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
|
|---|---|---|---|---|
from collections.abc import Sequence
def UpperCamelCase ( __lowerCamelCase : Sequence[float] , __lowerCamelCase : bool = False ):
if not arr:
return 0
snake_case : Tuple = 0 if allow_empty_subarrays else float("-inf" )
snake_case : int = 0.0
for num in arr:
snake_case : Dict = max(0 if allow_empty_subarrays else num , curr_sum + num )
snake_case : str = max(__lowerCamelCase , __lowerCamelCase )
return max_sum
if __name__ == "__main__":
from doctest import testmod
testmod()
__lowerCamelCase = [-2, 1, -3, 4, -1, 2, 1, -5, 4]
print(F'{max_subarray_sum(nums) = }')
| 59
|
import logging
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
import datasets
import numpy as np
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PreTrainedTokenizer,
TFAutoModelForSequenceClassification,
TFTrainer,
TFTrainingArguments,
)
from transformers.utils import logging as hf_logging
hf_logging.set_verbosity_info()
hf_logging.enable_default_handler()
hf_logging.enable_explicit_format()
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : PreTrainedTokenizer , __lowerCamelCase : int , __lowerCamelCase : Optional[int] = None , ):
snake_case : int = {}
if train_file is not None:
snake_case : List[Any] = [train_file]
if eval_file is not None:
snake_case : Optional[int] = [eval_file]
if test_file is not None:
snake_case : Any = [test_file]
snake_case : int = datasets.load_dataset("csv" , data_files=__lowerCamelCase )
snake_case : str = list(ds[list(files.keys() )[0]].features.keys() )
snake_case : int = features_name.pop(__lowerCamelCase )
snake_case : str = list(set(ds[list(files.keys() )[0]][label_name] ) )
snake_case : str = {label: i for i, label in enumerate(__lowerCamelCase )}
snake_case : List[Any] = tokenizer.model_input_names
snake_case : List[Any] = {}
if len(__lowerCamelCase ) == 1:
for k in files.keys():
snake_case : Tuple = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
example[features_name[0]] , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" ) , batched=__lowerCamelCase , )
elif len(__lowerCamelCase ) == 2:
for k in files.keys():
snake_case : List[Any] = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
(example[features_name[0]], example[features_name[1]]) , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" , ) , batched=__lowerCamelCase , )
def gen_train():
for ex in transformed_ds[datasets.Split.TRAIN]:
snake_case : Dict = {k: v for k, v in ex.items() if k in input_names}
snake_case : Union[str, Any] = labelaid[ex[label_name]]
yield (d, label)
def gen_val():
for ex in transformed_ds[datasets.Split.VALIDATION]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : Any = labelaid[ex[label_name]]
yield (d, label)
def gen_test():
for ex in transformed_ds[datasets.Split.TEST]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : List[str] = labelaid[ex[label_name]]
yield (d, label)
snake_case : int = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TRAIN in transformed_ds
else None
)
if train_ds is not None:
snake_case : Optional[Any] = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) )
snake_case : Tuple = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.VALIDATION in transformed_ds
else None
)
if val_ds is not None:
snake_case : List[str] = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) )
snake_case : Optional[int] = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TEST in transformed_ds
else None
)
if test_ds is not None:
snake_case : str = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) )
return train_ds, val_ds, test_ds, labelaid
__lowerCamelCase = logging.getLogger(__name__)
@dataclass
class UpperCAmelCase :
A__ : int = field(metadata={"help": "Which column contains the label"} )
A__ : str = field(default=A_ ,metadata={"help": "The path of the training file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the development file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the test file"} )
A__ : int = field(
default=1_28 ,metadata={
"help": (
"The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
} ,)
A__ : bool = field(
default=A_ ,metadata={"help": "Overwrite the cached training and evaluation sets"} )
@dataclass
class UpperCAmelCase :
A__ : str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained config name or path if not the same as model_name"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"} )
A__ : bool = field(default=A_ ,metadata={"help": "Set this flag to use fast tokenization."} )
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
# or just modify its tokenizer_config.json.
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} ,)
def UpperCamelCase ( ):
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
snake_case : List[Any] = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) )
snake_case , snake_case , snake_case : int = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir )
and os.listdir(training_args.output_dir )
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
f"""Output directory ({training_args.output_dir}) already exists and is not empty. Use"""
" --overwrite_output_dir to overcome." )
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(
f"""n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, """
f"""16-bits training: {training_args.fpaa}""" )
logger.info(f"""Training/evaluation parameters {training_args}""" )
# Load pretrained model and tokenizer
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
snake_case : Tuple = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path , cache_dir=model_args.cache_dir , )
snake_case , snake_case , snake_case , snake_case : Tuple = get_tfds(
train_file=data_args.train_file , eval_file=data_args.dev_file , test_file=data_args.test_file , tokenizer=__lowerCamelCase , label_column_id=data_args.label_column_id , max_seq_length=data_args.max_seq_length , )
snake_case : Optional[Any] = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path , num_labels=len(__lowerCamelCase ) , labelaid=__lowerCamelCase , idalabel={id: label for label, id in labelaid.items()} , finetuning_task="text-classification" , cache_dir=model_args.cache_dir , )
with training_args.strategy.scope():
snake_case : int = TFAutoModelForSequenceClassification.from_pretrained(
model_args.model_name_or_path , from_pt=bool(".bin" in model_args.model_name_or_path ) , config=__lowerCamelCase , cache_dir=model_args.cache_dir , )
def compute_metrics(__lowerCamelCase : EvalPrediction ) -> Dict:
snake_case : Optional[int] = np.argmax(p.predictions , axis=1 )
return {"acc": (preds == p.label_ids).mean()}
# Initialize our Trainer
snake_case : int = TFTrainer(
model=__lowerCamelCase , args=__lowerCamelCase , train_dataset=__lowerCamelCase , eval_dataset=__lowerCamelCase , compute_metrics=__lowerCamelCase , )
# Training
if training_args.do_train:
trainer.train()
trainer.save_model()
tokenizer.save_pretrained(training_args.output_dir )
# Evaluation
snake_case : int = {}
if training_args.do_eval:
logger.info("*** Evaluate ***" )
snake_case : Any = trainer.evaluate()
snake_case : List[Any] = os.path.join(training_args.output_dir , "eval_results.txt" )
with open(__lowerCamelCase , "w" ) as writer:
logger.info("***** Eval results *****" )
for key, value in result.items():
logger.info(f""" {key} = {value}""" )
writer.write(f"""{key} = {value}\n""" )
results.update(__lowerCamelCase )
return results
if __name__ == "__main__":
main()
| 59
| 1
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""google/canine-s""": """https://huggingface.co/google/canine-s/resolve/main/config.json""",
# See all CANINE models at https://huggingface.co/models?filter=canine
}
class UpperCAmelCase ( A_ ):
A__ : List[Any] = "canine"
def __init__(self : Dict , snake_case__ : Dict=7_68 , snake_case__ : Tuple=12 , snake_case__ : Optional[int]=12 , snake_case__ : Optional[Any]=30_72 , snake_case__ : List[Any]="gelu" , snake_case__ : Dict=0.1 , snake_case__ : Optional[Any]=0.1 , snake_case__ : List[str]=1_63_84 , snake_case__ : List[Any]=16 , snake_case__ : List[Any]=0.02 , snake_case__ : Tuple=1e-12 , snake_case__ : Tuple=0 , snake_case__ : Optional[int]=0XE_0_0_0 , snake_case__ : Dict=0XE_0_0_1 , snake_case__ : int=4 , snake_case__ : Union[str, Any]=4 , snake_case__ : Union[str, Any]=8 , snake_case__ : List[str]=1_63_84 , snake_case__ : List[str]=1_28 , **snake_case__ : Optional[Any] , ) -> List[str]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , bos_token_id=snake_case__ , eos_token_id=snake_case__ , **snake_case__ )
snake_case : Any = max_position_embeddings
snake_case : List[Any] = hidden_size
snake_case : Optional[int] = num_hidden_layers
snake_case : Tuple = num_attention_heads
snake_case : Dict = intermediate_size
snake_case : List[str] = hidden_act
snake_case : Tuple = hidden_dropout_prob
snake_case : List[str] = attention_probs_dropout_prob
snake_case : str = initializer_range
snake_case : int = type_vocab_size
snake_case : List[Any] = layer_norm_eps
# Character config:
snake_case : str = downsampling_rate
snake_case : Dict = upsampling_kernel_size
snake_case : List[str] = num_hash_functions
snake_case : Optional[int] = num_hash_buckets
snake_case : Dict = local_transformer_stride
| 59
|
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Any ) -> List[str]:
'''simple docstring'''
snake_case : int = tempfile.mkdtemp()
# fmt: off
snake_case : Optional[int] = ["[UNK]", "[CLS]", "[SEP]", "[PAD]", "[MASK]", "want", "##want", "##ed", "wa", "un", "runn", "##ing", ",", "low", "lowest"]
# fmt: on
snake_case : List[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] ) )
snake_case : int = {
"do_resize": True,
"size": {"height": 18, "width": 18},
"do_normalize": True,
"image_mean": [0.5, 0.5, 0.5],
"image_std": [0.5, 0.5, 0.5],
}
snake_case : Optional[Any] = os.path.join(self.tmpdirname , snake_case__ )
with open(self.image_processor_file , "w" , encoding="utf-8" ) as fp:
json.dump(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : str ) -> Optional[int]:
'''simple docstring'''
return BertTokenizer.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : List[str] ) -> int:
'''simple docstring'''
return ViTImageProcessor.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
shutil.rmtree(self.tmpdirname )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : List[Any] = [np.random.randint(2_55 , size=(3, 30, 4_00) , dtype=np.uinta )]
snake_case : Optional[int] = [Image.fromarray(np.moveaxis(snake_case__ , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.get_tokenizer()
snake_case : Optional[Any] = self.get_image_processor()
snake_case : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
processor.save_pretrained(self.tmpdirname )
snake_case : Any = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
snake_case : Optional[int] = self.get_tokenizer(bos_token="(BOS)" , eos_token="(EOS)" )
snake_case : Tuple = self.get_image_processor(do_normalize=snake_case__ , padding_value=1.0 )
snake_case : List[str] = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token="(BOS)" , eos_token="(EOS)" , do_normalize=snake_case__ , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Optional[int] = self.get_tokenizer()
snake_case : List[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = self.prepare_image_inputs()
snake_case : str = image_processor(snake_case__ , return_tensors="np" )
snake_case : Any = processor(images=snake_case__ , return_tensors="np" )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.get_image_processor()
snake_case : int = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Tuple = "lower newer"
snake_case : Tuple = processor(text=snake_case__ )
snake_case : Union[str, Any] = tokenizer(snake_case__ )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[int]:
'''simple docstring'''
snake_case : List[Any] = self.get_image_processor()
snake_case : Dict = self.get_tokenizer()
snake_case : Dict = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : int = "lower newer"
snake_case : Dict = self.prepare_image_inputs()
snake_case : Union[str, Any] = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , ["input_ids", "token_type_ids", "attention_mask", "pixel_values"] )
# test if it raises when no input is passed
with self.assertRaises(snake_case__ ):
processor()
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : Tuple = self.get_image_processor()
snake_case : Optional[Any] = self.get_tokenizer()
snake_case : Tuple = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : List[str] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
snake_case : List[Any] = processor.batch_decode(snake_case__ )
snake_case : Union[str, Any] = tokenizer.batch_decode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Union[str, Any] = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = "lower newer"
snake_case : List[Any] = self.prepare_image_inputs()
snake_case : Tuple = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 59
| 1
|
import torch
from diffusers import StableDiffusionPipeline
__lowerCamelCase = """path-to-your-trained-model"""
__lowerCamelCase = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.floataa).to("""cuda""")
__lowerCamelCase = """A photo of sks dog in a bucket"""
__lowerCamelCase = pipe(prompt, num_inference_steps=50, guidance_scale=7.5).images[0]
image.save("""dog-bucket.png""")
| 59
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__lowerCamelCase = {
"""configuration_biogpt""": ["""BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP""", """BioGptConfig"""],
"""tokenization_biogpt""": ["""BioGptTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""BioGptForCausalLM""",
"""BioGptForTokenClassification""",
"""BioGptForSequenceClassification""",
"""BioGptModel""",
"""BioGptPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_biogpt import BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, BioGptConfig
from .tokenization_biogpt import BioGptTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_biogpt import (
BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
BioGptForCausalLM,
BioGptForSequenceClassification,
BioGptForTokenClassification,
BioGptModel,
BioGptPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
| 1
|
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Union[str, Any] = len(__lowerCamelCase )
while cur > 1:
# Find the maximum number in arr
snake_case : Dict = arr.index(max(arr[0:cur] ) )
# Reverse from 0 to mi
snake_case : Tuple = arr[mi::-1] + arr[mi + 1 : len(__lowerCamelCase )]
# Reverse whole list
snake_case : Union[str, Any] = arr[cur - 1 :: -1] + arr[cur : len(__lowerCamelCase )]
cur -= 1
return arr
if __name__ == "__main__":
__lowerCamelCase = input("""Enter numbers separated by a comma:\n""").strip()
__lowerCamelCase = [int(item) for item in user_input.split(""",""")]
print(pancake_sort(unsorted))
| 59
|
import collections
import inspect
import unittest
from typing import Dict, List, Tuple
from transformers import MaskFormerSwinConfig
from transformers.testing_utils import require_torch, require_torch_multi_gpu, torch_device
from transformers.utils import is_torch_available
from ...test_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import MaskFormerSwinBackbone
from transformers.models.maskformer import MaskFormerSwinModel
class UpperCAmelCase :
def __init__(self : Dict , snake_case__ : Dict , snake_case__ : Any=13 , snake_case__ : Any=32 , snake_case__ : Optional[Any]=2 , snake_case__ : Union[str, Any]=3 , snake_case__ : List[Any]=16 , snake_case__ : int=[1, 2, 1] , snake_case__ : Dict=[2, 2, 4] , snake_case__ : Dict=2 , snake_case__ : Tuple=2.0 , snake_case__ : Optional[int]=True , snake_case__ : Union[str, Any]=0.0 , snake_case__ : Any=0.0 , snake_case__ : Union[str, Any]=0.1 , snake_case__ : int="gelu" , snake_case__ : Optional[int]=False , snake_case__ : List[Any]=True , snake_case__ : List[str]=0.02 , snake_case__ : int=1e-5 , snake_case__ : List[str]=True , snake_case__ : Union[str, Any]=None , snake_case__ : List[Any]=True , snake_case__ : Optional[Any]=10 , snake_case__ : Optional[Any]=8 , snake_case__ : Any=["stage1", "stage2", "stage3"] , snake_case__ : Tuple=[1, 2, 3] , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = parent
snake_case : Optional[int] = batch_size
snake_case : Union[str, Any] = image_size
snake_case : Dict = patch_size
snake_case : Optional[Any] = num_channels
snake_case : Union[str, Any] = embed_dim
snake_case : int = depths
snake_case : List[str] = num_heads
snake_case : Union[str, Any] = window_size
snake_case : Union[str, Any] = mlp_ratio
snake_case : List[Any] = qkv_bias
snake_case : List[Any] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Union[str, Any] = drop_path_rate
snake_case : int = hidden_act
snake_case : Optional[int] = use_absolute_embeddings
snake_case : int = patch_norm
snake_case : Union[str, Any] = layer_norm_eps
snake_case : Any = initializer_range
snake_case : Optional[Any] = is_training
snake_case : Tuple = scope
snake_case : Optional[int] = use_labels
snake_case : Optional[Any] = type_sequence_label_size
snake_case : Union[str, Any] = encoder_stride
snake_case : Any = out_features
snake_case : Tuple = out_indices
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case : Optional[Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case : int = None
if self.use_labels:
snake_case : Tuple = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : Dict = self.get_config()
return config, pixel_values, labels
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> int:
'''simple docstring'''
return MaskFormerSwinConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , depths=self.depths , num_heads=self.num_heads , window_size=self.window_size , mlp_ratio=self.mlp_ratio , qkv_bias=self.qkv_bias , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , drop_path_rate=self.drop_path_rate , hidden_act=self.hidden_act , use_absolute_embeddings=self.use_absolute_embeddings , path_norm=self.patch_norm , layer_norm_eps=self.layer_norm_eps , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , out_features=self.out_features , out_indices=self.out_indices , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : List[Any] , snake_case__ : List[str] , snake_case__ : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
snake_case : Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case : int = int(config.embed_dim * 2 ** (len(config.depths ) - 1) )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, expected_seq_len, expected_dim) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = MaskFormerSwinBackbone(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [13, 16, 16, 16] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , [16, 32, 64] )
# verify ValueError
with self.parent.assertRaises(snake_case__ ):
snake_case : Tuple = ["stem"]
snake_case : List[Any] = MaskFormerSwinBackbone(config=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self.prepare_config_and_inputs()
snake_case , snake_case , snake_case : List[Any] = config_and_inputs
snake_case : int = {"pixel_values": pixel_values}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : List[str] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
A__ : str = {"feature-extraction": MaskFormerSwinModel} if is_torch_available() else {}
A__ : Optional[Any] = False
A__ : List[Any] = False
A__ : List[str] = False
A__ : List[str] = False
A__ : Union[str, Any] = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = MaskFormerSwinModelTester(self )
snake_case : Optional[int] = ConfigTester(self , config_class=snake_case__ , embed_dim=37 )
@require_torch_multi_gpu
@unittest.skip(
reason=(
"`MaskFormerSwinModel` outputs `hidden_states_spatial_dimensions` which doesn't work well with"
" `nn.DataParallel`"
) )
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : str ) -> List[str]:
'''simple docstring'''
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
return
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Dict:
'''simple docstring'''
snake_case : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*snake_case__ )
@unittest.skip("Swin does not use inputs_embeds" )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
pass
@unittest.skip("Swin does not support feedforward chunking" )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Dict:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : int = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case : List[Any] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , nn.Linear ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case , snake_case : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : str = model_class(snake_case__ )
snake_case : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case : Optional[Any] = [*signature.parameters.keys()]
snake_case : Tuple = ["pixel_values"]
self.assertListEqual(arg_names[:1] , snake_case__ )
@unittest.skip(reason="MaskFormerSwin is only used as backbone and doesn't support output_attentions" )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="MaskFormerSwin is only used as an internal backbone" )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
with torch.no_grad():
snake_case : Any = model(**self._prepare_for_class(snake_case__ , snake_case__ ) )
snake_case : int = outputs.hidden_states
snake_case : Union[str, Any] = getattr(
self.model_tester , "expected_num_hidden_layers" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case__ ) , snake_case__ )
# Swin has a different seq_length
snake_case : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : Tuple = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : int = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
for model_class in self.all_model_classes:
snake_case : int = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Dict = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case , snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Any = 3
snake_case : List[str] = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
snake_case : Tuple = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : str = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case : str = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes:
snake_case : str = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Optional[Any] = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
@unittest.skip(reason="MaskFormerSwin doesn't have pretrained checkpoints" )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : int ) -> str:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case__ : Union[str, Any] ):
snake_case : Any = 0
return t
def check_equivalence(snake_case__ : Union[str, Any] , snake_case__ : int , snake_case__ : List[str] , snake_case__ : Optional[int]={} ):
with torch.no_grad():
snake_case : Optional[Any] = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ )
snake_case : Tuple = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ ).to_tuple()
def recursive_check(snake_case__ : List[str] , snake_case__ : Optional[Any] ):
if isinstance(snake_case__ , (List, Tuple) ):
for tuple_iterable_value, dict_iterable_value in zip(snake_case__ , snake_case__ ):
recursive_check(snake_case__ , snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values() , dict_object.values() ):
recursive_check(snake_case__ , snake_case__ )
elif tuple_object is None:
return
else:
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(snake_case__ ) , set_nan_tensor_to_zero(snake_case__ ) , atol=1e-5 ) , msg=(
"Tuple and dict output are not equal. Difference:"
f""" {torch.max(torch.abs(tuple_object - dict_object ) )}. Tuple has `nan`:"""
f""" {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}. Dict has"""
f""" `nan`: {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}."""
) , )
recursive_check(snake_case__ , snake_case__ )
for model_class in self.all_model_classes:
snake_case : Optional[int] = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Union[str, Any] = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : Optional[Any] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Dict = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : List[Any] = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
snake_case : Any = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : List[str] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
@require_torch
class UpperCAmelCase ( unittest.TestCase ,A_ ):
A__ : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
A__ : int = MaskFormerSwinConfig
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModelTester(self )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case : Dict = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Optional[int] = inputs_dict["pixel_values"].shape[0]
for backbone_class in self.all_model_classes:
snake_case : Optional[int] = backbone_class(snake_case__ )
backbone.to(snake_case__ )
backbone.eval()
snake_case : Union[str, Any] = backbone(**snake_case__ )
# Test default outputs and verify feature maps
self.assertIsInstance(outputs.feature_maps , snake_case__ )
self.assertTrue(len(outputs.feature_maps ) == len(backbone.channels ) )
for feature_map, n_channels in zip(outputs.feature_maps , backbone.channels ):
self.assertTrue(feature_map.shape[:2] , (batch_size, n_channels) )
self.assertIsNone(outputs.hidden_states )
self.assertIsNone(outputs.attentions )
# Test output_hidden_states=True
snake_case : Optional[int] = backbone(**snake_case__ , output_hidden_states=snake_case__ )
self.assertIsNotNone(outputs.hidden_states )
self.assertTrue(len(outputs.hidden_states ) , len(backbone.stage_names ) )
# We skip the stem layer
for hidden_states, n_channels in zip(outputs.hidden_states[1:] , backbone.channels ):
for hidden_state in hidden_states:
# Hidden states are in the format (batch_size, (height * width), n_channels)
snake_case , snake_case , snake_case : Dict = hidden_state.shape
self.assertTrue((h_batch_size, h_n_channels) , (batch_size, n_channels) )
# Test output_attentions=True
if self.has_attentions:
snake_case : Optional[Any] = backbone(**snake_case__ , output_attentions=snake_case__ )
self.assertIsNotNone(outputs.attentions )
| 59
| 1
|
def UpperCamelCase ( __lowerCamelCase : list ):
if any(not isinstance(__lowerCamelCase , __lowerCamelCase ) or x < 0 for x in sequence ):
raise TypeError("Sequence must be list of non-negative integers" )
for _ in range(len(__lowerCamelCase ) ):
for i, (rod_upper, rod_lower) in enumerate(zip(__lowerCamelCase , sequence[1:] ) ):
if rod_upper > rod_lower:
sequence[i] -= rod_upper - rod_lower
sequence[i + 1] += rod_upper - rod_lower
return sequence
if __name__ == "__main__":
assert bead_sort([5, 4, 3, 2, 1]) == [1, 2, 3, 4, 5]
assert bead_sort([7, 9, 4, 3, 5]) == [3, 4, 5, 7, 9]
| 59
|
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : List[str] = []
snake_case : Optional[int] = []
snake_case : Any = []
for rt in rc.restypes:
snake_case : List[Any] = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
snake_case : str = {name: i for i, name in enumerate(__lowerCamelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
snake_case : Optional[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : List[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : int = torch.tensor(
__lowerCamelCase , dtype=torch.floataa , device=protein["aatype"].device , )
snake_case : int = protein["aatype"].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
snake_case : List[Any] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : str = restype_atomaa_mask[protein_aatype]
snake_case : str = residx_atomaa_mask
snake_case : Any = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
snake_case : List[str] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : List[Any] = residx_atomaa_to_atomaa.long()
# create the corresponding mask
snake_case : Union[str, Any] = torch.zeros([21, 37] , dtype=torch.floataa , device=protein["aatype"].device )
for restype, restype_letter in enumerate(rc.restypes ):
snake_case : Optional[int] = rc.restype_atoa[restype_letter]
snake_case : Any = rc.residue_atoms[restype_name]
for atom_name in atom_names:
snake_case : List[Any] = rc.atom_order[atom_name]
snake_case : Optional[Any] = 1
snake_case : List[Any] = restype_atomaa_mask[protein_aatype]
snake_case : int = residx_atomaa_mask
return protein
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : Dict = tree_map(lambda __lowerCamelCase : torch.tensor(__lowerCamelCase , device=batch["aatype"].device ) , __lowerCamelCase , np.ndarray )
snake_case : List[str] = tensor_tree_map(lambda __lowerCamelCase : np.array(__lowerCamelCase ) , make_atomaa_masks(__lowerCamelCase ) )
return out
| 59
| 1
|
import argparse
import json
import os
import time
import zipfile
from get_ci_error_statistics import download_artifact, get_artifacts_links
from transformers import logging
__lowerCamelCase = logging.get_logger(__name__)
def UpperCamelCase ( __lowerCamelCase : Any , __lowerCamelCase : List[Any] ):
snake_case : int = set()
snake_case : Tuple = []
def parse_line(__lowerCamelCase : Optional[Any] ):
for line in fp:
if isinstance(__lowerCamelCase , __lowerCamelCase ):
snake_case : Tuple = line.decode("UTF-8" )
if "warnings summary (final)" in line:
continue
# This means we are outside the body of a warning
elif not line.startswith(" " ):
# process a single warning and move it to `selected_warnings`.
if len(__lowerCamelCase ) > 0:
snake_case : List[str] = "\n".join(__lowerCamelCase )
# Only keep the warnings specified in `targets`
if any(f""": {x}: """ in warning for x in targets ):
selected_warnings.add(__lowerCamelCase )
buffer.clear()
continue
else:
snake_case : Tuple = line.strip()
buffer.append(__lowerCamelCase )
if from_gh:
for filename in os.listdir(__lowerCamelCase ):
snake_case : List[Any] = os.path.join(__lowerCamelCase , __lowerCamelCase )
if not os.path.isdir(__lowerCamelCase ):
# read the file
if filename != "warnings.txt":
continue
with open(__lowerCamelCase ) as fp:
parse_line(__lowerCamelCase )
else:
try:
with zipfile.ZipFile(__lowerCamelCase ) as z:
for filename in z.namelist():
if not os.path.isdir(__lowerCamelCase ):
# read the file
if filename != "warnings.txt":
continue
with z.open(__lowerCamelCase ) as fp:
parse_line(__lowerCamelCase )
except Exception:
logger.warning(
f"""{artifact_path} is either an invalid zip file or something else wrong. This file is skipped.""" )
return selected_warnings
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : str ):
snake_case : Union[str, Any] = set()
snake_case : List[Any] = [os.path.join(__lowerCamelCase , __lowerCamelCase ) for p in os.listdir(__lowerCamelCase ) if (p.endswith(".zip" ) or from_gh)]
for p in paths:
selected_warnings.update(extract_warnings_from_single_artifact(__lowerCamelCase , __lowerCamelCase ) )
return selected_warnings
if __name__ == "__main__":
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] ):
return values.split("," )
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument("""--workflow_run_id""", type=str, required=True, help="""A GitHub Actions workflow run id.""")
parser.add_argument(
"""--output_dir""",
type=str,
required=True,
help="""Where to store the downloaded artifacts and other result files.""",
)
parser.add_argument("""--token""", default=None, type=str, help="""A token that has actions:read permission.""")
# optional parameters
parser.add_argument(
"""--targets""",
default="""DeprecationWarning,UserWarning,FutureWarning""",
type=list_str,
help="""Comma-separated list of target warning(s) which we want to extract.""",
)
parser.add_argument(
"""--from_gh""",
action="""store_true""",
help="""If running from a GitHub action workflow and collecting warnings from its artifacts.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.from_gh
if from_gh:
# The artifacts have to be downloaded using `actions/download-artifact@v3`
pass
else:
os.makedirs(args.output_dir, exist_ok=True)
# get download links
__lowerCamelCase = get_artifacts_links(args.workflow_run_id, token=args.token)
with open(os.path.join(args.output_dir, """artifacts.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(artifacts, fp, ensure_ascii=False, indent=4)
# download artifacts
for idx, (name, url) in enumerate(artifacts.items()):
print(name)
print(url)
print("""=""" * 80)
download_artifact(name, url, args.output_dir, args.token)
# Be gentle to GitHub
time.sleep(1)
# extract warnings from artifacts
__lowerCamelCase = extract_warnings(args.output_dir, args.targets)
__lowerCamelCase = sorted(selected_warnings)
with open(os.path.join(args.output_dir, """selected_warnings.json"""), """w""", encoding="""UTF-8""") as fp:
json.dump(selected_warnings, fp, ensure_ascii=False, indent=4)
| 59
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from .tokenization_lxmert import LxmertTokenizer
__lowerCamelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {
"""unc-nlp/lxmert-base-uncased""": """https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""unc-nlp/lxmert-base-uncased""": (
"""https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/tokenizer.json"""
),
},
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": 5_12,
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": {"""do_lower_case""": True},
}
class UpperCAmelCase ( A_ ):
A__ : Any = VOCAB_FILES_NAMES
A__ : Optional[int] = PRETRAINED_VOCAB_FILES_MAP
A__ : Tuple = PRETRAINED_INIT_CONFIGURATION
A__ : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : List[Any] = LxmertTokenizer
def __init__(self : Dict , snake_case__ : Tuple=None , snake_case__ : Optional[Any]=None , snake_case__ : Optional[Any]=True , snake_case__ : Tuple="[UNK]" , snake_case__ : Optional[Any]="[SEP]" , snake_case__ : Optional[Any]="[PAD]" , snake_case__ : List[Any]="[CLS]" , snake_case__ : Tuple="[MASK]" , snake_case__ : Dict=True , snake_case__ : Union[str, Any]=None , **snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
snake_case : int = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , snake_case__ ) != do_lower_case
or normalizer_state.get("strip_accents" , snake_case__ ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , snake_case__ ) != tokenize_chinese_chars
):
snake_case : Union[str, Any] = getattr(snake_case__ , normalizer_state.pop("type" ) )
snake_case : str = do_lower_case
snake_case : List[Any] = strip_accents
snake_case : Optional[int] = tokenize_chinese_chars
snake_case : int = normalizer_class(**snake_case__ )
snake_case : Optional[Any] = do_lower_case
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Optional[Any] , snake_case__ : Dict=None ) -> Any:
'''simple docstring'''
snake_case : Optional[int] = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Optional[Any] = [self.sep_token_id]
snake_case : 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 ) * [0] + len(token_ids_a + sep ) * [1]
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
snake_case : List[Any] = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 59
| 1
|
import warnings
from collections import OrderedDict
from typing import Any, Mapping, Optional
from ... import PreTrainedTokenizer
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig, OnnxConfigWithPast, OnnxSeqaSeqConfigWithPast
from ...onnx.utils import compute_effective_axis_dimension
from ...utils import TensorType, is_torch_available, logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""facebook/bart-large""": """https://huggingface.co/facebook/bart-large/resolve/main/config.json""",
# See all BART models at https://huggingface.co/models?filter=bart
}
class UpperCAmelCase ( A_ ):
A__ : List[str] = "bart"
A__ : str = ["past_key_values"]
A__ : Dict = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__(self : List[Any] , snake_case__ : Optional[Any]=5_02_65 , snake_case__ : Dict=10_24 , snake_case__ : List[str]=12 , snake_case__ : int=40_96 , snake_case__ : int=16 , snake_case__ : Optional[int]=12 , snake_case__ : Any=40_96 , snake_case__ : int=16 , snake_case__ : str=0.0 , snake_case__ : List[str]=0.0 , snake_case__ : Optional[int]="gelu" , snake_case__ : Any=10_24 , snake_case__ : Union[str, Any]=0.1 , snake_case__ : Dict=0.0 , snake_case__ : int=0.0 , snake_case__ : int=0.02 , snake_case__ : int=0.0 , snake_case__ : List[str]=False , snake_case__ : Tuple=True , snake_case__ : str=3 , snake_case__ : Any=1 , snake_case__ : Optional[Any]=0 , snake_case__ : int=2 , snake_case__ : Tuple=True , snake_case__ : List[Any]=2 , snake_case__ : Optional[Any]=2 , **snake_case__ : int , ) -> Optional[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = vocab_size
snake_case : Tuple = max_position_embeddings
snake_case : List[Any] = d_model
snake_case : Optional[Any] = encoder_ffn_dim
snake_case : Dict = encoder_layers
snake_case : str = encoder_attention_heads
snake_case : str = decoder_ffn_dim
snake_case : List[Any] = decoder_layers
snake_case : Optional[Any] = decoder_attention_heads
snake_case : Tuple = dropout
snake_case : Any = attention_dropout
snake_case : List[Any] = activation_dropout
snake_case : Optional[Any] = activation_function
snake_case : List[str] = init_std
snake_case : Optional[Any] = encoder_layerdrop
snake_case : Any = decoder_layerdrop
snake_case : Any = classifier_dropout
snake_case : Dict = use_cache
snake_case : int = encoder_layers
snake_case : Optional[Any] = scale_embedding # scale factor will be sqrt(d_model) if True
super().__init__(
num_labels=snake_case__ , pad_token_id=snake_case__ , bos_token_id=snake_case__ , eos_token_id=snake_case__ , is_encoder_decoder=snake_case__ , decoder_start_token_id=snake_case__ , forced_eos_token_id=snake_case__ , **snake_case__ , )
# ensure backward compatibility for BART CNN models
if self.forced_bos_token_id is None and kwargs.get("force_bos_token_to_be_generated" , snake_case__ ):
snake_case : Optional[Any] = self.bos_token_id
warnings.warn(
f"""Please make sure the config includes `forced_bos_token_id={self.bos_token_id}` in future versions. """
"The config can simply be saved and uploaded again to be fixed." )
class UpperCAmelCase ( A_ ):
@property
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
if self.task in ["default", "seq2seq-lm"]:
snake_case : List[Any] = OrderedDict(
[
("input_ids", {0: "batch", 1: "encoder_sequence"}),
("attention_mask", {0: "batch", 1: "encoder_sequence"}),
] )
if self.use_past:
snake_case : int = {0: "batch"}
snake_case : Dict = {0: "batch", 1: "past_decoder_sequence + sequence"}
else:
snake_case : Tuple = {0: "batch", 1: "decoder_sequence"}
snake_case : str = {0: "batch", 1: "decoder_sequence"}
if self.use_past:
self.fill_with_past_key_values_(snake_case__ , direction="inputs" )
elif self.task == "causal-lm":
# TODO: figure this case out.
snake_case : int = OrderedDict(
[
("input_ids", {0: "batch", 1: "encoder_sequence"}),
("attention_mask", {0: "batch", 1: "encoder_sequence"}),
] )
if self.use_past:
snake_case , snake_case : Dict = self.num_layers
for i in range(snake_case__ ):
snake_case : List[Any] = {0: "batch", 2: "past_sequence + sequence"}
snake_case : Optional[int] = {0: "batch", 2: "past_sequence + sequence"}
else:
snake_case : Union[str, Any] = OrderedDict(
[
("input_ids", {0: "batch", 1: "encoder_sequence"}),
("attention_mask", {0: "batch", 1: "encoder_sequence"}),
("decoder_input_ids", {0: "batch", 1: "decoder_sequence"}),
("decoder_attention_mask", {0: "batch", 1: "decoder_sequence"}),
] )
return common_inputs
@property
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
if self.task in ["default", "seq2seq-lm"]:
snake_case : Optional[Any] = super().outputs
else:
snake_case : List[str] = super(snake_case__ , self ).outputs
if self.use_past:
snake_case , snake_case : int = self.num_layers
for i in range(snake_case__ ):
snake_case : Dict = {0: "batch", 2: "past_sequence + sequence"}
snake_case : int = {0: "batch", 2: "past_sequence + sequence"}
return common_outputs
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : PreTrainedTokenizer , snake_case__ : int = -1 , snake_case__ : int = -1 , snake_case__ : bool = False , snake_case__ : Optional[TensorType] = None , ) -> Mapping[str, Any]:
'''simple docstring'''
snake_case : Optional[int] = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# Generate decoder inputs
snake_case : Any = seq_length if not self.use_past else 1
snake_case : str = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ )
snake_case : Optional[int] = {f"""decoder_{name}""": tensor for name, tensor in decoder_inputs.items()}
snake_case : int = dict(**snake_case__ , **snake_case__ )
if self.use_past:
if not is_torch_available():
raise ValueError("Cannot generate dummy past_keys inputs without PyTorch installed." )
else:
import torch
snake_case , snake_case : Optional[int] = common_inputs["input_ids"].shape
snake_case : Any = common_inputs["decoder_input_ids"].shape[1]
snake_case , snake_case : Tuple = self.num_attention_heads
snake_case : Union[str, Any] = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case : Any = decoder_seq_length + 3
snake_case : List[Any] = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
snake_case : Tuple = torch.cat(
[common_inputs["decoder_attention_mask"], torch.ones(snake_case__ , snake_case__ )] , dim=1 )
snake_case : Union[str, Any] = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
snake_case , snake_case : Optional[int] = self.num_layers
snake_case : Optional[int] = min(snake_case__ , snake_case__ )
snake_case : int = max(snake_case__ , snake_case__ ) - min_num_layers
snake_case : Optional[Any] = "encoder" if num_encoder_layers > num_decoder_layers else "decoder"
for _ in range(snake_case__ ):
common_inputs["past_key_values"].append(
(
torch.zeros(snake_case__ ),
torch.zeros(snake_case__ ),
torch.zeros(snake_case__ ),
torch.zeros(snake_case__ ),
) )
