
new_model.bert.load_state_dict(original.bert.state_dict() )
new_model.decoder.load_state_dict(original.decoder.state_dict() )
new_model.generator.load_state_dict(original.generator.state_dict() )
# ----------------------------------
# Make sure the outpus are identical
# ----------------------------------
logging.info("Make sure that the models' outputs are identical" )
__magic_name__ : Any = BertTokenizer.from_pretrained("bert-base-uncased" )
# prepare the model inputs
__magic_name__ : Tuple = tokenizer.encode("This is sample éàalj'-." )
encoder_input_ids.extend([tokenizer.pad_token_id] * (512 - len(_snake_case )) )
__magic_name__ : int = torch.tensor(_snake_case ).unsqueeze(0 )
__magic_name__ : int = tokenizer.encode("This is sample 3 éàalj'-." )
decoder_input_ids.extend([tokenizer.pad_token_id] * (512 - len(_snake_case )) )
__magic_name__ : Dict = torch.tensor(_snake_case ).unsqueeze(0 )
# failsafe to make sure the weights reset does not affect the
# loaded weights.
assert torch.max(torch.abs(original.generator[0].weight - new_model.generator[0].weight ) ) == 0
# forward pass
__magic_name__ : Dict = encoder_input_ids
__magic_name__ : List[Any] = decoder_input_ids
__magic_name__ : List[str] = None
__magic_name__ : Any = None
__magic_name__ : List[Any] = None
__magic_name__ : Optional[int] = None
__magic_name__ : str = None
# The original model does not apply the geneator layer immediatly but rather in
# the beam search (where it combines softmax + linear layer). Since we already
# apply the softmax in our generation process we only apply the linear layer here.
# We make sure that the outputs of the full stack are identical
__magic_name__ : List[Any] = original(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )[0]
__magic_name__ : Optional[int] = original.generator(_snake_case )
__magic_name__ : str = new_model(
_snake_case , _snake_case , _snake_case , _snake_case , _snake_case )[0]
__magic_name__ : Any = new_model.generator(_snake_case )
__magic_name__ : Tuple = torch.max(torch.abs(output_converted_model - output_original_model ) ).item()
print("Maximum absolute difference beween weights: {:.2f}".format(_snake_case ) )
__magic_name__ : Optional[Any] = torch.max(torch.abs(output_converted_generator - output_original_generator ) ).item()
print("Maximum absolute difference beween weights: {:.2f}".format(_snake_case ) )
__magic_name__ : Union[str, Any] = torch.allclose(_snake_case , _snake_case , atol=1E-3 )
if are_identical:
logging.info("all weights are equal up to 1e-3" )
else:
raise ValueError("the weights are different. The new model is likely different from the original one." )
# The model has been saved with torch.save(model) and this is bound to the exact
# directory structure. We save the state_dict instead.
logging.info("saving the model's state dictionary" )
torch.save(
new_model.state_dict() , "./bertabs-finetuned-cnndm-extractive-abstractive-summarization/pytorch_model.bin" )
if __name__ == "__main__":
snake_case : Optional[Any] = argparse.ArgumentParser()
parser.add_argument(
"--bertabs_checkpoint_path",
default=None,
type=str,
required=True,
help="Path the official PyTorch dump.",
)
parser.add_argument(
"--pytorch_dump_folder_path",
default=None,
type=str,
required=True,
help="Path to the output PyTorch model.",
)
snake_case : str = parser.parse_args()
convert_bertabs_checkpoints(
args.bertabs_checkpoint_path,
args.pytorch_dump_folder_path,
)
| 281 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import (
BitConfig,
ViTHybridConfig,
ViTHybridForImageClassification,
ViTHybridImageProcessor,
ViTHybridModel,
)
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
snake_case : Optional[Any] = logging.get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any]=False ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
# fmt: off
# stem:
rename_keys.append(("cls_token", "vit.embeddings.cls_token") )
rename_keys.append(("pos_embed", "vit.embeddings.position_embeddings") )
rename_keys.append(("patch_embed.proj.weight", "vit.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("patch_embed.proj.bias", "vit.embeddings.patch_embeddings.projection.bias") )
# backbone
rename_keys.append(("patch_embed.backbone.stem.conv.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.convolution.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.bias", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.bias") )
for stage_idx in range(len(config.backbone_config.depths ) ):
for layer_idx in range(config.backbone_config.depths[stage_idx] ):
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.conv.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.conv.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.bias''') )
# transformer encoder
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''vit.encoder.layer.{i}.output.dense.bias''') )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("norm.weight", "layernorm.weight"),
("norm.bias", "layernorm.bias"),
("pre_logits.fc.weight", "pooler.dense.weight"),
("pre_logits.fc.bias", "pooler.dense.bias"),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
__magic_name__ : int = [(pair[0], pair[1][4:]) if pair[1].startswith("vit" ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
("norm.weight", "vit.layernorm.weight"),
("norm.bias", "vit.layernorm.bias"),
("head.weight", "classifier.weight"),
("head.bias", "classifier.bias"),
] )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Any , _snake_case : Dict=False ) -> int:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
__magic_name__ : int = ""
else:
__magic_name__ : Union[str, Any] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__magic_name__ : Optional[Any] = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
__magic_name__ : int = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__magic_name__ : Dict = in_proj_weight[
: config.hidden_size, :
]
__magic_name__ : List[str] = in_proj_bias[: config.hidden_size]
__magic_name__ : List[str] = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__magic_name__ : Optional[Any] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__magic_name__ : Optional[Any] = in_proj_weight[
-config.hidden_size :, :
]
__magic_name__ : int = in_proj_bias[-config.hidden_size :]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
__magic_name__ : List[str] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : int , _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : int = dct.pop(_snake_case )
__magic_name__ : List[Any] = val
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
__magic_name__ : List[str] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Any , _snake_case : int=False ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=_snake_case , )
__magic_name__ : List[str] = ViTHybridConfig(backbone_config=_snake_case , image_size=384 , num_labels=1000 )
__magic_name__ : str = False
# load original model from timm
__magic_name__ : Union[str, Any] = timm.create_model(_snake_case , pretrained=_snake_case )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__magic_name__ : List[Any] = timm_model.state_dict()
if base_model:
remove_classification_head_(_snake_case )
__magic_name__ : Tuple = create_rename_keys(_snake_case , _snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
read_in_q_k_v(_snake_case , _snake_case , _snake_case )
__magic_name__ : List[str] = "huggingface/label-files"
__magic_name__ : int = "imagenet-1k-id2label.json"
__magic_name__ : Optional[int] = json.load(open(hf_hub_download(_snake_case , _snake_case , repo_type="dataset" ) , "r" ) )
__magic_name__ : int = {int(_snake_case ): v for k, v in idalabel.items()}
__magic_name__ : List[str] = idalabel
__magic_name__ : List[str] = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
__magic_name__ : List[str] = ViTHybridModel(_snake_case ).eval()
else:
__magic_name__ : str = ViTHybridForImageClassification(_snake_case ).eval()
model.load_state_dict(_snake_case )
# create image processor
__magic_name__ : List[Any] = create_transform(**resolve_data_config({} , model=_snake_case ) )
__magic_name__ : int = transform.transforms
__magic_name__ : List[str] = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
__magic_name__ : int = ViTHybridImageProcessor(
do_resize=_snake_case , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=_snake_case , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=_snake_case , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
__magic_name__ : List[Any] = prepare_img()
__magic_name__ : Any = transform(_snake_case ).unsqueeze(0 )
__magic_name__ : Tuple = processor(_snake_case , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(_snake_case , _snake_case )
# verify logits
with torch.no_grad():
__magic_name__ : Optional[int] = model(_snake_case )
__magic_name__ : List[str] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
__magic_name__ : List[str] = timm_model.forward_features(_snake_case )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(_snake_case , outputs.pooler_output , atol=1E-3 )
else:
__magic_name__ : Any = timm_model(_snake_case )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_snake_case , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(_snake_case ).mkdir(exist_ok=_snake_case )
print(F'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(_snake_case )
print(F'''Saving processor to {pytorch_dump_folder_path}''' )
processor.save_pretrained(_snake_case )
if push_to_hub:
print(F'''Pushing model and processor to the hub {vit_name}''' )
model.push_to_hub(F'''ybelkada/{vit_name}''' )
processor.push_to_hub(F'''ybelkada/{vit_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_r50_s16_384",
type=str,
help="Name of the hybrid ViT timm model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether to upload the model to the HuggingFace hub."
)
snake_case : List[Any] = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
from __future__ import annotations
from math import gcd
def lowerCAmelCase_ ( _snake_case : int , _snake_case : int = 2 , _snake_case : int = 1 , _snake_case : int = 3 , ) -> int | None:
'''simple docstring'''
if num < 2:
raise ValueError("The input value cannot be less than 2" )
# Because of the relationship between ``f(f(x))`` and ``f(x)``, this
# algorithm struggles to find factors that are divisible by two.
# As a workaround, we specifically check for two and even inputs.
# See: https://math.stackexchange.com/a/2856214/165820
if num > 2 and num % 2 == 0:
return 2
# Pollard's Rho algorithm requires a function that returns pseudorandom
# values between 0 <= X < ``num``. It doesn't need to be random in the
# sense that the output value is cryptographically secure or difficult
# to calculate, it only needs to be random in the sense that all output
# values should be equally likely to appear.
# For this reason, Pollard suggested using ``f(x) = (x**2 - 1) % num``
# However, the success of Pollard's algorithm isn't guaranteed and is
# determined in part by the initial seed and the chosen random function.
# To make retries easier, we will instead use ``f(x) = (x**2 + C) % num``
# where ``C`` is a value that we can modify between each attempt.
def rand_fn(_snake_case : int , _snake_case : int , _snake_case : int ) -> int:
return (pow(_snake_case , 2 ) + step) % modulus
for _ in range(_snake_case ):
# These track the position within the cycle detection logic.
__magic_name__ : str = seed
__magic_name__ : int = seed
while True:
# At each iteration, the tortoise moves one step and the hare moves two.
__magic_name__ : List[str] = rand_fn(_snake_case , _snake_case , _snake_case )
__magic_name__ : Dict = rand_fn(_snake_case , _snake_case , _snake_case )
__magic_name__ : int = rand_fn(_snake_case , _snake_case , _snake_case )
# At some point both the tortoise and the hare will enter a cycle whose
# length ``p`` is a divisor of ``num``. Once in that cycle, at some point
# the tortoise and hare will end up on the same value modulo ``p``.
# We can detect when this happens because the position difference between
# the tortoise and the hare will share a common divisor with ``num``.
__magic_name__ : List[str] = gcd(hare - tortoise , _snake_case )
if divisor == 1:
# No common divisor yet, just keep searching.
continue
else:
# We found a common divisor!
if divisor == num:
# Unfortunately, the divisor is ``num`` itself and is useless.
break
else:
# The divisor is a nontrivial factor of ``num``!
return divisor
# If we made it here, then this attempt failed.
# We need to pick a new starting seed for the tortoise and hare
# in addition to a new step value for the random function.
# To keep this example implementation deterministic, the
# new values will be generated based on currently available
# values instead of using something like ``random.randint``.
# We can use the hare's position as the new seed.
# This is actually what Richard Brent's the "optimized" variant does.
__magic_name__ : int = hare
# The new step value for the random function can just be incremented.
# At first the results will be similar to what the old function would
# have produced, but the value will quickly diverge after a bit.
step += 1
# We haven't found a divisor within the requested number of attempts.
# We were unlucky or ``num`` itself is actually prime.
return None
if __name__ == "__main__":
import argparse
snake_case : List[str] = argparse.ArgumentParser()
parser.add_argument(
"num",
type=int,
help="The value to find a divisor of",
)
parser.add_argument(
"--attempts",
type=int,
default=3,
help="The number of attempts before giving up",
)
snake_case : List[Any] = parser.parse_args()
snake_case : Any = pollard_rho(args.num, attempts=args.attempts)
if divisor is None:
print(F"{args.num} is probably prime")
else:
snake_case : Tuple = args.num // divisor
print(F"{args.num} = {divisor} * {quotient}")
| 281 |
# This script creates a super tiny model that is useful inside tests, when we just want to test that
# the machinery works, without needing to the check the quality of the outcomes.
#
# This version creates a tiny model through reduction of a normal pre-trained model, but keeping the
# full vocab, merges file, and thus also resulting in a larger model due to a large vocab size.
# This gives ~3MB in total for all files.
#
# If you want a 50 times smaller than this see `fsmt-make-super-tiny-model.py`, which is slightly more complicated
#
#
# It will be used then as "stas/tiny-wmt19-en-de"
# Build
from transformers import FSMTTokenizer, FSMTConfig, FSMTForConditionalGeneration
snake_case : List[str] = "facebook/wmt19-en-de"
snake_case : Dict = FSMTTokenizer.from_pretrained(mname)
# get the correct vocab sizes, etc. from the master model
snake_case : List[str] = FSMTConfig.from_pretrained(mname)
config.update(
dict(
d_model=4,
encoder_layers=1,
decoder_layers=1,
encoder_ffn_dim=4,
decoder_ffn_dim=4,
encoder_attention_heads=1,
decoder_attention_heads=1,
)
)
snake_case : int = FSMTForConditionalGeneration(config)
print(F"num of params {tiny_model.num_parameters()}")
# Test
snake_case : Optional[Any] = tokenizer(["Making tiny model"], return_tensors="pt")
snake_case : List[str] = tiny_model(**batch)
print("test output:", len(outputs.logits[0]))
# Save
snake_case : Dict = "tiny-wmt19-en-de"
tiny_model.half() # makes it smaller
tiny_model.save_pretrained(mname_tiny)
tokenizer.save_pretrained(mname_tiny)
print(F"Generated {mname_tiny}")
# Upload
# transformers-cli upload tiny-wmt19-en-de
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : int = 1000 ) -> int:
'''simple docstring'''
__magic_name__ , __magic_name__ : Union[str, Any] = 1, 1
__magic_name__ : Optional[int] = []
for i in range(1 , n + 1 ):
__magic_name__ : Tuple = prev_numerator + 2 * prev_denominator
__magic_name__ : Any = prev_numerator + prev_denominator
if len(str(_snake_case ) ) > len(str(_snake_case ) ):
result.append(_snake_case )
__magic_name__ : Any = numerator
__magic_name__ : int = denominator
return len(_snake_case )
if __name__ == "__main__":
print(F"{solution() = }")
| 281 |
import argparse
import csv
import logging
import os
import random
import numpy as np
import torch
from torch.utils.data import DataLoader, RandomSampler, SequentialSampler, TensorDataset
from tqdm import tqdm, trange
from transformers import (
CONFIG_NAME,
WEIGHTS_NAME,
AdamW,
OpenAIGPTDoubleHeadsModel,
OpenAIGPTTokenizer,
get_linear_schedule_with_warmup,
)
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s", datefmt="%m/%d/%Y %H:%M:%S", level=logging.INFO
)
snake_case : Optional[int] = logging.getLogger(__name__)
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : List[str] = np.argmax(_snake_case , axis=1 )
return np.sum(outputs == labels )
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
with open(_snake_case , encoding="utf_8" ) as f:
__magic_name__ : List[str] = csv.reader(_snake_case )
__magic_name__ : List[Any] = []
next(_snake_case ) # skip the first line
for line in tqdm(_snake_case ):
output.append((" ".join(line[1:5] ), line[5], line[6], int(line[-1] ) - 1) )
return output
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Tuple , _snake_case : Union[str, Any] , _snake_case : List[Any] , _snake_case : Tuple , _snake_case : Optional[int] ) -> int:
'''simple docstring'''
__magic_name__ : Optional[int] = []
for dataset in encoded_datasets:
__magic_name__ : Union[str, Any] = len(_snake_case )
__magic_name__ : Dict = np.zeros((n_batch, 2, input_len) , dtype=np.intaa )
__magic_name__ : List[str] = np.zeros((n_batch, 2) , dtype=np.intaa )
__magic_name__ : Optional[int] = np.full((n_batch, 2, input_len) , fill_value=-100 , dtype=np.intaa )
__magic_name__ : int = np.zeros((n_batch,) , dtype=np.intaa )
for (
i,
(story, conta, conta, mc_label),
) in enumerate(_snake_case ):
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : str = with_conta
__magic_name__ : Tuple = with_conta
__magic_name__ : Union[str, Any] = len(_snake_case ) - 1
__magic_name__ : int = len(_snake_case ) - 1
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[int] = mc_label
__magic_name__ : str = (input_ids, mc_token_ids, lm_labels, mc_labels)
tensor_datasets.append(tuple(torch.tensor(_snake_case ) for t in all_inputs ) )
return tensor_datasets
def lowerCAmelCase_ ( ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser()
parser.add_argument("--model_name" , type=_snake_case , default="openai-gpt" , help="pretrained model name" )
parser.add_argument("--do_train" , action="store_true" , help="Whether to run training." )
parser.add_argument("--do_eval" , action="store_true" , help="Whether to run eval on the dev set." )
parser.add_argument(
"--output_dir" , default=_snake_case , type=_snake_case , required=_snake_case , help="The output directory where the model predictions and checkpoints will be written." , )
parser.add_argument("--train_dataset" , type=_snake_case , default="" )
parser.add_argument("--eval_dataset" , type=_snake_case , default="" )
parser.add_argument("--seed" , type=_snake_case , default=42 )
parser.add_argument("--num_train_epochs" , type=_snake_case , default=3 )
parser.add_argument("--train_batch_size" , type=_snake_case , default=8 )
parser.add_argument("--eval_batch_size" , type=_snake_case , default=16 )
parser.add_argument("--adam_epsilon" , default=1E-8 , type=_snake_case , help="Epsilon for Adam optimizer." )
parser.add_argument("--max_grad_norm" , type=_snake_case , default=1 )
parser.add_argument(
"--max_steps" , default=-1 , type=_snake_case , help=(
"If > 0: set total number of training steps to perform. Override num_train_epochs."
) , )
parser.add_argument(
"--gradient_accumulation_steps" , type=_snake_case , default=1 , help="Number of updates steps to accumulate before performing a backward/update pass." , )
parser.add_argument("--learning_rate" , type=_snake_case , default=6.25E-5 )
parser.add_argument("--warmup_steps" , default=0 , type=_snake_case , help="Linear warmup over warmup_steps." )
parser.add_argument("--lr_schedule" , type=_snake_case , default="warmup_linear" )
parser.add_argument("--weight_decay" , type=_snake_case , default=0.01 )
parser.add_argument("--lm_coef" , type=_snake_case , default=0.9 )
parser.add_argument("--n_valid" , type=_snake_case , default=374 )
parser.add_argument("--server_ip" , type=_snake_case , default="" , help="Can be used for distant debugging." )
parser.add_argument("--server_port" , type=_snake_case , default="" , help="Can be used for distant debugging." )
__magic_name__ : List[Any] = parser.parse_args()
print(_snake_case )
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach" )
ptvsd.enable_attach(address=(args.server_ip, args.server_port) , redirect_output=_snake_case )
ptvsd.wait_for_attach()
random.seed(args.seed )
np.random.seed(args.seed )
torch.manual_seed(args.seed )
torch.cuda.manual_seed_all(args.seed )
__magic_name__ : Dict = torch.device("cuda" if torch.cuda.is_available() else "cpu" )
__magic_name__ : Optional[int] = torch.cuda.device_count()
logger.info("device: {}, n_gpu {}".format(_snake_case , _snake_case ) )
if not args.do_train and not args.do_eval:
raise ValueError("At least one of `do_train` or `do_eval` must be True." )
if not os.path.exists(args.output_dir ):
os.makedirs(args.output_dir )
# Load tokenizer and model
# This loading functions also add new tokens and embeddings called `special tokens`
# These new embeddings will be fine-tuned on the RocStories dataset
__magic_name__ : List[Any] = ["_start_", "_delimiter_", "_classify_"]
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.model_name )
tokenizer.add_tokens(_snake_case )
__magic_name__ : Optional[Any] = tokenizer.convert_tokens_to_ids(_snake_case )
__magic_name__ : List[str] = OpenAIGPTDoubleHeadsModel.from_pretrained(args.model_name )
model.resize_token_embeddings(len(_snake_case ) )
model.to(_snake_case )
# Load and encode the datasets
def tokenize_and_encode(_snake_case : str ):
if isinstance(_snake_case , _snake_case ):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(_snake_case ) )
elif isinstance(_snake_case , _snake_case ):
return obj
return [tokenize_and_encode(_snake_case ) for o in obj]
logger.info("Encoding dataset..." )
__magic_name__ : Optional[int] = load_rocstories_dataset(args.train_dataset )
__magic_name__ : str = load_rocstories_dataset(args.eval_dataset )
__magic_name__ : int = (train_dataset, eval_dataset)
__magic_name__ : List[str] = tokenize_and_encode(_snake_case )
# Compute the max input length for the Transformer
__magic_name__ : Optional[Any] = model.config.n_positions // 2 - 2
__magic_name__ : Optional[int] = max(
len(story[:max_length] ) + max(len(conta[:max_length] ) , len(conta[:max_length] ) ) + 3
for dataset in encoded_datasets
for story, conta, conta, _ in dataset )
__magic_name__ : List[str] = min(_snake_case , model.config.n_positions ) # Max size of input for the pre-trained model
# Prepare inputs tensors and dataloaders
__magic_name__ : List[Any] = pre_process_datasets(_snake_case , _snake_case , _snake_case , *_snake_case )
__magic_name__ , __magic_name__ : Optional[int] = tensor_datasets[0], tensor_datasets[1]
__magic_name__ : Tuple = TensorDataset(*_snake_case )
__magic_name__ : Union[str, Any] = RandomSampler(_snake_case )
__magic_name__ : Dict = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.train_batch_size )
__magic_name__ : Any = TensorDataset(*_snake_case )
__magic_name__ : Optional[Any] = SequentialSampler(_snake_case )
__magic_name__ : int = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.eval_batch_size )
# Prepare optimizer
if args.do_train:
if args.max_steps > 0:
__magic_name__ : Tuple = args.max_steps
__magic_name__ : List[str] = args.max_steps // (len(_snake_case ) // args.gradient_accumulation_steps) + 1
else:
__magic_name__ : List[str] = len(_snake_case ) // args.gradient_accumulation_steps * args.num_train_epochs
__magic_name__ : str = list(model.named_parameters() )
__magic_name__ : Dict = ["bias", "LayerNorm.bias", "LayerNorm.weight"]
__magic_name__ : str = [
{
"params": [p for n, p in param_optimizer if not any(nd in n for nd in no_decay )],
"weight_decay": args.weight_decay,
},
{"params": [p for n, p in param_optimizer if any(nd in n for nd in no_decay )], "weight_decay": 0.0},
]
__magic_name__ : str = AdamW(_snake_case , lr=args.learning_rate , eps=args.adam_epsilon )
__magic_name__ : List[str] = get_linear_schedule_with_warmup(
_snake_case , num_warmup_steps=args.warmup_steps , num_training_steps=_snake_case )
if args.do_train:
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = 0, 0, None
model.train()
for _ in trange(int(args.num_train_epochs ) , desc="Epoch" ):
__magic_name__ : List[str] = 0
__magic_name__ : Tuple = 0
__magic_name__ : Dict = tqdm(_snake_case , desc="Training" )
for step, batch in enumerate(_snake_case ):
__magic_name__ : Optional[Any] = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = batch
__magic_name__ : Optional[Any] = model(_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Optional[Any] = args.lm_coef * losses[0] + losses[1]
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
tr_loss += loss.item()
__magic_name__ : List[str] = (
loss.item() if exp_average_loss is None else 0.7 * exp_average_loss + 0.3 * loss.item()
)
nb_tr_steps += 1
__magic_name__ : int = "Training loss: {:.2e} lr: {:.2e}".format(_snake_case , scheduler.get_lr()[0] )
# Save a trained model
if args.do_train:
# Save a trained model, configuration and tokenizer
__magic_name__ : Dict = model.module if hasattr(_snake_case , "module" ) else model # Only save the model itself
# If we save using the predefined names, we can load using `from_pretrained`
__magic_name__ : List[Any] = os.path.join(args.output_dir , _snake_case )
__magic_name__ : Dict = os.path.join(args.output_dir , _snake_case )
torch.save(model_to_save.state_dict() , _snake_case )
model_to_save.config.to_json_file(_snake_case )
tokenizer.save_vocabulary(args.output_dir )
# Load a trained model and vocabulary that you have fine-tuned
__magic_name__ : Dict = OpenAIGPTDoubleHeadsModel.from_pretrained(args.output_dir )
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.output_dir )
model.to(_snake_case )
if args.do_eval:
model.eval()
__magic_name__ , __magic_name__ : Any = 0, 0
__magic_name__ , __magic_name__ : Union[str, Any] = 0, 0
for batch in tqdm(_snake_case , desc="Evaluating" ):
__magic_name__ : int = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = batch
with torch.no_grad():
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = model(
_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Tuple = mc_logits.detach().cpu().numpy()
__magic_name__ : Any = mc_labels.to("cpu" ).numpy()
__magic_name__ : str = accuracy(_snake_case , _snake_case )
eval_loss += mc_loss.mean().item()
eval_accuracy += tmp_eval_accuracy
nb_eval_examples += input_ids.size(0 )
nb_eval_steps += 1
__magic_name__ : Tuple = eval_loss / nb_eval_steps
__magic_name__ : List[Any] = eval_accuracy / nb_eval_examples
__magic_name__ : int = tr_loss / nb_tr_steps if args.do_train else None
__magic_name__ : Any = {"eval_loss": eval_loss, "eval_accuracy": eval_accuracy, "train_loss": train_loss}
__magic_name__ : int = os.path.join(args.output_dir , "eval_results.txt" )
with open(_snake_case , "w" ) as writer:
logger.info("***** Eval results *****" )
for key in sorted(result.keys() ):
logger.info(" %s = %s" , _snake_case , str(result[key] ) )
writer.write("%s = %s\n" % (key, str(result[key] )) )
if __name__ == "__main__":
main()
| 281 | 1 |
# 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.speechta import SpeechTaForTextToSpeech, SpeechTaHifiGan, SpeechTaProcessor
from ..utils import is_datasets_available
from .base import PipelineTool
if is_datasets_available():
from datasets import load_dataset
class _snake_case ( snake_case ):
UpperCamelCase__ = 'microsoft/speecht5_tts'
UpperCamelCase__ = (
'This is a tool that reads an English text out loud. It takes an input named `text` which should contain the '
'text to read (in English) and returns a waveform object containing the sound.'
)
UpperCamelCase__ = 'text_reader'
UpperCamelCase__ = SpeechTaProcessor
UpperCamelCase__ = SpeechTaForTextToSpeech
UpperCamelCase__ = SpeechTaHifiGan
UpperCamelCase__ = ['text']
UpperCamelCase__ = ['audio']
def SCREAMING_SNAKE_CASE ( self ):
if self.post_processor is None:
__magic_name__ : Union[str, Any] = "microsoft/speecht5_hifigan"
super().setup()
def SCREAMING_SNAKE_CASE ( self , _a , _a=None ):
__magic_name__ : Tuple = self.pre_processor(text=_a , return_tensors="pt" , truncation=_a )
if speaker_embeddings is None:
if not is_datasets_available():
raise ImportError("Datasets needs to be installed if not passing speaker embeddings." )
__magic_name__ : int = load_dataset("Matthijs/cmu-arctic-xvectors" , split="validation" )
__magic_name__ : Dict = torch.tensor(embeddings_dataset[7_305]["xvector"] ).unsqueeze(0 )
return {"input_ids": inputs["input_ids"], "speaker_embeddings": speaker_embeddings}
def SCREAMING_SNAKE_CASE ( self , _a ):
with torch.no_grad():
return self.model.generate_speech(**_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
with torch.no_grad():
return self.post_processor(_a ).cpu().detach()
| 281 |
from . import __version__
# Backward compatibility imports, to make sure all those objects can be found in file_utils
from .utils import (
CLOUDFRONT_DISTRIB_PREFIX,
CONFIG_NAME,
DISABLE_TELEMETRY,
DUMMY_INPUTS,
DUMMY_MASK,
ENV_VARS_TRUE_AND_AUTO_VALUES,
ENV_VARS_TRUE_VALUES,
FEATURE_EXTRACTOR_NAME,
FLAX_WEIGHTS_NAME,
HF_MODULES_CACHE,
HUGGINGFACE_CO_PREFIX,
HUGGINGFACE_CO_RESOLVE_ENDPOINT,
MODEL_CARD_NAME,
MULTIPLE_CHOICE_DUMMY_INPUTS,
PYTORCH_PRETRAINED_BERT_CACHE,
PYTORCH_TRANSFORMERS_CACHE,
S3_BUCKET_PREFIX,
SENTENCEPIECE_UNDERLINE,
SPIECE_UNDERLINE,
TF2_WEIGHTS_NAME,
TF_WEIGHTS_NAME,
TORCH_FX_REQUIRED_VERSION,
TRANSFORMERS_CACHE,
TRANSFORMERS_DYNAMIC_MODULE_NAME,
USE_JAX,
USE_TF,
USE_TORCH,
WEIGHTS_INDEX_NAME,
WEIGHTS_NAME,
ContextManagers,
DummyObject,
EntryNotFoundError,
ExplicitEnum,
ModelOutput,
PaddingStrategy,
PushToHubMixin,
RepositoryNotFoundError,
RevisionNotFoundError,
TensorType,
_LazyModule,
add_code_sample_docstrings,
add_end_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
cached_property,
copy_func,
default_cache_path,
define_sagemaker_information,
get_cached_models,
get_file_from_repo,
get_full_repo_name,
get_torch_version,
has_file,
http_user_agent,
is_apex_available,
is_bsa_available,
is_coloredlogs_available,
is_datasets_available,
is_detectrona_available,
is_faiss_available,
is_flax_available,
is_ftfy_available,
is_in_notebook,
is_ipex_available,
is_librosa_available,
is_offline_mode,
is_onnx_available,
is_pandas_available,
is_phonemizer_available,
is_protobuf_available,
is_psutil_available,
is_pyanvml_available,
is_pyctcdecode_available,
is_pytesseract_available,
is_pytorch_quantization_available,
is_rjieba_available,
is_sagemaker_dp_enabled,
is_sagemaker_mp_enabled,
is_scipy_available,
is_sentencepiece_available,
is_seqio_available,
is_sklearn_available,
is_soundfile_availble,
is_spacy_available,
is_speech_available,
is_tensor,
is_tensorflow_probability_available,
is_tfaonnx_available,
is_tf_available,
is_timm_available,
is_tokenizers_available,
is_torch_available,
is_torch_bfaa_available,
is_torch_cuda_available,
is_torch_fx_available,
is_torch_fx_proxy,
is_torch_mps_available,
is_torch_tfaa_available,
is_torch_tpu_available,
is_torchaudio_available,
is_training_run_on_sagemaker,
is_vision_available,
replace_return_docstrings,
requires_backends,
to_numpy,
to_py_obj,
torch_only_method,
)
| 281 | 1 |
import qiskit
def lowerCAmelCase_ ( _snake_case : int , _snake_case : int ) -> qiskit.result.counts.Counts:
'''simple docstring'''
__magic_name__ : int = qiskit.Aer.get_backend("aer_simulator" )
# Create a Quantum Circuit acting on the q register
__magic_name__ : str = qiskit.QuantumCircuit(_snake_case , _snake_case )
# Apply X (NOT) Gate to Qubits 0 & 1
circuit.x(0 )
circuit.x(1 )
# Map the quantum measurement to the classical bits
circuit.measure([0, 1] , [0, 1] )
# Execute the circuit on the qasm simulator
__magic_name__ : int = qiskit.execute(_snake_case , _snake_case , shots=1000 )
# Return the histogram data of the results of the experiment.
return job.result().get_counts(_snake_case )
if __name__ == "__main__":
snake_case : List[str] = single_qubit_measure(2, 2)
print(F"Total count for various states are: {counts}")
| 281 |
import importlib
import os
import fsspec
import pytest
from fsspec import register_implementation
from fsspec.registry import _registry as _fsspec_registry
from datasets.filesystems import COMPRESSION_FILESYSTEMS, HfFileSystem, extract_path_from_uri, is_remote_filesystem
from .utils import require_lza, require_zstandard
def lowerCAmelCase_ ( _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
assert "mock" in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
assert "mock" not in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Dict = "mock-s3-bucket"
__magic_name__ : Any = F'''s3://{mock_bucket}'''
__magic_name__ : str = extract_path_from_uri(_snake_case )
assert dataset_path.startswith("s3://" ) is False
__magic_name__ : Tuple = "./local/path"
__magic_name__ : Optional[Any] = extract_path_from_uri(_snake_case )
assert dataset_path == new_dataset_path
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : str = is_remote_filesystem(_snake_case )
assert is_remote is True
__magic_name__ : Optional[int] = fsspec.filesystem("file" )
__magic_name__ : int = is_remote_filesystem(_snake_case )
assert is_remote is False
@pytest.mark.parametrize("compression_fs_class" , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : int , _snake_case : Tuple , _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Any ) -> int:
'''simple docstring'''
__magic_name__ : Any = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_file, "bz2": bza_file, "lz4": lza_file}
__magic_name__ : str = input_paths[compression_fs_class.protocol]
if input_path is None:
__magic_name__ : Dict = F'''for \'{compression_fs_class.protocol}\' compression protocol, '''
if compression_fs_class.protocol == "lz4":
reason += require_lza.kwargs["reason"]
elif compression_fs_class.protocol == "zstd":
reason += require_zstandard.kwargs["reason"]
pytest.skip(_snake_case )
__magic_name__ : str = fsspec.filesystem(compression_fs_class.protocol , fo=_snake_case )
assert isinstance(_snake_case , _snake_case )
__magic_name__ : int = os.path.basename(_snake_case )
__magic_name__ : Optional[int] = expected_filename[: expected_filename.rindex("." )]
assert fs.glob("*" ) == [expected_filename]
with fs.open(_snake_case , "r" , encoding="utf-8" ) as f, open(_snake_case , encoding="utf-8" ) as expected_file:
assert f.read() == expected_file.read()
@pytest.mark.parametrize("protocol" , ["zip", "gzip"] )
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : Optional[Any] , _snake_case : Optional[Any] ) -> str:
'''simple docstring'''
__magic_name__ : int = {"zip": zip_jsonl_path, "gzip": jsonl_gz_path}
__magic_name__ : int = compressed_file_paths[protocol]
__magic_name__ : Tuple = "dataset.jsonl"
__magic_name__ : List[str] = F'''{protocol}://{member_file_path}::{compressed_file_path}'''
__magic_name__ , *__magic_name__ : Optional[Any] = fsspec.get_fs_token_paths(_snake_case )
assert fs.isfile(_snake_case )
assert not fs.isfile("non_existing_" + member_file_path )
@pytest.mark.integration
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : List[str] , _snake_case : Tuple ) -> str:
'''simple docstring'''
__magic_name__ : int = hf_api.dataset_info(_snake_case , token=_snake_case )
__magic_name__ : Optional[Any] = HfFileSystem(repo_info=_snake_case , token=_snake_case )
assert sorted(hffs.glob("*" ) ) == [".gitattributes", "data"]
assert hffs.isdir("data" )
assert hffs.isfile(".gitattributes" ) and hffs.isfile("data/text_data.txt" )
with open(_snake_case ) as f:
assert hffs.open("data/text_data.txt" , "r" ).read() == f.read()
def lowerCAmelCase_ ( ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Optional[Any] = "bz2"
# Import module
import datasets.filesystems
# Overwrite protocol and reload
register_implementation(_snake_case , _snake_case , clobber=_snake_case )
with pytest.warns(_snake_case ) as warning_info:
importlib.reload(datasets.filesystems )
assert len(_snake_case ) == 1
assert (
str(warning_info[0].message )
== F'''A filesystem protocol was already set for {protocol} and will be overwritten.'''
)
| 281 | 1 |
from math import sqrt
def lowerCAmelCase_ ( _snake_case : int ) -> bool:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (
number >= 0
), "'number' must been an int and positive"
__magic_name__ : Union[str, Any] = True
# 0 and 1 are none primes.
if number <= 1:
__magic_name__ : int = False
for divisor in range(2 , int(round(sqrt(_snake_case ) ) ) + 1 ):
# if 'number' divisible by 'divisor' then sets 'status'
# of false and break up the loop.
if number % divisor == 0:
__magic_name__ : int = False
break
# precondition
assert isinstance(_snake_case , _snake_case ), "'status' must been from type bool"
return status
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (n > 2), "'N' must been an int and > 2"
# beginList: contains all natural numbers from 2 up to N
__magic_name__ : Any = list(range(2 , n + 1 ) )
__magic_name__ : Optional[int] = [] # this list will be returns.
# actual sieve of erathostenes
for i in range(len(_snake_case ) ):
for j in range(i + 1 , len(_snake_case ) ):
if (begin_list[i] != 0) and (begin_list[j] % begin_list[i] == 0):
__magic_name__ : Union[str, Any] = 0
# filters actual prime numbers.
__magic_name__ : List[str] = [x for x in begin_list if x != 0]
# precondition
assert isinstance(_snake_case , _snake_case ), "'ans' must been from type list"
return ans
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (n > 2), "'N' must been an int and > 2"
__magic_name__ : 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(_snake_case ):
ans.append(_snake_case )
# precondition
assert isinstance(_snake_case , _snake_case ), "'ans' must been from type list"
return ans
def lowerCAmelCase_ ( _snake_case : str ) -> int:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and number >= 0, "'number' must been an int and >= 0"
__magic_name__ : int = [] # this list will be returns of the function.
# potential prime number factors.
__magic_name__ : Dict = 2
__magic_name__ : int = number
if number == 0 or number == 1:
ans.append(_snake_case )
# if 'number' not prime then builds the prime factorization of 'number'
elif not is_prime(_snake_case ):
while quotient != 1:
if is_prime(_snake_case ) and (quotient % factor == 0):
ans.append(_snake_case )
quotient /= factor
else:
factor += 1
else:
ans.append(_snake_case )
# precondition
assert isinstance(_snake_case , _snake_case ), "'ans' must been from type list"
return ans
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (
number >= 0
), "'number' bust been an int and >= 0"
__magic_name__ : List[str] = 0
# prime factorization of 'number'
__magic_name__ : Dict = prime_factorization(_snake_case )
__magic_name__ : List[str] = max(_snake_case )
# precondition
assert isinstance(_snake_case , _snake_case ), "'ans' must been from type int"
return ans
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (
number >= 0
), "'number' bust been an int and >= 0"
__magic_name__ : Any = 0
# prime factorization of 'number'
__magic_name__ : List[str] = prime_factorization(_snake_case )
__magic_name__ : Optional[int] = min(_snake_case )
# precondition
assert isinstance(_snake_case , _snake_case ), "'ans' must been from type int"
return ans
def lowerCAmelCase_ ( _snake_case : str ) -> Any:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ), "'number' must been an int"
assert isinstance(number % 2 == 0 , _snake_case ), "compare bust been from type bool"
return number % 2 == 0
def lowerCAmelCase_ ( _snake_case : str ) -> Dict:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ), "'number' must been an int"
assert isinstance(number % 2 != 0 , _snake_case ), "compare bust been from type bool"
return number % 2 != 0
def lowerCAmelCase_ ( _snake_case : Dict ) -> List[str]:
'''simple docstring'''
assert (
isinstance(_snake_case , _snake_case ) and (number > 2) and is_even(_snake_case )
), "'number' must been an int, even and > 2"
__magic_name__ : Optional[Any] = [] # this list will returned
# creates a list of prime numbers between 2 up to 'number'
__magic_name__ : List[Any] = get_prime_numbers(_snake_case )
__magic_name__ : List[str] = len(_snake_case )
# run variable for while-loops.
__magic_name__ : Dict = 0
__magic_name__ : Any = None
# exit variable. for break up the loops
__magic_name__ : Tuple = True
while i < len_pn and loop:
__magic_name__ : Optional[Any] = i + 1
while j < len_pn and loop:
if prime_numbers[i] + prime_numbers[j] == number:
__magic_name__ : int = False
ans.append(prime_numbers[i] )
ans.append(prime_numbers[j] )
j += 1
i += 1
# precondition
assert (
isinstance(_snake_case , _snake_case )
and (len(_snake_case ) == 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 lowerCAmelCase_ ( _snake_case : int , _snake_case : Optional[Any] ) -> List[Any]:
'''simple docstring'''
assert (
isinstance(_snake_case , _snake_case )
and isinstance(_snake_case , _snake_case )
and (numbera >= 0)
and (numbera >= 0)
), "'number1' and 'number2' must been positive integer."
__magic_name__ : str = 0
while numbera != 0:
__magic_name__ : str = numbera % numbera
__magic_name__ : str = numbera
__magic_name__ : Union[str, Any] = rest
# precondition
assert isinstance(_snake_case , _snake_case ) and (
numbera >= 0
), "'number' must been from type int and positive"
return numbera
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : List[str] ) -> int:
'''simple docstring'''
assert (
isinstance(_snake_case , _snake_case )
and isinstance(_snake_case , _snake_case )
and (numbera >= 1)
and (numbera >= 1)
), "'number1' and 'number2' must been positive integer."
__magic_name__ : 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'
__magic_name__ : str = prime_factorization(_snake_case )
__magic_name__ : Union[str, Any] = prime_factorization(_snake_case )
elif numbera == 1 or numbera == 1:
__magic_name__ : Optional[int] = []
__magic_name__ : Optional[Any] = []
__magic_name__ : int = max(_snake_case , _snake_case )
__magic_name__ : str = 0
__magic_name__ : Optional[int] = 0
__magic_name__ : Tuple = [] # 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:
__magic_name__ : Any = prime_fac_a.count(_snake_case )
__magic_name__ : Dict = prime_fac_a.count(_snake_case )
for _ in range(max(_snake_case , _snake_case ) ):
ans *= n
else:
__magic_name__ : Union[str, Any] = prime_fac_a.count(_snake_case )
for _ in range(_snake_case ):
ans *= n
done.append(_snake_case )
# iterates through primeFac2
for n in prime_fac_a:
if n not in done:
__magic_name__ : str = prime_fac_a.count(_snake_case )
for _ in range(_snake_case ):
ans *= n
done.append(_snake_case )
# precondition
assert isinstance(_snake_case , _snake_case ) and (
ans >= 0
), "'ans' must been from type int and positive"
return ans
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> List[str]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (n >= 0), "'number' must been a positive int"
__magic_name__ : int = 0
__magic_name__ : List[str] = 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(_snake_case ):
ans += 1
# precondition
assert isinstance(_snake_case , _snake_case ) and is_prime(
_snake_case ), "'ans' must been a prime number and from type int"
return ans
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : int ) -> Any:
'''simple docstring'''
assert (
is_prime(_snake_case ) and is_prime(_snake_case ) and (p_number_a < p_number_a)
), "The arguments must been prime numbers and 'pNumber1' < 'pNumber2'"
__magic_name__ : Tuple = p_number_a + 1 # jump to the next number
__magic_name__ : List[Any] = [] # this list will be returns.
# if number is not prime then
# fetch the next prime number.
while not is_prime(_snake_case ):
number += 1
while number < p_number_a:
ans.append(_snake_case )
number += 1
# fetch the next prime number.
while not is_prime(_snake_case ):
number += 1
# precondition
assert (
isinstance(_snake_case , _snake_case )
and ans[0] != p_number_a
and ans[len(_snake_case ) - 1] != p_number_a
), "'ans' must been a list without the arguments"
# 'ans' contains not 'pNumber1' and 'pNumber2' !
return ans
def lowerCAmelCase_ ( _snake_case : str ) -> Optional[int]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (n >= 1), "'n' must been int and >= 1"
__magic_name__ : Any = [] # will be returned.
for divisor in range(1 , n + 1 ):
if n % divisor == 0:
ans.append(_snake_case )
# precondition
assert ans[0] == 1 and ans[len(_snake_case ) - 1] == n, "Error in function getDivisiors(...)"
return ans
def lowerCAmelCase_ ( _snake_case : Tuple ) -> Optional[Any]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (
number > 1
), "'number' must been an int and >= 1"
__magic_name__ : int = get_divisors(_snake_case )
# precondition
assert (
isinstance(_snake_case , _snake_case )
and (divisors[0] == 1)
and (divisors[len(_snake_case ) - 1] == number)
), "Error in help-function getDivisiors(...)"
# summed all divisors up to 'number' (exclusive), hence [:-1]
return sum(divisors[:-1] ) == number
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> str:
'''simple docstring'''
assert (
isinstance(_snake_case , _snake_case )
and isinstance(_snake_case , _snake_case )
and (denominator != 0)
), "The arguments must been from type int and 'denominator' != 0"
# build the greatest common divisor of numerator and denominator.
__magic_name__ : Any = gcd(abs(_snake_case ) , abs(_snake_case ) )
# precondition
assert (
isinstance(_snake_case , _snake_case )
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 lowerCAmelCase_ ( _snake_case : Dict ) -> List[str]:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (n >= 0), "'n' must been a int and >= 0"
__magic_name__ : int = 1 # this will be return.
for factor in range(1 , n + 1 ):
ans *= factor
return ans
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Dict:
'''simple docstring'''
assert isinstance(_snake_case , _snake_case ) and (n >= 0), "'n' must been an int and >= 0"
__magic_name__ : Dict = 0
__magic_name__ : str = 1
__magic_name__ : List[Any] = 1 # this will be return
for _ in range(n - 1 ):
__magic_name__ : Optional[Any] = ans
ans += fiba
__magic_name__ : Dict = tmp
return ans
| 281 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : List[Any] = {
"YituTech/conv-bert-base": "https://huggingface.co/YituTech/conv-bert-base/resolve/main/config.json",
"YituTech/conv-bert-medium-small": (
"https://huggingface.co/YituTech/conv-bert-medium-small/resolve/main/config.json"
),
"YituTech/conv-bert-small": "https://huggingface.co/YituTech/conv-bert-small/resolve/main/config.json",
# See all ConvBERT models at https://huggingface.co/models?filter=convbert
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'convbert'
def __init__( self , _a=30_522 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=1 , _a=0 , _a=2 , _a=768 , _a=2 , _a=9 , _a=1 , _a=None , **_a , ):
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a , )
__magic_name__ : Tuple = vocab_size
__magic_name__ : List[Any] = hidden_size
__magic_name__ : Union[str, Any] = num_hidden_layers
__magic_name__ : List[Any] = num_attention_heads
__magic_name__ : str = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : Tuple = max_position_embeddings
__magic_name__ : str = type_vocab_size
__magic_name__ : List[str] = initializer_range
__magic_name__ : Tuple = layer_norm_eps
__magic_name__ : List[Any] = embedding_size
__magic_name__ : List[Any] = head_ratio
__magic_name__ : str = conv_kernel_size
__magic_name__ : Dict = num_groups
__magic_name__ : str = classifier_dropout
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : Dict = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
("token_type_ids", dynamic_axis),
] )
| 281 | 1 |
snake_case : Optional[int] = {"a": ["c", "b"], "b": ["d", "e"], "c": [], "d": [], "e": []}
snake_case : Optional[int] = ["a", "b", "c", "d", "e"]
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] , _snake_case : Dict ) -> int:
'''simple docstring'''
__magic_name__ : Any = start
# add current to visited
visited.append(_snake_case )
__magic_name__ : int = edges[current]
for neighbor in neighbors:
# if neighbor not in visited, visit
if neighbor not in visited:
__magic_name__ : str = topological_sort(_snake_case , _snake_case , _snake_case )
# if all neighbors visited add current to sort
sort.append(_snake_case )
# if all vertices haven't been visited select a new one to visit
if len(_snake_case ) != len(_snake_case ):
for vertice in vertices:
if vertice not in visited:
__magic_name__ : Optional[int] = topological_sort(_snake_case , _snake_case , _snake_case )
# return sort
return sort
if __name__ == "__main__":
snake_case : str = topological_sort("a", [], [])
print(sort)
| 281 |
import argparse
import requests
import torch
# pip3 install salesforce-lavis
# I'm actually installing a slightly modified version: pip3 install git+https://github.com/nielsrogge/LAVIS.git@fix_lavis
from lavis.models import load_model_and_preprocess
from PIL import Image
from transformers import (
AutoTokenizer,
BlipaConfig,
BlipaForConditionalGeneration,
BlipaProcessor,
BlipaVisionConfig,
BlipImageProcessor,
OPTConfig,
TaConfig,
)
from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD
def lowerCAmelCase_ ( ) -> str:
'''simple docstring'''
__magic_name__ : int = "https://storage.googleapis.com/sfr-vision-language-research/LAVIS/assets/merlion.png"
__magic_name__ : Union[str, Any] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw ).convert("RGB" )
return image
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = []
# fmt: off
# vision encoder
rename_keys.append(("visual_encoder.cls_token", "vision_model.embeddings.class_embedding") )
rename_keys.append(("visual_encoder.pos_embed", "vision_model.embeddings.position_embedding") )
rename_keys.append(("visual_encoder.patch_embed.proj.weight", "vision_model.embeddings.patch_embedding.weight") )
rename_keys.append(("visual_encoder.patch_embed.proj.bias", "vision_model.embeddings.patch_embedding.bias") )
rename_keys.append(("ln_vision.weight", "vision_model.post_layernorm.weight") )
rename_keys.append(("ln_vision.bias", "vision_model.post_layernorm.bias") )
for i in range(config.vision_config.num_hidden_layers ):
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.weight''', F'''vision_model.encoder.layers.{i}.layer_norm1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.bias''', F'''vision_model.encoder.layers.{i}.layer_norm1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.weight''', F'''vision_model.encoder.layers.{i}.layer_norm2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.bias''', F'''vision_model.encoder.layers.{i}.layer_norm2.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.qkv.weight''', F'''vision_model.encoder.layers.{i}.self_attn.qkv.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.weight''', F'''vision_model.encoder.layers.{i}.self_attn.projection.weight''',) )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.bias''', F'''vision_model.encoder.layers.{i}.self_attn.projection.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc2.bias''') )
# QFormer
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.weight", "qformer.layernorm.weight") )
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.bias", "qformer.layernorm.bias") )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Optional[Any] ) -> int:
'''simple docstring'''
__magic_name__ : Tuple = dct.pop(_snake_case )
__magic_name__ : int = val
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
for i in range(config.vision_config.num_hidden_layers ):
# read in original q and v biases
__magic_name__ : List[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.q_bias''' )
__magic_name__ : Optional[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.v_bias''' )
# next, set bias in the state dict
__magic_name__ : Optional[int] = torch.cat((q_bias, torch.zeros_like(_snake_case , requires_grad=_snake_case ), v_bias) )
__magic_name__ : Union[str, Any] = qkv_bias
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : str ) -> int:
'''simple docstring'''
__magic_name__ : List[Any] = 364 if "coco" in model_name else 224
__magic_name__ : Union[str, Any] = BlipaVisionConfig(image_size=_snake_case ).to_dict()
# make sure the models have proper bos_token_id and eos_token_id set (important for generation)
# seems like flan-T5 models don't have bos_token_id properly set?
if "opt-2.7b" in model_name:
__magic_name__ : List[str] = OPTConfig.from_pretrained("facebook/opt-2.7b" , eos_token_id=_snake_case ).to_dict()
elif "opt-6.7b" in model_name:
__magic_name__ : Any = OPTConfig.from_pretrained("facebook/opt-6.7b" , eos_token_id=_snake_case ).to_dict()
elif "t5-xl" in model_name:
__magic_name__ : Dict = TaConfig.from_pretrained("google/flan-t5-xl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
elif "t5-xxl" in model_name:
__magic_name__ : int = TaConfig.from_pretrained("google/flan-t5-xxl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
__magic_name__ : List[Any] = BlipaConfig(vision_config=_snake_case , text_config=_snake_case )
return config, image_size
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : str=None , _snake_case : Dict=False ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Optional[int] = (
AutoTokenizer.from_pretrained("facebook/opt-2.7b" )
if "opt" in model_name
else AutoTokenizer.from_pretrained("google/flan-t5-xl" )
)
__magic_name__ : List[Any] = tokenizer("\n" , add_special_tokens=_snake_case ).input_ids[0]
__magic_name__ , __magic_name__ : Tuple = get_blipa_config(_snake_case , eos_token_id=_snake_case )
__magic_name__ : Union[str, Any] = BlipaForConditionalGeneration(_snake_case ).eval()
__magic_name__ : Any = {
"blip2-opt-2.7b": ("blip2_opt", "pretrain_opt2.7b"),
"blip2-opt-6.7b": ("blip2_opt", "pretrain_opt6.7b"),
"blip2-opt-2.7b-coco": ("blip2_opt", "caption_coco_opt2.7b"),
"blip2-opt-6.7b-coco": ("blip2_opt", "caption_coco_opt6.7b"),
"blip2-flan-t5-xl": ("blip2_t5", "pretrain_flant5xl"),
"blip2-flan-t5-xl-coco": ("blip2_t5", "caption_coco_flant5xl"),
"blip2-flan-t5-xxl": ("blip2_t5", "pretrain_flant5xxl"),
}
__magic_name__ , __magic_name__ : Union[str, Any] = model_name_to_original[model_name]
# load original model
print("Loading original model..." )
__magic_name__ : Union[str, Any] = "cuda" if torch.cuda.is_available() else "cpu"
__magic_name__ , __magic_name__ , __magic_name__ : Optional[Any] = load_model_and_preprocess(
name=_snake_case , model_type=_snake_case , is_eval=_snake_case , device=_snake_case )
original_model.eval()
print("Done!" )
# update state dict keys
__magic_name__ : Dict = original_model.state_dict()
__magic_name__ : str = create_rename_keys(_snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
# some keys can be renamed efficiently
for key, val in state_dict.copy().items():
__magic_name__ : Any = state_dict.pop(_snake_case )
if key.startswith("Qformer.bert" ):
__magic_name__ : Optional[int] = key.replace("Qformer.bert" , "qformer" )
if "attention.self" in key:
__magic_name__ : Any = key.replace("self" , "attention" )
if "opt_proj" in key:
__magic_name__ : Union[str, Any] = key.replace("opt_proj" , "language_projection" )
if "t5_proj" in key:
__magic_name__ : Optional[int] = key.replace("t5_proj" , "language_projection" )
if key.startswith("opt" ):
__magic_name__ : List[str] = key.replace("opt" , "language" )
if key.startswith("t5" ):
__magic_name__ : Tuple = key.replace("t5" , "language" )
__magic_name__ : Dict = val
# read in qv biases
read_in_q_v_bias(_snake_case , _snake_case )
__magic_name__ , __magic_name__ : Tuple = hf_model.load_state_dict(_snake_case , strict=_snake_case )
assert len(_snake_case ) == 0
assert unexpected_keys == ["qformer.embeddings.position_ids"]
__magic_name__ : List[Any] = load_demo_image()
__magic_name__ : Tuple = vis_processors["eval"](_snake_case ).unsqueeze(0 ).to(_snake_case )
__magic_name__ : Dict = tokenizer(["\n"] , return_tensors="pt" ).input_ids.to(_snake_case )
# create processor
__magic_name__ : Optional[Any] = BlipImageProcessor(
size={"height": image_size, "width": image_size} , image_mean=_snake_case , image_std=_snake_case )
__magic_name__ : Dict = BlipaProcessor(image_processor=_snake_case , tokenizer=_snake_case )
__magic_name__ : Union[str, Any] = processor(images=_snake_case , return_tensors="pt" ).pixel_values.to(_snake_case )
# make sure processor creates exact same pixel values
assert torch.allclose(_snake_case , _snake_case )
original_model.to(_snake_case )
hf_model.to(_snake_case )
with torch.no_grad():
if "opt" in model_name:
__magic_name__ : List[Any] = original_model({"image": original_pixel_values, "text_input": [""]} ).logits
__magic_name__ : Optional[int] = hf_model(_snake_case , _snake_case ).logits
else:
__magic_name__ : int = original_model(
{"image": original_pixel_values, "text_input": ["\n"], "text_output": ["\n"]} ).logits
__magic_name__ : Tuple = input_ids.masked_fill(input_ids == tokenizer.pad_token_id , -100 )
__magic_name__ : List[str] = hf_model(_snake_case , _snake_case , labels=_snake_case ).logits
assert original_logits.shape == logits.shape
print("First values of original logits:" , original_logits[0, :3, :3] )
print("First values of HF logits:" , logits[0, :3, :3] )
# assert values
if model_name == "blip2-flan-t5-xl":
__magic_name__ : List[str] = torch.tensor(
[[-41.5_850, -4.4_440, -8.9_922], [-47.4_322, -5.9_143, -1.7_340]] , device=_snake_case )
assert torch.allclose(logits[0, :3, :3] , _snake_case , atol=1E-4 )
elif model_name == "blip2-flan-t5-xl-coco":
__magic_name__ : Tuple = torch.tensor(
[[-57.0_109, -9.8_967, -12.6_280], [-68.6_578, -12.7_191, -10.5_065]] , device=_snake_case )
else:
# cast to same type
__magic_name__ : str = logits.dtype
assert torch.allclose(original_logits.to(_snake_case ) , _snake_case , atol=1E-2 )
print("Looks ok!" )
print("Generating a caption..." )
__magic_name__ : Optional[int] = ""
__magic_name__ : Dict = tokenizer(_snake_case , return_tensors="pt" ).input_ids.to(_snake_case )
__magic_name__ : int = original_model.generate({"image": original_pixel_values} )
__magic_name__ : Optional[Any] = hf_model.generate(
_snake_case , _snake_case , do_sample=_snake_case , num_beams=5 , max_length=30 , min_length=1 , top_p=0.9 , repetition_penalty=1.0 , length_penalty=1.0 , temperature=1 , )
print("Original generation:" , _snake_case )
__magic_name__ : Tuple = input_ids.shape[1]
__magic_name__ : int = processor.batch_decode(outputs[:, prompt_length:] , skip_special_tokens=_snake_case )
__magic_name__ : Union[str, Any] = [text.strip() for text in output_text]
print("HF generation:" , _snake_case )
if pytorch_dump_folder_path is not None:
processor.save_pretrained(_snake_case )
hf_model.save_pretrained(_snake_case )
if push_to_hub:
processor.push_to_hub(F'''nielsr/{model_name}''' )
hf_model.push_to_hub(F'''nielsr/{model_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
snake_case : Union[str, Any] = [
"blip2-opt-2.7b",
"blip2-opt-6.7b",
"blip2-opt-2.7b-coco",
"blip2-opt-6.7b-coco",
"blip2-flan-t5-xl",
"blip2-flan-t5-xl-coco",
"blip2-flan-t5-xxl",
]
parser.add_argument(
"--model_name",
default="blip2-opt-2.7b",
choices=choices,
type=str,
help="Path to hf config.json of model to convert",
)
parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.")
parser.add_argument(
"--push_to_hub",
action="store_true",
help="Whether to push the model and processor to the hub after converting",
)
snake_case : int = parser.parse_args()
convert_blipa_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
import json
import sys
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Optional[int] ) -> Optional[int]:
'''simple docstring'''
with open(_snake_case , encoding="utf-8" ) as f:
__magic_name__ : List[str] = json.load(_snake_case )
__magic_name__ : Tuple = ["<details>", "<summary>Show updated benchmarks!</summary>", " "]
for benchmark_name in sorted(_snake_case ):
__magic_name__ : Tuple = results[benchmark_name]
__magic_name__ : str = benchmark_name.split("/" )[-1]
output_md.append(F'''### Benchmark: {benchmark_file_name}''' )
__magic_name__ : int = "| metric |"
__magic_name__ : Any = "|--------|"
__magic_name__ : int = "| new / old (diff) |"
for metric_name in sorted(_snake_case ):
__magic_name__ : List[Any] = benchmark_res[metric_name]
__magic_name__ : Any = metric_vals["new"]
__magic_name__ : Optional[int] = metric_vals.get("old" , _snake_case )
__magic_name__ : Any = metric_vals.get("diff" , _snake_case )
__magic_name__ : Optional[Any] = F''' {new_val:f}''' if isinstance(_snake_case , (int, float) ) else "None"
if old_val is not None:
val_str += F''' / {old_val:f}''' if isinstance(_snake_case , (int, float) ) else "None"
if dif_val is not None:
val_str += F''' ({dif_val:f})''' if isinstance(_snake_case , (int, float) ) else "None"
title += " " + metric_name + " |"
lines += "---|"
value += val_str + " |"
output_md += [title, lines, value, " "]
output_md.append("</details>" )
with open(_snake_case , "w" , encoding="utf-8" ) as f:
f.writelines("\n".join(_snake_case ) )
if __name__ == "__main__":
snake_case : Dict = sys.argv[1]
snake_case : List[Any] = sys.argv[2]
format_json_to_md(input_json_file, output_md_file)
| 281 |
import os
import re
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : Union[str, Any] = {
"vocab_file": "vocab.txt",
"merges_file": "bpe.codes",
}
snake_case : Dict = {
"vocab_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/vocab.txt",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/vocab.txt",
},
"merges_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/bpe.codes",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/bpe.codes",
},
}
snake_case : Union[str, Any] = {
"vinai/phobert-base": 256,
"vinai/phobert-large": 256,
}
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = set()
__magic_name__ : Any = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
__magic_name__ : int = char
__magic_name__ : List[str] = set(_snake_case )
return pairs
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a , _a="<s>" , _a="</s>" , _a="</s>" , _a="<s>" , _a="<unk>" , _a="<pad>" , _a="<mask>" , **_a , ):
super().__init__(
bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , cls_token=_a , pad_token=_a , mask_token=_a , **_a , )
__magic_name__ : Dict = vocab_file
__magic_name__ : Tuple = merges_file
__magic_name__ : List[Any] = {}
__magic_name__ : List[Any] = 0
__magic_name__ : Tuple = 1
__magic_name__ : int = 2
__magic_name__ : Union[str, Any] = 3
self.add_from_file(_a )
__magic_name__ : Optional[int] = {v: k for k, v in self.encoder.items()}
with open(_a , encoding="utf-8" ) as merges_handle:
__magic_name__ : List[str] = merges_handle.read().split("\n" )[:-1]
__magic_name__ : Union[str, Any] = [tuple(merge.split()[:-1] ) for merge in merges]
__magic_name__ : Union[str, Any] = dict(zip(_a , range(len(_a ) ) ) )
__magic_name__ : Optional[int] = {}
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
__magic_name__ : Optional[Any] = [self.cls_token_id]
__magic_name__ : Union[str, Any] = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is None:
return [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[Any] = [self.sep_token_id]
__magic_name__ : 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 + sep + token_ids_a + sep ) * [0]
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.encoder )
def SCREAMING_SNAKE_CASE ( self ):
return dict(self.encoder , **self.added_tokens_encoder )
def SCREAMING_SNAKE_CASE ( self , _a ):
if token in self.cache:
return self.cache[token]
__magic_name__ : List[Any] = tuple(_a )
__magic_name__ : List[Any] = tuple(list(word[:-1] ) + [word[-1] + "</w>"] )
__magic_name__ : Any = get_pairs(_a )
if not pairs:
return token
while True:
__magic_name__ : str = min(_a , key=lambda _a : self.bpe_ranks.get(_a , float("inf" ) ) )
if bigram not in self.bpe_ranks:
break
__magic_name__ , __magic_name__ : List[str] = bigram
__magic_name__ : List[str] = []
__magic_name__ : List[str] = 0
while i < len(_a ):
try:
__magic_name__ : Any = word.index(_a , _a )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
__magic_name__ : Tuple = j
if word[i] == first and i < len(_a ) - 1 and word[i + 1] == second:
new_word.append(first + second )
i += 2
else:
new_word.append(word[i] )
i += 1
__magic_name__ : Union[str, Any] = tuple(_a )
__magic_name__ : Optional[int] = new_word
if len(_a ) == 1:
break
else:
__magic_name__ : List[Any] = get_pairs(_a )
__magic_name__ : Optional[int] = "@@ ".join(_a )
__magic_name__ : Tuple = word[:-4]
__magic_name__ : str = word
return word
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = []
__magic_name__ : Dict = re.findall(r"\S+\n?" , _a )
for token in words:
split_tokens.extend(list(self.bpe(_a ).split(" " ) ) )
return split_tokens
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.encoder.get(_a , self.encoder.get(self.unk_token ) )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.decoder.get(_a , self.unk_token )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Tuple = " ".join(_a ).replace("@@ " , "" ).strip()
return out_string
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : Optional[int] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
__magic_name__ : Union[str, Any] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["merges_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ):
copyfile(self.vocab_file , _a )
if os.path.abspath(self.merges_file ) != os.path.abspath(_a ):
copyfile(self.merges_file , _a )
return out_vocab_file, out_merge_file
def SCREAMING_SNAKE_CASE ( self , _a ):
if isinstance(_a , _a ):
try:
with open(_a , "r" , encoding="utf-8" ) as fd:
self.add_from_file(_a )
except FileNotFoundError as fnfe:
raise fnfe
except UnicodeError:
raise Exception(f'''Incorrect encoding detected in {f}, please rebuild the dataset''' )
return
__magic_name__ : List[Any] = f.readlines()
for lineTmp in lines:
__magic_name__ : Optional[Any] = lineTmp.strip()
__magic_name__ : Union[str, Any] = line.rfind(" " )
if idx == -1:
raise ValueError("Incorrect dictionary format, expected '<token> <cnt>'" )
__magic_name__ : Optional[int] = line[:idx]
__magic_name__ : Dict = len(self.encoder )
| 281 | 1 |
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
snake_case : Optional[Any] = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def lowerCAmelCase_ ( _snake_case : Dict ) -> Dict:
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def lowerCAmelCase_ ( _snake_case : Tuple , _snake_case : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
if args.student_type == "roberta":
__magic_name__ : str = False
elif args.student_type == "gpt2":
__magic_name__ : Union[str, Any] = False
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Optional[int] ) -> List[Any]:
'''simple docstring'''
if args.student_type == "roberta":
__magic_name__ : Tuple = False
def lowerCAmelCase_ ( ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : Optional[int] = argparse.ArgumentParser(description="Training" )
parser.add_argument("--force" , action="store_true" , help="Overwrite dump_path if it already exists." )
parser.add_argument(
"--dump_path" , type=_snake_case , required=_snake_case , help="The output directory (log, checkpoints, parameters, etc.)" )
parser.add_argument(
"--data_file" , type=_snake_case , required=_snake_case , help="The binarized file (tokenized + tokens_to_ids) and grouped by sequence." , )
parser.add_argument(
"--student_type" , type=_snake_case , choices=["distilbert", "roberta", "gpt2"] , required=_snake_case , help="The student type (DistilBERT, RoBERTa)." , )
parser.add_argument("--student_config" , type=_snake_case , required=_snake_case , help="Path to the student configuration." )
parser.add_argument(
"--student_pretrained_weights" , default=_snake_case , type=_snake_case , help="Load student initialization checkpoint." )
parser.add_argument(
"--teacher_type" , choices=["bert", "roberta", "gpt2"] , required=_snake_case , help="Teacher type (BERT, RoBERTa)." )
parser.add_argument("--teacher_name" , type=_snake_case , required=_snake_case , help="The teacher model." )
parser.add_argument("--temperature" , default=2.0 , type=_snake_case , help="Temperature for the softmax temperature." )
parser.add_argument(
"--alpha_ce" , default=0.5 , type=_snake_case , help="Linear weight for the distillation loss. Must be >=0." )
parser.add_argument(
"--alpha_mlm" , default=0.0 , type=_snake_case , help="Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag." , )
parser.add_argument("--alpha_clm" , default=0.5 , type=_snake_case , help="Linear weight for the CLM loss. Must be >=0." )
parser.add_argument("--alpha_mse" , default=0.0 , type=_snake_case , help="Linear weight of the MSE loss. Must be >=0." )
parser.add_argument(
"--alpha_cos" , default=0.0 , type=_snake_case , help="Linear weight of the cosine embedding loss. Must be >=0." )
parser.add_argument(
"--mlm" , action="store_true" , help="The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM." )
parser.add_argument(
"--mlm_mask_prop" , default=0.15 , type=_snake_case , help="Proportion of tokens for which we need to make a prediction." , )
parser.add_argument("--word_mask" , default=0.8 , type=_snake_case , help="Proportion of tokens to mask out." )
parser.add_argument("--word_keep" , default=0.1 , type=_snake_case , help="Proportion of tokens to keep." )
parser.add_argument("--word_rand" , default=0.1 , type=_snake_case , help="Proportion of tokens to randomly replace." )
parser.add_argument(
"--mlm_smoothing" , default=0.7 , type=_snake_case , help="Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec)." , )
parser.add_argument("--token_counts" , type=_snake_case , help="The token counts in the data_file for MLM." )
parser.add_argument(
"--restrict_ce_to_mask" , action="store_true" , help="If true, compute the distillation loss only the [MLM] prediction distribution." , )
parser.add_argument(
"--freeze_pos_embs" , action="store_true" , help="Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only." , )
parser.add_argument(
"--freeze_token_type_embds" , action="store_true" , help="Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only." , )
parser.add_argument("--n_epoch" , type=_snake_case , default=3 , help="Number of pass on the whole dataset." )
parser.add_argument("--batch_size" , type=_snake_case , default=5 , help="Batch size (for each process)." )
parser.add_argument(
"--group_by_size" , action="store_false" , help="If true, group sequences that have similar length into the same batch. Default is true." , )
parser.add_argument(
"--gradient_accumulation_steps" , type=_snake_case , default=50 , help="Gradient accumulation for larger training batches." , )
parser.add_argument("--warmup_prop" , default=0.05 , type=_snake_case , help="Linear warmup proportion." )
parser.add_argument("--weight_decay" , default=0.0 , type=_snake_case , help="Weight decay if we apply some." )
parser.add_argument("--learning_rate" , default=5E-4 , type=_snake_case , help="The initial learning rate for Adam." )
parser.add_argument("--adam_epsilon" , default=1E-6 , type=_snake_case , help="Epsilon for Adam optimizer." )
parser.add_argument("--max_grad_norm" , default=5.0 , type=_snake_case , help="Max gradient norm." )
parser.add_argument("--initializer_range" , default=0.02 , type=_snake_case , help="Random initialization range." )
parser.add_argument(
"--fp16" , action="store_true" , help="Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit" , )
parser.add_argument(
"--fp16_opt_level" , type=_snake_case , default="O1" , help=(
"For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."
"See details at https://nvidia.github.io/apex/amp.html"
) , )
parser.add_argument("--n_gpu" , type=_snake_case , default=1 , help="Number of GPUs in the node." )
parser.add_argument("--local_rank" , type=_snake_case , default=-1 , help="Distributed training - Local rank" )
parser.add_argument("--seed" , type=_snake_case , default=56 , help="Random seed" )
parser.add_argument("--log_interval" , type=_snake_case , default=500 , help="Tensorboard logging interval." )
parser.add_argument("--checkpoint_interval" , type=_snake_case , default=4000 , help="Checkpoint interval." )
__magic_name__ : Any = parser.parse_args()
sanity_checks(_snake_case )
# ARGS #
init_gpu_params(_snake_case )
set_seed(_snake_case )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
F'''Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite'''
" itUse `--force` if you want to overwrite it" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(F'''Experiment will be dumped and logged in {args.dump_path}''' )
# SAVE PARAMS #
logger.info(F'''Param: {args}''' )
with open(os.path.join(args.dump_path , "parameters.json" ) , "w" ) as f:
json.dump(vars(_snake_case ) , _snake_case , indent=4 )
git_log(args.dump_path )
__magic_name__ , __magic_name__ , __magic_name__ : int = MODEL_CLASSES[args.student_type]
__magic_name__ , __magic_name__ , __magic_name__ : Tuple = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
__magic_name__ : Union[str, Any] = teacher_tokenizer_class.from_pretrained(args.teacher_name )
__magic_name__ : Any = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
__magic_name__ : Dict = tokenizer.all_special_tokens.index(_snake_case )
__magic_name__ : str = tokenizer.all_special_ids[idx]
logger.info(F'''Special tokens {special_tok_ids}''' )
__magic_name__ : Optional[int] = special_tok_ids
__magic_name__ : Optional[int] = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(F'''Loading data from {args.data_file}''' )
with open(args.data_file , "rb" ) as fp:
__magic_name__ : List[Any] = pickle.load(_snake_case )
if args.mlm:
logger.info(F'''Loading token counts from {args.token_counts} (already pre-computed)''' )
with open(args.token_counts , "rb" ) as fp:
__magic_name__ : List[str] = pickle.load(_snake_case )
__magic_name__ : int = np.maximum(_snake_case , 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
__magic_name__ : Optional[int] = 0.0 # do not predict special tokens
__magic_name__ : List[Any] = torch.from_numpy(_snake_case )
else:
__magic_name__ : Union[str, Any] = None
__magic_name__ : List[Any] = LmSeqsDataset(params=_snake_case , data=_snake_case )
logger.info("Data loader created." )
# STUDENT #
logger.info(F'''Loading student config from {args.student_config}''' )
__magic_name__ : Optional[int] = student_config_class.from_pretrained(args.student_config )
__magic_name__ : int = True
if args.student_pretrained_weights is not None:
logger.info(F'''Loading pretrained weights from {args.student_pretrained_weights}''' )
__magic_name__ : List[Any] = student_model_class.from_pretrained(args.student_pretrained_weights , config=_snake_case )
else:
__magic_name__ : Tuple = student_model_class(_snake_case )
if args.n_gpu > 0:
student.to(F'''cuda:{args.local_rank}''' )
logger.info("Student loaded." )
# TEACHER #
__magic_name__ : Tuple = teacher_model_class.from_pretrained(args.teacher_name , output_hidden_states=_snake_case )
if args.n_gpu > 0:
teacher.to(F'''cuda:{args.local_rank}''' )
logger.info(F'''Teacher loaded from {args.teacher_name}.''' )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_snake_case , _snake_case )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_snake_case , _snake_case )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
__magic_name__ : List[str] = Distiller(
params=_snake_case , dataset=_snake_case , token_probs=_snake_case , student=_snake_case , teacher=_snake_case )
distiller.train()
logger.info("Let's go get some drinks." )
if __name__ == "__main__":
main()
| 281 |
from itertools import zip_longest
import requests
from bsa import BeautifulSoup
from pandas import DataFrame
def lowerCAmelCase_ ( _snake_case : str = "laptop" ) -> DataFrame:
'''simple docstring'''
__magic_name__ : Tuple = F'''https://www.amazon.in/laptop/s?k={product}'''
__magic_name__ : Dict = {
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36\n (KHTML, like Gecko)Chrome/44.0.2403.157 Safari/537.36",
"Accept-Language": "en-US, en;q=0.5",
}
__magic_name__ : Tuple = BeautifulSoup(requests.get(_snake_case , headers=_snake_case ).text )
# Initialize a Pandas dataframe with the column titles
__magic_name__ : int = DataFrame(
columns=[
"Product Title",
"Product Link",
"Current Price of the product",
"Product Rating",
"MRP of the product",
"Discount",
] )
# Loop through each entry and store them in the dataframe
for item, _ in zip_longest(
soup.find_all(
"div" , attrs={"class": "s-result-item", "data-component-type": "s-search-result"} , ) , soup.find_all("div" , attrs={"class": "a-row a-size-base a-color-base"} ) , ):
try:
__magic_name__ : Dict = item.ha.text
__magic_name__ : Optional[int] = "https://www.amazon.in/" + item.ha.a["href"]
__magic_name__ : Optional[Any] = item.find("span" , attrs={"class": "a-offscreen"} ).text
try:
__magic_name__ : Union[str, Any] = item.find("span" , attrs={"class": "a-icon-alt"} ).text
except AttributeError:
__magic_name__ : Dict = "Not available"
try:
__magic_name__ : Optional[int] = (
"₹"
+ item.find(
"span" , attrs={"class": "a-price a-text-price"} ).text.split("₹" )[1]
)
except AttributeError:
__magic_name__ : List[str] = ""
try:
__magic_name__ : int = float(
(
(
float(product_mrp.strip("₹" ).replace("," , "" ) )
- float(product_price.strip("₹" ).replace("," , "" ) )
)
/ float(product_mrp.strip("₹" ).replace("," , "" ) )
)
* 100 )
except ValueError:
__magic_name__ : str = float("nan" )
except AttributeError:
pass
__magic_name__ : Optional[int] = [
product_title,
product_link,
product_price,
product_rating,
product_mrp,
discount,
]
__magic_name__ : Optional[Any] = " "
__magic_name__ : str = " "
data_frame.index += 1
return data_frame
if __name__ == "__main__":
snake_case : Any = "headphones"
get_amazon_product_data(product).to_csv(F"Amazon Product Data for {product}.csv")
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : int , _snake_case : list[int] , _snake_case : int ) -> int:
'''simple docstring'''
def count_of_possible_combinations(_snake_case : int ) -> int:
if target < 0:
return 0
if target == 0:
return 1
return sum(count_of_possible_combinations(target - item ) for item in array )
return count_of_possible_combinations(_snake_case )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : list[int] , _snake_case : int ) -> int:
'''simple docstring'''
def count_of_possible_combinations_with_dp_array(
_snake_case : int , _snake_case : list[int] ) -> int:
if target < 0:
return 0
if target == 0:
return 1
if dp_array[target] != -1:
return dp_array[target]
__magic_name__ : Union[str, Any] = sum(
count_of_possible_combinations_with_dp_array(target - item , _snake_case )
for item in array )
__magic_name__ : List[str] = answer
return answer
__magic_name__ : List[str] = [-1] * (target + 1)
return count_of_possible_combinations_with_dp_array(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : list[int] , _snake_case : int ) -> int:
'''simple docstring'''
__magic_name__ : str = [0] * (target + 1)
__magic_name__ : Any = 1
for i in range(1 , target + 1 ):
for j in range(_snake_case ):
if i - array[j] >= 0:
dp_array[i] += dp_array[i - array[j]]
return dp_array[target]
if __name__ == "__main__":
import doctest
doctest.testmod()
snake_case : List[str] = 3
snake_case : List[Any] = 5
snake_case : Any = [1, 2, 5]
print(combination_sum_iv(n, array, target))
| 281 |
from __future__ import annotations
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Optional[Any] = data
__magic_name__ : Node | None = None
__magic_name__ : Node | None = None
def lowerCAmelCase_ ( _snake_case : Node | None ) -> None: # In Order traversal of the tree
'''simple docstring'''
if tree:
display(tree.left )
print(tree.data )
display(tree.right )
def lowerCAmelCase_ ( _snake_case : Node | None ) -> int:
'''simple docstring'''
return 1 + max(depth_of_tree(tree.left ) , depth_of_tree(tree.right ) ) if tree else 0
def lowerCAmelCase_ ( _snake_case : Node ) -> bool:
'''simple docstring'''
if not tree:
return True
if tree.left and tree.right:
return is_full_binary_tree(tree.left ) and is_full_binary_tree(tree.right )
else:
return not tree.left and not tree.right
def lowerCAmelCase_ ( ) -> None: # Main function for testing.
'''simple docstring'''
__magic_name__ : int = Node(1 )
__magic_name__ : Union[str, Any] = Node(2 )
__magic_name__ : Tuple = Node(3 )
__magic_name__ : Optional[Any] = Node(4 )
__magic_name__ : Union[str, Any] = Node(5 )
__magic_name__ : Any = Node(6 )
__magic_name__ : int = Node(7 )
__magic_name__ : List[str] = Node(8 )
__magic_name__ : Union[str, Any] = Node(9 )
print(is_full_binary_tree(_snake_case ) )
print(depth_of_tree(_snake_case ) )
print("Tree is: " )
display(_snake_case )
if __name__ == "__main__":
main()
| 281 | 1 |
import warnings
from ..trainer import Trainer
from ..utils import logging
snake_case : Union[str, Any] = logging.get_logger(__name__)
class _snake_case ( snake_case ):
def __init__( self , _a=None , **_a ):
warnings.warn(
"`SageMakerTrainer` is deprecated and will be removed in v5 of Transformers. You can use `Trainer` "
"instead." , _a , )
super().__init__(args=_a , **_a )
| 281 |
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> bool:
'''simple docstring'''
__magic_name__ : Union[str, Any] = len(_snake_case ) + 1
__magic_name__ : List[str] = len(_snake_case ) + 1
# dp is a 2d matrix where dp[i][j] denotes whether prefix string of
# length i of input_string matches with prefix string of length j of
# given pattern.
# "dp" stands for dynamic programming.
__magic_name__ : str = [[0 for i in range(_snake_case )] for j in range(_snake_case )]
# since string of zero length match pattern of zero length
__magic_name__ : Optional[int] = 1
# since pattern of zero length will never match with string of non-zero length
for i in range(1 , _snake_case ):
__magic_name__ : Optional[int] = 0
# since string of zero length will match with pattern where there
# is at least one * alternatively
for j in range(1 , _snake_case ):
__magic_name__ : Union[str, Any] = dp[0][j - 2] if pattern[j - 1] == "*" else 0
# now using bottom-up approach to find for all remaining lengths
for i in range(1 , _snake_case ):
for j in range(1 , _snake_case ):
if input_string[i - 1] == pattern[j - 1] or pattern[j - 1] == ".":
__magic_name__ : Optional[int] = dp[i - 1][j - 1]
elif pattern[j - 1] == "*":
if dp[i][j - 2] == 1:
__magic_name__ : Optional[Any] = 1
elif pattern[j - 2] in (input_string[i - 1], "."):
__magic_name__ : List[Any] = dp[i - 1][j]
else:
__magic_name__ : Union[str, Any] = 0
else:
__magic_name__ : Dict = 0
return bool(dp[-1][-1] )
if __name__ == "__main__":
import doctest
doctest.testmod()
# inputing the strings
# input_string = input("input a string :")
# pattern = input("input a pattern :")
snake_case : Optional[Any] = "aab"
snake_case : List[str] = "c*a*b"
# using function to check whether given string matches the given pattern
if match_pattern(input_string, pattern):
print(F"{input_string} matches the given pattern {pattern}")
else:
print(F"{input_string} does not match with the given pattern {pattern}")
| 281 | 1 |
import unittest
import numpy as np
from transformers import RobertaPreLayerNormConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
import jax.numpy as jnp
from transformers.models.roberta_prelayernorm.modeling_flax_roberta_prelayernorm import (
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormModel,
)
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=True , _a=True , _a=True , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=16 , _a=2 , _a=0.02 , _a=4 , ):
__magic_name__ : List[Any] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Dict = seq_length
__magic_name__ : Union[str, Any] = is_training
__magic_name__ : Optional[Any] = use_attention_mask
__magic_name__ : Optional[Any] = use_token_type_ids
__magic_name__ : int = use_labels
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : Any = intermediate_size
__magic_name__ : List[Any] = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[Any] = max_position_embeddings
__magic_name__ : Tuple = type_vocab_size
__magic_name__ : List[str] = type_sequence_label_size
__magic_name__ : Dict = initializer_range
__magic_name__ : List[Any] = num_choices
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = None
if self.use_attention_mask:
__magic_name__ : List[str] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : str = None
if self.use_token_type_ids:
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__magic_name__ : List[str] = RobertaPreLayerNormConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=_a , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[Any] = config_and_inputs
__magic_name__ : List[str] = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": attention_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : Tuple = True
__magic_name__ : int = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
token_type_ids,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
# Copied from tests.models.roberta.test_modelling_flax_roberta.FlaxRobertaPreLayerNormModelTest with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta-base->andreasmadsen/efficient_mlm_m0.40
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = True
UpperCamelCase__ = (
(
FlaxRobertaPreLayerNormModel,
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
)
if is_flax_available()
else ()
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxRobertaPreLayerNormModelTester(self )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Optional[Any] = model_class_name.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(_a )
@require_flax
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxRobertaPreLayerNormForMaskedLM.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Union[str, Any] = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : List[str] = model(_a )[0]
__magic_name__ : str = [1, 11, 50_265]
self.assertEqual(list(output.shape ) , _a )
# compare the actual values for a slice.
__magic_name__ : List[str] = np.array(
[[[40.48_80, 18.01_99, -5.23_67], [-1.88_77, -4.08_85, 10.70_85], [-2.26_13, -5.61_10, 7.26_65]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = FlaxRobertaPreLayerNormModel.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Tuple = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : Tuple = model(_a )[0]
# compare the actual values for a slice.
__magic_name__ : Dict = np.array(
[[[0.02_08, -0.03_56, 0.02_37], [-0.15_69, -0.04_11, -0.26_26], [0.18_79, 0.01_25, -0.00_89]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
| 281 |
import hashlib
import unittest
from typing import Dict
import numpy as np
from transformers import (
MODEL_FOR_MASK_GENERATION_MAPPING,
TF_MODEL_FOR_MASK_GENERATION_MAPPING,
is_vision_available,
pipeline,
)
from transformers.pipelines import MaskGenerationPipeline
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
)
if is_vision_available():
from PIL import Image
else:
class _snake_case :
@staticmethod
def SCREAMING_SNAKE_CASE ( *_a , **_a ):
pass
def lowerCAmelCase_ ( _snake_case : Image ) -> str:
'''simple docstring'''
__magic_name__ : Optional[int] = hashlib.mda(image.tobytes() )
return m.hexdigest()[:10]
def lowerCAmelCase_ ( _snake_case : Image ) -> Dict:
'''simple docstring'''
__magic_name__ : List[Any] = np.array(_snake_case )
__magic_name__ : Optional[int] = npimg.shape
return {"hash": hashimage(_snake_case ), "shape": shape}
@is_pipeline_test
@require_vision
@require_torch
class _snake_case ( unittest.TestCase ):
UpperCamelCase__ = dict(
(list(MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if MODEL_FOR_MASK_GENERATION_MAPPING else []) )
UpperCamelCase__ = dict(
(list(TF_MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if TF_MODEL_FOR_MASK_GENERATION_MAPPING else []) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Dict = MaskGenerationPipeline(model=_a , image_processor=_a )
return image_segmenter, [
"./tests/fixtures/tests_samples/COCO/000000039769.png",
"./tests/fixtures/tests_samples/COCO/000000039769.png",
]
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
pass
@require_tf
@unittest.skip("Image segmentation not implemented in TF" )
def SCREAMING_SNAKE_CASE ( self ):
pass
@slow
@require_torch
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = pipeline("mask-generation" , model="facebook/sam-vit-huge" )
__magic_name__ : str = image_segmenter("http://images.cocodataset.org/val2017/000000039769.jpg" , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Dict = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
# fmt: off
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.0_21},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
{"mask": {"hash": "e2d0b7a0b7", "shape": (480, 640)}, "scores": 0.99_67},
{"mask": {"hash": "453c7844bd", "shape": (480, 640)}, "scores": 0.9_93},
{"mask": {"hash": "3d44f2926d", "shape": (480, 640)}, "scores": 0.99_09},
{"mask": {"hash": "64033ddc3f", "shape": (480, 640)}, "scores": 0.98_79},
{"mask": {"hash": "801064ff79", "shape": (480, 640)}, "scores": 0.98_34},
{"mask": {"hash": "6172f276ef", "shape": (480, 640)}, "scores": 0.97_16},
{"mask": {"hash": "b49e60e084", "shape": (480, 640)}, "scores": 0.96_12},
{"mask": {"hash": "a811e775fd", "shape": (480, 640)}, "scores": 0.95_99},
{"mask": {"hash": "a6a8ebcf4b", "shape": (480, 640)}, "scores": 0.95_52},
{"mask": {"hash": "9d8257e080", "shape": (480, 640)}, "scores": 0.95_32},
{"mask": {"hash": "32de6454a8", "shape": (480, 640)}, "scores": 0.95_16},
{"mask": {"hash": "af3d4af2c8", "shape": (480, 640)}, "scores": 0.94_99},
{"mask": {"hash": "3c6db475fb", "shape": (480, 640)}, "scores": 0.94_83},
{"mask": {"hash": "c290813fb9", "shape": (480, 640)}, "scores": 0.94_64},
{"mask": {"hash": "b6f0b8f606", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "92ce16bfdf", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "c749b25868", "shape": (480, 640)}, "scores": 0.94_08},
{"mask": {"hash": "efb6cab859", "shape": (480, 640)}, "scores": 0.93_35},
{"mask": {"hash": "1ff2eafb30", "shape": (480, 640)}, "scores": 0.93_26},
{"mask": {"hash": "788b798e24", "shape": (480, 640)}, "scores": 0.92_62},
{"mask": {"hash": "abea804f0e", "shape": (480, 640)}, "scores": 0.89_99},
{"mask": {"hash": "7b9e8ddb73", "shape": (480, 640)}, "scores": 0.89_86},
{"mask": {"hash": "cd24047c8a", "shape": (480, 640)}, "scores": 0.89_84},
{"mask": {"hash": "6943e6bcbd", "shape": (480, 640)}, "scores": 0.88_73},
{"mask": {"hash": "b5f47c9191", "shape": (480, 640)}, "scores": 0.88_71}
] , )
# fmt: on
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = "facebook/sam-vit-huge"
__magic_name__ : str = pipeline("mask-generation" , model=_a )
__magic_name__ : Tuple = image_segmenter(
"http://images.cocodataset.org/val2017/000000039769.jpg" , pred_iou_thresh=1 , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Any = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.02_10},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
] , )
| 281 | 1 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
snake_case : Tuple = {"configuration_ibert": ["IBERT_PRETRAINED_CONFIG_ARCHIVE_MAP", "IBertConfig", "IBertOnnxConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : Tuple = [
"IBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"IBertForMaskedLM",
"IBertForMultipleChoice",
"IBertForQuestionAnswering",
"IBertForSequenceClassification",
"IBertForTokenClassification",
"IBertModel",
"IBertPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_ibert import IBERT_PRETRAINED_CONFIG_ARCHIVE_MAP, IBertConfig, IBertOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_ibert import (
IBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
IBertForMaskedLM,
IBertForMultipleChoice,
IBertForQuestionAnswering,
IBertForSequenceClassification,
IBertForTokenClassification,
IBertModel,
IBertPreTrainedModel,
)
else:
import sys
snake_case : List[str] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 281 |
import absl # noqa: F401 # Here to have a nice missing dependency error message early on
import nltk # noqa: F401 # Here to have a nice missing dependency error message early on
import numpy # noqa: F401 # Here to have a nice missing dependency error message early on
import six # noqa: F401 # Here to have a nice missing dependency error message early on
from rouge_score import rouge_scorer, scoring
import datasets
snake_case : List[Any] = "\\n@inproceedings{lin-2004-rouge,\n title = \"{ROUGE}: A Package for Automatic Evaluation of Summaries\",\n author = \"Lin, Chin-Yew\",\n booktitle = \"Text Summarization Branches Out\",\n month = jul,\n year = \"2004\",\n address = \"Barcelona, Spain\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W04-1013\",\n pages = \"74--81\",\n}\n"
snake_case : Any = "\\nROUGE, or Recall-Oriented Understudy for Gisting Evaluation, is a set of metrics and a software package used for\nevaluating automatic summarization and machine translation software in natural language processing.\nThe metrics compare an automatically produced summary or translation against a reference or a set of references (human-produced) summary or translation.\n\nNote that ROUGE is case insensitive, meaning that upper case letters are treated the same way as lower case letters.\n\nThis metrics is a wrapper around Google Research reimplementation of ROUGE:\nhttps://github.com/google-research/google-research/tree/master/rouge\n"
snake_case : str = "\nCalculates average rouge scores for a list of hypotheses and references\nArgs:\n predictions: list of predictions to score. Each prediction\n should be a string with tokens separated by spaces.\n references: list of reference for each prediction. Each\n reference should be a string with tokens separated by spaces.\n rouge_types: A list of rouge types to calculate.\n Valid names:\n `\"rouge{n}\"` (e.g. `\"rouge1\"`, `\"rouge2\"`) where: {n} is the n-gram based scoring,\n `\"rougeL\"`: Longest common subsequence based scoring.\n `\"rougeLSum\"`: rougeLsum splits text using `\"\n\"`.\n See details in https://github.com/huggingface/datasets/issues/617\n use_stemmer: Bool indicating whether Porter stemmer should be used to strip word suffixes.\n use_aggregator: Return aggregates if this is set to True\nReturns:\n rouge1: rouge_1 (precision, recall, f1),\n rouge2: rouge_2 (precision, recall, f1),\n rougeL: rouge_l (precision, recall, f1),\n rougeLsum: rouge_lsum (precision, recall, f1)\nExamples:\n\n >>> rouge = datasets.load_metric('rouge')\n >>> predictions = [\"hello there\", \"general kenobi\"]\n >>> references = [\"hello there\", \"general kenobi\"]\n >>> results = rouge.compute(predictions=predictions, references=references)\n >>> print(list(results.keys()))\n ['rouge1', 'rouge2', 'rougeL', 'rougeLsum']\n >>> print(results[\"rouge1\"])\n AggregateScore(low=Score(precision=1.0, recall=1.0, fmeasure=1.0), mid=Score(precision=1.0, recall=1.0, fmeasure=1.0), high=Score(precision=1.0, recall=1.0, fmeasure=1.0))\n >>> print(results[\"rouge1\"].mid.fmeasure)\n 1.0\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class _snake_case ( datasets.Metric ):
def SCREAMING_SNAKE_CASE ( self ):
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , 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/google-research/tree/master/rouge"] , reference_urls=[
"https://en.wikipedia.org/wiki/ROUGE_(metric)",
"https://github.com/google-research/google-research/tree/master/rouge",
] , )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a=None , _a=True , _a=False ):
if rouge_types is None:
__magic_name__ : str = ["rouge1", "rouge2", "rougeL", "rougeLsum"]
__magic_name__ : List[str] = rouge_scorer.RougeScorer(rouge_types=_a , use_stemmer=_a )
if use_aggregator:
__magic_name__ : Dict = scoring.BootstrapAggregator()
else:
__magic_name__ : str = []
for ref, pred in zip(_a , _a ):
__magic_name__ : Union[str, Any] = scorer.score(_a , _a )
if use_aggregator:
aggregator.add_scores(_a )
else:
scores.append(_a )
if use_aggregator:
__magic_name__ : Any = aggregator.aggregate()
else:
__magic_name__ : List[Any] = {}
for key in scores[0]:
__magic_name__ : str = [score[key] for score in scores]
return result
| 281 | 1 |
import logging
import os
from typing import List, TextIO, Union
from conllu import parse_incr
from utils_ner import InputExample, Split, TokenClassificationTask
snake_case : List[str] = logging.getLogger(__name__)
class _snake_case ( snake_case ):
def __init__( self , _a=-1 ):
# in NER datasets, the last column is usually reserved for NER label
__magic_name__ : Dict = label_idx
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
if isinstance(_a , _a ):
__magic_name__ : List[str] = mode.value
__magic_name__ : str = os.path.join(_a , f'''{mode}.txt''' )
__magic_name__ : List[Any] = 1
__magic_name__ : Optional[Any] = []
with open(_a , encoding="utf-8" ) as f:
__magic_name__ : List[str] = []
__magic_name__ : Optional[Any] = []
for line in f:
if line.startswith("-DOCSTART-" ) or line == "" or line == "\n":
if words:
examples.append(InputExample(guid=f'''{mode}-{guid_index}''' , words=_a , labels=_a ) )
guid_index += 1
__magic_name__ : List[Any] = []
__magic_name__ : Dict = []
else:
__magic_name__ : Optional[Any] = line.split(" " )
words.append(splits[0] )
if len(_a ) > 1:
labels.append(splits[self.label_idx].replace("\n" , "" ) )
else:
# Examples could have no label for mode = "test"
labels.append("O" )
if words:
examples.append(InputExample(guid=f'''{mode}-{guid_index}''' , words=_a , labels=_a ) )
return examples
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Tuple = 0
for line in test_input_reader:
if line.startswith("-DOCSTART-" ) or line == "" or line == "\n":
writer.write(_a )
if not preds_list[example_id]:
example_id += 1
elif preds_list[example_id]:
__magic_name__ : List[Any] = line.split()[0] + " " + preds_list[example_id].pop(0 ) + "\n"
writer.write(_a )
else:
logger.warning("Maximum sequence length exceeded: No prediction for '%s'." , line.split()[0] )
def SCREAMING_SNAKE_CASE ( self , _a ):
if path:
with open(_a , "r" ) as f:
__magic_name__ : Tuple = f.read().splitlines()
if "O" not in labels:
__magic_name__ : Union[str, Any] = ["O"] + labels
return labels
else:
return ["O", "B-MISC", "I-MISC", "B-PER", "I-PER", "B-ORG", "I-ORG", "B-LOC", "I-LOC"]
class _snake_case ( snake_case ):
def __init__( self ):
# in CONLL2003 dataset chunk column is second-to-last
super().__init__(label_idx=-2 )
def SCREAMING_SNAKE_CASE ( self , _a ):
if path:
with open(_a , "r" ) as f:
__magic_name__ : Tuple = f.read().splitlines()
if "O" not in labels:
__magic_name__ : Union[str, Any] = ["O"] + labels
return labels
else:
return [
"O",
"B-ADVP",
"B-INTJ",
"B-LST",
"B-PRT",
"B-NP",
"B-SBAR",
"B-VP",
"B-ADJP",
"B-CONJP",
"B-PP",
"I-ADVP",
"I-INTJ",
"I-LST",
"I-PRT",
"I-NP",
"I-SBAR",
"I-VP",
"I-ADJP",
"I-CONJP",
"I-PP",
]
class _snake_case ( snake_case ):
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
if isinstance(_a , _a ):
__magic_name__ : List[Any] = mode.value
__magic_name__ : Union[str, Any] = os.path.join(_a , f'''{mode}.txt''' )
__magic_name__ : Optional[int] = 1
__magic_name__ : List[str] = []
with open(_a , encoding="utf-8" ) as f:
for sentence in parse_incr(_a ):
__magic_name__ : Union[str, Any] = []
__magic_name__ : Union[str, Any] = []
for token in sentence:
words.append(token["form"] )
labels.append(token["upos"] )
assert len(_a ) == len(_a )
if words:
examples.append(InputExample(guid=f'''{mode}-{guid_index}''' , words=_a , labels=_a ) )
guid_index += 1
return examples
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Any = 0
for sentence in parse_incr(_a ):
__magic_name__ : List[Any] = preds_list[example_id]
__magic_name__ : Optional[Any] = ""
for token in sentence:
out += f'''{token["form"]} ({token["upos"]}|{s_p.pop(0 )}) '''
out += "\n"
writer.write(_a )
example_id += 1
def SCREAMING_SNAKE_CASE ( self , _a ):
if path:
with open(_a , "r" ) as f:
return f.read().splitlines()
else:
return [
"ADJ",
"ADP",
"ADV",
"AUX",
"CCONJ",
"DET",
"INTJ",
"NOUN",
"NUM",
"PART",
"PRON",
"PROPN",
"PUNCT",
"SCONJ",
"SYM",
"VERB",
"X",
]
| 281 |
snake_case : Optional[int] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"
def lowerCAmelCase_ ( _snake_case : bytes ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ):
__magic_name__ : Tuple = F'''a bytes-like object is required, not \'{data.__class__.__name__}\''''
raise TypeError(_snake_case )
__magic_name__ : Optional[int] = "".join(bin(_snake_case )[2:].zfill(8 ) for byte in data )
__magic_name__ : List[Any] = len(_snake_case ) % 6 != 0
if padding_needed:
# The padding that will be added later
__magic_name__ : List[str] = B"=" * ((6 - len(_snake_case ) % 6) // 2)
# Append binary_stream with arbitrary binary digits (0's by default) to make its
# length a multiple of 6.
binary_stream += "0" * (6 - len(_snake_case ) % 6)
else:
__magic_name__ : List[str] = B""
# Encode every 6 binary digits to their corresponding Base64 character
return (
"".join(
B64_CHARSET[int(binary_stream[index : index + 6] , 2 )]
for index in range(0 , len(_snake_case ) , 6 ) ).encode()
+ padding
)
def lowerCAmelCase_ ( _snake_case : str ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ) and not isinstance(_snake_case , _snake_case ):
__magic_name__ : List[str] = (
"argument should be a bytes-like object or ASCII string, "
F'''not \'{encoded_data.__class__.__name__}\''''
)
raise TypeError(_snake_case )
# In case encoded_data is a bytes-like object, make sure it contains only
# ASCII characters so we convert it to a string object
if isinstance(_snake_case , _snake_case ):
try:
__magic_name__ : List[Any] = encoded_data.decode("utf-8" )
except UnicodeDecodeError:
raise ValueError("base64 encoded data should only contain ASCII characters" )
__magic_name__ : List[str] = encoded_data.count("=" )
# Check if the encoded string contains non base64 characters
if padding:
assert all(
char in B64_CHARSET for char in encoded_data[:-padding] ), "Invalid base64 character(s) found."
else:
assert all(
char in B64_CHARSET for char in encoded_data ), "Invalid base64 character(s) found."