# TODO: test this.
snake_case : Union[str, Any] = encoder_shape if remaining_side_name == "encoder" else decoder_shape
for _ in range(snake_case__ , snake_case__ ):
common_inputs["past_key_values"].append((torch.zeros(snake_case__ ), torch.zeros(snake_case__ )) )
return common_inputs
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : PreTrainedTokenizer , snake_case__ : int = -1 , snake_case__ : int = -1 , snake_case__ : bool = False , snake_case__ : Optional[TensorType] = None , ) -> Mapping[str, Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ )
if self.use_past:
if not is_torch_available():
raise ValueError("Cannot generate dummy past_keys inputs without PyTorch installed." )
else:
import torch
snake_case , snake_case : Optional[int] = common_inputs["input_ids"].shape
# Not using the same length for past_key_values
snake_case : Optional[Any] = seqlen + 2
snake_case , snake_case : Union[str, Any] = self.num_layers
snake_case , snake_case : Tuple = self.num_attention_heads
snake_case : Tuple = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
snake_case : Any = common_inputs["attention_mask"].dtype
snake_case : int = torch.cat(
[common_inputs["attention_mask"], torch.ones(snake_case__ , snake_case__ , dtype=snake_case__ )] , dim=1 )
snake_case : Dict = [
(torch.zeros(snake_case__ ), torch.zeros(snake_case__ )) for _ in range(snake_case__ )
]
return common_inputs
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : PreTrainedTokenizer , snake_case__ : int = -1 , snake_case__ : int = -1 , snake_case__ : bool = False , snake_case__ : Optional[TensorType] = None , ) -> Mapping[str, Any]:
'''simple docstring'''
snake_case : int = compute_effective_axis_dimension(
snake_case__ , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 )
# If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX
snake_case : Optional[Any] = tokenizer.num_special_tokens_to_add(snake_case__ )
snake_case : int = compute_effective_axis_dimension(
snake_case__ , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=snake_case__ )
# Generate dummy inputs according to compute batch and sequence
snake_case : Dict = [" ".join([tokenizer.unk_token] ) * seq_length] * batch_size
snake_case : int = dict(tokenizer(snake_case__ , return_tensors=snake_case__ ) )
return common_inputs
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : PreTrainedTokenizer , snake_case__ : int = -1 , snake_case__ : int = -1 , snake_case__ : bool = False , snake_case__ : Optional[TensorType] = None , ) -> Mapping[str, Any]:
'''simple docstring'''
if self.task in ["default", "seq2seq-lm"]:
snake_case : int = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
snake_case__ , batch_size=snake_case__ , seq_length=snake_case__ , is_pair=snake_case__ , framework=snake_case__ )
elif self.task == "causal-lm":
snake_case : str = self._generate_dummy_inputs_for_causal_lm(
snake_case__ , batch_size=snake_case__ , seq_length=snake_case__ , is_pair=snake_case__ , framework=snake_case__ )
else:
snake_case : Tuple = self._generate_dummy_inputs_for_sequence_classification_and_question_answering(
snake_case__ , batch_size=snake_case__ , seq_length=snake_case__ , is_pair=snake_case__ , framework=snake_case__ )
return common_inputs
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : Optional[Any] , snake_case__ : Optional[int] ) -> List[Any]:
'''simple docstring'''
if self.task in ["default", "seq2seq-lm"]:
snake_case : str = super()._flatten_past_key_values_(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
else:
snake_case : Tuple = super(snake_case__ , self )._flatten_past_key_values_(
snake_case__ , snake_case__ , snake_case__ , snake_case__ )
| 59
|
import torch
from diffusers import DDIMParallelScheduler
from .test_schedulers import SchedulerCommonTest
class UpperCAmelCase ( A_ ):
A__ : Dict = (DDIMParallelScheduler,)
A__ : Tuple = (("eta", 0.0), ("num_inference_steps", 50))
def _SCREAMING_SNAKE_CASE (self : Tuple , **snake_case__ : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case : Any = {
"num_train_timesteps": 10_00,
"beta_start": 0.0001,
"beta_end": 0.02,
"beta_schedule": "linear",
"clip_sample": True,
}
config.update(**snake_case__ )
return config
def _SCREAMING_SNAKE_CASE (self : Dict , **snake_case__ : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = self.scheduler_classes[0]
snake_case : Any = self.get_scheduler_config(**snake_case__ )
snake_case : Any = scheduler_class(**snake_case__ )
snake_case , snake_case : Union[str, Any] = 10, 0.0
snake_case : List[Any] = self.dummy_model()
snake_case : Any = self.dummy_sample_deter
scheduler.set_timesteps(snake_case__ )
for t in scheduler.timesteps:
snake_case : Optional[int] = model(snake_case__ , snake_case__ )
snake_case : List[str] = scheduler.step(snake_case__ , snake_case__ , snake_case__ , snake_case__ ).prev_sample
return sample
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> str:
'''simple docstring'''
for timesteps in [1_00, 5_00, 10_00]:
self.check_over_configs(num_train_timesteps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=snake_case__ )
snake_case : Optional[int] = self.scheduler_classes[0]
snake_case : Optional[int] = self.get_scheduler_config(steps_offset=1 )
snake_case : Union[str, Any] = scheduler_class(**snake_case__ )
scheduler.set_timesteps(5 )
assert torch.equal(scheduler.timesteps , torch.LongTensor([8_01, 6_01, 4_01, 2_01, 1] ) )
def _SCREAMING_SNAKE_CASE (self : int ) -> Tuple:
'''simple docstring'''
for beta_start, beta_end in zip([0.0001, 0.001, 0.01, 0.1] , [0.002, 0.02, 0.2, 2] ):
self.check_over_configs(beta_start=snake_case__ , beta_end=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
for timestep_spacing in ["trailing", "leading"]:
self.check_over_configs(timestep_spacing=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[Any]:
'''simple docstring'''
for rescale_betas_zero_snr in [True, False]:
self.check_over_configs(rescale_betas_zero_snr=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
self.check_over_configs(thresholding=snake_case__ )
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(
thresholding=snake_case__ , prediction_type=snake_case__ , sample_max_value=snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
for t in [1, 10, 49]:
self.check_over_forward(time_step=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Any:
'''simple docstring'''
for t, num_inference_steps in zip([1, 10, 50] , [10, 50, 5_00] ):
self.check_over_forward(time_step=snake_case__ , num_inference_steps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Optional[Any]:
'''simple docstring'''
for t, eta in zip([1, 10, 49] , [0.0, 0.5, 1.0] ):
self.check_over_forward(time_step=snake_case__ , eta=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Dict = self.scheduler_classes[0]
snake_case : Tuple = self.get_scheduler_config()
snake_case : Dict = scheduler_class(**snake_case__ )
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_20 , 4_00 ) - 0.14771 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_80 , 9_60 ) - 0.32460 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_87 , 4_86 ) - 0.00979 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_99 , 9_98 ) - 0.02 ) ) < 1e-5
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
snake_case : Union[str, Any] = self.scheduler_classes[0]
snake_case : List[Any] = self.get_scheduler_config()
snake_case : int = scheduler_class(**snake_case__ )
snake_case , snake_case : Any = 10, 0.0
scheduler.set_timesteps(snake_case__ )
snake_case : Optional[Any] = self.dummy_model()
snake_case : str = self.dummy_sample_deter
snake_case : Dict = self.dummy_sample_deter + 0.1
snake_case : Dict = self.dummy_sample_deter - 0.1
snake_case : Optional[Any] = samplea.shape[0]
snake_case : str = torch.stack([samplea, samplea, samplea] , dim=0 )
snake_case : Tuple = torch.arange(snake_case__ )[0:3, None].repeat(1 , snake_case__ )
snake_case : Tuple = model(samples.flatten(0 , 1 ) , timesteps.flatten(0 , 1 ) )
snake_case : List[str] = scheduler.batch_step_no_noise(snake_case__ , timesteps.flatten(0 , 1 ) , samples.flatten(0 , 1 ) , snake_case__ )
snake_case : Dict = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 1147.7904 ) < 1e-2
assert abs(result_mean.item() - 0.4982 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[Any] = self.full_loop()
snake_case : Optional[Any] = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 172.0067 ) < 1e-2
assert abs(result_mean.item() - 0.223967 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : str ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(prediction_type="v_prediction" )
snake_case : int = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[int] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 52.5302 ) < 1e-2
assert abs(result_mean.item() - 0.0684 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : str = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.8295 ) < 1e-2
assert abs(result_mean.item() - 0.1951 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : Tuple = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.0784 ) < 1e-2
assert abs(result_mean.item() - 0.1941 ) < 1e-3
| 59
| 1
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__lowerCamelCase = {
"""configuration_biogpt""": ["""BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP""", """BioGptConfig"""],
"""tokenization_biogpt""": ["""BioGptTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""BioGptForCausalLM""",
"""BioGptForTokenClassification""",
"""BioGptForSequenceClassification""",
"""BioGptModel""",
"""BioGptPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_biogpt import BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, BioGptConfig
from .tokenization_biogpt import BioGptTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_biogpt import (
BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
BioGptForCausalLM,
BioGptForSequenceClassification,
BioGptForTokenClassification,
BioGptModel,
BioGptPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
|
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )]
snake_case : int = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1 or len(__lowerCamelCase ) <= key:
return input_string
for position, character in enumerate(__lowerCamelCase ):
snake_case : Any = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append(__lowerCamelCase )
snake_case : List[str] = ["".join(__lowerCamelCase ) for row in temp_grid]
snake_case : Tuple = "".join(__lowerCamelCase )
return output_string
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : Dict = []
snake_case : Union[str, Any] = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1:
return input_string
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )] # generates template
for position in range(len(__lowerCamelCase ) ):
snake_case : List[str] = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append("*" )
snake_case : Tuple = 0
for row in temp_grid: # fills in the characters
snake_case : Union[str, Any] = input_string[counter : counter + len(__lowerCamelCase )]
grid.append(list(__lowerCamelCase ) )
counter += len(__lowerCamelCase )
snake_case : str = "" # reads as zigzag
for position in range(len(__lowerCamelCase ) ):
snake_case : Optional[int] = position % (lowest * 2) # puts it in bounds
snake_case : Tuple = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
output_string += grid[num][0]
grid[num].pop(0 )
return output_string
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : Tuple = {}
for key_guess in range(1 , len(__lowerCamelCase ) ): # tries every key
snake_case : Any = decrypt(__lowerCamelCase , __lowerCamelCase )
return results
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
from __future__ import annotations
import unittest
import numpy as np
from transformers import OPTConfig, is_tf_available
from transformers.testing_utils import require_sentencepiece, require_tf, slow
from ...test_configuration_common import ConfigTester
from ...test_modeling_tf_common import TFModelTesterMixin, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_tf_available():
import tensorflow as tf
from transformers import GPTaTokenizer, TFOPTForCausalLM, TFOPTModel
def UpperCamelCase ( __lowerCamelCase : Tuple , __lowerCamelCase : Optional[int] , __lowerCamelCase : int=None , __lowerCamelCase : Optional[Any]=None ):
if attention_mask is None:
snake_case : Optional[Any] = tf.cast(tf.math.not_equal(__lowerCamelCase , config.pad_token_id ) , tf.inta )
return {"input_ids": input_ids, "attention_mask": attention_mask}
@require_tf
class UpperCAmelCase :
A__ : List[str] = OPTConfig
A__ : Optional[int] = {}
A__ : List[Any] = "gelu"
def __init__(self : int , snake_case__ : Dict , snake_case__ : Tuple=13 , snake_case__ : Optional[Any]=7 , snake_case__ : Union[str, Any]=True , snake_case__ : Union[str, Any]=False , snake_case__ : Tuple=99 , snake_case__ : Any=16 , snake_case__ : Union[str, Any]=2 , snake_case__ : Union[str, Any]=4 , snake_case__ : List[Any]=4 , snake_case__ : Union[str, Any]="gelu" , snake_case__ : Union[str, Any]=0.1 , snake_case__ : List[str]=0.1 , snake_case__ : int=20 , snake_case__ : Optional[int]=2 , snake_case__ : Optional[Any]=1 , snake_case__ : List[Any]=0 , snake_case__ : str=16 , snake_case__ : Tuple=16 , ) -> Any:
'''simple docstring'''
snake_case : Optional[int] = parent
snake_case : int = batch_size
snake_case : List[str] = seq_length
snake_case : str = is_training
snake_case : List[Any] = use_labels
snake_case : Dict = vocab_size
snake_case : List[Any] = hidden_size
snake_case : Union[str, Any] = num_hidden_layers
snake_case : str = num_attention_heads
snake_case : Optional[int] = intermediate_size
snake_case : str = hidden_act
snake_case : int = hidden_dropout_prob
snake_case : Optional[Any] = attention_probs_dropout_prob
snake_case : Optional[int] = max_position_embeddings
snake_case : List[str] = eos_token_id
snake_case : int = pad_token_id
snake_case : Optional[Any] = bos_token_id
snake_case : List[str] = embed_dim
snake_case : List[str] = word_embed_proj_dim
snake_case : Optional[Any] = False
def _SCREAMING_SNAKE_CASE (self : int ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = ids_tensor([self.batch_size, self.seq_length - 1] , self.vocab_size )
snake_case : Union[str, Any] = tf.expand_dims(tf.constant([self.eos_token_id] * self.batch_size ) , 1 )
snake_case : List[Any] = tf.concat([input_ids, eos_tensor] , axis=1 )
snake_case : Dict = self.config_cls(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , ffn_dim=self.intermediate_size , dropout=self.hidden_dropout_prob , attention_dropout=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , eos_token_id=self.eos_token_id , bos_token_id=self.bos_token_id , pad_token_id=self.pad_token_id , embed_dim=self.embed_dim , word_embed_proj_dim=self.word_embed_proj_dim , is_encoder_decoder=snake_case__ , **self.config_updates , )
snake_case : Dict = prepare_opt_inputs_dict(snake_case__ , snake_case__ )
return config, inputs_dict
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : List[str] , snake_case__ : List[str] ) -> Dict:
'''simple docstring'''
snake_case : Any = TFOPTModel(config=snake_case__ )
snake_case : Optional[int] = inputs_dict["input_ids"]
snake_case : Union[str, Any] = input_ids[:1, :]
snake_case : List[str] = inputs_dict["attention_mask"][:1, :]
snake_case : Tuple = 1
# first forward pass
snake_case : int = model(snake_case__ , attention_mask=snake_case__ , use_cache=snake_case__ )
snake_case , snake_case : Dict = outputs.to_tuple()
# create hypothetical next token and extent to next_input_ids
snake_case : Dict = ids_tensor((self.batch_size, 3) , config.vocab_size )
snake_case : Any = tf.cast(ids_tensor((self.batch_size, 3) , 2 ) , tf.inta )
# append to next input_ids and
snake_case : Optional[Any] = tf.concat([input_ids, next_tokens] , axis=-1 )
snake_case : Union[str, Any] = tf.concat([attention_mask, next_attn_mask] , axis=-1 )
snake_case : int = model(snake_case__ , attention_mask=snake_case__ )[0]
snake_case : Union[str, Any] = model(snake_case__ , attention_mask=snake_case__ , past_key_values=snake_case__ )[0]
self.parent.assertEqual(next_tokens.shape[1] , output_from_past.shape[1] )
# select random slice
snake_case : Dict = int(ids_tensor((1,) , output_from_past.shape[-1] ) )
snake_case : List[str] = output_from_no_past[:, -3:, random_slice_idx]
snake_case : List[Any] = output_from_past[:, :, random_slice_idx]
# test that outputs are equal for slice
tf.debugging.assert_near(snake_case__ , snake_case__ , rtol=1e-3 )
@require_tf
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : Optional[int] = (TFOPTModel, TFOPTForCausalLM) if is_tf_available() else ()
A__ : Any = (TFOPTForCausalLM,) if is_tf_available() else ()
A__ : str = (
{"feature-extraction": TFOPTModel, "text-generation": TFOPTForCausalLM} if is_tf_available() else {}
)
A__ : int = False
A__ : Optional[int] = False
A__ : Optional[int] = False
A__ : List[Any] = 10
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
snake_case : Dict = TFOPTModelTester(self )
snake_case : Any = ConfigTester(self , config_class=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> List[Any]:
'''simple docstring'''
self.config_tester.run_common_tests()
def _SCREAMING_SNAKE_CASE (self : str ) -> List[str]:
'''simple docstring'''
snake_case : Tuple = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> str:
'''simple docstring'''
snake_case , snake_case : int = self.model_tester.prepare_config_and_inputs_for_common()
def _get_word_embedding_weight(snake_case__ : str , snake_case__ : Union[str, Any] ):
if hasattr(snake_case__ , "weight" ):
return embedding_layer.weight
else:
# Here we build the word embeddings weights if not exists.
# And then we retry to get the attribute once built.
model.build()
if hasattr(snake_case__ , "weight" ):
return embedding_layer.weight
else:
return None
for model_class in self.all_model_classes:
for size in [config.vocab_size - 10, config.vocab_size + 10]:
# build the embeddings
snake_case : List[str] = model_class(config=snake_case__ )
snake_case : int = _get_word_embedding_weight(snake_case__ , model.get_input_embeddings() )
snake_case : List[str] = _get_word_embedding_weight(snake_case__ , model.get_output_embeddings() )
# reshape the embeddings
model.resize_token_embeddings(snake_case__ )
snake_case : Tuple = _get_word_embedding_weight(snake_case__ , model.get_input_embeddings() )
snake_case : Tuple = _get_word_embedding_weight(snake_case__ , model.get_output_embeddings() )
# check that the resized embeddings size matches the desired size.
snake_case : List[str] = size if size is not None else config.vocab_size
self.assertEqual(new_input_embeddings.shape[0] , snake_case__ )
# check that weights remain the same after resizing
snake_case : Optional[int] = True
for pa, pa in zip(old_input_embeddings.value() , new_input_embeddings.value() ):
if tf.math.reduce_sum(tf.math.abs(pa - pa ) ) > 0:
snake_case : int = False
self.assertTrue(snake_case__ )
if old_output_embeddings is not None and new_output_embeddings is not None:
self.assertEqual(new_output_embeddings.shape[0] , snake_case__ )
snake_case : Optional[Any] = True
for pa, pa in zip(old_output_embeddings.value() , new_output_embeddings.value() ):
if tf.math.reduce_sum(tf.math.abs(pa - pa ) ) > 0:
snake_case : Union[str, Any] = False
self.assertTrue(snake_case__ )
def UpperCamelCase ( __lowerCamelCase : List[str] ):
return tf.constant(__lowerCamelCase , dtype=tf.intaa )
@require_tf
class UpperCAmelCase ( unittest.TestCase ):
A__ : str = 99
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = tf.ones((4, 1) , dtype=tf.intaa ) * 2
snake_case : Dict = tf.concat([ids_tensor((4, 6) , self.vocab_size - 3 ) + 3, eos_column_vector] , axis=1 )
snake_case : Optional[int] = input_ids.shape[0]
snake_case : List[str] = OPTConfig(
vocab_size=self.vocab_size , hidden_size=24 , num_hidden_layers=2 , num_attention_heads=2 , ffn_dim=32 , max_position_embeddings=48 , eos_token_id=2 , pad_token_id=1 , bos_token_id=0 , )
return config, input_ids, batch_size
@require_sentencepiece
@require_tf
class UpperCAmelCase ( unittest.TestCase ):
@slow
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = TFOPTModel.from_pretrained("facebook/opt-350m" )
snake_case : int = _long_tensor([[0, 3_14_14, 2_32, 3_28, 7_40, 11_40, 1_26_95, 69, 4_60_78, 15_88, 2]] )
snake_case : List[Any] = tf.not_equal(snake_case__ , model.config.pad_token_id )
with tf.GradientTape():
snake_case : List[Any] = model(input_ids=snake_case__ , attention_mask=snake_case__ ).last_hidden_state
snake_case : List[Any] = (1, 11, 5_12)
self.assertEqual(output.shape , snake_case__ )
snake_case : List[Any] = tf.constant(
[[-0.2873, -1.9218, -0.3033], [-1.2710, -0.1338, -0.1902], [0.4095, 0.1214, -1.3121]] )
self.assertTrue(np.allclose(output[:, :3, :3] , snake_case__ , atol=4e-3 ) )
snake_case : Optional[int] = tf.function(snake_case__ , jit_compile=snake_case__ )
snake_case : Optional[int] = xla_generate(snake_case__ , snake_case__ )[0]
self.assertTrue(np.allclose(output[:, :3, :3] , snake_case__ , atol=4e-2 ) )
@require_tf
@slow
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
super().setUp()
snake_case : Optional[int] = "facebook/opt-350m"
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : Any = TFOPTForCausalLM.from_pretrained(self.path_model )
snake_case : Tuple = GPTaTokenizer.from_pretrained(self.path_model )
snake_case : Optional[int] = [
"Today is a beautiful day and I want to",
"In the city of",
"Paris is the capital of France and",
"Computers and mobile phones have taken",
]
# verify that prompt without BOS token is identical to Metaseq -> add_special_tokens=False
snake_case : str = tokenizer(snake_case__ , return_tensors="tf" , padding=snake_case__ , add_special_tokens=snake_case__ )
snake_case : Dict = tf.math.reduce_mean(model(inputs.input_ids , attention_mask=inputs.attention_mask )[0] , axis=-1 )
snake_case : str = tf.constant(
[
[1.3851, -13.8923, -10.5229, -10.7533, -0.2309, -10.2384, -0.5365, -9.0947, -5.1670],
[-4.7073, -10.6276, -3.9415, -21.5242, -0.2822, -0.2822, -0.2822, -0.2822, -0.2822],
[0.6247, -3.4229, -8.9179, -1.4297, -14.1650, 1.4146, -9.0218, -0.2703, -0.2703],
[6.4783, -1.9913, -10.7926, -2.3336, 1.5092, -0.9974, -6.8213, 1.3477, 1.3477],
] )
self.assertTrue(np.allclose(snake_case__ , snake_case__ , atol=1e-4 ) )
snake_case : Optional[int] = tf.function(snake_case__ , jit_compile=snake_case__ )
snake_case : Dict = tf.math.reduce_mean(xla_generate(inputs.input_ids , attention_mask=inputs.attention_mask )[0] , axis=-1 )
self.assertTrue(np.allclose(snake_case__ , snake_case__ , atol=1e-4 ) )
@require_tf
@slow
class UpperCAmelCase ( unittest.TestCase ):
@property
def _SCREAMING_SNAKE_CASE (self : Any ) -> Dict:
'''simple docstring'''
return [
"Today is a beautiful day and I want",
"In the city of",
"Paris is the capital of France and",
"Computers and mobile phones have taken",
]
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Any:
'''simple docstring'''
snake_case : Any = "facebook/opt-125m"
snake_case : List[Any] = [
"Today is a beautiful day and I want to",
"In the city of New York, the city",
"Paris is the capital of France and the capital",
"Computers and mobile phones have taken over the",
]
snake_case : str = []
snake_case : Any = GPTaTokenizer.from_pretrained(snake_case__ )
snake_case : Optional[Any] = TFOPTForCausalLM.from_pretrained(snake_case__ )
for prompt in self.prompts:
snake_case : str = tokenizer(snake_case__ , return_tensors="tf" ).input_ids
snake_case : Optional[Any] = model.generate(snake_case__ , max_length=10 )
snake_case : Any = tokenizer.batch_decode(snake_case__ , skip_special_tokens=snake_case__ )
predicted_outputs += generated_string
self.assertListEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> str:
'''simple docstring'''
snake_case : Tuple = "facebook/opt-350m"
snake_case : Union[str, Any] = GPTaTokenizer.from_pretrained(snake_case__ )
snake_case : List[Any] = TFOPTForCausalLM.from_pretrained(snake_case__ )
snake_case : Tuple = "left"
# use different length sentences to test batching
snake_case : List[Any] = [
"Hello, my dog is a little",
"Today, I",
]
snake_case : int = tokenizer(snake_case__ , return_tensors="tf" , padding=snake_case__ )
snake_case : int = inputs["input_ids"]
snake_case : int = model.generate(input_ids=snake_case__ , attention_mask=inputs["attention_mask"] )
snake_case : List[Any] = tokenizer(sentences[0] , return_tensors="tf" ).input_ids
snake_case : Dict = model.generate(input_ids=snake_case__ )
snake_case : Union[str, Any] = inputs_non_padded.shape[-1] - tf.math.reduce_sum(
tf.cast(inputs["attention_mask"][-1] , tf.intaa ) )
snake_case : Optional[int] = tokenizer(sentences[1] , return_tensors="tf" ).input_ids
snake_case : Any = model.generate(input_ids=snake_case__ , max_length=model.config.max_length - num_paddings )
snake_case : Tuple = tokenizer.batch_decode(snake_case__ , skip_special_tokens=snake_case__ )
snake_case : Dict = tokenizer.decode(output_non_padded[0] , skip_special_tokens=snake_case__ )
snake_case : Optional[Any] = tokenizer.decode(output_padded[0] , skip_special_tokens=snake_case__ )
snake_case : Any = [
"Hello, my dog is a little bit of a dork.\nI'm a little bit",
"Today, I was in the middle of a conversation with a friend about the",
]
self.assertListEqual(snake_case__ , snake_case__ )
self.assertListEqual(snake_case__ , [non_padded_sentence, padded_sentence] )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : Any = "facebook/opt-350m"
snake_case : Optional[int] = [
"Today is a beautiful day and I want to",
"In the city of San Francisco, the city",
"Paris is the capital of France and the capital",
"Computers and mobile phones have taken over the",
]
snake_case : int = []
snake_case : int = GPTaTokenizer.from_pretrained(snake_case__ )
snake_case : Any = TFOPTForCausalLM.from_pretrained(snake_case__ )
for prompt in self.prompts:
snake_case : Optional[Any] = tokenizer(snake_case__ , return_tensors="tf" ).input_ids
snake_case : int = model.generate(snake_case__ , max_length=10 )
snake_case : Union[str, Any] = tokenizer.batch_decode(snake_case__ , skip_special_tokens=snake_case__ )
predicted_outputs += generated_string
self.assertListEqual(snake_case__ , snake_case__ )
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
if is_sentencepiece_available():
from ..ta.tokenization_ta import TaTokenizer
else:
from ...utils.dummy_sentencepiece_objects import TaTokenizer
__lowerCamelCase = TaTokenizer
if is_tokenizers_available():
from ..ta.tokenization_ta_fast import TaTokenizerFast
else:
from ...utils.dummy_tokenizers_objects import TaTokenizerFast
__lowerCamelCase = TaTokenizerFast
__lowerCamelCase = {"""configuration_mt5""": ["""MT5Config""", """MT5OnnxConfig"""]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""MT5EncoderModel""",
"""MT5ForConditionalGeneration""",
"""MT5ForQuestionAnswering""",
"""MT5Model""",
"""MT5PreTrainedModel""",
"""MT5Stack""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""TFMT5EncoderModel""", """TFMT5ForConditionalGeneration""", """TFMT5Model"""]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""FlaxMT5EncoderModel""", """FlaxMT5ForConditionalGeneration""", """FlaxMT5Model"""]
if TYPE_CHECKING:
from .configuration_mta import MTaConfig, MTaOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mta import (
MTaEncoderModel,
MTaForConditionalGeneration,
MTaForQuestionAnswering,
MTaModel,
MTaPreTrainedModel,
MTaStack,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel
else:
import sys
__lowerCamelCase = _LazyModule(
__name__,
globals()["""__file__"""],
_import_structure,
extra_objects={"""MT5Tokenizer""": MTaTokenizer, """MT5TokenizerFast""": MTaTokenizerFast},
module_spec=__spec__,
)
| 59
| 1
|
from typing import Optional, Tuple
import jax
import jax.numpy as jnp
from flax import linen as nn
from flax.core.frozen_dict import FrozenDict
from transformers import CLIPConfig, FlaxPreTrainedModel
from transformers.models.clip.modeling_flax_clip import FlaxCLIPVisionModule
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Dict , __lowerCamelCase : Any=1E-12 ):
snake_case : str = jnp.divide(emb_a.T , jnp.clip(jnp.linalg.norm(__lowerCamelCase , axis=1 ) , a_min=__lowerCamelCase ) ).T
snake_case : str = jnp.divide(emb_a.T , jnp.clip(jnp.linalg.norm(__lowerCamelCase , axis=1 ) , a_min=__lowerCamelCase ) ).T
return jnp.matmul(__lowerCamelCase , norm_emb_a.T )
class UpperCAmelCase ( nn.Module ):
A__ : CLIPConfig
A__ : jnp.dtype = jnp.floataa
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
snake_case : List[Any] = FlaxCLIPVisionModule(self.config.vision_config )
snake_case : List[Any] = nn.Dense(self.config.projection_dim , use_bias=snake_case__ , dtype=self.dtype )
snake_case : Tuple = self.param("concept_embeds" , jax.nn.initializers.ones , (17, self.config.projection_dim) )
snake_case : Optional[int] = self.param(
"special_care_embeds" , jax.nn.initializers.ones , (3, self.config.projection_dim) )
snake_case : Any = self.param("concept_embeds_weights" , jax.nn.initializers.ones , (17,) )
snake_case : Any = self.param("special_care_embeds_weights" , jax.nn.initializers.ones , (3,) )
def __call__(self : Tuple , snake_case__ : Tuple ) -> Any:
'''simple docstring'''
snake_case : str = self.vision_model(snake_case__ )[1]
snake_case : int = self.visual_projection(snake_case__ )
snake_case : str = jax_cosine_distance(snake_case__ , self.special_care_embeds )
snake_case : Dict = jax_cosine_distance(snake_case__ , self.concept_embeds )
# increase this value to create a stronger `nfsw` filter
# at the cost of increasing the possibility of filtering benign image inputs
snake_case : List[str] = 0.0
snake_case : Optional[int] = special_cos_dist - self.special_care_embeds_weights[None, :] + adjustment
snake_case : int = jnp.round(snake_case__ , 3 )
snake_case : Optional[Any] = jnp.any(special_scores > 0 , axis=1 , keepdims=snake_case__ )
# Use a lower threshold if an image has any special care concept
snake_case : str = is_special_care * 0.01
snake_case : Any = cos_dist - self.concept_embeds_weights[None, :] + special_adjustment
snake_case : Optional[Any] = jnp.round(snake_case__ , 3 )
snake_case : int = jnp.any(concept_scores > 0 , axis=1 )
return has_nsfw_concepts
class UpperCAmelCase ( A_ ):
A__ : str = CLIPConfig
A__ : int = "clip_input"
A__ : List[str] = FlaxStableDiffusionSafetyCheckerModule
def __init__(self : str , snake_case__ : CLIPConfig , snake_case__ : Optional[Tuple] = None , snake_case__ : int = 0 , snake_case__ : jnp.dtype = jnp.floataa , snake_case__ : bool = True , **snake_case__ : Dict , ) -> Dict:
'''simple docstring'''
if input_shape is None:
snake_case : int = (1, 2_24, 2_24, 3)
snake_case : Union[str, Any] = self.module_class(config=snake_case__ , dtype=snake_case__ , **snake_case__ )
super().__init__(snake_case__ , snake_case__ , input_shape=snake_case__ , seed=snake_case__ , dtype=snake_case__ , _do_init=_do_init )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : jax.random.KeyArray , snake_case__ : Tuple , snake_case__ : FrozenDict = None ) -> FrozenDict:
'''simple docstring'''
snake_case : str = jax.random.normal(snake_case__ , snake_case__ )
snake_case , snake_case : int = jax.random.split(snake_case__ )
snake_case : Optional[int] = {"params": params_rng, "dropout": dropout_rng}
snake_case : Optional[int] = self.module.init(snake_case__ , snake_case__ )["params"]
return random_params
def __call__(self : Union[str, Any] , snake_case__ : str , snake_case__ : dict = None , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : int = jnp.transpose(snake_case__ , (0, 2, 3, 1) )
return self.module.apply(
{"params": params or self.params} , jnp.array(snake_case__ , dtype=jnp.floataa ) , rngs={} , )
| 59
|
import os
import shutil
from pathlib import Path
from typing import Optional, Union
import numpy as np
from huggingface_hub import hf_hub_download
from ..utils import ONNX_EXTERNAL_WEIGHTS_NAME, ONNX_WEIGHTS_NAME, is_onnx_available, logging
if is_onnx_available():
import onnxruntime as ort
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""tensor(bool)""": np.bool_,
"""tensor(int8)""": np.inta,
"""tensor(uint8)""": np.uinta,
"""tensor(int16)""": np.intaa,
"""tensor(uint16)""": np.uintaa,
"""tensor(int32)""": np.intaa,
"""tensor(uint32)""": np.uintaa,
"""tensor(int64)""": np.intaa,
"""tensor(uint64)""": np.uintaa,
"""tensor(float16)""": np.floataa,
"""tensor(float)""": np.floataa,
"""tensor(double)""": np.floataa,
}
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=None , **snake_case__ : Optional[Any] ) -> List[str]:
'''simple docstring'''
logger.info("`diffusers.OnnxRuntimeModel` is experimental and might change in the future." )
snake_case : Optional[Any] = model
snake_case : Dict = kwargs.get("model_save_dir" , snake_case__ )
snake_case : int = kwargs.get("latest_model_name" , snake_case__ )
def __call__(self : Tuple , **snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Union[str, Any] = {k: np.array(snake_case__ ) for k, v in kwargs.items()}
return self.model.run(snake_case__ , snake_case__ )
@staticmethod
def _SCREAMING_SNAKE_CASE (snake_case__ : Union[str, Path] , snake_case__ : Optional[int]=None , snake_case__ : Optional[int]=None ) -> Any:
'''simple docstring'''
if provider is None:
logger.info("No onnxruntime provider specified, using CPUExecutionProvider" )
snake_case : Optional[int] = "CPUExecutionProvider"
return ort.InferenceSession(snake_case__ , providers=[provider] , sess_options=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Union[str, Path] , snake_case__ : Optional[str] = None , **snake_case__ : Any ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = file_name if file_name is not None else ONNX_WEIGHTS_NAME
snake_case : Any = self.model_save_dir.joinpath(self.latest_model_name )
snake_case : str = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
snake_case : List[str] = self.model_save_dir.joinpath(snake_case__ )
if src_path.exists():
snake_case : Tuple = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, os.PathLike] , **snake_case__ : Optional[int] , ) -> str:
'''simple docstring'''
if os.path.isfile(snake_case__ ):
logger.error(f"""Provided path ({save_directory}) should be a directory, not a file""" )
return
os.makedirs(snake_case__ , exist_ok=snake_case__ )
# saving model weights/files
self._save_pretrained(snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Tuple , snake_case__ : Union[str, Path] , snake_case__ : Optional[Union[bool, str, None]] = None , snake_case__ : Optional[Union[str, None]] = None , snake_case__ : bool = False , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional["ort.SessionOptions"] = None , **snake_case__ : Tuple , ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(snake_case__ ):
snake_case : Any = OnnxRuntimeModel.load_model(
os.path.join(snake_case__ , snake_case__ ) , provider=snake_case__ , sess_options=snake_case__ )
snake_case : Union[str, Any] = Path(snake_case__ )
# load model from hub
else:
# download model
snake_case : Dict = hf_hub_download(
repo_id=snake_case__ , filename=snake_case__ , use_auth_token=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , )
snake_case : List[Any] = Path(snake_case__ ).parent
snake_case : Union[str, Any] = Path(snake_case__ ).name
snake_case : Dict = OnnxRuntimeModel.load_model(snake_case__ , provider=snake_case__ , sess_options=snake_case__ )
return cls(model=snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : Union[str, Path] , snake_case__ : bool = True , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , **snake_case__ : Dict , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = None
if len(str(snake_case__ ).split("@" ) ) == 2:
snake_case , snake_case : int = model_id.split("@" )
return cls._from_pretrained(
model_id=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , use_auth_token=snake_case__ , **snake_case__ , )
| 59
| 1
|
class UpperCAmelCase :
def __init__(self : Dict , snake_case__ : Optional[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Optional[int] = val
snake_case : Any = None
snake_case : Tuple = None
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : str ) -> Tuple:
'''simple docstring'''
if self.val:
if val < self.val:
if self.left is None:
snake_case : Any = Node(snake_case__ )
else:
self.left.insert(snake_case__ )
elif val > self.val:
if self.right is None:
snake_case : Tuple = Node(snake_case__ )
else:
self.right.insert(snake_case__ )
else:
snake_case : Optional[Any] = val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : int ):
# Recursive traversal
if root:
inorder(root.left , __lowerCamelCase )
res.append(root.val )
inorder(root.right , __lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : Tuple ):
# Build BST
if len(__lowerCamelCase ) == 0:
return arr
snake_case : Optional[int] = Node(arr[0] )
for i in range(1 , len(__lowerCamelCase ) ):
root.insert(arr[i] )
# Traverse BST in order.
snake_case : Optional[int] = []
inorder(__lowerCamelCase , __lowerCamelCase )
return res
if __name__ == "__main__":
print(tree_sort([10, 1, 3, 2, 9, 14, 13]))
| 59
|
import argparse
import json
from dataclasses import dataclass, field
from functools import partial
from pathlib import Path
from typing import Callable, Dict, List, Tuple
import timm
import torch
import torch.nn as nn
from classy_vision.models.regnet import RegNet, RegNetParams, RegNetYaagf, RegNetYaagf, RegNetYaaagf
from huggingface_hub import cached_download, hf_hub_url
from torch import Tensor
from vissl.models.model_helpers import get_trunk_forward_outputs
from transformers import AutoImageProcessor, RegNetConfig, RegNetForImageClassification, RegNetModel
from transformers.utils import logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger()
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : List[nn.Module] = field(default_factory=A_ )
A__ : list = field(default_factory=A_ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : Tensor , snake_case__ : Tensor ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = len(list(m.modules() ) ) == 1 or isinstance(snake_case__ , nn.Convad ) or isinstance(snake_case__ , nn.BatchNormad )
if has_not_submodules:
self.traced.append(snake_case__ )
def __call__(self : List[Any] , snake_case__ : Tensor ) -> List[Any]:
'''simple docstring'''
for m in self.module.modules():
self.handles.append(m.register_forward_hook(self._forward_hook ) )
self.module(snake_case__ )
[x.remove() for x in self.handles]
return self
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[int]:
'''simple docstring'''
return list(filter(lambda snake_case__ : len(list(x.state_dict().keys() ) ) > 0 , self.traced ) )
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : nn.Module
A__ : int = 1
A__ : List = field(default_factory=A_ )
A__ : List = field(default_factory=A_ )
A__ : bool = True
def __call__(self : List[Any] , snake_case__ : Tensor ) -> Any:
'''simple docstring'''
snake_case : str = Tracker(self.dest )(snake_case__ ).parametrized
snake_case : Optional[int] = Tracker(self.src )(snake_case__ ).parametrized
snake_case : List[str] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.src_skip , snake_case__ ) )
snake_case : Optional[Any] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.dest_skip , snake_case__ ) )
if len(snake_case__ ) != len(snake_case__ ) and self.raise_if_mismatch:
raise Exception(
f"""Numbers of operations are different. Source module has {len(snake_case__ )} operations while"""
f""" destination module has {len(snake_case__ )}.""" )
for dest_m, src_m in zip(snake_case__ , snake_case__ ):
dest_m.load_state_dict(src_m.state_dict() )
if self.verbose == 1:
print(f"""Transfered from={src_m} to={dest_m}""" )
class UpperCAmelCase ( nn.Module ):
def __init__(self : Tuple , snake_case__ : nn.Module ) -> Optional[Any]:
'''simple docstring'''
super().__init__()
snake_case : List[Tuple[str, nn.Module]] = []
# - get the stem
feature_blocks.append(("conv1", model.stem) )
# - get all the feature blocks
for k, v in model.trunk_output.named_children():
assert k.startswith("block" ), f"""Unexpected layer name {k}"""
snake_case : Union[str, Any] = len(snake_case__ ) + 1
feature_blocks.append((f"""res{block_index}""", v) )
snake_case : Optional[Any] = nn.ModuleDict(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Tensor ) -> Dict:
'''simple docstring'''
return get_trunk_forward_outputs(
snake_case__ , out_feat_keys=snake_case__ , feature_blocks=self._feature_blocks , )
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> str:
'''simple docstring'''
snake_case : List[Any] = x.split("-" )
return x_split[0] + x_split[1] + "_" + "".join(x_split[2:] )
def __getitem__(self : Optional[int] , snake_case__ : str ) -> Callable[[], Tuple[nn.Module, Dict]]:
'''simple docstring'''
if x not in self:
snake_case : Dict = self.convert_name_to_timm(snake_case__ )
snake_case : Union[str, Any] = partial(lambda: (timm.create_model(snake_case__ , pretrained=snake_case__ ).eval(), None) )
else:
snake_case : List[str] = super().__getitem__(snake_case__ )
return val
class UpperCAmelCase ( A_ ):
def __getitem__(self : Dict , snake_case__ : str ) -> Callable[[], nn.Module]:
'''simple docstring'''
if "seer" in x and "in1k" not in x:
snake_case : str = RegNetModel
else:
snake_case : Optional[Any] = RegNetForImageClassification
return val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Tuple[str, str]] ):
for from_key, to_key in keys:
snake_case : str = from_state_dict[from_key].clone()
print(f"""Copied key={from_key} to={to_key}""" )
return to_state_dict
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : RegNetConfig , __lowerCamelCase : Path , __lowerCamelCase : bool = True , ):
print(f"""Converting {name}...""" )
with torch.no_grad():
snake_case , snake_case : int = from_model_func()
snake_case : str = our_model_func(__lowerCamelCase ).eval()
snake_case : int = ModuleTransfer(src=__lowerCamelCase , dest=__lowerCamelCase , raise_if_mismatch=__lowerCamelCase )
snake_case : Dict = torch.randn((1, 3, 224, 224) )
module_transfer(__lowerCamelCase )
if from_state_dict is not None:
snake_case : str = []
# for seer - in1k finetuned we have to manually copy the head
if "seer" in name and "in1k" in name:
snake_case : Tuple = [("0.clf.0.weight", "classifier.1.weight"), ("0.clf.0.bias", "classifier.1.bias")]
snake_case : Optional[Any] = manually_copy_vissl_head(__lowerCamelCase , our_model.state_dict() , __lowerCamelCase )
our_model.load_state_dict(__lowerCamelCase )
snake_case : Any = our_model(__lowerCamelCase , output_hidden_states=__lowerCamelCase )
snake_case : Union[str, Any] = (
our_outputs.logits if isinstance(__lowerCamelCase , __lowerCamelCase ) else our_outputs.last_hidden_state
)
snake_case : Union[str, Any] = from_model(__lowerCamelCase )
snake_case : Dict = from_output[-1] if type(__lowerCamelCase ) is list else from_output
# now since I don't want to use any config files, vissl seer model doesn't actually have an head, so let's just check the last hidden state
if "seer" in name and "in1k" in name:
snake_case : Any = our_outputs.hidden_states[-1]
assert torch.allclose(__lowerCamelCase , __lowerCamelCase ), "The model logits don't match the original one."
if push_to_hub:
our_model.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add model" , use_temp_dir=__lowerCamelCase , )
snake_case : List[str] = 224 if "seer" not in name else 384
# we can use the convnext one
snake_case : int = AutoImageProcessor.from_pretrained("facebook/convnext-base-224-22k-1k" , size=__lowerCamelCase )
image_processor.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add image processor" , use_temp_dir=__lowerCamelCase , )
print(f"""Pushed {name}""" )
def UpperCamelCase ( __lowerCamelCase : Path , __lowerCamelCase : str = None , __lowerCamelCase : bool = True ):
snake_case : Union[str, Any] = "imagenet-1k-id2label.json"
snake_case : List[str] = 1000
snake_case : List[str] = (1, num_labels)
snake_case : Any = "huggingface/label-files"
snake_case : List[str] = num_labels
snake_case : Optional[Any] = json.load(open(cached_download(hf_hub_url(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) ) , "r" ) )
snake_case : List[Any] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : str = idalabel
snake_case : List[Any] = {v: k for k, v in idalabel.items()}
snake_case : Dict = partial(__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase )
snake_case : Optional[Any] = {
"regnet-x-002": ImageNetPreTrainedConfig(
depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 , layer_type="x" ),
"regnet-x-004": ImageNetPreTrainedConfig(
depths=[1, 2, 7, 12] , hidden_sizes=[32, 64, 160, 384] , groups_width=16 , layer_type="x" ),
"regnet-x-006": ImageNetPreTrainedConfig(
depths=[1, 3, 5, 7] , hidden_sizes=[48, 96, 240, 528] , groups_width=24 , layer_type="x" ),
"regnet-x-008": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 5] , hidden_sizes=[64, 128, 288, 672] , groups_width=16 , layer_type="x" ),
"regnet-x-016": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 2] , hidden_sizes=[72, 168, 408, 912] , groups_width=24 , layer_type="x" ),
"regnet-x-032": ImageNetPreTrainedConfig(
depths=[2, 6, 15, 2] , hidden_sizes=[96, 192, 432, 1008] , groups_width=48 , layer_type="x" ),
"regnet-x-040": ImageNetPreTrainedConfig(
depths=[2, 5, 14, 2] , hidden_sizes=[80, 240, 560, 1360] , groups_width=40 , layer_type="x" ),
"regnet-x-064": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 392, 784, 1624] , groups_width=56 , layer_type="x" ),
"regnet-x-080": ImageNetPreTrainedConfig(
depths=[2, 5, 15, 1] , hidden_sizes=[80, 240, 720, 1920] , groups_width=120 , layer_type="x" ),
"regnet-x-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 , layer_type="x" ),
"regnet-x-160": ImageNetPreTrainedConfig(
depths=[2, 6, 13, 1] , hidden_sizes=[256, 512, 896, 2048] , groups_width=128 , layer_type="x" ),
"regnet-x-320": ImageNetPreTrainedConfig(
depths=[2, 7, 13, 1] , hidden_sizes=[336, 672, 1344, 2520] , groups_width=168 , layer_type="x" ),
# y variant
"regnet-y-002": ImageNetPreTrainedConfig(depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 ),
"regnet-y-004": ImageNetPreTrainedConfig(
depths=[1, 3, 6, 6] , hidden_sizes=[48, 104, 208, 440] , groups_width=8 ),
"regnet-y-006": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 4] , hidden_sizes=[48, 112, 256, 608] , groups_width=16 ),
"regnet-y-008": ImageNetPreTrainedConfig(
depths=[1, 3, 8, 2] , hidden_sizes=[64, 128, 320, 768] , groups_width=16 ),
"regnet-y-016": ImageNetPreTrainedConfig(
depths=[2, 6, 17, 2] , hidden_sizes=[48, 120, 336, 888] , groups_width=24 ),
"regnet-y-032": ImageNetPreTrainedConfig(
depths=[2, 5, 13, 1] , hidden_sizes=[72, 216, 576, 1512] , groups_width=24 ),
"regnet-y-040": ImageNetPreTrainedConfig(
depths=[2, 6, 12, 2] , hidden_sizes=[128, 192, 512, 1088] , groups_width=64 ),
"regnet-y-064": ImageNetPreTrainedConfig(
depths=[2, 7, 14, 2] , hidden_sizes=[144, 288, 576, 1296] , groups_width=72 ),
"regnet-y-080": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 448, 896, 2016] , groups_width=56 ),
"regnet-y-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 ),
"regnet-y-160": ImageNetPreTrainedConfig(
depths=[2, 4, 11, 1] , hidden_sizes=[224, 448, 1232, 3024] , groups_width=112 ),
"regnet-y-320": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
# models created by SEER -> https://arxiv.org/abs/2202.08360
"regnet-y-320-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer": RegNetConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer": RegNetConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
# finetuned on imagenet
"regnet-y-320-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer-in1k": ImageNetPreTrainedConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
}
snake_case : Union[str, Any] = NameToOurModelFuncMap()
snake_case : str = NameToFromModelFuncMap()
# add seer weights logic
def load_using_classy_vision(__lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] ) -> Tuple[nn.Module, Dict]:
snake_case : List[Any] = torch.hub.load_state_dict_from_url(__lowerCamelCase , model_dir=str(__lowerCamelCase ) , map_location="cpu" )
snake_case : Dict = model_func()
# check if we have a head, if yes add it
snake_case : str = files["classy_state_dict"]["base_model"]["model"]
snake_case : Dict = model_state_dict["trunk"]
model.load_state_dict(__lowerCamelCase )
return model.eval(), model_state_dict["heads"]
# pretrained
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet32d/seer_regnet32gf_model_iteration244000.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Optional[int] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet64/seer_regnet64gf_model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : List[str] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/swav_ig1b_regnet128Gf_cnstant_bs32_node16_sinkhorn10_proto16k_syncBN64_warmup8k/model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet10B/model_iteration124500_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
# IN1K finetuned
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet32_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet64_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : str = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet128_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Dict = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_10b_finetuned_in1k_model_phase28_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
if model_name:
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , names_to_config[model_name] , __lowerCamelCase , __lowerCamelCase , )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , )
return config, expected_shape
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help=(
"""The name of the model you wish to convert, it must be one of the supported regnet* architecture,"""
""" currently: regnetx-*, regnety-*. If `None`, all of them will the converted."""