# Check the padding
assert len(_snake_case ) % 4 == 0 and padding < 3, "Incorrect padding"
if padding:
# Remove padding if there is one
__magic_name__ : Optional[int] = encoded_data[:-padding]
__magic_name__ : Dict = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )[: -padding * 2]
else:
__magic_name__ : Union[str, Any] = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )
__magic_name__ : List[Any] = [
int(binary_stream[index : index + 8] , 2 )
for index in range(0 , len(_snake_case ) , 8 )
]
return bytes(_snake_case )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
class _snake_case :
def __init__( self , _a , _a=None , _a=None ):
__magic_name__ : Dict = data
__magic_name__ : List[str] = previous
__magic_name__ : Any = next_node
def __str__( self ):
return f'''{self.data}'''
def SCREAMING_SNAKE_CASE ( self ):
return self.data
def SCREAMING_SNAKE_CASE ( self ):
return self.next
def SCREAMING_SNAKE_CASE ( self ):
return self.previous
class _snake_case :
def __init__( self , _a ):
__magic_name__ : List[Any] = head
def __iter__( self ):
return self
def SCREAMING_SNAKE_CASE ( self ):
if not self.current:
raise StopIteration
else:
__magic_name__ : List[Any] = self.current.get_data()
__magic_name__ : int = self.current.get_next()
return value
class _snake_case :
def __init__( self ):
__magic_name__ : int = None # First node in list
__magic_name__ : Union[str, Any] = None # Last node in list
def __str__( self ):
__magic_name__ : Optional[int] = self.head
__magic_name__ : str = []
while current is not None:
nodes.append(current.get_data() )
__magic_name__ : List[Any] = current.get_next()
return " ".join(str(_a ) for node in nodes )
def __contains__( self , _a ):
__magic_name__ : List[str] = self.head
while current:
if current.get_data() == value:
return True
__magic_name__ : str = current.get_next()
return False
def __iter__( self ):
return LinkedListIterator(self.head )
def SCREAMING_SNAKE_CASE ( self ):
if self.head:
return self.head.get_data()
return None
def SCREAMING_SNAKE_CASE ( self ):
if self.tail:
return self.tail.get_data()
return None
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.head is None:
__magic_name__ : str = node
__magic_name__ : int = node
else:
self.insert_before_node(self.head , _a )
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.head is None:
self.set_head(_a )
else:
self.insert_after_node(self.tail , _a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[str] = Node(_a )
if self.head is None:
self.set_head(_a )
else:
self.set_tail(_a )
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Tuple = node
__magic_name__ : str = node.previous
if node.get_previous() is None:
__magic_name__ : Dict = node_to_insert
else:
__magic_name__ : str = node_to_insert
__magic_name__ : str = node_to_insert
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : List[Any] = node
__magic_name__ : Tuple = node.next
if node.get_next() is None:
__magic_name__ : Optional[int] = node_to_insert
else:
__magic_name__ : List[Any] = node_to_insert
__magic_name__ : int = node_to_insert
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : str = 1
__magic_name__ : str = Node(_a )
__magic_name__ : str = self.head
while node:
if current_position == position:
self.insert_before_node(_a , _a )
return
current_position += 1
__magic_name__ : Optional[int] = node.next
self.insert_after_node(self.tail , _a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Union[str, Any] = self.head
while node:
if node.get_data() == item:
return node
__magic_name__ : Any = node.get_next()
raise Exception("Node not found" )
def SCREAMING_SNAKE_CASE ( self , _a ):
if (node := self.get_node(_a )) is not None:
if node == self.head:
__magic_name__ : List[str] = self.head.get_next()
if node == self.tail:
__magic_name__ : int = self.tail.get_previous()
self.remove_node_pointers(_a )
@staticmethod
def SCREAMING_SNAKE_CASE ( _a ):
if node.get_next():
__magic_name__ : Dict = node.previous
if node.get_previous():
__magic_name__ : int = node.next
__magic_name__ : int = None
__magic_name__ : Tuple = None
def SCREAMING_SNAKE_CASE ( self ):
return self.head is None
def lowerCAmelCase_ ( ) -> None:
'''simple docstring'''
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
import unittest
import numpy as np
from transformers import RobertaPreLayerNormConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
import jax.numpy as jnp
from transformers.models.roberta_prelayernorm.modeling_flax_roberta_prelayernorm import (
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormModel,
)
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=True , _a=True , _a=True , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=16 , _a=2 , _a=0.02 , _a=4 , ):
__magic_name__ : List[Any] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Dict = seq_length
__magic_name__ : Union[str, Any] = is_training
__magic_name__ : Optional[Any] = use_attention_mask
__magic_name__ : Optional[Any] = use_token_type_ids
__magic_name__ : int = use_labels
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : Any = intermediate_size
__magic_name__ : List[Any] = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[Any] = max_position_embeddings
__magic_name__ : Tuple = type_vocab_size
__magic_name__ : List[str] = type_sequence_label_size
__magic_name__ : Dict = initializer_range
__magic_name__ : List[Any] = num_choices
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = None
if self.use_attention_mask:
__magic_name__ : List[str] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : str = None
if self.use_token_type_ids:
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__magic_name__ : List[str] = RobertaPreLayerNormConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=_a , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[Any] = config_and_inputs
__magic_name__ : List[str] = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": attention_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : Tuple = True
__magic_name__ : int = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
token_type_ids,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
# Copied from tests.models.roberta.test_modelling_flax_roberta.FlaxRobertaPreLayerNormModelTest with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta-base->andreasmadsen/efficient_mlm_m0.40
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = True
UpperCamelCase__ = (
(
FlaxRobertaPreLayerNormModel,
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
)
if is_flax_available()
else ()
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxRobertaPreLayerNormModelTester(self )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Optional[Any] = model_class_name.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(_a )
@require_flax
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxRobertaPreLayerNormForMaskedLM.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Union[str, Any] = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : List[str] = model(_a )[0]
__magic_name__ : str = [1, 11, 50_265]
self.assertEqual(list(output.shape ) , _a )
# compare the actual values for a slice.
__magic_name__ : List[str] = np.array(
[[[40.48_80, 18.01_99, -5.23_67], [-1.88_77, -4.08_85, 10.70_85], [-2.26_13, -5.61_10, 7.26_65]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = FlaxRobertaPreLayerNormModel.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Tuple = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : Tuple = model(_a )[0]
# compare the actual values for a slice.
__magic_name__ : Dict = np.array(
[[[0.02_08, -0.03_56, 0.02_37], [-0.15_69, -0.04_11, -0.26_26], [0.18_79, 0.01_25, -0.00_89]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
| 281 | 1 |
import argparse
import json
import os
import evaluate
import torch
from datasets import load_dataset
from torch.optim import AdamW
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer, get_linear_schedule_with_warmup, set_seed
from accelerate import Accelerator, DistributedType
from accelerate.utils.deepspeed import DummyOptim, DummyScheduler
snake_case : Any = 16
snake_case : Optional[Any] = 32
def lowerCAmelCase_ ( _snake_case : Accelerator , _snake_case : int = 16 , _snake_case : str = "bert-base-cased" ) -> str:
'''simple docstring'''
__magic_name__ : Optional[int] = AutoTokenizer.from_pretrained(_snake_case )
__magic_name__ : Optional[int] = load_dataset("glue" , "mrpc" )
def tokenize_function(_snake_case : Any ):
# max_length=None => use the model max length (it's actually the default)
__magic_name__ : List[Any] = tokenizer(examples["sentence1"] , examples["sentence2"] , truncation=_snake_case , max_length=_snake_case )
return outputs
# Apply the method we just defined to all the examples in all the splits of the dataset
__magic_name__ : Optional[int] = datasets.map(
_snake_case , batched=_snake_case , remove_columns=["idx", "sentence1", "sentence2"] , load_from_cache_file=_snake_case )
# We also rename the 'label' column to 'labels' which is the expected name for labels by the models of the
# transformers library
__magic_name__ : Dict = tokenized_datasets.rename_column("label" , "labels" )
def collate_fn(_snake_case : Dict ):
# On TPU it's best to pad everything to the same length or training will be very slow.
if accelerator.distributed_type == DistributedType.TPU:
return tokenizer.pad(_snake_case , padding="max_length" , max_length=128 , return_tensors="pt" )
return tokenizer.pad(_snake_case , padding="longest" , return_tensors="pt" )
# Instantiate dataloaders.
__magic_name__ : Optional[Any] = DataLoader(
tokenized_datasets["train"] , shuffle=_snake_case , collate_fn=_snake_case , batch_size=_snake_case )
__magic_name__ : List[Any] = DataLoader(
tokenized_datasets["validation"] , shuffle=_snake_case , collate_fn=_snake_case , batch_size=_snake_case )
return train_dataloader, eval_dataloader
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any , _snake_case : Optional[Any] , _snake_case : Union[str, Any] ) -> str:
'''simple docstring'''
model.eval()
__magic_name__ : Union[str, Any] = 0
for step, batch in enumerate(_snake_case ):
# We could avoid this line since we set the accelerator with `device_placement=True`.
batch.to(accelerator.device )
with torch.no_grad():
__magic_name__ : Any = model(**_snake_case )
__magic_name__ : Optional[int] = outputs.logits.argmax(dim=-1 )
# It is slightly faster to call this once, than multiple times
__magic_name__ , __magic_name__ : Dict = accelerator.gather(
(predictions, batch["labels"]) ) # If we are in a multiprocess environment, the last batch has duplicates
if accelerator.use_distributed:
if step == len(_snake_case ) - 1:
__magic_name__ : str = predictions[: len(eval_dataloader.dataset ) - samples_seen]
__magic_name__ : List[str] = references[: len(eval_dataloader.dataset ) - samples_seen]
else:
samples_seen += references.shape[0]
metric.add_batch(
predictions=_snake_case , references=_snake_case , )
__magic_name__ : List[str] = metric.compute()
return eval_metric["accuracy"]
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Tuple ) -> List[Any]:
'''simple docstring'''
__magic_name__ : List[Any] = Accelerator()
# Sample hyper-parameters for learning rate, batch size, seed and a few other HPs
__magic_name__ : Dict = config["lr"]
__magic_name__ : str = int(config["num_epochs"] )
__magic_name__ : Tuple = int(config["seed"] )
__magic_name__ : Tuple = int(config["batch_size"] )
__magic_name__ : Union[str, Any] = args.model_name_or_path
set_seed(_snake_case )
__magic_name__ , __magic_name__ : Any = get_dataloaders(_snake_case , _snake_case , _snake_case )
# Instantiate the model (we build the model here so that the seed also control new weights initialization)
__magic_name__ : Tuple = AutoModelForSequenceClassification.from_pretrained(_snake_case , return_dict=_snake_case )
# Instantiate optimizer
__magic_name__ : Tuple = (
AdamW
if accelerator.state.deepspeed_plugin is None
or "optimizer" not in accelerator.state.deepspeed_plugin.deepspeed_config
else DummyOptim
)
__magic_name__ : Tuple = optimizer_cls(params=model.parameters() , lr=_snake_case )
if accelerator.state.deepspeed_plugin is not None:
__magic_name__ : Optional[int] = accelerator.state.deepspeed_plugin.deepspeed_config[
"gradient_accumulation_steps"
]
else:
__magic_name__ : Union[str, Any] = 1
__magic_name__ : Union[str, Any] = (len(_snake_case ) * num_epochs) // gradient_accumulation_steps
# Instantiate scheduler
if (
accelerator.state.deepspeed_plugin is None
or "scheduler" not in accelerator.state.deepspeed_plugin.deepspeed_config
):
__magic_name__ : Tuple = get_linear_schedule_with_warmup(
optimizer=_snake_case , num_warmup_steps=0 , num_training_steps=_snake_case , )
else:
__magic_name__ : Optional[int] = DummyScheduler(_snake_case , total_num_steps=_snake_case , warmup_num_steps=0 )
# Prepare everything
# There is no specific order to remember, we just need to unpack the objects in the same order we gave them to the
# prepare method.
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Tuple = accelerator.prepare(
_snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
# We need to keep track of how many total steps we have iterated over
__magic_name__ : Union[str, Any] = 0
# We also need to keep track of the stating epoch so files are named properly
__magic_name__ : int = 0
__magic_name__ : int = evaluate.load("glue" , "mrpc" )
__magic_name__ : int = num_epochs
if args.partial_train_epoch is not None:
__magic_name__ : Tuple = args.partial_train_epoch
if args.resume_from_checkpoint:
accelerator.load_state(args.resume_from_checkpoint )
__magic_name__ : int = args.resume_from_checkpoint.split("epoch_" )[1]
__magic_name__ : List[Any] = ""
for char in epoch_string:
if char.isdigit():
state_epoch_num += char
else:
break
__magic_name__ : Dict = int(_snake_case ) + 1
__magic_name__ : Dict = evaluation_loop(_snake_case , _snake_case , _snake_case , _snake_case )
accelerator.print("resumed checkpoint performance:" , _snake_case )
accelerator.print("resumed checkpoint's scheduler's lr:" , lr_scheduler.get_lr()[0] )
accelerator.print("resumed optimizers's lr:" , optimizer.param_groups[0]["lr"] )
with open(os.path.join(args.output_dir , F'''state_{starting_epoch-1}.json''' ) , "r" ) as f:
__magic_name__ : str = json.load(_snake_case )
assert resumed_state["accuracy"] == accuracy, "Accuracy mismatch, loading from checkpoint failed"
assert (
resumed_state["lr"] == lr_scheduler.get_lr()[0]
), "Scheduler learning rate mismatch, loading from checkpoint failed"
assert (
resumed_state["optimizer_lr"] == optimizer.param_groups[0]["lr"]
), "Optimizer learning rate mismatch, loading from checkpoint failed"
assert resumed_state["epoch"] == starting_epoch - 1, "Epoch mismatch, loading from checkpoint failed"
return
# Now we train the model
__magic_name__ : Optional[Any] = {}
for epoch in range(_snake_case , _snake_case ):
model.train()
for step, batch in enumerate(_snake_case ):
__magic_name__ : List[str] = model(**_snake_case )
__magic_name__ : int = outputs.loss
__magic_name__ : Any = loss / gradient_accumulation_steps
accelerator.backward(_snake_case )
if step % gradient_accumulation_steps == 0:
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
overall_step += 1
__magic_name__ : int = F'''epoch_{epoch}'''
__magic_name__ : Optional[Any] = os.path.join(args.output_dir , _snake_case )
accelerator.save_state(_snake_case )
__magic_name__ : int = evaluation_loop(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ : Optional[Any] = accuracy
__magic_name__ : str = lr_scheduler.get_lr()[0]
__magic_name__ : List[Any] = optimizer.param_groups[0]["lr"]
__magic_name__ : Optional[Any] = epoch
__magic_name__ : Optional[int] = overall_step
accelerator.print(F'''epoch {epoch}:''' , _snake_case )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
with open(os.path.join(args.output_dir , F'''state_{epoch}.json''' ) , "w" ) as f:
json.dump(_snake_case , _snake_case )
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
__magic_name__ : str = argparse.ArgumentParser(description="Simple example of training script tracking peak GPU memory usage." )
parser.add_argument(
"--model_name_or_path" , type=_snake_case , default="bert-base-cased" , help="Path to pretrained model or model identifier from huggingface.co/models." , required=_snake_case , )
parser.add_argument(
"--output_dir" , type=_snake_case , default="." , help="Optional save directory where all checkpoint folders will be stored. Default is the current working directory." , )
parser.add_argument(
"--resume_from_checkpoint" , type=_snake_case , default=_snake_case , help="If the training should continue from a checkpoint folder." , )
parser.add_argument(
"--partial_train_epoch" , type=_snake_case , default=_snake_case , help="If passed, the training will stop after this number of epochs." , )
parser.add_argument(
"--num_epochs" , type=_snake_case , default=2 , help="Number of train epochs." , )
__magic_name__ : List[Any] = parser.parse_args()
__magic_name__ : Union[str, Any] = {"lr": 2E-5, "num_epochs": args.num_epochs, "seed": 42, "batch_size": 16}
training_function(_snake_case , _snake_case )
if __name__ == "__main__":
main()
| 281 |
def lowerCAmelCase_ ( _snake_case : list[list[int | float]] ) -> int:
'''simple docstring'''
__magic_name__ : Any = len(_snake_case )
__magic_name__ : Optional[Any] = len(matrix[0] )
__magic_name__ : Union[str, Any] = min(_snake_case , _snake_case )
for row in range(_snake_case ):
# Check if diagonal element is not zero
if matrix[row][row] != 0:
# Eliminate all the elements below the diagonal
for col in range(row + 1 , _snake_case ):
__magic_name__ : Optional[Any] = matrix[col][row] / matrix[row][row]
for i in range(_snake_case , _snake_case ):
matrix[col][i] -= multiplier * matrix[row][i]
else:
# Find a non-zero diagonal element to swap rows
__magic_name__ : str = True
for i in range(row + 1 , _snake_case ):
if matrix[i][row] != 0:
__magic_name__ , __magic_name__ : List[str] = matrix[i], matrix[row]
__magic_name__ : Union[str, Any] = False
break
if reduce:
rank -= 1
for i in range(_snake_case ):
__magic_name__ : Any = matrix[i][rank]
# Reduce the row pointer by one to stay on the same row
row -= 1
return rank
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow
if is_torch_available():
import torch
from transformers import XLMRobertaModel
@require_sentencepiece
@require_tokenizers
@require_torch
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = XLMRobertaModel.from_pretrained("xlm-roberta-base" )
__magic_name__ : str = torch.tensor([[0, 581, 10_269, 83, 99_942, 136, 60_742, 23, 70, 80_583, 18_276, 2]] )
# The dog is cute and lives in the garden house
__magic_name__ : Optional[Any] = torch.Size((1, 12, 768) ) # batch_size, sequence_length, embedding_vector_dim
__magic_name__ : Tuple = torch.tensor(
[[-0.01_01, 0.12_18, -0.08_03, 0.08_01, 0.13_27, 0.07_76, -0.12_15, 0.23_83, 0.33_38, 0.31_06, 0.03_00, 0.02_52]] )
# 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():
__magic_name__ : str = model(_a )["last_hidden_state"].detach()
self.assertEqual(output.shape , _a )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] , _a , atol=1e-3 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = XLMRobertaModel.from_pretrained("xlm-roberta-large" )
__magic_name__ : Optional[int] = torch.tensor([[0, 581, 10_269, 83, 99_942, 136, 60_742, 23, 70, 80_583, 18_276, 2]] )
# The dog is cute and lives in the garden house
__magic_name__ : List[str] = torch.Size((1, 12, 1_024) ) # batch_size, sequence_length, embedding_vector_dim
__magic_name__ : str = torch.tensor(
[[-0.06_99, -0.03_18, 0.07_05, -0.12_41, 0.09_99, -0.05_20, 0.10_04, -0.18_38, -0.47_04, 0.14_37, 0.08_21, 0.01_26]] )
# 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():
__magic_name__ : Dict = model(_a )["last_hidden_state"].detach()
self.assertEqual(output.shape , _a )
# compare the actual values for a slice of last dim
self.assertTrue(torch.allclose(output[:, :, -1] , _a , atol=1e-3 ) )
| 281 |
import argparse
import collections
import json
import os
import re
import string
import sys
import numpy as np
snake_case : Dict = re.compile(R"\b(a|an|the)\b", re.UNICODE)
snake_case : Optional[int] = None
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser("Official evaluation script for SQuAD version 2.0." )
parser.add_argument("data_file" , metavar="data.json" , help="Input data JSON file." )
parser.add_argument("pred_file" , metavar="pred.json" , help="Model predictions." )
parser.add_argument(
"--out-file" , "-o" , metavar="eval.json" , help="Write accuracy metrics to file (default is stdout)." )
parser.add_argument(
"--na-prob-file" , "-n" , metavar="na_prob.json" , help="Model estimates of probability of no answer." )
parser.add_argument(
"--na-prob-thresh" , "-t" , type=_snake_case , default=1.0 , help="Predict \"\" if no-answer probability exceeds this (default = 1.0)." , )
parser.add_argument(
"--out-image-dir" , "-p" , metavar="out_images" , default=_snake_case , help="Save precision-recall curves to directory." )
parser.add_argument("--verbose" , "-v" , action="store_true" )
if len(sys.argv ) == 1:
parser.print_help()
sys.exit(1 )
return parser.parse_args()
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Optional[int] = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : str = bool(qa["answers"]["text"] )
return qid_to_has_ans
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Optional[Any]:
'''simple docstring'''
def remove_articles(_snake_case : List[str] ):
return ARTICLES_REGEX.sub(" " , _snake_case )
def white_space_fix(_snake_case : Optional[int] ):
return " ".join(text.split() )
def remove_punc(_snake_case : Optional[int] ):
__magic_name__ : Dict = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(_snake_case : str ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(_snake_case ) ) ) )
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
if not s:
return []
return normalize_answer(_snake_case ).split()
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Dict ) -> Tuple:
'''simple docstring'''
return int(normalize_answer(_snake_case ) == normalize_answer(_snake_case ) )
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : int ) -> str:
'''simple docstring'''
__magic_name__ : Any = get_tokens(_snake_case )
__magic_name__ : Optional[int] = get_tokens(_snake_case )
__magic_name__ : Tuple = collections.Counter(_snake_case ) & collections.Counter(_snake_case )
__magic_name__ : Tuple = sum(common.values() )
if len(_snake_case ) == 0 or len(_snake_case ) == 0:
# If either is no-answer, then F1 is 1 if they agree, 0 otherwise
return int(gold_toks == pred_toks )
if num_same == 0:
return 0
__magic_name__ : Dict = 1.0 * num_same / len(_snake_case )
__magic_name__ : Optional[Any] = 1.0 * num_same / len(_snake_case )
__magic_name__ : List[Any] = (2 * precision * recall) / (precision + recall)
return fa
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = {}
__magic_name__ : int = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : Union[str, Any] = qa["id"]
__magic_name__ : Any = [t for t in qa["answers"]["text"] if normalize_answer(_snake_case )]
if not gold_answers:
# For unanswerable questions, only correct answer is empty string
__magic_name__ : Tuple = [""]
if qid not in preds:
print(F'''Missing prediction for {qid}''' )
continue
__magic_name__ : Any = preds[qid]
# Take max over all gold answers
__magic_name__ : List[Any] = max(compute_exact(_snake_case , _snake_case ) for a in gold_answers )
__magic_name__ : int = max(compute_fa(_snake_case , _snake_case ) for a in gold_answers )
return exact_scores, fa_scores
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] , _snake_case : Optional[int] , _snake_case : Dict ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : str = {}
for qid, s in scores.items():
__magic_name__ : Dict = na_probs[qid] > na_prob_thresh
if pred_na:
__magic_name__ : str = float(not qid_to_has_ans[qid] )
else:
__magic_name__ : Optional[int] = s
return new_scores
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : List[str] , _snake_case : Tuple=None ) -> Tuple:
'''simple docstring'''
if not qid_list:
__magic_name__ : Any = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores.values() ) / total),
("f1", 100.0 * sum(fa_scores.values() ) / total),
("total", total),
] )
else:
__magic_name__ : Tuple = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores[k] for k in qid_list ) / total),
("f1", 100.0 * sum(fa_scores[k] for k in qid_list ) / total),
("total", total),
] )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : str , _snake_case : str ) -> Dict:
'''simple docstring'''
for k in new_eval:
__magic_name__ : int = new_eval[k]
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : Optional[Any] , _snake_case : Union[str, Any] ) -> str:
'''simple docstring'''
plt.step(_snake_case , _snake_case , color="b" , alpha=0.2 , where="post" )
plt.fill_between(_snake_case , _snake_case , step="post" , alpha=0.2 , color="b" )
plt.xlabel("Recall" )
plt.ylabel("Precision" )
plt.xlim([0.0, 1.05] )
plt.ylim([0.0, 1.05] )
plt.title(_snake_case )
plt.savefig(_snake_case )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any , _snake_case : Optional[int] , _snake_case : List[Any] , _snake_case : Optional[int]=None , _snake_case : int=None ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
__magic_name__ : Optional[int] = 0.0
__magic_name__ : str = 1.0
__magic_name__ : str = 0.0
__magic_name__ : List[str] = [1.0]
__magic_name__ : str = [0.0]
__magic_name__ : Optional[Any] = 0.0
for i, qid in enumerate(_snake_case ):
if qid_to_has_ans[qid]:
true_pos += scores[qid]
__magic_name__ : List[str] = true_pos / float(i + 1 )
__magic_name__ : Any = true_pos / float(_snake_case )
if i == len(_snake_case ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]:
# i.e., if we can put a threshold after this point
avg_prec += cur_p * (cur_r - recalls[-1])
precisions.append(_snake_case )
recalls.append(_snake_case )
if out_image:
plot_pr_curve(_snake_case , _snake_case , _snake_case , _snake_case )
return {"ap": 100.0 * avg_prec}
def lowerCAmelCase_ ( _snake_case : Tuple , _snake_case : Optional[Any] , _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : Any , _snake_case : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
if out_image_dir and not os.path.exists(_snake_case ):
os.makedirs(_snake_case )
__magic_name__ : Any = sum(1 for v in qid_to_has_ans.values() if v )
if num_true_pos == 0:
return
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_exact.png" ) , title="Precision-Recall curve for Exact Match score" , )
__magic_name__ : Union[str, Any] = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_f1.png" ) , title="Precision-Recall curve for F1 score" , )
__magic_name__ : str = {k: float(_snake_case ) for k, v in qid_to_has_ans.items()}
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_oracle.png" ) , title="Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)" , )
merge_eval(_snake_case , _snake_case , "pr_exact" )
merge_eval(_snake_case , _snake_case , "pr_f1" )
merge_eval(_snake_case , _snake_case , "pr_oracle" )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : Optional[Any] , _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
if not qid_list:
return
__magic_name__ : Dict = [na_probs[k] for k in qid_list]
__magic_name__ : str = np.ones_like(_snake_case ) / float(len(_snake_case ) )
plt.hist(_snake_case , weights=_snake_case , bins=20 , range=(0.0, 1.0) )
plt.xlabel("Model probability of no-answer" )
plt.ylabel("Proportion of dataset" )
plt.title(F'''Histogram of no-answer probability: {name}''' )
plt.savefig(os.path.join(_snake_case , F'''na_prob_hist_{name}.png''' ) )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : List[str] , _snake_case : Dict ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] )
__magic_name__ : List[str] = num_no_ans
__magic_name__ : Dict = cur_score
__magic_name__ : Dict = 0.0
__magic_name__ : Any = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
for i, qid in enumerate(_snake_case ):
if qid not in scores:
continue
if qid_to_has_ans[qid]:
__magic_name__ : Union[str, Any] = scores[qid]
else:
if preds[qid]:
__magic_name__ : List[Any] = -1
else:
__magic_name__ : Optional[int] = 0
cur_score += diff
if cur_score > best_score:
__magic_name__ : Optional[int] = cur_score
__magic_name__ : List[Any] = na_probs[qid]
return 100.0 * best_score / len(_snake_case ), best_thresh
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str , _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Dict ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ , __magic_name__ : List[str] = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ , __magic_name__ : int = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ : Optional[int] = best_exact
__magic_name__ : List[Any] = exact_thresh
__magic_name__ : Dict = best_fa
__magic_name__ : Any = fa_thresh
def lowerCAmelCase_ ( ) -> int:
'''simple docstring'''
with open(OPTS.data_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
__magic_name__ : List[Any] = dataset_json["data"]
with open(OPTS.pred_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
if OPTS.na_prob_file:
with open(OPTS.na_prob_file ) as f:
__magic_name__ : Any = json.load(_snake_case )
else:
__magic_name__ : Any = {k: 0.0 for k in preds}
__magic_name__ : str = make_qid_to_has_ans(_snake_case ) # maps qid to True/False
__magic_name__ : Tuple = [k for k, v in qid_to_has_ans.items() if v]
__magic_name__ : Optional[Any] = [k for k, v in qid_to_has_ans.items() if not v]
__magic_name__ , __magic_name__ : Union[str, Any] = get_raw_scores(_snake_case , _snake_case )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case )
if has_ans_qids:
__magic_name__ : int = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "HasAns" )
if no_ans_qids:
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "NoAns" )
if OPTS.na_prob_file:
find_all_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
if OPTS.na_prob_file and OPTS.out_image_dir:
run_precision_recall_analysis(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , OPTS.out_image_dir )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "hasAns" )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "noAns" )
if OPTS.out_file:
with open(OPTS.out_file , "w" ) as f:
json.dump(_snake_case , _snake_case )
else:
print(json.dumps(_snake_case , indent=2 ) )
if __name__ == "__main__":
snake_case : int = parse_args()
if OPTS.out_image_dir:
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
main()
| 281 | 1 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_torch_available,
)
snake_case : List[Any] = {"configuration_vit_mae": ["VIT_MAE_PRETRAINED_CONFIG_ARCHIVE_MAP", "ViTMAEConfig"]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : int = [
"VIT_MAE_PRETRAINED_MODEL_ARCHIVE_LIST",
"ViTMAEForPreTraining",
"ViTMAELayer",
"ViTMAEModel",
"ViTMAEPreTrainedModel",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : str = [
"TFViTMAEForPreTraining",
"TFViTMAEModel",
"TFViTMAEPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_vit_mae import VIT_MAE_PRETRAINED_CONFIG_ARCHIVE_MAP, ViTMAEConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_vit_mae import (
VIT_MAE_PRETRAINED_MODEL_ARCHIVE_LIST,
ViTMAEForPreTraining,
ViTMAELayer,
ViTMAEModel,
ViTMAEPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_vit_mae import TFViTMAEForPreTraining, TFViTMAEModel, TFViTMAEPreTrainedModel
else:
import sys
snake_case : Dict = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 281 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
snake_case : str = "▁"
snake_case : List[Any] = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = BigBirdTokenizer
UpperCamelCase__ = BigBirdTokenizerFast
UpperCamelCase__ = True
UpperCamelCase__ = True
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
__magic_name__ : Optional[Any] = self.tokenizer_class(_a , keep_accents=_a )
tokenizer.save_pretrained(self.tmpdirname )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = "<s>"
__magic_name__ : Dict = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_a ) , _a )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_a ) , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : 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] , "[MASK]" )
self.assertEqual(len(_a ) , 1_004 )
def SCREAMING_SNAKE_CASE ( self ):
self.assertEqual(self.get_tokenizer().vocab_size , 1_000 )
def SCREAMING_SNAKE_CASE ( self ):
if not self.test_rust_tokenizer:
return
__magic_name__ : Dict = self.get_tokenizer()
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Any = "I was born in 92000, and this is falsé."
__magic_name__ : Dict = tokenizer.tokenize(_a )
__magic_name__ : Any = rust_tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : List[Any] = tokenizer.encode(_a , add_special_tokens=_a )
__magic_name__ : List[str] = rust_tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Dict = tokenizer.encode(_a )
__magic_name__ : Optional[int] = rust_tokenizer.encode(_a )
self.assertListEqual(_a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = BigBirdTokenizer(_a , keep_accents=_a )
__magic_name__ : str = tokenizer.tokenize("This is a test" )
self.assertListEqual(_a , ["▁This", "▁is", "▁a", "▁t", "est"] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_a ) , [285, 46, 10, 170, 382] , )
__magic_name__ : Dict = tokenizer.tokenize("I was born in 92000, and this is falsé." )
self.assertListEqual(
_a , [
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",
".",
] , )
__magic_name__ : Union[str, Any] = tokenizer.convert_tokens_to_ids(_a )
self.assertListEqual(
_a , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
__magic_name__ : int = tokenizer.convert_ids_to_tokens(_a )
self.assertListEqual(
_a , [
SPIECE_UNDERLINE + "I",
SPIECE_UNDERLINE + "was",
SPIECE_UNDERLINE + "b",
"or",
"n",
SPIECE_UNDERLINE + "in",
SPIECE_UNDERLINE + "",
"<unk>",
"2",
"0",
"0",
"0",
",",
SPIECE_UNDERLINE + "and",
SPIECE_UNDERLINE + "this",
SPIECE_UNDERLINE + "is",
SPIECE_UNDERLINE + "f",
"al",
"s",
"<unk>",
".",
] , )
@cached_property
def SCREAMING_SNAKE_CASE ( self ):
return BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "Hello World!"
__magic_name__ : Dict = [65, 18_536, 2_260, 101, 66]
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = (
"This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"
" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"
)
# fmt: off
__magic_name__ : List[str] = [65, 871, 419, 358, 946, 991, 2_521, 452, 358, 1_357, 387, 7_751, 3_536, 112, 985, 456, 126, 865, 938, 5_400, 5_734, 458, 1_368, 467, 786, 2_462, 5_246, 1_159, 633, 865, 4_519, 457, 582, 852, 2_557, 427, 916, 508, 405, 34_324, 497, 391, 408, 11_342, 1_244, 385, 100, 938, 985, 456, 574, 362, 12_597, 3_200, 3_129, 1_172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
__magic_name__ : Optional[Any] = list(self.big_tokenizer.get_vocab().keys() )[:10]
__magic_name__ : List[Any] = " ".join(_a )
__magic_name__ : Any = self.big_tokenizer.encode_plus(_a , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : Union[str, Any] = self.big_tokenizer.batch_encode_plus(
[sequence + " " + sequence] , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : List[str] = BigBirdConfig(attention_type="original_full" )
__magic_name__ : Optional[int] = BigBirdModel(_a )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_a )
model(**_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
__magic_name__ : int = tokenizer.decode(tokenizer("Paris is the [MASK]." ).input_ids )
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# fmt: off
__magic_name__ : Optional[Any] = {"input_ids": [[65, 39_286, 458, 36_335, 2_001, 456, 13_073, 13_266, 455, 113, 7_746, 1_741, 11_157, 391, 13_073, 13_266, 455, 113, 3_967, 35_412, 113, 4_936, 109, 3_870, 2_377, 113, 30_084, 45_720, 458, 134, 17_496, 112, 503, 11_672, 113, 118, 112, 5_665, 13_347, 38_687, 112, 1_496, 31_389, 112, 3_268, 47_264, 134, 962, 112, 16_377, 8_035, 23_130, 430, 12_169, 15_518, 28_592, 458, 146, 41_697, 109, 391, 12_169, 15_518, 16_689, 458, 146, 41_358, 109, 452, 726, 4_034, 111, 763, 35_412, 5_082, 388, 1_903, 111, 9_051, 391, 2_870, 48_918, 1_900, 1_123, 550, 998, 112, 9_586, 15_985, 455, 391, 410, 22_955, 37_636, 114, 66], [65, 448, 17_496, 419, 3_663, 385, 763, 113, 27_533, 2_870, 3_283, 13_043, 1_639, 24_713, 523, 656, 24_013, 18_550, 2_521, 517, 27_014, 21_244, 420, 1_212, 1_465, 391, 927, 4_833, 388, 578, 11_786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2_169, 7_687, 21_932, 18_146, 726, 363, 17_032, 3_391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], "attention_mask": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_a , model_name="google/bigbird-roberta-base" , revision="215c99f1600e06f83acce68422f2035b2b5c3510" , )
| 281 | 1 |
from __future__ import annotations
from collections.abc import Iterator
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Any = value
__magic_name__ : Node | None = None
__magic_name__ : Node | None = None
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Any = tree
def SCREAMING_SNAKE_CASE ( self , _a ):
if node is None:
return 0
return node.value + (
self.depth_first_search(node.left ) + self.depth_first_search(node.right )
)
def __iter__( self ):
yield self.depth_first_search(self.tree )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
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
snake_case : int = logging.get_logger(__name__)
snake_case : List[str] = {"vocab_file": "spiece.model"}
snake_case : List[str] = {
"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",
}
}
snake_case : Tuple = {
"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,
}
snake_case : List[str] = "▁"
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a=True , _a=True , _a=False , _a="[CLS]" , _a="[SEP]" , _a="<unk>" , _a="[SEP]" , _a="<pad>" , _a="[CLS]" , _a="[MASK]" , _a = None , **_a , ):
# 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.
__magic_name__ : str = (
AddedToken(_a , lstrip=_a , rstrip=_a , normalized=_a )
if isinstance(_a , _a )
else mask_token
)
__magic_name__ : Union[str, Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_a , remove_space=_a , keep_accents=_a , bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , pad_token=_a , cls_token=_a , mask_token=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Dict = do_lower_case
__magic_name__ : Tuple = remove_space
__magic_name__ : Union[str, Any] = keep_accents
__magic_name__ : Tuple = vocab_file
__magic_name__ : int = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_a )
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.sp_model )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ):
__magic_name__ : List[str] = self.__dict__.copy()
__magic_name__ : Any = None
return state
def __setstate__( self , _a ):
__magic_name__ : Union[str, Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : str = {}
__magic_name__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.remove_space:
__magic_name__ : List[Any] = " ".join(inputs.strip().split() )
else:
__magic_name__ : str = inputs
__magic_name__ : int = outputs.replace("``" , "\"" ).replace("''" , "\"" )
if not self.keep_accents:
__magic_name__ : str = unicodedata.normalize("NFKD" , _a )
__magic_name__ : Tuple = "".join([c for c in outputs if not unicodedata.combining(_a )] )
if self.do_lower_case:
__magic_name__ : int = outputs.lower()
return outputs
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = self.preprocess_text(_a )
__magic_name__ : Dict = self.sp_model.encode(_a , out_type=_a )
__magic_name__ : Any = []
for piece in pieces:
if len(_a ) > 1 and piece[-1] == str("," ) and piece[-2].isdigit():
__magic_name__ : Optional[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(_a , "" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
__magic_name__ : List[str] = cur_pieces[1:]
else:
__magic_name__ : Optional[int] = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(_a )
else:
new_pieces.append(_a )
return new_pieces
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.PieceToId(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.IdToPiece(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Any = []
__magic_name__ : Union[str, Any] = ""
__magic_name__ : int = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_a ) + token
__magic_name__ : List[Any] = True
__magic_name__ : Optional[int] = []
else:
current_sub_tokens.append(_a )
__magic_name__ : Optional[Any] = False
out_string += self.sp_model.decode(_a )
return out_string.strip()
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : List[str] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is not None:
return [1] + ([0] * len(_a )) + [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[int] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : List[str] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _a )
elif not os.path.isfile(self.vocab_file ):
with open(_a , "wb" ) as fi:
__magic_name__ : Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(_a )
return (out_vocab_file,)
| 281 | 1 |
import argparse
import logging
import pickle
import random
import time
import numpy as np
from transformers import BertTokenizer, GPTaTokenizer, RobertaTokenizer
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s", datefmt="%m/%d/%Y %H:%M:%S", level=logging.INFO
)
snake_case : Any = logging.getLogger(__name__)
def lowerCAmelCase_ ( ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Tuple = argparse.ArgumentParser(
description="Preprocess the data to avoid re-doing it several times by (tokenization + token_to_ids)." )
parser.add_argument("--file_path" , type=_snake_case , default="data/dump.txt" , help="The path to the data." )
parser.add_argument("--tokenizer_type" , type=_snake_case , default="bert" , choices=["bert", "roberta", "gpt2"] )
parser.add_argument("--tokenizer_name" , type=_snake_case , default="bert-base-uncased" , help="The tokenizer to use." )
parser.add_argument("--dump_file" , type=_snake_case , default="data/dump" , help="The dump file prefix." )
__magic_name__ : List[Any] = parser.parse_args()
logger.info(F'''Loading Tokenizer ({args.tokenizer_name})''' )
if args.tokenizer_type == "bert":
__magic_name__ : Dict = BertTokenizer.from_pretrained(args.tokenizer_name )
__magic_name__ : Dict = tokenizer.special_tokens_map["cls_token"] # `[CLS]`
__magic_name__ : str = tokenizer.special_tokens_map["sep_token"] # `[SEP]`
elif args.tokenizer_type == "roberta":
__magic_name__ : Optional[int] = RobertaTokenizer.from_pretrained(args.tokenizer_name )
__magic_name__ : List[Any] = tokenizer.special_tokens_map["cls_token"] # `<s>`
__magic_name__ : str = tokenizer.special_tokens_map["sep_token"] # `</s>`
elif args.tokenizer_type == "gpt2":
__magic_name__ : List[Any] = GPTaTokenizer.from_pretrained(args.tokenizer_name )
__magic_name__ : str = tokenizer.special_tokens_map["bos_token"] # `<|endoftext|>`
__magic_name__ : Union[str, Any] = tokenizer.special_tokens_map["eos_token"] # `<|endoftext|>`
logger.info(F'''Loading text from {args.file_path}''' )
with open(args.file_path , "r" , encoding="utf8" ) as fp:
__magic_name__ : Dict = fp.readlines()
logger.info("Start encoding" )
logger.info(F'''{len(_snake_case )} examples to process.''' )
__magic_name__ : Optional[Any] = []
__magic_name__ : Dict = 0
__magic_name__ : str = 10000
__magic_name__ : Union[str, Any] = time.time()
for text in data:
__magic_name__ : Union[str, Any] = F'''{bos} {text.strip()} {sep}'''
__magic_name__ : int = tokenizer.encode(_snake_case , add_special_tokens=_snake_case )
rslt.append(_snake_case )
iter += 1
if iter % interval == 0:
__magic_name__ : Tuple = time.time()
logger.info(F'''{iter} examples processed. - {(end-start):.2f}s/{interval}expl''' )
__magic_name__ : Optional[Any] = time.time()
logger.info("Finished binarization" )
logger.info(F'''{len(_snake_case )} examples processed.''' )
__magic_name__ : str = F'''{args.dump_file}.{args.tokenizer_name}.pickle'''
__magic_name__ : Optional[Any] = tokenizer.vocab_size
if vocab_size < (1 << 16):
__magic_name__ : Union[str, Any] = [np.uintaa(_snake_case ) for d in rslt]
else:
__magic_name__ : Dict = [np.intaa(_snake_case ) for d in rslt]
random.shuffle(rslt_ )
logger.info(F'''Dump to {dp_file}''' )
with open(_snake_case , "wb" ) as handle:
pickle.dump(rslt_ , _snake_case , protocol=pickle.HIGHEST_PROTOCOL )
if __name__ == "__main__":
main()
| 281 |
import unicodedata
from dataclasses import dataclass
from typing import Optional, Union
import numpy as np
from transformers.data.data_collator import DataCollatorMixin
from transformers.file_utils import PaddingStrategy
from transformers.tokenization_utils_base import PreTrainedTokenizerBase
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Union[str, Any] = np.full((len(_snake_case ), sequence_length, 2) , _snake_case )
else:
__magic_name__ : List[Any] = np.full((len(_snake_case ), sequence_length) , _snake_case )
for i, tensor in enumerate(_snake_case ):
if padding_side == "right":
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Optional[Any] = tensor[:sequence_length]
else:
__magic_name__ : Union[str, Any] = tensor[:sequence_length]
else:
if isinstance(_snake_case , _snake_case ):
__magic_name__ : List[Any] = tensor[:sequence_length]
else:
__magic_name__ : Optional[Any] = tensor[:sequence_length]
return out_tensor.tolist()
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = ord(_snake_case )
if (cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126):
return True
__magic_name__ : Any = unicodedata.category(_snake_case )
if cat.startswith("P" ):
return True
return False
@dataclass
class _snake_case ( snake_case ):
UpperCamelCase__ = 42
UpperCamelCase__ = True
UpperCamelCase__ = None
UpperCamelCase__ = None
UpperCamelCase__ = -100
UpperCamelCase__ = "pt"
def SCREAMING_SNAKE_CASE ( self , _a ):
import torch
__magic_name__ : List[str] = "label" if "label" in features[0].keys() else "labels"
__magic_name__ : Union[str, Any] = [feature[label_name] for feature in features] if label_name in features[0].keys() else None
__magic_name__ : Optional[int] = self.tokenizer.pad(
_a , padding=self.padding , max_length=self.max_length , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors="pt" if labels is None else None , )
if labels is None:
return batch
__magic_name__ : Dict = torch.tensor(batch["entity_ids"] ).shape[1]
__magic_name__ : List[Any] = self.tokenizer.padding_side
if padding_side == "right":
__magic_name__ : str = [
list(_a ) + [self.label_pad_token_id] * (sequence_length - len(_a )) for label in labels
]
else:
__magic_name__ : int = [
[self.label_pad_token_id] * (sequence_length - len(_a )) + list(_a ) for label in labels
]
__magic_name__ : Dict = [feature["ner_tags"] for feature in features]
__magic_name__ : List[Any] = padding_tensor(_a , -1 , _a , _a )
__magic_name__ : Any = [feature["original_entity_spans"] for feature in features]
__magic_name__ : Any = padding_tensor(_a , (-1, -1) , _a , _a )
__magic_name__ : List[Any] = {k: torch.tensor(_a , dtype=torch.intaa ) for k, v in batch.items()}
return batch
| 281 | 1 |
import time
from contextlib import contextmanager
from pathlib import Path
import pytest
import requests
from huggingface_hub.hf_api import HfApi, HfFolder
snake_case : Tuple = "__DUMMY_TRANSFORMERS_USER__"
snake_case : List[Any] = "Dummy User"
snake_case : int = "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"
snake_case : int = "https://hub-ci.huggingface.co"
snake_case : Tuple = CI_HUB_ENDPOINT + "/datasets/{repo_id}/resolve/{revision}/{path}"
snake_case : Tuple = CI_HUB_ENDPOINT + "/{repo_id}/resolve/{revision}/{filename}"
snake_case : Union[str, Any] = Path("~/.huggingface/hub_ci_token").expanduser()
@pytest.fixture
def lowerCAmelCase_ ( _snake_case : Dict ) -> Any:
'''simple docstring'''
monkeypatch.setattr(
"huggingface_hub.file_download.HUGGINGFACE_CO_URL_TEMPLATE" , _snake_case )
@pytest.fixture
def lowerCAmelCase_ ( _snake_case : int ) -> List[str]:
'''simple docstring'''
monkeypatch.setattr("datasets.config.HF_ENDPOINT" , _snake_case )
monkeypatch.setattr("datasets.config.HUB_DATASETS_URL" , _snake_case )
@pytest.fixture
def lowerCAmelCase_ ( _snake_case : Dict ) -> int:
'''simple docstring'''
monkeypatch.setattr("huggingface_hub.hf_api.HfFolder.path_token" , _snake_case )
@pytest.fixture
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str ) -> Optional[int]:
'''simple docstring'''
HfFolder.save_token(_snake_case )
yield
HfFolder.delete_token()
@pytest.fixture(scope="session" )
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
return HfApi(endpoint=_snake_case )
@pytest.fixture(scope="session" )
def lowerCAmelCase_ ( _snake_case : HfApi ) -> Any:
'''simple docstring'''
__magic_name__ : List[Any] = HfFolder.get_token()
HfFolder.save_token(_snake_case )
yield CI_HUB_USER_TOKEN
if previous_token is not None:
HfFolder.save_token(_snake_case )
@pytest.fixture
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Any:
'''simple docstring'''
def _cleanup_repo(_snake_case : Optional[Any] ):
hf_api.delete_repo(_snake_case , token=_snake_case , repo_type="dataset" )
return _cleanup_repo
@pytest.fixture
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
@contextmanager
def _temporary_repo(_snake_case : Optional[Any] ):
try:
yield repo_id
finally:
cleanup_repo(_snake_case )
return _temporary_repo
@pytest.fixture(scope="session" )
def lowerCAmelCase_ ( _snake_case : HfApi , _snake_case : Union[str, Any] , _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = F'''repo_txt_data-{int(time.time() * 10E3 )}'''
__magic_name__ : List[Any] = F'''{CI_HUB_USER}/{repo_name}'''
hf_api.create_repo(_snake_case , token=_snake_case , repo_type="dataset" , private=_snake_case )
hf_api.upload_file(
token=_snake_case , path_or_fileobj=str(_snake_case ) , path_in_repo="data/text_data.txt" , repo_id=_snake_case , repo_type="dataset" , )
yield repo_id
try:
hf_api.delete_repo(_snake_case , token=_snake_case , repo_type="dataset" )
except (requests.exceptions.HTTPError, ValueError): # catch http error and token invalid error
pass
@pytest.fixture()
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Optional[int] , _snake_case : List[Any] ) -> Any:
'''simple docstring'''
return hf_private_dataset_repo_txt_data_
@pytest.fixture(scope="session" )
def lowerCAmelCase_ ( _snake_case : HfApi , _snake_case : Union[str, Any] , _snake_case : Optional[Any] ) -> int:
'''simple docstring'''
__magic_name__ : Union[str, Any] = F'''repo_zipped_txt_data-{int(time.time() * 10E3 )}'''
__magic_name__ : List[str] = F'''{CI_HUB_USER}/{repo_name}'''
hf_api.create_repo(_snake_case , token=_snake_case , repo_type="dataset" , private=_snake_case )
hf_api.upload_file(
token=_snake_case , path_or_fileobj=str(_snake_case ) , path_in_repo="data.zip" , repo_id=_snake_case , repo_type="dataset" , )
yield repo_id
try:
hf_api.delete_repo(_snake_case , token=_snake_case , repo_type="dataset" )
except (requests.exceptions.HTTPError, ValueError): # catch http error and token invalid error
pass
@pytest.fixture()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : int ) -> Optional[Any]:
'''simple docstring'''
return hf_private_dataset_repo_zipped_txt_data_
@pytest.fixture(scope="session" )
def lowerCAmelCase_ ( _snake_case : HfApi , _snake_case : Tuple , _snake_case : Optional[int] ) -> int:
'''simple docstring'''
__magic_name__ : int = F'''repo_zipped_img_data-{int(time.time() * 10E3 )}'''
__magic_name__ : int = F'''{CI_HUB_USER}/{repo_name}'''
hf_api.create_repo(_snake_case , token=_snake_case , repo_type="dataset" , private=_snake_case )
hf_api.upload_file(
token=_snake_case , path_or_fileobj=str(_snake_case ) , path_in_repo="data.zip" , repo_id=_snake_case , repo_type="dataset" , )
yield repo_id
try:
hf_api.delete_repo(_snake_case , token=_snake_case , repo_type="dataset" )
except (requests.exceptions.HTTPError, ValueError): # catch http error and token invalid error
pass
@pytest.fixture()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : int , _snake_case : Optional[Any] ) -> int:
'''simple docstring'''
return hf_private_dataset_repo_zipped_img_data_
| 281 |
import math
def lowerCAmelCase_ ( _snake_case : float , _snake_case : float ) -> float:
'''simple docstring'''
return math.pow(_snake_case , 2 ) - a
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
return 2 * x
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
__magic_name__ : Optional[int] = 2.0
while start <= a:
__magic_name__ : str = math.pow(_snake_case , 2 )
return start
def lowerCAmelCase_ ( _snake_case : float , _snake_case : int = 9999 , _snake_case : float = 0.00_000_000_000_001 ) -> float:
'''simple docstring'''
if a < 0:
raise ValueError("math domain error" )
__magic_name__ : Optional[int] = get_initial_point(_snake_case )
for _ in range(_snake_case ):
__magic_name__ : int = value
__magic_name__ : str = value - fx(_snake_case , _snake_case ) / fx_derivative(_snake_case )
if abs(prev_value - value ) < tolerance:
return value
return value
if __name__ == "__main__":
from doctest import testmod
testmod()
| 281 | 1 |
from ...configuration_utils import PretrainedConfig
class _snake_case ( snake_case ):
UpperCamelCase__ = 'bert-generation'
def __init__( self , _a=50_358 , _a=1_024 , _a=24 , _a=16 , _a=4_096 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=0.02 , _a=1e-12 , _a=0 , _a=2 , _a=1 , _a="absolute" , _a=True , **_a , ):
super().__init__(pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a )
__magic_name__ : str = vocab_size
__magic_name__ : Dict = hidden_size
__magic_name__ : str = num_hidden_layers
__magic_name__ : List[str] = num_attention_heads
__magic_name__ : Optional[int] = hidden_act
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Dict = hidden_dropout_prob
__magic_name__ : List[Any] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Optional[int] = initializer_range
__magic_name__ : List[Any] = layer_norm_eps
__magic_name__ : Optional[Any] = position_embedding_type
__magic_name__ : List[Any] = use_cache
| 281 |
from __future__ import annotations
import unittest
from transformers import LEDConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFLEDForConditionalGeneration, TFLEDModel
@require_tf
class _snake_case :
UpperCamelCase__ = LEDConfig
UpperCamelCase__ = {}
UpperCamelCase__ = 'gelu'
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=False , _a=99 , _a=32 , _a=2 , _a=4 , _a=37 , _a=0.1 , _a=0.1 , _a=20 , _a=2 , _a=1 , _a=0 , _a=4 , ):
__magic_name__ : int = parent
__magic_name__ : Optional[int] = batch_size
__magic_name__ : Tuple = seq_length
__magic_name__ : List[Any] = is_training
__magic_name__ : Dict = use_labels
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : int = hidden_size
__magic_name__ : Optional[int] = num_hidden_layers
__magic_name__ : Optional[int] = num_attention_heads
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Any = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Any = eos_token_id
__magic_name__ : str = pad_token_id
__magic_name__ : int = bos_token_id
__magic_name__ : Optional[int] = attention_window
# `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size
# [num_attention_heads, encoder_seq_length, encoder_key_length], but TFLongformerSelfAttention
# returns attention of shape [num_attention_heads, encoder_seq_length, self.attention_window + 1]
# because its local attention only attends to `self.attention_window` and one before and one after
__magic_name__ : Tuple = self.attention_window + 2
# because of padding `encoder_seq_length`, is different from `seq_length`. Relevant for
# the `test_attention_outputs` and `test_hidden_states_output` tests
__magic_name__ : Tuple = (
self.seq_length + (self.attention_window - self.seq_length % self.attention_window) % self.attention_window
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
__magic_name__ : int = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
__magic_name__ : int = tf.concat([input_ids, eos_tensor] , axis=1 )
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : Dict = self.config_cls(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_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_ids=[2] , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.pad_token_id , attention_window=self.attention_window , **self.config_updates , )
__magic_name__ : List[str] = prepare_led_inputs_dict(_a , _a , _a )
__magic_name__ : Union[str, Any] = tf.concat(
[tf.zeros_like(_a )[:, :-1], tf.ones_like(_a )[:, -1:]] , axis=-1 , )
__magic_name__ : List[Any] = global_attention_mask
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Dict = TFLEDModel(config=_a ).get_decoder()
__magic_name__ : Optional[int] = inputs_dict["input_ids"]
__magic_name__ : Union[str, Any] = input_ids[:1, :]
__magic_name__ : str = inputs_dict["attention_mask"][:1, :]
__magic_name__ : int = 1
# first forward pass
__magic_name__ : Tuple = model(_a , attention_mask=_a , use_cache=_a )
__magic_name__ , __magic_name__ : str = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
__magic_name__ : Optional[int] = ids_tensor((self.batch_size, 3) , config.vocab_size )
__magic_name__ : Any = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
__magic_name__ : Optional[Any] = tf.concat([input_ids, next_tokens] , axis=-1 )
__magic_name__ : List[Any] = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
__magic_name__ : List[str] = model(_a , attention_mask=_a )[0]
__magic_name__ : Dict = model(_a , attention_mask=_a , past_key_values=_a )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
__magic_name__ : List[Any] = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
__magic_name__ : Union[str, Any] = output_from_no_past[:, -3:, random_slice_idx]
__magic_name__ : List[str] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(_a , _a , rtol=1e-3 )
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : List[Any] , _snake_case : Any , _snake_case : str=None , _snake_case : List[str]=None , _snake_case : int=None , _snake_case : Any=None , ) -> int:
'''simple docstring'''
if attention_mask is None:
__magic_name__ : str = tf.cast(tf.math.not_equal(_snake_case , config.pad_token_id ) , tf.inta )
if decoder_attention_mask is None:
__magic_name__ : List[Any] = tf.concat(
[
tf.ones(decoder_input_ids[:, :1].shape , dtype=tf.inta ),
tf.cast(tf.math.not_equal(decoder_input_ids[:, 1:] , config.pad_token_id ) , tf.inta ),
] , axis=-1 , )
if head_mask is None:
__magic_name__ : Dict = tf.ones((config.encoder_layers, config.encoder_attention_heads) )
if decoder_head_mask is None:
__magic_name__ : Any = tf.ones((config.decoder_layers, config.decoder_attention_heads) )
return {
"input_ids": input_ids,
"attention_mask": attention_mask,
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
}
@require_tf
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (TFLEDForConditionalGeneration, TFLEDModel) if is_tf_available() else ()
UpperCamelCase__ = (TFLEDForConditionalGeneration,) if is_tf_available() else ()
UpperCamelCase__ = (
{
'conversational': TFLEDForConditionalGeneration,
'feature-extraction': TFLEDModel,
'summarization': TFLEDForConditionalGeneration,
'text2text-generation': TFLEDForConditionalGeneration,
'translation': TFLEDForConditionalGeneration,
}
if is_tf_available()
else {}
)
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = TFLEDModelTester(self )
__magic_name__ : List[Any] = ConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : int = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[str] = tf.zeros_like(inputs_dict["attention_mask"] )
__magic_name__ : Optional[Any] = 2
__magic_name__ : Tuple = tf.where(
tf.range(self.model_tester.seq_length )[None, :] < num_global_attn_indices , 1 , inputs_dict["global_attention_mask"] , )
__magic_name__ : Any = True
__magic_name__ : str = self.model_tester.seq_length
__magic_name__ : Dict = self.model_tester.encoder_seq_length
def check_decoder_attentions_output(_a ):
__magic_name__ : str = outputs.decoder_attentions
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(decoder_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
def check_encoder_attentions_output(_a ):
__magic_name__ : Any = [t.numpy() for t in outputs.encoder_attentions]
__magic_name__ : Tuple = [t.numpy() for t in outputs.encoder_global_attentions]
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
self.assertListEqual(
list(global_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, encoder_seq_length, num_global_attn_indices] , )
for model_class in self.all_model_classes:
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = False
__magic_name__ : Tuple = False
__magic_name__ : Optional[int] = model_class(_a )
__magic_name__ : str = model(self._prepare_for_class(_a , _a ) )
__magic_name__ : Any = len(_a )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
if self.is_encoder_decoder:
__magic_name__ : Tuple = model_class(_a )
__magic_name__ : Optional[Any] = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_decoder_attentions_output(_a )
# Check that output attentions can also be changed via the config
del inputs_dict["output_attentions"]
__magic_name__ : Dict = True
__magic_name__ : str = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
# Check attention is always last and order is fine
__magic_name__ : Union[str, Any] = True
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(out_len + (2 if self.is_encoder_decoder else 1) , len(_a ) )
self.assertEqual(model.config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
@unittest.skip("LED keeps using potentially symbolic tensors in conditionals and breaks tracing." )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
# TODO: Head-masking not yet implement
pass
def lowerCAmelCase_ ( _snake_case : int ) -> Optional[int]:
'''simple docstring'''
return tf.constant(_snake_case , dtype=tf.intaa )
snake_case : Optional[int] = 1E-4
@slow
@require_tf
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" ).led
# change to intended input here
__magic_name__ : Optional[int] = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : str = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Any = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : List[Any] = model(**_a )[0]
__magic_name__ : List[str] = (1, 1_024, 768)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : int = tf.convert_to_tensor(
[[2.30_50, 2.82_79, 0.65_31], [-1.84_57, -0.14_55, -3.56_61], [-1.01_86, 0.45_86, -2.20_43]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" )
# change to intended input here
__magic_name__ : int = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Tuple = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Optional[Any] = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : Union[str, Any] = model(**_a )[0]
__magic_name__ : Optional[int] = (1, 1_024, model.config.vocab_size)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : str = tf.convert_to_tensor(
[[33.65_07, 6.45_72, 16.80_89], [5.87_39, -2.42_38, 11.29_02], [-3.21_39, -4.31_49, 4.27_83]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 , rtol=1e-3 )
| 281 | 1 |
import os
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : Optional[int] = {"vocab_file": "sentencepiece.bpe.model"}
snake_case : List[str] = {
"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"
),
},
}
snake_case : Optional[int] = {
"moussaKam/mbarthez": 1_024,
"moussaKam/barthez": 1_024,
"moussaKam/barthez-orangesum-title": 1_024,
}
snake_case : Optional[Any] = "▁"
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ = ['input_ids', 'attention_mask']
def __init__( self , _a , _a="<s>" , _a="</s>" , _a="</s>" , _a="<s>" , _a="<unk>" , _a="<pad>" , _a="<mask>" , _a = None , **_a , ):
# Mask token behave like a normal word, i.e. include the space before it
__magic_name__ : List[str] = AddedToken(_a , lstrip=_a , rstrip=_a ) if isinstance(_a , _a ) else mask_token
__magic_name__ : List[Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , cls_token=_a , pad_token=_a , mask_token=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Union[str, Any] = vocab_file
__magic_name__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(str(_a ) )
__magic_name__ : List[Any] = {"<s>": 0, "<pad>": 1, "</s>": 2, "<unk>": 3}
__magic_name__ : str = len(self.sp_model ) - 1
__magic_name__ : Optional[int] = {v: k for k, v in self.fairseq_tokens_to_ids.items()}
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
__magic_name__ : Any = [self.cls_token_id]
__magic_name__ : Any = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is None:
return [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : str = [self.sep_token_id]
__magic_name__ : Optional[Any] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep + sep + token_ids_a + sep ) * [0]
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.sp_model )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.encode(_a , out_type=_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
if token in self.fairseq_tokens_to_ids:
return self.fairseq_tokens_to_ids[token]
__magic_name__ : List[str] = self.sp_model.PieceToId(_a )
return spm_id if spm_id else self.unk_token_id
def SCREAMING_SNAKE_CASE ( self , _a ):
if index in self.fairseq_ids_to_tokens:
return self.fairseq_ids_to_tokens[index]
return self.sp_model.IdToPiece(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Any = []
__magic_name__ : Tuple = ""
__magic_name__ : Any = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_a ) + token
__magic_name__ : List[str] = True
__magic_name__ : Union[str, Any] = []
else:
current_sub_tokens.append(_a )
__magic_name__ : Optional[int] = False
out_string += self.sp_model.decode(_a )
return out_string.strip()
def __getstate__( self ):
__magic_name__ : str = self.__dict__.copy()
__magic_name__ : Tuple = None
return state
def __setstate__( self , _a ):
__magic_name__ : List[Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : str = {}
__magic_name__ : Dict = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : Dict = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _a )
elif not os.path.isfile(self.vocab_file ):
with open(_a , "wb" ) as fi:
__magic_name__ : Dict = self.sp_model.serialized_model_proto()
fi.write(_a )
return (out_vocab_file,)
| 281 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import (
BitConfig,
ViTHybridConfig,
ViTHybridForImageClassification,
ViTHybridImageProcessor,
ViTHybridModel,
)
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
snake_case : Optional[Any] = logging.get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any]=False ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
# fmt: off
# stem:
rename_keys.append(("cls_token", "vit.embeddings.cls_token") )
rename_keys.append(("pos_embed", "vit.embeddings.position_embeddings") )
rename_keys.append(("patch_embed.proj.weight", "vit.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("patch_embed.proj.bias", "vit.embeddings.patch_embeddings.projection.bias") )
# backbone
rename_keys.append(("patch_embed.backbone.stem.conv.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.convolution.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.bias", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.bias") )
for stage_idx in range(len(config.backbone_config.depths ) ):
for layer_idx in range(config.backbone_config.depths[stage_idx] ):
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.conv.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.conv.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.bias''') )
# transformer encoder
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''vit.encoder.layer.{i}.output.dense.bias''') )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("norm.weight", "layernorm.weight"),
("norm.bias", "layernorm.bias"),
("pre_logits.fc.weight", "pooler.dense.weight"),
("pre_logits.fc.bias", "pooler.dense.bias"),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
__magic_name__ : int = [(pair[0], pair[1][4:]) if pair[1].startswith("vit" ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
("norm.weight", "vit.layernorm.weight"),
("norm.bias", "vit.layernorm.bias"),
("head.weight", "classifier.weight"),
("head.bias", "classifier.bias"),
] )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Any , _snake_case : Dict=False ) -> int:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
__magic_name__ : int = ""
else:
__magic_name__ : Union[str, Any] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__magic_name__ : Optional[Any] = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
__magic_name__ : int = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__magic_name__ : Dict = in_proj_weight[
: config.hidden_size, :
]
__magic_name__ : List[str] = in_proj_bias[: config.hidden_size]
__magic_name__ : List[str] = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__magic_name__ : Optional[Any] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__magic_name__ : Optional[Any] = in_proj_weight[
-config.hidden_size :, :
]
__magic_name__ : int = in_proj_bias[-config.hidden_size :]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
__magic_name__ : List[str] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : int , _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : int = dct.pop(_snake_case )
__magic_name__ : List[Any] = val
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
__magic_name__ : List[str] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Any , _snake_case : int=False ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=_snake_case , )
__magic_name__ : List[str] = ViTHybridConfig(backbone_config=_snake_case , image_size=384 , num_labels=1000 )
__magic_name__ : str = False
# load original model from timm
__magic_name__ : Union[str, Any] = timm.create_model(_snake_case , pretrained=_snake_case )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__magic_name__ : List[Any] = timm_model.state_dict()
if base_model:
remove_classification_head_(_snake_case )
__magic_name__ : Tuple = create_rename_keys(_snake_case , _snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
read_in_q_k_v(_snake_case , _snake_case , _snake_case )
__magic_name__ : List[str] = "huggingface/label-files"
__magic_name__ : int = "imagenet-1k-id2label.json"
__magic_name__ : Optional[int] = json.load(open(hf_hub_download(_snake_case , _snake_case , repo_type="dataset" ) , "r" ) )
__magic_name__ : int = {int(_snake_case ): v for k, v in idalabel.items()}
__magic_name__ : List[str] = idalabel
__magic_name__ : List[str] = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
__magic_name__ : List[str] = ViTHybridModel(_snake_case ).eval()
else:
__magic_name__ : str = ViTHybridForImageClassification(_snake_case ).eval()
model.load_state_dict(_snake_case )
# create image processor
__magic_name__ : List[Any] = create_transform(**resolve_data_config({} , model=_snake_case ) )
__magic_name__ : int = transform.transforms
__magic_name__ : List[str] = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
__magic_name__ : int = ViTHybridImageProcessor(
do_resize=_snake_case , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=_snake_case , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=_snake_case , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
__magic_name__ : List[Any] = prepare_img()
__magic_name__ : Any = transform(_snake_case ).unsqueeze(0 )
__magic_name__ : Tuple = processor(_snake_case , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(_snake_case , _snake_case )
# verify logits
with torch.no_grad():
__magic_name__ : Optional[int] = model(_snake_case )
__magic_name__ : List[str] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
__magic_name__ : List[str] = timm_model.forward_features(_snake_case )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(_snake_case , outputs.pooler_output , atol=1E-3 )
else:
__magic_name__ : Any = timm_model(_snake_case )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_snake_case , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(_snake_case ).mkdir(exist_ok=_snake_case )
print(F'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(_snake_case )
print(F'''Saving processor to {pytorch_dump_folder_path}''' )
processor.save_pretrained(_snake_case )
if push_to_hub:
print(F'''Pushing model and processor to the hub {vit_name}''' )
model.push_to_hub(F'''ybelkada/{vit_name}''' )
processor.push_to_hub(F'''ybelkada/{vit_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_r50_s16_384",
type=str,
help="Name of the hybrid ViT timm model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether to upload the model to the HuggingFace hub."
)
snake_case : List[Any] = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
import os
snake_case : List[Any] = {"I": 1, "V": 5, "X": 10, "L": 50, "C": 100, "D": 500, "M": 1_000}
def lowerCAmelCase_ ( _snake_case : str ) -> int:
'''simple docstring'''
__magic_name__ : Tuple = 0
__magic_name__ : Union[str, Any] = 0
while index < len(_snake_case ) - 1:
__magic_name__ : Any = SYMBOLS[numerals[index]]
__magic_name__ : str = SYMBOLS[numerals[index + 1]]
if current_value < next_value:
total_value -= current_value
else:
total_value += current_value
index += 1
total_value += SYMBOLS[numerals[index]]
return total_value
def lowerCAmelCase_ ( _snake_case : int ) -> str:
'''simple docstring'''
__magic_name__ : str = ""
__magic_name__ : List[Any] = num // 1000
numerals += m_count * "M"
num %= 1000
__magic_name__ : Optional[int] = num // 100
if c_count == 9:
numerals += "CM"
c_count -= 9
elif c_count == 4:
numerals += "CD"
c_count -= 4
if c_count >= 5:
numerals += "D"
c_count -= 5
numerals += c_count * "C"
num %= 100
__magic_name__ : Optional[int] = num // 10
if x_count == 9:
numerals += "XC"
x_count -= 9
elif x_count == 4:
numerals += "XL"
x_count -= 4
if x_count >= 5:
numerals += "L"
x_count -= 5
numerals += x_count * "X"
num %= 10
if num == 9:
numerals += "IX"
num -= 9
elif num == 4:
numerals += "IV"
num -= 4
if num >= 5:
numerals += "V"
num -= 5
numerals += num * "I"
return numerals
def lowerCAmelCase_ ( _snake_case : str = "/p089_roman.txt" ) -> int:
'''simple docstring'''
__magic_name__ : int = 0
with open(os.path.dirname(_snake_case ) + roman_numerals_filename ) as filea:
__magic_name__ : List[Any] = filea.readlines()
for line in lines:
__magic_name__ : Tuple = line.strip()
__magic_name__ : List[Any] = parse_roman_numerals(_snake_case )
__magic_name__ : List[str] = generate_roman_numerals(_snake_case )
savings += len(_snake_case ) - len(_snake_case )
return savings
if __name__ == "__main__":
print(F"{solution() = }")
| 281 |
# This script creates a super tiny model that is useful inside tests, when we just want to test that
# the machinery works, without needing to the check the quality of the outcomes.
#
# This version creates a tiny model through reduction of a normal pre-trained model, but keeping the
# full vocab, merges file, and thus also resulting in a larger model due to a large vocab size.
# This gives ~3MB in total for all files.
#
# If you want a 50 times smaller than this see `fsmt-make-super-tiny-model.py`, which is slightly more complicated
#
#
# It will be used then as "stas/tiny-wmt19-en-de"
# Build
from transformers import FSMTTokenizer, FSMTConfig, FSMTForConditionalGeneration
snake_case : List[str] = "facebook/wmt19-en-de"
snake_case : Dict = FSMTTokenizer.from_pretrained(mname)
# get the correct vocab sizes, etc. from the master model
snake_case : List[str] = FSMTConfig.from_pretrained(mname)
config.update(
dict(
d_model=4,
encoder_layers=1,
decoder_layers=1,
encoder_ffn_dim=4,
decoder_ffn_dim=4,
encoder_attention_heads=1,
decoder_attention_heads=1,
)
)
snake_case : int = FSMTForConditionalGeneration(config)
print(F"num of params {tiny_model.num_parameters()}")
# Test
snake_case : Optional[Any] = tokenizer(["Making tiny model"], return_tensors="pt")
snake_case : List[str] = tiny_model(**batch)
print("test output:", len(outputs.logits[0]))
# Save
snake_case : Dict = "tiny-wmt19-en-de"
tiny_model.half() # makes it smaller
tiny_model.save_pretrained(mname_tiny)
tokenizer.save_pretrained(mname_tiny)
print(F"Generated {mname_tiny}")
# Upload
# transformers-cli upload tiny-wmt19-en-de
| 281 | 1 |
from pathlib import Path
from typing import List
from transformers import is_torch_available, is_vision_available
from transformers.testing_utils import get_tests_dir, is_tool_test
from transformers.tools.agent_types import AGENT_TYPE_MAPPING, AgentAudio, AgentImage, AgentText
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
snake_case : List[str] = ["text", "image", "audio"]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
for input_type in input_types:
if input_type == "text":
inputs.append("Text input" )
elif input_type == "image":
inputs.append(
Image.open(Path(get_tests_dir("fixtures/tests_samples/COCO" ) ) / "000000039769.png" ).resize((512, 512) ) )
elif input_type == "audio":
inputs.append(torch.ones(3000 ) )
elif isinstance(_snake_case , _snake_case ):
inputs.append(create_inputs(_snake_case ) )
else:
raise ValueError(F'''Invalid type requested: {input_type}''' )
return inputs
def lowerCAmelCase_ ( _snake_case : List ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
for output in outputs:
if isinstance(_snake_case , (str, AgentText) ):
output_types.append("text" )
elif isinstance(_snake_case , (Image.Image, AgentImage) ):
output_types.append("image" )
elif isinstance(_snake_case , (torch.Tensor, AgentAudio) ):
output_types.append("audio" )
else:
raise ValueError(F'''Invalid output: {output}''' )
return output_types
@is_tool_test
class _snake_case :
def SCREAMING_SNAKE_CASE ( self ):
self.assertTrue(hasattr(self.tool , "inputs" ) )
self.assertTrue(hasattr(self.tool , "outputs" ) )
__magic_name__ : Union[str, Any] = self.tool.inputs
for _input in inputs:
if isinstance(_input , _a ):
for __input in _input:
self.assertTrue(__input in authorized_types )
else:
self.assertTrue(_input in authorized_types )
__magic_name__ : List[Any] = self.tool.outputs
for _output in outputs:
self.assertTrue(_output in authorized_types )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = create_inputs(self.tool.inputs )
__magic_name__ : List[str] = self.tool(*_a )
# There is a single output
if len(self.tool.outputs ) == 1:
__magic_name__ : Any = [outputs]
self.assertListEqual(output_types(_a ) , self.tool.outputs )
def SCREAMING_SNAKE_CASE ( self ):
self.assertTrue(hasattr(self.tool , "description" ) )
self.assertTrue(hasattr(self.tool , "default_checkpoint" ) )
self.assertTrue(self.tool.description.startswith("This is a tool that" ) )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = create_inputs(self.tool.inputs )
__magic_name__ : Dict = self.tool(*_a )
if not isinstance(_a , _a ):
__magic_name__ : List[Any] = [outputs]
self.assertEqual(len(_a ) , len(self.tool.outputs ) )
for output, output_type in zip(_a , self.tool.outputs ):
__magic_name__ : str = AGENT_TYPE_MAPPING[output_type]
self.assertTrue(isinstance(_a , _a ) )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = create_inputs(self.tool.inputs )
__magic_name__ : Optional[Any] = []
for _input, input_type in zip(_a , self.tool.inputs ):
if isinstance(_a , _a ):
_inputs.append([AGENT_TYPE_MAPPING[_input_type](_input ) for _input_type in input_type] )
else:
_inputs.append(AGENT_TYPE_MAPPING[input_type](_input ) )
# Should not raise an error
__magic_name__ : List[Any] = self.tool(*_a )
if not isinstance(_a , _a ):
__magic_name__ : str = [outputs]
self.assertEqual(len(_a ) , len(self.tool.outputs ) )
| 281 |
import argparse
import csv
import logging
import os
import random
import numpy as np
import torch
from torch.utils.data import DataLoader, RandomSampler, SequentialSampler, TensorDataset
from tqdm import tqdm, trange
from transformers import (
CONFIG_NAME,
WEIGHTS_NAME,
AdamW,
OpenAIGPTDoubleHeadsModel,
OpenAIGPTTokenizer,
get_linear_schedule_with_warmup,
)
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s", datefmt="%m/%d/%Y %H:%M:%S", level=logging.INFO
)
snake_case : Optional[int] = logging.getLogger(__name__)
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : List[str] = np.argmax(_snake_case , axis=1 )
return np.sum(outputs == labels )
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
with open(_snake_case , encoding="utf_8" ) as f:
__magic_name__ : List[str] = csv.reader(_snake_case )
__magic_name__ : List[Any] = []
next(_snake_case ) # skip the first line
for line in tqdm(_snake_case ):
output.append((" ".join(line[1:5] ), line[5], line[6], int(line[-1] ) - 1) )
return output
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Tuple , _snake_case : Union[str, Any] , _snake_case : List[Any] , _snake_case : Tuple , _snake_case : Optional[int] ) -> int:
'''simple docstring'''
__magic_name__ : Optional[int] = []
for dataset in encoded_datasets:
__magic_name__ : Union[str, Any] = len(_snake_case )
__magic_name__ : Dict = np.zeros((n_batch, 2, input_len) , dtype=np.intaa )
__magic_name__ : List[str] = np.zeros((n_batch, 2) , dtype=np.intaa )
__magic_name__ : Optional[int] = np.full((n_batch, 2, input_len) , fill_value=-100 , dtype=np.intaa )
__magic_name__ : int = np.zeros((n_batch,) , dtype=np.intaa )
for (
i,
(story, conta, conta, mc_label),
) in enumerate(_snake_case ):
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : str = with_conta
__magic_name__ : Tuple = with_conta
__magic_name__ : Union[str, Any] = len(_snake_case ) - 1
__magic_name__ : int = len(_snake_case ) - 1
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[int] = mc_label
__magic_name__ : str = (input_ids, mc_token_ids, lm_labels, mc_labels)
tensor_datasets.append(tuple(torch.tensor(_snake_case ) for t in all_inputs ) )
return tensor_datasets
def lowerCAmelCase_ ( ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser()
parser.add_argument("--model_name" , type=_snake_case , default="openai-gpt" , help="pretrained model name" )
parser.add_argument("--do_train" , action="store_true" , help="Whether to run training." )
parser.add_argument("--do_eval" , action="store_true" , help="Whether to run eval on the dev set." )
parser.add_argument(
"--output_dir" , default=_snake_case , type=_snake_case , required=_snake_case , help="The output directory where the model predictions and checkpoints will be written." , )
parser.add_argument("--train_dataset" , type=_snake_case , default="" )
parser.add_argument("--eval_dataset" , type=_snake_case , default="" )
parser.add_argument("--seed" , type=_snake_case , default=42 )
parser.add_argument("--num_train_epochs" , type=_snake_case , default=3 )
parser.add_argument("--train_batch_size" , type=_snake_case , default=8 )
parser.add_argument("--eval_batch_size" , type=_snake_case , default=16 )
parser.add_argument("--adam_epsilon" , default=1E-8 , type=_snake_case , help="Epsilon for Adam optimizer." )
parser.add_argument("--max_grad_norm" , type=_snake_case , default=1 )
parser.add_argument(
"--max_steps" , default=-1 , type=_snake_case , help=(
"If > 0: set total number of training steps to perform. Override num_train_epochs."
) , )
parser.add_argument(
"--gradient_accumulation_steps" , type=_snake_case , default=1 , help="Number of updates steps to accumulate before performing a backward/update pass." , )
parser.add_argument("--learning_rate" , type=_snake_case , default=6.25E-5 )
parser.add_argument("--warmup_steps" , default=0 , type=_snake_case , help="Linear warmup over warmup_steps." )
parser.add_argument("--lr_schedule" , type=_snake_case , default="warmup_linear" )
parser.add_argument("--weight_decay" , type=_snake_case , default=0.01 )
parser.add_argument("--lm_coef" , type=_snake_case , default=0.9 )
parser.add_argument("--n_valid" , type=_snake_case , default=374 )
parser.add_argument("--server_ip" , type=_snake_case , default="" , help="Can be used for distant debugging." )
parser.add_argument("--server_port" , type=_snake_case , default="" , help="Can be used for distant debugging." )
__magic_name__ : List[Any] = parser.parse_args()
print(_snake_case )
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach" )
ptvsd.enable_attach(address=(args.server_ip, args.server_port) , redirect_output=_snake_case )
ptvsd.wait_for_attach()
random.seed(args.seed )
np.random.seed(args.seed )
torch.manual_seed(args.seed )
torch.cuda.manual_seed_all(args.seed )
__magic_name__ : Dict = torch.device("cuda" if torch.cuda.is_available() else "cpu" )
__magic_name__ : Optional[int] = torch.cuda.device_count()
logger.info("device: {}, n_gpu {}".format(_snake_case , _snake_case ) )
if not args.do_train and not args.do_eval:
raise ValueError("At least one of `do_train` or `do_eval` must be True." )
if not os.path.exists(args.output_dir ):
os.makedirs(args.output_dir )
# Load tokenizer and model
# This loading functions also add new tokens and embeddings called `special tokens`
# These new embeddings will be fine-tuned on the RocStories dataset
__magic_name__ : List[Any] = ["_start_", "_delimiter_", "_classify_"]
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.model_name )
tokenizer.add_tokens(_snake_case )
__magic_name__ : Optional[Any] = tokenizer.convert_tokens_to_ids(_snake_case )
__magic_name__ : List[str] = OpenAIGPTDoubleHeadsModel.from_pretrained(args.model_name )
model.resize_token_embeddings(len(_snake_case ) )
model.to(_snake_case )
# Load and encode the datasets
def tokenize_and_encode(_snake_case : str ):
if isinstance(_snake_case , _snake_case ):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(_snake_case ) )
elif isinstance(_snake_case , _snake_case ):
return obj
return [tokenize_and_encode(_snake_case ) for o in obj]
logger.info("Encoding dataset..." )
__magic_name__ : Optional[int] = load_rocstories_dataset(args.train_dataset )
__magic_name__ : str = load_rocstories_dataset(args.eval_dataset )
__magic_name__ : int = (train_dataset, eval_dataset)
__magic_name__ : List[str] = tokenize_and_encode(_snake_case )
# Compute the max input length for the Transformer
__magic_name__ : Optional[Any] = model.config.n_positions // 2 - 2
__magic_name__ : Optional[int] = max(
len(story[:max_length] ) + max(len(conta[:max_length] ) , len(conta[:max_length] ) ) + 3
for dataset in encoded_datasets
for story, conta, conta, _ in dataset )
__magic_name__ : List[str] = min(_snake_case , model.config.n_positions ) # Max size of input for the pre-trained model
# Prepare inputs tensors and dataloaders
__magic_name__ : List[Any] = pre_process_datasets(_snake_case , _snake_case , _snake_case , *_snake_case )
__magic_name__ , __magic_name__ : Optional[int] = tensor_datasets[0], tensor_datasets[1]
__magic_name__ : Tuple = TensorDataset(*_snake_case )
__magic_name__ : Union[str, Any] = RandomSampler(_snake_case )
__magic_name__ : Dict = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.train_batch_size )
__magic_name__ : Any = TensorDataset(*_snake_case )
__magic_name__ : Optional[Any] = SequentialSampler(_snake_case )
__magic_name__ : int = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.eval_batch_size )
# Prepare optimizer
if args.do_train:
if args.max_steps > 0:
__magic_name__ : Tuple = args.max_steps
__magic_name__ : List[str] = args.max_steps // (len(_snake_case ) // args.gradient_accumulation_steps) + 1
else:
__magic_name__ : List[str] = len(_snake_case ) // args.gradient_accumulation_steps * args.num_train_epochs
__magic_name__ : str = list(model.named_parameters() )
__magic_name__ : Dict = ["bias", "LayerNorm.bias", "LayerNorm.weight"]
__magic_name__ : str = [
{
"params": [p for n, p in param_optimizer if not any(nd in n for nd in no_decay )],
"weight_decay": args.weight_decay,
},
{"params": [p for n, p in param_optimizer if any(nd in n for nd in no_decay )], "weight_decay": 0.0},
]
__magic_name__ : str = AdamW(_snake_case , lr=args.learning_rate , eps=args.adam_epsilon )
__magic_name__ : List[str] = get_linear_schedule_with_warmup(
_snake_case , num_warmup_steps=args.warmup_steps , num_training_steps=_snake_case )
if args.do_train:
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = 0, 0, None
model.train()
for _ in trange(int(args.num_train_epochs ) , desc="Epoch" ):
__magic_name__ : List[str] = 0
__magic_name__ : Tuple = 0
__magic_name__ : Dict = tqdm(_snake_case , desc="Training" )
for step, batch in enumerate(_snake_case ):
__magic_name__ : Optional[Any] = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = batch
__magic_name__ : Optional[Any] = model(_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Optional[Any] = args.lm_coef * losses[0] + losses[1]
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
tr_loss += loss.item()
__magic_name__ : List[str] = (
loss.item() if exp_average_loss is None else 0.7 * exp_average_loss + 0.3 * loss.item()
)
nb_tr_steps += 1
__magic_name__ : int = "Training loss: {:.2e} lr: {:.2e}".format(_snake_case , scheduler.get_lr()[0] )
# Save a trained model
if args.do_train:
# Save a trained model, configuration and tokenizer
__magic_name__ : Dict = model.module if hasattr(_snake_case , "module" ) else model # Only save the model itself
# If we save using the predefined names, we can load using `from_pretrained`
__magic_name__ : List[Any] = os.path.join(args.output_dir , _snake_case )
__magic_name__ : Dict = os.path.join(args.output_dir , _snake_case )
torch.save(model_to_save.state_dict() , _snake_case )
model_to_save.config.to_json_file(_snake_case )
tokenizer.save_vocabulary(args.output_dir )
# Load a trained model and vocabulary that you have fine-tuned
__magic_name__ : Dict = OpenAIGPTDoubleHeadsModel.from_pretrained(args.output_dir )
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.output_dir )
model.to(_snake_case )
if args.do_eval:
model.eval()
__magic_name__ , __magic_name__ : Any = 0, 0
__magic_name__ , __magic_name__ : Union[str, Any] = 0, 0
for batch in tqdm(_snake_case , desc="Evaluating" ):
__magic_name__ : int = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = batch
with torch.no_grad():
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = model(
_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Tuple = mc_logits.detach().cpu().numpy()
__magic_name__ : Any = mc_labels.to("cpu" ).numpy()
__magic_name__ : str = accuracy(_snake_case , _snake_case )
eval_loss += mc_loss.mean().item()
eval_accuracy += tmp_eval_accuracy
nb_eval_examples += input_ids.size(0 )
nb_eval_steps += 1
__magic_name__ : Tuple = eval_loss / nb_eval_steps
__magic_name__ : List[Any] = eval_accuracy / nb_eval_examples
__magic_name__ : int = tr_loss / nb_tr_steps if args.do_train else None
__magic_name__ : Any = {"eval_loss": eval_loss, "eval_accuracy": eval_accuracy, "train_loss": train_loss}
__magic_name__ : int = os.path.join(args.output_dir , "eval_results.txt" )
with open(_snake_case , "w" ) as writer:
logger.info("***** Eval results *****" )
for key in sorted(result.keys() ):
logger.info(" %s = %s" , _snake_case , str(result[key] ) )
writer.write("%s = %s\n" % (key, str(result[key] )) )
if __name__ == "__main__":
main()
| 281 | 1 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
snake_case : str = logging.get_logger(__name__)
snake_case : int = {
"microsoft/swinv2-tiny-patch4-window8-256": (
"https://huggingface.co/microsoft/swinv2-tiny-patch4-window8-256/resolve/main/config.json"
),
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'swinv2'
UpperCamelCase__ = {
'num_attention_heads': 'num_heads',
'num_hidden_layers': 'num_layers',
}
def __init__( self , _a=224 , _a=4 , _a=3 , _a=96 , _a=[2, 2, 6, 2] , _a=[3, 6, 12, 24] , _a=7 , _a=4.0 , _a=True , _a=0.0 , _a=0.0 , _a=0.1 , _a="gelu" , _a=False , _a=0.02 , _a=1e-5 , _a=32 , **_a , ):
super().__init__(**_a )
__magic_name__ : Tuple = image_size
__magic_name__ : str = patch_size
__magic_name__ : Optional[int] = num_channels
__magic_name__ : Tuple = embed_dim
__magic_name__ : Optional[Any] = depths
__magic_name__ : int = len(_a )
__magic_name__ : str = num_heads
__magic_name__ : str = window_size
__magic_name__ : Optional[int] = mlp_ratio
__magic_name__ : Optional[Any] = qkv_bias
__magic_name__ : int = hidden_dropout_prob
__magic_name__ : int = attention_probs_dropout_prob
__magic_name__ : Tuple = drop_path_rate
__magic_name__ : Union[str, Any] = hidden_act
__magic_name__ : Optional[int] = use_absolute_embeddings
__magic_name__ : str = layer_norm_eps
__magic_name__ : str = initializer_range
__magic_name__ : Optional[int] = encoder_stride
# we set the hidden_size attribute in order to make Swinv2 work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
__magic_name__ : List[Any] = int(embed_dim * 2 ** (len(_a ) - 1) )
__magic_name__ : Tuple = (0, 0, 0, 0)
| 281 |
from . import __version__
# Backward compatibility imports, to make sure all those objects can be found in file_utils
from .utils import (
CLOUDFRONT_DISTRIB_PREFIX,
CONFIG_NAME,
DISABLE_TELEMETRY,
DUMMY_INPUTS,
DUMMY_MASK,
ENV_VARS_TRUE_AND_AUTO_VALUES,
ENV_VARS_TRUE_VALUES,
FEATURE_EXTRACTOR_NAME,
FLAX_WEIGHTS_NAME,
HF_MODULES_CACHE,
HUGGINGFACE_CO_PREFIX,
HUGGINGFACE_CO_RESOLVE_ENDPOINT,
MODEL_CARD_NAME,
MULTIPLE_CHOICE_DUMMY_INPUTS,
PYTORCH_PRETRAINED_BERT_CACHE,
PYTORCH_TRANSFORMERS_CACHE,
S3_BUCKET_PREFIX,
SENTENCEPIECE_UNDERLINE,
SPIECE_UNDERLINE,
TF2_WEIGHTS_NAME,
TF_WEIGHTS_NAME,
TORCH_FX_REQUIRED_VERSION,
TRANSFORMERS_CACHE,
TRANSFORMERS_DYNAMIC_MODULE_NAME,
USE_JAX,
USE_TF,
USE_TORCH,
WEIGHTS_INDEX_NAME,
WEIGHTS_NAME,
ContextManagers,
DummyObject,
EntryNotFoundError,
ExplicitEnum,
ModelOutput,
PaddingStrategy,
PushToHubMixin,
RepositoryNotFoundError,
RevisionNotFoundError,
TensorType,
_LazyModule,
add_code_sample_docstrings,
add_end_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
cached_property,
copy_func,
default_cache_path,
define_sagemaker_information,
get_cached_models,
get_file_from_repo,
get_full_repo_name,
get_torch_version,
has_file,
http_user_agent,
is_apex_available,
is_bsa_available,
is_coloredlogs_available,
is_datasets_available,
is_detectrona_available,
is_faiss_available,
is_flax_available,
is_ftfy_available,
is_in_notebook,
is_ipex_available,
is_librosa_available,
is_offline_mode,
is_onnx_available,
is_pandas_available,
is_phonemizer_available,
is_protobuf_available,
is_psutil_available,
is_pyanvml_available,
is_pyctcdecode_available,
is_pytesseract_available,
is_pytorch_quantization_available,
is_rjieba_available,
is_sagemaker_dp_enabled,
is_sagemaker_mp_enabled,
is_scipy_available,
is_sentencepiece_available,
is_seqio_available,
is_sklearn_available,
is_soundfile_availble,
is_spacy_available,
is_speech_available,
is_tensor,
is_tensorflow_probability_available,
is_tfaonnx_available,
is_tf_available,
is_timm_available,
is_tokenizers_available,
is_torch_available,
is_torch_bfaa_available,
is_torch_cuda_available,
is_torch_fx_available,
is_torch_fx_proxy,
is_torch_mps_available,
is_torch_tfaa_available,
is_torch_tpu_available,
is_torchaudio_available,
is_training_run_on_sagemaker,
is_vision_available,
replace_return_docstrings,
requires_backends,
to_numpy,
to_py_obj,
torch_only_method,
)
| 281 | 1 |
import gc
import random
import tempfile
import unittest
import numpy as np
import torch
from PIL import Image
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMInverseScheduler,
DDIMScheduler,
DPMSolverMultistepInverseScheduler,
DPMSolverMultistepScheduler,
StableDiffusionDiffEditPipeline,
UNetaDConditionModel,
)
from diffusers.utils import load_image, slow
from diffusers.utils.testing_utils import enable_full_determinism, floats_tensor, require_torch_gpu, torch_device
from ..pipeline_params import TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS, TEXT_GUIDED_IMAGE_INPAINTING_PARAMS
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = StableDiffusionDiffEditPipeline
UpperCamelCase__ = TEXT_GUIDED_IMAGE_INPAINTING_PARAMS - {'height', 'width', 'image'} | {'image_latents'}
UpperCamelCase__ = TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS - {'image'} | {'image_latents'}
UpperCamelCase__ = frozenset(
[] ) # TO-DO: update image_params once pipeline is refactored with VaeImageProcessor.preprocess
UpperCamelCase__ = frozenset([] )
def SCREAMING_SNAKE_CASE ( self ):
torch.manual_seed(0 )
__magic_name__ : List[str] = 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 , attention_head_dim=(2, 4) , use_linear_projection=_a , )
__magic_name__ : Optional[Any] = DDIMScheduler(
beta_start=0.0_00_85 , beta_end=0.0_12 , beta_schedule="scaled_linear" , clip_sample=_a , set_alpha_to_one=_a , )
__magic_name__ : Tuple = DDIMInverseScheduler(
beta_start=0.0_00_85 , beta_end=0.0_12 , beta_schedule="scaled_linear" , clip_sample=_a , set_alpha_to_zero=_a , )
torch.manual_seed(0 )
__magic_name__ : List[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 , sample_size=128 , )
torch.manual_seed(0 )
__magic_name__ : List[Any] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1e-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , hidden_act="gelu" , projection_dim=512 , )
__magic_name__ : Dict = CLIPTextModel(_a )
__magic_name__ : List[Any] = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip" )
__magic_name__ : Optional[Any] = {
"unet": unet,
"scheduler": scheduler,
"inverse_scheduler": inverse_scheduler,
"vae": vae,
"text_encoder": text_encoder,
"tokenizer": tokenizer,
"safety_checker": None,
"feature_extractor": None,
}
return components
def SCREAMING_SNAKE_CASE ( self , _a , _a=0 ):
__magic_name__ : Optional[int] = floats_tensor((1, 16, 16) , rng=random.Random(_a ) ).to(_a )
__magic_name__ : Union[str, Any] = floats_tensor((1, 2, 4, 16, 16) , rng=random.Random(_a ) ).to(_a )
if str(_a ).startswith("mps" ):
__magic_name__ : List[str] = torch.manual_seed(_a )
else:
__magic_name__ : Optional[int] = torch.Generator(device=_a ).manual_seed(_a )
__magic_name__ : Dict = {
"prompt": "a dog and a newt",
"mask_image": mask,
"image_latents": latents,
"generator": generator,
"num_inference_steps": 2,
"inpaint_strength": 1.0,
"guidance_scale": 6.0,
"output_type": "numpy",
}
return inputs
def SCREAMING_SNAKE_CASE ( self , _a , _a=0 ):
__magic_name__ : List[str] = floats_tensor((1, 3, 32, 32) , rng=random.Random(_a ) ).to(_a )
__magic_name__ : Optional[int] = image.cpu().permute(0 , 2 , 3 , 1 )[0]
__magic_name__ : List[Any] = Image.fromarray(np.uinta(_a ) ).convert("RGB" )
if str(_a ).startswith("mps" ):
__magic_name__ : List[str] = torch.manual_seed(_a )
else:
__magic_name__ : Tuple = torch.Generator(device=_a ).manual_seed(_a )
__magic_name__ : List[Any] = {
"image": image,
"source_prompt": "a cat and a frog",
"target_prompt": "a dog and a newt",
"generator": generator,
"num_inference_steps": 2,
"num_maps_per_mask": 2,
"mask_encode_strength": 1.0,
"guidance_scale": 6.0,
"output_type": "numpy",
}
return inputs
def SCREAMING_SNAKE_CASE ( self , _a , _a=0 ):
__magic_name__ : str = floats_tensor((1, 3, 32, 32) , rng=random.Random(_a ) ).to(_a )
__magic_name__ : Tuple = image.cpu().permute(0 , 2 , 3 , 1 )[0]
__magic_name__ : str = Image.fromarray(np.uinta(_a ) ).convert("RGB" )
if str(_a ).startswith("mps" ):
__magic_name__ : int = torch.manual_seed(_a )
else:
__magic_name__ : Dict = torch.Generator(device=_a ).manual_seed(_a )
__magic_name__ : Tuple = {
"image": image,
"prompt": "a cat and a frog",
"generator": generator,
"num_inference_steps": 2,
"inpaint_strength": 1.0,
"guidance_scale": 6.0,
"decode_latents": True,
"output_type": "numpy",
}
return inputs
def SCREAMING_SNAKE_CASE ( self ):
if not hasattr(self.pipeline_class , "_optional_components" ):
return
__magic_name__ : Dict = self.get_dummy_components()
__magic_name__ : Any = self.pipeline_class(**_a )
pipe.to(_a )
pipe.set_progress_bar_config(disable=_a )
# set all optional components to None and update pipeline config accordingly
for optional_component in pipe._optional_components:
setattr(_a , _a , _a )
pipe.register_modules(**{optional_component: None for optional_component in pipe._optional_components} )
__magic_name__ : List[Any] = self.get_dummy_inputs(_a )
__magic_name__ : int = pipe(**_a )[0]
with tempfile.TemporaryDirectory() as tmpdir:
pipe.save_pretrained(_a )
__magic_name__ : Optional[int] = self.pipeline_class.from_pretrained(_a )
pipe_loaded.to(_a )
pipe_loaded.set_progress_bar_config(disable=_a )
for optional_component in pipe._optional_components:
self.assertTrue(
getattr(_a , _a ) is None , f'''`{optional_component}` did not stay set to None after loading.''' , )
__magic_name__ : str = self.get_dummy_inputs(_a )
__magic_name__ : Optional[int] = pipe_loaded(**_a )[0]
__magic_name__ : Dict = np.abs(output - output_loaded ).max()
self.assertLess(_a , 1e-4 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = "cpu"
__magic_name__ : List[Any] = self.get_dummy_components()
__magic_name__ : Optional[int] = self.pipeline_class(**_a )
pipe.to(_a )
pipe.set_progress_bar_config(disable=_a )
__magic_name__ : Optional[int] = self.get_dummy_mask_inputs(_a )
__magic_name__ : Optional[Any] = pipe.generate_mask(**_a )
__magic_name__ : List[Any] = mask[0, -3:, -3:]
self.assertEqual(mask.shape , (1, 16, 16) )
__magic_name__ : Tuple = np.array([0] * 9 )
__magic_name__ : Union[str, Any] = np.abs(mask_slice.flatten() - expected_slice ).max()
self.assertLessEqual(_a , 1e-3 )
self.assertEqual(mask[0, -3, -4] , 0 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "cpu"
__magic_name__ : Dict = self.get_dummy_components()
__magic_name__ : List[Any] = self.pipeline_class(**_a )
pipe.to(_a )
pipe.set_progress_bar_config(disable=_a )
__magic_name__ : str = self.get_dummy_inversion_inputs(_a )
__magic_name__ : Tuple = pipe.invert(**_a ).images
__magic_name__ : Union[str, Any] = image[0, -1, -3:, -3:]
self.assertEqual(image.shape , (2, 32, 32, 3) )
__magic_name__ : Union[str, Any] = np.array(
[0.51_50, 0.51_34, 0.50_43, 0.53_76, 0.46_94, 0.5_10_50, 0.50_15, 0.44_07, 0.47_99] , )
__magic_name__ : List[str] = np.abs(image_slice.flatten() - expected_slice ).max()
self.assertLessEqual(_a , 1e-3 )
def SCREAMING_SNAKE_CASE ( self ):
super().test_inference_batch_single_identical(expected_max_diff=5e-3 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = "cpu"
__magic_name__ : Any = self.get_dummy_components()
__magic_name__ : Any = {"beta_start": 0.0_00_85, "beta_end": 0.0_12, "beta_schedule": "scaled_linear"}
__magic_name__ : Optional[int] = DPMSolverMultistepScheduler(**_a )
__magic_name__ : List[str] = DPMSolverMultistepInverseScheduler(**_a )
__magic_name__ : Optional[Any] = self.pipeline_class(**_a )
pipe.to(_a )
pipe.set_progress_bar_config(disable=_a )
__magic_name__ : Optional[Any] = self.get_dummy_inversion_inputs(_a )
__magic_name__ : Optional[Any] = pipe.invert(**_a ).images
__magic_name__ : Union[str, Any] = image[0, -1, -3:, -3:]
self.assertEqual(image.shape , (2, 32, 32, 3) )
__magic_name__ : Dict = np.array(
[0.51_50, 0.51_34, 0.50_43, 0.53_76, 0.46_94, 0.5_10_50, 0.50_15, 0.44_07, 0.47_99] , )
__magic_name__ : Optional[Any] = np.abs(image_slice.flatten() - expected_slice ).max()
self.assertLessEqual(_a , 1e-3 )
@require_torch_gpu
@slow
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
@classmethod
def SCREAMING_SNAKE_CASE ( cls ):
__magic_name__ : Any = load_image(
"https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/diffedit/fruit.png" )
__magic_name__ : Any = raw_image.convert("RGB" ).resize((768, 768) )
__magic_name__ : Union[str, Any] = raw_image
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = torch.manual_seed(0 )
__magic_name__ : Dict = StableDiffusionDiffEditPipeline.from_pretrained(
"stabilityai/stable-diffusion-2-1" , safety_checker=_a , torch_dtype=torch.floataa )
__magic_name__ : int = DDIMScheduler.from_config(pipe.scheduler.config )
__magic_name__ : Optional[Any] = DDIMInverseScheduler.from_config(pipe.scheduler.config )
pipe.enable_model_cpu_offload()
pipe.set_progress_bar_config(disable=_a )
__magic_name__ : Dict = "a bowl of fruit"
__magic_name__ : Any = "a bowl of pears"
__magic_name__ : str = pipe.generate_mask(
image=self.raw_image , source_prompt=_a , target_prompt=_a , generator=_a , )
__magic_name__ : Dict = pipe.invert(
prompt=_a , image=self.raw_image , inpaint_strength=0.7 , generator=_a ).latents
__magic_name__ : Dict = pipe(
prompt=_a , mask_image=_a , image_latents=_a , generator=_a , negative_prompt=_a , inpaint_strength=0.7 , output_type="numpy" , ).images[0]
__magic_name__ : Optional[Any] = (
np.array(
load_image(
"https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main"
"/diffedit/pears.png" ).resize((768, 768) ) )
/ 255
)
assert np.abs((expected_image - image).max() ) < 5e-1
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = torch.manual_seed(0 )
__magic_name__ : Optional[Any] = StableDiffusionDiffEditPipeline.from_pretrained(
"stabilityai/stable-diffusion-2-1" , safety_checker=_a , torch_dtype=torch.floataa )
__magic_name__ : List[Any] = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config )
__magic_name__ : Optional[int] = DPMSolverMultistepInverseScheduler.from_config(pipe.scheduler.config )
pipe.enable_model_cpu_offload()
pipe.set_progress_bar_config(disable=_a )
__magic_name__ : str = "a bowl of fruit"
__magic_name__ : Tuple = "a bowl of pears"
__magic_name__ : Union[str, Any] = pipe.generate_mask(
image=self.raw_image , source_prompt=_a , target_prompt=_a , generator=_a , )
__magic_name__ : Optional[int] = pipe.invert(
prompt=_a , image=self.raw_image , inpaint_strength=0.7 , generator=_a , num_inference_steps=25 , ).latents
__magic_name__ : Optional[Any] = pipe(
prompt=_a , mask_image=_a , image_latents=_a , generator=_a , negative_prompt=_a , inpaint_strength=0.7 , num_inference_steps=25 , output_type="numpy" , ).images[0]
__magic_name__ : Any = (
np.array(
load_image(
"https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main"
"/diffedit/pears.png" ).resize((768, 768) ) )
/ 255
)
assert np.abs((expected_image - image).max() ) < 5e-1
| 281 |
import importlib
import os
import fsspec
import pytest
from fsspec import register_implementation
from fsspec.registry import _registry as _fsspec_registry
from datasets.filesystems import COMPRESSION_FILESYSTEMS, HfFileSystem, extract_path_from_uri, is_remote_filesystem
from .utils import require_lza, require_zstandard
def lowerCAmelCase_ ( _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
assert "mock" in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
assert "mock" not in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Dict = "mock-s3-bucket"
__magic_name__ : Any = F'''s3://{mock_bucket}'''
__magic_name__ : str = extract_path_from_uri(_snake_case )
assert dataset_path.startswith("s3://" ) is False
__magic_name__ : Tuple = "./local/path"
__magic_name__ : Optional[Any] = extract_path_from_uri(_snake_case )
assert dataset_path == new_dataset_path
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : str = is_remote_filesystem(_snake_case )
assert is_remote is True
__magic_name__ : Optional[int] = fsspec.filesystem("file" )
__magic_name__ : int = is_remote_filesystem(_snake_case )
assert is_remote is False
@pytest.mark.parametrize("compression_fs_class" , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : int , _snake_case : Tuple , _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Any ) -> int:
'''simple docstring'''
__magic_name__ : Any = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_file, "bz2": bza_file, "lz4": lza_file}
__magic_name__ : str = input_paths[compression_fs_class.protocol]
if input_path is None:
__magic_name__ : Dict = F'''for \'{compression_fs_class.protocol}\' compression protocol, '''
if compression_fs_class.protocol == "lz4":
reason += require_lza.kwargs["reason"]
elif compression_fs_class.protocol == "zstd":
reason += require_zstandard.kwargs["reason"]
pytest.skip(_snake_case )
__magic_name__ : str = fsspec.filesystem(compression_fs_class.protocol , fo=_snake_case )
assert isinstance(_snake_case , _snake_case )
__magic_name__ : int = os.path.basename(_snake_case )
__magic_name__ : Optional[int] = expected_filename[: expected_filename.rindex("." )]
assert fs.glob("*" ) == [expected_filename]
with fs.open(_snake_case , "r" , encoding="utf-8" ) as f, open(_snake_case , encoding="utf-8" ) as expected_file:
assert f.read() == expected_file.read()
@pytest.mark.parametrize("protocol" , ["zip", "gzip"] )
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : Optional[Any] , _snake_case : Optional[Any] ) -> str:
'''simple docstring'''
__magic_name__ : int = {"zip": zip_jsonl_path, "gzip": jsonl_gz_path}
__magic_name__ : int = compressed_file_paths[protocol]
__magic_name__ : Tuple = "dataset.jsonl"
__magic_name__ : List[str] = F'''{protocol}://{member_file_path}::{compressed_file_path}'''
__magic_name__ , *__magic_name__ : Optional[Any] = fsspec.get_fs_token_paths(_snake_case )
assert fs.isfile(_snake_case )
assert not fs.isfile("non_existing_" + member_file_path )
@pytest.mark.integration
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : List[str] , _snake_case : Tuple ) -> str:
'''simple docstring'''
__magic_name__ : int = hf_api.dataset_info(_snake_case , token=_snake_case )
__magic_name__ : Optional[Any] = HfFileSystem(repo_info=_snake_case , token=_snake_case )
assert sorted(hffs.glob("*" ) ) == [".gitattributes", "data"]
assert hffs.isdir("data" )
assert hffs.isfile(".gitattributes" ) and hffs.isfile("data/text_data.txt" )
with open(_snake_case ) as f:
assert hffs.open("data/text_data.txt" , "r" ).read() == f.read()
def lowerCAmelCase_ ( ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Optional[Any] = "bz2"
# Import module
import datasets.filesystems
# Overwrite protocol and reload
register_implementation(_snake_case , _snake_case , clobber=_snake_case )
with pytest.warns(_snake_case ) as warning_info:
importlib.reload(datasets.filesystems )
assert len(_snake_case ) == 1
assert (
str(warning_info[0].message )
== F'''A filesystem protocol was already set for {protocol} and will be overwritten.'''