),
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default=None,
type=Path,
required=True,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument(
"""--push_to_hub""",
default=True,
type=bool,
required=False,
help="""If True, push model and image processor to the hub.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 59
| 1
|
from dataclasses import dataclass
from typing import List, Optional, Union
import numpy as np
import PIL
import torch
from transformers import CLIPImageProcessor, CLIPVisionModel
from ...models import PriorTransformer
from ...pipelines import DiffusionPipeline
from ...schedulers import HeunDiscreteScheduler
from ...utils import (
BaseOutput,
is_accelerate_available,
logging,
randn_tensor,
replace_example_docstring,
)
from .renderer import ShapERenderer
__lowerCamelCase = logging.get_logger(__name__) # pylint: disable=invalid-name
__lowerCamelCase = """
Examples:
```py
>>> from PIL import Image
>>> import torch
>>> from diffusers import DiffusionPipeline
>>> from diffusers.utils import export_to_gif, load_image
>>> device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")
>>> repo = \"openai/shap-e-img2img\"
>>> pipe = DiffusionPipeline.from_pretrained(repo, torch_dtype=torch.float16)
>>> pipe = pipe.to(device)
>>> guidance_scale = 3.0
>>> image_url = \"https://hf.co/datasets/diffusers/docs-images/resolve/main/shap-e/corgi.png\"
>>> image = load_image(image_url).convert(\"RGB\")
>>> images = pipe(
... image,
... guidance_scale=guidance_scale,
... num_inference_steps=64,
... frame_size=256,
... ).images
>>> gif_path = export_to_gif(images[0], \"corgi_3d.gif\")
```
"""
@dataclass
class UpperCAmelCase ( A_ ):
A__ : Union[PIL.Image.Image, np.ndarray]
class UpperCAmelCase ( A_ ):
def __init__(self : Any , snake_case__ : PriorTransformer , snake_case__ : CLIPVisionModel , snake_case__ : CLIPImageProcessor , snake_case__ : HeunDiscreteScheduler , snake_case__ : ShapERenderer , ) -> Optional[int]:
'''simple docstring'''
super().__init__()
self.register_modules(
prior=snake_case__ , image_encoder=snake_case__ , image_processor=snake_case__ , scheduler=snake_case__ , renderer=snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : str , snake_case__ : Dict , snake_case__ : str , snake_case__ : Dict , snake_case__ : int , snake_case__ : Optional[int] ) -> Tuple:
'''simple docstring'''
if latents is None:
snake_case : Tuple = randn_tensor(snake_case__ , generator=snake_case__ , device=snake_case__ , dtype=snake_case__ )
else:
if latents.shape != shape:
raise ValueError(f"""Unexpected latents shape, got {latents.shape}, expected {shape}""" )
snake_case : List[Any] = latents.to(snake_case__ )
snake_case : List[Any] = latents * scheduler.init_noise_sigma
return latents
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Union[str, Any]=0 ) -> Tuple:
'''simple docstring'''
if is_accelerate_available():
from accelerate import cpu_offload
else:
raise ImportError("Please install accelerate via `pip install accelerate`" )
snake_case : Tuple = torch.device(f"""cuda:{gpu_id}""" )
snake_case : List[str] = [self.image_encoder, self.prior]
for cpu_offloaded_model in models:
if cpu_offloaded_model is not None:
cpu_offload(snake_case__ , snake_case__ )
@property
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
if self.device != torch.device("meta" ) or not hasattr(self.image_encoder , "_hf_hook" ):
return self.device
for module in self.image_encoder.modules():
if (
hasattr(snake_case__ , "_hf_hook" )
and hasattr(module._hf_hook , "execution_device" )
and module._hf_hook.execution_device is not None
):
return torch.device(module._hf_hook.execution_device )
return self.device
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : int , snake_case__ : Dict , snake_case__ : Any , snake_case__ : Optional[int] , ) -> Optional[Any]:
'''simple docstring'''
if isinstance(snake_case__ , snake_case__ ) and isinstance(image[0] , torch.Tensor ):
snake_case : str = torch.cat(snake_case__ , axis=0 ) if image[0].ndim == 4 else torch.stack(snake_case__ , axis=0 )
if not isinstance(snake_case__ , torch.Tensor ):
snake_case : Union[str, Any] = self.image_processor(snake_case__ , return_tensors="pt" ).pixel_values[0].unsqueeze(0 )
snake_case : int = image.to(dtype=self.image_encoder.dtype , device=snake_case__ )
snake_case : Tuple = self.image_encoder(snake_case__ )["last_hidden_state"]
snake_case : List[str] = image_embeds[:, 1:, :].contiguous() # batch_size, dim, 256
snake_case : List[str] = image_embeds.repeat_interleave(snake_case__ , dim=0 )
if do_classifier_free_guidance:
snake_case : int = torch.zeros_like(snake_case__ )
# For classifier free guidance, we need to do two forward passes.
# Here we concatenate the unconditional and text embeddings into a single batch
# to avoid doing two forward passes
snake_case : Optional[Any] = torch.cat([negative_image_embeds, image_embeds] )
return image_embeds
@torch.no_grad()
@replace_example_docstring(snake_case__ )
def __call__(self : Tuple , snake_case__ : Union[PIL.Image.Image, List[PIL.Image.Image]] , snake_case__ : int = 1 , snake_case__ : int = 25 , snake_case__ : Optional[Union[torch.Generator, List[torch.Generator]]] = None , snake_case__ : Optional[torch.FloatTensor] = None , snake_case__ : float = 4.0 , snake_case__ : int = 64 , snake_case__ : Optional[str] = "pil" , snake_case__ : bool = True , ) -> List[Any]:
'''simple docstring'''
if isinstance(snake_case__ , PIL.Image.Image ):
snake_case : List[Any] = 1
elif isinstance(snake_case__ , torch.Tensor ):
snake_case : Union[str, Any] = image.shape[0]
elif isinstance(snake_case__ , snake_case__ ) and isinstance(image[0] , (torch.Tensor, PIL.Image.Image) ):
snake_case : Optional[Any] = len(snake_case__ )
else:
raise ValueError(
f"""`image` has to be of type `PIL.Image.Image`, `torch.Tensor`, `List[PIL.Image.Image]` or `List[torch.Tensor]` but is {type(snake_case__ )}""" )
snake_case : List[Any] = self._execution_device
snake_case : int = batch_size * num_images_per_prompt
snake_case : Optional[Any] = guidance_scale > 1.0
snake_case : str = self._encode_image(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# prior
self.scheduler.set_timesteps(snake_case__ , device=snake_case__ )
snake_case : List[str] = self.scheduler.timesteps
snake_case : Dict = self.prior.config.num_embeddings
snake_case : int = self.prior.config.embedding_dim
snake_case : Dict = self.prepare_latents(
(batch_size, num_embeddings * embedding_dim) , image_embeds.dtype , snake_case__ , snake_case__ , snake_case__ , self.scheduler , )
# YiYi notes: for testing only to match ldm, we can directly create a latents with desired shape: batch_size, num_embeddings, embedding_dim
snake_case : Optional[Any] = latents.reshape(latents.shape[0] , snake_case__ , snake_case__ )
for i, t in enumerate(self.progress_bar(snake_case__ ) ):
# expand the latents if we are doing classifier free guidance
snake_case : Any = torch.cat([latents] * 2 ) if do_classifier_free_guidance else latents
snake_case : Tuple = self.scheduler.scale_model_input(snake_case__ , snake_case__ )
snake_case : Dict = self.prior(
snake_case__ , timestep=snake_case__ , proj_embedding=snake_case__ , ).predicted_image_embedding
# remove the variance
snake_case , snake_case : str = noise_pred.split(
scaled_model_input.shape[2] , dim=2 ) # batch_size, num_embeddings, embedding_dim
if do_classifier_free_guidance is not None:
snake_case , snake_case : str = noise_pred.chunk(2 )
snake_case : Any = noise_pred_uncond + guidance_scale * (noise_pred - noise_pred_uncond)
snake_case : Union[str, Any] = self.scheduler.step(
snake_case__ , timestep=snake_case__ , sample=snake_case__ , ).prev_sample
if output_type == "latent":
return ShapEPipelineOutput(images=snake_case__ )
snake_case : Any = []
for i, latent in enumerate(snake_case__ ):
print()
snake_case : Tuple = self.renderer.decode(
latent[None, :] , snake_case__ , size=snake_case__ , ray_batch_size=40_96 , n_coarse_samples=64 , n_fine_samples=1_28 , )
images.append(snake_case__ )
snake_case : Any = torch.stack(snake_case__ )
if output_type not in ["np", "pil"]:
raise ValueError(f"""Only the output types `pil` and `np` are supported not output_type={output_type}""" )
snake_case : Optional[int] = images.cpu().numpy()
if output_type == "pil":
snake_case : Union[str, Any] = [self.numpy_to_pil(snake_case__ ) for image in images]
# Offload last model to CPU
if hasattr(self , "final_offload_hook" ) and self.final_offload_hook is not None:
self.final_offload_hook.offload()
if not return_dict:
return (images,)
return ShapEPipelineOutput(images=snake_case__ )
| 59
|
import warnings
from typing import Dict
import numpy as np
from ..utils import ExplicitEnum, add_end_docstrings, is_tf_available, is_torch_available
from .base import PIPELINE_INIT_ARGS, GenericTensor, Pipeline
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
return 1.0 / (1.0 + np.exp(-_outputs ))
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Tuple = np.max(_outputs , axis=-1 , keepdims=__lowerCamelCase )
snake_case : int = np.exp(_outputs - maxes )
return shifted_exp / shifted_exp.sum(axis=-1 , keepdims=__lowerCamelCase )
class UpperCAmelCase ( A_ ):
A__ : Any = "sigmoid"
A__ : str = "softmax"
A__ : int = "none"
@add_end_docstrings(
A_ ,r"\n return_all_scores (`bool`, *optional*, defaults to `False`):\n Whether to return all prediction scores or just the one of the predicted class.\n function_to_apply (`str`, *optional*, defaults to `\"default\"`):\n The function to apply to the model outputs in order to retrieve the scores. Accepts four different values:\n\n - `\"default\"`: if the model has a single label, will apply the sigmoid function on the output. If the model\n has several labels, will apply the softmax function on the output.\n - `\"sigmoid\"`: Applies the sigmoid function on the output.\n - `\"softmax\"`: Applies the softmax function on the output.\n - `\"none\"`: Does not apply any function on the output.\n " ,)
class UpperCAmelCase ( A_ ):
A__ : int = False
A__ : Union[str, Any] = ClassificationFunction.NONE
def __init__(self : List[str] , **snake_case__ : int ) -> str:
'''simple docstring'''
super().__init__(**snake_case__ )
self.check_model_type(
TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if self.framework == "tf"
else MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str]=None , snake_case__ : Optional[Any]=None , snake_case__ : Union[str, Any]="" , **snake_case__ : List[str] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = tokenizer_kwargs
snake_case : List[Any] = {}
if hasattr(self.model.config , "return_all_scores" ) and return_all_scores is None:
snake_case : Optional[int] = self.model.config.return_all_scores
if isinstance(snake_case__ , snake_case__ ) or top_k is None:
snake_case : List[Any] = top_k
snake_case : str = False
elif return_all_scores is not None:
warnings.warn(
"`return_all_scores` is now deprecated, if want a similar functionality use `top_k=None` instead of"
" `return_all_scores=True` or `top_k=1` instead of `return_all_scores=False`." , snake_case__ , )
if return_all_scores:
snake_case : List[str] = None
else:
snake_case : Optional[int] = 1
if isinstance(snake_case__ , snake_case__ ):
snake_case : Dict = ClassificationFunction[function_to_apply.upper()]
if function_to_apply is not None:
snake_case : Optional[int] = function_to_apply
return preprocess_params, {}, postprocess_params
def __call__(self : Dict , *snake_case__ : List[str] , **snake_case__ : int ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = super().__call__(*snake_case__ , **snake_case__ )
# TODO try and retrieve it in a nicer way from _sanitize_parameters.
snake_case : Tuple = "top_k" not in kwargs
if isinstance(args[0] , snake_case__ ) and _legacy:
# This pipeline is odd, and return a list when single item is run
return [result]
else:
return result
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Tuple , **snake_case__ : Union[str, Any] ) -> Dict[str, GenericTensor]:
'''simple docstring'''
snake_case : int = self.framework
if isinstance(snake_case__ , snake_case__ ):
return self.tokenizer(**snake_case__ , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ) and len(snake_case__ ) == 1 and isinstance(inputs[0] , snake_case__ ) and len(inputs[0] ) == 2:
# It used to be valid to use a list of list of list for text pairs, keeping this path for BC
return self.tokenizer(
text=inputs[0][0] , text_pair=inputs[0][1] , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
# This is likely an invalid usage of the pipeline attempting to pass text pairs.
raise ValueError(
"The pipeline received invalid inputs, if you are trying to send text pairs, you can try to send a"
" dictionary `{\"text\": \"My text\", \"text_pair\": \"My pair\"}` in order to send a text pair." )
return self.tokenizer(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Union[str, Any] ) -> int:
'''simple docstring'''
return self.model(**snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Optional[Any] , snake_case__ : List[str]=None , snake_case__ : Dict=1 , snake_case__ : Tuple=True ) -> str:
'''simple docstring'''
if function_to_apply is None:
if self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels == 1:
snake_case : Tuple = ClassificationFunction.SIGMOID
elif self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels > 1:
snake_case : Tuple = ClassificationFunction.SOFTMAX
elif hasattr(self.model.config , "function_to_apply" ) and function_to_apply is None:
snake_case : Tuple = self.model.config.function_to_apply
else:
snake_case : int = ClassificationFunction.NONE
snake_case : Any = model_outputs["logits"][0]
snake_case : List[str] = outputs.numpy()
if function_to_apply == ClassificationFunction.SIGMOID:
snake_case : Optional[Any] = sigmoid(snake_case__ )
elif function_to_apply == ClassificationFunction.SOFTMAX:
snake_case : Union[str, Any] = softmax(snake_case__ )
elif function_to_apply == ClassificationFunction.NONE:
snake_case : Optional[Any] = outputs
else:
raise ValueError(f"""Unrecognized `function_to_apply` argument: {function_to_apply}""" )
if top_k == 1 and _legacy:
return {"label": self.model.config.idalabel[scores.argmax().item()], "score": scores.max().item()}
snake_case : Optional[int] = [
{"label": self.model.config.idalabel[i], "score": score.item()} for i, score in enumerate(snake_case__ )
]
if not _legacy:
dict_scores.sort(key=lambda snake_case__ : x["score"] , reverse=snake_case__ )
if top_k is not None:
snake_case : Optional[int] = dict_scores[:top_k]
return dict_scores
| 59
| 1
|
from __future__ import annotations
def UpperCamelCase ( __lowerCamelCase : list[int] ):
return len(set(__lowerCamelCase ) ) == len(__lowerCamelCase )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
|
from __future__ import annotations
__lowerCamelCase = list[list[int]]
# assigning initial values to the grid
__lowerCamelCase = [
[3, 0, 6, 5, 0, 8, 4, 0, 0],
[5, 2, 0, 0, 0, 0, 0, 0, 0],
[0, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
# a grid with no solution
__lowerCamelCase = [
[5, 0, 6, 5, 0, 8, 4, 0, 3],
[5, 2, 0, 0, 0, 0, 0, 0, 2],
[1, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
def UpperCamelCase ( __lowerCamelCase : Matrix , __lowerCamelCase : int , __lowerCamelCase : int , __lowerCamelCase : int ):
for i in range(9 ):
if grid[row][i] == n or grid[i][column] == n:
return False
for i in range(3 ):
for j in range(3 ):
if grid[(row - row % 3) + i][(column - column % 3) + j] == n:
return False
return True
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for i in range(9 ):
for j in range(9 ):
if grid[i][j] == 0:
return i, j
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
if location := find_empty_location(__lowerCamelCase ):
snake_case , snake_case : Union[str, Any] = location
else:
# If the location is ``None``, then the grid is solved.
return grid
for digit in range(1 , 10 ):
if is_safe(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase ):
snake_case : List[Any] = digit
if sudoku(__lowerCamelCase ) is not None:
return grid
snake_case : Union[str, Any] = 0
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for row in grid:
for cell in row:
print(__lowerCamelCase , end=" " )
print()
if __name__ == "__main__":
# make a copy of grid so that you can compare with the unmodified grid
for example_grid in (initial_grid, no_solution):
print("""\nExample grid:\n""" + """=""" * 20)
print_solution(example_grid)
print("""\nExample grid solution:""")
__lowerCamelCase = sudoku(example_grid)
if solution is not None:
print_solution(solution)
else:
print("""Cannot find a solution.""")
| 59
| 1
|
from typing import Optional, Tuple, Union
import flax
import flax.linen as nn
import jax
import jax.numpy as jnp
from flax.core.frozen_dict import FrozenDict
from ..configuration_utils import ConfigMixin, flax_register_to_config
from ..utils import BaseOutput
from .embeddings_flax import FlaxTimestepEmbedding, FlaxTimesteps
from .modeling_flax_utils import FlaxModelMixin
from .unet_ad_blocks_flax import (
FlaxCrossAttnDownBlockaD,
FlaxCrossAttnUpBlockaD,
FlaxDownBlockaD,
FlaxUNetMidBlockaDCrossAttn,
FlaxUpBlockaD,
)
@flax.struct.dataclass
class UpperCAmelCase ( A_ ):
A__ : jnp.ndarray
@flax_register_to_config
class UpperCAmelCase ( nn.Module ,A_ ,A_ ):
A__ : int = 32
A__ : int = 4
A__ : int = 4
A__ : Tuple[str] = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
A__ : Tuple[str] = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
A__ : Union[bool, Tuple[bool]] = False
A__ : Tuple[int] = (3_20, 6_40, 12_80, 12_80)
A__ : int = 2
A__ : Union[int, Tuple[int]] = 8
A__ : Optional[Union[int, Tuple[int]]] = None
A__ : int = 12_80
A__ : float = 0.0
A__ : bool = False
A__ : jnp.dtype = jnp.floataa
A__ : bool = True
A__ : int = 0
A__ : bool = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : jax.random.KeyArray ) -> FrozenDict:
'''simple docstring'''
snake_case : Dict = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case : Any = jnp.zeros(snake_case__ , dtype=jnp.floataa )
snake_case : List[str] = jnp.ones((1,) , dtype=jnp.intaa )
snake_case : str = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case , snake_case : Optional[int] = jax.random.split(snake_case__ )
snake_case : Union[str, Any] = {"params": params_rng, "dropout": dropout_rng}
return self.init(snake_case__ , snake_case__ , snake_case__ , snake_case__ )["params"]
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : str = self.block_out_channels
snake_case : Optional[Any] = block_out_channels[0] * 4
if self.num_attention_heads is not None:
raise ValueError(
"At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19." )
# If `num_attention_heads` is not defined (which is the case for most models)
# it will default to `attention_head_dim`. This looks weird upon first reading it and it is.
# The reason for this behavior is to correct for incorrectly named variables that were introduced
# when this library was created. The incorrect naming was only discovered much later in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131
# Changing `attention_head_dim` to `num_attention_heads` for 40,000+ configurations is too backwards breaking
# which is why we correct for the naming here.
snake_case : Tuple = self.num_attention_heads or self.attention_head_dim
# input
snake_case : Tuple = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case : Union[str, Any] = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case : Dict = FlaxTimestepEmbedding(snake_case__ , dtype=self.dtype )
snake_case : List[str] = self.only_cross_attention
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (only_cross_attention,) * len(self.down_block_types )
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case : List[Any] = []
snake_case : Optional[int] = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case : List[Any] = output_channel
snake_case : Dict = block_out_channels[i]
snake_case : Optional[Any] = i == len(snake_case__ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case : List[Any] = FlaxCrossAttnDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , num_attention_heads=num_attention_heads[i] , add_downsample=not is_final_block , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Union[str, Any] = FlaxDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , add_downsample=not is_final_block , dtype=self.dtype , )
down_blocks.append(snake_case__ )
snake_case : Dict = down_blocks
# mid
snake_case : Optional[int] = FlaxUNetMidBlockaDCrossAttn(
in_channels=block_out_channels[-1] , dropout=self.dropout , num_attention_heads=num_attention_heads[-1] , use_linear_projection=self.use_linear_projection , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
# up
snake_case : Optional[Any] = []
snake_case : Optional[int] = list(reversed(snake_case__ ) )
snake_case : Dict = list(reversed(snake_case__ ) )
snake_case : Tuple = list(reversed(snake_case__ ) )
snake_case : Optional[Any] = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case : Optional[int] = output_channel
snake_case : List[Any] = reversed_block_out_channels[i]
snake_case : Union[str, Any] = reversed_block_out_channels[min(i + 1 , len(snake_case__ ) - 1 )]
snake_case : int = i == len(snake_case__ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case : Any = FlaxCrossAttnUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , num_attention_heads=reversed_num_attention_heads[i] , add_upsample=not is_final_block , dropout=self.dropout , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Optional[int] = FlaxUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , add_upsample=not is_final_block , dropout=self.dropout , dtype=self.dtype , )
up_blocks.append(snake_case__ )
snake_case : Optional[int] = output_channel
snake_case : Tuple = up_blocks
# out
snake_case : Optional[int] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case : List[str] = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__(self : Dict , snake_case__ : Dict , snake_case__ : Dict , snake_case__ : Optional[int] , snake_case__ : Tuple=None , snake_case__ : Union[str, Any]=None , snake_case__ : bool = True , snake_case__ : bool = False , ) -> Union[FlaxUNetaDConditionOutput, Tuple]:
'''simple docstring'''
if not isinstance(snake_case__ , jnp.ndarray ):
snake_case : List[Any] = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(snake_case__ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case : Any = timesteps.astype(dtype=jnp.floataa )
snake_case : int = jnp.expand_dims(snake_case__ , 0 )
snake_case : str = self.time_proj(snake_case__ )
snake_case : str = self.time_embedding(snake_case__ )
# 2. pre-process
snake_case : int = jnp.transpose(snake_case__ , (0, 2, 3, 1) )
snake_case : List[Any] = self.conv_in(snake_case__ )
# 3. down
snake_case : Optional[int] = (sample,)
for down_block in self.down_blocks:
if isinstance(snake_case__ , snake_case__ ):
snake_case , snake_case : List[Any] = down_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
else:
snake_case , snake_case : str = down_block(snake_case__ , snake_case__ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case : Tuple = ()
for down_block_res_sample, down_block_additional_residual in zip(
snake_case__ , snake_case__ ):
down_block_res_sample += down_block_additional_residual
new_down_block_res_samples += (down_block_res_sample,)
snake_case : Optional[int] = new_down_block_res_samples
# 4. mid
snake_case : Optional[int] = self.mid_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
if mid_block_additional_residual is not None:
sample += mid_block_additional_residual
# 5. up
for up_block in self.up_blocks:
snake_case : int = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case : Optional[Any] = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(snake_case__ , snake_case__ ):
snake_case : Optional[Any] = up_block(
snake_case__ , temb=snake_case__ , encoder_hidden_states=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train , )
else:
snake_case : Dict = up_block(snake_case__ , temb=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train )
# 6. post-process
snake_case : List[str] = self.conv_norm_out(snake_case__ )
snake_case : Any = nn.silu(snake_case__ )
snake_case : Optional[int] = self.conv_out(snake_case__ )
snake_case : Union[str, Any] = jnp.transpose(snake_case__ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=snake_case__ )
| 59
|
import logging
import numpy as np
import pytest
from scipy.linalg import eigh
logging.basicConfig(level=logging.INFO, format="""%(message)s""")
def UpperCamelCase ( __lowerCamelCase : np.ndarray ):
return input_array.reshape((input_array.size, 1) )
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Any = np.nan
for i in range(__lowerCamelCase ):
snake_case : List[str] = features[:, labels == i]
snake_case : Dict = data.mean(1 )
# Centralize the data of class i
snake_case : Optional[Any] = data - column_reshape(__lowerCamelCase )
if i > 0:
# If covariance_sum is not None
covariance_sum += np.dot(__lowerCamelCase , centered_data.T )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Optional[Any] = features.mean(1 )
snake_case : Tuple = np.nan
for i in range(__lowerCamelCase ):
snake_case : Tuple = features[:, labels == i]
snake_case : Tuple = data.shape[1]
snake_case : List[str] = data.mean(1 )
if i > 0:
# If covariance_sum is not None
covariance_sum += device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[int] = device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
# Check if the features have been loaded
if features.any():
snake_case : Tuple = features.mean(1 )
# Center the dataset
snake_case : List[str] = features - np.reshape(__lowerCamelCase , (data_mean.size, 1) )
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T ) / features.shape[1]
snake_case , snake_case : Dict = np.linalg.eigh(__lowerCamelCase )
# Take all the columns in the reverse order (-1), and then takes only the first
snake_case : Optional[Any] = eigenvectors[:, ::-1][:, 0:dimensions]
# Project the database on the new space
snake_case : Union[str, Any] = np.dot(filtered_eigenvectors.T , __lowerCamelCase )
logging.info("Principal Component Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int , __lowerCamelCase : int ):
assert classes > dimensions
# Check if features have been already loaded
if features.any:
snake_case , snake_case : str = eigh(
covariance_between_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , covariance_within_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , )
snake_case : str = eigenvectors[:, ::-1][:, :dimensions]
snake_case , snake_case , snake_case : int = np.linalg.svd(__lowerCamelCase )
snake_case : List[Any] = svd_matrix[:, 0:dimensions]
snake_case : Optional[Any] = np.dot(filtered_svd_matrix.T , __lowerCamelCase )
logging.info("Linear Discriminant Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( ):
# Create dummy dataset with 2 classes and 3 features
snake_case : str = np.array([[1, 2, 3, 4, 5], [2, 3, 4, 5, 6], [3, 4, 5, 6, 7]] )
snake_case : Union[str, Any] = np.array([0, 0, 0, 1, 1] )
snake_case : List[Any] = 2
snake_case : Any = 2
# Assert that the function raises an AssertionError if dimensions > classes
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : str = linear_discriminant_analysis(
__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
if isinstance(__lowerCamelCase , np.ndarray ):
raise AssertionError(
"Did not raise AssertionError for dimensions > classes" )
assert error_info.type is AssertionError
def UpperCamelCase ( ):
snake_case : List[str] = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]] )
snake_case : List[str] = 2
snake_case : int = np.array([[6.9282_0323, 8.6602_5404, 10.3923_0485], [3.0, 3.0, 3.0]] )
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : Union[str, Any] = principal_component_analysis(__lowerCamelCase , __lowerCamelCase )
if not np.allclose(__lowerCamelCase , __lowerCamelCase ):
raise AssertionError
assert error_info.type is AssertionError
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
from typing import Dict, List, Optional
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""nielsr/canine-s""": 20_48,
}
# Unicode defines 1,114,112 total “codepoints”
__lowerCamelCase = 1_11_41_12
# Below: Constants defining canonical codepoints for special, pseudo-characters.
# Copied from https://github.com/google-research/language/blob/master/language/canine/special_codepoints.py
__lowerCamelCase = 0
__lowerCamelCase = 0xe_000
__lowerCamelCase = 0xe_001
__lowerCamelCase = 0xe_002
__lowerCamelCase = 0xe_003
__lowerCamelCase = 0xe_004
# Maps special codepoints to human-readable names.
__lowerCamelCase = {
# Special symbols are represented using codepoints values that are valid,
# but designated as "Private Use", meaning that they will never be assigned
# characters by the Unicode Consortium, and are thus safe for use here.
#
# NOTE: Do *NOT* add any sort of [UNK_CHAR] here. They are explicitly
# excluded and should fail with a hard error.
CLS: "[CLS]",
SEP: "[SEP]",
BOS: "[BOS]",
MASK: "[MASK]",
PAD: "[PAD]",
RESERVED: "[RESERVED]",
}
# Maps special codepoint human-readable names to their codepoint values.
__lowerCamelCase = {name: codepoint for codepoint, name in SPECIAL_CODEPOINTS.items()}
class UpperCAmelCase ( A_ ):
A__ : str = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__(self : List[Any] , snake_case__ : Optional[Any]=chr(snake_case__ ) , snake_case__ : Optional[int]=chr(snake_case__ ) , snake_case__ : Optional[int]=chr(snake_case__ ) , snake_case__ : Dict=chr(snake_case__ ) , snake_case__ : List[Any]=chr(snake_case__ ) , snake_case__ : Tuple=chr(snake_case__ ) , snake_case__ : Optional[int]=False , snake_case__ : Optional[int]=20_48 , **snake_case__ : str , ) -> Tuple:
'''simple docstring'''
snake_case : Dict = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else bos_token
snake_case : List[Any] = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else eos_token
snake_case : Any = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else sep_token
snake_case : Union[str, Any] = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else cls_token
snake_case : str = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else pad_token
# Mask token behave like a normal word, i.e. include the space before it
snake_case : Union[str, Any] = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else mask_token
super().__init__(
bos_token=snake_case__ , eos_token=snake_case__ , sep_token=snake_case__ , cls_token=snake_case__ , pad_token=snake_case__ , mask_token=snake_case__ , add_prefix_space=snake_case__ , model_max_length=snake_case__ , **snake_case__ , )
# Creates a mapping for looking up the IDs of special symbols.
snake_case : Dict[str, int] = {}
for codepoint, name in SPECIAL_CODEPOINTS.items():
snake_case : List[str] = codepoint
# Creates a mapping for looking up the string forms of special symbol IDs.
snake_case : Dict[int, str] = {
codepoint: name for name, codepoint in self._special_codepoints.items()
}
snake_case : Dict = UNICODE_VOCAB_SIZE
snake_case : Dict = len(self._special_codepoints )
@property
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> int:
'''simple docstring'''
return self._unicode_vocab_size
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : str ) -> List[str]:
'''simple docstring'''
return list(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : str ) -> int:
'''simple docstring'''
try:
return ord(snake_case__ )
except TypeError:
raise ValueError(f"""invalid token: '{token}'""" )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : int ) -> str:
'''simple docstring'''
try:
if index in SPECIAL_CODEPOINTS:
return SPECIAL_CODEPOINTS[index]
return chr(snake_case__ )
except TypeError:
raise ValueError(f"""invalid id: {index}""" )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Tuple ) -> str:
'''simple docstring'''
return "".join(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Any = [self.sep_token_id]
snake_case : List[Any] = [self.cls_token_id]
snake_case : str = cls + token_ids_a + sep
if token_ids_a is not None:
result += token_ids_a + sep
return result
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None , snake_case__ : bool = False ) -> List[int]:
'''simple docstring'''
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=snake_case__ , token_ids_a=snake_case__ , already_has_special_tokens=snake_case__ )
snake_case : Tuple = [1] + ([0] * len(snake_case__ )) + [1]
if token_ids_a is not None:
result += ([0] * len(snake_case__ )) + [1]
return result
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : str = [self.sep_token_id]
snake_case : Optional[int] = [self.cls_token_id]
snake_case : int = len(cls + token_ids_a + sep ) * [0]
if token_ids_a is not None:
result += len(token_ids_a + sep ) * [1]
return result
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : str , snake_case__ : Optional[str] = None ) -> List[Any]:
'''simple docstring'''
return ()
| 59
|
import pytest
from datasets import inspect_metric, list_metrics, load_metric
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : Optional[int] ):
monkeypatch.setattr("datasets.utils.deprecation_utils._emitted_deprecation_warnings" , set() )
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : str ):
class UpperCAmelCase :
def __init__(self : Optional[int] , snake_case__ : str ) -> Any:
'''simple docstring'''
snake_case : List[str] = metric_id
class UpperCAmelCase :
A__ : List[str] = [MetricMock(A_ ) for metric_id in ["accuracy", "mse", "precision", "codeparrot/apps_metric"]]
def _SCREAMING_SNAKE_CASE (self : int ) -> List[str]:
'''simple docstring'''
return self._metrics
monkeypatch.setattr("datasets.inspect.huggingface_hub" , HfhMock() )
@pytest.mark.parametrize(
"func, args" , [(load_metric, ("metrics/mse",)), (list_metrics, ()), (inspect_metric, ("metrics/mse", "tmp_path"))] )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Any , __lowerCamelCase : Any , __lowerCamelCase : int , __lowerCamelCase : Any ):
if "tmp_path" in args:
snake_case : str = tuple(arg if arg != "tmp_path" else tmp_path for arg in args )
with pytest.warns(__lowerCamelCase , match="https://huggingface.co/docs/evaluate" ):
func(*__lowerCamelCase )
| 59
| 1
|
import argparse
import os
import sys
from unittest.mock import patch
import pytorch_lightning as pl
import timeout_decorator
import torch
from distillation import SummarizationDistiller, distill_main
from finetune import SummarizationModule, main
from transformers import MarianMTModel
from transformers.file_utils import cached_path
from transformers.testing_utils import TestCasePlus, require_torch_gpu, slow
from utils import load_json
__lowerCamelCase = """sshleifer/mar_enro_6_3_student"""
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
super().setUp()
snake_case : List[str] = cached_path(
"https://cdn-datasets.huggingface.co/translation/wmt_en_ro-tr40k-va0.5k-te0.5k.tar.gz" , extract_compressed_file=snake_case__ , )
snake_case : Union[str, Any] = f"""{data_cached}/wmt_en_ro-tr40k-va0.5k-te0.5k"""
@slow
@require_torch_gpu
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> List[str]:
'''simple docstring'''
MarianMTModel.from_pretrained(snake_case__ )
@slow
@require_torch_gpu
def _SCREAMING_SNAKE_CASE (self : Dict ) -> List[str]:
'''simple docstring'''
snake_case : Tuple = {
"$MAX_LEN": 64,
"$BS": 64,
"$GAS": 1,
"$ENRO_DIR": self.data_dir,
"facebook/mbart-large-cc25": MARIAN_MODEL,
# "val_check_interval=0.25": "val_check_interval=1.0",
"--learning_rate=3e-5": "--learning_rate 3e-4",
"--num_train_epochs 6": "--num_train_epochs 1",
}
# Clean up bash script
snake_case : int = (self.test_file_dir / "train_mbart_cc25_enro.sh").open().read().split("finetune.py" )[1].strip()
snake_case : Optional[int] = bash_script.replace("\\\n" , "" ).strip().replace("\"$@\"" , "" )
for k, v in env_vars_to_replace.items():
snake_case : Optional[int] = bash_script.replace(snake_case__ , str(snake_case__ ) )
snake_case : int = self.get_auto_remove_tmp_dir()
# bash_script = bash_script.replace("--fp16 ", "")
snake_case : List[str] = f"""
--output_dir {output_dir}
--tokenizer_name Helsinki-NLP/opus-mt-en-ro
--sortish_sampler
--do_predict
--gpus 1
--freeze_encoder
--n_train 40000
--n_val 500
--n_test 500
--fp16_opt_level O1
--num_sanity_val_steps 0
--eval_beams 2
""".split()
# XXX: args.gpus > 1 : handle multi_gpu in the future
snake_case : Optional[Any] = ["finetune.py"] + bash_script.split() + args
with patch.object(snake_case__ , "argv" , snake_case__ ):
snake_case : Dict = argparse.ArgumentParser()
snake_case : List[str] = pl.Trainer.add_argparse_args(snake_case__ )
snake_case : Dict = SummarizationModule.add_model_specific_args(snake_case__ , os.getcwd() )
snake_case : Union[str, Any] = parser.parse_args()
snake_case : int = main(snake_case__ )
# Check metrics
snake_case : List[Any] = load_json(model.metrics_save_path )
snake_case : Union[str, Any] = metrics["val"][0]
snake_case : Dict = metrics["val"][-1]
self.assertEqual(len(metrics["val"] ) , (args.max_epochs / args.val_check_interval) )
assert isinstance(last_step_stats[f"""val_avg_{model.val_metric}"""] , snake_case__ )
self.assertGreater(last_step_stats["val_avg_gen_time"] , 0.01 )
# model hanging on generate. Maybe bad config was saved. (XXX: old comment/assert?)
self.assertLessEqual(last_step_stats["val_avg_gen_time"] , 1.0 )
# test learning requirements:
# 1. BLEU improves over the course of training by more than 2 pts
self.assertGreater(last_step_stats["val_avg_bleu"] - first_step_stats["val_avg_bleu"] , 2 )
# 2. BLEU finishes above 17
self.assertGreater(last_step_stats["val_avg_bleu"] , 17 )
# 3. test BLEU and val BLEU within ~1.1 pt.
self.assertLess(abs(metrics["val"][-1]["val_avg_bleu"] - metrics["test"][-1]["test_avg_bleu"] ) , 1.1 )
# check lightning ckpt can be loaded and has a reasonable statedict
snake_case : int = os.listdir(snake_case__ )
snake_case : Any = [x for x in contents if x.endswith(".ckpt" )][0]
snake_case : List[Any] = os.path.join(args.output_dir , snake_case__ )
snake_case : str = torch.load(snake_case__ , map_location="cpu" )
snake_case : Any = "model.model.decoder.layers.0.encoder_attn_layer_norm.weight"
assert expected_key in ckpt["state_dict"]
assert ckpt["state_dict"]["model.model.decoder.layers.0.encoder_attn_layer_norm.weight"].dtype == torch.floataa
# TODO: turn on args.do_predict when PL bug fixed.
if args.do_predict:
snake_case : Any = {os.path.basename(snake_case__ ) for p in contents}
assert "test_generations.txt" in contents
assert "test_results.txt" in contents
# assert len(metrics["val"]) == desired_n_evals
assert len(metrics["test"] ) == 1
class UpperCAmelCase ( A_ ):
@timeout_decorator.timeout(6_00 )
@slow
@require_torch_gpu
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
snake_case : Any = f"""{self.test_file_dir_str}/test_data/wmt_en_ro"""
snake_case : List[Any] = {
"--fp16_opt_level=O1": "",
"$MAX_LEN": 1_28,
"$BS": 16,
"$GAS": 1,
"$ENRO_DIR": data_dir,
"$m": "sshleifer/student_marian_en_ro_6_1",
"val_check_interval=0.25": "val_check_interval=1.0",
}
# Clean up bash script
snake_case : Dict = (
(self.test_file_dir / "distil_marian_no_teacher.sh").open().read().split("distillation.py" )[1].strip()
)
snake_case : Union[str, Any] = bash_script.replace("\\\n" , "" ).strip().replace("\"$@\"" , "" )
snake_case : str = bash_script.replace("--fp16 " , " " )
for k, v in env_vars_to_replace.items():
snake_case : Union[str, Any] = bash_script.replace(snake_case__ , str(snake_case__ ) )
snake_case : List[str] = self.get_auto_remove_tmp_dir()
snake_case : Any = bash_script.replace("--fp16" , "" )
snake_case : Tuple = 6
snake_case : str = (
["distillation.py"]
+ bash_script.split()
+ [
f"""--output_dir={output_dir}""",
"--gpus=1",
"--learning_rate=1e-3",
f"""--num_train_epochs={epochs}""",
"--warmup_steps=10",
"--val_check_interval=1.0",
"--do_predict",
]
)
with patch.object(snake_case__ , "argv" , snake_case__ ):
snake_case : List[str] = argparse.ArgumentParser()
snake_case : int = pl.Trainer.add_argparse_args(snake_case__ )
snake_case : Tuple = SummarizationDistiller.add_model_specific_args(snake_case__ , os.getcwd() )
snake_case : List[Any] = parser.parse_args()
# assert args.gpus == gpus THIS BREAKS for multi_gpu
snake_case : Dict = distill_main(snake_case__ )
# Check metrics
snake_case : Tuple = load_json(model.metrics_save_path )
snake_case : Optional[int] = metrics["val"][0]
snake_case : int = metrics["val"][-1]
assert len(metrics["val"] ) >= (args.max_epochs / args.val_check_interval) # +1 accounts for val_sanity_check
assert last_step_stats["val_avg_gen_time"] >= 0.01
assert first_step_stats["val_avg_bleu"] < last_step_stats["val_avg_bleu"] # model learned nothing
assert 1.0 >= last_step_stats["val_avg_gen_time"] # model hanging on generate. Maybe bad config was saved.
assert isinstance(last_step_stats[f"""val_avg_{model.val_metric}"""] , snake_case__ )
# check lightning ckpt can be loaded and has a reasonable statedict
snake_case : int = os.listdir(snake_case__ )
snake_case : str = [x for x in contents if x.endswith(".ckpt" )][0]
snake_case : List[Any] = os.path.join(args.output_dir , snake_case__ )
snake_case : str = torch.load(snake_case__ , map_location="cpu" )
snake_case : Dict = "model.model.decoder.layers.0.encoder_attn_layer_norm.weight"
assert expected_key in ckpt["state_dict"]
assert ckpt["state_dict"]["model.model.decoder.layers.0.encoder_attn_layer_norm.weight"].dtype == torch.floataa
# TODO: turn on args.do_predict when PL bug fixed.
if args.do_predict:
snake_case : Dict = {os.path.basename(snake_case__ ) for p in contents}
assert "test_generations.txt" in contents
assert "test_results.txt" in contents
# assert len(metrics["val"]) == desired_n_evals
assert len(metrics["test"] ) == 1
| 59
|
import argparse
import dataclasses
import json
import logging
import os
import shutil
from typing import List, Optional
import datasets
from accelerate import Accelerator
from datasets import load_dataset
from finetuning import finetune
from tqdm.auto import tqdm
import transformers
from transformers import AutoConfig, set_seed
from transformers.trainer_utils import IntervalStrategy
__lowerCamelCase = logging.getLogger(__name__)
__lowerCamelCase = """pytorch_model.bin"""
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co."} ,)
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the training data."} )
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the data to predict on."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "A csv or a json file containing the validation data."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "The name of the task to train on."} ,)
A__ : Optional[List[str]] = dataclasses.field(
default=A_ ,metadata={"help": "The list of labels for the task."} )
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "The output directory where the model predictions and checkpoints will be written."} )
A__ : Optional[str] = dataclasses.field(
default="accuracy" ,metadata={"help": "The evaluation metric used for the task."} )
A__ : Optional[str] = dataclasses.field(
default="no" ,metadata={
"help": "The evaluation strategy to adopt during training. Possible values are: [\"no\", \"step\", \"epoch]"
} ,)
A__ : Optional[int] = dataclasses.field(
default=10 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={
"help": "How much the specified evaluation metric must improve to satisfy early stopping conditions."
} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the confidence score."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the validation performance."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to fine-tune on labeled data after pseudo training."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={"help": "Confidence threshold for pseudo-labeled data filtering."} ,)
A__ : Optional[int] = dataclasses.field(
default=1_00 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[int] = dataclasses.field(
default=A_ ,metadata={"help": "Random seed for initialization."} ,)
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Tuple , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Tuple , __lowerCamelCase : Dict , __lowerCamelCase : Optional[int] ):
snake_case : Tuple = datasets.concatenate_datasets([infer_input, infer_output] , axis=1 )
if args.do_filter_by_confidence:
snake_case : Optional[int] = dataset.filter(lambda __lowerCamelCase : example["probability"] > args.confidence_threshold )
if args.do_filter_by_val_performance:
assert eval_result >= 0.0 and eval_result <= 1.0
snake_case : int = int(eval_result * len(__lowerCamelCase ) )
print(__lowerCamelCase )
snake_case : List[str] = dataset.sort("probability" , reverse=__lowerCamelCase )
snake_case : Tuple = dataset.select(range(__lowerCamelCase ) )
snake_case : List[Any] = dataset.remove_columns(["label", "probability"] )
snake_case : Any = dataset.rename_column("prediction" , "label" )
snake_case : str = dataset.map(lambda __lowerCamelCase : {"label": idalabel[example["label"]]} )
snake_case : List[str] = dataset.shuffle(seed=args.seed )
snake_case : int = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.data_file_extension == "csv":
dataset.to_csv(__lowerCamelCase , index=__lowerCamelCase )
else:
dataset.to_json(__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : str , __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple , **__lowerCamelCase : List[Any] ):
snake_case : int = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(accelerator.state )
# Setup logging, we only want one process per machine to log things on the
# screen. accelerator.is_local_main_process is only True for one process per
# machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
snake_case : Dict = STModelArguments(model_name_or_path=__lowerCamelCase )
snake_case : Tuple = STDataArguments(train_file=__lowerCamelCase , infer_file=__lowerCamelCase )
snake_case : str = STTrainingArguments(output_dir=__lowerCamelCase )
snake_case : int = argparse.Namespace()
for arg_class in (model_args, data_args, training_args):
for key, value in vars(__lowerCamelCase ).items():
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
for key, value in kwargs.items():
if hasattr(__lowerCamelCase , __lowerCamelCase ):
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
# Sanity checks
snake_case : List[str] = {}
snake_case : Optional[int] = None
# You need to provide the training data and the data to predict on
assert args.train_file is not None
assert args.infer_file is not None
snake_case : str = args.train_file
snake_case : Tuple = args.infer_file
if args.evaluation_strategy != IntervalStrategy.NO.value:
assert args.eval_file is not None
snake_case : Tuple = args.eval_file
for key in data_files:
snake_case : List[Any] = data_files[key].split("." )[-1]
assert extension in ["csv", "json"], f"""`{key}_file` should be a csv or a json file."""
if args.data_file_extension is None:
snake_case : Union[str, Any] = extension
else:
assert extension == args.data_file_extension, f"""`{key}_file` should be a {args.data_file_extension} file`."""
assert (
args.eval_metric in datasets.list_metrics()
), f"""{args.eval_metric} not in the list of supported metrics {datasets.list_metrics()}."""