)
| 281 | 1 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import (
BitConfig,
ViTHybridConfig,
ViTHybridForImageClassification,
ViTHybridImageProcessor,
ViTHybridModel,
)
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
snake_case : Optional[Any] = logging.get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any]=False ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
# fmt: off
# stem:
rename_keys.append(("cls_token", "vit.embeddings.cls_token") )
rename_keys.append(("pos_embed", "vit.embeddings.position_embeddings") )
rename_keys.append(("patch_embed.proj.weight", "vit.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("patch_embed.proj.bias", "vit.embeddings.patch_embeddings.projection.bias") )
# backbone
rename_keys.append(("patch_embed.backbone.stem.conv.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.convolution.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.bias", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.bias") )
for stage_idx in range(len(config.backbone_config.depths ) ):
for layer_idx in range(config.backbone_config.depths[stage_idx] ):
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.conv.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.conv.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.bias''') )
# transformer encoder
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''vit.encoder.layer.{i}.output.dense.bias''') )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("norm.weight", "layernorm.weight"),
("norm.bias", "layernorm.bias"),
("pre_logits.fc.weight", "pooler.dense.weight"),
("pre_logits.fc.bias", "pooler.dense.bias"),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
__magic_name__ : int = [(pair[0], pair[1][4:]) if pair[1].startswith("vit" ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
("norm.weight", "vit.layernorm.weight"),
("norm.bias", "vit.layernorm.bias"),
("head.weight", "classifier.weight"),
("head.bias", "classifier.bias"),
] )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Any , _snake_case : Dict=False ) -> int:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
__magic_name__ : int = ""
else:
__magic_name__ : Union[str, Any] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__magic_name__ : Optional[Any] = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
__magic_name__ : int = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__magic_name__ : Dict = in_proj_weight[
: config.hidden_size, :
]
__magic_name__ : List[str] = in_proj_bias[: config.hidden_size]
__magic_name__ : List[str] = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__magic_name__ : Optional[Any] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__magic_name__ : Optional[Any] = in_proj_weight[
-config.hidden_size :, :
]
__magic_name__ : int = in_proj_bias[-config.hidden_size :]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
__magic_name__ : List[str] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : int , _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : int = dct.pop(_snake_case )
__magic_name__ : List[Any] = val
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
__magic_name__ : List[str] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Any , _snake_case : int=False ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=_snake_case , )
__magic_name__ : List[str] = ViTHybridConfig(backbone_config=_snake_case , image_size=384 , num_labels=1000 )
__magic_name__ : str = False
# load original model from timm
__magic_name__ : Union[str, Any] = timm.create_model(_snake_case , pretrained=_snake_case )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__magic_name__ : List[Any] = timm_model.state_dict()
if base_model:
remove_classification_head_(_snake_case )
__magic_name__ : Tuple = create_rename_keys(_snake_case , _snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
read_in_q_k_v(_snake_case , _snake_case , _snake_case )
__magic_name__ : List[str] = "huggingface/label-files"
__magic_name__ : int = "imagenet-1k-id2label.json"
__magic_name__ : Optional[int] = json.load(open(hf_hub_download(_snake_case , _snake_case , repo_type="dataset" ) , "r" ) )
__magic_name__ : int = {int(_snake_case ): v for k, v in idalabel.items()}
__magic_name__ : List[str] = idalabel
__magic_name__ : List[str] = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
__magic_name__ : List[str] = ViTHybridModel(_snake_case ).eval()
else:
__magic_name__ : str = ViTHybridForImageClassification(_snake_case ).eval()
model.load_state_dict(_snake_case )
# create image processor
__magic_name__ : List[Any] = create_transform(**resolve_data_config({} , model=_snake_case ) )
__magic_name__ : int = transform.transforms
__magic_name__ : List[str] = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
__magic_name__ : int = ViTHybridImageProcessor(
do_resize=_snake_case , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=_snake_case , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=_snake_case , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
__magic_name__ : List[Any] = prepare_img()
__magic_name__ : Any = transform(_snake_case ).unsqueeze(0 )
__magic_name__ : Tuple = processor(_snake_case , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(_snake_case , _snake_case )
# verify logits
with torch.no_grad():
__magic_name__ : Optional[int] = model(_snake_case )
__magic_name__ : List[str] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
__magic_name__ : List[str] = timm_model.forward_features(_snake_case )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(_snake_case , outputs.pooler_output , atol=1E-3 )
else:
__magic_name__ : Any = timm_model(_snake_case )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_snake_case , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(_snake_case ).mkdir(exist_ok=_snake_case )
print(F'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(_snake_case )
print(F'''Saving processor to {pytorch_dump_folder_path}''' )
processor.save_pretrained(_snake_case )
if push_to_hub:
print(F'''Pushing model and processor to the hub {vit_name}''' )
model.push_to_hub(F'''ybelkada/{vit_name}''' )
processor.push_to_hub(F'''ybelkada/{vit_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_r50_s16_384",
type=str,
help="Name of the hybrid ViT timm model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether to upload the model to the HuggingFace hub."
)
snake_case : List[Any] = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : List[Any] = {
"YituTech/conv-bert-base": "https://huggingface.co/YituTech/conv-bert-base/resolve/main/config.json",
"YituTech/conv-bert-medium-small": (
"https://huggingface.co/YituTech/conv-bert-medium-small/resolve/main/config.json"
),
"YituTech/conv-bert-small": "https://huggingface.co/YituTech/conv-bert-small/resolve/main/config.json",
# See all ConvBERT models at https://huggingface.co/models?filter=convbert
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'convbert'
def __init__( self , _a=30_522 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=2 , _a=0.02 , _a=1e-12 , _a=1 , _a=0 , _a=2 , _a=768 , _a=2 , _a=9 , _a=1 , _a=None , **_a , ):
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , **_a , )
__magic_name__ : Tuple = vocab_size
__magic_name__ : List[Any] = hidden_size
__magic_name__ : Union[str, Any] = num_hidden_layers
__magic_name__ : List[Any] = num_attention_heads
__magic_name__ : str = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : Tuple = max_position_embeddings
__magic_name__ : str = type_vocab_size
__magic_name__ : List[str] = initializer_range
__magic_name__ : Tuple = layer_norm_eps
__magic_name__ : List[Any] = embedding_size
__magic_name__ : List[Any] = head_ratio
__magic_name__ : str = conv_kernel_size
__magic_name__ : Dict = num_groups
__magic_name__ : str = classifier_dropout
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : Dict = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
("token_type_ids", dynamic_axis),
] )
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : int , _snake_case : int ) -> int:
'''simple docstring'''
return 1 if input_a == input_a else 0
def lowerCAmelCase_ ( ) -> None:
'''simple docstring'''
assert xnor_gate(0 , 0 ) == 1
assert xnor_gate(0 , 1 ) == 0
assert xnor_gate(1 , 0 ) == 0
assert xnor_gate(1 , 1 ) == 1
if __name__ == "__main__":
print(xnor_gate(0, 0))
print(xnor_gate(0, 1))
print(xnor_gate(1, 0))
print(xnor_gate(1, 1))
| 281 |
import argparse
import requests
import torch
# pip3 install salesforce-lavis
# I'm actually installing a slightly modified version: pip3 install git+https://github.com/nielsrogge/LAVIS.git@fix_lavis
from lavis.models import load_model_and_preprocess
from PIL import Image
from transformers import (
AutoTokenizer,
BlipaConfig,
BlipaForConditionalGeneration,
BlipaProcessor,
BlipaVisionConfig,
BlipImageProcessor,
OPTConfig,
TaConfig,
)
from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD
def lowerCAmelCase_ ( ) -> str:
'''simple docstring'''
__magic_name__ : int = "https://storage.googleapis.com/sfr-vision-language-research/LAVIS/assets/merlion.png"
__magic_name__ : Union[str, Any] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw ).convert("RGB" )
return image
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = []
# fmt: off
# vision encoder
rename_keys.append(("visual_encoder.cls_token", "vision_model.embeddings.class_embedding") )
rename_keys.append(("visual_encoder.pos_embed", "vision_model.embeddings.position_embedding") )
rename_keys.append(("visual_encoder.patch_embed.proj.weight", "vision_model.embeddings.patch_embedding.weight") )
rename_keys.append(("visual_encoder.patch_embed.proj.bias", "vision_model.embeddings.patch_embedding.bias") )
rename_keys.append(("ln_vision.weight", "vision_model.post_layernorm.weight") )
rename_keys.append(("ln_vision.bias", "vision_model.post_layernorm.bias") )
for i in range(config.vision_config.num_hidden_layers ):
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.weight''', F'''vision_model.encoder.layers.{i}.layer_norm1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm1.bias''', F'''vision_model.encoder.layers.{i}.layer_norm1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.weight''', F'''vision_model.encoder.layers.{i}.layer_norm2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.norm2.bias''', F'''vision_model.encoder.layers.{i}.layer_norm2.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.qkv.weight''', F'''vision_model.encoder.layers.{i}.self_attn.qkv.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.weight''', F'''vision_model.encoder.layers.{i}.self_attn.projection.weight''',) )
rename_keys.append((F'''visual_encoder.blocks.{i}.attn.proj.bias''', F'''vision_model.encoder.layers.{i}.self_attn.projection.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc1.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc1.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc1.bias''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.weight''', F'''vision_model.encoder.layers.{i}.mlp.fc2.weight''') )
rename_keys.append((F'''visual_encoder.blocks.{i}.mlp.fc2.bias''', F'''vision_model.encoder.layers.{i}.mlp.fc2.bias''') )
# QFormer
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.weight", "qformer.layernorm.weight") )
rename_keys.append(("Qformer.bert.embeddings.LayerNorm.bias", "qformer.layernorm.bias") )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Optional[Any] ) -> int:
'''simple docstring'''
__magic_name__ : Tuple = dct.pop(_snake_case )
__magic_name__ : int = val
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
for i in range(config.vision_config.num_hidden_layers ):
# read in original q and v biases
__magic_name__ : List[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.q_bias''' )
__magic_name__ : Optional[Any] = state_dict.pop(F'''visual_encoder.blocks.{i}.attn.v_bias''' )
# next, set bias in the state dict
__magic_name__ : Optional[int] = torch.cat((q_bias, torch.zeros_like(_snake_case , requires_grad=_snake_case ), v_bias) )
__magic_name__ : Union[str, Any] = qkv_bias
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : str ) -> int:
'''simple docstring'''
__magic_name__ : List[Any] = 364 if "coco" in model_name else 224
__magic_name__ : Union[str, Any] = BlipaVisionConfig(image_size=_snake_case ).to_dict()
# make sure the models have proper bos_token_id and eos_token_id set (important for generation)
# seems like flan-T5 models don't have bos_token_id properly set?
if "opt-2.7b" in model_name:
__magic_name__ : List[str] = OPTConfig.from_pretrained("facebook/opt-2.7b" , eos_token_id=_snake_case ).to_dict()
elif "opt-6.7b" in model_name:
__magic_name__ : Any = OPTConfig.from_pretrained("facebook/opt-6.7b" , eos_token_id=_snake_case ).to_dict()
elif "t5-xl" in model_name:
__magic_name__ : Dict = TaConfig.from_pretrained("google/flan-t5-xl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
elif "t5-xxl" in model_name:
__magic_name__ : int = TaConfig.from_pretrained("google/flan-t5-xxl" , dense_act_fn="gelu" , bos_token_id=1 ).to_dict()
__magic_name__ : List[Any] = BlipaConfig(vision_config=_snake_case , text_config=_snake_case )
return config, image_size
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : str=None , _snake_case : Dict=False ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Optional[int] = (
AutoTokenizer.from_pretrained("facebook/opt-2.7b" )
if "opt" in model_name
else AutoTokenizer.from_pretrained("google/flan-t5-xl" )
)
__magic_name__ : List[Any] = tokenizer("\n" , add_special_tokens=_snake_case ).input_ids[0]
__magic_name__ , __magic_name__ : Tuple = get_blipa_config(_snake_case , eos_token_id=_snake_case )
__magic_name__ : Union[str, Any] = BlipaForConditionalGeneration(_snake_case ).eval()
__magic_name__ : Any = {
"blip2-opt-2.7b": ("blip2_opt", "pretrain_opt2.7b"),
"blip2-opt-6.7b": ("blip2_opt", "pretrain_opt6.7b"),
"blip2-opt-2.7b-coco": ("blip2_opt", "caption_coco_opt2.7b"),
"blip2-opt-6.7b-coco": ("blip2_opt", "caption_coco_opt6.7b"),
"blip2-flan-t5-xl": ("blip2_t5", "pretrain_flant5xl"),
"blip2-flan-t5-xl-coco": ("blip2_t5", "caption_coco_flant5xl"),
"blip2-flan-t5-xxl": ("blip2_t5", "pretrain_flant5xxl"),
}
__magic_name__ , __magic_name__ : Union[str, Any] = model_name_to_original[model_name]
# load original model
print("Loading original model..." )
__magic_name__ : Union[str, Any] = "cuda" if torch.cuda.is_available() else "cpu"
__magic_name__ , __magic_name__ , __magic_name__ : Optional[Any] = load_model_and_preprocess(
name=_snake_case , model_type=_snake_case , is_eval=_snake_case , device=_snake_case )
original_model.eval()
print("Done!" )
# update state dict keys
__magic_name__ : Dict = original_model.state_dict()
__magic_name__ : str = create_rename_keys(_snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
# some keys can be renamed efficiently
for key, val in state_dict.copy().items():
__magic_name__ : Any = state_dict.pop(_snake_case )
if key.startswith("Qformer.bert" ):
__magic_name__ : Optional[int] = key.replace("Qformer.bert" , "qformer" )
if "attention.self" in key:
__magic_name__ : Any = key.replace("self" , "attention" )
if "opt_proj" in key:
__magic_name__ : Union[str, Any] = key.replace("opt_proj" , "language_projection" )
if "t5_proj" in key:
__magic_name__ : Optional[int] = key.replace("t5_proj" , "language_projection" )
if key.startswith("opt" ):
__magic_name__ : List[str] = key.replace("opt" , "language" )
if key.startswith("t5" ):
__magic_name__ : Tuple = key.replace("t5" , "language" )
__magic_name__ : Dict = val
# read in qv biases
read_in_q_v_bias(_snake_case , _snake_case )
__magic_name__ , __magic_name__ : Tuple = hf_model.load_state_dict(_snake_case , strict=_snake_case )
assert len(_snake_case ) == 0
assert unexpected_keys == ["qformer.embeddings.position_ids"]
__magic_name__ : List[Any] = load_demo_image()
__magic_name__ : Tuple = vis_processors["eval"](_snake_case ).unsqueeze(0 ).to(_snake_case )
__magic_name__ : Dict = tokenizer(["\n"] , return_tensors="pt" ).input_ids.to(_snake_case )
# create processor
__magic_name__ : Optional[Any] = BlipImageProcessor(
size={"height": image_size, "width": image_size} , image_mean=_snake_case , image_std=_snake_case )
__magic_name__ : Dict = BlipaProcessor(image_processor=_snake_case , tokenizer=_snake_case )
__magic_name__ : Union[str, Any] = processor(images=_snake_case , return_tensors="pt" ).pixel_values.to(_snake_case )
# make sure processor creates exact same pixel values
assert torch.allclose(_snake_case , _snake_case )
original_model.to(_snake_case )
hf_model.to(_snake_case )
with torch.no_grad():
if "opt" in model_name:
__magic_name__ : List[Any] = original_model({"image": original_pixel_values, "text_input": [""]} ).logits
__magic_name__ : Optional[int] = hf_model(_snake_case , _snake_case ).logits
else:
__magic_name__ : int = original_model(
{"image": original_pixel_values, "text_input": ["\n"], "text_output": ["\n"]} ).logits
__magic_name__ : Tuple = input_ids.masked_fill(input_ids == tokenizer.pad_token_id , -100 )
__magic_name__ : List[str] = hf_model(_snake_case , _snake_case , labels=_snake_case ).logits
assert original_logits.shape == logits.shape
print("First values of original logits:" , original_logits[0, :3, :3] )
print("First values of HF logits:" , logits[0, :3, :3] )
# assert values
if model_name == "blip2-flan-t5-xl":
__magic_name__ : List[str] = torch.tensor(
[[-41.5_850, -4.4_440, -8.9_922], [-47.4_322, -5.9_143, -1.7_340]] , device=_snake_case )
assert torch.allclose(logits[0, :3, :3] , _snake_case , atol=1E-4 )
elif model_name == "blip2-flan-t5-xl-coco":
__magic_name__ : Tuple = torch.tensor(
[[-57.0_109, -9.8_967, -12.6_280], [-68.6_578, -12.7_191, -10.5_065]] , device=_snake_case )
else:
# cast to same type
__magic_name__ : str = logits.dtype
assert torch.allclose(original_logits.to(_snake_case ) , _snake_case , atol=1E-2 )
print("Looks ok!" )
print("Generating a caption..." )
__magic_name__ : Optional[int] = ""
__magic_name__ : Dict = tokenizer(_snake_case , return_tensors="pt" ).input_ids.to(_snake_case )
__magic_name__ : int = original_model.generate({"image": original_pixel_values} )
__magic_name__ : Optional[Any] = hf_model.generate(
_snake_case , _snake_case , do_sample=_snake_case , num_beams=5 , max_length=30 , min_length=1 , top_p=0.9 , repetition_penalty=1.0 , length_penalty=1.0 , temperature=1 , )
print("Original generation:" , _snake_case )
__magic_name__ : Tuple = input_ids.shape[1]
__magic_name__ : int = processor.batch_decode(outputs[:, prompt_length:] , skip_special_tokens=_snake_case )
__magic_name__ : Union[str, Any] = [text.strip() for text in output_text]
print("HF generation:" , _snake_case )
if pytorch_dump_folder_path is not None:
processor.save_pretrained(_snake_case )
hf_model.save_pretrained(_snake_case )
if push_to_hub:
processor.push_to_hub(F'''nielsr/{model_name}''' )
hf_model.push_to_hub(F'''nielsr/{model_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
snake_case : Union[str, Any] = [
"blip2-opt-2.7b",
"blip2-opt-6.7b",
"blip2-opt-2.7b-coco",
"blip2-opt-6.7b-coco",
"blip2-flan-t5-xl",
"blip2-flan-t5-xl-coco",
"blip2-flan-t5-xxl",
]
parser.add_argument(
"--model_name",
default="blip2-opt-2.7b",
choices=choices,
type=str,
help="Path to hf config.json of model to convert",
)
parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.")
parser.add_argument(
"--push_to_hub",
action="store_true",
help="Whether to push the model and processor to the hub after converting",
)
snake_case : int = parser.parse_args()
convert_blipa_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
import os
import sys
import unittest
snake_case : int = os.path.abspath(os.path.dirname(os.path.dirname(os.path.dirname(__file__))))
sys.path.append(os.path.join(git_repo_path, "utils"))
import check_dummies # noqa: E402
from check_dummies import create_dummy_files, create_dummy_object, find_backend, read_init # noqa: E402
# Align TRANSFORMERS_PATH in check_dummies with the current path
snake_case : Union[str, Any] = os.path.join(git_repo_path, "src", "diffusers")
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = find_backend(" if not is_torch_available():" )
self.assertEqual(_a , "torch" )
# backend_with_underscore = find_backend(" if not is_tensorflow_text_available():")
# self.assertEqual(backend_with_underscore, "tensorflow_text")
__magic_name__ : Dict = find_backend(" if not (is_torch_available() and is_transformers_available()):" )
self.assertEqual(_a , "torch_and_transformers" )
# double_backend_with_underscore = find_backend(
# " if not (is_sentencepiece_available() and is_tensorflow_text_available()):"
# )
# self.assertEqual(double_backend_with_underscore, "sentencepiece_and_tensorflow_text")
__magic_name__ : List[Any] = find_backend(
" if not (is_torch_available() and is_transformers_available() and is_onnx_available()):" )
self.assertEqual(_a , "torch_and_transformers_and_onnx" )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = read_init()
# We don't assert on the exact list of keys to allow for smooth grow of backend-specific objects
self.assertIn("torch" , _a )
self.assertIn("torch_and_transformers" , _a )
self.assertIn("flax_and_transformers" , _a )
self.assertIn("torch_and_transformers_and_onnx" , _a )
# Likewise, we can't assert on the exact content of a key
self.assertIn("UNet2DModel" , objects["torch"] )
self.assertIn("FlaxUNet2DConditionModel" , objects["flax"] )
self.assertIn("StableDiffusionPipeline" , objects["torch_and_transformers"] )
self.assertIn("FlaxStableDiffusionPipeline" , objects["flax_and_transformers"] )
self.assertIn("LMSDiscreteScheduler" , objects["torch_and_scipy"] )
self.assertIn("OnnxStableDiffusionPipeline" , objects["torch_and_transformers_and_onnx"] )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = create_dummy_object("CONSTANT" , "'torch'" )
self.assertEqual(_a , "\nCONSTANT = None\n" )
__magic_name__ : List[Any] = create_dummy_object("function" , "'torch'" )
self.assertEqual(
_a , "\ndef function(*args, **kwargs):\n requires_backends(function, 'torch')\n" )
__magic_name__ : Any = "\nclass FakeClass(metaclass=DummyObject):\n _backends = 'torch'\n\n def __init__(self, *args, **kwargs):\n requires_backends(self, 'torch')\n\n @classmethod\n def from_config(cls, *args, **kwargs):\n requires_backends(cls, 'torch')\n\n @classmethod\n def from_pretrained(cls, *args, **kwargs):\n requires_backends(cls, 'torch')\n"
__magic_name__ : str = create_dummy_object("FakeClass" , "'torch'" )
self.assertEqual(_a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = "# This file is autogenerated by the command `make fix-copies`, do not edit.\nfrom ..utils import DummyObject, requires_backends\n\n\nCONSTANT = None\n\n\ndef function(*args, **kwargs):\n requires_backends(function, [\"torch\"])\n\n\nclass FakeClass(metaclass=DummyObject):\n _backends = [\"torch\"]\n\n def __init__(self, *args, **kwargs):\n requires_backends(self, [\"torch\"])\n\n @classmethod\n def from_config(cls, *args, **kwargs):\n requires_backends(cls, [\"torch\"])\n\n @classmethod\n def from_pretrained(cls, *args, **kwargs):\n requires_backends(cls, [\"torch\"])\n"
__magic_name__ : Optional[int] = create_dummy_files({"torch": ["CONSTANT", "function", "FakeClass"]} )
self.assertEqual(dummy_files["torch"] , _a )
| 281 |
import os
import re
from shutil import copyfile
from typing import List, Optional, Tuple
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
snake_case : Dict = logging.get_logger(__name__)
snake_case : Union[str, Any] = {
"vocab_file": "vocab.txt",
"merges_file": "bpe.codes",
}
snake_case : Dict = {
"vocab_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/vocab.txt",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/vocab.txt",
},
"merges_file": {
"vinai/phobert-base": "https://huggingface.co/vinai/phobert-base/resolve/main/bpe.codes",
"vinai/phobert-large": "https://huggingface.co/vinai/phobert-large/resolve/main/bpe.codes",
},
}
snake_case : Union[str, Any] = {
"vinai/phobert-base": 256,
"vinai/phobert-large": 256,
}
def lowerCAmelCase_ ( _snake_case : str ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : List[str] = set()
__magic_name__ : Any = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
__magic_name__ : int = char
__magic_name__ : List[str] = set(_snake_case )
return pairs
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a , _a="<s>" , _a="</s>" , _a="</s>" , _a="<s>" , _a="<unk>" , _a="<pad>" , _a="<mask>" , **_a , ):
super().__init__(
bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , cls_token=_a , pad_token=_a , mask_token=_a , **_a , )
__magic_name__ : Dict = vocab_file
__magic_name__ : Tuple = merges_file
__magic_name__ : List[Any] = {}
__magic_name__ : List[Any] = 0
__magic_name__ : Tuple = 1
__magic_name__ : int = 2
__magic_name__ : Union[str, Any] = 3
self.add_from_file(_a )
__magic_name__ : Optional[int] = {v: k for k, v in self.encoder.items()}
with open(_a , encoding="utf-8" ) as merges_handle:
__magic_name__ : List[str] = merges_handle.read().split("\n" )[:-1]
__magic_name__ : Union[str, Any] = [tuple(merge.split()[:-1] ) for merge in merges]
__magic_name__ : Union[str, Any] = dict(zip(_a , range(len(_a ) ) ) )
__magic_name__ : Optional[int] = {}
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
__magic_name__ : Optional[Any] = [self.cls_token_id]
__magic_name__ : Union[str, Any] = [self.sep_token_id]
return cls + token_ids_a + sep + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is None:
return [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[Any] = [self.sep_token_id]
__magic_name__ : 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 + sep + token_ids_a + sep ) * [0]
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.encoder )
def SCREAMING_SNAKE_CASE ( self ):
return dict(self.encoder , **self.added_tokens_encoder )
def SCREAMING_SNAKE_CASE ( self , _a ):
if token in self.cache:
return self.cache[token]
__magic_name__ : List[Any] = tuple(_a )
__magic_name__ : List[Any] = tuple(list(word[:-1] ) + [word[-1] + "</w>"] )
__magic_name__ : Any = get_pairs(_a )
if not pairs:
return token
while True:
__magic_name__ : str = min(_a , key=lambda _a : self.bpe_ranks.get(_a , float("inf" ) ) )
if bigram not in self.bpe_ranks:
break
__magic_name__ , __magic_name__ : List[str] = bigram
__magic_name__ : List[str] = []
__magic_name__ : List[str] = 0
while i < len(_a ):
try:
__magic_name__ : Any = word.index(_a , _a )
except ValueError:
new_word.extend(word[i:] )
break
else:
new_word.extend(word[i:j] )
__magic_name__ : Tuple = j
if word[i] == first and i < len(_a ) - 1 and word[i + 1] == second:
new_word.append(first + second )
i += 2
else:
new_word.append(word[i] )
i += 1
__magic_name__ : Union[str, Any] = tuple(_a )
__magic_name__ : Optional[int] = new_word
if len(_a ) == 1:
break
else:
__magic_name__ : List[Any] = get_pairs(_a )
__magic_name__ : Optional[int] = "@@ ".join(_a )
__magic_name__ : Tuple = word[:-4]
__magic_name__ : str = word
return word
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = []
__magic_name__ : Dict = re.findall(r"\S+\n?" , _a )
for token in words:
split_tokens.extend(list(self.bpe(_a ).split(" " ) ) )
return split_tokens
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.encoder.get(_a , self.encoder.get(self.unk_token ) )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.decoder.get(_a , self.unk_token )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Tuple = " ".join(_a ).replace("@@ " , "" ).strip()
return out_string
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : Optional[int] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
__magic_name__ : Union[str, Any] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["merges_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ):
copyfile(self.vocab_file , _a )
if os.path.abspath(self.merges_file ) != os.path.abspath(_a ):
copyfile(self.merges_file , _a )
return out_vocab_file, out_merge_file
def SCREAMING_SNAKE_CASE ( self , _a ):
if isinstance(_a , _a ):
try:
with open(_a , "r" , encoding="utf-8" ) as fd:
self.add_from_file(_a )
except FileNotFoundError as fnfe:
raise fnfe
except UnicodeError:
raise Exception(f'''Incorrect encoding detected in {f}, please rebuild the dataset''' )
return
__magic_name__ : List[Any] = f.readlines()
for lineTmp in lines:
__magic_name__ : Optional[Any] = lineTmp.strip()
__magic_name__ : Union[str, Any] = line.rfind(" " )
if idx == -1:
raise ValueError("Incorrect dictionary format, expected '<token> <cnt>'" )
__magic_name__ : Optional[int] = line[:idx]
__magic_name__ : Dict = len(self.encoder )
| 281 | 1 |
import math
def lowerCAmelCase_ ( ) -> None:
'''simple docstring'''
__magic_name__ : Dict = input("Enter message: " )
__magic_name__ : Dict = int(input(F'''Enter key [2-{len(_snake_case ) - 1}]: ''' ) )
__magic_name__ : Optional[int] = input("Encryption/Decryption [e/d]: " )
if mode.lower().startswith("e" ):
__magic_name__ : Union[str, Any] = encrypt_message(_snake_case , _snake_case )
elif mode.lower().startswith("d" ):
__magic_name__ : Dict = decrypt_message(_snake_case , _snake_case )
# Append pipe symbol (vertical bar) to identify spaces at the end.
print(F'''Output:\n{text + "|"}''' )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str ) -> str:
'''simple docstring'''
__magic_name__ : Optional[int] = [""] * key
for col in range(_snake_case ):
__magic_name__ : Union[str, Any] = col
while pointer < len(_snake_case ):
cipher_text[col] += message[pointer]
pointer += key
return "".join(_snake_case )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str ) -> str:
'''simple docstring'''
__magic_name__ : Tuple = math.ceil(len(_snake_case ) / key )
__magic_name__ : int = key
__magic_name__ : Tuple = (num_cols * num_rows) - len(_snake_case )
__magic_name__ : int = [""] * num_cols
__magic_name__ : Dict = 0
__magic_name__ : Dict = 0
for symbol in message:
plain_text[col] += symbol
col += 1
if (
(col == num_cols)
or (col == num_cols - 1)
and (row >= num_rows - num_shaded_boxes)
):
__magic_name__ : Any = 0
row += 1
return "".join(_snake_case )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 281 |
from itertools import zip_longest
import requests
from bsa import BeautifulSoup
from pandas import DataFrame
def lowerCAmelCase_ ( _snake_case : str = "laptop" ) -> DataFrame:
'''simple docstring'''
__magic_name__ : Tuple = F'''https://www.amazon.in/laptop/s?k={product}'''
__magic_name__ : Dict = {
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36\n (KHTML, like Gecko)Chrome/44.0.2403.157 Safari/537.36",
"Accept-Language": "en-US, en;q=0.5",
}
__magic_name__ : Tuple = BeautifulSoup(requests.get(_snake_case , headers=_snake_case ).text )
# Initialize a Pandas dataframe with the column titles
__magic_name__ : int = DataFrame(
columns=[
"Product Title",
"Product Link",
"Current Price of the product",
"Product Rating",
"MRP of the product",
"Discount",
] )
# Loop through each entry and store them in the dataframe
for item, _ in zip_longest(
soup.find_all(
"div" , attrs={"class": "s-result-item", "data-component-type": "s-search-result"} , ) , soup.find_all("div" , attrs={"class": "a-row a-size-base a-color-base"} ) , ):
try:
__magic_name__ : Dict = item.ha.text
__magic_name__ : Optional[int] = "https://www.amazon.in/" + item.ha.a["href"]
__magic_name__ : Optional[Any] = item.find("span" , attrs={"class": "a-offscreen"} ).text
try:
__magic_name__ : Union[str, Any] = item.find("span" , attrs={"class": "a-icon-alt"} ).text
except AttributeError:
__magic_name__ : Dict = "Not available"
try:
__magic_name__ : Optional[int] = (
"₹"
+ item.find(
"span" , attrs={"class": "a-price a-text-price"} ).text.split("₹" )[1]
)
except AttributeError:
__magic_name__ : List[str] = ""
try:
__magic_name__ : int = float(
(
(
float(product_mrp.strip("₹" ).replace("," , "" ) )
- float(product_price.strip("₹" ).replace("," , "" ) )
)
/ float(product_mrp.strip("₹" ).replace("," , "" ) )
)
* 100 )
except ValueError:
__magic_name__ : str = float("nan" )
except AttributeError:
pass
__magic_name__ : Optional[int] = [
product_title,
product_link,
product_price,
product_rating,
product_mrp,
discount,
]
__magic_name__ : Optional[Any] = " "
__magic_name__ : str = " "
data_frame.index += 1
return data_frame
if __name__ == "__main__":
snake_case : Any = "headphones"
get_amazon_product_data(product).to_csv(F"Amazon Product Data for {product}.csv")
| 281 | 1 |
from typing import List, Optional
import numpy as np
from ...processing_utils import ProcessorMixin
from ...utils import to_numpy
class _snake_case ( snake_case ):
UpperCamelCase__ = 'EncodecFeatureExtractor'
UpperCamelCase__ = ('T5Tokenizer', 'T5TokenizerFast')
def __init__( self , _a , _a ):
super().__init__(_a , _a )
__magic_name__ : int = self.feature_extractor
__magic_name__ : List[str] = False
def SCREAMING_SNAKE_CASE ( self , _a=None , _a=None , _a=True ):
return self.tokenizer.get_decoder_prompt_ids(task=_a , language=_a , no_timestamps=_a )
def __call__( self , *_a , **_a ):
# For backward compatibility
if self._in_target_context_manager:
return self.current_processor(*_a , **_a )
__magic_name__ : Union[str, Any] = kwargs.pop("audio" , _a )
__magic_name__ : List[str] = kwargs.pop("sampling_rate" , _a )
__magic_name__ : Tuple = kwargs.pop("text" , _a )
if len(_a ) > 0:
__magic_name__ : Union[str, Any] = args[0]
__magic_name__ : str = args[1:]
if audio is None and text is None:
raise ValueError("You need to specify either an `audio` or `text` input to process." )
if text is not None:
__magic_name__ : Dict = self.tokenizer(_a , **_a )
if audio is not None:
__magic_name__ : Any = self.feature_extractor(_a , *_a , sampling_rate=_a , **_a )
if audio is None:
return inputs
elif text is None:
return audio_inputs
else:
__magic_name__ : Union[str, Any] = audio_inputs["input_values"]
if "padding_mask" in audio_inputs:
__magic_name__ : Optional[Any] = audio_inputs["padding_mask"]
return inputs
def SCREAMING_SNAKE_CASE ( self , *_a , **_a ):
__magic_name__ : Union[str, Any] = kwargs.pop("audio" , _a )
__magic_name__ : List[Any] = kwargs.pop("padding_mask" , _a )
if len(_a ) > 0:
__magic_name__ : List[str] = args[0]
__magic_name__ : List[Any] = args[1:]
if audio_values is not None:
return self._decode_audio(_a , padding_mask=_a )
else:
return self.tokenizer.batch_decode(*_a , **_a )
def SCREAMING_SNAKE_CASE ( self , *_a , **_a ):
return self.tokenizer.decode(*_a , **_a )
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[Any] = to_numpy(_a )
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = audio_values.shape
if padding_mask is None:
return list(_a )
__magic_name__ : Tuple = to_numpy(_a )
# match the sequence length of the padding mask to the generated audio arrays by padding with the **non-padding**
# token (so that the generated audio values are **not** treated as padded tokens)
__magic_name__ : str = seq_len - padding_mask.shape[-1]
__magic_name__ : str = 1 - self.feature_extractor.padding_value
__magic_name__ : List[Any] = np.pad(_a , ((0, 0), (0, difference)) , "constant" , constant_values=_a )
__magic_name__ : Tuple = audio_values.tolist()
for i in range(_a ):
__magic_name__ : Dict = np.asarray(audio_values[i] )[
padding_mask[i][None, :] != self.feature_extractor.padding_value
]
__magic_name__ : List[str] = sliced_audio.reshape(_a , -1 )
return audio_values
| 281 |
from __future__ import annotations
class _snake_case :
def __init__( self , _a ):
__magic_name__ : Optional[Any] = data
__magic_name__ : Node | None = None
__magic_name__ : Node | None = None
def lowerCAmelCase_ ( _snake_case : Node | None ) -> None: # In Order traversal of the tree
'''simple docstring'''
if tree:
display(tree.left )
print(tree.data )
display(tree.right )
def lowerCAmelCase_ ( _snake_case : Node | None ) -> int:
'''simple docstring'''
return 1 + max(depth_of_tree(tree.left ) , depth_of_tree(tree.right ) ) if tree else 0
def lowerCAmelCase_ ( _snake_case : Node ) -> bool:
'''simple docstring'''
if not tree:
return True
if tree.left and tree.right:
return is_full_binary_tree(tree.left ) and is_full_binary_tree(tree.right )
else:
return not tree.left and not tree.right
def lowerCAmelCase_ ( ) -> None: # Main function for testing.
'''simple docstring'''
__magic_name__ : int = Node(1 )
__magic_name__ : Union[str, Any] = Node(2 )
__magic_name__ : Tuple = Node(3 )
__magic_name__ : Optional[Any] = Node(4 )
__magic_name__ : Union[str, Any] = Node(5 )
__magic_name__ : Any = Node(6 )
__magic_name__ : int = Node(7 )
__magic_name__ : List[str] = Node(8 )
__magic_name__ : Union[str, Any] = Node(9 )
print(is_full_binary_tree(_snake_case ) )
print(depth_of_tree(_snake_case ) )
print("Tree is: " )
display(_snake_case )
if __name__ == "__main__":
main()
| 281 | 1 |
from timeit import timeit
snake_case : List[str] = {
"MALAYALAM": True,
"String": False,
"rotor": True,
"level": True,
"A": True,
"BB": True,
"ABC": False,
"amanaplanacanalpanama": True, # "a man a plan a canal panama"
}
# Ensure our test data is valid
assert all((key == key[::-1]) is value for key, value in test_data.items())
def lowerCAmelCase_ ( _snake_case : str ) -> bool:
'''simple docstring'''
__magic_name__ : Union[str, Any] = 0
__magic_name__ : Union[str, Any] = len(_snake_case ) - 1
while start_i < end_i:
if s[start_i] == s[end_i]:
start_i += 1
end_i -= 1
else:
return False
return True
def lowerCAmelCase_ ( _snake_case : str ) -> bool:
'''simple docstring'''
__magic_name__ : Dict = len(_snake_case ) // 2
__magic_name__ : Tuple = len(_snake_case )
# We need to traverse till half of the length of string
# as we can get access of the i'th last element from
# i'th index.
# eg: [0,1,2,3,4,5] => 4th index can be accessed
# with the help of 1st index (i==n-i-1)
# where n is length of string
return all(s[i] == s[n - i - 1] for i in range(_snake_case ) )
def lowerCAmelCase_ ( _snake_case : str ) -> bool:
'''simple docstring'''
if len(_snake_case ) <= 2:
return True
if s[0] == s[len(_snake_case ) - 1]:
return is_palindrome_recursive(s[1:-1] )
else:
return False
def lowerCAmelCase_ ( _snake_case : str ) -> bool:
'''simple docstring'''
return s == s[::-1]
def lowerCAmelCase_ ( _snake_case : str ) -> None:
'''simple docstring'''
__magic_name__ : Dict = F'''all({name}(key) is value for key, value in test_data.items())'''
__magic_name__ : Dict = F'''from __main__ import test_data, {name}'''
__magic_name__ : Optional[int] = 500000
__magic_name__ : Dict = timeit(stmt=_snake_case , setup=_snake_case , number=_snake_case )
print(F'''{name:<35} finished {number:,} runs in {result:.5f} seconds''' )
if __name__ == "__main__":
for key, value in test_data.items():
assert is_palindrome(key) is is_palindrome_recursive(key)
assert is_palindrome(key) is is_palindrome_slice(key)
print(F"{key:21} {value}")
print("a man a plan a canal panama")
# finished 500,000 runs in 0.46793 seconds
benchmark_function("is_palindrome_slice")
# finished 500,000 runs in 0.85234 seconds
benchmark_function("is_palindrome")
# finished 500,000 runs in 1.32028 seconds
benchmark_function("is_palindrome_recursive")
# finished 500,000 runs in 2.08679 seconds
benchmark_function("is_palindrome_traversal")
| 281 |
def lowerCAmelCase_ ( _snake_case : str , _snake_case : str ) -> bool:
'''simple docstring'''
__magic_name__ : Union[str, Any] = len(_snake_case ) + 1
__magic_name__ : List[str] = len(_snake_case ) + 1
# dp is a 2d matrix where dp[i][j] denotes whether prefix string of
# length i of input_string matches with prefix string of length j of
# given pattern.
# "dp" stands for dynamic programming.
__magic_name__ : str = [[0 for i in range(_snake_case )] for j in range(_snake_case )]
# since string of zero length match pattern of zero length
__magic_name__ : Optional[int] = 1
# since pattern of zero length will never match with string of non-zero length
for i in range(1 , _snake_case ):
__magic_name__ : Optional[int] = 0
# since string of zero length will match with pattern where there
# is at least one * alternatively
for j in range(1 , _snake_case ):
__magic_name__ : Union[str, Any] = dp[0][j - 2] if pattern[j - 1] == "*" else 0
# now using bottom-up approach to find for all remaining lengths
for i in range(1 , _snake_case ):
for j in range(1 , _snake_case ):
if input_string[i - 1] == pattern[j - 1] or pattern[j - 1] == ".":
__magic_name__ : Optional[int] = dp[i - 1][j - 1]
elif pattern[j - 1] == "*":
if dp[i][j - 2] == 1:
__magic_name__ : Optional[Any] = 1
elif pattern[j - 2] in (input_string[i - 1], "."):
__magic_name__ : List[Any] = dp[i - 1][j]
else:
__magic_name__ : Union[str, Any] = 0
else:
__magic_name__ : Dict = 0
return bool(dp[-1][-1] )
if __name__ == "__main__":
import doctest
doctest.testmod()
# inputing the strings
# input_string = input("input a string :")
# pattern = input("input a pattern :")
snake_case : Optional[Any] = "aab"
snake_case : List[str] = "c*a*b"
# using function to check whether given string matches the given pattern
if match_pattern(input_string, pattern):
print(F"{input_string} matches the given pattern {pattern}")
else:
print(F"{input_string} does not match with the given pattern {pattern}")
| 281 | 1 |
import importlib
import os
import fsspec
import pytest
from fsspec import register_implementation
from fsspec.registry import _registry as _fsspec_registry
from datasets.filesystems import COMPRESSION_FILESYSTEMS, HfFileSystem, extract_path_from_uri, is_remote_filesystem
from .utils import require_lza, require_zstandard
def lowerCAmelCase_ ( _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
assert "mock" in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Tuple:
'''simple docstring'''
assert "mock" not in _fsspec_registry
assert "bz2" in _fsspec_registry
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Dict = "mock-s3-bucket"
__magic_name__ : Any = F'''s3://{mock_bucket}'''
__magic_name__ : str = extract_path_from_uri(_snake_case )
assert dataset_path.startswith("s3://" ) is False
__magic_name__ : Tuple = "./local/path"
__magic_name__ : Optional[Any] = extract_path_from_uri(_snake_case )
assert dataset_path == new_dataset_path
def lowerCAmelCase_ ( _snake_case : List[str] ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ : str = is_remote_filesystem(_snake_case )
assert is_remote is True
__magic_name__ : Optional[int] = fsspec.filesystem("file" )
__magic_name__ : int = is_remote_filesystem(_snake_case )
assert is_remote is False
@pytest.mark.parametrize("compression_fs_class" , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : int , _snake_case : Tuple , _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Any ) -> int:
'''simple docstring'''
__magic_name__ : Any = {"gzip": gz_file, "xz": xz_file, "zstd": zstd_file, "bz2": bza_file, "lz4": lza_file}
__magic_name__ : str = input_paths[compression_fs_class.protocol]
if input_path is None:
__magic_name__ : Dict = F'''for \'{compression_fs_class.protocol}\' compression protocol, '''
if compression_fs_class.protocol == "lz4":
reason += require_lza.kwargs["reason"]
elif compression_fs_class.protocol == "zstd":
reason += require_zstandard.kwargs["reason"]
pytest.skip(_snake_case )
__magic_name__ : str = fsspec.filesystem(compression_fs_class.protocol , fo=_snake_case )
assert isinstance(_snake_case , _snake_case )
__magic_name__ : int = os.path.basename(_snake_case )
__magic_name__ : Optional[int] = expected_filename[: expected_filename.rindex("." )]
assert fs.glob("*" ) == [expected_filename]
with fs.open(_snake_case , "r" , encoding="utf-8" ) as f, open(_snake_case , encoding="utf-8" ) as expected_file:
assert f.read() == expected_file.read()
@pytest.mark.parametrize("protocol" , ["zip", "gzip"] )
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : Optional[Any] , _snake_case : Optional[Any] ) -> str:
'''simple docstring'''
__magic_name__ : int = {"zip": zip_jsonl_path, "gzip": jsonl_gz_path}
__magic_name__ : int = compressed_file_paths[protocol]
__magic_name__ : Tuple = "dataset.jsonl"
__magic_name__ : List[str] = F'''{protocol}://{member_file_path}::{compressed_file_path}'''
__magic_name__ , *__magic_name__ : Optional[Any] = fsspec.get_fs_token_paths(_snake_case )
assert fs.isfile(_snake_case )
assert not fs.isfile("non_existing_" + member_file_path )
@pytest.mark.integration
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : List[str] , _snake_case : Tuple ) -> str:
'''simple docstring'''
__magic_name__ : int = hf_api.dataset_info(_snake_case , token=_snake_case )
__magic_name__ : Optional[Any] = HfFileSystem(repo_info=_snake_case , token=_snake_case )
assert sorted(hffs.glob("*" ) ) == [".gitattributes", "data"]
assert hffs.isdir("data" )
assert hffs.isfile(".gitattributes" ) and hffs.isfile("data/text_data.txt" )
with open(_snake_case ) as f:
assert hffs.open("data/text_data.txt" , "r" ).read() == f.read()
def lowerCAmelCase_ ( ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : Optional[Any] = "bz2"
# Import module
import datasets.filesystems
# Overwrite protocol and reload
register_implementation(_snake_case , _snake_case , clobber=_snake_case )
with pytest.warns(_snake_case ) as warning_info:
importlib.reload(datasets.filesystems )
assert len(_snake_case ) == 1
assert (
str(warning_info[0].message )
== F'''A filesystem protocol was already set for {protocol} and will be overwritten.'''
)
| 281 |
import hashlib
import unittest
from typing import Dict
import numpy as np
from transformers import (
MODEL_FOR_MASK_GENERATION_MAPPING,
TF_MODEL_FOR_MASK_GENERATION_MAPPING,
is_vision_available,
pipeline,
)
from transformers.pipelines import MaskGenerationPipeline
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
)
if is_vision_available():
from PIL import Image
else:
class _snake_case :
@staticmethod
def SCREAMING_SNAKE_CASE ( *_a , **_a ):
pass
def lowerCAmelCase_ ( _snake_case : Image ) -> str:
'''simple docstring'''
__magic_name__ : Optional[int] = hashlib.mda(image.tobytes() )
return m.hexdigest()[:10]
def lowerCAmelCase_ ( _snake_case : Image ) -> Dict:
'''simple docstring'''
__magic_name__ : List[Any] = np.array(_snake_case )
__magic_name__ : Optional[int] = npimg.shape
return {"hash": hashimage(_snake_case ), "shape": shape}
@is_pipeline_test
@require_vision
@require_torch
class _snake_case ( unittest.TestCase ):
UpperCamelCase__ = dict(
(list(MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if MODEL_FOR_MASK_GENERATION_MAPPING else []) )
UpperCamelCase__ = dict(
(list(TF_MODEL_FOR_MASK_GENERATION_MAPPING.items() ) if TF_MODEL_FOR_MASK_GENERATION_MAPPING else []) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a ):
__magic_name__ : Dict = MaskGenerationPipeline(model=_a , image_processor=_a )
return image_segmenter, [
"./tests/fixtures/tests_samples/COCO/000000039769.png",
"./tests/fixtures/tests_samples/COCO/000000039769.png",
]
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
pass
@require_tf
@unittest.skip("Image segmentation not implemented in TF" )
def SCREAMING_SNAKE_CASE ( self ):
pass
@slow
@require_torch
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = pipeline("mask-generation" , model="facebook/sam-vit-huge" )
__magic_name__ : str = image_segmenter("http://images.cocodataset.org/val2017/000000039769.jpg" , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Dict = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
# fmt: off
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.0_21},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
{"mask": {"hash": "e2d0b7a0b7", "shape": (480, 640)}, "scores": 0.99_67},
{"mask": {"hash": "453c7844bd", "shape": (480, 640)}, "scores": 0.9_93},
{"mask": {"hash": "3d44f2926d", "shape": (480, 640)}, "scores": 0.99_09},
{"mask": {"hash": "64033ddc3f", "shape": (480, 640)}, "scores": 0.98_79},
{"mask": {"hash": "801064ff79", "shape": (480, 640)}, "scores": 0.98_34},
{"mask": {"hash": "6172f276ef", "shape": (480, 640)}, "scores": 0.97_16},
{"mask": {"hash": "b49e60e084", "shape": (480, 640)}, "scores": 0.96_12},
{"mask": {"hash": "a811e775fd", "shape": (480, 640)}, "scores": 0.95_99},
{"mask": {"hash": "a6a8ebcf4b", "shape": (480, 640)}, "scores": 0.95_52},
{"mask": {"hash": "9d8257e080", "shape": (480, 640)}, "scores": 0.95_32},
{"mask": {"hash": "32de6454a8", "shape": (480, 640)}, "scores": 0.95_16},
{"mask": {"hash": "af3d4af2c8", "shape": (480, 640)}, "scores": 0.94_99},
{"mask": {"hash": "3c6db475fb", "shape": (480, 640)}, "scores": 0.94_83},
{"mask": {"hash": "c290813fb9", "shape": (480, 640)}, "scores": 0.94_64},
{"mask": {"hash": "b6f0b8f606", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "92ce16bfdf", "shape": (480, 640)}, "scores": 0.9_43},
{"mask": {"hash": "c749b25868", "shape": (480, 640)}, "scores": 0.94_08},
{"mask": {"hash": "efb6cab859", "shape": (480, 640)}, "scores": 0.93_35},
{"mask": {"hash": "1ff2eafb30", "shape": (480, 640)}, "scores": 0.93_26},
{"mask": {"hash": "788b798e24", "shape": (480, 640)}, "scores": 0.92_62},
{"mask": {"hash": "abea804f0e", "shape": (480, 640)}, "scores": 0.89_99},
{"mask": {"hash": "7b9e8ddb73", "shape": (480, 640)}, "scores": 0.89_86},
{"mask": {"hash": "cd24047c8a", "shape": (480, 640)}, "scores": 0.89_84},
{"mask": {"hash": "6943e6bcbd", "shape": (480, 640)}, "scores": 0.88_73},
{"mask": {"hash": "b5f47c9191", "shape": (480, 640)}, "scores": 0.88_71}
] , )
# fmt: on
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = "facebook/sam-vit-huge"
__magic_name__ : str = pipeline("mask-generation" , model=_a )
__magic_name__ : Tuple = image_segmenter(
"http://images.cocodataset.org/val2017/000000039769.jpg" , pred_iou_thresh=1 , points_per_batch=256 )
# Shortening by hashing
__magic_name__ : Any = []
for i, o in enumerate(outputs["masks"] ):
new_outupt += [{"mask": mask_to_test_readable(_a ), "scores": outputs["scores"][i]}]
self.assertEqual(
nested_simplify(_a , decimals=4 ) , [
{"mask": {"hash": "115ad19f5f", "shape": (480, 640)}, "scores": 1.04_44},
{"mask": {"hash": "6affa964c6", "shape": (480, 640)}, "scores": 1.02_10},
{"mask": {"hash": "dfe28a0388", "shape": (480, 640)}, "scores": 1.01_67},
{"mask": {"hash": "c0a5f4a318", "shape": (480, 640)}, "scores": 1.01_32},
{"mask": {"hash": "fe8065c197", "shape": (480, 640)}, "scores": 1.00_53},
] , )
| 281 | 1 |
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
snake_case : str = logging.get_logger(__name__)
snake_case : Any = {
"google/bigbird-roberta-base": "https://huggingface.co/google/bigbird-roberta-base/resolve/main/config.json",
"google/bigbird-roberta-large": "https://huggingface.co/google/bigbird-roberta-large/resolve/main/config.json",
"google/bigbird-base-trivia-itc": "https://huggingface.co/google/bigbird-base-trivia-itc/resolve/main/config.json",
# See all BigBird models at https://huggingface.co/models?filter=big_bird
}
class _snake_case ( snake_case ):
UpperCamelCase__ = 'big_bird'
def __init__( self , _a=50_358 , _a=768 , _a=12 , _a=12 , _a=3_072 , _a="gelu_new" , _a=0.1 , _a=0.1 , _a=4_096 , _a=2 , _a=0.02 , _a=1e-12 , _a=True , _a=0 , _a=1 , _a=2 , _a=66 , _a="block_sparse" , _a=True , _a=False , _a=64 , _a=3 , _a=None , **_a , ):
super().__init__(
pad_token_id=_a , bos_token_id=_a , eos_token_id=_a , sep_token_id=_a , **_a , )
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Dict = max_position_embeddings
__magic_name__ : Dict = hidden_size
__magic_name__ : int = num_hidden_layers
__magic_name__ : Dict = num_attention_heads
__magic_name__ : List[Any] = intermediate_size
__magic_name__ : Any = hidden_act
__magic_name__ : Optional[Any] = hidden_dropout_prob
__magic_name__ : List[Any] = attention_probs_dropout_prob
__magic_name__ : List[Any] = initializer_range
__magic_name__ : List[str] = type_vocab_size
__magic_name__ : Tuple = layer_norm_eps
__magic_name__ : Tuple = use_cache
__magic_name__ : Union[str, Any] = rescale_embeddings
__magic_name__ : Optional[int] = attention_type
__magic_name__ : Optional[int] = use_bias
__magic_name__ : Optional[Any] = block_size
__magic_name__ : List[Any] = num_random_blocks
__magic_name__ : List[str] = classifier_dropout
class _snake_case ( snake_case ):
@property
def SCREAMING_SNAKE_CASE ( self ):
if self.task == "multiple-choice":
__magic_name__ : List[Any] = {0: "batch", 1: "choice", 2: "sequence"}
else:
__magic_name__ : Dict = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
] )
| 281 |
import absl # noqa: F401 # Here to have a nice missing dependency error message early on
import nltk # noqa: F401 # Here to have a nice missing dependency error message early on
import numpy # noqa: F401 # Here to have a nice missing dependency error message early on
import six # noqa: F401 # Here to have a nice missing dependency error message early on
from rouge_score import rouge_scorer, scoring
import datasets
snake_case : List[Any] = "\\n@inproceedings{lin-2004-rouge,\n title = \"{ROUGE}: A Package for Automatic Evaluation of Summaries\",\n author = \"Lin, Chin-Yew\",\n booktitle = \"Text Summarization Branches Out\",\n month = jul,\n year = \"2004\",\n address = \"Barcelona, Spain\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://www.aclweb.org/anthology/W04-1013\",\n pages = \"74--81\",\n}\n"
snake_case : Any = "\\nROUGE, or Recall-Oriented Understudy for Gisting Evaluation, is a set of metrics and a software package used for\nevaluating automatic summarization and machine translation software in natural language processing.\nThe metrics compare an automatically produced summary or translation against a reference or a set of references (human-produced) summary or translation.\n\nNote that ROUGE is case insensitive, meaning that upper case letters are treated the same way as lower case letters.\n\nThis metrics is a wrapper around Google Research reimplementation of ROUGE:\nhttps://github.com/google-research/google-research/tree/master/rouge\n"
snake_case : str = "\nCalculates average rouge scores for a list of hypotheses and references\nArgs:\n predictions: list of predictions to score. Each prediction\n should be a string with tokens separated by spaces.\n references: list of reference for each prediction. Each\n reference should be a string with tokens separated by spaces.\n rouge_types: A list of rouge types to calculate.\n Valid names:\n `\"rouge{n}\"` (e.g. `\"rouge1\"`, `\"rouge2\"`) where: {n} is the n-gram based scoring,\n `\"rougeL\"`: Longest common subsequence based scoring.\n `\"rougeLSum\"`: rougeLsum splits text using `\"\n\"`.\n See details in https://github.com/huggingface/datasets/issues/617\n use_stemmer: Bool indicating whether Porter stemmer should be used to strip word suffixes.\n use_aggregator: Return aggregates if this is set to True\nReturns:\n rouge1: rouge_1 (precision, recall, f1),\n rouge2: rouge_2 (precision, recall, f1),\n rougeL: rouge_l (precision, recall, f1),\n rougeLsum: rouge_lsum (precision, recall, f1)\nExamples:\n\n >>> rouge = datasets.load_metric('rouge')\n >>> predictions = [\"hello there\", \"general kenobi\"]\n >>> references = [\"hello there\", \"general kenobi\"]\n >>> results = rouge.compute(predictions=predictions, references=references)\n >>> print(list(results.keys()))\n ['rouge1', 'rouge2', 'rougeL', 'rougeLsum']\n >>> print(results[\"rouge1\"])\n AggregateScore(low=Score(precision=1.0, recall=1.0, fmeasure=1.0), mid=Score(precision=1.0, recall=1.0, fmeasure=1.0), high=Score(precision=1.0, recall=1.0, fmeasure=1.0))\n >>> print(results[\"rouge1\"].mid.fmeasure)\n 1.0\n"
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class _snake_case ( datasets.Metric ):
def SCREAMING_SNAKE_CASE ( self ):
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , 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/google-research/tree/master/rouge"] , reference_urls=[
"https://en.wikipedia.org/wiki/ROUGE_(metric)",
"https://github.com/google-research/google-research/tree/master/rouge",
] , )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a=None , _a=True , _a=False ):
if rouge_types is None:
__magic_name__ : str = ["rouge1", "rouge2", "rougeL", "rougeLsum"]
__magic_name__ : List[str] = rouge_scorer.RougeScorer(rouge_types=_a , use_stemmer=_a )
if use_aggregator:
__magic_name__ : Dict = scoring.BootstrapAggregator()
else:
__magic_name__ : str = []
for ref, pred in zip(_a , _a ):
__magic_name__ : Union[str, Any] = scorer.score(_a , _a )
if use_aggregator:
aggregator.add_scores(_a )
else:
scores.append(_a )
if use_aggregator:
__magic_name__ : Any = aggregator.aggregate()
else:
__magic_name__ : List[Any] = {}
for key in scores[0]:
__magic_name__ : str = [score[key] for score in scores]
return result
| 281 | 1 |
import logging
import torch
from torch import nn
from torch.nn import CrossEntropyLoss, MSELoss
from transformers.file_utils import add_start_docstrings, add_start_docstrings_to_model_forward
from transformers.models.bert.modeling_bert import (
BERT_INPUTS_DOCSTRING,
BERT_START_DOCSTRING,
BertEncoder,
BertModel,
BertPreTrainedModel,
)
snake_case : Dict = logging.getLogger(__name__)
class _snake_case ( snake_case ):
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a=None , _a=None ):
__magic_name__ : List[Any] = self.layer[current_layer](_a , _a , head_mask[current_layer] )
__magic_name__ : Optional[Any] = layer_outputs[0]
return hidden_states
@add_start_docstrings(
'The bare Bert Model transformer with PABEE outputting raw hidden-states without any specific head on top.' , snake_case , )
class _snake_case ( snake_case ):
def __init__( self , _a ):
super().__init__(_a )
__magic_name__ : Tuple = BertEncoderWithPabee(_a )
self.init_weights()
__magic_name__ : Optional[Any] = 0
__magic_name__ : Dict = 0
__magic_name__ : Optional[Any] = 0
__magic_name__ : Optional[Any] = 0
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = threshold
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[str] = patience
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = 0
__magic_name__ : Optional[int] = 0
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = self.inference_layers_num / self.inference_instances_num
__magic_name__ : Any = (
f'''*** Patience = {self.patience} Avg. Inference Layers = {avg_inf_layers:.2f} Speed Up ='''
f''' {1 - avg_inf_layers / self.config.num_hidden_layers:.2f} ***'''
)
print(_a )
@add_start_docstrings_to_model_forward(_a )
def SCREAMING_SNAKE_CASE ( self , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , _a=False , ):
if input_ids is not None and inputs_embeds is not None:
raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time" )
elif input_ids is not None:
__magic_name__ : Dict = input_ids.size()
elif inputs_embeds is not None:
__magic_name__ : Optional[Any] = inputs_embeds.size()[:-1]
else:
raise ValueError("You have to specify either input_ids or inputs_embeds" )
__magic_name__ : Optional[Any] = input_ids.device if input_ids is not None else inputs_embeds.device
if attention_mask is None:
__magic_name__ : Optional[int] = torch.ones(_a , device=_a )
if token_type_ids is None:
__magic_name__ : List[str] = torch.zeros(_a , dtype=torch.long , device=_a )
# We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length]
# ourselves in which case we just need to make it broadcastable to all heads.
__magic_name__ : torch.Tensor = self.get_extended_attention_mask(_a , _a , _a )
# If a 2D ou 3D attention mask is provided for the cross-attention
# we need to make broadcastable to [batch_size, num_heads, seq_length, seq_length]
if self.config.is_decoder and encoder_hidden_states is not None:
__magic_name__ , __magic_name__ , __magic_name__ : Optional[int] = encoder_hidden_states.size()
__magic_name__ : Tuple = (encoder_batch_size, encoder_sequence_length)
if encoder_attention_mask is None:
__magic_name__ : Dict = torch.ones(_a , device=_a )
__magic_name__ : Dict = self.invert_attention_mask(_a )
else:
__magic_name__ : Any = None
# Prepare head mask if needed
# 1.0 in head_mask indicate we keep the head
# attention_probs has shape bsz x n_heads x N x N
# input head_mask has shape [num_heads] or [num_hidden_layers x num_heads]
# and head_mask is converted to shape [num_hidden_layers x batch x num_heads x seq_length x seq_length]
__magic_name__ : str = self.get_head_mask(_a , self.config.num_hidden_layers )
__magic_name__ : List[Any] = self.embeddings(
input_ids=_a , position_ids=_a , token_type_ids=_a , inputs_embeds=_a )
__magic_name__ : Union[str, Any] = embedding_output
if self.training:
__magic_name__ : List[Any] = []
for i in range(self.config.num_hidden_layers ):
__magic_name__ : Any = self.encoder.adaptive_forward(
_a , current_layer=_a , attention_mask=_a , head_mask=_a )
__magic_name__ : Tuple = self.pooler(_a )
__magic_name__ : List[str] = output_layers[i](output_dropout(_a ) )
res.append(_a )
elif self.patience == 0: # Use all layers for inference
__magic_name__ : List[str] = self.encoder(
_a , attention_mask=_a , head_mask=_a , encoder_hidden_states=_a , encoder_attention_mask=_a , )
__magic_name__ : Optional[int] = self.pooler(encoder_outputs[0] )
__magic_name__ : str = [output_layers[self.config.num_hidden_layers - 1](_a )]
else:
__magic_name__ : Union[str, Any] = 0
__magic_name__ : Optional[Any] = None
__magic_name__ : Any = 0
for i in range(self.config.num_hidden_layers ):
calculated_layer_num += 1
__magic_name__ : List[str] = self.encoder.adaptive_forward(
_a , current_layer=_a , attention_mask=_a , head_mask=_a )
__magic_name__ : Union[str, Any] = self.pooler(_a )
__magic_name__ : Union[str, Any] = output_layers[i](_a )
if regression:
__magic_name__ : List[Any] = logits.detach()
if patient_result is not None:
__magic_name__ : Dict = patient_result.detach()
if (patient_result is not None) and torch.abs(patient_result - labels ) < self.regression_threshold:
patient_counter += 1
else:
__magic_name__ : str = 0
else:
__magic_name__ : str = logits.detach().argmax(dim=1 )
if patient_result is not None:
__magic_name__ : Dict = patient_result.detach().argmax(dim=1 )
if (patient_result is not None) and torch.all(labels.eq(_a ) ):
patient_counter += 1
else:
__magic_name__ : Tuple = 0
__magic_name__ : Dict = logits
if patient_counter == self.patience:
break
__magic_name__ : List[Any] = [patient_result]
self.inference_layers_num += calculated_layer_num
self.inference_instances_num += 1
return res
@add_start_docstrings(
'Bert Model transformer with PABEE and a sequence classification/regression head on top (a linear layer on top of\n the pooled output) e.g. for GLUE tasks. ' , snake_case , )
class _snake_case ( snake_case ):
def __init__( self , _a ):
super().__init__(_a )
__magic_name__ : Optional[Any] = config.num_labels
__magic_name__ : Optional[int] = BertModelWithPabee(_a )
__magic_name__ : Union[str, Any] = nn.Dropout(config.hidden_dropout_prob )
__magic_name__ : List[str] = nn.ModuleList(
[nn.Linear(config.hidden_size , self.config.num_labels ) for _ in range(config.num_hidden_layers )] )
self.init_weights()
@add_start_docstrings_to_model_forward(_a )
def SCREAMING_SNAKE_CASE ( self , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , _a=None , ):
__magic_name__ : Optional[Any] = self.bert(
input_ids=_a , attention_mask=_a , token_type_ids=_a , position_ids=_a , head_mask=_a , inputs_embeds=_a , output_dropout=self.dropout , output_layers=self.classifiers , regression=self.num_labels == 1 , )
__magic_name__ : Optional[Any] = (logits[-1],)
if labels is not None:
__magic_name__ : str = None
__magic_name__ : str = 0
for ix, logits_item in enumerate(_a ):
if self.num_labels == 1:
# We are doing regression
__magic_name__ : List[Any] = MSELoss()
__magic_name__ : Optional[int] = loss_fct(logits_item.view(-1 ) , labels.view(-1 ) )
else:
__magic_name__ : List[str] = CrossEntropyLoss()
__magic_name__ : Dict = loss_fct(logits_item.view(-1 , self.num_labels ) , labels.view(-1 ) )
if total_loss is None:
__magic_name__ : List[Any] = loss
else:
total_loss += loss * (ix + 1)
total_weights += ix + 1
__magic_name__ : List[str] = (total_loss / total_weights,) + outputs
return outputs
| 281 |
snake_case : Optional[int] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"
def lowerCAmelCase_ ( _snake_case : bytes ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ):
__magic_name__ : Tuple = F'''a bytes-like object is required, not \'{data.__class__.__name__}\''''
raise TypeError(_snake_case )
__magic_name__ : Optional[int] = "".join(bin(_snake_case )[2:].zfill(8 ) for byte in data )
__magic_name__ : List[Any] = len(_snake_case ) % 6 != 0
if padding_needed:
# The padding that will be added later
__magic_name__ : List[str] = B"=" * ((6 - len(_snake_case ) % 6) // 2)
# Append binary_stream with arbitrary binary digits (0's by default) to make its
# length a multiple of 6.
binary_stream += "0" * (6 - len(_snake_case ) % 6)
else:
__magic_name__ : List[str] = B""
# Encode every 6 binary digits to their corresponding Base64 character
return (
"".join(
B64_CHARSET[int(binary_stream[index : index + 6] , 2 )]
for index in range(0 , len(_snake_case ) , 6 ) ).encode()
+ padding
)
def lowerCAmelCase_ ( _snake_case : str ) -> bytes:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ) and not isinstance(_snake_case , _snake_case ):
__magic_name__ : List[str] = (
"argument should be a bytes-like object or ASCII string, "
F'''not \'{encoded_data.__class__.__name__}\''''
)
raise TypeError(_snake_case )
# In case encoded_data is a bytes-like object, make sure it contains only
# ASCII characters so we convert it to a string object
if isinstance(_snake_case , _snake_case ):
try:
__magic_name__ : List[Any] = encoded_data.decode("utf-8" )
except UnicodeDecodeError:
raise ValueError("base64 encoded data should only contain ASCII characters" )
__magic_name__ : List[str] = encoded_data.count("=" )
# Check if the encoded string contains non base64 characters
if padding:
assert all(
char in B64_CHARSET for char in encoded_data[:-padding] ), "Invalid base64 character(s) found."
else:
assert all(
char in B64_CHARSET for char in encoded_data ), "Invalid base64 character(s) found."
# Check the padding
assert len(_snake_case ) % 4 == 0 and padding < 3, "Incorrect padding"
if padding:
# Remove padding if there is one
__magic_name__ : Optional[int] = encoded_data[:-padding]
__magic_name__ : Dict = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )[: -padding * 2]
else:
__magic_name__ : Union[str, Any] = "".join(
bin(B64_CHARSET.index(_snake_case ) )[2:].zfill(6 ) for char in encoded_data )
__magic_name__ : List[Any] = [
int(binary_stream[index : index + 8] , 2 )
for index in range(0 , len(_snake_case ) , 8 )
]
return bytes(_snake_case )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
snake_case : Union[str, Any] = {
"configuration_bigbird_pegasus": [
"BIGBIRD_PEGASUS_PRETRAINED_CONFIG_ARCHIVE_MAP",
"BigBirdPegasusConfig",
"BigBirdPegasusOnnxConfig",
],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
snake_case : Tuple = [
"BIGBIRD_PEGASUS_PRETRAINED_MODEL_ARCHIVE_LIST",
"BigBirdPegasusForCausalLM",
"BigBirdPegasusForConditionalGeneration",
"BigBirdPegasusForQuestionAnswering",
"BigBirdPegasusForSequenceClassification",
"BigBirdPegasusModel",
"BigBirdPegasusPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_bigbird_pegasus import (
BIGBIRD_PEGASUS_PRETRAINED_CONFIG_ARCHIVE_MAP,
BigBirdPegasusConfig,
BigBirdPegasusOnnxConfig,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_bigbird_pegasus import (
BIGBIRD_PEGASUS_PRETRAINED_MODEL_ARCHIVE_LIST,
BigBirdPegasusForCausalLM,
BigBirdPegasusForConditionalGeneration,
BigBirdPegasusForQuestionAnswering,
BigBirdPegasusForSequenceClassification,
BigBirdPegasusModel,
BigBirdPegasusPreTrainedModel,
)
else:
import sys
snake_case : str = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 281 |
import unittest
import numpy as np
from transformers import RobertaPreLayerNormConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
import jax.numpy as jnp
from transformers.models.roberta_prelayernorm.modeling_flax_roberta_prelayernorm import (
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormModel,
)
class _snake_case ( unittest.TestCase ):
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=True , _a=True , _a=True , _a=99 , _a=32 , _a=5 , _a=4 , _a=37 , _a="gelu" , _a=0.1 , _a=0.1 , _a=512 , _a=16 , _a=2 , _a=0.02 , _a=4 , ):
__magic_name__ : List[Any] = parent
__magic_name__ : Optional[Any] = batch_size
__magic_name__ : Dict = seq_length
__magic_name__ : Union[str, Any] = is_training
__magic_name__ : Optional[Any] = use_attention_mask
__magic_name__ : Optional[Any] = use_token_type_ids
__magic_name__ : int = use_labels
__magic_name__ : List[Any] = vocab_size
__magic_name__ : Union[str, Any] = hidden_size
__magic_name__ : Optional[Any] = num_hidden_layers
__magic_name__ : int = num_attention_heads
__magic_name__ : Any = intermediate_size
__magic_name__ : List[Any] = hidden_act
__magic_name__ : List[Any] = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[Any] = max_position_embeddings
__magic_name__ : Tuple = type_vocab_size
__magic_name__ : List[str] = type_sequence_label_size
__magic_name__ : Dict = initializer_range
__magic_name__ : List[Any] = num_choices
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : List[Any] = None
if self.use_attention_mask:
__magic_name__ : List[str] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : str = None
if self.use_token_type_ids:
__magic_name__ : str = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
__magic_name__ : List[str] = RobertaPreLayerNormConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=_a , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[Any] = config_and_inputs
__magic_name__ : List[str] = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": attention_mask}
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.prepare_config_and_inputs()
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = config_and_inputs
__magic_name__ : Tuple = True
__magic_name__ : int = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
token_type_ids,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
# Copied from tests.models.roberta.test_modelling_flax_roberta.FlaxRobertaPreLayerNormModelTest with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta-base->andreasmadsen/efficient_mlm_m0.40
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = True
UpperCamelCase__ = (
(
FlaxRobertaPreLayerNormModel,
FlaxRobertaPreLayerNormForCausalLM,
FlaxRobertaPreLayerNormForMaskedLM,
FlaxRobertaPreLayerNormForSequenceClassification,
FlaxRobertaPreLayerNormForTokenClassification,
FlaxRobertaPreLayerNormForMultipleChoice,
FlaxRobertaPreLayerNormForQuestionAnswering,
)
if is_flax_available()
else ()
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = FlaxRobertaPreLayerNormModelTester(self )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_class_name in self.all_model_classes:
__magic_name__ : Optional[Any] = model_class_name.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(_a )
@require_flax
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = FlaxRobertaPreLayerNormForMaskedLM.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Union[str, Any] = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : List[str] = model(_a )[0]
__magic_name__ : str = [1, 11, 50_265]
self.assertEqual(list(output.shape ) , _a )
# compare the actual values for a slice.
__magic_name__ : List[str] = np.array(
[[[40.48_80, 18.01_99, -5.23_67], [-1.88_77, -4.08_85, 10.70_85], [-2.26_13, -5.61_10, 7.26_65]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = FlaxRobertaPreLayerNormModel.from_pretrained("andreasmadsen/efficient_mlm_m0.40" , from_pt=_a )