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed )
logger.info("Creating the initial data directory for self-training..." )
snake_case : List[Any] = f"""{args.output_dir}/self-train_iter-{{}}""".format
snake_case : Optional[int] = data_dir_format(0 )
if accelerator.is_main_process:
if args.output_dir is not None:
os.makedirs(args.output_dir , exist_ok=__lowerCamelCase )
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : Dict = None
snake_case : Union[str, Any] = None
snake_case : Tuple = 0
snake_case : List[Any] = False
# Show the progress bar
snake_case : List[Any] = tqdm(range(args.max_selftrain_iterations ) , disable=not accelerator.is_local_main_process )
# Self-train
for iteration in range(0 , int(args.max_selftrain_iterations ) ):
snake_case : str = data_dir_format(__lowerCamelCase )
assert os.path.exists(__lowerCamelCase )
# Stage 1: initial fine-tuning for iteration = 0 or pseudo-training for
# iteration > 0
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-1" )
snake_case : Optional[Any] = {
"accelerator": accelerator,
"model_name_or_path": args.model_name_or_path,
"cache_dir": args.cache_dir,
"do_train": True,
"train_file": data_files["train"] if iteration == 0 else data_files["train_pseudo"],
"do_eval": True if args.eval_file is not None else False,
"eval_file": data_files["eval"],
"do_predict": True,
"infer_file": data_files["infer"],
"task_name": args.task_name,
"label_list": args.label_list,
"output_dir": current_output_dir,
"eval_metric": args.eval_metric,
"evaluation_strategy": args.evaluation_strategy,
"early_stopping_patience": args.early_stopping_patience,
"early_stopping_threshold": args.early_stopping_threshold,
"seed": args.seed,
}
# Add additional training arguments
for key, value in kwargs.items():
if key not in arguments_dict and not hasattr(__lowerCamelCase , __lowerCamelCase ):
arguments_dict.update({key: value} )
snake_case : int = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 1." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 1 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 1." , __lowerCamelCase )
if iteration > 0 and args.finetune_on_labeled_data:
# Stage 2 (optional): fine-tuning on the original labeled data
snake_case : str = os.path.join(__lowerCamelCase , "best-checkpoint" )
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-2" )
# Update arguments_dict
snake_case : List[str] = model_path
snake_case : Optional[Any] = data_files["train"]
snake_case : Optional[Any] = current_output_dir
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 2." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 2 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 2." , __lowerCamelCase )
snake_case : int = iteration
snake_case : Tuple = data_dir_format(iteration + 1 )
snake_case : Tuple = AutoConfig.from_pretrained(os.path.join(__lowerCamelCase , "best-checkpoint" ) )
snake_case : Optional[int] = config.idalabel
snake_case : List[Any] = os.path.join(__lowerCamelCase , "eval_results_best-checkpoint.json" )
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "test_results_best-checkpoint.json" )
assert os.path.exists(__lowerCamelCase )
with open(__lowerCamelCase , "r" ) as f:
snake_case : Dict = float(json.load(__lowerCamelCase )[args.eval_metric] )
snake_case : Optional[int] = os.path.join(__lowerCamelCase , "infer_output_best-checkpoint.csv" )
assert os.path.exists(__lowerCamelCase )
# Loading the dataset from local csv or json files.
snake_case : Optional[Any] = load_dataset(args.data_file_extension , data_files={"data": data_files["infer"]} )["data"]
snake_case : Dict = load_dataset("csv" , data_files={"data": infer_output_file} )["data"]
if accelerator.is_main_process:
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) )
if os.path.exists(__lowerCamelCase ):
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""test_results_iter-{iteration}.json""" ) )
create_pseudo_labeled_data(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : str = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.evaluation_strategy != IntervalStrategy.NO.value:
snake_case : List[Any] = eval_result
if best_iteration is None:
snake_case : List[Any] = new_iteration
snake_case : int = new_eval_result
else:
if new_eval_result - best_eval_result > args.early_stopping_threshold:
snake_case : int = new_iteration
snake_case : Union[str, Any] = new_eval_result
snake_case : str = 0
else:
if new_eval_result == best_eval_result:
snake_case : Any = new_iteration
snake_case : Union[str, Any] = new_eval_result
early_stopping_patience_counter += 1
if early_stopping_patience_counter >= args.early_stopping_patience:
snake_case : Tuple = True
progress_bar.update(1 )
if should_training_stop:
break
if best_iteration is not None:
# Save the best iteration
logger.info("Best iteration: %d" , __lowerCamelCase )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
else:
# Assume that the last iteration is the best
logger.info("Best iteration: %d" , args.max_selftrain_iterations - 1 )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{args.max_selftrain_iterations - 1}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
| 59
| 1
|
import qiskit
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : int ):
snake_case : Dict = qiskit.Aer.get_backend("aer_simulator" )
# Create a Quantum Circuit acting on the q register
snake_case : Any = qiskit.QuantumCircuit(__lowerCamelCase , __lowerCamelCase )
# Map the quantum measurement to the classical bits
circuit.measure([0] , [0] )
# Execute the circuit on the simulator
snake_case : Optional[Any] = qiskit.execute(__lowerCamelCase , __lowerCamelCase , shots=1000 )
# Return the histogram data of the results of the experiment.
return job.result().get_counts(__lowerCamelCase )
if __name__ == "__main__":
print(F'Total count for various states are: {single_qubit_measure(1, 1)}')
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__lowerCamelCase = {"""configuration_xglm""": ["""XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP""", """XGLMConfig"""]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizer"""]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizerFast"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""XGLMForCausalLM""",
"""XGLMModel""",
"""XGLMPreTrainedModel""",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""FlaxXGLMForCausalLM""",
"""FlaxXGLMModel""",
"""FlaxXGLMPreTrainedModel""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""TFXGLMForCausalLM""",
"""TFXGLMModel""",
"""TFXGLMPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm import XGLMTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm_fast import XGLMTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_xglm import (
TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST,
TFXGLMForCausalLM,
TFXGLMModel,
TFXGLMPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure)
| 59
| 1
|
import math
import os
from copy import deepcopy
import datasets
import evaluate
import torch
import transformers
from datasets import load_dataset
from torch.utils.data import DataLoader
from transformers import AutoModelForSequenceClassification, AutoTokenizer
from accelerate import Accelerator
from accelerate.test_utils import RegressionDataset, RegressionModel
from accelerate.utils import is_tpu_available, set_seed
__lowerCamelCase = """true"""
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Any=82 , __lowerCamelCase : int=16 ):
set_seed(42 )
snake_case : Optional[int] = RegressionModel()
snake_case : Optional[Any] = deepcopy(__lowerCamelCase )
snake_case : Any = RegressionDataset(length=__lowerCamelCase )
snake_case : List[Any] = DataLoader(__lowerCamelCase , batch_size=__lowerCamelCase )
model.to(accelerator.device )
snake_case , snake_case : Tuple = accelerator.prepare(__lowerCamelCase , __lowerCamelCase )
return model, ddp_model, dataloader
def UpperCamelCase ( __lowerCamelCase : Accelerator , __lowerCamelCase : List[str]=False ):
snake_case : Tuple = AutoTokenizer.from_pretrained("hf-internal-testing/mrpc-bert-base-cased" )
snake_case : Any = load_dataset("glue" , "mrpc" , split="validation" )
def tokenize_function(__lowerCamelCase : Optional[Any] ):
snake_case : List[Any] = tokenizer(examples["sentence1"] , examples["sentence2"] , truncation=__lowerCamelCase , max_length=__lowerCamelCase )
return outputs
with accelerator.main_process_first():
snake_case : Optional[Any] = dataset.map(
__lowerCamelCase , batched=__lowerCamelCase , remove_columns=["idx", "sentence1", "sentence2"] , )
snake_case : Dict = tokenized_datasets.rename_column("label" , "labels" )
def collate_fn(__lowerCamelCase : Any ):
if use_longest:
return tokenizer.pad(__lowerCamelCase , padding="longest" , return_tensors="pt" )
return tokenizer.pad(__lowerCamelCase , padding="max_length" , max_length=128 , return_tensors="pt" )
return DataLoader(__lowerCamelCase , shuffle=__lowerCamelCase , collate_fn=__lowerCamelCase , batch_size=16 )
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Dict ):
snake_case : Optional[Any] = Accelerator(dispatch_batches=__lowerCamelCase , split_batches=__lowerCamelCase )
snake_case : List[str] = get_dataloader(__lowerCamelCase , not dispatch_batches )
snake_case : List[Any] = AutoModelForSequenceClassification.from_pretrained(
"hf-internal-testing/mrpc-bert-base-cased" , return_dict=__lowerCamelCase )
snake_case , snake_case : str = accelerator.prepare(__lowerCamelCase , __lowerCamelCase )
return {"ddp": [ddp_model, ddp_dataloader, "cuda:0"], "no": [model, dataloader, accelerator.device]}, accelerator
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Dict ):
snake_case : Optional[Any] = []
for batch in dataloader:
snake_case , snake_case : Optional[int] = batch.values()
with torch.no_grad():
snake_case : List[Any] = model(__lowerCamelCase )
snake_case , snake_case : Union[str, Any] = accelerator.gather_for_metrics((logit, target) )
logits_and_targets.append((logit, target) )
snake_case , snake_case : str = [], []
for logit, targ in logits_and_targets:
logits.append(__lowerCamelCase )
targs.append(__lowerCamelCase )
snake_case , snake_case : Union[str, Any] = torch.cat(__lowerCamelCase ), torch.cat(__lowerCamelCase )
return logits, targs
def UpperCamelCase ( __lowerCamelCase : Accelerator , __lowerCamelCase : Any=82 , __lowerCamelCase : int=False , __lowerCamelCase : Any=False , __lowerCamelCase : Optional[int]=16 ):
snake_case , snake_case , snake_case : Optional[Any] = get_basic_setup(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
snake_case , snake_case : Dict = generate_predictions(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
assert (
len(__lowerCamelCase ) == num_samples
), f"""Unexpected number of inputs:\n Expected: {num_samples}\n Actual: {len(__lowerCamelCase )}"""
def UpperCamelCase ( __lowerCamelCase : bool = False , __lowerCamelCase : bool = False ):
snake_case : Union[str, Any] = evaluate.load("glue" , "mrpc" )
snake_case , snake_case : int = get_mrpc_setup(__lowerCamelCase , __lowerCamelCase )
# First do baseline
snake_case , snake_case , snake_case : List[str] = setup["no"]
model.to(__lowerCamelCase )
model.eval()
for batch in dataloader:
batch.to(__lowerCamelCase )
with torch.inference_mode():
snake_case : int = model(**__lowerCamelCase )
snake_case : Optional[int] = outputs.logits.argmax(dim=-1 )
metric.add_batch(predictions=__lowerCamelCase , references=batch["labels"] )
snake_case : Tuple = metric.compute()
# Then do distributed
snake_case , snake_case , snake_case : Optional[int] = setup["ddp"]
model.eval()
for batch in dataloader:
with torch.inference_mode():
snake_case : str = model(**__lowerCamelCase )
snake_case : List[Any] = outputs.logits.argmax(dim=-1 )
snake_case : Optional[int] = batch["labels"]
snake_case , snake_case : Dict = accelerator.gather_for_metrics((preds, references) )
metric.add_batch(predictions=__lowerCamelCase , references=__lowerCamelCase )
snake_case : Optional[Any] = metric.compute()
for key in "accuracy f1".split():
assert math.isclose(
baseline[key] , distributed[key] ), f"""Baseline and Distributed are not the same for key {key}:\n\tBaseline: {baseline[key]}\n\tDistributed: {distributed[key]}\n"""
def UpperCamelCase ( ):
snake_case : List[Any] = Accelerator(split_batches=__lowerCamelCase , dispatch_batches=__lowerCamelCase )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_warning()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
# These are a bit slower so they should only be ran on the GPU or TPU
if torch.cuda.is_available() or is_tpu_available():
if accelerator.is_local_main_process:
print("**Testing gather_for_metrics**" )
for split_batches in [True, False]:
for dispatch_batches in [True, False]:
if accelerator.is_local_main_process:
print(f"""With: `split_batches={split_batches}`, `dispatch_batches={dispatch_batches}`""" )
test_mrpc(__lowerCamelCase , __lowerCamelCase )
accelerator.state._reset_state()
if accelerator.is_local_main_process:
print("**Test torch metrics**" )
for split_batches in [True, False]:
for dispatch_batches in [True, False]:
snake_case : Optional[Any] = Accelerator(split_batches=__lowerCamelCase , dispatch_batches=__lowerCamelCase )
if accelerator.is_local_main_process:
print(f"""With: `split_batches={split_batches}`, `dispatch_batches={dispatch_batches}`, length=99""" )
test_torch_metrics(__lowerCamelCase , 99 )
accelerator.state._reset_state()
if accelerator.is_local_main_process:
print("**Test last batch is not dropped when perfectly divisible**" )
snake_case : Optional[int] = Accelerator()
test_torch_metrics(__lowerCamelCase , 512 )
accelerator.state._reset_state()
def UpperCamelCase ( __lowerCamelCase : Tuple ):
# For xla_spawn (TPUs)
main()
if __name__ == "__main__":
main()
| 59
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
# See all MEGATRON_BERT models at https://huggingface.co/models?filter=bert
}
class UpperCAmelCase ( A_ ):
A__ : List[str] = "megatron-bert"
def __init__(self : Optional[int] , snake_case__ : List[str]=2_90_56 , snake_case__ : List[Any]=10_24 , snake_case__ : str=24 , snake_case__ : Tuple=16 , snake_case__ : Union[str, Any]=40_96 , snake_case__ : str="gelu" , snake_case__ : str=0.1 , snake_case__ : Optional[int]=0.1 , snake_case__ : Tuple=5_12 , snake_case__ : Union[str, Any]=2 , snake_case__ : Dict=0.02 , snake_case__ : List[Any]=1e-12 , snake_case__ : int=0 , snake_case__ : Tuple="absolute" , snake_case__ : Any=True , **snake_case__ : Union[str, Any] , ) -> Optional[Any]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , **snake_case__ )
snake_case : Tuple = vocab_size
snake_case : str = hidden_size
snake_case : str = num_hidden_layers
snake_case : str = num_attention_heads
snake_case : Optional[int] = hidden_act
snake_case : int = intermediate_size
snake_case : List[str] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : List[str] = type_vocab_size
snake_case : List[str] = initializer_range
snake_case : Tuple = layer_norm_eps
snake_case : int = position_embedding_type
snake_case : str = use_cache
| 59
| 1
|
from typing import Dict, List, Optional, Union
import numpy as np
from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
from ...image_transforms import (
center_crop,
convert_to_rgb,
get_resize_output_image_size,
normalize,
rescale,
resize,
to_channel_dimension_format,
)
from ...image_utils import (
OPENAI_CLIP_MEAN,
OPENAI_CLIP_STD,
ChannelDimension,
ImageInput,
PILImageResampling,
make_list_of_images,
to_numpy_array,
valid_images,
)
from ...utils import TensorType, is_vision_available, logging
__lowerCamelCase = logging.get_logger(__name__)
if is_vision_available():
import PIL
class UpperCAmelCase ( A_ ):
A__ : str = ["pixel_values"]
def __init__(self : List[str] , snake_case__ : bool = True , snake_case__ : Dict[str, int] = None , snake_case__ : PILImageResampling = PILImageResampling.BICUBIC , snake_case__ : bool = True , snake_case__ : Dict[str, int] = None , snake_case__ : bool = True , snake_case__ : Union[int, float] = 1 / 2_55 , snake_case__ : bool = True , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : bool = True , **snake_case__ : List[Any] , ) -> None:
'''simple docstring'''
super().__init__(**snake_case__ )
snake_case : Any = size if size is not None else {"shortest_edge": 2_24}
snake_case : str = get_size_dict(snake_case__ , default_to_square=snake_case__ )
snake_case : Dict = crop_size if crop_size is not None else {"height": 2_24, "width": 2_24}
snake_case : Optional[int] = get_size_dict(snake_case__ , default_to_square=snake_case__ , param_name="crop_size" )
snake_case : List[str] = do_resize
snake_case : List[str] = size
snake_case : Optional[int] = resample
snake_case : List[str] = do_center_crop
snake_case : List[Any] = crop_size
snake_case : Any = do_rescale
snake_case : Union[str, Any] = rescale_factor
snake_case : Dict = do_normalize
snake_case : Tuple = image_mean if image_mean is not None else OPENAI_CLIP_MEAN
snake_case : Union[str, Any] = image_std if image_std is not None else OPENAI_CLIP_STD
snake_case : Union[str, Any] = do_convert_rgb
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : np.ndarray , snake_case__ : Dict[str, int] , snake_case__ : PILImageResampling = PILImageResampling.BICUBIC , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : Optional[int] , ) -> np.ndarray:
'''simple docstring'''
snake_case : List[str] = get_size_dict(snake_case__ , default_to_square=snake_case__ )
if "shortest_edge" not in size:
raise ValueError(f"""The `size` parameter must contain the key `shortest_edge`. Got {size.keys()}""" )
snake_case : Dict = get_resize_output_image_size(snake_case__ , size=size["shortest_edge"] , default_to_square=snake_case__ )
return resize(snake_case__ , size=snake_case__ , resample=snake_case__ , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : np.ndarray , snake_case__ : Dict[str, int] , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : Optional[Any] , ) -> np.ndarray:
'''simple docstring'''
snake_case : Union[str, Any] = get_size_dict(snake_case__ )
if "height" not in size or "width" not in size:
raise ValueError(f"""The `size` parameter must contain the keys (height, width). Got {size.keys()}""" )
return center_crop(snake_case__ , size=(size["height"], size["width"]) , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : np.ndarray , snake_case__ : Union[int, float] , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : Optional[int] , ) -> Any:
'''simple docstring'''
return rescale(snake_case__ , scale=snake_case__ , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : np.ndarray , snake_case__ : Union[float, List[float]] , snake_case__ : Union[float, List[float]] , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : Optional[int] , ) -> np.ndarray:
'''simple docstring'''
return normalize(snake_case__ , mean=snake_case__ , std=snake_case__ , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : ImageInput , snake_case__ : bool = None , snake_case__ : Dict[str, int] = None , snake_case__ : PILImageResampling = None , snake_case__ : bool = None , snake_case__ : int = None , snake_case__ : bool = None , snake_case__ : float = None , snake_case__ : bool = None , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : bool = None , snake_case__ : Optional[Union[str, TensorType]] = None , snake_case__ : Optional[ChannelDimension] = ChannelDimension.FIRST , **snake_case__ : Any , ) -> PIL.Image.Image:
'''simple docstring'''
snake_case : Optional[int] = do_resize if do_resize is not None else self.do_resize
snake_case : Union[str, Any] = size if size is not None else self.size
snake_case : List[str] = get_size_dict(snake_case__ , param_name="size" , default_to_square=snake_case__ )
snake_case : Any = resample if resample is not None else self.resample
snake_case : Dict = do_center_crop if do_center_crop is not None else self.do_center_crop
snake_case : List[Any] = crop_size if crop_size is not None else self.crop_size
snake_case : Optional[int] = get_size_dict(snake_case__ , param_name="crop_size" , default_to_square=snake_case__ )
snake_case : Any = do_rescale if do_rescale is not None else self.do_rescale
snake_case : str = rescale_factor if rescale_factor is not None else self.rescale_factor
snake_case : List[str] = do_normalize if do_normalize is not None else self.do_normalize
snake_case : Union[str, Any] = image_mean if image_mean is not None else self.image_mean
snake_case : str = image_std if image_std is not None else self.image_std
snake_case : Union[str, Any] = do_convert_rgb if do_convert_rgb is not None else self.do_convert_rgb
snake_case : int = make_list_of_images(snake_case__ )
if not valid_images(snake_case__ ):
raise ValueError(
"Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "
"torch.Tensor, tf.Tensor or jax.ndarray." )
if do_resize and size is None:
raise ValueError("Size must be specified if do_resize is True." )
if do_center_crop and crop_size is None:
raise ValueError("Crop size must be specified if do_center_crop is True." )
if do_rescale and rescale_factor is None:
raise ValueError("Rescale factor must be specified if do_rescale is True." )
if do_normalize and (image_mean is None or image_std is None):
raise ValueError("Image mean and std must be specified if do_normalize is True." )
# PIL RGBA images are converted to RGB
if do_convert_rgb:
snake_case : Union[str, Any] = [convert_to_rgb(snake_case__ ) for image in images]
# All transformations expect numpy arrays.
snake_case : Optional[int] = [to_numpy_array(snake_case__ ) for image in images]
if do_resize:
snake_case : Optional[Any] = [self.resize(image=snake_case__ , size=snake_case__ , resample=snake_case__ ) for image in images]
if do_center_crop:
snake_case : Tuple = [self.center_crop(image=snake_case__ , size=snake_case__ ) for image in images]
if do_rescale:
snake_case : Optional[Any] = [self.rescale(image=snake_case__ , scale=snake_case__ ) for image in images]
if do_normalize:
snake_case : int = [self.normalize(image=snake_case__ , mean=snake_case__ , std=snake_case__ ) for image in images]
snake_case : str = [to_channel_dimension_format(snake_case__ , snake_case__ ) for image in images]
snake_case : Optional[Any] = {"pixel_values": images}
return BatchFeature(data=snake_case__ , tensor_type=snake_case__ )
| 59
|
import gc
import unittest
from parameterized import parameterized
from diffusers import FlaxUNetaDConditionModel
from diffusers.utils import is_flax_available
from diffusers.utils.testing_utils import load_hf_numpy, require_flax, slow
if is_flax_available():
import jax
import jax.numpy as jnp
@slow
@require_flax
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, Any] , snake_case__ : List[str] ) -> List[str]:
'''simple docstring'''
return f"""gaussian_noise_s={seed}_shape={'_'.join([str(snake_case__ ) for s in shape] )}.npy"""
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> int:
'''simple docstring'''
super().tearDown()
gc.collect()
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Optional[Any]=0 , snake_case__ : Any=(4, 4, 64, 64) , snake_case__ : List[Any]=False ) -> int:
'''simple docstring'''
snake_case : Optional[Any] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Optional[int] = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return image
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Tuple=False , snake_case__ : List[Any]="CompVis/stable-diffusion-v1-4" ) -> List[Any]:
'''simple docstring'''
snake_case : List[str] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : str = "bf16" if fpaa else None
snake_case , snake_case : Optional[int] = FlaxUNetaDConditionModel.from_pretrained(
snake_case__ , subfolder="unet" , dtype=snake_case__ , revision=snake_case__ )
return model, params
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any]=0 , snake_case__ : Union[str, Any]=(4, 77, 7_68) , snake_case__ : Dict=False ) -> List[str]:
'''simple docstring'''
snake_case : Any = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Any = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return hidden_states
@parameterized.expand(
[
# fmt: off
[83, 4, [-0.2323, -0.1304, 0.0813, -0.3093, -0.0919, -0.1571, -0.1125, -0.5806]],
[17, 0.55, [-0.0831, -0.2443, 0.0901, -0.0919, 0.3396, 0.0103, -0.3743, 0.0701]],
[8, 0.89, [-0.4863, 0.0859, 0.0875, -0.1658, 0.9199, -0.0114, 0.4839, 0.4639]],
[3, 10_00, [-0.5649, 0.2402, -0.5518, 0.1248, 1.1328, -0.2443, -0.0325, -1.0078]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : int , snake_case__ : Optional[int] , snake_case__ : Dict ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : List[str] = self.get_unet_model(model_id="CompVis/stable-diffusion-v1-4" , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_latents(snake_case__ , fpaa=snake_case__ )
snake_case : List[str] = self.get_encoder_hidden_states(snake_case__ , fpaa=snake_case__ )
snake_case : Dict = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : Optional[Any] = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Optional[int] = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, in the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[83, 4, [0.1514, 0.0807, 0.1624, 0.1016, -0.1896, 0.0263, 0.0677, 0.2310]],
[17, 0.55, [0.1164, -0.0216, 0.0170, 0.1589, -0.3120, 0.1005, -0.0581, -0.1458]],
[8, 0.89, [-0.1758, -0.0169, 0.1004, -0.1411, 0.1312, 0.1103, -0.1996, 0.2139]],
[3, 10_00, [0.1214, 0.0352, -0.0731, -0.1562, -0.0994, -0.0906, -0.2340, -0.0539]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[int] , snake_case__ : str , snake_case__ : Tuple ) -> str:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.get_unet_model(model_id="stabilityai/stable-diffusion-2" , fpaa=snake_case__ )
snake_case : List[str] = self.get_latents(snake_case__ , shape=(4, 4, 96, 96) , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_encoder_hidden_states(snake_case__ , shape=(4, 77, 10_24) , fpaa=snake_case__ )
snake_case : Optional[int] = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : int = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Dict = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, on the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
| 59
| 1
|
from __future__ import annotations
import collections
import pprint
from pathlib import Path
def UpperCamelCase ( __lowerCamelCase : str ):
return "".join(sorted(__lowerCamelCase ) )
def UpperCamelCase ( __lowerCamelCase : str ):
return word_by_signature[signature(__lowerCamelCase )]
__lowerCamelCase = Path(__file__).parent.joinpath("""words.txt""").read_text(encoding="""utf-8""")
__lowerCamelCase = sorted({word.strip().lower() for word in data.splitlines()})
__lowerCamelCase = collections.defaultdict(list)
for word in word_list:
word_by_signature[signature(word)].append(word)
if __name__ == "__main__":
__lowerCamelCase = {word: anagram(word) for word in word_list if len(anagram(word)) > 1}
with open("""anagrams.txt""", """w""") as file:
file.write("""all_anagrams = \n """)
file.write(pprint.pformat(all_anagrams))
| 59
|
import argparse
import re
from typing import Dict
import torch
from datasets import Audio, Dataset, load_dataset, load_metric
from transformers import AutoFeatureExtractor, pipeline
def UpperCamelCase ( __lowerCamelCase : Dataset , __lowerCamelCase : Dict[str, str] ):
snake_case : int = args.log_outputs
snake_case : Dict = "_".join(args.dataset.split("/" ) + [args.config, args.split] )
# load metric
snake_case : List[str] = load_metric("wer" )
snake_case : Tuple = load_metric("cer" )
# compute metrics
snake_case : List[Any] = wer.compute(references=result["target"] , predictions=result["prediction"] )
snake_case : int = cer.compute(references=result["target"] , predictions=result["prediction"] )
# print & log results
snake_case : int = f"""WER: {wer_result}\nCER: {cer_result}"""
print(__lowerCamelCase )
with open(f"""{dataset_id}_eval_results.txt""" , "w" ) as f:
f.write(__lowerCamelCase )
# log all results in text file. Possibly interesting for analysis
if log_outputs is not None:
snake_case : int = f"""log_{dataset_id}_predictions.txt"""
snake_case : List[Any] = f"""log_{dataset_id}_targets.txt"""
with open(__lowerCamelCase , "w" ) as p, open(__lowerCamelCase , "w" ) as t:
# mapping function to write output
def write_to_file(__lowerCamelCase : str , __lowerCamelCase : Optional[int] ):
p.write(f"""{i}""" + "\n" )
p.write(batch["prediction"] + "\n" )
t.write(f"""{i}""" + "\n" )
t.write(batch["target"] + "\n" )
result.map(__lowerCamelCase , with_indices=__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : List[Any] = "[,?.!\-\;\:\"“%‘”�—’…–]" # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
snake_case : List[Any] = re.sub(__lowerCamelCase , "" , text.lower() )
# In addition, we can normalize the target text, e.g. removing new lines characters etc...
# note that order is important here!
snake_case : Optional[Any] = ["\n\n", "\n", " ", " "]
for t in token_sequences_to_ignore:
snake_case : Dict = " ".join(text.split(__lowerCamelCase ) )
return text
def UpperCamelCase ( __lowerCamelCase : int ):
# load dataset
snake_case : str = load_dataset(args.dataset , args.config , split=args.split , use_auth_token=__lowerCamelCase )
# for testing: only process the first two examples as a test
# dataset = dataset.select(range(10))
# load processor
snake_case : List[Any] = AutoFeatureExtractor.from_pretrained(args.model_id )
snake_case : Union[str, Any] = feature_extractor.sampling_rate
# resample audio
snake_case : Union[str, Any] = dataset.cast_column("audio" , Audio(sampling_rate=__lowerCamelCase ) )
# load eval pipeline
if args.device is None:
snake_case : List[str] = 0 if torch.cuda.is_available() else -1
snake_case : str = pipeline("automatic-speech-recognition" , model=args.model_id , device=args.device )
# map function to decode audio
def map_to_pred(__lowerCamelCase : int ):
snake_case : Dict = asr(
batch["audio"]["array"] , chunk_length_s=args.chunk_length_s , stride_length_s=args.stride_length_s )
snake_case : str = prediction["text"]
snake_case : Tuple = normalize_text(batch["sentence"] )
return batch
# run inference on all examples
snake_case : Dict = dataset.map(__lowerCamelCase , remove_columns=dataset.column_names )
# compute and log_results
# do not change function below
log_results(__lowerCamelCase , __lowerCamelCase )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
"""--model_id""", type=str, required=True, help="""Model identifier. Should be loadable with 🤗 Transformers"""
)
parser.add_argument(
"""--dataset""",
type=str,
required=True,
help="""Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets""",
)
parser.add_argument(
"""--config""", type=str, required=True, help="""Config of the dataset. *E.g.* `'en'` for Common Voice"""
)
parser.add_argument("""--split""", type=str, required=True, help="""Split of the dataset. *E.g.* `'test'`""")
parser.add_argument(
"""--chunk_length_s""", type=float, default=None, help="""Chunk length in seconds. Defaults to 5 seconds."""
)
parser.add_argument(
"""--stride_length_s""", type=float, default=None, help="""Stride of the audio chunks. Defaults to 1 second."""
)
parser.add_argument(
"""--log_outputs""", action="""store_true""", help="""If defined, write outputs to log file for analysis."""
)
parser.add_argument(
"""--device""",
type=int,
default=None,
help="""The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.""",
)
__lowerCamelCase = parser.parse_args()
main(args)
| 59
| 1
|
from itertools import product
from cva import COLOR_BGR2GRAY, cvtColor, imread, imshow, waitKey
from numpy import dot, exp, mgrid, pi, ravel, square, uinta, zeros
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Dict ):
snake_case : int = k_size // 2
snake_case , snake_case : Tuple = mgrid[0 - center : k_size - center, 0 - center : k_size - center]
snake_case : Optional[int] = 1 / (2 * pi * sigma) * exp(-(square(__lowerCamelCase ) + square(__lowerCamelCase )) / (2 * square(__lowerCamelCase )) )
return g
def UpperCamelCase ( __lowerCamelCase : Any , __lowerCamelCase : Any , __lowerCamelCase : Dict ):
snake_case , snake_case : List[str] = image.shape[0], image.shape[1]
# dst image height and width
snake_case : Union[str, Any] = height - k_size + 1
snake_case : Any = width - k_size + 1
# im2col, turn the k_size*k_size pixels into a row and np.vstack all rows
snake_case : Optional[int] = zeros((dst_height * dst_width, k_size * k_size) )
snake_case : List[str] = 0
for i, j in product(range(__lowerCamelCase ) , range(__lowerCamelCase ) ):
snake_case : Tuple = ravel(image[i : i + k_size, j : j + k_size] )
snake_case : List[Any] = window
row += 1
# turn the kernel into shape(k*k, 1)
snake_case : Tuple = gen_gaussian_kernel(__lowerCamelCase , __lowerCamelCase )
snake_case : List[str] = ravel(__lowerCamelCase )
# reshape and get the dst image
snake_case : int = dot(__lowerCamelCase , __lowerCamelCase ).reshape(__lowerCamelCase , __lowerCamelCase ).astype(__lowerCamelCase )
return dst
if __name__ == "__main__":
# read original image
__lowerCamelCase = imread(R"""../image_data/lena.jpg""")
# turn image in gray scale value
__lowerCamelCase = cvtColor(img, COLOR_BGR2GRAY)
# get values with two different mask size
__lowerCamelCase = gaussian_filter(gray, 3, sigma=1)
__lowerCamelCase = gaussian_filter(gray, 5, sigma=0.8)
# show result images
imshow("""gaussian filter with 3x3 mask""", gaussianaxa)
imshow("""gaussian filter with 5x5 mask""", gaussianaxa)
waitKey()
| 59
|
from typing import Optional, Tuple, Union
import flax
import flax.linen as nn
import jax
import jax.numpy as jnp
from flax.core.frozen_dict import FrozenDict
from ..configuration_utils import ConfigMixin, flax_register_to_config
from ..utils import BaseOutput
from .embeddings_flax import FlaxTimestepEmbedding, FlaxTimesteps
from .modeling_flax_utils import FlaxModelMixin
from .unet_ad_blocks_flax import (
FlaxCrossAttnDownBlockaD,
FlaxCrossAttnUpBlockaD,
FlaxDownBlockaD,
FlaxUNetMidBlockaDCrossAttn,
FlaxUpBlockaD,
)
@flax.struct.dataclass
class UpperCAmelCase ( A_ ):
A__ : jnp.ndarray
@flax_register_to_config
class UpperCAmelCase ( nn.Module ,A_ ,A_ ):
A__ : int = 32
A__ : int = 4
A__ : int = 4
A__ : Tuple[str] = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
A__ : Tuple[str] = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
A__ : Union[bool, Tuple[bool]] = False
A__ : Tuple[int] = (3_20, 6_40, 12_80, 12_80)
A__ : int = 2
A__ : Union[int, Tuple[int]] = 8
A__ : Optional[Union[int, Tuple[int]]] = None
A__ : int = 12_80
A__ : float = 0.0
A__ : bool = False
A__ : jnp.dtype = jnp.floataa
A__ : bool = True
A__ : int = 0
A__ : bool = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : jax.random.KeyArray ) -> FrozenDict:
'''simple docstring'''
snake_case : Dict = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case : Any = jnp.zeros(snake_case__ , dtype=jnp.floataa )
snake_case : List[str] = jnp.ones((1,) , dtype=jnp.intaa )
snake_case : str = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case , snake_case : Optional[int] = jax.random.split(snake_case__ )
snake_case : Union[str, Any] = {"params": params_rng, "dropout": dropout_rng}
return self.init(snake_case__ , snake_case__ , snake_case__ , snake_case__ )["params"]
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : str = self.block_out_channels
snake_case : Optional[Any] = block_out_channels[0] * 4
if self.num_attention_heads is not None:
raise ValueError(
"At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19." )
# If `num_attention_heads` is not defined (which is the case for most models)
# it will default to `attention_head_dim`. This looks weird upon first reading it and it is.
# The reason for this behavior is to correct for incorrectly named variables that were introduced
# when this library was created. The incorrect naming was only discovered much later in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131
# Changing `attention_head_dim` to `num_attention_heads` for 40,000+ configurations is too backwards breaking
# which is why we correct for the naming here.
snake_case : Tuple = self.num_attention_heads or self.attention_head_dim
# input
snake_case : Tuple = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case : Union[str, Any] = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case : Dict = FlaxTimestepEmbedding(snake_case__ , dtype=self.dtype )
snake_case : List[str] = self.only_cross_attention
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (only_cross_attention,) * len(self.down_block_types )
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case : List[Any] = []
snake_case : Optional[int] = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case : List[Any] = output_channel
snake_case : Dict = block_out_channels[i]
snake_case : Optional[Any] = i == len(snake_case__ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case : List[Any] = FlaxCrossAttnDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , num_attention_heads=num_attention_heads[i] , add_downsample=not is_final_block , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Union[str, Any] = FlaxDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , add_downsample=not is_final_block , dtype=self.dtype , )
down_blocks.append(snake_case__ )
snake_case : Dict = down_blocks
# mid
snake_case : Optional[int] = FlaxUNetMidBlockaDCrossAttn(
in_channels=block_out_channels[-1] , dropout=self.dropout , num_attention_heads=num_attention_heads[-1] , use_linear_projection=self.use_linear_projection , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
# up
snake_case : Optional[Any] = []
snake_case : Optional[int] = list(reversed(snake_case__ ) )
snake_case : Dict = list(reversed(snake_case__ ) )
snake_case : Tuple = list(reversed(snake_case__ ) )
snake_case : Optional[Any] = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case : Optional[int] = output_channel
snake_case : List[Any] = reversed_block_out_channels[i]
snake_case : Union[str, Any] = reversed_block_out_channels[min(i + 1 , len(snake_case__ ) - 1 )]
snake_case : int = i == len(snake_case__ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case : Any = FlaxCrossAttnUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , num_attention_heads=reversed_num_attention_heads[i] , add_upsample=not is_final_block , dropout=self.dropout , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Optional[int] = FlaxUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , add_upsample=not is_final_block , dropout=self.dropout , dtype=self.dtype , )
up_blocks.append(snake_case__ )
snake_case : Optional[int] = output_channel
snake_case : Tuple = up_blocks
# out
snake_case : Optional[int] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case : List[str] = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__(self : Dict , snake_case__ : Dict , snake_case__ : Dict , snake_case__ : Optional[int] , snake_case__ : Tuple=None , snake_case__ : Union[str, Any]=None , snake_case__ : bool = True , snake_case__ : bool = False , ) -> Union[FlaxUNetaDConditionOutput, Tuple]:
'''simple docstring'''
if not isinstance(snake_case__ , jnp.ndarray ):
snake_case : List[Any] = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(snake_case__ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case : Any = timesteps.astype(dtype=jnp.floataa )
snake_case : int = jnp.expand_dims(snake_case__ , 0 )
snake_case : str = self.time_proj(snake_case__ )
snake_case : str = self.time_embedding(snake_case__ )
# 2. pre-process
snake_case : int = jnp.transpose(snake_case__ , (0, 2, 3, 1) )
snake_case : List[Any] = self.conv_in(snake_case__ )
# 3. down
snake_case : Optional[int] = (sample,)
for down_block in self.down_blocks:
if isinstance(snake_case__ , snake_case__ ):
snake_case , snake_case : List[Any] = down_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
else:
snake_case , snake_case : str = down_block(snake_case__ , snake_case__ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case : Tuple = ()
for down_block_res_sample, down_block_additional_residual in zip(
snake_case__ , snake_case__ ):
down_block_res_sample += down_block_additional_residual
new_down_block_res_samples += (down_block_res_sample,)
snake_case : Optional[int] = new_down_block_res_samples
# 4. mid
snake_case : Optional[int] = self.mid_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
if mid_block_additional_residual is not None:
sample += mid_block_additional_residual
# 5. up
for up_block in self.up_blocks:
snake_case : int = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case : Optional[Any] = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(snake_case__ , snake_case__ ):
snake_case : Optional[Any] = up_block(
snake_case__ , temb=snake_case__ , encoder_hidden_states=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train , )
else:
snake_case : Dict = up_block(snake_case__ , temb=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train )
# 6. post-process
snake_case : List[str] = self.conv_norm_out(snake_case__ )
snake_case : Any = nn.silu(snake_case__ )
snake_case : Optional[int] = self.conv_out(snake_case__ )
snake_case : Union[str, Any] = jnp.transpose(snake_case__ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=snake_case__ )
| 59
| 1
|
import gc
import unittest
import numpy as np
import torch
from torch.backends.cuda import sdp_kernel
from diffusers import (
CMStochasticIterativeScheduler,
ConsistencyModelPipeline,
UNetaDModel,
)
from diffusers.utils import randn_tensor, slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_a, require_torch_gpu
from ..pipeline_params import UNCONDITIONAL_IMAGE_GENERATION_BATCH_PARAMS, UNCONDITIONAL_IMAGE_GENERATION_PARAMS
from ..test_pipelines_common import PipelineTesterMixin
enable_full_determinism()
class UpperCAmelCase ( A_ ,unittest.TestCase ):
A__ : Union[str, Any] = ConsistencyModelPipeline
A__ : Union[str, Any] = UNCONDITIONAL_IMAGE_GENERATION_PARAMS
A__ : Any = UNCONDITIONAL_IMAGE_GENERATION_BATCH_PARAMS
# Override required_optional_params to remove num_images_per_prompt
A__ : List[str] = frozenset(
[
"num_inference_steps",
"generator",
"latents",
"output_type",
"return_dict",
"callback",
"callback_steps",
] )
@property
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = UNetaDModel.from_pretrained(
"diffusers/consistency-models-test" , subfolder="test_unet" , )
return unet
@property
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : List[Any] = UNetaDModel.from_pretrained(
"diffusers/consistency-models-test" , subfolder="test_unet_class_cond" , )
return unet
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any]=False ) -> List[Any]:
'''simple docstring'''
if class_cond:
snake_case : List[Any] = self.dummy_cond_unet
else:
snake_case : Optional[int] = self.dummy_uncond_unet
# Default to CM multistep sampler
snake_case : Any = CMStochasticIterativeScheduler(
num_train_timesteps=40 , sigma_min=0.002 , sigma_max=80.0 , )
snake_case : List[Any] = {
"unet": unet,
"scheduler": scheduler,
}
return components
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Any , snake_case__ : Optional[Any]=0 ) -> Optional[Any]:
'''simple docstring'''
if str(snake_case__ ).startswith("mps" ):
snake_case : List[str] = torch.manual_seed(snake_case__ )
else:
snake_case : Any = torch.Generator(device=snake_case__ ).manual_seed(snake_case__ )
snake_case : Optional[int] = {
"batch_size": 1,
"num_inference_steps": None,
"timesteps": [22, 0],
"generator": generator,
"output_type": "np",
}
return inputs
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = "cpu" # ensure determinism for the device-dependent torch.Generator
snake_case : Dict = self.get_dummy_components()
snake_case : Optional[Any] = ConsistencyModelPipeline(**snake_case__ )
snake_case : str = pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : Union[str, Any] = self.get_dummy_inputs(snake_case__ )
snake_case : Optional[int] = pipe(**snake_case__ ).images
assert image.shape == (1, 32, 32, 3)
snake_case : Union[str, Any] = image[0, -3:, -3:, -1]
snake_case : Dict = np.array([0.3572, 0.6273, 0.4031, 0.3961, 0.4321, 0.5730, 0.5266, 0.4780, 0.5004] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case : List[Any] = "cpu" # ensure determinism for the device-dependent torch.Generator
snake_case : int = self.get_dummy_components(class_cond=snake_case__ )
snake_case : Any = ConsistencyModelPipeline(**snake_case__ )
snake_case : List[Any] = pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : Union[str, Any] = self.get_dummy_inputs(snake_case__ )
snake_case : Dict = 0
snake_case : List[Any] = pipe(**snake_case__ ).images
assert image.shape == (1, 32, 32, 3)
snake_case : Optional[Any] = image[0, -3:, -3:, -1]
snake_case : Optional[Any] = np.array([0.3572, 0.6273, 0.4031, 0.3961, 0.4321, 0.5730, 0.5266, 0.4780, 0.5004] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : List[Any] = "cpu" # ensure determinism for the device-dependent torch.Generator
snake_case : Union[str, Any] = self.get_dummy_components()
snake_case : Optional[Any] = ConsistencyModelPipeline(**snake_case__ )
snake_case : Optional[int] = pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : Dict = self.get_dummy_inputs(snake_case__ )
snake_case : List[Any] = 1
snake_case : Dict = None
snake_case : Tuple = pipe(**snake_case__ ).images
assert image.shape == (1, 32, 32, 3)
snake_case : Any = image[0, -3:, -3:, -1]
snake_case : List[str] = np.array([0.5004, 0.5004, 0.4994, 0.5008, 0.4976, 0.5018, 0.4990, 0.4982, 0.4987] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = "cpu" # ensure determinism for the device-dependent torch.Generator
snake_case : str = self.get_dummy_components(class_cond=snake_case__ )
snake_case : List[Any] = ConsistencyModelPipeline(**snake_case__ )
snake_case : Union[str, Any] = pipe.to(snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : List[str] = self.get_dummy_inputs(snake_case__ )
snake_case : Any = 1
snake_case : Optional[int] = None
snake_case : List[Any] = 0
snake_case : int = pipe(**snake_case__ ).images
assert image.shape == (1, 32, 32, 3)
snake_case : Union[str, Any] = image[0, -3:, -3:, -1]
snake_case : Dict = np.array([0.5004, 0.5004, 0.4994, 0.5008, 0.4976, 0.5018, 0.4990, 0.4982, 0.4987] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
@slow
@require_torch_gpu
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Tuple:
'''simple docstring'''
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : Union[str, Any]=0 , snake_case__ : Optional[Any]=False , snake_case__ : str="cpu" , snake_case__ : Optional[int]=torch.floataa , snake_case__ : Optional[Any]=(1, 3, 64, 64) ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = torch.manual_seed(snake_case__ )
snake_case : List[str] = {
"num_inference_steps": None,
"timesteps": [22, 0],
"class_labels": 0,
"generator": generator,
"output_type": "np",
}
if get_fixed_latents:
snake_case : Tuple = self.get_fixed_latents(seed=snake_case__ , device=snake_case__ , dtype=snake_case__ , shape=snake_case__ )
snake_case : int = latents
return inputs
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : List[Any]=0 , snake_case__ : Union[str, Any]="cpu" , snake_case__ : int=torch.floataa , snake_case__ : Any=(1, 3, 64, 64) ) -> int:
'''simple docstring'''
if type(snake_case__ ) == str:
snake_case : Optional[Any] = torch.device(snake_case__ )
snake_case : List[Any] = torch.Generator(device=snake_case__ ).manual_seed(snake_case__ )
snake_case : List[Any] = randn_tensor(snake_case__ , generator=snake_case__ , device=snake_case__ , dtype=snake_case__ )
return latents
def _SCREAMING_SNAKE_CASE (self : Any ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Optional[Any] = UNetaDModel.from_pretrained("diffusers/consistency_models" , subfolder="diffusers_cd_imagenet64_l2" )
snake_case : Tuple = CMStochasticIterativeScheduler(
num_train_timesteps=40 , sigma_min=0.002 , sigma_max=80.0 , )
snake_case : List[Any] = ConsistencyModelPipeline(unet=snake_case__ , scheduler=snake_case__ )
pipe.to(torch_device=snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : str = self.get_inputs()
snake_case : Optional[int] = pipe(**snake_case__ ).images
assert image.shape == (1, 64, 64, 3)
snake_case : Optional[Any] = image[0, -3:, -3:, -1]
snake_case : List[str] = np.array([0.0888, 0.0881, 0.0666, 0.0479, 0.0292, 0.0195, 0.0201, 0.0163, 0.0254] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 2e-2
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = UNetaDModel.from_pretrained("diffusers/consistency_models" , subfolder="diffusers_cd_imagenet64_l2" )
snake_case : Tuple = CMStochasticIterativeScheduler(
num_train_timesteps=40 , sigma_min=0.002 , sigma_max=80.0 , )
snake_case : Dict = ConsistencyModelPipeline(unet=snake_case__ , scheduler=snake_case__ )
pipe.to(torch_device=snake_case__ )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : int = self.get_inputs()
snake_case : Union[str, Any] = 1
snake_case : Optional[int] = None
snake_case : int = pipe(**snake_case__ ).images
assert image.shape == (1, 64, 64, 3)
snake_case : Dict = image[0, -3:, -3:, -1]
snake_case : Union[str, Any] = np.array([0.0340, 0.0152, 0.0063, 0.0267, 0.0221, 0.0107, 0.0416, 0.0186, 0.0217] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 2e-2
@require_torch_a
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : str = UNetaDModel.from_pretrained("diffusers/consistency_models" , subfolder="diffusers_cd_imagenet64_l2" )
snake_case : Optional[int] = CMStochasticIterativeScheduler(
num_train_timesteps=40 , sigma_min=0.002 , sigma_max=80.0 , )
snake_case : Optional[int] = ConsistencyModelPipeline(unet=snake_case__ , scheduler=snake_case__ )
pipe.to(torch_device=snake_case__ , torch_dtype=torch.floataa )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : str = self.get_inputs(get_fixed_latents=snake_case__ , device=snake_case__ )
# Ensure usage of flash attention in torch 2.0
with sdp_kernel(enable_flash=snake_case__ , enable_math=snake_case__ , enable_mem_efficient=snake_case__ ):
snake_case : List[Any] = pipe(**snake_case__ ).images
assert image.shape == (1, 64, 64, 3)
snake_case : Optional[Any] = image[0, -3:, -3:, -1]
snake_case : Optional[Any] = np.array([0.1875, 0.1428, 0.1289, 0.2151, 0.2092, 0.1477, 0.1877, 0.1641, 0.1353] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
@require_torch_a
def _SCREAMING_SNAKE_CASE (self : str ) -> List[Any]:
'''simple docstring'''
snake_case : Optional[int] = UNetaDModel.from_pretrained("diffusers/consistency_models" , subfolder="diffusers_cd_imagenet64_l2" )
snake_case : str = CMStochasticIterativeScheduler(
num_train_timesteps=40 , sigma_min=0.002 , sigma_max=80.0 , )
snake_case : str = ConsistencyModelPipeline(unet=snake_case__ , scheduler=snake_case__ )
pipe.to(torch_device=snake_case__ , torch_dtype=torch.floataa )
pipe.set_progress_bar_config(disable=snake_case__ )
snake_case : Dict = self.get_inputs(get_fixed_latents=snake_case__ , device=snake_case__ )
snake_case : Any = 1
snake_case : Union[str, Any] = None
# Ensure usage of flash attention in torch 2.0
with sdp_kernel(enable_flash=snake_case__ , enable_math=snake_case__ , enable_mem_efficient=snake_case__ ):
snake_case : Union[str, Any] = pipe(**snake_case__ ).images
assert image.shape == (1, 64, 64, 3)
snake_case : int = image[0, -3:, -3:, -1]
snake_case : List[Any] = np.array([0.1663, 0.1948, 0.2275, 0.1680, 0.1204, 0.1245, 0.1858, 0.1338, 0.2095] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1e-3
| 59
|
__lowerCamelCase = {
"joule": 1.0,
"kilojoule": 10_00,
"megajoule": 1_00_00_00,
"gigajoule": 10_00_00_00_00,
"wattsecond": 1.0,
"watthour": 36_00,
"kilowatthour": 3_60_00_00,
"newtonmeter": 1.0,
"calorie_nutr": 41_86.8,
"kilocalorie_nutr": 4_18_68_00.00,
"electronvolt": 1.602_176_634e-19,
"britishthermalunit_it": 10_55.0_55_85,
"footpound": 1.35_5818,
}
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : float ):
if to_type not in ENERGY_CONVERSION or from_type not in ENERGY_CONVERSION:
snake_case : List[Any] = (
f"""Incorrect 'from_type' or 'to_type' value: {from_type!r}, {to_type!r}\n"""
f"""Valid values are: {', '.join(__lowerCamelCase )}"""
)
raise ValueError(__lowerCamelCase )
return value * ENERGY_CONVERSION[from_type] / ENERGY_CONVERSION[to_type]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import logging
from .tokenization_mobilebert import MobileBertTokenizer
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {"""mobilebert-uncased""": """https://huggingface.co/google/mobilebert-uncased/resolve/main/vocab.txt"""},
"""tokenizer_file""": {
"""mobilebert-uncased""": """https://huggingface.co/google/mobilebert-uncased/resolve/main/tokenizer.json"""
},
}
__lowerCamelCase = {"""mobilebert-uncased""": 5_12}
__lowerCamelCase = {}
class UpperCAmelCase ( A_ ):
A__ : Union[str, Any] = VOCAB_FILES_NAMES
A__ : Dict = PRETRAINED_VOCAB_FILES_MAP
A__ : Any = PRETRAINED_INIT_CONFIGURATION
A__ : int = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : Optional[Any] = MobileBertTokenizer
def __init__(self : str , snake_case__ : List[str]=None , snake_case__ : int=None , snake_case__ : List[str]=True , snake_case__ : str="[UNK]" , snake_case__ : int="[SEP]" , snake_case__ : Optional[Any]="[PAD]" , snake_case__ : Any="[CLS]" , snake_case__ : Tuple="[MASK]" , snake_case__ : Tuple=True , snake_case__ : str=None , **snake_case__ : Optional[Any] , ) -> Optional[int]:
'''simple docstring'''
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
snake_case : Dict = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , snake_case__ ) != do_lower_case
or normalizer_state.get("strip_accents" , snake_case__ ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , snake_case__ ) != tokenize_chinese_chars
):
snake_case : Tuple = getattr(snake_case__ , normalizer_state.pop("type" ) )
snake_case : Optional[Any] = do_lower_case
snake_case : Dict = strip_accents
snake_case : Union[str, Any] = tokenize_chinese_chars
snake_case : Tuple = normalizer_class(**snake_case__ )
snake_case : Optional[Any] = do_lower_case
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Union[str, Any] , snake_case__ : List[str]=None ) -> Union[str, Any]:
'''simple docstring'''
snake_case : str = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Dict = [self.sep_token_id]
snake_case : Dict = [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 : List[str] , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
snake_case : List[Any] = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 59
|
import logging
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
import datasets
import numpy as np
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PreTrainedTokenizer,
TFAutoModelForSequenceClassification,
TFTrainer,
TFTrainingArguments,
)
from transformers.utils import logging as hf_logging
hf_logging.set_verbosity_info()
hf_logging.enable_default_handler()
hf_logging.enable_explicit_format()
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : PreTrainedTokenizer , __lowerCamelCase : int , __lowerCamelCase : Optional[int] = None , ):
snake_case : int = {}
if train_file is not None:
snake_case : List[Any] = [train_file]
if eval_file is not None:
snake_case : Optional[int] = [eval_file]
if test_file is not None:
snake_case : Any = [test_file]
snake_case : int = datasets.load_dataset("csv" , data_files=__lowerCamelCase )
snake_case : str = list(ds[list(files.keys() )[0]].features.keys() )
snake_case : int = features_name.pop(__lowerCamelCase )
snake_case : str = list(set(ds[list(files.keys() )[0]][label_name] ) )
snake_case : str = {label: i for i, label in enumerate(__lowerCamelCase )}
snake_case : List[Any] = tokenizer.model_input_names
snake_case : List[Any] = {}
if len(__lowerCamelCase ) == 1:
for k in files.keys():
snake_case : Tuple = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
example[features_name[0]] , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" ) , batched=__lowerCamelCase , )
elif len(__lowerCamelCase ) == 2:
for k in files.keys():
snake_case : List[Any] = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
(example[features_name[0]], example[features_name[1]]) , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" , ) , batched=__lowerCamelCase , )
def gen_train():
for ex in transformed_ds[datasets.Split.TRAIN]:
snake_case : Dict = {k: v for k, v in ex.items() if k in input_names}
snake_case : Union[str, Any] = labelaid[ex[label_name]]
yield (d, label)
def gen_val():
for ex in transformed_ds[datasets.Split.VALIDATION]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : Any = labelaid[ex[label_name]]
yield (d, label)
def gen_test():
for ex in transformed_ds[datasets.Split.TEST]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : List[str] = labelaid[ex[label_name]]
yield (d, label)
snake_case : int = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TRAIN in transformed_ds
else None
)
if train_ds is not None:
snake_case : Optional[Any] = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) )
snake_case : Tuple = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.VALIDATION in transformed_ds
else None
)
if val_ds is not None:
snake_case : List[str] = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) )
snake_case : Optional[int] = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TEST in transformed_ds
else None
)
if test_ds is not None:
snake_case : str = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) )
return train_ds, val_ds, test_ds, labelaid
__lowerCamelCase = logging.getLogger(__name__)
@dataclass
class UpperCAmelCase :
A__ : int = field(metadata={"help": "Which column contains the label"} )
A__ : str = field(default=A_ ,metadata={"help": "The path of the training file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the development file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the test file"} )
A__ : int = field(
default=1_28 ,metadata={
"help": (
"The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
} ,)
A__ : bool = field(
default=A_ ,metadata={"help": "Overwrite the cached training and evaluation sets"} )
@dataclass
class UpperCAmelCase :
A__ : str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained config name or path if not the same as model_name"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"} )
A__ : bool = field(default=A_ ,metadata={"help": "Set this flag to use fast tokenization."} )