__magic_name__ : Tuple = np.array([[0, 31_414, 232, 328, 740, 1_140, 12_695, 69, 46_078, 1_588, 2]] , dtype=jnp.intaa )
__magic_name__ : Tuple = model(_a )[0]
# compare the actual values for a slice.
__magic_name__ : Dict = np.array(
[[[0.02_08, -0.03_56, 0.02_37], [-0.15_69, -0.04_11, -0.26_26], [0.18_79, 0.01_25, -0.00_89]]] , dtype=np.floataa )
self.assertTrue(np.allclose(output[:, :3, :3] , _a , atol=1e-4 ) )
| 281 | 1 |
from heapq import heappop, heappush
import numpy as np
def lowerCAmelCase_ ( _snake_case : np.ndarray , _snake_case : tuple[int, int] , _snake_case : tuple[int, int] , _snake_case : bool , ) -> tuple[float | int, list[tuple[int, int]]]:
'''simple docstring'''
__magic_name__ , __magic_name__ : Union[str, Any] = grid.shape
__magic_name__ : List[Any] = [-1, 1, 0, 0]
__magic_name__ : Union[str, Any] = [0, 0, -1, 1]
if allow_diagonal:
dx += [-1, -1, 1, 1]
dy += [-1, 1, -1, 1]
__magic_name__ , __magic_name__ : List[Any] = [(0, source)], set()
__magic_name__ : Tuple = np.full((rows, cols) , np.inf )
__magic_name__ : int = 0
__magic_name__ : Optional[int] = np.empty((rows, cols) , dtype=_snake_case )
__magic_name__ : Optional[Any] = None
while queue:
((__magic_name__) , (__magic_name__)) : Any = heappop(_snake_case )
if (x, y) in visited:
continue
visited.add((x, y) )
if (x, y) == destination:
__magic_name__ : Optional[int] = []
while (x, y) != source:
path.append((x, y) )
__magic_name__ , __magic_name__ : List[Any] = predecessors[x, y]
path.append(_snake_case ) # add the source manually
path.reverse()
return matrix[destination], path
for i in range(len(_snake_case ) ):
__magic_name__ , __magic_name__ : Optional[Any] = x + dx[i], y + dy[i]
if 0 <= nx < rows and 0 <= ny < cols:
__magic_name__ : Optional[int] = grid[nx][ny]
if next_node == 1 and matrix[nx, ny] > dist + 1:
heappush(_snake_case , (dist + 1, (nx, ny)) )
__magic_name__ : List[str] = dist + 1
__magic_name__ : Any = (x, y)
return np.inf, []
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
def lowerCAmelCase_ ( _snake_case : list[list[int | float]] ) -> int:
'''simple docstring'''
__magic_name__ : Any = len(_snake_case )
__magic_name__ : Optional[Any] = len(matrix[0] )
__magic_name__ : Union[str, Any] = min(_snake_case , _snake_case )
for row in range(_snake_case ):
# Check if diagonal element is not zero
if matrix[row][row] != 0:
# Eliminate all the elements below the diagonal
for col in range(row + 1 , _snake_case ):
__magic_name__ : Optional[Any] = matrix[col][row] / matrix[row][row]
for i in range(_snake_case , _snake_case ):
matrix[col][i] -= multiplier * matrix[row][i]
else:
# Find a non-zero diagonal element to swap rows
__magic_name__ : str = True
for i in range(row + 1 , _snake_case ):
if matrix[i][row] != 0:
__magic_name__ , __magic_name__ : List[str] = matrix[i], matrix[row]
__magic_name__ : Union[str, Any] = False
break
if reduce:
rank -= 1
for i in range(_snake_case ):
__magic_name__ : Any = matrix[i][rank]
# Reduce the row pointer by one to stay on the same row
row -= 1
return rank
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 | 1 |
import os
import torch
from ..logging import get_logger
from .constants import FSDP_PYTORCH_VERSION, MODEL_NAME, OPTIMIZER_NAME
from .versions import is_torch_version
if is_torch_version(">=", FSDP_PYTORCH_VERSION):
import torch.distributed.checkpoint as dist_cp
from torch.distributed.checkpoint.default_planner import DefaultLoadPlanner, DefaultSavePlanner
from torch.distributed.checkpoint.optimizer import load_sharded_optimizer_state_dict
from torch.distributed.fsdp.fully_sharded_data_parallel import FullyShardedDataParallel as FSDP
from torch.distributed.fsdp.fully_sharded_data_parallel import StateDictType
snake_case : Union[str, Any] = get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Tuple , _snake_case : Any , _snake_case : int , _snake_case : Optional[int]=0 ) -> int:
'''simple docstring'''
os.makedirs(_snake_case , exist_ok=_snake_case )
with FSDP.state_dict_type(
_snake_case , fsdp_plugin.state_dict_type , fsdp_plugin.state_dict_config , fsdp_plugin.optim_state_dict_config ):
__magic_name__ : Optional[Any] = model.state_dict()
if fsdp_plugin.state_dict_type == StateDictType.FULL_STATE_DICT:
__magic_name__ : Optional[Any] = F'''{MODEL_NAME}.bin''' if model_index == 0 else F'''{MODEL_NAME}_{model_index}.bin'''
__magic_name__ : List[Any] = os.path.join(_snake_case , _snake_case )
if accelerator.process_index == 0:
logger.info(F'''Saving model to {output_model_file}''' )
torch.save(_snake_case , _snake_case )
logger.info(F'''Model saved to {output_model_file}''' )
elif fsdp_plugin.state_dict_type == StateDictType.LOCAL_STATE_DICT:
__magic_name__ : Optional[int] = (
F'''{MODEL_NAME}_rank{accelerator.process_index}.bin'''
if model_index == 0
else F'''{MODEL_NAME}_{model_index}_rank{accelerator.process_index}.bin'''
)
__magic_name__ : Tuple = os.path.join(_snake_case , _snake_case )
logger.info(F'''Saving model to {output_model_file}''' )
torch.save(_snake_case , _snake_case )
logger.info(F'''Model saved to {output_model_file}''' )
elif fsdp_plugin.state_dict_type == StateDictType.SHARDED_STATE_DICT:
__magic_name__ : Any = os.path.join(_snake_case , F'''{MODEL_NAME}_{model_index}''' )
os.makedirs(_snake_case , exist_ok=_snake_case )
logger.info(F'''Saving model to {ckpt_dir}''' )
__magic_name__ : Union[str, Any] = {"model": state_dict}
dist_cp.save_state_dict(
state_dict=_snake_case , storage_writer=dist_cp.FileSystemWriter(_snake_case ) , planner=DefaultSavePlanner() , )
logger.info(F'''Model saved to {ckpt_dir}''' )
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : int , _snake_case : Optional[Any] , _snake_case : int , _snake_case : Union[str, Any]=0 ) -> int:
'''simple docstring'''
accelerator.wait_for_everyone()
with FSDP.state_dict_type(
_snake_case , fsdp_plugin.state_dict_type , fsdp_plugin.state_dict_config , fsdp_plugin.optim_state_dict_config ):
if fsdp_plugin.state_dict_type == StateDictType.FULL_STATE_DICT:
if type(_snake_case ) != FSDP and accelerator.process_index != 0:
if not fsdp_plugin.sync_module_states:
raise ValueError(
"Set the `sync_module_states` flag to `True` so that model states are synced across processes when "
"initializing FSDP object" )
return
__magic_name__ : int = F'''{MODEL_NAME}.bin''' if model_index == 0 else F'''{MODEL_NAME}_{model_index}.bin'''
__magic_name__ : List[Any] = os.path.join(_snake_case , _snake_case )
logger.info(F'''Loading model from {input_model_file}''' )
__magic_name__ : Any = torch.load(_snake_case )
logger.info(F'''Model loaded from {input_model_file}''' )
elif fsdp_plugin.state_dict_type == StateDictType.LOCAL_STATE_DICT:
__magic_name__ : str = (
F'''{MODEL_NAME}_rank{accelerator.process_index}.bin'''
if model_index == 0
else F'''{MODEL_NAME}_{model_index}_rank{accelerator.process_index}.bin'''
)
__magic_name__ : Optional[int] = os.path.join(_snake_case , _snake_case )
logger.info(F'''Loading model from {input_model_file}''' )
__magic_name__ : int = torch.load(_snake_case )
logger.info(F'''Model loaded from {input_model_file}''' )
elif fsdp_plugin.state_dict_type == StateDictType.SHARDED_STATE_DICT:
__magic_name__ : Any = (
os.path.join(_snake_case , F'''{MODEL_NAME}_{model_index}''' )
if F'''{MODEL_NAME}''' not in input_dir
else input_dir
)
logger.info(F'''Loading model from {ckpt_dir}''' )
__magic_name__ : int = {"model": model.state_dict()}
dist_cp.load_state_dict(
state_dict=_snake_case , storage_reader=dist_cp.FileSystemReader(_snake_case ) , planner=DefaultLoadPlanner() , )
__magic_name__ : str = state_dict["model"]
logger.info(F'''Model loaded from {ckpt_dir}''' )
model.load_state_dict(_snake_case )
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : List[str] , _snake_case : Optional[int] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Tuple=0 ) -> int:
'''simple docstring'''
os.makedirs(_snake_case , exist_ok=_snake_case )
with FSDP.state_dict_type(
_snake_case , fsdp_plugin.state_dict_type , fsdp_plugin.state_dict_config , fsdp_plugin.optim_state_dict_config ):
__magic_name__ : Optional[int] = FSDP.optim_state_dict(_snake_case , _snake_case )
if fsdp_plugin.state_dict_type == StateDictType.FULL_STATE_DICT:
if accelerator.process_index == 0:
__magic_name__ : List[Any] = (
F'''{OPTIMIZER_NAME}.bin''' if optimizer_index == 0 else F'''{OPTIMIZER_NAME}_{optimizer_index}.bin'''
)
__magic_name__ : Any = os.path.join(_snake_case , _snake_case )
logger.info(F'''Saving Optimizer state to {output_optimizer_file}''' )
torch.save(_snake_case , _snake_case )
logger.info(F'''Optimizer state saved in {output_optimizer_file}''' )
else:
__magic_name__ : Optional[int] = os.path.join(_snake_case , F'''{OPTIMIZER_NAME}_{optimizer_index}''' )
os.makedirs(_snake_case , exist_ok=_snake_case )
logger.info(F'''Saving Optimizer state to {ckpt_dir}''' )
dist_cp.save_state_dict(
state_dict={"optimizer": optim_state} , storage_writer=dist_cp.FileSystemWriter(_snake_case ) , planner=DefaultSavePlanner() , )
logger.info(F'''Optimizer state saved in {ckpt_dir}''' )
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Any , _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : Any , _snake_case : Union[str, Any]=0 ) -> int:
'''simple docstring'''
accelerator.wait_for_everyone()
with FSDP.state_dict_type(
_snake_case , fsdp_plugin.state_dict_type , fsdp_plugin.state_dict_config , fsdp_plugin.optim_state_dict_config ):
if fsdp_plugin.state_dict_type == StateDictType.FULL_STATE_DICT:
__magic_name__ : Optional[Any] = None
# below check should work but currently it isn't working (mostly opytorch issue),
# in the meantime disabling it at the cost of excess memory usage
# if accelerator.process_index == 0 or not fsdp_plugin.optim_state_dict_config.rank0_only:
__magic_name__ : Optional[int] = (
F'''{OPTIMIZER_NAME}.bin''' if optimizer_index == 0 else F'''{OPTIMIZER_NAME}_{optimizer_index}.bin'''
)
__magic_name__ : str = os.path.join(_snake_case , _snake_case )
logger.info(F'''Loading Optimizer state from {input_optimizer_file}''' )
__magic_name__ : Optional[int] = torch.load(_snake_case )
logger.info(F'''Optimizer state loaded from {input_optimizer_file}''' )
else:
__magic_name__ : Dict = (
os.path.join(_snake_case , F'''{OPTIMIZER_NAME}_{optimizer_index}''' )
if F'''{OPTIMIZER_NAME}''' not in input_dir
else input_dir
)
logger.info(F'''Loading Optimizer from {ckpt_dir}''' )
__magic_name__ : List[str] = load_sharded_optimizer_state_dict(
model_state_dict=model.state_dict() , optimizer_key="optimizer" , storage_reader=dist_cp.FileSystemReader(_snake_case ) , )
__magic_name__ : Union[str, Any] = optim_state["optimizer"]
logger.info(F'''Optimizer loaded from {ckpt_dir}''' )
__magic_name__ : Dict = FSDP.optim_state_dict_to_load(_snake_case , _snake_case , _snake_case )
optimizer.load_state_dict(_snake_case )
| 281 |
import argparse
import collections
import json
import os
import re
import string
import sys
import numpy as np
snake_case : Dict = re.compile(R"\b(a|an|the)\b", re.UNICODE)
snake_case : Optional[int] = None
def lowerCAmelCase_ ( ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser("Official evaluation script for SQuAD version 2.0." )
parser.add_argument("data_file" , metavar="data.json" , help="Input data JSON file." )
parser.add_argument("pred_file" , metavar="pred.json" , help="Model predictions." )
parser.add_argument(
"--out-file" , "-o" , metavar="eval.json" , help="Write accuracy metrics to file (default is stdout)." )
parser.add_argument(
"--na-prob-file" , "-n" , metavar="na_prob.json" , help="Model estimates of probability of no answer." )
parser.add_argument(
"--na-prob-thresh" , "-t" , type=_snake_case , default=1.0 , help="Predict \"\" if no-answer probability exceeds this (default = 1.0)." , )
parser.add_argument(
"--out-image-dir" , "-p" , metavar="out_images" , default=_snake_case , help="Save precision-recall curves to directory." )
parser.add_argument("--verbose" , "-v" , action="store_true" )
if len(sys.argv ) == 1:
parser.print_help()
sys.exit(1 )
return parser.parse_args()
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Optional[int] = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : str = bool(qa["answers"]["text"] )
return qid_to_has_ans
def lowerCAmelCase_ ( _snake_case : Union[str, Any] ) -> Optional[Any]:
'''simple docstring'''
def remove_articles(_snake_case : List[str] ):
return ARTICLES_REGEX.sub(" " , _snake_case )
def white_space_fix(_snake_case : Optional[int] ):
return " ".join(text.split() )
def remove_punc(_snake_case : Optional[int] ):
__magic_name__ : Dict = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(_snake_case : str ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(_snake_case ) ) ) )
def lowerCAmelCase_ ( _snake_case : Any ) -> Optional[Any]:
'''simple docstring'''
if not s:
return []
return normalize_answer(_snake_case ).split()
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Dict ) -> Tuple:
'''simple docstring'''
return int(normalize_answer(_snake_case ) == normalize_answer(_snake_case ) )
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : int ) -> str:
'''simple docstring'''
__magic_name__ : Any = get_tokens(_snake_case )
__magic_name__ : Optional[int] = get_tokens(_snake_case )
__magic_name__ : Tuple = collections.Counter(_snake_case ) & collections.Counter(_snake_case )
__magic_name__ : Tuple = sum(common.values() )
if len(_snake_case ) == 0 or len(_snake_case ) == 0:
# If either is no-answer, then F1 is 1 if they agree, 0 otherwise
return int(gold_toks == pred_toks )
if num_same == 0:
return 0
__magic_name__ : Dict = 1.0 * num_same / len(_snake_case )
__magic_name__ : Optional[Any] = 1.0 * num_same / len(_snake_case )
__magic_name__ : List[Any] = (2 * precision * recall) / (precision + recall)
return fa
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = {}
__magic_name__ : int = {}
for article in dataset:
for p in article["paragraphs"]:
for qa in p["qas"]:
__magic_name__ : Union[str, Any] = qa["id"]
__magic_name__ : Any = [t for t in qa["answers"]["text"] if normalize_answer(_snake_case )]
if not gold_answers:
# For unanswerable questions, only correct answer is empty string
__magic_name__ : Tuple = [""]
if qid not in preds:
print(F'''Missing prediction for {qid}''' )
continue
__magic_name__ : Any = preds[qid]
# Take max over all gold answers
__magic_name__ : List[Any] = max(compute_exact(_snake_case , _snake_case ) for a in gold_answers )
__magic_name__ : int = max(compute_fa(_snake_case , _snake_case ) for a in gold_answers )
return exact_scores, fa_scores
def lowerCAmelCase_ ( _snake_case : Optional[Any] , _snake_case : List[Any] , _snake_case : Optional[int] , _snake_case : Dict ) -> Union[str, Any]:
'''simple docstring'''
__magic_name__ : str = {}
for qid, s in scores.items():
__magic_name__ : Dict = na_probs[qid] > na_prob_thresh
if pred_na:
__magic_name__ : str = float(not qid_to_has_ans[qid] )
else:
__magic_name__ : Optional[int] = s
return new_scores
def lowerCAmelCase_ ( _snake_case : List[Any] , _snake_case : List[str] , _snake_case : Tuple=None ) -> Tuple:
'''simple docstring'''
if not qid_list:
__magic_name__ : Any = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores.values() ) / total),
("f1", 100.0 * sum(fa_scores.values() ) / total),
("total", total),
] )
else:
__magic_name__ : Tuple = len(_snake_case )
return collections.OrderedDict(
[
("exact", 100.0 * sum(exact_scores[k] for k in qid_list ) / total),
("f1", 100.0 * sum(fa_scores[k] for k in qid_list ) / total),
("total", total),
] )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : str , _snake_case : str ) -> Dict:
'''simple docstring'''
for k in new_eval:
__magic_name__ : int = new_eval[k]
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Dict , _snake_case : Optional[Any] , _snake_case : Union[str, Any] ) -> str:
'''simple docstring'''
plt.step(_snake_case , _snake_case , color="b" , alpha=0.2 , where="post" )
plt.fill_between(_snake_case , _snake_case , step="post" , alpha=0.2 , color="b" )
plt.xlabel("Recall" )
plt.ylabel("Precision" )
plt.xlim([0.0, 1.05] )
plt.ylim([0.0, 1.05] )
plt.title(_snake_case )
plt.savefig(_snake_case )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Any , _snake_case : Optional[int] , _snake_case : List[Any] , _snake_case : Optional[int]=None , _snake_case : int=None ) -> str:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
__magic_name__ : Optional[int] = 0.0
__magic_name__ : str = 1.0
__magic_name__ : str = 0.0
__magic_name__ : List[str] = [1.0]
__magic_name__ : str = [0.0]
__magic_name__ : Optional[Any] = 0.0
for i, qid in enumerate(_snake_case ):
if qid_to_has_ans[qid]:
true_pos += scores[qid]
__magic_name__ : List[str] = true_pos / float(i + 1 )
__magic_name__ : Any = true_pos / float(_snake_case )
if i == len(_snake_case ) - 1 or na_probs[qid] != na_probs[qid_list[i + 1]]:
# i.e., if we can put a threshold after this point
avg_prec += cur_p * (cur_r - recalls[-1])
precisions.append(_snake_case )
recalls.append(_snake_case )
if out_image:
plot_pr_curve(_snake_case , _snake_case , _snake_case , _snake_case )
return {"ap": 100.0 * avg_prec}
def lowerCAmelCase_ ( _snake_case : Tuple , _snake_case : Optional[Any] , _snake_case : Optional[int] , _snake_case : Optional[Any] , _snake_case : Any , _snake_case : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
if out_image_dir and not os.path.exists(_snake_case ):
os.makedirs(_snake_case )
__magic_name__ : Any = sum(1 for v in qid_to_has_ans.values() if v )
if num_true_pos == 0:
return
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_exact.png" ) , title="Precision-Recall curve for Exact Match score" , )
__magic_name__ : Union[str, Any] = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_f1.png" ) , title="Precision-Recall curve for F1 score" , )
__magic_name__ : str = {k: float(_snake_case ) for k, v in qid_to_has_ans.items()}
__magic_name__ : str = make_precision_recall_eval(
_snake_case , _snake_case , _snake_case , _snake_case , out_image=os.path.join(_snake_case , "pr_oracle.png" ) , title="Oracle Precision-Recall curve (binary task of HasAns vs. NoAns)" , )
merge_eval(_snake_case , _snake_case , "pr_exact" )
merge_eval(_snake_case , _snake_case , "pr_f1" )
merge_eval(_snake_case , _snake_case , "pr_oracle" )
def lowerCAmelCase_ ( _snake_case : int , _snake_case : Optional[Any] , _snake_case : List[str] , _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
if not qid_list:
return
__magic_name__ : Dict = [na_probs[k] for k in qid_list]
__magic_name__ : str = np.ones_like(_snake_case ) / float(len(_snake_case ) )
plt.hist(_snake_case , weights=_snake_case , bins=20 , range=(0.0, 1.0) )
plt.xlabel("Model probability of no-answer" )
plt.ylabel("Proportion of dataset" )
plt.title(F'''Histogram of no-answer probability: {name}''' )
plt.savefig(os.path.join(_snake_case , F'''na_prob_hist_{name}.png''' ) )
plt.clf()
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Tuple , _snake_case : List[str] , _snake_case : Dict ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = sum(1 for k in qid_to_has_ans if not qid_to_has_ans[k] )
__magic_name__ : List[str] = num_no_ans
__magic_name__ : Dict = cur_score
__magic_name__ : Dict = 0.0
__magic_name__ : Any = sorted(_snake_case , key=lambda _snake_case : na_probs[k] )
for i, qid in enumerate(_snake_case ):
if qid not in scores:
continue
if qid_to_has_ans[qid]:
__magic_name__ : Union[str, Any] = scores[qid]
else:
if preds[qid]:
__magic_name__ : List[Any] = -1
else:
__magic_name__ : Optional[int] = 0
cur_score += diff
if cur_score > best_score:
__magic_name__ : Optional[int] = cur_score
__magic_name__ : List[Any] = na_probs[qid]
return 100.0 * best_score / len(_snake_case ), best_thresh
def lowerCAmelCase_ ( _snake_case : int , _snake_case : str , _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Dict ) -> Optional[Any]:
'''simple docstring'''
__magic_name__ , __magic_name__ : List[str] = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ , __magic_name__ : int = find_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case )
__magic_name__ : Optional[int] = best_exact
__magic_name__ : List[Any] = exact_thresh
__magic_name__ : Dict = best_fa
__magic_name__ : Any = fa_thresh
def lowerCAmelCase_ ( ) -> int:
'''simple docstring'''
with open(OPTS.data_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
__magic_name__ : List[Any] = dataset_json["data"]
with open(OPTS.pred_file ) as f:
__magic_name__ : Optional[Any] = json.load(_snake_case )
if OPTS.na_prob_file:
with open(OPTS.na_prob_file ) as f:
__magic_name__ : Any = json.load(_snake_case )
else:
__magic_name__ : Any = {k: 0.0 for k in preds}
__magic_name__ : str = make_qid_to_has_ans(_snake_case ) # maps qid to True/False
__magic_name__ : Tuple = [k for k, v in qid_to_has_ans.items() if v]
__magic_name__ : Optional[Any] = [k for k, v in qid_to_has_ans.items() if not v]
__magic_name__ , __magic_name__ : Union[str, Any] = get_raw_scores(_snake_case , _snake_case )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : Optional[Any] = apply_no_ans_threshold(_snake_case , _snake_case , _snake_case , OPTS.na_prob_thresh )
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case )
if has_ans_qids:
__magic_name__ : int = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "HasAns" )
if no_ans_qids:
__magic_name__ : List[Any] = make_eval_dict(_snake_case , _snake_case , qid_list=_snake_case )
merge_eval(_snake_case , _snake_case , "NoAns" )
if OPTS.na_prob_file:
find_all_best_thresh(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , _snake_case )
if OPTS.na_prob_file and OPTS.out_image_dir:
run_precision_recall_analysis(_snake_case , _snake_case , _snake_case , _snake_case , _snake_case , OPTS.out_image_dir )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "hasAns" )
histogram_na_prob(_snake_case , _snake_case , OPTS.out_image_dir , "noAns" )
if OPTS.out_file:
with open(OPTS.out_file , "w" ) as f:
json.dump(_snake_case , _snake_case )
else:
print(json.dumps(_snake_case , indent=2 ) )
if __name__ == "__main__":
snake_case : int = parse_args()
if OPTS.out_image_dir:
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
main()
| 281 | 1 |
import shutil
import tempfile
import unittest
import numpy as np
import pytest
from transformers import is_speech_available, is_vision_available
from transformers.testing_utils import require_torch
if is_vision_available():
from transformers import TvltImageProcessor
if is_speech_available():
from transformers import TvltFeatureExtractor
from transformers import TvltProcessor
@require_torch
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "ZinengTang/tvlt-base"
__magic_name__ : int = tempfile.mkdtemp()
def SCREAMING_SNAKE_CASE ( self , **_a ):
return TvltImageProcessor.from_pretrained(self.checkpoint , **_a )
def SCREAMING_SNAKE_CASE ( self , **_a ):
return TvltFeatureExtractor.from_pretrained(self.checkpoint , **_a )
def SCREAMING_SNAKE_CASE ( self ):
shutil.rmtree(self.tmpdirname )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = self.get_image_processor()
__magic_name__ : str = self.get_feature_extractor()
__magic_name__ : List[Any] = TvltProcessor(image_processor=_a , feature_extractor=_a )
processor.save_pretrained(self.tmpdirname )
__magic_name__ : Any = TvltProcessor.from_pretrained(self.tmpdirname )
self.assertIsInstance(processor.feature_extractor , _a )
self.assertIsInstance(processor.image_processor , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = self.get_image_processor()
__magic_name__ : Union[str, Any] = self.get_feature_extractor()
__magic_name__ : Any = TvltProcessor(image_processor=_a , feature_extractor=_a )
__magic_name__ : List[Any] = np.ones([12_000] )
__magic_name__ : Optional[int] = feature_extractor(_a , return_tensors="np" )
__magic_name__ : List[str] = processor(audio=_a , return_tensors="np" )
for key in audio_dict.keys():
self.assertAlmostEqual(audio_dict[key].sum() , input_processor[key].sum() , delta=1e-2 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.get_image_processor()
__magic_name__ : Union[str, Any] = self.get_feature_extractor()
__magic_name__ : int = TvltProcessor(image_processor=_a , feature_extractor=_a )
__magic_name__ : int = np.ones([3, 224, 224] )
__magic_name__ : int = image_processor(_a , return_tensors="np" )
__magic_name__ : List[str] = processor(images=_a , return_tensors="np" )
for key in image_dict.keys():
self.assertAlmostEqual(image_dict[key].sum() , input_processor[key].sum() , delta=1e-2 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = self.get_image_processor()
__magic_name__ : Dict = self.get_feature_extractor()
__magic_name__ : Tuple = TvltProcessor(image_processor=_a , feature_extractor=_a )
__magic_name__ : Tuple = np.ones([12_000] )
__magic_name__ : List[Any] = np.ones([3, 224, 224] )
__magic_name__ : str = processor(audio=_a , images=_a )
self.assertListEqual(list(inputs.keys() ) , ["audio_values", "audio_mask", "pixel_values", "pixel_mask"] )
# test if it raises when no input is passed
with pytest.raises(_a ):
processor()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = self.get_image_processor()
__magic_name__ : str = self.get_feature_extractor()
__magic_name__ : List[Any] = TvltProcessor(image_processor=_a , feature_extractor=_a )
self.assertListEqual(
processor.model_input_names , image_processor.model_input_names + feature_extractor.model_input_names , msg="`processor` and `image_processor`+`feature_extractor` model input names do not match" , )
| 281 |
import unittest
from transformers import BigBirdTokenizer, BigBirdTokenizerFast
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, require_torch, slow
from transformers.utils import cached_property
from ...test_tokenization_common import TokenizerTesterMixin
snake_case : str = "▁"
snake_case : List[Any] = get_tests_dir("fixtures/test_sentencepiece.model")
@require_sentencepiece
@require_tokenizers
class _snake_case ( snake_case , unittest.TestCase ):
UpperCamelCase__ = BigBirdTokenizer
UpperCamelCase__ = BigBirdTokenizerFast
UpperCamelCase__ = True
UpperCamelCase__ = True
def SCREAMING_SNAKE_CASE ( self ):
super().setUp()
__magic_name__ : Optional[Any] = self.tokenizer_class(_a , keep_accents=_a )
tokenizer.save_pretrained(self.tmpdirname )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = "<s>"
__magic_name__ : Dict = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(_a ) , _a )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(_a ) , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : 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] , "[MASK]" )
self.assertEqual(len(_a ) , 1_004 )
def SCREAMING_SNAKE_CASE ( self ):
self.assertEqual(self.get_tokenizer().vocab_size , 1_000 )
def SCREAMING_SNAKE_CASE ( self ):
if not self.test_rust_tokenizer:
return
__magic_name__ : Dict = self.get_tokenizer()
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Any = "I was born in 92000, and this is falsé."
__magic_name__ : Dict = tokenizer.tokenize(_a )
__magic_name__ : Any = rust_tokenizer.tokenize(_a )
self.assertListEqual(_a , _a )
__magic_name__ : List[Any] = tokenizer.encode(_a , add_special_tokens=_a )
__magic_name__ : List[str] = rust_tokenizer.encode(_a , add_special_tokens=_a )
self.assertListEqual(_a , _a )
__magic_name__ : str = self.get_rust_tokenizer()
__magic_name__ : Dict = tokenizer.encode(_a )
__magic_name__ : Optional[int] = rust_tokenizer.encode(_a )
self.assertListEqual(_a , _a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = BigBirdTokenizer(_a , keep_accents=_a )
__magic_name__ : str = tokenizer.tokenize("This is a test" )
self.assertListEqual(_a , ["▁This", "▁is", "▁a", "▁t", "est"] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(_a ) , [285, 46, 10, 170, 382] , )
__magic_name__ : Dict = tokenizer.tokenize("I was born in 92000, and this is falsé." )
self.assertListEqual(
_a , [
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",
".",
] , )
__magic_name__ : Union[str, Any] = tokenizer.convert_tokens_to_ids(_a )
self.assertListEqual(
_a , [8, 21, 84, 55, 24, 19, 7, 0, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 0, 4] , )
__magic_name__ : int = tokenizer.convert_ids_to_tokens(_a )
self.assertListEqual(
_a , [
SPIECE_UNDERLINE + "I",
SPIECE_UNDERLINE + "was",
SPIECE_UNDERLINE + "b",
"or",
"n",
SPIECE_UNDERLINE + "in",
SPIECE_UNDERLINE + "",
"<unk>",
"2",
"0",
"0",
"0",
",",
SPIECE_UNDERLINE + "and",
SPIECE_UNDERLINE + "this",
SPIECE_UNDERLINE + "is",
SPIECE_UNDERLINE + "f",
"al",
"s",
"<unk>",
".",
] , )
@cached_property
def SCREAMING_SNAKE_CASE ( self ):
return BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = "Hello World!"
__magic_name__ : Dict = [65, 18_536, 2_260, 101, 66]
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = (
"This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"
" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"
)
# fmt: off
__magic_name__ : List[str] = [65, 871, 419, 358, 946, 991, 2_521, 452, 358, 1_357, 387, 7_751, 3_536, 112, 985, 456, 126, 865, 938, 5_400, 5_734, 458, 1_368, 467, 786, 2_462, 5_246, 1_159, 633, 865, 4_519, 457, 582, 852, 2_557, 427, 916, 508, 405, 34_324, 497, 391, 408, 11_342, 1_244, 385, 100, 938, 985, 456, 574, 362, 12_597, 3_200, 3_129, 1_172, 66] # noqa: E231
# fmt: on
self.assertListEqual(_a , self.big_tokenizer.encode(_a ) )
@require_torch
@slow
def SCREAMING_SNAKE_CASE ( self ):
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
__magic_name__ : Optional[Any] = list(self.big_tokenizer.get_vocab().keys() )[:10]
__magic_name__ : List[Any] = " ".join(_a )
__magic_name__ : Any = self.big_tokenizer.encode_plus(_a , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : Union[str, Any] = self.big_tokenizer.batch_encode_plus(
[sequence + " " + sequence] , return_tensors="pt" , return_token_type_ids=_a )
__magic_name__ : List[str] = BigBirdConfig(attention_type="original_full" )
__magic_name__ : Optional[int] = BigBirdModel(_a )
assert model.get_input_embeddings().weight.shape[0] >= self.big_tokenizer.vocab_size
with torch.no_grad():
model(**_a )
model(**_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
__magic_name__ : int = tokenizer.decode(tokenizer("Paris is the [MASK]." ).input_ids )
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]" )
@slow
def SCREAMING_SNAKE_CASE ( self ):
# fmt: off
__magic_name__ : Optional[Any] = {"input_ids": [[65, 39_286, 458, 36_335, 2_001, 456, 13_073, 13_266, 455, 113, 7_746, 1_741, 11_157, 391, 13_073, 13_266, 455, 113, 3_967, 35_412, 113, 4_936, 109, 3_870, 2_377, 113, 30_084, 45_720, 458, 134, 17_496, 112, 503, 11_672, 113, 118, 112, 5_665, 13_347, 38_687, 112, 1_496, 31_389, 112, 3_268, 47_264, 134, 962, 112, 16_377, 8_035, 23_130, 430, 12_169, 15_518, 28_592, 458, 146, 41_697, 109, 391, 12_169, 15_518, 16_689, 458, 146, 41_358, 109, 452, 726, 4_034, 111, 763, 35_412, 5_082, 388, 1_903, 111, 9_051, 391, 2_870, 48_918, 1_900, 1_123, 550, 998, 112, 9_586, 15_985, 455, 391, 410, 22_955, 37_636, 114, 66], [65, 448, 17_496, 419, 3_663, 385, 763, 113, 27_533, 2_870, 3_283, 13_043, 1_639, 24_713, 523, 656, 24_013, 18_550, 2_521, 517, 27_014, 21_244, 420, 1_212, 1_465, 391, 927, 4_833, 388, 578, 11_786, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [65, 484, 2_169, 7_687, 21_932, 18_146, 726, 363, 17_032, 3_391, 114, 66, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], "attention_mask": [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=_a , model_name="google/bigbird-roberta-base" , revision="215c99f1600e06f83acce68422f2035b2b5c3510" , )
| 281 | 1 |
import re
import warnings
from contextlib import contextmanager
from ...processing_utils import ProcessorMixin
class _snake_case ( snake_case ):
UpperCamelCase__ = ['image_processor', 'tokenizer']
UpperCamelCase__ = 'AutoImageProcessor'
UpperCamelCase__ = 'AutoTokenizer'
def __init__( self , _a=None , _a=None , **_a ):
__magic_name__ : Optional[int] = None
if "feature_extractor" in kwargs:
warnings.warn(
"The `feature_extractor` argument is deprecated and will be removed in v5, use `image_processor`"
" instead." , _a , )
__magic_name__ : Union[str, Any] = kwargs.pop("feature_extractor" )
__magic_name__ : List[str] = image_processor if image_processor is not None else feature_extractor
if image_processor is None:
raise ValueError("You need to specify an `image_processor`." )
if tokenizer is None:
raise ValueError("You need to specify a `tokenizer`." )
super().__init__(_a , _a )
__magic_name__ : Tuple = self.image_processor
__magic_name__ : Union[str, Any] = False
def __call__( self , *_a , **_a ):
# For backward compatibility
if self._in_target_context_manager:
return self.current_processor(*_a , **_a )
__magic_name__ : str = kwargs.pop("images" , _a )
__magic_name__ : Dict = kwargs.pop("text" , _a )
if len(_a ) > 0:
__magic_name__ : List[Any] = args[0]
__magic_name__ : Any = args[1:]
if images is None and text is None:
raise ValueError("You need to specify either an `images` or `text` input to process." )
if images is not None:
__magic_name__ : Tuple = self.image_processor(_a , *_a , **_a )
if text is not None:
__magic_name__ : str = self.tokenizer(_a , **_a )
if text is None:
return inputs
elif images is None:
return encodings
else:
__magic_name__ : List[str] = encodings["input_ids"]
return inputs
def SCREAMING_SNAKE_CASE ( self , *_a , **_a ):
return self.tokenizer.batch_decode(*_a , **_a )
def SCREAMING_SNAKE_CASE ( self , *_a , **_a ):
return self.tokenizer.decode(*_a , **_a )
@contextmanager
def SCREAMING_SNAKE_CASE ( self ):
warnings.warn(
"`as_target_processor` is deprecated and will be removed in v5 of Transformers. You can process your "
"labels by using the argument `text` of the regular `__call__` method (either in the same call as "
"your images inputs, or in a separate call." )
__magic_name__ : Any = True
__magic_name__ : Dict = self.tokenizer
yield
__magic_name__ : Any = self.image_processor
__magic_name__ : Optional[int] = False
def SCREAMING_SNAKE_CASE ( self , _a , _a=False , _a=None ):
if added_vocab is None:
__magic_name__ : Optional[int] = self.tokenizer.get_added_vocab()
__magic_name__ : int = {}
while tokens:
__magic_name__ : Optional[Any] = re.search(r"<s_(.*?)>" , _a , re.IGNORECASE )
if start_token is None:
break
__magic_name__ : List[str] = start_token.group(1 )
__magic_name__ : Optional[Any] = re.search(rf'''</s_{key}>''' , _a , re.IGNORECASE )
__magic_name__ : Union[str, Any] = start_token.group()
if end_token is None:
__magic_name__ : str = tokens.replace(_a , "" )
else:
__magic_name__ : Optional[int] = end_token.group()
__magic_name__ : Dict = re.escape(_a )
__magic_name__ : List[str] = re.escape(_a )
__magic_name__ : str = re.search(f'''{start_token_escaped}(.*?){end_token_escaped}''' , _a , re.IGNORECASE )
if content is not None:
__magic_name__ : int = content.group(1 ).strip()
if r"<s_" in content and r"</s_" in content: # non-leaf node
__magic_name__ : List[str] = self.tokenajson(_a , is_inner_value=_a , added_vocab=_a )
if value:
if len(_a ) == 1:
__magic_name__ : Any = value[0]
__magic_name__ : int = value
else: # leaf nodes
__magic_name__ : str = []
for leaf in content.split(r"<sep/>" ):
__magic_name__ : Optional[Any] = leaf.strip()
if leaf in added_vocab and leaf[0] == "<" and leaf[-2:] == "/>":
__magic_name__ : Optional[Any] = leaf[1:-2] # for categorical special tokens
output[key].append(_a )
if len(output[key] ) == 1:
__magic_name__ : Any = output[key][0]
__magic_name__ : Optional[Any] = tokens[tokens.find(_a ) + len(_a ) :].strip()
if tokens[:6] == r"<sep/>": # non-leaf nodes
return [output] + self.tokenajson(tokens[6:] , is_inner_value=_a , added_vocab=_a )
if len(_a ):
return [output] if is_inner_value else output
else:
return [] if is_inner_value else {"text_sequence": tokens}
@property
def SCREAMING_SNAKE_CASE ( self ):
warnings.warn(
"`feature_extractor_class` is deprecated and will be removed in v5. Use `image_processor_class` instead." , _a , )
return self.image_processor_class
@property
def SCREAMING_SNAKE_CASE ( self ):
warnings.warn(
"`feature_extractor` is deprecated and will be removed in v5. Use `image_processor` instead." , _a , )
return self.image_processor
| 281 |
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
snake_case : int = logging.get_logger(__name__)
snake_case : List[str] = {"vocab_file": "spiece.model"}
snake_case : List[str] = {
"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",
}
}
snake_case : Tuple = {
"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,
}
snake_case : List[str] = "▁"
class _snake_case ( snake_case ):
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self , _a , _a=True , _a=True , _a=False , _a="[CLS]" , _a="[SEP]" , _a="<unk>" , _a="[SEP]" , _a="<pad>" , _a="[CLS]" , _a="[MASK]" , _a = None , **_a , ):
# 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.
__magic_name__ : str = (
AddedToken(_a , lstrip=_a , rstrip=_a , normalized=_a )
if isinstance(_a , _a )
else mask_token
)
__magic_name__ : Union[str, Any] = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_a , remove_space=_a , keep_accents=_a , bos_token=_a , eos_token=_a , unk_token=_a , sep_token=_a , pad_token=_a , cls_token=_a , mask_token=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Dict = do_lower_case
__magic_name__ : Tuple = remove_space
__magic_name__ : Union[str, Any] = keep_accents
__magic_name__ : Tuple = vocab_file
__magic_name__ : int = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_a )
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.sp_model )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = {self.convert_ids_to_tokens(_a ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self ):
__magic_name__ : List[str] = self.__dict__.copy()
__magic_name__ : Any = None
return state
def __setstate__( self , _a ):
__magic_name__ : Union[str, Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : str = {}
__magic_name__ : Optional[int] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def SCREAMING_SNAKE_CASE ( self , _a ):
if self.remove_space:
__magic_name__ : List[Any] = " ".join(inputs.strip().split() )
else:
__magic_name__ : str = inputs
__magic_name__ : int = outputs.replace("``" , "\"" ).replace("''" , "\"" )
if not self.keep_accents:
__magic_name__ : str = unicodedata.normalize("NFKD" , _a )
__magic_name__ : Tuple = "".join([c for c in outputs if not unicodedata.combining(_a )] )
if self.do_lower_case:
__magic_name__ : int = outputs.lower()
return outputs
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Optional[Any] = self.preprocess_text(_a )
__magic_name__ : Dict = self.sp_model.encode(_a , out_type=_a )
__magic_name__ : Any = []
for piece in pieces:
if len(_a ) > 1 and piece[-1] == str("," ) and piece[-2].isdigit():
__magic_name__ : Optional[Any] = self.sp_model.EncodeAsPieces(piece[:-1].replace(_a , "" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
__magic_name__ : List[str] = cur_pieces[1:]
else:
__magic_name__ : Optional[int] = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(_a )
else:
new_pieces.append(_a )
return new_pieces
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.PieceToId(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.sp_model.IdToPiece(_a )
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : Any = []
__magic_name__ : Union[str, Any] = ""
__magic_name__ : int = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(_a ) + token
__magic_name__ : List[Any] = True
__magic_name__ : Optional[int] = []
else:
current_sub_tokens.append(_a )
__magic_name__ : Optional[Any] = False
out_string += self.sp_model.decode(_a )
return out_string.strip()
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : List[str] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return cls + token_ids_a + sep
return cls + token_ids_a + sep + token_ids_a + sep
def SCREAMING_SNAKE_CASE ( self , _a , _a = None , _a = False ):
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=_a , token_ids_a=_a , already_has_special_tokens=_a )
if token_ids_a is not None:
return [1] + ([0] * len(_a )) + [1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Optional[int] = [self.sep_token_id]
__magic_name__ : Union[str, Any] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
if not os.path.isdir(_a ):
logger.error(f'''Vocabulary path ({save_directory}) should be a directory''' )
return
__magic_name__ : List[str] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(_a ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , _a )
elif not os.path.isfile(self.vocab_file ):
with open(_a , "wb" ) as fi:
__magic_name__ : Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(_a )
return (out_vocab_file,)
| 281 | 1 |
import io
import os
import unicodedata
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
snake_case : Union[str, Any] = logging.get_logger(__name__)
snake_case : List[Any] = "▁"
snake_case : Union[str, Any] = {"vocab_file": "vocab.txt", "sentencepiece_model_ckpt": "sentencepiece.bpe.model"}
snake_case : Optional[int] = {
"sentencepiece_model_file": "sentencepiece.bpe.model",
"vocab_file": "vocab.txt",
}
snake_case : int = {
"vocab_file": {
"ernie-m-base": "https://huggingface.co/susnato/ernie-m-base_pytorch/blob/main/vocab.txt",
"ernie-m-large": "https://huggingface.co/susnato/ernie-m-base_pytorch/blob/main/vocab.txt",
},
"sentencepiece_model_file": {
"ernie-m-base": "https://huggingface.co/susnato/ernie-m-base_pytorch/blob/main/sentencepiece.bpe.model",
"ernie-m-large": "https://huggingface.co/susnato/ernie-m-base_pytorch/blob/main/sentencepiece.bpe.model",
},
}
snake_case : Optional[int] = {
"ernie-m-base": 514,
"ernie-m-large": 514,
}
snake_case : Any = {
"ernie-m-base": {"do_lower_case": False},
"ernie-m-large": {"do_lower_case": False},
}
class _snake_case ( snake_case ):
UpperCamelCase__ = ["input_ids"]
UpperCamelCase__ = VOCAB_FILES_NAMES
UpperCamelCase__ = PRETRAINED_INIT_CONFIGURATION
UpperCamelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
UpperCamelCase__ = PRETRAINED_VOCAB_FILES_MAP
UpperCamelCase__ = RESOURCE_FILES_NAMES
def __init__( self , _a , _a=None , _a=False , _a="utf8" , _a="[UNK]" , _a="[SEP]" , _a="[PAD]" , _a="[CLS]" , _a="[MASK]" , _a = None , **_a , ):
# 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.
__magic_name__ : Any = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=_a , unk_token=_a , sep_token=_a , pad_token=_a , cls_token=_a , mask_token=_a , vocab_file=_a , encoding=_a , sp_model_kwargs=self.sp_model_kwargs , **_a , )
__magic_name__ : Union[str, Any] = do_lower_case
__magic_name__ : List[Any] = sentencepiece_model_ckpt
__magic_name__ : List[str] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(_a )
# to mimic paddlenlp.transformers.ernie_m.tokenizer.ErnieMTokenizer functioning
if vocab_file is not None:
__magic_name__ : Dict = self.load_vocab(filepath=_a )
else:
__magic_name__ : Optional[int] = {self.sp_model.id_to_piece(_a ): id for id in range(self.sp_model.get_piece_size() )}
__magic_name__ : str = {v: k for k, v in self.vocab.items()}
def SCREAMING_SNAKE_CASE ( self , _a ):
if text is None:
return None
__magic_name__ : Optional[Any] = self.tokenize(_a )
__magic_name__ , __magic_name__ : Optional[Any] = "", []
for i, ch in enumerate(_a ):
if ch in self.SP_CHAR_MAPPING:
__magic_name__ : Any = self.SP_CHAR_MAPPING.get(_a )
else:
__magic_name__ : str = unicodedata.normalize("NFKC" , _a )
if self.is_whitespace(_a ):
continue
normalized_text += ch
char_mapping.extend([i] * len(_a ) )
__magic_name__ , __magic_name__ , __magic_name__ : Optional[Any] = normalized_text, [], 0
if self.do_lower_case:
__magic_name__ : Optional[int] = text.lower()
for token in split_tokens:
if token[:1] == "▁":
__magic_name__ : List[Any] = token[1:]
__magic_name__ : str = text[offset:].index(_a ) + offset
__magic_name__ : List[str] = start + len(_a )
token_mapping.append((char_mapping[start], char_mapping[end - 1] + 1) )
__magic_name__ : Union[str, Any] = end
return token_mapping
@property
def SCREAMING_SNAKE_CASE ( self ):
return len(self.vocab )
def SCREAMING_SNAKE_CASE ( self ):
return dict(self.vocab , **self.added_tokens_encoder )
def __getstate__( self ):
__magic_name__ : int = self.__dict__.copy()
__magic_name__ : int = None
return state
def __setstate__( self , _a ):
__magic_name__ : str = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
__magic_name__ : int = {}
__magic_name__ : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.sentencepiece_model_ckpt )
def SCREAMING_SNAKE_CASE ( self , _a ):
return "".join((self.SP_CHAR_MAPPING.get(_a , _a ) for c in text) )
def SCREAMING_SNAKE_CASE ( self , _a , _a=False , _a=64 , _a=0.1 ):
if self.sp_model_kwargs.get("enable_sampling" ) is True:
__magic_name__ : Union[str, Any] = True
if self.sp_model_kwargs.get("alpha" ) is not None:
__magic_name__ : Union[str, Any] = self.sp_model_kwargs.get("alpha" )
if self.sp_model_kwargs.get("nbest_size" ) is not None:
__magic_name__ : List[Any] = self.sp_model_kwargs.get("nbest_size" )
if not enable_sampling:
__magic_name__ : Tuple = self.sp_model.EncodeAsPieces(_a )
else:
__magic_name__ : Union[str, Any] = self.sp_model.SampleEncodeAsPieces(_a , _a , _a )
__magic_name__ : Optional[int] = []
for pi, piece in enumerate(_a ):
if piece == SPIECE_UNDERLINE:
if not pieces[pi + 1].startswith(_a ) and pi != 0:
new_pieces.append(_a )
continue
else:
continue
__magic_name__ : Dict = 0
for i, chunk in enumerate(_a ):
if chunk == SPIECE_UNDERLINE:
continue
if self.is_ch_char(_a ) or self.is_punct(_a ):
if i > lst_i and piece[lst_i:i] != SPIECE_UNDERLINE:
new_pieces.append(piece[lst_i:i] )
new_pieces.append(_a )
__magic_name__ : Any = i + 1
elif chunk.isdigit() and i > 0 and not piece[i - 1].isdigit():
if i > lst_i and piece[lst_i:i] != SPIECE_UNDERLINE:
new_pieces.append(piece[lst_i:i] )
__magic_name__ : str = i
elif not chunk.isdigit() and i > 0 and piece[i - 1].isdigit():
if i > lst_i and piece[lst_i:i] != SPIECE_UNDERLINE:
new_pieces.append(piece[lst_i:i] )
__magic_name__ : List[str] = i
if len(_a ) > lst_i:
new_pieces.append(piece[lst_i:] )
return new_pieces
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[Any] = "".join(_a ).replace(_a , " " ).strip()
return out_string
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[Any] = self.convert_ids_to_tokens(_a )
__magic_name__ : Optional[Any] = "".join(_a ).replace(_a , " " ).strip()
return out_string
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.vocab.get(_a , self.vocab.get(self.unk_token ) )
def SCREAMING_SNAKE_CASE ( self , _a ):
return self.reverse_vocab.get(_a , self.unk_token )
def SCREAMING_SNAKE_CASE ( self , _a , _a=None ):
if token_ids_a is None:
return [self.cls_token_id] + token_ids_a + [self.sep_token_id]
__magic_name__ : int = [self.cls_token_id]
__magic_name__ : Union[str, Any] = [self.sep_token_id]
return _cls + token_ids_a + _sep + _sep + token_ids_a + _sep
def SCREAMING_SNAKE_CASE ( self , _a , _a=None ):
if offset_mapping_a is None:
return [(0, 0)] + offset_mapping_a + [(0, 0)]
return [(0, 0)] + offset_mapping_a + [(0, 0), (0, 0)] + offset_mapping_a + [(0, 0)]
def SCREAMING_SNAKE_CASE ( self , _a , _a=None , _a=False ):
if already_has_special_tokens:
if token_ids_a is not None:
raise ValueError(
"You should not supply a second sequence if the provided sequence of "
"ids is already formatted with special tokens for the model." )
return [1 if x in [self.sep_token_id, self.cls_token_id] else 0 for x in token_ids_a]
if token_ids_a is not None:
return [1] + ([0] * len(_a )) + [1, 1] + ([0] * len(_a )) + [1]
return [1] + ([0] * len(_a )) + [1]
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
# called when `add_special_tokens` is True, so align with `build_inputs_with_special_tokens` method
if token_ids_a is None:
# [CLS] X [SEP]
return (len(_a ) + 2) * [0]
# [CLS] A [SEP] [SEP] B [SEP]
return [0] * (len(_a ) + 1) + [1] * (len(_a ) + 3)
def SCREAMING_SNAKE_CASE ( self , _a ):
if "\u4e00" <= char <= "\u9fff":
return True
return False
def SCREAMING_SNAKE_CASE ( self , _a ):
if ("a" <= char <= "z") or ("A" <= char <= "Z"):
return True
return False
def SCREAMING_SNAKE_CASE ( self , _a ):
if char in ",;:.?!~，；：。？！《》【】":
return True
return False
def SCREAMING_SNAKE_CASE ( self , _a ):
if char == " " or char == "\t" or char == "\n" or char == "\r":
return True
if len(_a ) == 1:
__magic_name__ : List[Any] = unicodedata.category(_a )
if cat == "Zs":
return True
return False
def SCREAMING_SNAKE_CASE ( self , _a ):
__magic_name__ : List[str] = {}
with io.open(_a , "r" , encoding="utf-8" ) as f:
for index, line in enumerate(_a ):
__magic_name__ : str = line.rstrip("\n" )
__magic_name__ : Optional[Any] = int(_a )
return token_to_idx
def SCREAMING_SNAKE_CASE ( self , _a , _a = None ):
__magic_name__ : Tuple = 0
if os.path.isdir(_a ):
__magic_name__ : List[Any] = os.path.join(
_a , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
else:
__magic_name__ : int = (filename_prefix + "-" if filename_prefix else "") + save_directory
with open(_a , "w" , encoding="utf-8" ) as writer:
for token, token_index in sorted(self.vocab.items() , key=lambda _a : kv[1] ):
if index != token_index:
logger.warning(
f'''Saving vocabulary to {vocab_file}: vocabulary indices are not consecutive.'''