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
# or just modify its tokenizer_config.json.
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} ,)
def UpperCamelCase ( ):
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
snake_case : List[Any] = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) )
snake_case , snake_case , snake_case : int = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir )
and os.listdir(training_args.output_dir )
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
f"""Output directory ({training_args.output_dir}) already exists and is not empty. Use"""
" --overwrite_output_dir to overcome." )
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(
f"""n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, """
f"""16-bits training: {training_args.fpaa}""" )
logger.info(f"""Training/evaluation parameters {training_args}""" )
# Load pretrained model and tokenizer
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
snake_case : Tuple = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path , cache_dir=model_args.cache_dir , )
snake_case , snake_case , snake_case , snake_case : Tuple = get_tfds(
train_file=data_args.train_file , eval_file=data_args.dev_file , test_file=data_args.test_file , tokenizer=__lowerCamelCase , label_column_id=data_args.label_column_id , max_seq_length=data_args.max_seq_length , )
snake_case : Optional[Any] = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path , num_labels=len(__lowerCamelCase ) , labelaid=__lowerCamelCase , idalabel={id: label for label, id in labelaid.items()} , finetuning_task="text-classification" , cache_dir=model_args.cache_dir , )
with training_args.strategy.scope():
snake_case : int = TFAutoModelForSequenceClassification.from_pretrained(
model_args.model_name_or_path , from_pt=bool(".bin" in model_args.model_name_or_path ) , config=__lowerCamelCase , cache_dir=model_args.cache_dir , )
def compute_metrics(__lowerCamelCase : EvalPrediction ) -> Dict:
snake_case : Optional[int] = np.argmax(p.predictions , axis=1 )
return {"acc": (preds == p.label_ids).mean()}
# Initialize our Trainer
snake_case : int = TFTrainer(
model=__lowerCamelCase , args=__lowerCamelCase , train_dataset=__lowerCamelCase , eval_dataset=__lowerCamelCase , compute_metrics=__lowerCamelCase , )
# Training
if training_args.do_train:
trainer.train()
trainer.save_model()
tokenizer.save_pretrained(training_args.output_dir )
# Evaluation
snake_case : int = {}
if training_args.do_eval:
logger.info("*** Evaluate ***" )
snake_case : Any = trainer.evaluate()
snake_case : List[Any] = os.path.join(training_args.output_dir , "eval_results.txt" )
with open(__lowerCamelCase , "w" ) as writer:
logger.info("***** Eval results *****" )
for key, value in result.items():
logger.info(f""" {key} = {value}""" )
writer.write(f"""{key} = {value}\n""" )
results.update(__lowerCamelCase )
return results
if __name__ == "__main__":
main()
| 59
| 1
|
import argparse
import json
from dataclasses import dataclass, field
from functools import partial
from pathlib import Path
from typing import Callable, Dict, List, Tuple
import timm
import torch
import torch.nn as nn
from classy_vision.models.regnet import RegNet, RegNetParams, RegNetYaagf, RegNetYaagf, RegNetYaaagf
from huggingface_hub import cached_download, hf_hub_url
from torch import Tensor
from vissl.models.model_helpers import get_trunk_forward_outputs
from transformers import AutoImageProcessor, RegNetConfig, RegNetForImageClassification, RegNetModel
from transformers.utils import logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger()
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : List[nn.Module] = field(default_factory=A_ )
A__ : list = field(default_factory=A_ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : Tensor , snake_case__ : Tensor ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = len(list(m.modules() ) ) == 1 or isinstance(snake_case__ , nn.Convad ) or isinstance(snake_case__ , nn.BatchNormad )
if has_not_submodules:
self.traced.append(snake_case__ )
def __call__(self : List[Any] , snake_case__ : Tensor ) -> List[Any]:
'''simple docstring'''
for m in self.module.modules():
self.handles.append(m.register_forward_hook(self._forward_hook ) )
self.module(snake_case__ )
[x.remove() for x in self.handles]
return self
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[int]:
'''simple docstring'''
return list(filter(lambda snake_case__ : len(list(x.state_dict().keys() ) ) > 0 , self.traced ) )
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : nn.Module
A__ : int = 1
A__ : List = field(default_factory=A_ )
A__ : List = field(default_factory=A_ )
A__ : bool = True
def __call__(self : List[Any] , snake_case__ : Tensor ) -> Any:
'''simple docstring'''
snake_case : str = Tracker(self.dest )(snake_case__ ).parametrized
snake_case : Optional[int] = Tracker(self.src )(snake_case__ ).parametrized
snake_case : List[str] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.src_skip , snake_case__ ) )
snake_case : Optional[Any] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.dest_skip , snake_case__ ) )
if len(snake_case__ ) != len(snake_case__ ) and self.raise_if_mismatch:
raise Exception(
f"""Numbers of operations are different. Source module has {len(snake_case__ )} operations while"""
f""" destination module has {len(snake_case__ )}.""" )
for dest_m, src_m in zip(snake_case__ , snake_case__ ):
dest_m.load_state_dict(src_m.state_dict() )
if self.verbose == 1:
print(f"""Transfered from={src_m} to={dest_m}""" )
class UpperCAmelCase ( nn.Module ):
def __init__(self : Tuple , snake_case__ : nn.Module ) -> Optional[Any]:
'''simple docstring'''
super().__init__()
snake_case : List[Tuple[str, nn.Module]] = []
# - get the stem
feature_blocks.append(("conv1", model.stem) )
# - get all the feature blocks
for k, v in model.trunk_output.named_children():
assert k.startswith("block" ), f"""Unexpected layer name {k}"""
snake_case : Union[str, Any] = len(snake_case__ ) + 1
feature_blocks.append((f"""res{block_index}""", v) )
snake_case : Optional[Any] = nn.ModuleDict(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Tensor ) -> Dict:
'''simple docstring'''
return get_trunk_forward_outputs(
snake_case__ , out_feat_keys=snake_case__ , feature_blocks=self._feature_blocks , )
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> str:
'''simple docstring'''
snake_case : List[Any] = x.split("-" )
return x_split[0] + x_split[1] + "_" + "".join(x_split[2:] )
def __getitem__(self : Optional[int] , snake_case__ : str ) -> Callable[[], Tuple[nn.Module, Dict]]:
'''simple docstring'''
if x not in self:
snake_case : Dict = self.convert_name_to_timm(snake_case__ )
snake_case : Union[str, Any] = partial(lambda: (timm.create_model(snake_case__ , pretrained=snake_case__ ).eval(), None) )
else:
snake_case : List[str] = super().__getitem__(snake_case__ )
return val
class UpperCAmelCase ( A_ ):
def __getitem__(self : Dict , snake_case__ : str ) -> Callable[[], nn.Module]:
'''simple docstring'''
if "seer" in x and "in1k" not in x:
snake_case : str = RegNetModel
else:
snake_case : Optional[Any] = RegNetForImageClassification
return val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Tuple[str, str]] ):
for from_key, to_key in keys:
snake_case : str = from_state_dict[from_key].clone()
print(f"""Copied key={from_key} to={to_key}""" )
return to_state_dict
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : RegNetConfig , __lowerCamelCase : Path , __lowerCamelCase : bool = True , ):
print(f"""Converting {name}...""" )
with torch.no_grad():
snake_case , snake_case : int = from_model_func()
snake_case : str = our_model_func(__lowerCamelCase ).eval()
snake_case : int = ModuleTransfer(src=__lowerCamelCase , dest=__lowerCamelCase , raise_if_mismatch=__lowerCamelCase )
snake_case : Dict = torch.randn((1, 3, 224, 224) )
module_transfer(__lowerCamelCase )
if from_state_dict is not None:
snake_case : str = []
# for seer - in1k finetuned we have to manually copy the head
if "seer" in name and "in1k" in name:
snake_case : Tuple = [("0.clf.0.weight", "classifier.1.weight"), ("0.clf.0.bias", "classifier.1.bias")]
snake_case : Optional[Any] = manually_copy_vissl_head(__lowerCamelCase , our_model.state_dict() , __lowerCamelCase )
our_model.load_state_dict(__lowerCamelCase )
snake_case : Any = our_model(__lowerCamelCase , output_hidden_states=__lowerCamelCase )
snake_case : Union[str, Any] = (
our_outputs.logits if isinstance(__lowerCamelCase , __lowerCamelCase ) else our_outputs.last_hidden_state
)
snake_case : Union[str, Any] = from_model(__lowerCamelCase )
snake_case : Dict = from_output[-1] if type(__lowerCamelCase ) is list else from_output
# now since I don't want to use any config files, vissl seer model doesn't actually have an head, so let's just check the last hidden state
if "seer" in name and "in1k" in name:
snake_case : Any = our_outputs.hidden_states[-1]
assert torch.allclose(__lowerCamelCase , __lowerCamelCase ), "The model logits don't match the original one."
if push_to_hub:
our_model.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add model" , use_temp_dir=__lowerCamelCase , )
snake_case : List[str] = 224 if "seer" not in name else 384
# we can use the convnext one
snake_case : int = AutoImageProcessor.from_pretrained("facebook/convnext-base-224-22k-1k" , size=__lowerCamelCase )
image_processor.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add image processor" , use_temp_dir=__lowerCamelCase , )
print(f"""Pushed {name}""" )
def UpperCamelCase ( __lowerCamelCase : Path , __lowerCamelCase : str = None , __lowerCamelCase : bool = True ):
snake_case : Union[str, Any] = "imagenet-1k-id2label.json"
snake_case : List[str] = 1000
snake_case : List[str] = (1, num_labels)
snake_case : Any = "huggingface/label-files"
snake_case : List[str] = num_labels
snake_case : Optional[Any] = json.load(open(cached_download(hf_hub_url(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) ) , "r" ) )
snake_case : List[Any] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : str = idalabel
snake_case : List[Any] = {v: k for k, v in idalabel.items()}
snake_case : Dict = partial(__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase )
snake_case : Optional[Any] = {
"regnet-x-002": ImageNetPreTrainedConfig(
depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 , layer_type="x" ),
"regnet-x-004": ImageNetPreTrainedConfig(
depths=[1, 2, 7, 12] , hidden_sizes=[32, 64, 160, 384] , groups_width=16 , layer_type="x" ),
"regnet-x-006": ImageNetPreTrainedConfig(
depths=[1, 3, 5, 7] , hidden_sizes=[48, 96, 240, 528] , groups_width=24 , layer_type="x" ),
"regnet-x-008": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 5] , hidden_sizes=[64, 128, 288, 672] , groups_width=16 , layer_type="x" ),
"regnet-x-016": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 2] , hidden_sizes=[72, 168, 408, 912] , groups_width=24 , layer_type="x" ),
"regnet-x-032": ImageNetPreTrainedConfig(
depths=[2, 6, 15, 2] , hidden_sizes=[96, 192, 432, 1008] , groups_width=48 , layer_type="x" ),
"regnet-x-040": ImageNetPreTrainedConfig(
depths=[2, 5, 14, 2] , hidden_sizes=[80, 240, 560, 1360] , groups_width=40 , layer_type="x" ),
"regnet-x-064": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 392, 784, 1624] , groups_width=56 , layer_type="x" ),
"regnet-x-080": ImageNetPreTrainedConfig(
depths=[2, 5, 15, 1] , hidden_sizes=[80, 240, 720, 1920] , groups_width=120 , layer_type="x" ),
"regnet-x-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 , layer_type="x" ),
"regnet-x-160": ImageNetPreTrainedConfig(
depths=[2, 6, 13, 1] , hidden_sizes=[256, 512, 896, 2048] , groups_width=128 , layer_type="x" ),
"regnet-x-320": ImageNetPreTrainedConfig(
depths=[2, 7, 13, 1] , hidden_sizes=[336, 672, 1344, 2520] , groups_width=168 , layer_type="x" ),
# y variant
"regnet-y-002": ImageNetPreTrainedConfig(depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 ),
"regnet-y-004": ImageNetPreTrainedConfig(
depths=[1, 3, 6, 6] , hidden_sizes=[48, 104, 208, 440] , groups_width=8 ),
"regnet-y-006": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 4] , hidden_sizes=[48, 112, 256, 608] , groups_width=16 ),
"regnet-y-008": ImageNetPreTrainedConfig(
depths=[1, 3, 8, 2] , hidden_sizes=[64, 128, 320, 768] , groups_width=16 ),
"regnet-y-016": ImageNetPreTrainedConfig(
depths=[2, 6, 17, 2] , hidden_sizes=[48, 120, 336, 888] , groups_width=24 ),
"regnet-y-032": ImageNetPreTrainedConfig(
depths=[2, 5, 13, 1] , hidden_sizes=[72, 216, 576, 1512] , groups_width=24 ),
"regnet-y-040": ImageNetPreTrainedConfig(
depths=[2, 6, 12, 2] , hidden_sizes=[128, 192, 512, 1088] , groups_width=64 ),
"regnet-y-064": ImageNetPreTrainedConfig(
depths=[2, 7, 14, 2] , hidden_sizes=[144, 288, 576, 1296] , groups_width=72 ),
"regnet-y-080": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 448, 896, 2016] , groups_width=56 ),
"regnet-y-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 ),
"regnet-y-160": ImageNetPreTrainedConfig(
depths=[2, 4, 11, 1] , hidden_sizes=[224, 448, 1232, 3024] , groups_width=112 ),
"regnet-y-320": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
# models created by SEER -> https://arxiv.org/abs/2202.08360
"regnet-y-320-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer": RegNetConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer": RegNetConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
# finetuned on imagenet
"regnet-y-320-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer-in1k": ImageNetPreTrainedConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
}
snake_case : Union[str, Any] = NameToOurModelFuncMap()
snake_case : str = NameToFromModelFuncMap()
# add seer weights logic
def load_using_classy_vision(__lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] ) -> Tuple[nn.Module, Dict]:
snake_case : List[Any] = torch.hub.load_state_dict_from_url(__lowerCamelCase , model_dir=str(__lowerCamelCase ) , map_location="cpu" )
snake_case : Dict = model_func()
# check if we have a head, if yes add it
snake_case : str = files["classy_state_dict"]["base_model"]["model"]
snake_case : Dict = model_state_dict["trunk"]
model.load_state_dict(__lowerCamelCase )
return model.eval(), model_state_dict["heads"]
# pretrained
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet32d/seer_regnet32gf_model_iteration244000.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Optional[int] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet64/seer_regnet64gf_model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : List[str] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/swav_ig1b_regnet128Gf_cnstant_bs32_node16_sinkhorn10_proto16k_syncBN64_warmup8k/model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet10B/model_iteration124500_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
# IN1K finetuned
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet32_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet64_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : str = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet128_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Dict = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_10b_finetuned_in1k_model_phase28_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
if model_name:
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , names_to_config[model_name] , __lowerCamelCase , __lowerCamelCase , )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , )
return config, expected_shape
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help=(
"""The name of the model you wish to convert, it must be one of the supported regnet* architecture,"""
""" currently: regnetx-*, regnety-*. If `None`, all of them will the converted."""
),
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default=None,
type=Path,
required=True,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument(
"""--push_to_hub""",
default=True,
type=bool,
required=False,
help="""If True, push model and image processor to the hub.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 59
|
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Any ) -> List[str]:
'''simple docstring'''
snake_case : int = tempfile.mkdtemp()
# fmt: off
snake_case : Optional[int] = ["[UNK]", "[CLS]", "[SEP]", "[PAD]", "[MASK]", "want", "##want", "##ed", "wa", "un", "runn", "##ing", ",", "low", "lowest"]
# fmt: on
snake_case : List[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] ) )
snake_case : int = {
"do_resize": True,
"size": {"height": 18, "width": 18},
"do_normalize": True,
"image_mean": [0.5, 0.5, 0.5],
"image_std": [0.5, 0.5, 0.5],
}
snake_case : Optional[Any] = os.path.join(self.tmpdirname , snake_case__ )
with open(self.image_processor_file , "w" , encoding="utf-8" ) as fp:
json.dump(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : str ) -> Optional[int]:
'''simple docstring'''
return BertTokenizer.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : List[str] ) -> int:
'''simple docstring'''
return ViTImageProcessor.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
shutil.rmtree(self.tmpdirname )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : List[Any] = [np.random.randint(2_55 , size=(3, 30, 4_00) , dtype=np.uinta )]
snake_case : Optional[int] = [Image.fromarray(np.moveaxis(snake_case__ , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.get_tokenizer()
snake_case : Optional[Any] = self.get_image_processor()
snake_case : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
processor.save_pretrained(self.tmpdirname )
snake_case : Any = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
snake_case : Optional[int] = self.get_tokenizer(bos_token="(BOS)" , eos_token="(EOS)" )
snake_case : Tuple = self.get_image_processor(do_normalize=snake_case__ , padding_value=1.0 )
snake_case : List[str] = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token="(BOS)" , eos_token="(EOS)" , do_normalize=snake_case__ , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Optional[int] = self.get_tokenizer()
snake_case : List[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = self.prepare_image_inputs()
snake_case : str = image_processor(snake_case__ , return_tensors="np" )
snake_case : Any = processor(images=snake_case__ , return_tensors="np" )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.get_image_processor()
snake_case : int = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Tuple = "lower newer"
snake_case : Tuple = processor(text=snake_case__ )
snake_case : Union[str, Any] = tokenizer(snake_case__ )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[int]:
'''simple docstring'''
snake_case : List[Any] = self.get_image_processor()
snake_case : Dict = self.get_tokenizer()
snake_case : Dict = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : int = "lower newer"
snake_case : Dict = self.prepare_image_inputs()
snake_case : Union[str, Any] = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , ["input_ids", "token_type_ids", "attention_mask", "pixel_values"] )
# test if it raises when no input is passed
with self.assertRaises(snake_case__ ):
processor()
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : Tuple = self.get_image_processor()
snake_case : Optional[Any] = self.get_tokenizer()
snake_case : Tuple = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : List[str] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
snake_case : List[Any] = processor.batch_decode(snake_case__ )
snake_case : Union[str, Any] = tokenizer.batch_decode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Union[str, Any] = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = "lower newer"
snake_case : List[Any] = self.prepare_image_inputs()
snake_case : Tuple = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 59
| 1
|
from ..utils import DummyObject, requires_backends
class UpperCAmelCase ( metaclass=A_ ):
A__ : List[str] = ["keras_nlp"]
def __init__(self : Any , *snake_case__ : Any , **snake_case__ : Dict ) -> Optional[int]:
'''simple docstring'''
requires_backends(self , ["keras_nlp"] )
| 59
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__lowerCamelCase = {
"""configuration_biogpt""": ["""BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP""", """BioGptConfig"""],
"""tokenization_biogpt""": ["""BioGptTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""BioGptForCausalLM""",
"""BioGptForTokenClassification""",
"""BioGptForSequenceClassification""",
"""BioGptModel""",
"""BioGptPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_biogpt import BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, BioGptConfig
from .tokenization_biogpt import BioGptTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_biogpt import (
BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
BioGptForCausalLM,
BioGptForSequenceClassification,
BioGptForTokenClassification,
BioGptModel,
BioGptPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
| 1
|
import io
import json
import unittest
from parameterized import parameterized
from transformers import FSMTForConditionalGeneration, FSMTTokenizer
from transformers.testing_utils import get_tests_dir, require_torch, slow, torch_device
from utils import calculate_bleu
__lowerCamelCase = get_tests_dir() + """/test_data/fsmt/fsmt_val_data.json"""
with io.open(filename, """r""", encoding="""utf-8""") as f:
__lowerCamelCase = json.load(f)
@require_torch
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[Any] ) -> Optional[Any]:
'''simple docstring'''
return FSMTTokenizer.from_pretrained(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Tuple ) -> int:
'''simple docstring'''
snake_case : str = FSMTForConditionalGeneration.from_pretrained(snake_case__ ).to(snake_case__ )
if torch_device == "cuda":
model.half()
return model
@parameterized.expand(
[
["en-ru", 26.0],
["ru-en", 22.0],
["en-de", 22.0],
["de-en", 29.0],
] )
@slow
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Optional[int] , snake_case__ : Tuple ) -> Dict:
'''simple docstring'''
snake_case : List[str] = f"""facebook/wmt19-{pair}"""
snake_case : str = self.get_tokenizer(snake_case__ )
snake_case : int = self.get_model(snake_case__ )
snake_case : Any = bleu_data[pair]["src"]
snake_case : Tuple = bleu_data[pair]["tgt"]
snake_case : int = tokenizer(snake_case__ , return_tensors="pt" , truncation=snake_case__ , padding="longest" ).to(snake_case__ )
snake_case : Optional[int] = model.generate(
input_ids=batch.input_ids , num_beams=8 , )
snake_case : Optional[int] = tokenizer.batch_decode(
snake_case__ , skip_special_tokens=snake_case__ , clean_up_tokenization_spaces=snake_case__ )
snake_case : Dict = calculate_bleu(snake_case__ , snake_case__ )
print(snake_case__ )
self.assertGreaterEqual(scores["bleu"] , snake_case__ )
| 59
|
import collections
import inspect
import unittest
from typing import Dict, List, Tuple
from transformers import MaskFormerSwinConfig
from transformers.testing_utils import require_torch, require_torch_multi_gpu, torch_device
from transformers.utils import is_torch_available
from ...test_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import MaskFormerSwinBackbone
from transformers.models.maskformer import MaskFormerSwinModel
class UpperCAmelCase :
def __init__(self : Dict , snake_case__ : Dict , snake_case__ : Any=13 , snake_case__ : Any=32 , snake_case__ : Optional[Any]=2 , snake_case__ : Union[str, Any]=3 , snake_case__ : List[Any]=16 , snake_case__ : int=[1, 2, 1] , snake_case__ : Dict=[2, 2, 4] , snake_case__ : Dict=2 , snake_case__ : Tuple=2.0 , snake_case__ : Optional[int]=True , snake_case__ : Union[str, Any]=0.0 , snake_case__ : Any=0.0 , snake_case__ : Union[str, Any]=0.1 , snake_case__ : int="gelu" , snake_case__ : Optional[int]=False , snake_case__ : List[Any]=True , snake_case__ : List[str]=0.02 , snake_case__ : int=1e-5 , snake_case__ : List[str]=True , snake_case__ : Union[str, Any]=None , snake_case__ : List[Any]=True , snake_case__ : Optional[Any]=10 , snake_case__ : Optional[Any]=8 , snake_case__ : Any=["stage1", "stage2", "stage3"] , snake_case__ : Tuple=[1, 2, 3] , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = parent
snake_case : Optional[int] = batch_size
snake_case : Union[str, Any] = image_size
snake_case : Dict = patch_size
snake_case : Optional[Any] = num_channels
snake_case : Union[str, Any] = embed_dim
snake_case : int = depths
snake_case : List[str] = num_heads
snake_case : Union[str, Any] = window_size
snake_case : Union[str, Any] = mlp_ratio
snake_case : List[Any] = qkv_bias
snake_case : List[Any] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Union[str, Any] = drop_path_rate
snake_case : int = hidden_act
snake_case : Optional[int] = use_absolute_embeddings
snake_case : int = patch_norm
snake_case : Union[str, Any] = layer_norm_eps
snake_case : Any = initializer_range
snake_case : Optional[Any] = is_training
snake_case : Tuple = scope
snake_case : Optional[int] = use_labels
snake_case : Optional[Any] = type_sequence_label_size
snake_case : Union[str, Any] = encoder_stride
snake_case : Any = out_features
snake_case : Tuple = out_indices
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case : Optional[Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case : int = None
if self.use_labels:
snake_case : Tuple = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : Dict = self.get_config()
return config, pixel_values, labels
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> int:
'''simple docstring'''
return MaskFormerSwinConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , depths=self.depths , num_heads=self.num_heads , window_size=self.window_size , mlp_ratio=self.mlp_ratio , qkv_bias=self.qkv_bias , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , drop_path_rate=self.drop_path_rate , hidden_act=self.hidden_act , use_absolute_embeddings=self.use_absolute_embeddings , path_norm=self.patch_norm , layer_norm_eps=self.layer_norm_eps , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , out_features=self.out_features , out_indices=self.out_indices , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : List[Any] , snake_case__ : List[str] , snake_case__ : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
snake_case : Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case : int = int(config.embed_dim * 2 ** (len(config.depths ) - 1) )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, expected_seq_len, expected_dim) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = MaskFormerSwinBackbone(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [13, 16, 16, 16] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , [16, 32, 64] )
# verify ValueError
with self.parent.assertRaises(snake_case__ ):
snake_case : Tuple = ["stem"]
snake_case : List[Any] = MaskFormerSwinBackbone(config=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self.prepare_config_and_inputs()
snake_case , snake_case , snake_case : List[Any] = config_and_inputs
snake_case : int = {"pixel_values": pixel_values}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : List[str] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
A__ : str = {"feature-extraction": MaskFormerSwinModel} if is_torch_available() else {}
A__ : Optional[Any] = False
A__ : List[Any] = False
A__ : List[str] = False
A__ : List[str] = False
A__ : Union[str, Any] = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = MaskFormerSwinModelTester(self )
snake_case : Optional[int] = ConfigTester(self , config_class=snake_case__ , embed_dim=37 )
@require_torch_multi_gpu
@unittest.skip(
reason=(
"`MaskFormerSwinModel` outputs `hidden_states_spatial_dimensions` which doesn't work well with"
" `nn.DataParallel`"
) )
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : str ) -> List[str]:
'''simple docstring'''
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
return
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Dict:
'''simple docstring'''
snake_case : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*snake_case__ )
@unittest.skip("Swin does not use inputs_embeds" )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
pass
@unittest.skip("Swin does not support feedforward chunking" )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Dict:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : int = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case : List[Any] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , nn.Linear ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case , snake_case : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : str = model_class(snake_case__ )
snake_case : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case : Optional[Any] = [*signature.parameters.keys()]
snake_case : Tuple = ["pixel_values"]
self.assertListEqual(arg_names[:1] , snake_case__ )
@unittest.skip(reason="MaskFormerSwin is only used as backbone and doesn't support output_attentions" )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="MaskFormerSwin is only used as an internal backbone" )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
with torch.no_grad():
snake_case : Any = model(**self._prepare_for_class(snake_case__ , snake_case__ ) )
snake_case : int = outputs.hidden_states
snake_case : Union[str, Any] = getattr(
self.model_tester , "expected_num_hidden_layers" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case__ ) , snake_case__ )
# Swin has a different seq_length
snake_case : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : Tuple = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : int = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
for model_class in self.all_model_classes:
snake_case : int = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Dict = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case , snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Any = 3
snake_case : List[str] = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
snake_case : Tuple = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : str = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case : str = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes:
snake_case : str = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Optional[Any] = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
@unittest.skip(reason="MaskFormerSwin doesn't have pretrained checkpoints" )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : int ) -> str:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case__ : Union[str, Any] ):
snake_case : Any = 0
return t
def check_equivalence(snake_case__ : Union[str, Any] , snake_case__ : int , snake_case__ : List[str] , snake_case__ : Optional[int]={} ):
with torch.no_grad():
snake_case : Optional[Any] = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ )
snake_case : Tuple = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ ).to_tuple()
def recursive_check(snake_case__ : List[str] , snake_case__ : Optional[Any] ):
if isinstance(snake_case__ , (List, Tuple) ):
for tuple_iterable_value, dict_iterable_value in zip(snake_case__ , snake_case__ ):
recursive_check(snake_case__ , snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values() , dict_object.values() ):
recursive_check(snake_case__ , snake_case__ )
elif tuple_object is None:
return
else:
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(snake_case__ ) , set_nan_tensor_to_zero(snake_case__ ) , atol=1e-5 ) , msg=(
"Tuple and dict output are not equal. Difference:"
f""" {torch.max(torch.abs(tuple_object - dict_object ) )}. Tuple has `nan`:"""
f""" {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}. Dict has"""
f""" `nan`: {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}."""
) , )
recursive_check(snake_case__ , snake_case__ )
for model_class in self.all_model_classes:
snake_case : Optional[int] = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Union[str, Any] = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : Optional[Any] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Dict = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : List[Any] = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
snake_case : Any = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : List[str] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
@require_torch
class UpperCAmelCase ( unittest.TestCase ,A_ ):
A__ : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
A__ : int = MaskFormerSwinConfig
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModelTester(self )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case : Dict = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Optional[int] = inputs_dict["pixel_values"].shape[0]
for backbone_class in self.all_model_classes:
snake_case : Optional[int] = backbone_class(snake_case__ )
backbone.to(snake_case__ )
backbone.eval()
snake_case : Union[str, Any] = backbone(**snake_case__ )
# Test default outputs and verify feature maps
self.assertIsInstance(outputs.feature_maps , snake_case__ )
self.assertTrue(len(outputs.feature_maps ) == len(backbone.channels ) )
for feature_map, n_channels in zip(outputs.feature_maps , backbone.channels ):
self.assertTrue(feature_map.shape[:2] , (batch_size, n_channels) )
self.assertIsNone(outputs.hidden_states )
self.assertIsNone(outputs.attentions )
# Test output_hidden_states=True
snake_case : Optional[int] = backbone(**snake_case__ , output_hidden_states=snake_case__ )
self.assertIsNotNone(outputs.hidden_states )
self.assertTrue(len(outputs.hidden_states ) , len(backbone.stage_names ) )
# We skip the stem layer
for hidden_states, n_channels in zip(outputs.hidden_states[1:] , backbone.channels ):
for hidden_state in hidden_states:
# Hidden states are in the format (batch_size, (height * width), n_channels)
snake_case , snake_case , snake_case : Dict = hidden_state.shape
self.assertTrue((h_batch_size, h_n_channels) , (batch_size, n_channels) )
# Test output_attentions=True
if self.has_attentions:
snake_case : Optional[Any] = backbone(**snake_case__ , output_attentions=snake_case__ )
self.assertIsNotNone(outputs.attentions )
| 59
| 1
|
import numpy as np
import torch
from torch.utils.data import Dataset
from utils import logger
class UpperCAmelCase ( A_ ):
def __init__(self : Optional[int] , snake_case__ : str , snake_case__ : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = params
snake_case : Union[str, Any] = np.array(snake_case__ )
snake_case : str = np.array([len(snake_case__ ) for t in data] )
self.check()
self.remove_long_sequences()
self.remove_empty_sequences()
self.remove_unknown_sequences()
self.check()
self.print_statistics()
def __getitem__(self : List[str] , snake_case__ : Union[str, Any] ) -> Any:
'''simple docstring'''
return (self.token_ids[index], self.lengths[index])
def __len__(self : str ) -> int:
'''simple docstring'''
return len(self.lengths )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> int:
'''simple docstring'''
assert len(self.token_ids ) == len(self.lengths )
assert all(self.lengths[i] == len(self.token_ids[i] ) for i in range(len(self.lengths ) ) )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
snake_case : str = self.params.max_model_input_size
snake_case : int = self.lengths > max_len
logger.info(f"""Splitting {sum(snake_case__ )} too long sequences.""" )
def divide_chunks(snake_case__ : Optional[Any] , snake_case__ : str ):
return [l[i : i + n] for i in range(0 , len(snake_case__ ) , snake_case__ )]
snake_case : List[str] = []
snake_case : Tuple = []
if self.params.mlm:
snake_case , snake_case : List[str] = self.params.special_tok_ids["cls_token"], self.params.special_tok_ids["sep_token"]
else:
snake_case , snake_case : Optional[int] = self.params.special_tok_ids["bos_token"], self.params.special_tok_ids["eos_token"]
for seq_, len_ in zip(self.token_ids , self.lengths ):
assert (seq_[0] == cls_id) and (seq_[-1] == sep_id), seq_
if len_ <= max_len:
new_tok_ids.append(seq_ )
new_lengths.append(len_ )
else:
snake_case : List[str] = []
for sub_s in divide_chunks(seq_ , max_len - 2 ):
if sub_s[0] != cls_id:
snake_case : List[str] = np.insert(snake_case__ , 0 , snake_case__ )
if sub_s[-1] != sep_id:
snake_case : str = np.insert(snake_case__ , len(snake_case__ ) , snake_case__ )
assert len(snake_case__ ) <= max_len
assert (sub_s[0] == cls_id) and (sub_s[-1] == sep_id), sub_s
sub_seqs.append(snake_case__ )
new_tok_ids.extend(snake_case__ )
new_lengths.extend([len(snake_case__ ) for l in sub_seqs] )
snake_case : Tuple = np.array(snake_case__ )
snake_case : Dict = np.array(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Tuple = len(self )
snake_case : List[str] = self.lengths > 11
snake_case : str = self.token_ids[indices]
snake_case : List[str] = self.lengths[indices]
snake_case : Dict = len(self )
logger.info(f"""Remove {init_size - new_size} too short (<=11 tokens) sequences.""" )
def _SCREAMING_SNAKE_CASE (self : str ) -> List[Any]:
'''simple docstring'''
if "unk_token" not in self.params.special_tok_ids:
return
else:
snake_case : List[str] = self.params.special_tok_ids["unk_token"]
snake_case : Tuple = len(self )
snake_case : Any = np.array([np.count_nonzero(a == unk_token_id ) for a in self.token_ids] )
snake_case : Dict = (unk_occs / self.lengths) < 0.5
snake_case : List[str] = self.token_ids[indices]
snake_case : int = self.lengths[indices]
snake_case : str = len(self )
logger.info(f"""Remove {init_size - new_size} sequences with a high level of unknown tokens (50%).""" )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> List[str]:
'''simple docstring'''
if not self.params.is_master:
return
logger.info(f"""{len(self )} sequences""" )
# data_len = sum(self.lengths)
# nb_unique_tokens = len(Counter(list(chain(*self.token_ids))))
# logger.info(f'{data_len} tokens ({nb_unique_tokens} unique)')
# unk_idx = self.params.special_tok_ids['unk_token']
# nb_unknown = sum([(t==unk_idx).sum() for t in self.token_ids])
# logger.info(f'{nb_unknown} unknown tokens (covering {100*nb_unknown/data_len:.2f}% of the data)')
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, Any] ) -> Any:
'''simple docstring'''
snake_case : Any = [t[0] for t in batch]
snake_case : List[str] = [t[1] for t in batch]
assert len(snake_case__ ) == len(snake_case__ )
# Max for paddings
snake_case : Dict = max(snake_case__ )
# Pad token ids
if self.params.mlm:
snake_case : int = self.params.special_tok_ids["pad_token"]
else:
snake_case : Optional[int] = self.params.special_tok_ids["unk_token"]
snake_case : int = [list(t.astype(snake_case__ ) ) + [pad_idx] * (max_seq_len_ - len(snake_case__ )) for t in token_ids]
assert len(tk_ ) == len(snake_case__ )
assert all(len(snake_case__ ) == max_seq_len_ for t in tk_ )
snake_case : Tuple = torch.tensor(tk_ ) # (bs, max_seq_len_)
snake_case : Optional[int] = torch.tensor(snake_case__ ) # (bs)
return tk_t, lg_t
| 59
|
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : List[str] = []
snake_case : Optional[int] = []
snake_case : Any = []
for rt in rc.restypes:
snake_case : List[Any] = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
snake_case : str = {name: i for i, name in enumerate(__lowerCamelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
snake_case : Optional[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : List[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : int = torch.tensor(
__lowerCamelCase , dtype=torch.floataa , device=protein["aatype"].device , )
snake_case : int = protein["aatype"].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
snake_case : List[Any] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : str = restype_atomaa_mask[protein_aatype]
snake_case : str = residx_atomaa_mask
snake_case : Any = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
snake_case : List[str] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : List[Any] = residx_atomaa_to_atomaa.long()
# create the corresponding mask
snake_case : Union[str, Any] = torch.zeros([21, 37] , dtype=torch.floataa , device=protein["aatype"].device )
for restype, restype_letter in enumerate(rc.restypes ):
snake_case : Optional[int] = rc.restype_atoa[restype_letter]
snake_case : Any = rc.residue_atoms[restype_name]
for atom_name in atom_names:
snake_case : List[Any] = rc.atom_order[atom_name]
snake_case : Optional[Any] = 1
snake_case : List[Any] = restype_atomaa_mask[protein_aatype]
snake_case : int = residx_atomaa_mask
return protein
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : Dict = tree_map(lambda __lowerCamelCase : torch.tensor(__lowerCamelCase , device=batch["aatype"].device ) , __lowerCamelCase , np.ndarray )
snake_case : List[str] = tensor_tree_map(lambda __lowerCamelCase : np.array(__lowerCamelCase ) , make_atomaa_masks(__lowerCamelCase ) )
return out
| 59
| 1
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_torch_available,
)
__lowerCamelCase = {
"""configuration_speecht5""": [
"""SPEECHT5_PRETRAINED_CONFIG_ARCHIVE_MAP""",
"""SPEECHT5_PRETRAINED_HIFIGAN_CONFIG_ARCHIVE_MAP""",
"""SpeechT5Config""",
"""SpeechT5HifiGanConfig""",
],
"""feature_extraction_speecht5""": ["""SpeechT5FeatureExtractor"""],
"""processing_speecht5""": ["""SpeechT5Processor"""],
}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""SpeechT5Tokenizer"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""SPEECHT5_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""SpeechT5ForSpeechToText""",
"""SpeechT5ForSpeechToSpeech""",
"""SpeechT5ForTextToSpeech""",
"""SpeechT5Model""",
"""SpeechT5PreTrainedModel""",
"""SpeechT5HifiGan""",
]
if TYPE_CHECKING:
from .configuration_speechta import (
SPEECHT5_PRETRAINED_CONFIG_ARCHIVE_MAP,
SPEECHT5_PRETRAINED_HIFIGAN_CONFIG_ARCHIVE_MAP,
SpeechTaConfig,
SpeechTaHifiGanConfig,
)
from .feature_extraction_speechta import SpeechTaFeatureExtractor
from .processing_speechta import SpeechTaProcessor
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_speechta import SpeechTaTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_speechta import (
SPEECHT5_PRETRAINED_MODEL_ARCHIVE_LIST,
SpeechTaForSpeechToSpeech,
SpeechTaForSpeechToText,
SpeechTaForTextToSpeech,
SpeechTaHifiGan,
SpeechTaModel,
SpeechTaPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from .tokenization_lxmert import LxmertTokenizer
__lowerCamelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {
"""unc-nlp/lxmert-base-uncased""": """https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""unc-nlp/lxmert-base-uncased""": (
"""https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/tokenizer.json"""
),
},
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": 5_12,
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": {"""do_lower_case""": True},
}
class UpperCAmelCase ( A_ ):
A__ : Any = VOCAB_FILES_NAMES
A__ : Optional[int] = PRETRAINED_VOCAB_FILES_MAP
A__ : Tuple = PRETRAINED_INIT_CONFIGURATION
A__ : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : List[Any] = LxmertTokenizer
def __init__(self : Dict , snake_case__ : Tuple=None , snake_case__ : Optional[Any]=None , snake_case__ : Optional[Any]=True , snake_case__ : Tuple="[UNK]" , snake_case__ : Optional[Any]="[SEP]" , snake_case__ : Optional[Any]="[PAD]" , snake_case__ : List[Any]="[CLS]" , snake_case__ : Tuple="[MASK]" , snake_case__ : Dict=True , snake_case__ : Union[str, Any]=None , **snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
snake_case : int = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , snake_case__ ) != do_lower_case
or normalizer_state.get("strip_accents" , snake_case__ ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , snake_case__ ) != tokenize_chinese_chars
):
snake_case : Union[str, Any] = getattr(snake_case__ , normalizer_state.pop("type" ) )
snake_case : str = do_lower_case
snake_case : List[Any] = strip_accents
snake_case : Optional[int] = tokenize_chinese_chars
snake_case : int = normalizer_class(**snake_case__ )
snake_case : Optional[Any] = do_lower_case
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Optional[Any] , snake_case__ : Dict=None ) -> Any:
'''simple docstring'''
snake_case : Optional[int] = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Optional[Any] = [self.sep_token_id]
snake_case : 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 ) * [0] + len(token_ids_a + sep ) * [1]
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
snake_case : List[Any] = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 59
| 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()
__lowerCamelCase = logging.get_logger(__name__)
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple=False ):
snake_case : List[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"
snake_case : Dict = [(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 UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : int , __lowerCamelCase : Optional[int]=False ):
for i in range(config.num_hidden_layers ):
if base_model:
snake_case : Dict = ""
else:
snake_case : Optional[int] = "vit."
# read in weights + bias of input projection layer (in timm, this is a single matrix + bias)
snake_case : Optional[Any] = state_dict.pop(f"""blocks.{i}.attn.qkv.weight""" )
snake_case : Any = state_dict.pop(f"""blocks.{i}.attn.qkv.bias""" )
# next, add query, keys and values (in that order) to the state dict
snake_case : Any = in_proj_weight[
: config.hidden_size, :
]
snake_case : Union[str, Any] = in_proj_bias[: config.hidden_size]
snake_case : int = in_proj_weight[
config.hidden_size : config.hidden_size * 2, :
]
snake_case : Optional[int] = in_proj_bias[
config.hidden_size : config.hidden_size * 2
]
snake_case : int = in_proj_weight[
-config.hidden_size :, :
]
snake_case : List[Any] = in_proj_bias[-config.hidden_size :]
def UpperCamelCase ( __lowerCamelCase : Optional[Any] ):
snake_case : Optional[int] = ["head.weight", "head.bias"]
for k in ignore_keys:
state_dict.pop(__lowerCamelCase , __lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : List[str] , __lowerCamelCase : Tuple ):
snake_case : Optional[int] = dct.pop(__lowerCamelCase )
snake_case : List[Any] = val
def UpperCamelCase ( ):
snake_case : List[str] = "http://images.cocodataset.org/val2017/000000039769.jpg"
snake_case : str = Image.open(requests.get(__lowerCamelCase , stream=__lowerCamelCase ).raw )
return im
@torch.no_grad()
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : str , __lowerCamelCase : Optional[Any]=False ):
snake_case : int = BitConfig(
global_padding="same" , layer_type="bottleneck" , depths=(3, 4, 9) , out_features=["stage3"] , embedding_dynamic_padding=__lowerCamelCase , )
snake_case : Optional[Any] = ViTHybridConfig(backbone_config=__lowerCamelCase , image_size=384 , num_labels=1000 )
snake_case : Optional[Any] = False
# load original model from timm
snake_case : Any = timm.create_model(__lowerCamelCase , pretrained=__lowerCamelCase )
timm_model.eval()
# load state_dict of original model, remove and rename some keys
snake_case : Any = timm_model.state_dict()
if base_model:
remove_classification_head_(__lowerCamelCase )
snake_case : str = create_rename_keys(__lowerCamelCase , __lowerCamelCase )
for src, dest in rename_keys:
rename_key(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
read_in_q_k_v(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
snake_case : Dict = "huggingface/label-files"
snake_case : List[str] = "imagenet-1k-id2label.json"
snake_case : str = json.load(open(hf_hub_download(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) , "r" ) )
snake_case : int = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : Union[str, Any] = idalabel
snake_case : int = {v: k for k, v in idalabel.items()}
# load HuggingFace model
if vit_name[-5:] == "in21k":
snake_case : Dict = ViTHybridModel(__lowerCamelCase ).eval()
else:
snake_case : int = ViTHybridForImageClassification(__lowerCamelCase ).eval()
model.load_state_dict(__lowerCamelCase )
# create image processor
snake_case : Union[str, Any] = create_transform(**resolve_data_config({} , model=__lowerCamelCase ) )
snake_case : int = transform.transforms
snake_case : int = {
"bilinear": PILImageResampling.BILINEAR,
"bicubic": PILImageResampling.BICUBIC,
"nearest": PILImageResampling.NEAREST,
}
snake_case : Optional[Any] = ViTHybridImageProcessor(
do_resize=__lowerCamelCase , size={"shortest_edge": timm_transforms[0].size} , resample=pillow_resamplings[timm_transforms[0].interpolation.value] , do_center_crop=__lowerCamelCase , crop_size={"height": timm_transforms[1].size[0], "width": timm_transforms[1].size[1]} , do_normalize=__lowerCamelCase , image_mean=timm_transforms[-1].mean.tolist() , image_std=timm_transforms[-1].std.tolist() , )
snake_case : int = prepare_img()
snake_case : List[str] = transform(__lowerCamelCase ).unsqueeze(0 )
snake_case : Union[str, Any] = processor(__lowerCamelCase , return_tensors="pt" ).pixel_values
# verify pixel values
assert torch.allclose(__lowerCamelCase , __lowerCamelCase )
# verify logits
with torch.no_grad():
snake_case : Union[str, Any] = model(__lowerCamelCase )
snake_case : Optional[Any] = outputs.logits
print("Predicted class:" , logits.argmax(-1 ).item() )
if base_model:
snake_case : List[str] = timm_model.forward_features(__lowerCamelCase )
assert timm_pooled_output.shape == outputs.pooler_output.shape
assert torch.allclose(__lowerCamelCase , outputs.pooler_output , atol=1E-3 )
else:
snake_case : str = timm_model(__lowerCamelCase )
assert timm_logits.shape == outputs.logits.shape
assert torch.allclose(__lowerCamelCase , outputs.logits , atol=1E-3 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
Path(__lowerCamelCase ).mkdir(exist_ok=__lowerCamelCase )
print(f"""Saving model {vit_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(__lowerCamelCase )
print(f"""Saving processor to {pytorch_dump_folder_path}""" )
processor.save_pretrained(__lowerCamelCase )
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__":
__lowerCamelCase = 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."""
)
__lowerCamelCase = parser.parse_args()
convert_vit_checkpoint(args.vit_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 59
|
import torch
from diffusers import DDIMParallelScheduler
from .test_schedulers import SchedulerCommonTest
class UpperCAmelCase ( A_ ):
A__ : Dict = (DDIMParallelScheduler,)
A__ : Tuple = (("eta", 0.0), ("num_inference_steps", 50))
def _SCREAMING_SNAKE_CASE (self : Tuple , **snake_case__ : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case : Any = {
"num_train_timesteps": 10_00,
"beta_start": 0.0001,
"beta_end": 0.02,
"beta_schedule": "linear",
"clip_sample": True,
}
config.update(**snake_case__ )
return config
def _SCREAMING_SNAKE_CASE (self : Dict , **snake_case__ : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = self.scheduler_classes[0]
snake_case : Any = self.get_scheduler_config(**snake_case__ )
snake_case : Any = scheduler_class(**snake_case__ )
snake_case , snake_case : Union[str, Any] = 10, 0.0
snake_case : List[Any] = self.dummy_model()
snake_case : Any = self.dummy_sample_deter
scheduler.set_timesteps(snake_case__ )
for t in scheduler.timesteps:
snake_case : Optional[int] = model(snake_case__ , snake_case__ )
snake_case : List[str] = scheduler.step(snake_case__ , snake_case__ , snake_case__ , snake_case__ ).prev_sample
return sample
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> str:
'''simple docstring'''
for timesteps in [1_00, 5_00, 10_00]:
self.check_over_configs(num_train_timesteps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=snake_case__ )
snake_case : Optional[int] = self.scheduler_classes[0]
snake_case : Optional[int] = self.get_scheduler_config(steps_offset=1 )
snake_case : Union[str, Any] = scheduler_class(**snake_case__ )
scheduler.set_timesteps(5 )
assert torch.equal(scheduler.timesteps , torch.LongTensor([8_01, 6_01, 4_01, 2_01, 1] ) )
def _SCREAMING_SNAKE_CASE (self : int ) -> Tuple:
'''simple docstring'''
for beta_start, beta_end in zip([0.0001, 0.001, 0.01, 0.1] , [0.002, 0.02, 0.2, 2] ):
self.check_over_configs(beta_start=snake_case__ , beta_end=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
for timestep_spacing in ["trailing", "leading"]:
self.check_over_configs(timestep_spacing=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[Any]:
'''simple docstring'''
for rescale_betas_zero_snr in [True, False]:
self.check_over_configs(rescale_betas_zero_snr=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
self.check_over_configs(thresholding=snake_case__ )
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(
thresholding=snake_case__ , prediction_type=snake_case__ , sample_max_value=snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
for t in [1, 10, 49]:
self.check_over_forward(time_step=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Any:
'''simple docstring'''
for t, num_inference_steps in zip([1, 10, 50] , [10, 50, 5_00] ):
self.check_over_forward(time_step=snake_case__ , num_inference_steps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Optional[Any]:
'''simple docstring'''
for t, eta in zip([1, 10, 49] , [0.0, 0.5, 1.0] ):
self.check_over_forward(time_step=snake_case__ , eta=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Dict = self.scheduler_classes[0]
snake_case : Tuple = self.get_scheduler_config()
snake_case : Dict = scheduler_class(**snake_case__ )
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_20 , 4_00 ) - 0.14771 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_80 , 9_60 ) - 0.32460 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_87 , 4_86 ) - 0.00979 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_99 , 9_98 ) - 0.02 ) ) < 1e-5
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
snake_case : Union[str, Any] = self.scheduler_classes[0]
snake_case : List[Any] = self.get_scheduler_config()
snake_case : int = scheduler_class(**snake_case__ )
snake_case , snake_case : Any = 10, 0.0
scheduler.set_timesteps(snake_case__ )
snake_case : Optional[Any] = self.dummy_model()
snake_case : str = self.dummy_sample_deter
snake_case : Dict = self.dummy_sample_deter + 0.1
snake_case : Dict = self.dummy_sample_deter - 0.1
snake_case : Optional[Any] = samplea.shape[0]
snake_case : str = torch.stack([samplea, samplea, samplea] , dim=0 )
snake_case : Tuple = torch.arange(snake_case__ )[0:3, None].repeat(1 , snake_case__ )
snake_case : Tuple = model(samples.flatten(0 , 1 ) , timesteps.flatten(0 , 1 ) )
snake_case : List[str] = scheduler.batch_step_no_noise(snake_case__ , timesteps.flatten(0 , 1 ) , samples.flatten(0 , 1 ) , snake_case__ )
snake_case : Dict = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 1147.7904 ) < 1e-2
assert abs(result_mean.item() - 0.4982 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[Any] = self.full_loop()
snake_case : Optional[Any] = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 172.0067 ) < 1e-2
assert abs(result_mean.item() - 0.223967 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : str ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(prediction_type="v_prediction" )
snake_case : int = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[int] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 52.5302 ) < 1e-2
assert abs(result_mean.item() - 0.0684 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : str = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.8295 ) < 1e-2
assert abs(result_mean.item() - 0.1951 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : Tuple = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.0784 ) < 1e-2
assert abs(result_mean.item() - 0.1941 ) < 1e-3
| 59
| 1
|
import requests
__lowerCamelCase = """https://newsapi.org/v1/articles?source=bbc-news&sortBy=top&apiKey="""
def UpperCamelCase ( __lowerCamelCase : str ):
# fetching a list of articles in json format
snake_case : List[str] = requests.get(_NEWS_API + bbc_news_api_key ).json()
# each article in the list is a dict
for i, article in enumerate(bbc_news_page["articles"] , 1 ):
print(f"""{i}.) {article['title']}""" )
if __name__ == "__main__":
fetch_bbc_news(bbc_news_api_key="""<Your BBC News API key goes here>""")
| 59
|
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )]
snake_case : int = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1 or len(__lowerCamelCase ) <= key:
return input_string
for position, character in enumerate(__lowerCamelCase ):
snake_case : Any = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append(__lowerCamelCase )
snake_case : List[str] = ["".join(__lowerCamelCase ) for row in temp_grid]
snake_case : Tuple = "".join(__lowerCamelCase )
return output_string
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : Dict = []
snake_case : Union[str, Any] = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1:
return input_string
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )] # generates template
for position in range(len(__lowerCamelCase ) ):
snake_case : List[str] = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append("*" )
snake_case : Tuple = 0
for row in temp_grid: # fills in the characters
snake_case : Union[str, Any] = input_string[counter : counter + len(__lowerCamelCase )]
grid.append(list(__lowerCamelCase ) )
counter += len(__lowerCamelCase )
snake_case : str = "" # reads as zigzag
for position in range(len(__lowerCamelCase ) ):
snake_case : Optional[int] = position % (lowest * 2) # puts it in bounds
snake_case : Tuple = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
output_string += grid[num][0]
grid[num].pop(0 )
return output_string
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : Tuple = {}
for key_guess in range(1 , len(__lowerCamelCase ) ): # tries every key
snake_case : Any = decrypt(__lowerCamelCase , __lowerCamelCase )
return results
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
import argparse
import json
import requests
import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from transformers import SegformerImageProcessor, SwinConfig, UperNetConfig, UperNetForSemanticSegmentation
def UpperCamelCase ( __lowerCamelCase : List[str] ):
snake_case : Union[str, Any] = 384
snake_case : List[str] = 7
if "tiny" in model_name:
snake_case : Optional[int] = 96
snake_case : int = (2, 2, 6, 2)
snake_case : Tuple = (3, 6, 12, 24)
elif "small" in model_name:
snake_case : List[str] = 96
snake_case : Optional[int] = (2, 2, 18, 2)
snake_case : List[Any] = (3, 6, 12, 24)
elif "base" in model_name:
snake_case : str = 128
snake_case : Dict = (2, 2, 18, 2)
snake_case : int = (4, 8, 16, 32)
snake_case : Dict = 12
snake_case : Tuple = 512
elif "large" in model_name:
snake_case : List[Any] = 192
snake_case : Dict = (2, 2, 18, 2)
snake_case : Tuple = (6, 12, 24, 48)
snake_case : List[str] = 12
snake_case : str = 768
# set label information
snake_case : int = 150
snake_case : List[Any] = "huggingface/label-files"
snake_case : Optional[int] = "ade20k-id2label.json"
snake_case : List[Any] = json.load(open(hf_hub_download(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) , "r" ) )
snake_case : List[str] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : Optional[Any] = {v: k for k, v in idalabel.items()}
snake_case : Tuple = SwinConfig(
embed_dim=__lowerCamelCase , depths=__lowerCamelCase , num_heads=__lowerCamelCase , window_size=__lowerCamelCase , out_features=["stage1", "stage2", "stage3", "stage4"] , )
snake_case : Any = UperNetConfig(
backbone_config=__lowerCamelCase , auxiliary_in_channels=__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase , )
return config
def UpperCamelCase ( __lowerCamelCase : Any ):
snake_case : Union[str, Any] = []
# fmt: off
# stem
rename_keys.append(("backbone.patch_embed.projection.weight", "backbone.embeddings.patch_embeddings.projection.weight") )
rename_keys.append(("backbone.patch_embed.projection.bias", "backbone.embeddings.patch_embeddings.projection.bias") )
rename_keys.append(("backbone.patch_embed.norm.weight", "backbone.embeddings.norm.weight") )
rename_keys.append(("backbone.patch_embed.norm.bias", "backbone.embeddings.norm.bias") )
# stages
for i in range(len(config.backbone_config.depths ) ):
for j in range(config.backbone_config.depths[i] ):
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.norm1.weight""", f"""backbone.encoder.layers.{i}.blocks.{j}.layernorm_before.weight""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.norm1.bias""", f"""backbone.encoder.layers.{i}.blocks.{j}.layernorm_before.bias""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.attn.w_msa.relative_position_bias_table""", f"""backbone.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_bias_table""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.attn.w_msa.relative_position_index""", f"""backbone.encoder.layers.{i}.blocks.{j}.attention.self.relative_position_index""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.attn.w_msa.proj.weight""", f"""backbone.encoder.layers.{i}.blocks.{j}.attention.output.dense.weight""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.attn.w_msa.proj.bias""", f"""backbone.encoder.layers.{i}.blocks.{j}.attention.output.dense.bias""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.norm2.weight""", f"""backbone.encoder.layers.{i}.blocks.{j}.layernorm_after.weight""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.norm2.bias""", f"""backbone.encoder.layers.{i}.blocks.{j}.layernorm_after.bias""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.ffn.layers.0.0.weight""", f"""backbone.encoder.layers.{i}.blocks.{j}.intermediate.dense.weight""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.ffn.layers.0.0.bias""", f"""backbone.encoder.layers.{i}.blocks.{j}.intermediate.dense.bias""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.ffn.layers.1.weight""", f"""backbone.encoder.layers.{i}.blocks.{j}.output.dense.weight""") )
rename_keys.append((f"""backbone.stages.{i}.blocks.{j}.ffn.layers.1.bias""", f"""backbone.encoder.layers.{i}.blocks.{j}.output.dense.bias""") )
if i < 3:
rename_keys.append((f"""backbone.stages.{i}.downsample.reduction.weight""", f"""backbone.encoder.layers.{i}.downsample.reduction.weight""") )
rename_keys.append((f"""backbone.stages.{i}.downsample.norm.weight""", f"""backbone.encoder.layers.{i}.downsample.norm.weight""") )
rename_keys.append((f"""backbone.stages.{i}.downsample.norm.bias""", f"""backbone.encoder.layers.{i}.downsample.norm.bias""") )
rename_keys.append((f"""backbone.norm{i}.weight""", f"""backbone.hidden_states_norms.stage{i+1}.weight""") )
rename_keys.append((f"""backbone.norm{i}.bias""", f"""backbone.hidden_states_norms.stage{i+1}.bias""") )
# decode head
rename_keys.extend(
[
("decode_head.conv_seg.weight", "decode_head.classifier.weight"),
("decode_head.conv_seg.bias", "decode_head.classifier.bias"),
("auxiliary_head.conv_seg.weight", "auxiliary_head.classifier.weight"),
("auxiliary_head.conv_seg.bias", "auxiliary_head.classifier.bias"),
] )
# fmt: on
return rename_keys
def UpperCamelCase ( __lowerCamelCase : Tuple , __lowerCamelCase : Optional[Any] , __lowerCamelCase : List[str] ):
snake_case : int = dct.pop(__lowerCamelCase )
snake_case : List[str] = val
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Tuple ):
snake_case : List[str] = [int(backbone_config.embed_dim * 2**i ) for i in range(len(backbone_config.depths ) )]
for i in range(len(backbone_config.depths ) ):
snake_case : List[Any] = num_features[i]
for j in range(backbone_config.depths[i] ):
# fmt: off
# read in weights + bias of input projection layer (in original implementation, this is a single matrix + bias)
snake_case : Any = state_dict.pop(f"""backbone.stages.{i}.blocks.{j}.attn.w_msa.qkv.weight""" )
snake_case : Union[str, Any] = state_dict.pop(f"""backbone.stages.{i}.blocks.{j}.attn.w_msa.qkv.bias""" )
# next, add query, keys and values (in that order) to the state dict
snake_case : Dict = in_proj_weight[:dim, :]
snake_case : Union[str, Any] = in_proj_bias[: dim]
snake_case : Union[str, Any] = in_proj_weight[
dim : dim * 2, :
]
snake_case : Dict = in_proj_bias[
dim : dim * 2
]
snake_case : Dict = in_proj_weight[
-dim :, :
]
snake_case : List[str] = in_proj_bias[-dim :]
# fmt: on
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case , snake_case : Union[str, Any] = x.shape
snake_case : int = x.reshape(__lowerCamelCase , 4 , in_channel // 4 )
snake_case : Any = x[:, [0, 2, 1, 3], :].transpose(1 , 2 ).reshape(__lowerCamelCase , __lowerCamelCase )
return x
def UpperCamelCase ( __lowerCamelCase : Dict ):
snake_case , snake_case : Optional[int] = x.shape
snake_case : Dict = x.reshape(__lowerCamelCase , in_channel // 4 , 4 )
snake_case : Optional[int] = x[:, :, [0, 2, 1, 3]].transpose(1 , 2 ).reshape(__lowerCamelCase , __lowerCamelCase )
return x
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
snake_case : Optional[int] = x.shape[0]
snake_case : Tuple = x.reshape(4 , in_channel // 4 )
snake_case : Tuple = x[[0, 2, 1, 3], :].transpose(0 , 1 ).reshape(__lowerCamelCase )
return x
def UpperCamelCase ( __lowerCamelCase : Tuple ):
snake_case : Optional[int] = x.shape[0]
snake_case : Optional[int] = x.reshape(in_channel // 4 , 4 )
snake_case : Optional[Any] = x[:, [0, 2, 1, 3]].transpose(0 , 1 ).reshape(__lowerCamelCase )
return x
def UpperCamelCase ( __lowerCamelCase : Optional[Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Any] ):
snake_case : Optional[int] = {
"upernet-swin-tiny": "https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K_20210531_112542-e380ad3e.pth",
"upernet-swin-small": "https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K_20210526_192015-ee2fff1c.pth",
"upernet-swin-base": "https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K_20210531_125459-429057bf.pth",
"upernet-swin-large": "https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_large_patch4_window12_512x512_pretrain_384x384_22K_160k_ade20k/upernet_swin_large_patch4_window12_512x512_pretrain_384x384_22K_160k_ade20k_20220318_091743-9ba68901.pth",
}
snake_case : Union[str, Any] = model_name_to_url[model_name]
snake_case : int = torch.hub.load_state_dict_from_url(__lowerCamelCase , map_location="cpu" , file_name=__lowerCamelCase )[
"state_dict"
]
for name, param in state_dict.items():
print(__lowerCamelCase , param.shape )
snake_case : List[Any] = get_upernet_config(__lowerCamelCase )
snake_case : List[str] = UperNetForSemanticSegmentation(__lowerCamelCase )
model.eval()
# replace "bn" => "batch_norm"
for key in state_dict.copy().keys():
snake_case : List[Any] = state_dict.pop(__lowerCamelCase )
if "bn" in key:
snake_case : Dict = key.replace("bn" , "batch_norm" )
snake_case : Optional[Any] = val
# rename keys
snake_case : Union[str, Any] = create_rename_keys(__lowerCamelCase )
for src, dest in rename_keys:
rename_key(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
read_in_q_k_v(__lowerCamelCase , config.backbone_config )
# fix downsample parameters
for key, value in state_dict.items():
if "downsample" in key:
if "reduction" in key:
snake_case : Union[str, Any] = reverse_correct_unfold_reduction_order(__lowerCamelCase )
if "norm" in key:
snake_case : Union[str, Any] = reverse_correct_unfold_norm_order(__lowerCamelCase )
model.load_state_dict(__lowerCamelCase )
# verify on image
snake_case : Union[str, Any] = "https://huggingface.co/datasets/hf-internal-testing/fixtures_ade20k/resolve/main/ADE_val_00000001.jpg"
snake_case : List[str] = Image.open(requests.get(__lowerCamelCase , stream=__lowerCamelCase ).raw ).convert("RGB" )
snake_case : Optional[int] = SegformerImageProcessor()
snake_case : List[Any] = processor(__lowerCamelCase , return_tensors="pt" ).pixel_values
with torch.no_grad():
snake_case : int = model(__lowerCamelCase )
snake_case : int = outputs.logits
print(logits.shape )
print("First values of logits:" , logits[0, 0, :3, :3] )
# assert values
if model_name == "upernet-swin-tiny":
snake_case : int = torch.tensor(
[[-7.5958, -7.5958, -7.4302], [-7.5958, -7.5958, -7.4302], [-7.4797, -7.4797, -7.3068]] )
elif model_name == "upernet-swin-small":
snake_case : List[Any] = torch.tensor(
[[-7.1921, -7.1921, -6.9532], [-7.1921, -7.1921, -6.9532], [-7.0908, -7.0908, -6.8534]] )
elif model_name == "upernet-swin-base":
snake_case : Any = torch.tensor(
[[-6.5851, -6.5851, -6.4330], [-6.5851, -6.5851, -6.4330], [-6.4763, -6.4763, -6.3254]] )
elif model_name == "upernet-swin-large":
snake_case : List[str] = torch.tensor(
[[-7.5297, -7.5297, -7.3802], [-7.5297, -7.5297, -7.3802], [-7.4044, -7.4044, -7.2586]] )
print("Logits:" , outputs.logits[0, 0, :3, :3] )
assert torch.allclose(outputs.logits[0, 0, :3, :3] , __lowerCamelCase , atol=1E-4 )
print("Looks ok!" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(__lowerCamelCase )
print(f"""Saving processor to {pytorch_dump_folder_path}""" )
processor.save_pretrained(__lowerCamelCase )
if push_to_hub:
print(f"""Pushing model and processor for {model_name} to hub""" )
model.push_to_hub(f"""openmmlab/{model_name}""" )
processor.push_to_hub(f"""openmmlab/{model_name}""" )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default="""upernet-swin-tiny""",
type=str,
choices=[F'upernet-swin-{size}' for size in ["""tiny""", """small""", """base""", """large"""]],
help="""Name of the Swin + UperNet 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 or not to push the converted model to the 🤗 hub."""
)
__lowerCamelCase = parser.parse_args()
convert_upernet_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
if is_sentencepiece_available():
from ..ta.tokenization_ta import TaTokenizer
else:
from ...utils.dummy_sentencepiece_objects import TaTokenizer
__lowerCamelCase = TaTokenizer
if is_tokenizers_available():
from ..ta.tokenization_ta_fast import TaTokenizerFast
else:
from ...utils.dummy_tokenizers_objects import TaTokenizerFast
__lowerCamelCase = TaTokenizerFast
__lowerCamelCase = {"""configuration_mt5""": ["""MT5Config""", """MT5OnnxConfig"""]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""MT5EncoderModel""",
"""MT5ForConditionalGeneration""",
"""MT5ForQuestionAnswering""",
"""MT5Model""",
"""MT5PreTrainedModel""",
"""MT5Stack""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""TFMT5EncoderModel""", """TFMT5ForConditionalGeneration""", """TFMT5Model"""]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""FlaxMT5EncoderModel""", """FlaxMT5ForConditionalGeneration""", """FlaxMT5Model"""]
if TYPE_CHECKING:
from .configuration_mta import MTaConfig, MTaOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mta import (
MTaEncoderModel,
MTaForConditionalGeneration,
MTaForQuestionAnswering,
MTaModel,
MTaPreTrainedModel,
MTaStack,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel
else:
import sys
__lowerCamelCase = _LazyModule(
__name__,
globals()["""__file__"""],
_import_structure,
extra_objects={"""MT5Tokenizer""": MTaTokenizer, """MT5TokenizerFast""": MTaTokenizerFast},
module_spec=__spec__,
)
| 59
| 1
|
from typing import Dict, List, Optional, Union
import numpy as np
from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
from ...image_transforms import (
center_crop,
get_resize_output_image_size,
normalize,
rescale,
resize,
to_channel_dimension_format,
)
from ...image_utils import (
IMAGENET_DEFAULT_MEAN,
IMAGENET_DEFAULT_STD,
ChannelDimension,
ImageInput,
PILImageResampling,
is_batched,
to_numpy_array,
valid_images,
)
from ...utils import TensorType, logging
__lowerCamelCase = logging.get_logger(__name__)
class UpperCAmelCase ( A_ ):
A__ : int = ["pixel_values"]
def __init__(self : int , snake_case__ : bool = True , snake_case__ : Optional[Dict[str, int]] = None , snake_case__ : PILImageResampling = PILImageResampling.BICUBIC , snake_case__ : bool = True , snake_case__ : bool = True , snake_case__ : Union[int, float] = 1 / 2_55 , snake_case__ : Dict[str, int] = None , snake_case__ : bool = True , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : Optional[Union[float, List[float]]] = None , **snake_case__ : Optional[int] , ) -> None:
'''simple docstring'''
super().__init__(**snake_case__ )
snake_case : Union[str, Any] = size if size is not None else {"height": 2_24, "width": 2_24}
snake_case : Dict = get_size_dict(snake_case__ )
snake_case : Optional[Any] = crop_size if crop_size is not None else {"height": 2_24, "width": 2_24}
snake_case : List[Any] = get_size_dict(snake_case__ , default_to_square=snake_case__ , param_name="crop_size" )
snake_case : Tuple = do_resize
snake_case : Optional[Any] = do_rescale
snake_case : str = do_normalize
snake_case : int = do_center_crop
snake_case : Tuple = crop_size
snake_case : int = size
snake_case : Union[str, Any] = resample
snake_case : int = rescale_factor
snake_case : Any = image_mean if image_mean is not None else IMAGENET_DEFAULT_MEAN
snake_case : List[str] = image_std if image_std is not None else IMAGENET_DEFAULT_STD
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : np.ndarray , snake_case__ : Dict[str, int] , snake_case__ : PILImageResampling = PILImageResampling.BILINEAR , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : Any , ) -> np.ndarray:
'''simple docstring'''
snake_case : List[str] = get_size_dict(snake_case__ )
if "shortest_edge" in size:
snake_case : List[str] = get_resize_output_image_size(snake_case__ , size=size["shortest_edge"] , default_to_square=snake_case__ )
# size = get_resize_output_image_size(image, size["shortest_edge"], size["longest_edge"])
elif "height" in size and "width" in size:
snake_case : List[Any] = (size["height"], size["width"])
else:
raise ValueError(f"""Size must contain 'height' and 'width' keys or 'shortest_edge' key. Got {size.keys()}""" )
return resize(snake_case__ , size=snake_case__ , resample=snake_case__ , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : np.ndarray , snake_case__ : Dict[str, int] , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : List[str] , ) -> np.ndarray:
'''simple docstring'''
snake_case : Dict = get_size_dict(snake_case__ )
if "height" not in size or "width" not in size:
raise ValueError(f"""The `size` parameter must contain the keys (height, width). Got {size.keys()}""" )
return center_crop(snake_case__ , size=(size["height"], size["width"]) , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : np.ndarray , snake_case__ : float , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : List[Any] ) -> np.ndarray:
'''simple docstring'''
return rescale(snake_case__ , scale=snake_case__ , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : np.ndarray , snake_case__ : Union[float, List[float]] , snake_case__ : Union[float, List[float]] , snake_case__ : Optional[Union[str, ChannelDimension]] = None , **snake_case__ : List[str] , ) -> np.ndarray:
'''simple docstring'''
return normalize(snake_case__ , mean=snake_case__ , std=snake_case__ , data_format=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : ImageInput , snake_case__ : Optional[bool] = None , snake_case__ : Dict[str, int] = None , snake_case__ : PILImageResampling = None , snake_case__ : bool = None , snake_case__ : int = None , snake_case__ : Optional[bool] = None , snake_case__ : Optional[float] = None , snake_case__ : Optional[bool] = None , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : Optional[Union[float, List[float]]] = None , snake_case__ : Optional[Union[str, TensorType]] = None , snake_case__ : Union[str, ChannelDimension] = ChannelDimension.FIRST , **snake_case__ : Optional[Any] , ) -> BatchFeature:
'''simple docstring'''
snake_case : int = do_resize if do_resize is not None else self.do_resize
snake_case : int = do_rescale if do_rescale is not None else self.do_rescale
snake_case : Union[str, Any] = do_normalize if do_normalize is not None else self.do_normalize
snake_case : str = do_center_crop if do_center_crop is not None else self.do_center_crop
snake_case : Tuple = crop_size if crop_size is not None else self.crop_size
snake_case : Union[str, Any] = get_size_dict(snake_case__ , param_name="crop_size" , default_to_square=snake_case__ )
snake_case : List[str] = resample if resample is not None else self.resample
snake_case : Dict = rescale_factor if rescale_factor is not None else self.rescale_factor
snake_case : Optional[int] = image_mean if image_mean is not None else self.image_mean
snake_case : Tuple = image_std if image_std is not None else self.image_std
snake_case : str = size if size is not None else self.size
snake_case : List[str] = get_size_dict(snake_case__ )
if not is_batched(snake_case__ ):
snake_case : Optional[int] = [images]
if not valid_images(snake_case__ ):
raise ValueError(
"Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "
"torch.Tensor, tf.Tensor or jax.ndarray." )
if do_resize and size is None:
raise ValueError("Size must be specified if do_resize is True." )
if do_center_crop and crop_size is None:
raise ValueError("Crop size must be specified if do_center_crop is True." )
if do_rescale and rescale_factor is None:
raise ValueError("Rescale factor must be specified if do_rescale is True." )
# All transformations expect numpy arrays.
snake_case : List[Any] = [to_numpy_array(snake_case__ ) for image in images]
if do_resize:
snake_case : Tuple = [self.resize(image=snake_case__ , size=snake_case__ , resample=snake_case__ ) for image in images]
if do_center_crop:
snake_case : List[str] = [self.center_crop(image=snake_case__ , size=snake_case__ ) for image in images]
if do_rescale:
snake_case : Union[str, Any] = [self.rescale(image=snake_case__ , scale=snake_case__ ) for image in images]
if do_normalize:
snake_case : Optional[Any] = [self.normalize(image=snake_case__ , mean=snake_case__ , std=snake_case__ ) for image in images]
snake_case : Union[str, Any] = [to_channel_dimension_format(snake_case__ , snake_case__ ) for image in images]
snake_case : Tuple = {"pixel_values": images}
return BatchFeature(data=snake_case__ , tensor_type=snake_case__ )
| 59
|
import os
import shutil
from pathlib import Path
from typing import Optional, Union
import numpy as np
from huggingface_hub import hf_hub_download
from ..utils import ONNX_EXTERNAL_WEIGHTS_NAME, ONNX_WEIGHTS_NAME, is_onnx_available, logging
if is_onnx_available():
import onnxruntime as ort
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""tensor(bool)""": np.bool_,
"""tensor(int8)""": np.inta,
"""tensor(uint8)""": np.uinta,
"""tensor(int16)""": np.intaa,
"""tensor(uint16)""": np.uintaa,
"""tensor(int32)""": np.intaa,
"""tensor(uint32)""": np.uintaa,
"""tensor(int64)""": np.intaa,
"""tensor(uint64)""": np.uintaa,
"""tensor(float16)""": np.floataa,
"""tensor(float)""": np.floataa,
"""tensor(double)""": np.floataa,
}
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=None , **snake_case__ : Optional[Any] ) -> List[str]:
'''simple docstring'''
logger.info("`diffusers.OnnxRuntimeModel` is experimental and might change in the future." )
snake_case : Optional[Any] = model
snake_case : Dict = kwargs.get("model_save_dir" , snake_case__ )
snake_case : int = kwargs.get("latest_model_name" , snake_case__ )
def __call__(self : Tuple , **snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Union[str, Any] = {k: np.array(snake_case__ ) for k, v in kwargs.items()}
return self.model.run(snake_case__ , snake_case__ )
@staticmethod
def _SCREAMING_SNAKE_CASE (snake_case__ : Union[str, Path] , snake_case__ : Optional[int]=None , snake_case__ : Optional[int]=None ) -> Any:
'''simple docstring'''
if provider is None:
logger.info("No onnxruntime provider specified, using CPUExecutionProvider" )
snake_case : Optional[int] = "CPUExecutionProvider"
return ort.InferenceSession(snake_case__ , providers=[provider] , sess_options=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Union[str, Path] , snake_case__ : Optional[str] = None , **snake_case__ : Any ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = file_name if file_name is not None else ONNX_WEIGHTS_NAME
snake_case : Any = self.model_save_dir.joinpath(self.latest_model_name )
snake_case : str = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
snake_case : List[str] = self.model_save_dir.joinpath(snake_case__ )
if src_path.exists():
snake_case : Tuple = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, os.PathLike] , **snake_case__ : Optional[int] , ) -> str:
'''simple docstring'''
if os.path.isfile(snake_case__ ):
logger.error(f"""Provided path ({save_directory}) should be a directory, not a file""" )
return
os.makedirs(snake_case__ , exist_ok=snake_case__ )
# saving model weights/files
self._save_pretrained(snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Tuple , snake_case__ : Union[str, Path] , snake_case__ : Optional[Union[bool, str, None]] = None , snake_case__ : Optional[Union[str, None]] = None , snake_case__ : bool = False , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional["ort.SessionOptions"] = None , **snake_case__ : Tuple , ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(snake_case__ ):
snake_case : Any = OnnxRuntimeModel.load_model(
os.path.join(snake_case__ , snake_case__ ) , provider=snake_case__ , sess_options=snake_case__ )
snake_case : Union[str, Any] = Path(snake_case__ )
# load model from hub
else:
# download model
snake_case : Dict = hf_hub_download(
repo_id=snake_case__ , filename=snake_case__ , use_auth_token=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , )
snake_case : List[Any] = Path(snake_case__ ).parent
snake_case : Union[str, Any] = Path(snake_case__ ).name
snake_case : Dict = OnnxRuntimeModel.load_model(snake_case__ , provider=snake_case__ , sess_options=snake_case__ )
return cls(model=snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : Union[str, Path] , snake_case__ : bool = True , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , **snake_case__ : Dict , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = None
if len(str(snake_case__ ).split("@" ) ) == 2:
snake_case , snake_case : int = model_id.split("@" )
return cls._from_pretrained(
model_id=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , use_auth_token=snake_case__ , **snake_case__ , )
| 59
| 1
|
import sys
import tempfile
import unittest
import unittest.mock as mock
from pathlib import Path
from huggingface_hub import HfFolder, delete_repo
from requests.exceptions import HTTPError
from transformers import AutoFeatureExtractor, WavaVecaFeatureExtractor
from transformers.testing_utils import TOKEN, USER, get_tests_dir, is_staging_test
sys.path.append(str(Path(__file__).parent.parent / """utils"""))
from test_module.custom_feature_extraction import CustomFeatureExtractor # noqa E402
__lowerCamelCase = get_tests_dir("""fixtures""")
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : Dict = mock.Mock()
snake_case : Tuple = 5_00
snake_case : Tuple = {}
snake_case : int = HTTPError
snake_case : Optional[Any] = {}
# Download this model to make sure it's in the cache.
snake_case : Union[str, Any] = WavaVecaFeatureExtractor.from_pretrained("hf-internal-testing/tiny-random-wav2vec2" )
# Under the mock environment we get a 500 error when trying to reach the model.
with mock.patch("requests.Session.request" , return_value=snake_case__ ) as mock_head:
snake_case : int = WavaVecaFeatureExtractor.from_pretrained("hf-internal-testing/tiny-random-wav2vec2" )
# This check we did call the fake head request
mock_head.assert_called()
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = WavaVecaFeatureExtractor.from_pretrained(
"https://huggingface.co/hf-internal-testing/tiny-random-wav2vec2/resolve/main/preprocessor_config.json" )
@is_staging_test
class UpperCAmelCase ( unittest.TestCase ):
@classmethod
def _SCREAMING_SNAKE_CASE (cls : List[str] ) -> Optional[int]:
'''simple docstring'''
snake_case : int = TOKEN
HfFolder.save_token(snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : str ) -> Tuple:
'''simple docstring'''
try:
delete_repo(token=cls._token , repo_id="test-feature-extractor" )
except HTTPError:
pass
try:
delete_repo(token=cls._token , repo_id="valid_org/test-feature-extractor-org" )
except HTTPError:
pass
try:
delete_repo(token=cls._token , repo_id="test-dynamic-feature-extractor" )
except HTTPError:
pass
def _SCREAMING_SNAKE_CASE (self : Any ) -> str:
'''simple docstring'''
snake_case : Any = WavaVecaFeatureExtractor.from_pretrained(snake_case__ )
feature_extractor.push_to_hub("test-feature-extractor" , use_auth_token=self._token )
snake_case : str = WavaVecaFeatureExtractor.from_pretrained(f"""{USER}/test-feature-extractor""" )
for k, v in feature_extractor.__dict__.items():
self.assertEqual(snake_case__ , getattr(snake_case__ , snake_case__ ) )
# Reset repo
delete_repo(token=self._token , repo_id="test-feature-extractor" )
# Push to hub via save_pretrained
with tempfile.TemporaryDirectory() as tmp_dir:
feature_extractor.save_pretrained(
snake_case__ , repo_id="test-feature-extractor" , push_to_hub=snake_case__ , use_auth_token=self._token )
snake_case : List[Any] = WavaVecaFeatureExtractor.from_pretrained(f"""{USER}/test-feature-extractor""" )
for k, v in feature_extractor.__dict__.items():
self.assertEqual(snake_case__ , getattr(snake_case__ , snake_case__ ) )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
snake_case : List[Any] = WavaVecaFeatureExtractor.from_pretrained(snake_case__ )
feature_extractor.push_to_hub("valid_org/test-feature-extractor" , use_auth_token=self._token )
snake_case : List[str] = WavaVecaFeatureExtractor.from_pretrained("valid_org/test-feature-extractor" )
for k, v in feature_extractor.__dict__.items():
self.assertEqual(snake_case__ , getattr(snake_case__ , snake_case__ ) )
# Reset repo
delete_repo(token=self._token , repo_id="valid_org/test-feature-extractor" )
# Push to hub via save_pretrained
with tempfile.TemporaryDirectory() as tmp_dir:
feature_extractor.save_pretrained(
snake_case__ , repo_id="valid_org/test-feature-extractor-org" , push_to_hub=snake_case__ , use_auth_token=self._token )
snake_case : str = WavaVecaFeatureExtractor.from_pretrained("valid_org/test-feature-extractor-org" )
for k, v in feature_extractor.__dict__.items():
self.assertEqual(snake_case__ , getattr(snake_case__ , snake_case__ ) )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Any:
'''simple docstring'''
CustomFeatureExtractor.register_for_auto_class()
snake_case : Union[str, Any] = CustomFeatureExtractor.from_pretrained(snake_case__ )
feature_extractor.push_to_hub("test-dynamic-feature-extractor" , use_auth_token=self._token )
# This has added the proper auto_map field to the config
self.assertDictEqual(
feature_extractor.auto_map , {"AutoFeatureExtractor": "custom_feature_extraction.CustomFeatureExtractor"} , )
snake_case : Optional[int] = AutoFeatureExtractor.from_pretrained(
f"""{USER}/test-dynamic-feature-extractor""" , trust_remote_code=snake_case__ )
# Can't make an isinstance check because the new_feature_extractor is from the CustomFeatureExtractor class of a dynamic module
self.assertEqual(new_feature_extractor.__class__.__name__ , "CustomFeatureExtractor" )
| 59
|
import argparse
import json
from dataclasses import dataclass, field
from functools import partial
from pathlib import Path
from typing import Callable, Dict, List, Tuple
import timm
import torch
import torch.nn as nn
from classy_vision.models.regnet import RegNet, RegNetParams, RegNetYaagf, RegNetYaagf, RegNetYaaagf
from huggingface_hub import cached_download, hf_hub_url
from torch import Tensor
from vissl.models.model_helpers import get_trunk_forward_outputs
from transformers import AutoImageProcessor, RegNetConfig, RegNetForImageClassification, RegNetModel
from transformers.utils import logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger()
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : List[nn.Module] = field(default_factory=A_ )
A__ : list = field(default_factory=A_ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : Tensor , snake_case__ : Tensor ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = len(list(m.modules() ) ) == 1 or isinstance(snake_case__ , nn.Convad ) or isinstance(snake_case__ , nn.BatchNormad )
if has_not_submodules:
self.traced.append(snake_case__ )
def __call__(self : List[Any] , snake_case__ : Tensor ) -> List[Any]:
'''simple docstring'''
for m in self.module.modules():
self.handles.append(m.register_forward_hook(self._forward_hook ) )
self.module(snake_case__ )
[x.remove() for x in self.handles]
return self
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[int]:
'''simple docstring'''
return list(filter(lambda snake_case__ : len(list(x.state_dict().keys() ) ) > 0 , self.traced ) )
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : nn.Module
A__ : int = 1
A__ : List = field(default_factory=A_ )
A__ : List = field(default_factory=A_ )
A__ : bool = True
def __call__(self : List[Any] , snake_case__ : Tensor ) -> Any:
'''simple docstring'''
snake_case : str = Tracker(self.dest )(snake_case__ ).parametrized
snake_case : Optional[int] = Tracker(self.src )(snake_case__ ).parametrized
snake_case : List[str] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.src_skip , snake_case__ ) )
snake_case : Optional[Any] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.dest_skip , snake_case__ ) )
if len(snake_case__ ) != len(snake_case__ ) and self.raise_if_mismatch:
raise Exception(
f"""Numbers of operations are different. Source module has {len(snake_case__ )} operations while"""
f""" destination module has {len(snake_case__ )}.""" )
for dest_m, src_m in zip(snake_case__ , snake_case__ ):
dest_m.load_state_dict(src_m.state_dict() )
if self.verbose == 1:
print(f"""Transfered from={src_m} to={dest_m}""" )
class UpperCAmelCase ( nn.Module ):
def __init__(self : Tuple , snake_case__ : nn.Module ) -> Optional[Any]:
'''simple docstring'''
super().__init__()
snake_case : List[Tuple[str, nn.Module]] = []
# - get the stem
feature_blocks.append(("conv1", model.stem) )
# - get all the feature blocks
for k, v in model.trunk_output.named_children():
assert k.startswith("block" ), f"""Unexpected layer name {k}"""
snake_case : Union[str, Any] = len(snake_case__ ) + 1
feature_blocks.append((f"""res{block_index}""", v) )
snake_case : Optional[Any] = nn.ModuleDict(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Tensor ) -> Dict:
'''simple docstring'''
return get_trunk_forward_outputs(
snake_case__ , out_feat_keys=snake_case__ , feature_blocks=self._feature_blocks , )
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> str:
'''simple docstring'''
snake_case : List[Any] = x.split("-" )
return x_split[0] + x_split[1] + "_" + "".join(x_split[2:] )
def __getitem__(self : Optional[int] , snake_case__ : str ) -> Callable[[], Tuple[nn.Module, Dict]]:
'''simple docstring'''
if x not in self:
snake_case : Dict = self.convert_name_to_timm(snake_case__ )
snake_case : Union[str, Any] = partial(lambda: (timm.create_model(snake_case__ , pretrained=snake_case__ ).eval(), None) )
else:
snake_case : List[str] = super().__getitem__(snake_case__ )
return val
class UpperCAmelCase ( A_ ):
def __getitem__(self : Dict , snake_case__ : str ) -> Callable[[], nn.Module]:
'''simple docstring'''
if "seer" in x and "in1k" not in x:
snake_case : str = RegNetModel
else:
snake_case : Optional[Any] = RegNetForImageClassification
return val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Tuple[str, str]] ):
for from_key, to_key in keys:
snake_case : str = from_state_dict[from_key].clone()
print(f"""Copied key={from_key} to={to_key}""" )
return to_state_dict
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : RegNetConfig , __lowerCamelCase : Path , __lowerCamelCase : bool = True , ):
print(f"""Converting {name}...""" )
with torch.no_grad():
snake_case , snake_case : int = from_model_func()
snake_case : str = our_model_func(__lowerCamelCase ).eval()
snake_case : int = ModuleTransfer(src=__lowerCamelCase , dest=__lowerCamelCase , raise_if_mismatch=__lowerCamelCase )
snake_case : Dict = torch.randn((1, 3, 224, 224) )
module_transfer(__lowerCamelCase )
if from_state_dict is not None:
snake_case : str = []
# for seer - in1k finetuned we have to manually copy the head
if "seer" in name and "in1k" in name:
snake_case : Tuple = [("0.clf.0.weight", "classifier.1.weight"), ("0.clf.0.bias", "classifier.1.bias")]
snake_case : Optional[Any] = manually_copy_vissl_head(__lowerCamelCase , our_model.state_dict() , __lowerCamelCase )
our_model.load_state_dict(__lowerCamelCase )
snake_case : Any = our_model(__lowerCamelCase , output_hidden_states=__lowerCamelCase )
snake_case : Union[str, Any] = (
our_outputs.logits if isinstance(__lowerCamelCase , __lowerCamelCase ) else our_outputs.last_hidden_state
)
snake_case : Union[str, Any] = from_model(__lowerCamelCase )
snake_case : Dict = from_output[-1] if type(__lowerCamelCase ) is list else from_output
# now since I don't want to use any config files, vissl seer model doesn't actually have an head, so let's just check the last hidden state
if "seer" in name and "in1k" in name:
snake_case : Any = our_outputs.hidden_states[-1]
assert torch.allclose(__lowerCamelCase , __lowerCamelCase ), "The model logits don't match the original one."