" Please check that the vocabulary is not corrupted!" )
__magic_name__ : List[str] = token_index
writer.write(token + "\n" )
index += 1
__magic_name__ : List[Any] = os.path.join(_a , "sentencepiece.bpe.model" )
with open(_a , "wb" ) as fi:
__magic_name__ : int = self.sp_model.serialized_model_proto()
fi.write(_a )
return (vocab_file,)
| 281 |
import unicodedata
from dataclasses import dataclass
from typing import Optional, Union
import numpy as np
from transformers.data.data_collator import DataCollatorMixin
from transformers.file_utils import PaddingStrategy
from transformers.tokenization_utils_base import PreTrainedTokenizerBase
def lowerCAmelCase_ ( _snake_case : List[str] , _snake_case : Tuple , _snake_case : List[Any] , _snake_case : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Union[str, Any] = np.full((len(_snake_case ), sequence_length, 2) , _snake_case )
else:
__magic_name__ : List[Any] = np.full((len(_snake_case ), sequence_length) , _snake_case )
for i, tensor in enumerate(_snake_case ):
if padding_side == "right":
if isinstance(_snake_case , _snake_case ):
__magic_name__ : Optional[Any] = tensor[:sequence_length]
else:
__magic_name__ : Union[str, Any] = tensor[:sequence_length]
else:
if isinstance(_snake_case , _snake_case ):
__magic_name__ : List[Any] = tensor[:sequence_length]
else:
__magic_name__ : Optional[Any] = tensor[:sequence_length]
return out_tensor.tolist()
def lowerCAmelCase_ ( _snake_case : Optional[int] ) -> Tuple:
'''simple docstring'''
__magic_name__ : Union[str, Any] = ord(_snake_case )
if (cp >= 33 and cp <= 47) or (cp >= 58 and cp <= 64) or (cp >= 91 and cp <= 96) or (cp >= 123 and cp <= 126):
return True
__magic_name__ : Any = unicodedata.category(_snake_case )
if cat.startswith("P" ):
return True
return False
@dataclass
class _snake_case ( snake_case ):
UpperCamelCase__ = 42
UpperCamelCase__ = True
UpperCamelCase__ = None
UpperCamelCase__ = None
UpperCamelCase__ = -100
UpperCamelCase__ = "pt"
def SCREAMING_SNAKE_CASE ( self , _a ):
import torch
__magic_name__ : List[str] = "label" if "label" in features[0].keys() else "labels"
__magic_name__ : Union[str, Any] = [feature[label_name] for feature in features] if label_name in features[0].keys() else None
__magic_name__ : Optional[int] = self.tokenizer.pad(
_a , padding=self.padding , max_length=self.max_length , pad_to_multiple_of=self.pad_to_multiple_of , return_tensors="pt" if labels is None else None , )
if labels is None:
return batch
__magic_name__ : Dict = torch.tensor(batch["entity_ids"] ).shape[1]
__magic_name__ : List[Any] = self.tokenizer.padding_side
if padding_side == "right":
__magic_name__ : str = [
list(_a ) + [self.label_pad_token_id] * (sequence_length - len(_a )) for label in labels
]
else:
__magic_name__ : int = [
[self.label_pad_token_id] * (sequence_length - len(_a )) + list(_a ) for label in labels
]
__magic_name__ : Dict = [feature["ner_tags"] for feature in features]
__magic_name__ : List[Any] = padding_tensor(_a , -1 , _a , _a )
__magic_name__ : Any = [feature["original_entity_spans"] for feature in features]
__magic_name__ : Any = padding_tensor(_a , (-1, -1) , _a , _a )
__magic_name__ : List[Any] = {k: torch.tensor(_a , dtype=torch.intaa ) for k, v in batch.items()}
return batch
| 281 | 1 |
def lowerCAmelCase_ ( _snake_case : int ) -> int:
'''simple docstring'''
if not isinstance(_snake_case , _snake_case ):
raise TypeError("Input value must be an 'int' type" )
__magic_name__ : List[Any] = 0
while number:
position += 1
number >>= 1
return position
if __name__ == "__main__":
import doctest
doctest.testmod()
| 281 |
import math
def lowerCAmelCase_ ( _snake_case : float , _snake_case : float ) -> float:
'''simple docstring'''
return math.pow(_snake_case , 2 ) - a
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
return 2 * x
def lowerCAmelCase_ ( _snake_case : float ) -> float:
'''simple docstring'''
__magic_name__ : Optional[int] = 2.0
while start <= a:
__magic_name__ : str = math.pow(_snake_case , 2 )
return start
def lowerCAmelCase_ ( _snake_case : float , _snake_case : int = 9999 , _snake_case : float = 0.00_000_000_000_001 ) -> float:
'''simple docstring'''
if a < 0:
raise ValueError("math domain error" )
__magic_name__ : Optional[int] = get_initial_point(_snake_case )
for _ in range(_snake_case ):
__magic_name__ : int = value
__magic_name__ : str = value - fx(_snake_case , _snake_case ) / fx_derivative(_snake_case )
if abs(prev_value - value ) < tolerance:
return value
return value
if __name__ == "__main__":
from doctest import testmod
testmod()
| 281 | 1 |
from __future__ import annotations
def lowerCAmelCase_ ( _snake_case : str , _snake_case : list[str] | None = None , _snake_case : dict[str, float] | None = None , _snake_case : bool = False , ) -> tuple[int, float, str]:
'''simple docstring'''
__magic_name__ : Any = cipher_alphabet or [chr(_snake_case ) for i in range(97 , 123 )]
# If the argument is None or the user provided an empty dictionary
if not frequencies_dict:
# Frequencies of letters in the english language (how much they show up)
__magic_name__ : Optional[int] = {
"a": 0.08_497,
"b": 0.01_492,
"c": 0.02_202,
"d": 0.04_253,
"e": 0.11_162,
"f": 0.02_228,
"g": 0.02_015,
"h": 0.06_094,
"i": 0.07_546,
"j": 0.00_153,
"k": 0.01_292,
"l": 0.04_025,
"m": 0.02_406,
"n": 0.06_749,
"o": 0.07_507,
"p": 0.01_929,
"q": 0.00_095,
"r": 0.07_587,
"s": 0.06_327,
"t": 0.09_356,
"u": 0.02_758,
"v": 0.00_978,
"w": 0.02_560,
"x": 0.00_150,
"y": 0.01_994,
"z": 0.00_077,
}
else:
# Custom frequencies dictionary
__magic_name__ : Dict = frequencies_dict
if not case_sensitive:
__magic_name__ : List[str] = ciphertext.lower()
# Chi squared statistic values
__magic_name__ : dict[int, tuple[float, str]] = {}
# cycle through all of the shifts
for shift in range(len(_snake_case ) ):
__magic_name__ : int = ""
# decrypt the message with the shift
for letter in ciphertext:
try:
# Try to index the letter in the alphabet
__magic_name__ : Optional[Any] = (alphabet_letters.index(letter.lower() ) - shift) % len(
_snake_case )
decrypted_with_shift += (
alphabet_letters[new_key].upper()
if case_sensitive and letter.isupper()
else alphabet_letters[new_key]
)
except ValueError:
# Append the character if it isn't in the alphabet
decrypted_with_shift += letter
__magic_name__ : List[Any] = 0.0
# Loop through each letter in the decoded message with the shift
for letter in decrypted_with_shift:
if case_sensitive:
__magic_name__ : Optional[Any] = letter.lower()
if letter in frequencies:
# Get the amount of times the letter occurs in the message
__magic_name__ : List[str] = decrypted_with_shift.lower().count(_snake_case )
# Get the excepcted amount of times the letter should appear based
# on letter frequencies
__magic_name__ : Any = frequencies[letter] * occurrences
# Complete the chi squared statistic formula
__magic_name__ : List[Any] = ((occurrences - expected) ** 2) / expected
# Add the margin of error to the total chi squared statistic
chi_squared_statistic += chi_letter_value
else:
if letter.lower() in frequencies:
# Get the amount of times the letter occurs in the message
__magic_name__ : str = decrypted_with_shift.count(_snake_case )
# Get the excepcted amount of times the letter should appear based
# on letter frequencies
__magic_name__ : Optional[Any] = frequencies[letter] * occurrences
# Complete the chi squared statistic formula
__magic_name__ : List[Any] = ((occurrences - expected) ** 2) / expected
# Add the margin of error to the total chi squared statistic
chi_squared_statistic += chi_letter_value
# Add the data to the chi_squared_statistic_values dictionary
__magic_name__ : Optional[int] = (
chi_squared_statistic,
decrypted_with_shift,
)
# Get the most likely cipher by finding the cipher with the smallest chi squared
# statistic
def chi_squared_statistic_values_sorting_key(_snake_case : int ) -> tuple[float, str]:
return chi_squared_statistic_values[key]
__magic_name__ : int = min(
_snake_case , key=_snake_case , )
# Get all the data from the most likely cipher (key, decoded message)
(
(
__magic_name__
) , (
__magic_name__
) ,
) : List[str] = chi_squared_statistic_values[most_likely_cipher]
# Return the data on the most likely shift
return (
most_likely_cipher,
most_likely_cipher_chi_squared_value,
decoded_most_likely_cipher,
)
| 281 |
from __future__ import annotations
import unittest
from transformers import LEDConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import TFLEDForConditionalGeneration, TFLEDModel
@require_tf
class _snake_case :
UpperCamelCase__ = LEDConfig
UpperCamelCase__ = {}
UpperCamelCase__ = 'gelu'
def __init__( self , _a , _a=13 , _a=7 , _a=True , _a=False , _a=99 , _a=32 , _a=2 , _a=4 , _a=37 , _a=0.1 , _a=0.1 , _a=20 , _a=2 , _a=1 , _a=0 , _a=4 , ):
__magic_name__ : int = parent
__magic_name__ : Optional[int] = batch_size
__magic_name__ : Tuple = seq_length
__magic_name__ : List[Any] = is_training
__magic_name__ : Dict = use_labels
__magic_name__ : Optional[Any] = vocab_size
__magic_name__ : int = hidden_size
__magic_name__ : Optional[int] = num_hidden_layers
__magic_name__ : Optional[int] = num_attention_heads
__magic_name__ : Tuple = intermediate_size
__magic_name__ : Any = hidden_dropout_prob
__magic_name__ : Optional[int] = attention_probs_dropout_prob
__magic_name__ : List[str] = max_position_embeddings
__magic_name__ : Any = eos_token_id
__magic_name__ : str = pad_token_id
__magic_name__ : int = bos_token_id
__magic_name__ : Optional[int] = attention_window
# `ModelTesterMixin.test_attention_outputs` is expecting attention tensors to be of size
# [num_attention_heads, encoder_seq_length, encoder_key_length], but TFLongformerSelfAttention
# returns attention of shape [num_attention_heads, encoder_seq_length, self.attention_window + 1]
# because its local attention only attends to `self.attention_window` and one before and one after
__magic_name__ : Tuple = self.attention_window + 2
# because of padding `encoder_seq_length`, is different from `seq_length`. Relevant for
# the `test_attention_outputs` and `test_hidden_states_output` tests
__magic_name__ : Tuple = (
self.seq_length + (self.attention_window - self.seq_length % self.attention_window) % self.attention_window
)
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Union[str, Any] = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
__magic_name__ : int = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
__magic_name__ : int = tf.concat([input_ids, eos_tensor] , axis=1 )
__magic_name__ : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : Dict = self.config_cls(
vocab_size=self.vocab_size , d_model=self.hidden_size , encoder_layers=self.num_hidden_layers , decoder_layers=self.num_hidden_layers , encoder_attention_heads=self.num_attention_heads , decoder_attention_heads=self.num_attention_heads , encoder_ffn_dim=self.intermediate_size , decoder_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_ids=[2] , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , decoder_start_token_id=self.pad_token_id , attention_window=self.attention_window , **self.config_updates , )
__magic_name__ : List[str] = prepare_led_inputs_dict(_a , _a , _a )
__magic_name__ : Union[str, Any] = tf.concat(
[tf.zeros_like(_a )[:, :-1], tf.ones_like(_a )[:, -1:]] , axis=-1 , )
__magic_name__ : List[Any] = global_attention_mask
return config, inputs_dict
def SCREAMING_SNAKE_CASE ( self , _a , _a ):
__magic_name__ : Dict = TFLEDModel(config=_a ).get_decoder()
__magic_name__ : Optional[int] = inputs_dict["input_ids"]
__magic_name__ : Union[str, Any] = input_ids[:1, :]
__magic_name__ : str = inputs_dict["attention_mask"][:1, :]
__magic_name__ : int = 1
# first forward pass
__magic_name__ : Tuple = model(_a , attention_mask=_a , use_cache=_a )
__magic_name__ , __magic_name__ : str = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
__magic_name__ : Optional[int] = ids_tensor((self.batch_size, 3) , config.vocab_size )
__magic_name__ : Any = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
__magic_name__ : Optional[Any] = tf.concat([input_ids, next_tokens] , axis=-1 )
__magic_name__ : List[Any] = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
__magic_name__ : List[str] = model(_a , attention_mask=_a )[0]
__magic_name__ : Dict = model(_a , attention_mask=_a , past_key_values=_a )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
__magic_name__ : List[Any] = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
__magic_name__ : Union[str, Any] = output_from_no_past[:, -3:, random_slice_idx]
__magic_name__ : List[str] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(_a , _a , rtol=1e-3 )
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : List[Any] , _snake_case : Any , _snake_case : str=None , _snake_case : List[str]=None , _snake_case : int=None , _snake_case : Any=None , ) -> int:
'''simple docstring'''
if attention_mask is None:
__magic_name__ : str = tf.cast(tf.math.not_equal(_snake_case , config.pad_token_id ) , tf.inta )
if decoder_attention_mask is None:
__magic_name__ : List[Any] = tf.concat(
[
tf.ones(decoder_input_ids[:, :1].shape , dtype=tf.inta ),
tf.cast(tf.math.not_equal(decoder_input_ids[:, 1:] , config.pad_token_id ) , tf.inta ),
] , axis=-1 , )
if head_mask is None:
__magic_name__ : Dict = tf.ones((config.encoder_layers, config.encoder_attention_heads) )
if decoder_head_mask is None:
__magic_name__ : Any = tf.ones((config.decoder_layers, config.decoder_attention_heads) )
return {
"input_ids": input_ids,
"attention_mask": attention_mask,
"decoder_input_ids": decoder_input_ids,
"decoder_attention_mask": decoder_attention_mask,
"head_mask": head_mask,
"decoder_head_mask": decoder_head_mask,
}
@require_tf
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (TFLEDForConditionalGeneration, TFLEDModel) if is_tf_available() else ()
UpperCamelCase__ = (TFLEDForConditionalGeneration,) if is_tf_available() else ()
UpperCamelCase__ = (
{
'conversational': TFLEDForConditionalGeneration,
'feature-extraction': TFLEDModel,
'summarization': TFLEDForConditionalGeneration,
'text2text-generation': TFLEDForConditionalGeneration,
'translation': TFLEDForConditionalGeneration,
}
if is_tf_available()
else {}
)
UpperCamelCase__ = True
UpperCamelCase__ = False
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = TFLEDModelTester(self )
__magic_name__ : List[Any] = ConfigTester(self , config_class=_a )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ , __magic_name__ : int = self.model_tester.prepare_config_and_inputs_for_common()
__magic_name__ : List[str] = tf.zeros_like(inputs_dict["attention_mask"] )
__magic_name__ : Optional[Any] = 2
__magic_name__ : Tuple = tf.where(
tf.range(self.model_tester.seq_length )[None, :] < num_global_attn_indices , 1 , inputs_dict["global_attention_mask"] , )
__magic_name__ : Any = True
__magic_name__ : str = self.model_tester.seq_length
__magic_name__ : Dict = self.model_tester.encoder_seq_length
def check_decoder_attentions_output(_a ):
__magic_name__ : str = outputs.decoder_attentions
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(decoder_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
def check_encoder_attentions_output(_a ):
__magic_name__ : Any = [t.numpy() for t in outputs.encoder_attentions]
__magic_name__ : Tuple = [t.numpy() for t in outputs.encoder_global_attentions]
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertEqual(len(_a ) , self.model_tester.num_hidden_layers )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, seq_length, seq_length] , )
self.assertListEqual(
list(global_attentions[0].shape[-3:] ) , [self.model_tester.num_attention_heads, encoder_seq_length, num_global_attn_indices] , )
for model_class in self.all_model_classes:
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = False
__magic_name__ : Tuple = False
__magic_name__ : Optional[int] = model_class(_a )
__magic_name__ : str = model(self._prepare_for_class(_a , _a ) )
__magic_name__ : Any = len(_a )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
if self.is_encoder_decoder:
__magic_name__ : Tuple = model_class(_a )
__magic_name__ : Optional[Any] = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_decoder_attentions_output(_a )
# Check that output attentions can also be changed via the config
del inputs_dict["output_attentions"]
__magic_name__ : Dict = True
__magic_name__ : str = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
# Check attention is always last and order is fine
__magic_name__ : Union[str, Any] = True
__magic_name__ : Union[str, Any] = True
__magic_name__ : List[str] = model_class(_a )
__magic_name__ : Any = model(self._prepare_for_class(_a , _a ) )
self.assertEqual(out_len + (2 if self.is_encoder_decoder else 1) , len(_a ) )
self.assertEqual(model.config.output_hidden_states , _a )
check_encoder_attentions_output(_a )
@unittest.skip("LED keeps using potentially symbolic tensors in conditionals and breaks tracing." )
def SCREAMING_SNAKE_CASE ( self ):
pass
def SCREAMING_SNAKE_CASE ( self ):
# TODO: Head-masking not yet implement
pass
def lowerCAmelCase_ ( _snake_case : int ) -> Optional[int]:
'''simple docstring'''
return tf.constant(_snake_case , dtype=tf.intaa )
snake_case : Optional[int] = 1E-4
@slow
@require_tf
class _snake_case ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[Any] = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" ).led
# change to intended input here
__magic_name__ : Optional[int] = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : str = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Any = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : List[Any] = model(**_a )[0]
__magic_name__ : List[str] = (1, 1_024, 768)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : int = tf.convert_to_tensor(
[[2.30_50, 2.82_79, 0.65_31], [-1.84_57, -0.14_55, -3.56_61], [-1.01_86, 0.45_86, -2.20_43]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Tuple = TFLEDForConditionalGeneration.from_pretrained("allenai/led-base-16384" )
# change to intended input here
__magic_name__ : int = _long_tensor([512 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Tuple = _long_tensor([128 * [0, 31_414, 232, 328, 740, 1_140, 12_695, 69]] )
__magic_name__ : Optional[Any] = prepare_led_inputs_dict(model.config , _a , _a )
__magic_name__ : Union[str, Any] = model(**_a )[0]
__magic_name__ : Optional[int] = (1, 1_024, model.config.vocab_size)
self.assertEqual(output.shape , _a )
# change to expected output here
__magic_name__ : str = tf.convert_to_tensor(
[[33.65_07, 6.45_72, 16.80_89], [5.87_39, -2.42_38, 11.29_02], [-3.21_39, -4.31_49, 4.27_83]] , )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-3 , rtol=1e-3 )
| 281 | 1 |
import warnings
from ...utils import logging
from .image_processing_chinese_clip import ChineseCLIPImageProcessor
snake_case : Union[str, Any] = logging.get_logger(__name__)
class _snake_case ( snake_case ):
def __init__( self , *_a , **_a ):
warnings.warn(
"The class ChineseCLIPFeatureExtractor is deprecated and will be removed in version 5 of Transformers."
" Please use ChineseCLIPImageProcessor instead." , _a , )
super().__init__(*_a , **_a )
| 281 |
import argparse
import json
from pathlib import Path
import requests
import timm
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform
from transformers import (
BitConfig,
ViTHybridConfig,
ViTHybridForImageClassification,
ViTHybridImageProcessor,
ViTHybridModel,
)
from transformers.image_utils import PILImageResampling
from transformers.utils import logging
logging.set_verbosity_info()
snake_case : Optional[Any] = logging.get_logger(__name__)
def lowerCAmelCase_ ( _snake_case : Union[str, Any] , _snake_case : Union[str, Any]=False ) -> List[str]:
'''simple docstring'''
__magic_name__ : Union[str, Any] = []
# fmt: off
# stem:
rename_keys.append(("cls_token", "vit.embeddings.cls_token") )
rename_keys.append(("pos_embed", "vit.embeddings.position_embeddings") )
rename_keys.append(("patch_embed.proj.weight", "vit.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("patch_embed.proj.bias", "vit.embeddings.patch_embeddings.projection.bias") )
# backbone
rename_keys.append(("patch_embed.backbone.stem.conv.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.convolution.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.weight", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.weight") )
rename_keys.append(("patch_embed.backbone.stem.norm.bias", "vit.embeddings.patch_embeddings.backbone.bit.embedder.norm.bias") )
for stage_idx in range(len(config.backbone_config.depths ) ):
for layer_idx in range(config.backbone_config.depths[stage_idx] ):
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm1.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm1.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm2.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm2.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.conv3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.conv3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.{layer_idx}.norm3.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.{layer_idx}.norm3.bias''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.conv.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.conv.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.weight''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.weight''') )
rename_keys.append((F'''patch_embed.backbone.stages.{stage_idx}.blocks.0.downsample.norm.bias''', F'''vit.embeddings.patch_embeddings.backbone.bit.encoder.stages.{stage_idx}.layers.0.downsample.norm.bias''') )
# transformer encoder
for i in range(config.num_hidden_layers ):
# encoder layers: output projection, 2 feedforward neural networks and 2 layernorms
rename_keys.append((F'''blocks.{i}.norm1.weight''', F'''vit.encoder.layer.{i}.layernorm_before.weight''') )
rename_keys.append((F'''blocks.{i}.norm1.bias''', F'''vit.encoder.layer.{i}.layernorm_before.bias''') )
rename_keys.append((F'''blocks.{i}.attn.proj.weight''', F'''vit.encoder.layer.{i}.attention.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.attn.proj.bias''', F'''vit.encoder.layer.{i}.attention.output.dense.bias''') )
rename_keys.append((F'''blocks.{i}.norm2.weight''', F'''vit.encoder.layer.{i}.layernorm_after.weight''') )
rename_keys.append((F'''blocks.{i}.norm2.bias''', F'''vit.encoder.layer.{i}.layernorm_after.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.weight''', F'''vit.encoder.layer.{i}.intermediate.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc1.bias''', F'''vit.encoder.layer.{i}.intermediate.dense.bias''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.weight''', F'''vit.encoder.layer.{i}.output.dense.weight''') )
rename_keys.append((F'''blocks.{i}.mlp.fc2.bias''', F'''vit.encoder.layer.{i}.output.dense.bias''') )
if base_model:
# layernorm + pooler
rename_keys.extend(
[
("norm.weight", "layernorm.weight"),
("norm.bias", "layernorm.bias"),
("pre_logits.fc.weight", "pooler.dense.weight"),
("pre_logits.fc.bias", "pooler.dense.bias"),
] )
# if just the base model, we should remove "vit" from all keys that start with "vit"
__magic_name__ : int = [(pair[0], pair[1][4:]) if pair[1].startswith("vit" ) else pair for pair in rename_keys]
else:
# layernorm + classification head
rename_keys.extend(
[
("norm.weight", "vit.layernorm.weight"),
("norm.bias", "vit.layernorm.bias"),
("head.weight", "classifier.weight"),
("head.bias", "classifier.bias"),
] )
# fmt: on
return rename_keys
def lowerCAmelCase_ ( _snake_case : Any , _snake_case : Any , _snake_case : Dict=False ) -> int:
'''simple docstring'''
for i in range(config.num_hidden_layers ):
if base_model:
__magic_name__ : int = ""
else:
__magic_name__ : Union[str, Any] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
__magic_name__ : Optional[Any] = state_dict.pop(F'''blocks.{i}.attn.qkv.weight''' )
__magic_name__ : int = state_dict.pop(F'''blocks.{i}.attn.qkv.bias''' )
# next, add query, keys and values (in that order) to the state dict
__magic_name__ : Dict = in_proj_weight[
: config.hidden_size, :
]
__magic_name__ : List[str] = in_proj_bias[: config.hidden_size]
__magic_name__ : List[str] = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
__magic_name__ : Optional[Any] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
__magic_name__ : Optional[Any] = in_proj_weight[
-config.hidden_size :, :
]
__magic_name__ : int = in_proj_bias[-config.hidden_size :]
def lowerCAmelCase_ ( _snake_case : List[str] ) -> List[str]:
'''simple docstring'''
__magic_name__ : List[str] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(_snake_case , _snake_case )
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : int , _snake_case : Union[str, Any] ) -> Optional[int]:
'''simple docstring'''
__magic_name__ : int = dct.pop(_snake_case )
__magic_name__ : List[Any] = val
def lowerCAmelCase_ ( ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
__magic_name__ : List[str] = Image.open(requests.get(_snake_case , stream=_snake_case ).raw )
return im
@torch.no_grad()
def lowerCAmelCase_ ( _snake_case : Optional[int] , _snake_case : Any , _snake_case : int=False ) -> Dict:
'''simple docstring'''
__magic_name__ : List[str] = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=_snake_case , )
__magic_name__ : List[str] = ViTHybridConfig(backbone_config=_snake_case , image_size=384 , num_labels=1000 )
__magic_name__ : str = False
# load original model from timm
__magic_name__ : Union[str, Any] = timm.create_model(_snake_case , pretrained=_snake_case )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
__magic_name__ : List[Any] = timm_model.state_dict()
if base_model:
remove_classification_head_(_snake_case )
__magic_name__ : Tuple = create_rename_keys(_snake_case , _snake_case )
for src, dest in rename_keys:
rename_key(_snake_case , _snake_case , _snake_case )
read_in_q_k_v(_snake_case , _snake_case , _snake_case )
__magic_name__ : List[str] = "huggingface/label-files"
__magic_name__ : int = "imagenet-1k-id2label.json"
__magic_name__ : Optional[int] = json.load(open(hf_hub_download(_snake_case , _snake_case , repo_type="dataset" ) , "r" ) )
__magic_name__ : int = {int(_snake_case ): v for k, v in idalabel.items()}
__magic_name__ : List[str] = idalabel
__magic_name__ : List[str] = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
__magic_name__ : List[str] = ViTHybridModel(_snake_case ).eval()
else:
__magic_name__ : str = ViTHybridForImageClassification(_snake_case ).eval()
model.load_state_dict(_snake_case )
# create image processor
__magic_name__ : List[Any] = create_transform(**resolve_data_config({} , model=_snake_case ) )
__magic_name__ : int = transform.transforms
__magic_name__ : List[str] = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
__magic_name__ : int = ViTHybridImageProcessor(
do_resize=_snake_case , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=_snake_case , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=_snake_case , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
__magic_name__ : List[Any] = prepare_img()
__magic_name__ : Any = transform(_snake_case ).unsqueeze(0 )
__magic_name__ : Tuple = processor(_snake_case , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(_snake_case , _snake_case )
# verify logits
with torch.no_grad():
__magic_name__ : Optional[int] = model(_snake_case )
__magic_name__ : List[str] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
__magic_name__ : List[str] = timm_model.forward_features(_snake_case )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(_snake_case , outputs.pooler_output , atol=1E-3 )
else:
__magic_name__ : Any = timm_model(_snake_case )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(_snake_case , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(_snake_case ).mkdir(exist_ok=_snake_case )
print(F'''Saving model {vit_name} to {pytorch_dump_folder_path}''' )
model.save_pretrained(_snake_case )
print(F'''Saving processor to {pytorch_dump_folder_path}''' )
processor.save_pretrained(_snake_case )
if push_to_hub:
print(F'''Pushing model and processor to the hub {vit_name}''' )
model.push_to_hub(F'''ybelkada/{vit_name}''' )
processor.push_to_hub(F'''ybelkada/{vit_name}''' )
if __name__ == "__main__":
snake_case : Any = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--vit_name",
default="vit_base_r50_s16_384",
type=str,
help="Name of the hybrid ViT timm model you'd like to convert.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether to upload the model to the HuggingFace hub."
)
snake_case : List[Any] = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 281 | 1 |
import os
from huggingface_hub.constants import HUGGINGFACE_HUB_CACHE, hf_cache_home
snake_case : str = HUGGINGFACE_HUB_CACHE
snake_case : Tuple = "config.json"
snake_case : int = "diffusion_pytorch_model.bin"
snake_case : Dict = "diffusion_flax_model.msgpack"
snake_case : List[str] = "model.onnx"
snake_case : str = "diffusion_pytorch_model.safetensors"
snake_case : Tuple = "weights.pb"
snake_case : Optional[Any] = "https://huggingface.co"
snake_case : Dict = default_cache_path
snake_case : Any = "diffusers_modules"
snake_case : Optional[Any] = os.getenv("HF_MODULES_CACHE", os.path.join(hf_cache_home, "modules"))
snake_case : Any = ["fp16", "non-ema"]
snake_case : Any = ".self_attn"
| 281 |
# This script creates a super tiny model that is useful inside tests, when we just want to test that
# the machinery works, without needing to the check the quality of the outcomes.
#
# This version creates a tiny model through reduction of a normal pre-trained model, but keeping the
# full vocab, merges file, and thus also resulting in a larger model due to a large vocab size.
# This gives ~3MB in total for all files.
#
# If you want a 50 times smaller than this see `fsmt-make-super-tiny-model.py`, which is slightly more complicated
#
#
# It will be used then as "stas/tiny-wmt19-en-de"
# Build
from transformers import FSMTTokenizer, FSMTConfig, FSMTForConditionalGeneration
snake_case : List[str] = "facebook/wmt19-en-de"
snake_case : Dict = FSMTTokenizer.from_pretrained(mname)
# get the correct vocab sizes, etc. from the master model
snake_case : List[str] = FSMTConfig.from_pretrained(mname)
config.update(
dict(
d_model=4,
encoder_layers=1,
decoder_layers=1,
encoder_ffn_dim=4,
decoder_ffn_dim=4,
encoder_attention_heads=1,
decoder_attention_heads=1,
)
)
snake_case : int = FSMTForConditionalGeneration(config)
print(F"num of params {tiny_model.num_parameters()}")
# Test
snake_case : Optional[Any] = tokenizer(["Making tiny model"], return_tensors="pt")
snake_case : List[str] = tiny_model(**batch)
print("test output:", len(outputs.logits[0]))
# Save
snake_case : Dict = "tiny-wmt19-en-de"
tiny_model.half() # makes it smaller
tiny_model.save_pretrained(mname_tiny)
tokenizer.save_pretrained(mname_tiny)
print(F"Generated {mname_tiny}")
# Upload
# transformers-cli upload tiny-wmt19-en-de
| 281 | 1 |
import numpy as np
from matplotlib import pyplot as plt
from sklearn.datasets import load_iris
from sklearn.metrics import ConfusionMatrixDisplay
from sklearn.model_selection import train_test_split
from xgboost import XGBClassifier
def lowerCAmelCase_ ( _snake_case : dict ) -> tuple:
'''simple docstring'''
return (data["data"], data["target"])
def lowerCAmelCase_ ( _snake_case : np.ndarray , _snake_case : np.ndarray ) -> XGBClassifier:
'''simple docstring'''
__magic_name__ : str = XGBClassifier()
classifier.fit(_snake_case , _snake_case )
return classifier
def lowerCAmelCase_ ( ) -> None:
'''simple docstring'''
__magic_name__ : str = load_iris()
__magic_name__ , __magic_name__ : List[Any] = data_handling(_snake_case )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : List[str] = train_test_split(
_snake_case , _snake_case , test_size=0.25 )
__magic_name__ : Optional[int] = iris["target_names"]
# Create an XGBoost Classifier from the training data
__magic_name__ : Dict = xgboost(_snake_case , _snake_case )
# Display the confusion matrix of the classifier with both training and test sets
ConfusionMatrixDisplay.from_estimator(
_snake_case , _snake_case , _snake_case , display_labels=_snake_case , cmap="Blues" , normalize="true" , )
plt.title("Normalized Confusion Matrix - IRIS Dataset" )
plt.show()
if __name__ == "__main__":
import doctest
doctest.testmod(verbose=True)
main()
| 281 |
import argparse
import csv
import logging
import os
import random
import numpy as np
import torch
from torch.utils.data import DataLoader, RandomSampler, SequentialSampler, TensorDataset
from tqdm import tqdm, trange
from transformers import (
CONFIG_NAME,
WEIGHTS_NAME,
AdamW,
OpenAIGPTDoubleHeadsModel,
OpenAIGPTTokenizer,
get_linear_schedule_with_warmup,
)
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s", datefmt="%m/%d/%Y %H:%M:%S", level=logging.INFO
)
snake_case : Optional[int] = logging.getLogger(__name__)
def lowerCAmelCase_ ( _snake_case : Dict , _snake_case : Union[str, Any] ) -> Tuple:
'''simple docstring'''
__magic_name__ : List[str] = np.argmax(_snake_case , axis=1 )
return np.sum(outputs == labels )
def lowerCAmelCase_ ( _snake_case : Optional[Any] ) -> Dict:
'''simple docstring'''
with open(_snake_case , encoding="utf_8" ) as f:
__magic_name__ : List[str] = csv.reader(_snake_case )
__magic_name__ : List[Any] = []
next(_snake_case ) # skip the first line
for line in tqdm(_snake_case ):
output.append((" ".join(line[1:5] ), line[5], line[6], int(line[-1] ) - 1) )
return output
def lowerCAmelCase_ ( _snake_case : str , _snake_case : Tuple , _snake_case : Union[str, Any] , _snake_case : List[Any] , _snake_case : Tuple , _snake_case : Optional[int] ) -> int:
'''simple docstring'''
__magic_name__ : Optional[int] = []
for dataset in encoded_datasets:
__magic_name__ : Union[str, Any] = len(_snake_case )
__magic_name__ : Dict = np.zeros((n_batch, 2, input_len) , dtype=np.intaa )
__magic_name__ : List[str] = np.zeros((n_batch, 2) , dtype=np.intaa )
__magic_name__ : Optional[int] = np.full((n_batch, 2, input_len) , fill_value=-100 , dtype=np.intaa )
__magic_name__ : int = np.zeros((n_batch,) , dtype=np.intaa )
for (
i,
(story, conta, conta, mc_label),
) in enumerate(_snake_case ):
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : Dict = [start_token] + story[:cap_length] + [delimiter_token] + conta[:cap_length] + [clf_token]
__magic_name__ : str = with_conta
__magic_name__ : Tuple = with_conta
__magic_name__ : Union[str, Any] = len(_snake_case ) - 1
__magic_name__ : int = len(_snake_case ) - 1
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[Any] = with_conta
__magic_name__ : Optional[int] = mc_label
__magic_name__ : str = (input_ids, mc_token_ids, lm_labels, mc_labels)
tensor_datasets.append(tuple(torch.tensor(_snake_case ) for t in all_inputs ) )
return tensor_datasets
def lowerCAmelCase_ ( ) -> List[Any]:
'''simple docstring'''
__magic_name__ : Any = argparse.ArgumentParser()
parser.add_argument("--model_name" , type=_snake_case , default="openai-gpt" , help="pretrained model name" )
parser.add_argument("--do_train" , action="store_true" , help="Whether to run training." )
parser.add_argument("--do_eval" , action="store_true" , help="Whether to run eval on the dev set." )
parser.add_argument(
"--output_dir" , default=_snake_case , type=_snake_case , required=_snake_case , help="The output directory where the model predictions and checkpoints will be written." , )
parser.add_argument("--train_dataset" , type=_snake_case , default="" )
parser.add_argument("--eval_dataset" , type=_snake_case , default="" )
parser.add_argument("--seed" , type=_snake_case , default=42 )
parser.add_argument("--num_train_epochs" , type=_snake_case , default=3 )
parser.add_argument("--train_batch_size" , type=_snake_case , default=8 )
parser.add_argument("--eval_batch_size" , type=_snake_case , default=16 )
parser.add_argument("--adam_epsilon" , default=1E-8 , type=_snake_case , help="Epsilon for Adam optimizer." )
parser.add_argument("--max_grad_norm" , type=_snake_case , default=1 )
parser.add_argument(
"--max_steps" , default=-1 , type=_snake_case , help=(
"If > 0: set total number of training steps to perform. Override num_train_epochs."
) , )
parser.add_argument(
"--gradient_accumulation_steps" , type=_snake_case , default=1 , help="Number of updates steps to accumulate before performing a backward/update pass." , )
parser.add_argument("--learning_rate" , type=_snake_case , default=6.25E-5 )
parser.add_argument("--warmup_steps" , default=0 , type=_snake_case , help="Linear warmup over warmup_steps." )
parser.add_argument("--lr_schedule" , type=_snake_case , default="warmup_linear" )
parser.add_argument("--weight_decay" , type=_snake_case , default=0.01 )
parser.add_argument("--lm_coef" , type=_snake_case , default=0.9 )
parser.add_argument("--n_valid" , type=_snake_case , default=374 )
parser.add_argument("--server_ip" , type=_snake_case , default="" , help="Can be used for distant debugging." )
parser.add_argument("--server_port" , type=_snake_case , default="" , help="Can be used for distant debugging." )
__magic_name__ : List[Any] = parser.parse_args()
print(_snake_case )
if args.server_ip and args.server_port:
# Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script
import ptvsd
print("Waiting for debugger attach" )
ptvsd.enable_attach(address=(args.server_ip, args.server_port) , redirect_output=_snake_case )
ptvsd.wait_for_attach()
random.seed(args.seed )
np.random.seed(args.seed )
torch.manual_seed(args.seed )
torch.cuda.manual_seed_all(args.seed )
__magic_name__ : Dict = torch.device("cuda" if torch.cuda.is_available() else "cpu" )
__magic_name__ : Optional[int] = torch.cuda.device_count()
logger.info("device: {}, n_gpu {}".format(_snake_case , _snake_case ) )
if not args.do_train and not args.do_eval:
raise ValueError("At least one of `do_train` or `do_eval` must be True." )
if not os.path.exists(args.output_dir ):
os.makedirs(args.output_dir )
# Load tokenizer and model
# This loading functions also add new tokens and embeddings called `special tokens`
# These new embeddings will be fine-tuned on the RocStories dataset
__magic_name__ : List[Any] = ["_start_", "_delimiter_", "_classify_"]
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.model_name )
tokenizer.add_tokens(_snake_case )
__magic_name__ : Optional[Any] = tokenizer.convert_tokens_to_ids(_snake_case )
__magic_name__ : List[str] = OpenAIGPTDoubleHeadsModel.from_pretrained(args.model_name )
model.resize_token_embeddings(len(_snake_case ) )
model.to(_snake_case )
# Load and encode the datasets
def tokenize_and_encode(_snake_case : str ):
if isinstance(_snake_case , _snake_case ):
return tokenizer.convert_tokens_to_ids(tokenizer.tokenize(_snake_case ) )
elif isinstance(_snake_case , _snake_case ):
return obj
return [tokenize_and_encode(_snake_case ) for o in obj]
logger.info("Encoding dataset..." )
__magic_name__ : Optional[int] = load_rocstories_dataset(args.train_dataset )
__magic_name__ : str = load_rocstories_dataset(args.eval_dataset )
__magic_name__ : int = (train_dataset, eval_dataset)
__magic_name__ : List[str] = tokenize_and_encode(_snake_case )
# Compute the max input length for the Transformer
__magic_name__ : Optional[Any] = model.config.n_positions // 2 - 2
__magic_name__ : Optional[int] = max(
len(story[:max_length] ) + max(len(conta[:max_length] ) , len(conta[:max_length] ) ) + 3
for dataset in encoded_datasets
for story, conta, conta, _ in dataset )
__magic_name__ : List[str] = min(_snake_case , model.config.n_positions ) # Max size of input for the pre-trained model
# Prepare inputs tensors and dataloaders
__magic_name__ : List[Any] = pre_process_datasets(_snake_case , _snake_case , _snake_case , *_snake_case )
__magic_name__ , __magic_name__ : Optional[int] = tensor_datasets[0], tensor_datasets[1]
__magic_name__ : Tuple = TensorDataset(*_snake_case )
__magic_name__ : Union[str, Any] = RandomSampler(_snake_case )
__magic_name__ : Dict = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.train_batch_size )
__magic_name__ : Any = TensorDataset(*_snake_case )
__magic_name__ : Optional[Any] = SequentialSampler(_snake_case )
__magic_name__ : int = DataLoader(_snake_case , sampler=_snake_case , batch_size=args.eval_batch_size )
# Prepare optimizer
if args.do_train:
if args.max_steps > 0:
__magic_name__ : Tuple = args.max_steps
__magic_name__ : List[str] = args.max_steps // (len(_snake_case ) // args.gradient_accumulation_steps) + 1
else:
__magic_name__ : List[str] = len(_snake_case ) // args.gradient_accumulation_steps * args.num_train_epochs
__magic_name__ : str = list(model.named_parameters() )
__magic_name__ : Dict = ["bias", "LayerNorm.bias", "LayerNorm.weight"]
__magic_name__ : str = [
{
"params": [p for n, p in param_optimizer if not any(nd in n for nd in no_decay )],
"weight_decay": args.weight_decay,
},
{"params": [p for n, p in param_optimizer if any(nd in n for nd in no_decay )], "weight_decay": 0.0},
]
__magic_name__ : str = AdamW(_snake_case , lr=args.learning_rate , eps=args.adam_epsilon )
__magic_name__ : List[str] = get_linear_schedule_with_warmup(
_snake_case , num_warmup_steps=args.warmup_steps , num_training_steps=_snake_case )
if args.do_train:
__magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = 0, 0, None
model.train()
for _ in trange(int(args.num_train_epochs ) , desc="Epoch" ):
__magic_name__ : List[str] = 0
__magic_name__ : Tuple = 0
__magic_name__ : Dict = tqdm(_snake_case , desc="Training" )
for step, batch in enumerate(_snake_case ):
__magic_name__ : Optional[Any] = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = batch
__magic_name__ : Optional[Any] = model(_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Optional[Any] = args.lm_coef * losses[0] + losses[1]
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
tr_loss += loss.item()
__magic_name__ : List[str] = (
loss.item() if exp_average_loss is None else 0.7 * exp_average_loss + 0.3 * loss.item()
)
nb_tr_steps += 1
__magic_name__ : int = "Training loss: {:.2e} lr: {:.2e}".format(_snake_case , scheduler.get_lr()[0] )
# Save a trained model
if args.do_train:
# Save a trained model, configuration and tokenizer
__magic_name__ : Dict = model.module if hasattr(_snake_case , "module" ) else model # Only save the model itself
# If we save using the predefined names, we can load using `from_pretrained`
__magic_name__ : List[Any] = os.path.join(args.output_dir , _snake_case )
__magic_name__ : Dict = os.path.join(args.output_dir , _snake_case )
torch.save(model_to_save.state_dict() , _snake_case )
model_to_save.config.to_json_file(_snake_case )
tokenizer.save_vocabulary(args.output_dir )
# Load a trained model and vocabulary that you have fine-tuned
__magic_name__ : Dict = OpenAIGPTDoubleHeadsModel.from_pretrained(args.output_dir )
__magic_name__ : Optional[int] = OpenAIGPTTokenizer.from_pretrained(args.output_dir )
model.to(_snake_case )
if args.do_eval:
model.eval()
__magic_name__ , __magic_name__ : Any = 0, 0
__magic_name__ , __magic_name__ : Union[str, Any] = 0, 0
for batch in tqdm(_snake_case , desc="Evaluating" ):
__magic_name__ : int = tuple(t.to(_snake_case ) for t in batch )
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Union[str, Any] = batch
with torch.no_grad():
__magic_name__ , __magic_name__ , __magic_name__ , __magic_name__ : Dict = model(
_snake_case , mc_token_ids=_snake_case , lm_labels=_snake_case , mc_labels=_snake_case )
__magic_name__ : Tuple = mc_logits.detach().cpu().numpy()
__magic_name__ : Any = mc_labels.to("cpu" ).numpy()
__magic_name__ : str = accuracy(_snake_case , _snake_case )
eval_loss += mc_loss.mean().item()
eval_accuracy += tmp_eval_accuracy
nb_eval_examples += input_ids.size(0 )
nb_eval_steps += 1
__magic_name__ : Tuple = eval_loss / nb_eval_steps
__magic_name__ : List[Any] = eval_accuracy / nb_eval_examples
__magic_name__ : int = tr_loss / nb_tr_steps if args.do_train else None
__magic_name__ : Any = {"eval_loss": eval_loss, "eval_accuracy": eval_accuracy, "train_loss": train_loss}
__magic_name__ : int = os.path.join(args.output_dir , "eval_results.txt" )
with open(_snake_case , "w" ) as writer:
logger.info("***** Eval results *****" )
for key in sorted(result.keys() ):
logger.info(" %s = %s" , _snake_case , str(result[key] ) )
writer.write("%s = %s\n" % (key, str(result[key] )) )
if __name__ == "__main__":
main()
| 281 | 1 |
from __future__ import annotations
import unittest
from transformers import DistilBertConfig, is_tf_available
from transformers.testing_utils import require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers.models.distilbert.modeling_tf_distilbert import (
TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFDistilBertForMaskedLM,
TFDistilBertForMultipleChoice,
TFDistilBertForQuestionAnswering,
TFDistilBertForSequenceClassification,
TFDistilBertForTokenClassification,
TFDistilBertModel,
)
class _snake_case :
def __init__( self , _a , ):
__magic_name__ : Optional[Any] = parent
__magic_name__ : Optional[Any] = 13
__magic_name__ : List[Any] = 7
__magic_name__ : int = True
__magic_name__ : int = True
__magic_name__ : Dict = False
__magic_name__ : int = True
__magic_name__ : List[str] = 99
__magic_name__ : Tuple = 32
__magic_name__ : Tuple = 2
__magic_name__ : str = 4
__magic_name__ : Union[str, Any] = 37
__magic_name__ : Union[str, Any] = "gelu"
__magic_name__ : int = 0.1
__magic_name__ : Optional[Any] = 0.1
__magic_name__ : List[Any] = 512
__magic_name__ : int = 16
__magic_name__ : Optional[Any] = 2
__magic_name__ : Any = 0.02
__magic_name__ : Union[str, Any] = 3
__magic_name__ : Tuple = 4
__magic_name__ : Tuple = None
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
__magic_name__ : str = None
if self.use_input_mask:
__magic_name__ : Optional[int] = random_attention_mask([self.batch_size, self.seq_length] )
__magic_name__ : Tuple = None
__magic_name__ : int = None
__magic_name__ : Dict = None
if self.use_labels:
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
__magic_name__ : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
__magic_name__ : int = ids_tensor([self.batch_size] , self.num_choices )
__magic_name__ : Optional[Any] = DistilBertConfig(
vocab_size=self.vocab_size , dim=self.hidden_size , n_layers=self.num_hidden_layers , n_heads=self.num_attention_heads , hidden_dim=self.intermediate_size , hidden_act=self.hidden_act , dropout=self.hidden_dropout_prob , attention_dropout=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , initializer_range=self.initializer_range , )
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a ):
__magic_name__ : Optional[Any] = TFDistilBertModel(config=_a )
__magic_name__ : Dict = {"input_ids": input_ids, "attention_mask": input_mask}
__magic_name__ : str = model(_a )
__magic_name__ : List[str] = [input_ids, input_mask]
__magic_name__ : Dict = model(_a )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a ):
__magic_name__ : List[str] = TFDistilBertForMaskedLM(config=_a )
__magic_name__ : Any = {"input_ids": input_ids, "attention_mask": input_mask}
__magic_name__ : List[str] = model(_a )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a ):
__magic_name__ : str = TFDistilBertForQuestionAnswering(config=_a )
__magic_name__ : Optional[Any] = {
"input_ids": input_ids,
"attention_mask": input_mask,
}
__magic_name__ : str = model(_a )
self.parent.assertEqual(result.start_logits.shape , (self.batch_size, self.seq_length) )
self.parent.assertEqual(result.end_logits.shape , (self.batch_size, self.seq_length) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a ):
__magic_name__ : List[str] = self.num_labels
__magic_name__ : List[Any] = TFDistilBertForSequenceClassification(_a )
__magic_name__ : Optional[Any] = {"input_ids": input_ids, "attention_mask": input_mask}
__magic_name__ : int = model(_a )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a ):
__magic_name__ : List[Any] = self.num_choices
__magic_name__ : int = TFDistilBertForMultipleChoice(_a )
__magic_name__ : int = tf.tile(tf.expand_dims(_a , 1 ) , (1, self.num_choices, 1) )
__magic_name__ : Optional[int] = tf.tile(tf.expand_dims(_a , 1 ) , (1, self.num_choices, 1) )
__magic_name__ : Optional[int] = {
"input_ids": multiple_choice_inputs_ids,
"attention_mask": multiple_choice_input_mask,
}
__magic_name__ : Optional[Any] = model(_a )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def SCREAMING_SNAKE_CASE ( self , _a , _a , _a , _a , _a , _a ):
__magic_name__ : Dict = self.num_labels
__magic_name__ : List[str] = TFDistilBertForTokenClassification(_a )
__magic_name__ : int = {"input_ids": input_ids, "attention_mask": input_mask}
__magic_name__ : List[str] = model(_a )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = self.prepare_config_and_inputs()
((__magic_name__) , (__magic_name__) , (__magic_name__) , (__magic_name__) , (__magic_name__) , (__magic_name__)) : int = config_and_inputs
__magic_name__ : str = {"input_ids": input_ids, "attention_mask": input_mask}
return config, inputs_dict
@require_tf
class _snake_case ( snake_case , snake_case , unittest.TestCase ):
UpperCamelCase__ = (
(
TFDistilBertModel,
TFDistilBertForMaskedLM,
TFDistilBertForQuestionAnswering,
TFDistilBertForSequenceClassification,
TFDistilBertForTokenClassification,
TFDistilBertForMultipleChoice,
)
if is_tf_available()
else None
)
UpperCamelCase__ = (
{
'feature-extraction': TFDistilBertModel,
'fill-mask': TFDistilBertForMaskedLM,
'question-answering': TFDistilBertForQuestionAnswering,
'text-classification': TFDistilBertForSequenceClassification,
'token-classification': TFDistilBertForTokenClassification,
'zero-shot': TFDistilBertForSequenceClassification,
}
if is_tf_available()
else {}
)
UpperCamelCase__ = False
UpperCamelCase__ = False
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : List[str] = TFDistilBertModelTester(self )
__magic_name__ : List[str] = ConfigTester(self , config_class=_a , dim=37 )
def SCREAMING_SNAKE_CASE ( self ):
self.config_tester.run_common_tests()
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[int] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_model(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_masked_lm(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Dict = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_question_answering(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_sequence_classification(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_multiple_choice(*_a )
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_distilbert_for_token_classification(*_a )
@slow
def SCREAMING_SNAKE_CASE ( self ):
for model_name in list(TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1] ):
__magic_name__ : int = TFDistilBertModel.from_pretrained(_a )
self.assertIsNotNone(_a )
@require_tf
class _snake_case ( unittest.TestCase ):
@slow
def SCREAMING_SNAKE_CASE ( self ):
__magic_name__ : Optional[Any] = TFDistilBertModel.from_pretrained("distilbert-base-uncased" )
__magic_name__ : Union[str, Any] = tf.constant([[0, 1, 2, 3, 4, 5]] )
__magic_name__ : List[Any] = model(_a )[0]
__magic_name__ : Optional[Any] = [1, 6, 768]
self.assertEqual(output.shape , _a )
__magic_name__ : List[str] = tf.constant(
[
[
[0.19_26_18_85, -0.13_73_29_55, 0.4_11_97_99],
[0.22_15_01_56, -0.07_42_26_61, 0.39_03_72_04],
[0.22_75_60_18, -0.0_89_64_14, 0.3_70_14_67],
]
] )
tf.debugging.assert_near(output[:, :3, :3] , _a , atol=1e-4 )
| 281 |
from . import __version__
# Backward compatibility imports, to make sure all those objects can be found in file_utils
from .utils import (
CLOUDFRONT_DISTRIB_PREFIX,
CONFIG_NAME,
DISABLE_TELEMETRY,
DUMMY_INPUTS,
DUMMY_MASK,
ENV_VARS_TRUE_AND_AUTO_VALUES,
ENV_VARS_TRUE_VALUES,
FEATURE_EXTRACTOR_NAME,
FLAX_WEIGHTS_NAME,
HF_MODULES_CACHE,
HUGGINGFACE_CO_PREFIX,
HUGGINGFACE_CO_RESOLVE_ENDPOINT,
MODEL_CARD_NAME,
MULTIPLE_CHOICE_DUMMY_INPUTS,
PYTORCH_PRETRAINED_BERT_CACHE,
PYTORCH_TRANSFORMERS_CACHE,
S3_BUCKET_PREFIX,
SENTENCEPIECE_UNDERLINE,
SPIECE_UNDERLINE,
TF2_WEIGHTS_NAME,
TF_WEIGHTS_NAME,
TORCH_FX_REQUIRED_VERSION,
TRANSFORMERS_CACHE,
TRANSFORMERS_DYNAMIC_MODULE_NAME,
USE_JAX,
USE_TF,
USE_TORCH,
WEIGHTS_INDEX_NAME,
WEIGHTS_NAME,
ContextManagers,
DummyObject,
EntryNotFoundError,
ExplicitEnum,
ModelOutput,
PaddingStrategy,
PushToHubMixin,
RepositoryNotFoundError,
RevisionNotFoundError,
TensorType,
_LazyModule,
add_code_sample_docstrings,
add_end_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
cached_property,
copy_func,
default_cache_path,
define_sagemaker_information,
get_cached_models,
get_file_from_repo,
get_full_repo_name,
get_torch_version,
has_file,
http_user_agent,
is_apex_available,
is_bsa_available,
is_coloredlogs_available,
is_datasets_available,
is_detectrona_available,
is_faiss_available,
is_flax_available,
is_ftfy_available,
is_in_notebook,
is_ipex_available,
is_librosa_available,
is_offline_mode,
is_onnx_available,
is_pandas_available,
is_phonemizer_available,
is_protobuf_available,
is_psutil_available,
is_pyanvml_available,
is_pyctcdecode_available,
is_pytesseract_available,
is_pytorch_quantization_available,
is_rjieba_available,
is_sagemaker_dp_enabled,
is_sagemaker_mp_enabled,
is_scipy_available,
is_sentencepiece_available,
is_seqio_available,
is_sklearn_available,
is_soundfile_availble,
is_spacy_available,
is_speech_available,
is_tensor,
is_tensorflow_probability_available,
is_tfaonnx_available,
is_tf_available,
is_timm_available,
is_tokenizers_available,
is_torch_available,
is_torch_bfaa_available,
is_torch_cuda_available,
is_torch_fx_available,
is_torch_fx_proxy,
is_torch_mps_available,
is_torch_tfaa_available,
is_torch_tpu_available,
is_torchaudio_available,
is_training_run_on_sagemaker,
is_vision_available,
replace_return_docstrings,
requires_backends,
to_numpy,
to_py_obj,
torch_only_method,
)
| 281 | 1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.