if push_to_hub:
our_model.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add model" , use_temp_dir=__lowerCamelCase , )
snake_case : List[str] = 224 if "seer" not in name else 384
# we can use the convnext one
snake_case : int = AutoImageProcessor.from_pretrained("facebook/convnext-base-224-22k-1k" , size=__lowerCamelCase )
image_processor.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add image processor" , use_temp_dir=__lowerCamelCase , )
print(f"""Pushed {name}""" )
def UpperCamelCase ( __lowerCamelCase : Path , __lowerCamelCase : str = None , __lowerCamelCase : bool = True ):
snake_case : Union[str, Any] = "imagenet-1k-id2label.json"
snake_case : List[str] = 1000
snake_case : List[str] = (1, num_labels)
snake_case : Any = "huggingface/label-files"
snake_case : List[str] = num_labels
snake_case : Optional[Any] = json.load(open(cached_download(hf_hub_url(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) ) , "r" ) )
snake_case : List[Any] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : str = idalabel
snake_case : List[Any] = {v: k for k, v in idalabel.items()}
snake_case : Dict = partial(__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase )
snake_case : Optional[Any] = {
"regnet-x-002": ImageNetPreTrainedConfig(
depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 , layer_type="x" ),
"regnet-x-004": ImageNetPreTrainedConfig(
depths=[1, 2, 7, 12] , hidden_sizes=[32, 64, 160, 384] , groups_width=16 , layer_type="x" ),
"regnet-x-006": ImageNetPreTrainedConfig(
depths=[1, 3, 5, 7] , hidden_sizes=[48, 96, 240, 528] , groups_width=24 , layer_type="x" ),
"regnet-x-008": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 5] , hidden_sizes=[64, 128, 288, 672] , groups_width=16 , layer_type="x" ),
"regnet-x-016": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 2] , hidden_sizes=[72, 168, 408, 912] , groups_width=24 , layer_type="x" ),
"regnet-x-032": ImageNetPreTrainedConfig(
depths=[2, 6, 15, 2] , hidden_sizes=[96, 192, 432, 1008] , groups_width=48 , layer_type="x" ),
"regnet-x-040": ImageNetPreTrainedConfig(
depths=[2, 5, 14, 2] , hidden_sizes=[80, 240, 560, 1360] , groups_width=40 , layer_type="x" ),
"regnet-x-064": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 392, 784, 1624] , groups_width=56 , layer_type="x" ),
"regnet-x-080": ImageNetPreTrainedConfig(
depths=[2, 5, 15, 1] , hidden_sizes=[80, 240, 720, 1920] , groups_width=120 , layer_type="x" ),
"regnet-x-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 , layer_type="x" ),
"regnet-x-160": ImageNetPreTrainedConfig(
depths=[2, 6, 13, 1] , hidden_sizes=[256, 512, 896, 2048] , groups_width=128 , layer_type="x" ),
"regnet-x-320": ImageNetPreTrainedConfig(
depths=[2, 7, 13, 1] , hidden_sizes=[336, 672, 1344, 2520] , groups_width=168 , layer_type="x" ),
# y variant
"regnet-y-002": ImageNetPreTrainedConfig(depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 ),
"regnet-y-004": ImageNetPreTrainedConfig(
depths=[1, 3, 6, 6] , hidden_sizes=[48, 104, 208, 440] , groups_width=8 ),
"regnet-y-006": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 4] , hidden_sizes=[48, 112, 256, 608] , groups_width=16 ),
"regnet-y-008": ImageNetPreTrainedConfig(
depths=[1, 3, 8, 2] , hidden_sizes=[64, 128, 320, 768] , groups_width=16 ),
"regnet-y-016": ImageNetPreTrainedConfig(
depths=[2, 6, 17, 2] , hidden_sizes=[48, 120, 336, 888] , groups_width=24 ),
"regnet-y-032": ImageNetPreTrainedConfig(
depths=[2, 5, 13, 1] , hidden_sizes=[72, 216, 576, 1512] , groups_width=24 ),
"regnet-y-040": ImageNetPreTrainedConfig(
depths=[2, 6, 12, 2] , hidden_sizes=[128, 192, 512, 1088] , groups_width=64 ),
"regnet-y-064": ImageNetPreTrainedConfig(
depths=[2, 7, 14, 2] , hidden_sizes=[144, 288, 576, 1296] , groups_width=72 ),
"regnet-y-080": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 448, 896, 2016] , groups_width=56 ),
"regnet-y-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 ),
"regnet-y-160": ImageNetPreTrainedConfig(
depths=[2, 4, 11, 1] , hidden_sizes=[224, 448, 1232, 3024] , groups_width=112 ),
"regnet-y-320": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
# models created by SEER -> https://arxiv.org/abs/2202.08360
"regnet-y-320-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer": RegNetConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer": RegNetConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
# finetuned on imagenet
"regnet-y-320-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer-in1k": ImageNetPreTrainedConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
}
snake_case : Union[str, Any] = NameToOurModelFuncMap()
snake_case : str = NameToFromModelFuncMap()
# add seer weights logic
def load_using_classy_vision(__lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] ) -> Tuple[nn.Module, Dict]:
snake_case : List[Any] = torch.hub.load_state_dict_from_url(__lowerCamelCase , model_dir=str(__lowerCamelCase ) , map_location="cpu" )
snake_case : Dict = model_func()
# check if we have a head, if yes add it
snake_case : str = files["classy_state_dict"]["base_model"]["model"]
snake_case : Dict = model_state_dict["trunk"]
model.load_state_dict(__lowerCamelCase )
return model.eval(), model_state_dict["heads"]
# pretrained
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet32d/seer_regnet32gf_model_iteration244000.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Optional[int] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet64/seer_regnet64gf_model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : List[str] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/swav_ig1b_regnet128Gf_cnstant_bs32_node16_sinkhorn10_proto16k_syncBN64_warmup8k/model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet10B/model_iteration124500_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
# IN1K finetuned
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet32_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet64_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : str = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet128_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Dict = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_10b_finetuned_in1k_model_phase28_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
if model_name:
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , names_to_config[model_name] , __lowerCamelCase , __lowerCamelCase , )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , )
return config, expected_shape
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help=(
"""The name of the model you wish to convert, it must be one of the supported regnet* architecture,"""
""" currently: regnetx-*, regnety-*. If `None`, all of them will the converted."""
),
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default=None,
type=Path,
required=True,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument(
"""--push_to_hub""",
default=True,
type=bool,
required=False,
help="""If True, push model and image processor to the hub.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 59
| 1
|
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""distilbert-base-uncased""": """https://huggingface.co/distilbert-base-uncased/resolve/main/config.json""",
"""distilbert-base-uncased-distilled-squad""": (
"""https://huggingface.co/distilbert-base-uncased-distilled-squad/resolve/main/config.json"""
),
"""distilbert-base-cased""": """https://huggingface.co/distilbert-base-cased/resolve/main/config.json""",
"""distilbert-base-cased-distilled-squad""": (
"""https://huggingface.co/distilbert-base-cased-distilled-squad/resolve/main/config.json"""
),
"""distilbert-base-german-cased""": """https://huggingface.co/distilbert-base-german-cased/resolve/main/config.json""",
"""distilbert-base-multilingual-cased""": (
"""https://huggingface.co/distilbert-base-multilingual-cased/resolve/main/config.json"""
),
"""distilbert-base-uncased-finetuned-sst-2-english""": (
"""https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english/resolve/main/config.json"""
),
}
class UpperCAmelCase ( A_ ):
A__ : int = "distilbert"
A__ : List[str] = {
"hidden_size": "dim",
"num_attention_heads": "n_heads",
"num_hidden_layers": "n_layers",
}
def __init__(self : str , snake_case__ : Union[str, Any]=3_05_22 , snake_case__ : int=5_12 , snake_case__ : Optional[int]=False , snake_case__ : Optional[int]=6 , snake_case__ : Any=12 , snake_case__ : List[Any]=7_68 , snake_case__ : int=4 * 7_68 , snake_case__ : int=0.1 , snake_case__ : List[str]=0.1 , snake_case__ : str="gelu" , snake_case__ : Any=0.02 , snake_case__ : Optional[Any]=0.1 , snake_case__ : List[Any]=0.2 , snake_case__ : Dict=0 , **snake_case__ : List[str] , ) -> Optional[int]:
'''simple docstring'''
snake_case : Any = vocab_size
snake_case : int = max_position_embeddings
snake_case : Optional[Any] = sinusoidal_pos_embds
snake_case : List[str] = n_layers
snake_case : List[Any] = n_heads
snake_case : str = dim
snake_case : Tuple = hidden_dim
snake_case : Union[str, Any] = dropout
snake_case : List[str] = attention_dropout
snake_case : Any = activation
snake_case : int = initializer_range
snake_case : List[Any] = qa_dropout
snake_case : str = seq_classif_dropout
super().__init__(**snake_case__ , pad_token_id=snake_case__ )
class UpperCAmelCase ( A_ ):
@property
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
if self.task == "multiple-choice":
snake_case : str = {0: "batch", 1: "choice", 2: "sequence"}
else:
snake_case : int = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
] )
| 59
|
import warnings
from typing import Dict
import numpy as np
from ..utils import ExplicitEnum, add_end_docstrings, is_tf_available, is_torch_available
from .base import PIPELINE_INIT_ARGS, GenericTensor, Pipeline
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
return 1.0 / (1.0 + np.exp(-_outputs ))
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Tuple = np.max(_outputs , axis=-1 , keepdims=__lowerCamelCase )
snake_case : int = np.exp(_outputs - maxes )
return shifted_exp / shifted_exp.sum(axis=-1 , keepdims=__lowerCamelCase )
class UpperCAmelCase ( A_ ):
A__ : Any = "sigmoid"
A__ : str = "softmax"
A__ : int = "none"
@add_end_docstrings(
A_ ,r"\n return_all_scores (`bool`, *optional*, defaults to `False`):\n Whether to return all prediction scores or just the one of the predicted class.\n function_to_apply (`str`, *optional*, defaults to `\"default\"`):\n The function to apply to the model outputs in order to retrieve the scores. Accepts four different values:\n\n - `\"default\"`: if the model has a single label, will apply the sigmoid function on the output. If the model\n has several labels, will apply the softmax function on the output.\n - `\"sigmoid\"`: Applies the sigmoid function on the output.\n - `\"softmax\"`: Applies the softmax function on the output.\n - `\"none\"`: Does not apply any function on the output.\n " ,)
class UpperCAmelCase ( A_ ):
A__ : int = False
A__ : Union[str, Any] = ClassificationFunction.NONE
def __init__(self : List[str] , **snake_case__ : int ) -> str:
'''simple docstring'''
super().__init__(**snake_case__ )
self.check_model_type(
TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if self.framework == "tf"
else MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str]=None , snake_case__ : Optional[Any]=None , snake_case__ : Union[str, Any]="" , **snake_case__ : List[str] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = tokenizer_kwargs
snake_case : List[Any] = {}
if hasattr(self.model.config , "return_all_scores" ) and return_all_scores is None:
snake_case : Optional[int] = self.model.config.return_all_scores
if isinstance(snake_case__ , snake_case__ ) or top_k is None:
snake_case : List[Any] = top_k
snake_case : str = False
elif return_all_scores is not None:
warnings.warn(
"`return_all_scores` is now deprecated, if want a similar functionality use `top_k=None` instead of"
" `return_all_scores=True` or `top_k=1` instead of `return_all_scores=False`." , snake_case__ , )
if return_all_scores:
snake_case : List[str] = None
else:
snake_case : Optional[int] = 1
if isinstance(snake_case__ , snake_case__ ):
snake_case : Dict = ClassificationFunction[function_to_apply.upper()]
if function_to_apply is not None:
snake_case : Optional[int] = function_to_apply
return preprocess_params, {}, postprocess_params
def __call__(self : Dict , *snake_case__ : List[str] , **snake_case__ : int ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = super().__call__(*snake_case__ , **snake_case__ )
# TODO try and retrieve it in a nicer way from _sanitize_parameters.
snake_case : Tuple = "top_k" not in kwargs
if isinstance(args[0] , snake_case__ ) and _legacy:
# This pipeline is odd, and return a list when single item is run
return [result]
else:
return result
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Tuple , **snake_case__ : Union[str, Any] ) -> Dict[str, GenericTensor]:
'''simple docstring'''
snake_case : int = self.framework
if isinstance(snake_case__ , snake_case__ ):
return self.tokenizer(**snake_case__ , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ) and len(snake_case__ ) == 1 and isinstance(inputs[0] , snake_case__ ) and len(inputs[0] ) == 2:
# It used to be valid to use a list of list of list for text pairs, keeping this path for BC
return self.tokenizer(
text=inputs[0][0] , text_pair=inputs[0][1] , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
# This is likely an invalid usage of the pipeline attempting to pass text pairs.
raise ValueError(
"The pipeline received invalid inputs, if you are trying to send text pairs, you can try to send a"
" dictionary `{\"text\": \"My text\", \"text_pair\": \"My pair\"}` in order to send a text pair." )
return self.tokenizer(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Union[str, Any] ) -> int:
'''simple docstring'''
return self.model(**snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Optional[Any] , snake_case__ : List[str]=None , snake_case__ : Dict=1 , snake_case__ : Tuple=True ) -> str:
'''simple docstring'''
if function_to_apply is None:
if self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels == 1:
snake_case : Tuple = ClassificationFunction.SIGMOID
elif self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels > 1:
snake_case : Tuple = ClassificationFunction.SOFTMAX
elif hasattr(self.model.config , "function_to_apply" ) and function_to_apply is None:
snake_case : Tuple = self.model.config.function_to_apply
else:
snake_case : int = ClassificationFunction.NONE
snake_case : Any = model_outputs["logits"][0]
snake_case : List[str] = outputs.numpy()
if function_to_apply == ClassificationFunction.SIGMOID:
snake_case : Optional[Any] = sigmoid(snake_case__ )
elif function_to_apply == ClassificationFunction.SOFTMAX:
snake_case : Union[str, Any] = softmax(snake_case__ )
elif function_to_apply == ClassificationFunction.NONE:
snake_case : Optional[Any] = outputs
else:
raise ValueError(f"""Unrecognized `function_to_apply` argument: {function_to_apply}""" )
if top_k == 1 and _legacy:
return {"label": self.model.config.idalabel[scores.argmax().item()], "score": scores.max().item()}
snake_case : Optional[int] = [
{"label": self.model.config.idalabel[i], "score": score.item()} for i, score in enumerate(snake_case__ )
]
if not _legacy:
dict_scores.sort(key=lambda snake_case__ : x["score"] , reverse=snake_case__ )
if top_k is not None:
snake_case : Optional[int] = dict_scores[:top_k]
return dict_scores
| 59
| 1
|
from collections import OrderedDict
from typing import TYPE_CHECKING, Any, List, Mapping, Optional, Union
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import TensorType, logging
if TYPE_CHECKING:
from ...onnx.config import PatchingSpec
from ...tokenization_utils_base import PreTrainedTokenizerBase
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""allenai/longformer-base-4096""": """https://huggingface.co/allenai/longformer-base-4096/resolve/main/config.json""",
"""allenai/longformer-large-4096""": """https://huggingface.co/allenai/longformer-large-4096/resolve/main/config.json""",
"""allenai/longformer-large-4096-finetuned-triviaqa""": (
"""https://huggingface.co/allenai/longformer-large-4096-finetuned-triviaqa/resolve/main/config.json"""
),
"""allenai/longformer-base-4096-extra.pos.embd.only""": (
"""https://huggingface.co/allenai/longformer-base-4096-extra.pos.embd.only/resolve/main/config.json"""
),
"""allenai/longformer-large-4096-extra.pos.embd.only""": (
"""https://huggingface.co/allenai/longformer-large-4096-extra.pos.embd.only/resolve/main/config.json"""
),
}
class UpperCAmelCase ( A_ ):
A__ : Optional[Any] = "longformer"
def __init__(self : Optional[Any] , snake_case__ : Union[List[int], int] = 5_12 , snake_case__ : int = 2 , snake_case__ : int = 1 , snake_case__ : int = 0 , snake_case__ : int = 2 , snake_case__ : int = 3_05_22 , snake_case__ : int = 7_68 , snake_case__ : int = 12 , snake_case__ : int = 12 , snake_case__ : int = 30_72 , snake_case__ : str = "gelu" , snake_case__ : float = 0.1 , snake_case__ : float = 0.1 , snake_case__ : int = 5_12 , snake_case__ : int = 2 , snake_case__ : float = 0.02 , snake_case__ : float = 1e-12 , snake_case__ : bool = False , **snake_case__ : List[Any] , ) -> Union[str, Any]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , **snake_case__ )
snake_case : List[Any] = attention_window
snake_case : Any = sep_token_id
snake_case : str = bos_token_id
snake_case : List[str] = eos_token_id
snake_case : Optional[Any] = vocab_size
snake_case : List[str] = hidden_size
snake_case : Dict = num_hidden_layers
snake_case : Tuple = num_attention_heads
snake_case : str = hidden_act
snake_case : List[str] = intermediate_size
snake_case : Any = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : int = max_position_embeddings
snake_case : int = type_vocab_size
snake_case : Dict = initializer_range
snake_case : Union[str, Any] = layer_norm_eps
snake_case : List[str] = onnx_export
class UpperCAmelCase ( A_ ):
def __init__(self : Dict , snake_case__ : "PretrainedConfig" , snake_case__ : str = "default" , snake_case__ : "List[PatchingSpec]" = None ) -> Dict:
'''simple docstring'''
super().__init__(snake_case__ , snake_case__ , snake_case__ )
snake_case : int = True
@property
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
if self.task == "multiple-choice":
snake_case : List[str] = {0: "batch", 1: "choice", 2: "sequence"}
else:
snake_case : int = {0: "batch", 1: "sequence"}
return OrderedDict(
[
("input_ids", dynamic_axis),
("attention_mask", dynamic_axis),
("global_attention_mask", dynamic_axis),
] )
@property
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
snake_case : Any = super().outputs
if self.task == "default":
snake_case : List[Any] = {0: "batch"}
return outputs
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> float:
'''simple docstring'''
return 1e-4
@property
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
return max(super().default_onnx_opset , 14 )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : "PreTrainedTokenizerBase" , snake_case__ : int = -1 , snake_case__ : int = -1 , snake_case__ : bool = False , snake_case__ : Optional[TensorType] = None , ) -> Mapping[str, Any]:
'''simple docstring'''
snake_case : int = super().generate_dummy_inputs(
preprocessor=snake_case__ , batch_size=snake_case__ , seq_length=snake_case__ , is_pair=snake_case__ , framework=snake_case__ )
import torch
# for some reason, replacing this code by inputs["global_attention_mask"] = torch.randint(2, inputs["input_ids"].shape, dtype=torch.int64)
# makes the export fail randomly
snake_case : Optional[int] = torch.zeros_like(inputs["input_ids"] )
# make every second token global
snake_case : Union[str, Any] = 1
return inputs
| 59
|
from __future__ import annotations
__lowerCamelCase = list[list[int]]
# assigning initial values to the grid
__lowerCamelCase = [
[3, 0, 6, 5, 0, 8, 4, 0, 0],
[5, 2, 0, 0, 0, 0, 0, 0, 0],
[0, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
# a grid with no solution
__lowerCamelCase = [
[5, 0, 6, 5, 0, 8, 4, 0, 3],
[5, 2, 0, 0, 0, 0, 0, 0, 2],
[1, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
def UpperCamelCase ( __lowerCamelCase : Matrix , __lowerCamelCase : int , __lowerCamelCase : int , __lowerCamelCase : int ):
for i in range(9 ):
if grid[row][i] == n or grid[i][column] == n:
return False
for i in range(3 ):
for j in range(3 ):
if grid[(row - row % 3) + i][(column - column % 3) + j] == n:
return False
return True
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for i in range(9 ):
for j in range(9 ):
if grid[i][j] == 0:
return i, j
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
if location := find_empty_location(__lowerCamelCase ):
snake_case , snake_case : Union[str, Any] = location
else:
# If the location is ``None``, then the grid is solved.
return grid
for digit in range(1 , 10 ):
if is_safe(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase ):
snake_case : List[Any] = digit
if sudoku(__lowerCamelCase ) is not None:
return grid
snake_case : Union[str, Any] = 0
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for row in grid:
for cell in row:
print(__lowerCamelCase , end=" " )
print()
if __name__ == "__main__":
# make a copy of grid so that you can compare with the unmodified grid
for example_grid in (initial_grid, no_solution):
print("""\nExample grid:\n""" + """=""" * 20)
print_solution(example_grid)
print("""\nExample grid solution:""")
__lowerCamelCase = sudoku(example_grid)
if solution is not None:
print_solution(solution)
else:
print("""Cannot find a solution.""")
| 59
| 1
|
import random
import torch
from huggingface_hub import HfApi
from diffusers import UNetaDModel
__lowerCamelCase = HfApi()
__lowerCamelCase = {}
# fmt: off
__lowerCamelCase = torch.tensor([
-0.7515, -1.6883, 0.2420, 0.0300, 0.6347, 1.3433, -1.1743, -3.7467,
1.2342, -2.2485, 0.4636, 0.8076, -0.7991, 0.3969, 0.8498, 0.9189,
-1.8887, -3.3522, 0.7639, 0.2040, 0.6271, -2.7148, -1.6316, 3.0839,
0.3186, 0.2721, -0.9759, -1.2461, 2.6257, 1.3557
])
__lowerCamelCase = torch.tensor([
-2.3639, -2.5344, 0.0054, -0.6674, 1.5990, 1.0158, 0.3124, -2.1436,
1.8795, -2.5429, -0.1566, -0.3973, 1.2490, 2.6447, 1.2283, -0.5208,
-2.8154, -3.5119, 2.3838, 1.2033, 1.7201, -2.1256, -1.4576, 2.7948,
2.4204, -0.9752, -1.2546, 0.8027, 3.2758, 3.1365
])
__lowerCamelCase = torch.tensor([
-0.6531, -0.6891, -0.3172, -0.5375, -0.9140, -0.5367, -0.1175, -0.7869,
-0.3808, -0.4513, -0.2098, -0.0083, 0.3183, 0.5140, 0.2247, -0.1304,
-0.1302, -0.2802, -0.2084, -0.2025, -0.4967, -0.4873, -0.0861, 0.6925,
0.0250, 0.1290, -0.1543, 0.6316, 1.0460, 1.4943
])
__lowerCamelCase = torch.tensor([
0.0911, 0.1107, 0.0182, 0.0435, -0.0805, -0.0608, 0.0381, 0.2172,
-0.0280, 0.1327, -0.0299, -0.0255, -0.0050, -0.1170, -0.1046, 0.0309,
0.1367, 0.1728, -0.0533, -0.0748, -0.0534, 0.1624, 0.0384, -0.1805,
-0.0707, 0.0642, 0.0220, -0.0134, -0.1333, -0.1505
])
__lowerCamelCase = torch.tensor([
0.1321, 0.1337, 0.0440, 0.0622, -0.0591, -0.0370, 0.0503, 0.2133,
-0.0177, 0.1415, -0.0116, -0.0112, 0.0044, -0.0980, -0.0789, 0.0395,
0.1502, 0.1785, -0.0488, -0.0514, -0.0404, 0.1539, 0.0454, -0.1559,
-0.0665, 0.0659, 0.0383, -0.0005, -0.1266, -0.1386
])
__lowerCamelCase = torch.tensor([
0.1154, 0.1218, 0.0307, 0.0526, -0.0711, -0.0541, 0.0366, 0.2078,
-0.0267, 0.1317, -0.0226, -0.0193, -0.0014, -0.1055, -0.0902, 0.0330,
0.1391, 0.1709, -0.0562, -0.0693, -0.0560, 0.1482, 0.0381, -0.1683,
-0.0681, 0.0661, 0.0331, -0.0046, -0.1268, -0.1431
])
__lowerCamelCase = torch.tensor([
0.1192, 0.1240, 0.0414, 0.0606, -0.0557, -0.0412, 0.0430, 0.2042,
-0.0200, 0.1385, -0.0115, -0.0132, 0.0017, -0.0965, -0.0802, 0.0398,
0.1433, 0.1747, -0.0458, -0.0533, -0.0407, 0.1545, 0.0419, -0.1574,
-0.0645, 0.0626, 0.0341, -0.0010, -0.1199, -0.1390
])
__lowerCamelCase = torch.tensor([
0.1075, 0.1074, 0.0205, 0.0431, -0.0774, -0.0607, 0.0298, 0.2042,
-0.0320, 0.1267, -0.0281, -0.0250, -0.0064, -0.1091, -0.0946, 0.0290,
0.1328, 0.1650, -0.0580, -0.0738, -0.0586, 0.1440, 0.0337, -0.1746,
-0.0712, 0.0605, 0.0250, -0.0099, -0.1316, -0.1473
])
__lowerCamelCase = torch.tensor([
-1.4572, -2.0481, -0.0414, -0.6005, 1.4136, 0.5848, 0.4028, -2.7330,
1.2212, -2.1228, 0.2155, 0.4039, 0.7662, 2.0535, 0.7477, -0.3243,
-2.1758, -2.7648, 1.6947, 0.7026, 1.2338, -1.6078, -0.8682, 2.2810,
1.8574, -0.5718, -0.5586, -0.0186, 2.3415, 2.1251])
__lowerCamelCase = torch.tensor([
-1.3690, -1.9720, -0.4090, -0.6966, 1.4660, 0.9938, -0.1385, -2.7324,
0.7736, -1.8917, 0.2923, 0.4293, 0.1693, 1.4112, 1.1887, -0.3181,
-2.2160, -2.6381, 1.3170, 0.8163, 0.9240, -1.6544, -0.6099, 2.5259,
1.6430, -0.9090, -0.9392, -0.0126, 2.4268, 2.3266
])
__lowerCamelCase = torch.tensor([
-1.3525, -1.9628, -0.3956, -0.6860, 1.4664, 1.0014, -0.1259, -2.7212,
0.7772, -1.8811, 0.2996, 0.4388, 0.1704, 1.4029, 1.1701, -0.3027,
-2.2053, -2.6287, 1.3350, 0.8131, 0.9274, -1.6292, -0.6098, 2.5131,
1.6505, -0.8958, -0.9298, -0.0151, 2.4257, 2.3355
])
__lowerCamelCase = torch.tensor([
-2.0585, -2.7897, -0.2850, -0.8940, 1.9052, 0.5702, 0.6345, -3.8959,
1.5932, -3.2319, 0.1974, 0.0287, 1.7566, 2.6543, 0.8387, -0.5351,
-3.2736, -4.3375, 2.9029, 1.6390, 1.4640, -2.1701, -1.9013, 2.9341,
3.4981, -0.6255, -1.1644, -0.1591, 3.7097, 3.2066
])
__lowerCamelCase = torch.tensor([
-2.3139, -2.5594, -0.0197, -0.6785, 1.7001, 1.1606, 0.3075, -2.1740,
1.8071, -2.5630, -0.0926, -0.3811, 1.2116, 2.6246, 1.2731, -0.5398,
-2.8153, -3.6140, 2.3893, 1.3262, 1.6258, -2.1856, -1.3267, 2.8395,
2.3779, -1.0623, -1.2468, 0.8959, 3.3367, 3.2243
])
__lowerCamelCase = torch.tensor([
-2.0628, -2.7667, -0.2089, -0.8263, 2.0539, 0.5992, 0.6495, -3.8336,
1.6025, -3.2817, 0.1721, -0.0633, 1.7516, 2.7039, 0.8100, -0.5908,
-3.2113, -4.4343, 2.9257, 1.3632, 1.5562, -2.1489, -1.9894, 3.0560,
3.3396, -0.7328, -1.0417, 0.0383, 3.7093, 3.2343
])
__lowerCamelCase = torch.tensor([
-1.4574, -2.0569, -0.0473, -0.6117, 1.4018, 0.5769, 0.4129, -2.7344,
1.2241, -2.1397, 0.2000, 0.3937, 0.7616, 2.0453, 0.7324, -0.3391,
-2.1746, -2.7744, 1.6963, 0.6921, 1.2187, -1.6172, -0.8877, 2.2439,
1.8471, -0.5839, -0.5605, -0.0464, 2.3250, 2.1219
])
# fmt: on
__lowerCamelCase = api.list_models(filter="""diffusers""")
for mod in models:
if "google" in mod.author or mod.modelId == "CompVis/ldm-celebahq-256":
__lowerCamelCase = """/home/patrick/google_checkpoints/""" + mod.modelId.split("""/""")[-1]
print(F'Started running {mod.modelId}!!!')
if mod.modelId.startswith("""CompVis"""):
__lowerCamelCase = UNetaDModel.from_pretrained(local_checkpoint, subfolder="""unet""")
else:
__lowerCamelCase = UNetaDModel.from_pretrained(local_checkpoint)
torch.manual_seed(0)
random.seed(0)
__lowerCamelCase = torch.randn(1, model.config.in_channels, model.config.sample_size, model.config.sample_size)
__lowerCamelCase = torch.tensor([10] * noise.shape[0])
with torch.no_grad():
__lowerCamelCase = model(noise, time_step).sample
assert torch.allclose(
logits[0, 0, 0, :30], results["""_""".join("""_""".join(mod.modelId.split("""/""")).split("""-"""))], atol=1e-3
)
print(F'{mod.modelId} has passed successfully!!!')
| 59
|
import logging
import numpy as np
import pytest
from scipy.linalg import eigh
logging.basicConfig(level=logging.INFO, format="""%(message)s""")
def UpperCamelCase ( __lowerCamelCase : np.ndarray ):
return input_array.reshape((input_array.size, 1) )
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Any = np.nan
for i in range(__lowerCamelCase ):
snake_case : List[str] = features[:, labels == i]
snake_case : Dict = data.mean(1 )
# Centralize the data of class i
snake_case : Optional[Any] = data - column_reshape(__lowerCamelCase )
if i > 0:
# If covariance_sum is not None
covariance_sum += np.dot(__lowerCamelCase , centered_data.T )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Optional[Any] = features.mean(1 )
snake_case : Tuple = np.nan
for i in range(__lowerCamelCase ):
snake_case : Tuple = features[:, labels == i]
snake_case : Tuple = data.shape[1]
snake_case : List[str] = data.mean(1 )
if i > 0:
# If covariance_sum is not None
covariance_sum += device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[int] = device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
# Check if the features have been loaded
if features.any():
snake_case : Tuple = features.mean(1 )
# Center the dataset
snake_case : List[str] = features - np.reshape(__lowerCamelCase , (data_mean.size, 1) )
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T ) / features.shape[1]
snake_case , snake_case : Dict = np.linalg.eigh(__lowerCamelCase )
# Take all the columns in the reverse order (-1), and then takes only the first
snake_case : Optional[Any] = eigenvectors[:, ::-1][:, 0:dimensions]
# Project the database on the new space
snake_case : Union[str, Any] = np.dot(filtered_eigenvectors.T , __lowerCamelCase )
logging.info("Principal Component Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int , __lowerCamelCase : int ):
assert classes > dimensions
# Check if features have been already loaded
if features.any:
snake_case , snake_case : str = eigh(
covariance_between_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , covariance_within_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , )
snake_case : str = eigenvectors[:, ::-1][:, :dimensions]
snake_case , snake_case , snake_case : int = np.linalg.svd(__lowerCamelCase )
snake_case : List[Any] = svd_matrix[:, 0:dimensions]
snake_case : Optional[Any] = np.dot(filtered_svd_matrix.T , __lowerCamelCase )
logging.info("Linear Discriminant Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( ):
# Create dummy dataset with 2 classes and 3 features
snake_case : str = np.array([[1, 2, 3, 4, 5], [2, 3, 4, 5, 6], [3, 4, 5, 6, 7]] )
snake_case : Union[str, Any] = np.array([0, 0, 0, 1, 1] )
snake_case : List[Any] = 2
snake_case : Any = 2
# Assert that the function raises an AssertionError if dimensions > classes
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : str = linear_discriminant_analysis(
__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
if isinstance(__lowerCamelCase , np.ndarray ):
raise AssertionError(
"Did not raise AssertionError for dimensions > classes" )
assert error_info.type is AssertionError
def UpperCamelCase ( ):
snake_case : List[str] = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]] )
snake_case : List[str] = 2
snake_case : int = np.array([[6.9282_0323, 8.6602_5404, 10.3923_0485], [3.0, 3.0, 3.0]] )
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : Union[str, Any] = principal_component_analysis(__lowerCamelCase , __lowerCamelCase )
if not np.allclose(__lowerCamelCase , __lowerCamelCase ):
raise AssertionError
assert error_info.type is AssertionError
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
# See all MEGATRON_BERT models at https://huggingface.co/models?filter=bert
}
class UpperCAmelCase ( A_ ):
A__ : List[str] = "megatron-bert"
def __init__(self : Optional[int] , snake_case__ : List[str]=2_90_56 , snake_case__ : List[Any]=10_24 , snake_case__ : str=24 , snake_case__ : Tuple=16 , snake_case__ : Union[str, Any]=40_96 , snake_case__ : str="gelu" , snake_case__ : str=0.1 , snake_case__ : Optional[int]=0.1 , snake_case__ : Tuple=5_12 , snake_case__ : Union[str, Any]=2 , snake_case__ : Dict=0.02 , snake_case__ : List[Any]=1e-12 , snake_case__ : int=0 , snake_case__ : Tuple="absolute" , snake_case__ : Any=True , **snake_case__ : Union[str, Any] , ) -> Optional[Any]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , **snake_case__ )
snake_case : Tuple = vocab_size
snake_case : str = hidden_size
snake_case : str = num_hidden_layers
snake_case : str = num_attention_heads
snake_case : Optional[int] = hidden_act
snake_case : int = intermediate_size
snake_case : List[str] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : List[str] = type_vocab_size
snake_case : List[str] = initializer_range
snake_case : Tuple = layer_norm_eps
snake_case : int = position_embedding_type
snake_case : str = use_cache
| 59
|
import pytest
from datasets import inspect_metric, list_metrics, load_metric
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : Optional[int] ):
monkeypatch.setattr("datasets.utils.deprecation_utils._emitted_deprecation_warnings" , set() )
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : str ):
class UpperCAmelCase :
def __init__(self : Optional[int] , snake_case__ : str ) -> Any:
'''simple docstring'''
snake_case : List[str] = metric_id
class UpperCAmelCase :
A__ : List[str] = [MetricMock(A_ ) for metric_id in ["accuracy", "mse", "precision", "codeparrot/apps_metric"]]
def _SCREAMING_SNAKE_CASE (self : int ) -> List[str]:
'''simple docstring'''
return self._metrics
monkeypatch.setattr("datasets.inspect.huggingface_hub" , HfhMock() )
@pytest.mark.parametrize(
"func, args" , [(load_metric, ("metrics/mse",)), (list_metrics, ()), (inspect_metric, ("metrics/mse", "tmp_path"))] )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Any , __lowerCamelCase : Any , __lowerCamelCase : int , __lowerCamelCase : Any ):
if "tmp_path" in args:
snake_case : str = tuple(arg if arg != "tmp_path" else tmp_path for arg in args )
with pytest.warns(__lowerCamelCase , match="https://huggingface.co/docs/evaluate" ):
func(*__lowerCamelCase )
| 59
| 1
|
import os
import shutil
from pathlib import Path
from typing import Optional, Union
import numpy as np
from huggingface_hub import hf_hub_download
from ..utils import ONNX_EXTERNAL_WEIGHTS_NAME, ONNX_WEIGHTS_NAME, is_onnx_available, logging
if is_onnx_available():
import onnxruntime as ort
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""tensor(bool)""": np.bool_,
"""tensor(int8)""": np.inta,
"""tensor(uint8)""": np.uinta,
"""tensor(int16)""": np.intaa,
"""tensor(uint16)""": np.uintaa,
"""tensor(int32)""": np.intaa,
"""tensor(uint32)""": np.uintaa,
"""tensor(int64)""": np.intaa,
"""tensor(uint64)""": np.uintaa,
"""tensor(float16)""": np.floataa,
"""tensor(float)""": np.floataa,
"""tensor(double)""": np.floataa,
}
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=None , **snake_case__ : Optional[Any] ) -> List[str]:
'''simple docstring'''
logger.info("`diffusers.OnnxRuntimeModel` is experimental and might change in the future." )
snake_case : Optional[Any] = model
snake_case : Dict = kwargs.get("model_save_dir" , snake_case__ )
snake_case : int = kwargs.get("latest_model_name" , snake_case__ )
def __call__(self : Tuple , **snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Union[str, Any] = {k: np.array(snake_case__ ) for k, v in kwargs.items()}
return self.model.run(snake_case__ , snake_case__ )
@staticmethod
def _SCREAMING_SNAKE_CASE (snake_case__ : Union[str, Path] , snake_case__ : Optional[int]=None , snake_case__ : Optional[int]=None ) -> Any:
'''simple docstring'''
if provider is None:
logger.info("No onnxruntime provider specified, using CPUExecutionProvider" )
snake_case : Optional[int] = "CPUExecutionProvider"
return ort.InferenceSession(snake_case__ , providers=[provider] , sess_options=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Union[str, Path] , snake_case__ : Optional[str] = None , **snake_case__ : Any ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = file_name if file_name is not None else ONNX_WEIGHTS_NAME
snake_case : Any = self.model_save_dir.joinpath(self.latest_model_name )
snake_case : str = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
snake_case : List[str] = self.model_save_dir.joinpath(snake_case__ )
if src_path.exists():
snake_case : Tuple = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, os.PathLike] , **snake_case__ : Optional[int] , ) -> str:
'''simple docstring'''
if os.path.isfile(snake_case__ ):
logger.error(f"""Provided path ({save_directory}) should be a directory, not a file""" )
return
os.makedirs(snake_case__ , exist_ok=snake_case__ )
# saving model weights/files
self._save_pretrained(snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Tuple , snake_case__ : Union[str, Path] , snake_case__ : Optional[Union[bool, str, None]] = None , snake_case__ : Optional[Union[str, None]] = None , snake_case__ : bool = False , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional["ort.SessionOptions"] = None , **snake_case__ : Tuple , ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(snake_case__ ):
snake_case : Any = OnnxRuntimeModel.load_model(
os.path.join(snake_case__ , snake_case__ ) , provider=snake_case__ , sess_options=snake_case__ )
snake_case : Union[str, Any] = Path(snake_case__ )
# load model from hub
else:
# download model
snake_case : Dict = hf_hub_download(
repo_id=snake_case__ , filename=snake_case__ , use_auth_token=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , )
snake_case : List[Any] = Path(snake_case__ ).parent
snake_case : Union[str, Any] = Path(snake_case__ ).name
snake_case : Dict = OnnxRuntimeModel.load_model(snake_case__ , provider=snake_case__ , sess_options=snake_case__ )
return cls(model=snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : Union[str, Path] , snake_case__ : bool = True , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , **snake_case__ : Dict , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = None
if len(str(snake_case__ ).split("@" ) ) == 2:
snake_case , snake_case : int = model_id.split("@" )
return cls._from_pretrained(
model_id=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , use_auth_token=snake_case__ , **snake_case__ , )
| 59
|
import argparse
import dataclasses
import json
import logging
import os
import shutil
from typing import List, Optional
import datasets
from accelerate import Accelerator
from datasets import load_dataset
from finetuning import finetune
from tqdm.auto import tqdm
import transformers
from transformers import AutoConfig, set_seed
from transformers.trainer_utils import IntervalStrategy
__lowerCamelCase = logging.getLogger(__name__)
__lowerCamelCase = """pytorch_model.bin"""
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co."} ,)
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the training data."} )
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the data to predict on."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "A csv or a json file containing the validation data."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "The name of the task to train on."} ,)
A__ : Optional[List[str]] = dataclasses.field(
default=A_ ,metadata={"help": "The list of labels for the task."} )
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "The output directory where the model predictions and checkpoints will be written."} )
A__ : Optional[str] = dataclasses.field(
default="accuracy" ,metadata={"help": "The evaluation metric used for the task."} )
A__ : Optional[str] = dataclasses.field(
default="no" ,metadata={
"help": "The evaluation strategy to adopt during training. Possible values are: [\"no\", \"step\", \"epoch]"
} ,)
A__ : Optional[int] = dataclasses.field(
default=10 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={
"help": "How much the specified evaluation metric must improve to satisfy early stopping conditions."
} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the confidence score."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the validation performance."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to fine-tune on labeled data after pseudo training."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={"help": "Confidence threshold for pseudo-labeled data filtering."} ,)
A__ : Optional[int] = dataclasses.field(
default=1_00 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[int] = dataclasses.field(
default=A_ ,metadata={"help": "Random seed for initialization."} ,)
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Tuple , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Tuple , __lowerCamelCase : Dict , __lowerCamelCase : Optional[int] ):
snake_case : Tuple = datasets.concatenate_datasets([infer_input, infer_output] , axis=1 )
if args.do_filter_by_confidence:
snake_case : Optional[int] = dataset.filter(lambda __lowerCamelCase : example["probability"] > args.confidence_threshold )
if args.do_filter_by_val_performance:
assert eval_result >= 0.0 and eval_result <= 1.0
snake_case : int = int(eval_result * len(__lowerCamelCase ) )
print(__lowerCamelCase )
snake_case : List[str] = dataset.sort("probability" , reverse=__lowerCamelCase )
snake_case : Tuple = dataset.select(range(__lowerCamelCase ) )
snake_case : List[Any] = dataset.remove_columns(["label", "probability"] )
snake_case : Any = dataset.rename_column("prediction" , "label" )
snake_case : str = dataset.map(lambda __lowerCamelCase : {"label": idalabel[example["label"]]} )
snake_case : List[str] = dataset.shuffle(seed=args.seed )
snake_case : int = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.data_file_extension == "csv":
dataset.to_csv(__lowerCamelCase , index=__lowerCamelCase )
else:
dataset.to_json(__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : str , __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple , **__lowerCamelCase : List[Any] ):
snake_case : int = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(accelerator.state )
# Setup logging, we only want one process per machine to log things on the
# screen. accelerator.is_local_main_process is only True for one process per
# machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
snake_case : Dict = STModelArguments(model_name_or_path=__lowerCamelCase )
snake_case : Tuple = STDataArguments(train_file=__lowerCamelCase , infer_file=__lowerCamelCase )
snake_case : str = STTrainingArguments(output_dir=__lowerCamelCase )
snake_case : int = argparse.Namespace()
for arg_class in (model_args, data_args, training_args):
for key, value in vars(__lowerCamelCase ).items():
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
for key, value in kwargs.items():
if hasattr(__lowerCamelCase , __lowerCamelCase ):
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
# Sanity checks
snake_case : List[str] = {}
snake_case : Optional[int] = None
# You need to provide the training data and the data to predict on
assert args.train_file is not None
assert args.infer_file is not None
snake_case : str = args.train_file
snake_case : Tuple = args.infer_file
if args.evaluation_strategy != IntervalStrategy.NO.value:
assert args.eval_file is not None
snake_case : Tuple = args.eval_file
for key in data_files:
snake_case : List[Any] = data_files[key].split("." )[-1]
assert extension in ["csv", "json"], f"""`{key}_file` should be a csv or a json file."""
if args.data_file_extension is None:
snake_case : Union[str, Any] = extension
else:
assert extension == args.data_file_extension, f"""`{key}_file` should be a {args.data_file_extension} file`."""
assert (
args.eval_metric in datasets.list_metrics()
), f"""{args.eval_metric} not in the list of supported metrics {datasets.list_metrics()}."""
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed )
logger.info("Creating the initial data directory for self-training..." )
snake_case : List[Any] = f"""{args.output_dir}/self-train_iter-{{}}""".format
snake_case : Optional[int] = data_dir_format(0 )
if accelerator.is_main_process:
if args.output_dir is not None:
os.makedirs(args.output_dir , exist_ok=__lowerCamelCase )
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : Dict = None
snake_case : Union[str, Any] = None
snake_case : Tuple = 0
snake_case : List[Any] = False
# Show the progress bar
snake_case : List[Any] = tqdm(range(args.max_selftrain_iterations ) , disable=not accelerator.is_local_main_process )
# Self-train
for iteration in range(0 , int(args.max_selftrain_iterations ) ):
snake_case : str = data_dir_format(__lowerCamelCase )
assert os.path.exists(__lowerCamelCase )
# Stage 1: initial fine-tuning for iteration = 0 or pseudo-training for
# iteration > 0
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-1" )
snake_case : Optional[Any] = {
"accelerator": accelerator,
"model_name_or_path": args.model_name_or_path,
"cache_dir": args.cache_dir,
"do_train": True,
"train_file": data_files["train"] if iteration == 0 else data_files["train_pseudo"],
"do_eval": True if args.eval_file is not None else False,
"eval_file": data_files["eval"],
"do_predict": True,
"infer_file": data_files["infer"],
"task_name": args.task_name,
"label_list": args.label_list,
"output_dir": current_output_dir,
"eval_metric": args.eval_metric,
"evaluation_strategy": args.evaluation_strategy,
"early_stopping_patience": args.early_stopping_patience,
"early_stopping_threshold": args.early_stopping_threshold,
"seed": args.seed,
}
# Add additional training arguments
for key, value in kwargs.items():
if key not in arguments_dict and not hasattr(__lowerCamelCase , __lowerCamelCase ):
arguments_dict.update({key: value} )
snake_case : int = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 1." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 1 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 1." , __lowerCamelCase )
if iteration > 0 and args.finetune_on_labeled_data:
# Stage 2 (optional): fine-tuning on the original labeled data
snake_case : str = os.path.join(__lowerCamelCase , "best-checkpoint" )
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-2" )
# Update arguments_dict
snake_case : List[str] = model_path
snake_case : Optional[Any] = data_files["train"]
snake_case : Optional[Any] = current_output_dir
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 2." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 2 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 2." , __lowerCamelCase )
snake_case : int = iteration
snake_case : Tuple = data_dir_format(iteration + 1 )
snake_case : Tuple = AutoConfig.from_pretrained(os.path.join(__lowerCamelCase , "best-checkpoint" ) )
snake_case : Optional[int] = config.idalabel
snake_case : List[Any] = os.path.join(__lowerCamelCase , "eval_results_best-checkpoint.json" )
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "test_results_best-checkpoint.json" )
assert os.path.exists(__lowerCamelCase )
with open(__lowerCamelCase , "r" ) as f:
snake_case : Dict = float(json.load(__lowerCamelCase )[args.eval_metric] )
snake_case : Optional[int] = os.path.join(__lowerCamelCase , "infer_output_best-checkpoint.csv" )
assert os.path.exists(__lowerCamelCase )
# Loading the dataset from local csv or json files.
snake_case : Optional[Any] = load_dataset(args.data_file_extension , data_files={"data": data_files["infer"]} )["data"]
snake_case : Dict = load_dataset("csv" , data_files={"data": infer_output_file} )["data"]
if accelerator.is_main_process:
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) )
if os.path.exists(__lowerCamelCase ):
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""test_results_iter-{iteration}.json""" ) )
create_pseudo_labeled_data(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : str = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.evaluation_strategy != IntervalStrategy.NO.value:
snake_case : List[Any] = eval_result
if best_iteration is None:
snake_case : List[Any] = new_iteration
snake_case : int = new_eval_result
else:
if new_eval_result - best_eval_result > args.early_stopping_threshold:
snake_case : int = new_iteration
snake_case : Union[str, Any] = new_eval_result
snake_case : str = 0
else:
if new_eval_result == best_eval_result:
snake_case : Any = new_iteration
snake_case : Union[str, Any] = new_eval_result
early_stopping_patience_counter += 1
if early_stopping_patience_counter >= args.early_stopping_patience:
snake_case : Tuple = True
progress_bar.update(1 )
if should_training_stop:
break
if best_iteration is not None:
# Save the best iteration
logger.info("Best iteration: %d" , __lowerCamelCase )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
else:
# Assume that the last iteration is the best
logger.info("Best iteration: %d" , args.max_selftrain_iterations - 1 )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{args.max_selftrain_iterations - 1}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
| 59
| 1
|
from __future__ import annotations
from dataclasses import dataclass
@dataclass
class UpperCAmelCase :
A__ : float
A__ : TreeNode | None = None
A__ : TreeNode | None = None
def UpperCamelCase ( __lowerCamelCase : TreeNode | None ):
# Validation
def is_valid_tree(__lowerCamelCase : TreeNode | None ) -> bool:
if node is None:
return True
if not isinstance(__lowerCamelCase , __lowerCamelCase ):
return False
try:
float(node.data )
except (TypeError, ValueError):
return False
return is_valid_tree(node.left ) and is_valid_tree(node.right )
if not is_valid_tree(__lowerCamelCase ):
raise ValueError(
"Each node should be type of TreeNode and data should be float." )
def is_binary_search_tree_recursive_check(
__lowerCamelCase : TreeNode | None , __lowerCamelCase : float , __lowerCamelCase : float ) -> bool:
if node is None:
return True
return (
left_bound < node.data < right_bound
and is_binary_search_tree_recursive_check(node.left , __lowerCamelCase , node.data )
and is_binary_search_tree_recursive_check(
node.right , node.data , __lowerCamelCase )
)
return is_binary_search_tree_recursive_check(__lowerCamelCase , -float("inf" ) , float("inf" ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__lowerCamelCase = {"""configuration_xglm""": ["""XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP""", """XGLMConfig"""]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizer"""]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizerFast"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""XGLMForCausalLM""",
"""XGLMModel""",
"""XGLMPreTrainedModel""",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""FlaxXGLMForCausalLM""",
"""FlaxXGLMModel""",
"""FlaxXGLMPreTrainedModel""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""TFXGLMForCausalLM""",
"""TFXGLMModel""",
"""TFXGLMPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm import XGLMTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm_fast import XGLMTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_xglm import (
TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST,
TFXGLMForCausalLM,
TFXGLMModel,
TFXGLMPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure)
| 59
| 1
|
import os
from shutil import copyfile
from typing import List, Optional, Tuple
from tokenizers import processors
from ...tokenization_utils import AddedToken, BatchEncoding
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from ...utils import is_sentencepiece_available, logging
if is_sentencepiece_available():
from .tokenization_nllb import NllbTokenizer
else:
__lowerCamelCase = None
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {"""vocab_file""": """sentencepiece.bpe.model""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/sentencepiece.bpe.model"""
),
},
"""tokenizer_file""": {
"""facebook/nllb-200-distilled-600M""": (
"""https://huggingface.co/facebook/nllb-200-distilled-600M/resolve/main/tokenizer.json"""
),
},
}
__lowerCamelCase = {
"""facebook/nllb-large-en-ro""": 10_24,
"""facebook/nllb-200-distilled-600M""": 10_24,
}
# fmt: off
__lowerCamelCase = ["""ace_Arab""", """ace_Latn""", """acm_Arab""", """acq_Arab""", """aeb_Arab""", """afr_Latn""", """ajp_Arab""", """aka_Latn""", """amh_Ethi""", """apc_Arab""", """arb_Arab""", """ars_Arab""", """ary_Arab""", """arz_Arab""", """asm_Beng""", """ast_Latn""", """awa_Deva""", """ayr_Latn""", """azb_Arab""", """azj_Latn""", """bak_Cyrl""", """bam_Latn""", """ban_Latn""", """bel_Cyrl""", """bem_Latn""", """ben_Beng""", """bho_Deva""", """bjn_Arab""", """bjn_Latn""", """bod_Tibt""", """bos_Latn""", """bug_Latn""", """bul_Cyrl""", """cat_Latn""", """ceb_Latn""", """ces_Latn""", """cjk_Latn""", """ckb_Arab""", """crh_Latn""", """cym_Latn""", """dan_Latn""", """deu_Latn""", """dik_Latn""", """dyu_Latn""", """dzo_Tibt""", """ell_Grek""", """eng_Latn""", """epo_Latn""", """est_Latn""", """eus_Latn""", """ewe_Latn""", """fao_Latn""", """pes_Arab""", """fij_Latn""", """fin_Latn""", """fon_Latn""", """fra_Latn""", """fur_Latn""", """fuv_Latn""", """gla_Latn""", """gle_Latn""", """glg_Latn""", """grn_Latn""", """guj_Gujr""", """hat_Latn""", """hau_Latn""", """heb_Hebr""", """hin_Deva""", """hne_Deva""", """hrv_Latn""", """hun_Latn""", """hye_Armn""", """ibo_Latn""", """ilo_Latn""", """ind_Latn""", """isl_Latn""", """ita_Latn""", """jav_Latn""", """jpn_Jpan""", """kab_Latn""", """kac_Latn""", """kam_Latn""", """kan_Knda""", """kas_Arab""", """kas_Deva""", """kat_Geor""", """knc_Arab""", """knc_Latn""", """kaz_Cyrl""", """kbp_Latn""", """kea_Latn""", """khm_Khmr""", """kik_Latn""", """kin_Latn""", """kir_Cyrl""", """kmb_Latn""", """kon_Latn""", """kor_Hang""", """kmr_Latn""", """lao_Laoo""", """lvs_Latn""", """lij_Latn""", """lim_Latn""", """lin_Latn""", """lit_Latn""", """lmo_Latn""", """ltg_Latn""", """ltz_Latn""", """lua_Latn""", """lug_Latn""", """luo_Latn""", """lus_Latn""", """mag_Deva""", """mai_Deva""", """mal_Mlym""", """mar_Deva""", """min_Latn""", """mkd_Cyrl""", """plt_Latn""", """mlt_Latn""", """mni_Beng""", """khk_Cyrl""", """mos_Latn""", """mri_Latn""", """zsm_Latn""", """mya_Mymr""", """nld_Latn""", """nno_Latn""", """nob_Latn""", """npi_Deva""", """nso_Latn""", """nus_Latn""", """nya_Latn""", """oci_Latn""", """gaz_Latn""", """ory_Orya""", """pag_Latn""", """pan_Guru""", """pap_Latn""", """pol_Latn""", """por_Latn""", """prs_Arab""", """pbt_Arab""", """quy_Latn""", """ron_Latn""", """run_Latn""", """rus_Cyrl""", """sag_Latn""", """san_Deva""", """sat_Beng""", """scn_Latn""", """shn_Mymr""", """sin_Sinh""", """slk_Latn""", """slv_Latn""", """smo_Latn""", """sna_Latn""", """snd_Arab""", """som_Latn""", """sot_Latn""", """spa_Latn""", """als_Latn""", """srd_Latn""", """srp_Cyrl""", """ssw_Latn""", """sun_Latn""", """swe_Latn""", """swh_Latn""", """szl_Latn""", """tam_Taml""", """tat_Cyrl""", """tel_Telu""", """tgk_Cyrl""", """tgl_Latn""", """tha_Thai""", """tir_Ethi""", """taq_Latn""", """taq_Tfng""", """tpi_Latn""", """tsn_Latn""", """tso_Latn""", """tuk_Latn""", """tum_Latn""", """tur_Latn""", """twi_Latn""", """tzm_Tfng""", """uig_Arab""", """ukr_Cyrl""", """umb_Latn""", """urd_Arab""", """uzn_Latn""", """vec_Latn""", """vie_Latn""", """war_Latn""", """wol_Latn""", """xho_Latn""", """ydd_Hebr""", """yor_Latn""", """yue_Hant""", """zho_Hans""", """zho_Hant""", """zul_Latn"""]
class UpperCAmelCase ( A_ ):
A__ : Dict = VOCAB_FILES_NAMES
A__ : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : Any = PRETRAINED_VOCAB_FILES_MAP
A__ : List[Any] = ["input_ids", "attention_mask"]
A__ : int = NllbTokenizer
A__ : List[int] = []
A__ : List[int] = []
def __init__(self : Union[str, Any] , snake_case__ : Union[str, Any]=None , snake_case__ : int=None , snake_case__ : Any="<s>" , snake_case__ : List[Any]="</s>" , snake_case__ : List[str]="</s>" , snake_case__ : Optional[int]="<s>" , snake_case__ : str="<unk>" , snake_case__ : List[Any]="<pad>" , snake_case__ : int="<mask>" , snake_case__ : int=None , snake_case__ : Tuple=None , snake_case__ : str=None , snake_case__ : List[str]=False , **snake_case__ : Optional[int] , ) -> List[str]:
'''simple docstring'''
snake_case : Any = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else mask_token
snake_case : Union[str, Any] = legacy_behaviour
super().__init__(
vocab_file=snake_case__ , tokenizer_file=snake_case__ , bos_token=snake_case__ , eos_token=snake_case__ , sep_token=snake_case__ , cls_token=snake_case__ , unk_token=snake_case__ , pad_token=snake_case__ , mask_token=snake_case__ , src_lang=snake_case__ , tgt_lang=snake_case__ , additional_special_tokens=snake_case__ , legacy_behaviour=snake_case__ , **snake_case__ , )
snake_case : Optional[int] = vocab_file
snake_case : int = False if not self.vocab_file else True
snake_case : Any = FAIRSEQ_LANGUAGE_CODES.copy()
if additional_special_tokens is not None:
# Only add those special tokens if they are not already there.
_additional_special_tokens.extend(
[t for t in additional_special_tokens if t not in _additional_special_tokens] )
self.add_special_tokens({"additional_special_tokens": _additional_special_tokens} )
snake_case : Tuple = {
lang_code: self.convert_tokens_to_ids(snake_case__ ) for lang_code in FAIRSEQ_LANGUAGE_CODES
}
snake_case : Union[str, Any] = src_lang if src_lang is not None else "eng_Latn"
snake_case : int = self.convert_tokens_to_ids(self._src_lang )
snake_case : List[str] = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
return self._src_lang
@src_lang.setter
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : str ) -> None:
'''simple docstring'''
snake_case : int = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
if token_ids_a is None:
return self.prefix_tokens + token_ids_a + self.suffix_tokens
# We don't expect to process pairs, but leave the pair logic for API consistency
return self.prefix_tokens + token_ids_a + token_ids_a + self.suffix_tokens
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Dict = [self.sep_token_id]
snake_case : 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]
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Union[str, Any] , snake_case__ : str , snake_case__ : Optional[str] , snake_case__ : Optional[str] , **snake_case__ : str ) -> Dict:
'''simple docstring'''
if src_lang is None or tgt_lang is None:
raise ValueError("Translation requires a `src_lang` and a `tgt_lang` for this model" )
snake_case : Tuple = src_lang
snake_case : Optional[int] = self(snake_case__ , add_special_tokens=snake_case__ , return_tensors=snake_case__ , **snake_case__ )
snake_case : Union[str, Any] = self.convert_tokens_to_ids(snake_case__ )
snake_case : Optional[int] = tgt_lang_id
return inputs
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : List[str] , snake_case__ : str = "eng_Latn" , snake_case__ : Optional[List[str]] = None , snake_case__ : str = "fra_Latn" , **snake_case__ : List[str] , ) -> BatchEncoding:
'''simple docstring'''
snake_case : Union[str, Any] = src_lang
snake_case : Dict = tgt_lang
return super().prepare_seqaseq_batch(snake_case__ , snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
return self.set_src_lang_special_tokens(self.src_lang )
def _SCREAMING_SNAKE_CASE (self : str ) -> List[Any]:
'''simple docstring'''
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[Any] ) -> None:
'''simple docstring'''
snake_case : List[Any] = self.convert_tokens_to_ids(snake_case__ )
if self.legacy_behaviour:
snake_case : List[str] = []
snake_case : List[str] = [self.eos_token_id, self.cur_lang_code]
else:
snake_case : Any = [self.cur_lang_code]
snake_case : List[str] = [self.eos_token_id]
snake_case : Union[str, Any] = self.convert_ids_to_tokens(self.prefix_tokens )
snake_case : List[str] = self.convert_ids_to_tokens(self.suffix_tokens )
snake_case : str = processors.TemplateProcessing(
single=prefix_tokens_str + ["$A"] + suffix_tokens_str , pair=prefix_tokens_str + ["$A", "$B"] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : str ) -> None:
'''simple docstring'''
snake_case : Tuple = self.convert_tokens_to_ids(snake_case__ )
if self.legacy_behaviour:
snake_case : Any = []
snake_case : Optional[int] = [self.eos_token_id, self.cur_lang_code]
else:
snake_case : List[Any] = [self.cur_lang_code]
snake_case : Optional[Any] = [self.eos_token_id]
snake_case : str = self.convert_ids_to_tokens(self.prefix_tokens )
snake_case : Any = self.convert_ids_to_tokens(self.suffix_tokens )
snake_case : Optional[int] = processors.TemplateProcessing(
single=prefix_tokens_str + ["$A"] + suffix_tokens_str , pair=prefix_tokens_str + ["$A", "$B"] + suffix_tokens_str , special_tokens=list(zip(prefix_tokens_str + suffix_tokens_str , self.prefix_tokens + self.suffix_tokens ) ) , )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
if not self.can_save_slow_tokenizer:
raise ValueError(
"Your fast tokenizer does not have the necessary information to save the vocabulary for a slow "
"tokenizer." )
if not os.path.isdir(snake_case__ ):
logger.error(f"""Vocabulary path ({save_directory}) should be a directory.""" )
return
snake_case : Optional[Any] = os.path.join(
snake_case__ , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(snake_case__ ):
copyfile(self.vocab_file , snake_case__ )
return (out_vocab_file,)
| 59
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
# See all MEGATRON_BERT models at https://huggingface.co/models?filter=bert
}
class UpperCAmelCase ( A_ ):
A__ : List[str] = "megatron-bert"
def __init__(self : Optional[int] , snake_case__ : List[str]=2_90_56 , snake_case__ : List[Any]=10_24 , snake_case__ : str=24 , snake_case__ : Tuple=16 , snake_case__ : Union[str, Any]=40_96 , snake_case__ : str="gelu" , snake_case__ : str=0.1 , snake_case__ : Optional[int]=0.1 , snake_case__ : Tuple=5_12 , snake_case__ : Union[str, Any]=2 , snake_case__ : Dict=0.02 , snake_case__ : List[Any]=1e-12 , snake_case__ : int=0 , snake_case__ : Tuple="absolute" , snake_case__ : Any=True , **snake_case__ : Union[str, Any] , ) -> Optional[Any]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , **snake_case__ )
snake_case : Tuple = vocab_size
snake_case : str = hidden_size
snake_case : str = num_hidden_layers
snake_case : str = num_attention_heads
snake_case : Optional[int] = hidden_act
snake_case : int = intermediate_size
snake_case : List[str] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : List[str] = type_vocab_size
snake_case : List[str] = initializer_range
snake_case : Tuple = layer_norm_eps
snake_case : int = position_embedding_type
snake_case : str = use_cache
| 59
| 1
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__lowerCamelCase = {"""configuration_xglm""": ["""XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP""", """XGLMConfig"""]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizer"""]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizerFast"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""XGLMForCausalLM""",
"""XGLMModel""",
"""XGLMPreTrainedModel""",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""FlaxXGLMForCausalLM""",
"""FlaxXGLMModel""",
"""FlaxXGLMPreTrainedModel""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""TFXGLMForCausalLM""",
"""TFXGLMModel""",
"""TFXGLMPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm import XGLMTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm_fast import XGLMTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_xglm import (
TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST,
TFXGLMForCausalLM,
TFXGLMModel,
TFXGLMPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure)
| 59
|
import gc
import unittest
from parameterized import parameterized
from diffusers import FlaxUNetaDConditionModel
from diffusers.utils import is_flax_available
from diffusers.utils.testing_utils import load_hf_numpy, require_flax, slow
if is_flax_available():
import jax
import jax.numpy as jnp
@slow
@require_flax
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, Any] , snake_case__ : List[str] ) -> List[str]:
'''simple docstring'''
return f"""gaussian_noise_s={seed}_shape={'_'.join([str(snake_case__ ) for s in shape] )}.npy"""
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> int:
'''simple docstring'''
super().tearDown()
gc.collect()
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Optional[Any]=0 , snake_case__ : Any=(4, 4, 64, 64) , snake_case__ : List[Any]=False ) -> int:
'''simple docstring'''
snake_case : Optional[Any] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Optional[int] = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return image
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Tuple=False , snake_case__ : List[Any]="CompVis/stable-diffusion-v1-4" ) -> List[Any]:
'''simple docstring'''
snake_case : List[str] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : str = "bf16" if fpaa else None
snake_case , snake_case : Optional[int] = FlaxUNetaDConditionModel.from_pretrained(
snake_case__ , subfolder="unet" , dtype=snake_case__ , revision=snake_case__ )
return model, params
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any]=0 , snake_case__ : Union[str, Any]=(4, 77, 7_68) , snake_case__ : Dict=False ) -> List[str]:
'''simple docstring'''
snake_case : Any = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Any = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return hidden_states
@parameterized.expand(
[
# fmt: off
[83, 4, [-0.2323, -0.1304, 0.0813, -0.3093, -0.0919, -0.1571, -0.1125, -0.5806]],
[17, 0.55, [-0.0831, -0.2443, 0.0901, -0.0919, 0.3396, 0.0103, -0.3743, 0.0701]],
[8, 0.89, [-0.4863, 0.0859, 0.0875, -0.1658, 0.9199, -0.0114, 0.4839, 0.4639]],
[3, 10_00, [-0.5649, 0.2402, -0.5518, 0.1248, 1.1328, -0.2443, -0.0325, -1.0078]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : int , snake_case__ : Optional[int] , snake_case__ : Dict ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : List[str] = self.get_unet_model(model_id="CompVis/stable-diffusion-v1-4" , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_latents(snake_case__ , fpaa=snake_case__ )
snake_case : List[str] = self.get_encoder_hidden_states(snake_case__ , fpaa=snake_case__ )
snake_case : Dict = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : Optional[Any] = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Optional[int] = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, in the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[83, 4, [0.1514, 0.0807, 0.1624, 0.1016, -0.1896, 0.0263, 0.0677, 0.2310]],
[17, 0.55, [0.1164, -0.0216, 0.0170, 0.1589, -0.3120, 0.1005, -0.0581, -0.1458]],
[8, 0.89, [-0.1758, -0.0169, 0.1004, -0.1411, 0.1312, 0.1103, -0.1996, 0.2139]],
[3, 10_00, [0.1214, 0.0352, -0.0731, -0.1562, -0.0994, -0.0906, -0.2340, -0.0539]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[int] , snake_case__ : str , snake_case__ : Tuple ) -> str:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.get_unet_model(model_id="stabilityai/stable-diffusion-2" , fpaa=snake_case__ )
snake_case : List[str] = self.get_latents(snake_case__ , shape=(4, 4, 96, 96) , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_encoder_hidden_states(snake_case__ , shape=(4, 77, 10_24) , fpaa=snake_case__ )
snake_case : Optional[int] = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : int = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Dict = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, on the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
| 59
| 1
|
import unittest
from transformers import XLMConfig, is_torch_available
from transformers.testing_utils import require_torch, slow, torch_device
from ...generation.test_utils import GenerationTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import (
XLMForMultipleChoice,
XLMForQuestionAnswering,
XLMForQuestionAnsweringSimple,
XLMForSequenceClassification,
XLMForTokenClassification,
XLMModel,
XLMWithLMHeadModel,
)
from transformers.models.xlm.modeling_xlm import XLM_PRETRAINED_MODEL_ARCHIVE_LIST
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : str , snake_case__ : Optional[Any]=13 , snake_case__ : List[str]=7 , snake_case__ : Dict=True , snake_case__ : str=True , snake_case__ : Optional[Any]=True , snake_case__ : Tuple=True , snake_case__ : Union[str, Any]=True , snake_case__ : Optional[int]=False , snake_case__ : Tuple=False , snake_case__ : Tuple=False , snake_case__ : List[str]=2 , snake_case__ : List[Any]=99 , snake_case__ : List[Any]=0 , snake_case__ : List[Any]=32 , snake_case__ : Tuple=5 , snake_case__ : int=4 , snake_case__ : str=0.1 , snake_case__ : Dict=0.1 , snake_case__ : Optional[Any]=5_12 , snake_case__ : Optional[Any]=2 , snake_case__ : Tuple=0.02 , snake_case__ : Optional[int]=2 , snake_case__ : Union[str, Any]=4 , snake_case__ : int="last" , snake_case__ : str=True , snake_case__ : Dict=None , snake_case__ : int=0 , ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = parent
snake_case : List[str] = batch_size
snake_case : Dict = seq_length
snake_case : List[str] = is_training
snake_case : Dict = use_input_lengths
snake_case : List[str] = use_token_type_ids
snake_case : Dict = use_labels
snake_case : List[Any] = gelu_activation
snake_case : Any = sinusoidal_embeddings
snake_case : Tuple = causal
snake_case : str = asm
snake_case : Tuple = n_langs
snake_case : int = vocab_size
snake_case : Any = n_special
snake_case : List[Any] = hidden_size
snake_case : Optional[int] = num_hidden_layers
snake_case : List[Any] = num_attention_heads
snake_case : Any = hidden_dropout_prob
snake_case : Tuple = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : int = type_sequence_label_size
snake_case : List[str] = initializer_range
snake_case : str = num_labels
snake_case : Optional[Any] = num_choices
snake_case : Union[str, Any] = summary_type
snake_case : Any = use_proj
snake_case : int = scope
snake_case : Optional[Any] = bos_token_id
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case : Optional[int] = random_attention_mask([self.batch_size, self.seq_length] )
snake_case : Tuple = None
if self.use_input_lengths:
snake_case : List[str] = (
ids_tensor([self.batch_size] , vocab_size=2 ) + self.seq_length - 2
) # small variation of seq_length
snake_case : Any = None
if self.use_token_type_ids:
snake_case : str = ids_tensor([self.batch_size, self.seq_length] , self.n_langs )
snake_case : Optional[int] = None
snake_case : Optional[int] = None
snake_case : Union[str, Any] = None
if self.use_labels:
snake_case : Optional[Any] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : int = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case : Union[str, Any] = ids_tensor([self.batch_size] , 2 ).float()
snake_case : int = ids_tensor([self.batch_size] , self.num_choices )
snake_case : List[Any] = self.get_config()
return (
config,
input_ids,
token_type_ids,
input_lengths,
sequence_labels,
token_labels,
is_impossible_labels,
choice_labels,
input_mask,
)
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> int:
'''simple docstring'''
return XLMConfig(
vocab_size=self.vocab_size , n_special=self.n_special , emb_dim=self.hidden_size , n_layers=self.num_hidden_layers , n_heads=self.num_attention_heads , dropout=self.hidden_dropout_prob , attention_dropout=self.attention_probs_dropout_prob , gelu_activation=self.gelu_activation , sinusoidal_embeddings=self.sinusoidal_embeddings , asm=self.asm , causal=self.causal , n_langs=self.n_langs , max_position_embeddings=self.max_position_embeddings , initializer_range=self.initializer_range , summary_type=self.summary_type , use_proj=self.use_proj , num_labels=self.num_labels , bos_token_id=self.bos_token_id , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Any , snake_case__ : List[str] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : List[Any] , snake_case__ : Optional[int] , snake_case__ : int , snake_case__ : Optional[Any] , snake_case__ : str , ) -> Optional[Any]:
'''simple docstring'''
snake_case : str = XLMModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Dict = model(snake_case__ , lengths=snake_case__ , langs=snake_case__ )
snake_case : Optional[Any] = model(snake_case__ , langs=snake_case__ )
snake_case : Union[str, Any] = model(snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[Any] , snake_case__ : Optional[int] , snake_case__ : Tuple , snake_case__ : Optional[Any] , snake_case__ : Optional[int] , snake_case__ : List[str] , snake_case__ : Any , snake_case__ : Any , snake_case__ : List[Any] , ) -> str:
'''simple docstring'''
snake_case : Union[str, Any] = XLMWithLMHeadModel(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : str = model(snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.loss.shape , () )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : List[Any] , snake_case__ : Any , snake_case__ : List[str] , snake_case__ : Optional[Any] , snake_case__ : Tuple , snake_case__ : str , snake_case__ : int , snake_case__ : int , ) -> Optional[int]:
'''simple docstring'''
snake_case : int = XLMForQuestionAnsweringSimple(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : int = model(snake_case__ )
snake_case : Optional[int] = model(snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ )
snake_case : Tuple = outputs
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 : int , snake_case__ : int , snake_case__ : Optional[Any] , snake_case__ : Dict , snake_case__ : Optional[Any] , snake_case__ : Union[str, Any] , snake_case__ : Any , snake_case__ : List[Any] , snake_case__ : Dict , snake_case__ : Tuple , ) -> str:
'''simple docstring'''
snake_case : int = XLMForQuestionAnswering(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : int = model(snake_case__ )
snake_case : Optional[int] = model(
snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ , cls_index=snake_case__ , is_impossible=snake_case__ , p_mask=snake_case__ , )
snake_case : int = model(
snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ , cls_index=snake_case__ , is_impossible=snake_case__ , )
((snake_case) , ) : List[Any] = result_with_labels.to_tuple()
snake_case : Optional[Any] = model(snake_case__ , start_positions=snake_case__ , end_positions=snake_case__ )
((snake_case) , ) : int = result_with_labels.to_tuple()
self.parent.assertEqual(result_with_labels.loss.shape , () )
self.parent.assertEqual(result.start_top_log_probs.shape , (self.batch_size, model.config.start_n_top) )
self.parent.assertEqual(result.start_top_index.shape , (self.batch_size, model.config.start_n_top) )
self.parent.assertEqual(
result.end_top_log_probs.shape , (self.batch_size, model.config.start_n_top * model.config.end_n_top) )
self.parent.assertEqual(
result.end_top_index.shape , (self.batch_size, model.config.start_n_top * model.config.end_n_top) )
self.parent.assertEqual(result.cls_logits.shape , (self.batch_size,) )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Dict , snake_case__ : Any , snake_case__ : Any , snake_case__ : int , snake_case__ : Optional[int] , snake_case__ : Union[str, Any] , snake_case__ : int , snake_case__ : str , snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
snake_case : Union[str, Any] = XLMForSequenceClassification(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Dict = model(snake_case__ )
snake_case : List[str] = model(snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.loss.shape , () )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : Any , snake_case__ : Any , snake_case__ : int , snake_case__ : str , snake_case__ : Optional[Any] , snake_case__ : str , snake_case__ : List[str] , snake_case__ : List[Any] , snake_case__ : Optional[int] , ) -> Tuple:
'''simple docstring'''
snake_case : Optional[int] = self.num_labels
snake_case : Tuple = XLMForTokenClassification(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Dict = model(snake_case__ , attention_mask=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Any , snake_case__ : str , snake_case__ : Tuple , snake_case__ : Tuple , snake_case__ : Tuple , snake_case__ : Tuple , snake_case__ : Union[str, Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : List[str] = self.num_choices
snake_case : str = XLMForMultipleChoice(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Dict = input_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case : int = token_type_ids.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case : Tuple = input_mask.unsqueeze(1 ).expand(-1 , self.num_choices , -1 ).contiguous()
snake_case : Union[str, Any] = model(
snake_case__ , attention_mask=snake_case__ , token_type_ids=snake_case__ , labels=snake_case__ , )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_choices) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Any:
'''simple docstring'''
snake_case : Optional[Any] = self.prepare_config_and_inputs()
(
(
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) ,
) : Dict = config_and_inputs
snake_case : Dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "lengths": input_lengths}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,A_ ,unittest.TestCase ):
A__ : Dict = (
(
XLMModel,
XLMWithLMHeadModel,
XLMForQuestionAnswering,
XLMForSequenceClassification,
XLMForQuestionAnsweringSimple,
XLMForTokenClassification,
XLMForMultipleChoice,
)
if is_torch_available()
else ()
)
A__ : Any = (
(XLMWithLMHeadModel,) if is_torch_available() else ()
) # TODO (PVP): Check other models whether language generation is also applicable
A__ : str = (
{
"feature-extraction": XLMModel,
"fill-mask": XLMWithLMHeadModel,
"question-answering": XLMForQuestionAnsweringSimple,
"text-classification": XLMForSequenceClassification,
"text-generation": XLMWithLMHeadModel,
"token-classification": XLMForTokenClassification,
"zero-shot": XLMForSequenceClassification,
}
if is_torch_available()
else {}
)
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Tuple , snake_case__ : int , snake_case__ : List[Any] , snake_case__ : Dict , snake_case__ : List[str] ) -> int:
'''simple docstring'''
if (
pipeline_test_casse_name == "QAPipelineTests"
and tokenizer_name is not None
and not tokenizer_name.endswith("Fast" )
):
# `QAPipelineTests` fails for a few models when the slower tokenizer are used.
# (The slower tokenizers were never used for pipeline tests before the pipeline testing rework)
# TODO: check (and possibly fix) the `QAPipelineTests` with slower tokenizer
return True
return False
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[Any] , snake_case__ : Dict , snake_case__ : str=False ) -> int:
'''simple docstring'''
snake_case : Tuple = super()._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
if return_labels:
if model_class.__name__ == "XLMForQuestionAnswering":
snake_case : Tuple = torch.zeros(
self.model_tester.batch_size , dtype=torch.long , device=snake_case__ )
snake_case : Optional[int] = torch.zeros(
self.model_tester.batch_size , dtype=torch.long , device=snake_case__ )
return inputs_dict
def _SCREAMING_SNAKE_CASE (self : Dict ) -> int:
'''simple docstring'''
snake_case : Any = XLMModelTester(self )
snake_case : List[str] = ConfigTester(self , config_class=snake_case__ , emb_dim=37 )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[Any]:
'''simple docstring'''
self.config_tester.run_common_tests()
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_lm_head(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
snake_case : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_simple_qa(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_qa(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_sequence_classif(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_token_classif(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_xlm_for_multiple_choice(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Union[str, Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : int , snake_case__ : Optional[Any] , snake_case__ : Optional[Any]=False , snake_case__ : int=1 ) -> Optional[int]:
'''simple docstring'''
self.assertIsInstance(snake_case__ , snake_case__ )
self.assertListEqual(
[isinstance(snake_case__ , snake_case__ ) for iter_attentions in attentions] , [True] * len(snake_case__ ) )
self.assertEqual(len(snake_case__ ) , (max_length - min_length) * num_beam_groups )
for idx, iter_attentions in enumerate(snake_case__ ):
# adds PAD dummy token
snake_case : int = min_length + idx + 1
snake_case : Any = min_length + idx + 1
snake_case : str = (
batch_size * num_beam_groups,
config.num_attention_heads,
tgt_len,
src_len,
)
# check attn size
self.assertListEqual(
[layer_attention.shape for layer_attention in iter_attentions] , [expected_shape] * len(snake_case__ ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Union[str, Any] , snake_case__ : List[str] , snake_case__ : Dict , snake_case__ : Tuple , snake_case__ : Dict , snake_case__ : Optional[int]=False , snake_case__ : str=1 ) -> Dict:
'''simple docstring'''
self.assertIsInstance(snake_case__ , snake_case__ )
self.assertListEqual(
[isinstance(snake_case__ , snake_case__ ) for iter_hidden_states in hidden_states] , [True] * len(snake_case__ ) , )
self.assertEqual(len(snake_case__ ) , (max_length - min_length) * num_beam_groups )
for idx, iter_hidden_states in enumerate(snake_case__ ):
# adds PAD dummy token
snake_case : str = min_length + idx + 1
snake_case : Tuple = (batch_size * num_beam_groups, seq_len, config.hidden_size)
# check hidden size
self.assertListEqual(
[layer_hidden_states.shape for layer_hidden_states in iter_hidden_states] , [expected_shape] * len(snake_case__ ) , )
pass
@slow
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
for model_name in XLM_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case : Dict = XLMModel.from_pretrained(snake_case__ )
self.assertIsNotNone(snake_case__ )
@require_torch
class UpperCAmelCase ( unittest.TestCase ):
@slow
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : int = XLMWithLMHeadModel.from_pretrained("xlm-mlm-en-2048" )
model.to(snake_case__ )
snake_case : Any = torch.tensor([[14, 4_47]] , dtype=torch.long , device=snake_case__ ) # the president
snake_case : Dict = [
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
14,
4_47,
] # the president the president the president the president the president the president the president the president the president the president
# TODO(PVP): this and other input_ids I tried for generation give pretty bad results. Not sure why. Model might just not be made for auto-regressive inference
snake_case : Optional[int] = model.generate(snake_case__ , do_sample=snake_case__ )
self.assertListEqual(output_ids[0].cpu().numpy().tolist() , snake_case__ )
| 59
|
import argparse
import re
from typing import Dict
import torch
from datasets import Audio, Dataset, load_dataset, load_metric
from transformers import AutoFeatureExtractor, pipeline
def UpperCamelCase ( __lowerCamelCase : Dataset , __lowerCamelCase : Dict[str, str] ):
snake_case : int = args.log_outputs
snake_case : Dict = "_".join(args.dataset.split("/" ) + [args.config, args.split] )
# load metric
snake_case : List[str] = load_metric("wer" )
snake_case : Tuple = load_metric("cer" )
# compute metrics
snake_case : List[Any] = wer.compute(references=result["target"] , predictions=result["prediction"] )
snake_case : int = cer.compute(references=result["target"] , predictions=result["prediction"] )
# print & log results
snake_case : int = f"""WER: {wer_result}\nCER: {cer_result}"""
print(__lowerCamelCase )
with open(f"""{dataset_id}_eval_results.txt""" , "w" ) as f:
f.write(__lowerCamelCase )
# log all results in text file. Possibly interesting for analysis
if log_outputs is not None:
snake_case : int = f"""log_{dataset_id}_predictions.txt"""
snake_case : List[Any] = f"""log_{dataset_id}_targets.txt"""
with open(__lowerCamelCase , "w" ) as p, open(__lowerCamelCase , "w" ) as t:
# mapping function to write output
def write_to_file(__lowerCamelCase : str , __lowerCamelCase : Optional[int] ):
p.write(f"""{i}""" + "\n" )
p.write(batch["prediction"] + "\n" )
t.write(f"""{i}""" + "\n" )
t.write(batch["target"] + "\n" )
result.map(__lowerCamelCase , with_indices=__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : List[Any] = "[,?.!\-\;\:\"“%‘”�—’…–]" # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
snake_case : List[Any] = re.sub(__lowerCamelCase , "" , text.lower() )
# In addition, we can normalize the target text, e.g. removing new lines characters etc...
# note that order is important here!
snake_case : Optional[Any] = ["\n\n", "\n", " ", " "]
for t in token_sequences_to_ignore:
snake_case : Dict = " ".join(text.split(__lowerCamelCase ) )
return text
def UpperCamelCase ( __lowerCamelCase : int ):
# load dataset
snake_case : str = load_dataset(args.dataset , args.config , split=args.split , use_auth_token=__lowerCamelCase )
# for testing: only process the first two examples as a test
# dataset = dataset.select(range(10))
# load processor
snake_case : List[Any] = AutoFeatureExtractor.from_pretrained(args.model_id )
snake_case : Union[str, Any] = feature_extractor.sampling_rate
# resample audio
snake_case : Union[str, Any] = dataset.cast_column("audio" , Audio(sampling_rate=__lowerCamelCase ) )
# load eval pipeline
if args.device is None:
snake_case : List[str] = 0 if torch.cuda.is_available() else -1
snake_case : str = pipeline("automatic-speech-recognition" , model=args.model_id , device=args.device )
# map function to decode audio
def map_to_pred(__lowerCamelCase : int ):
snake_case : Dict = asr(
batch["audio"]["array"] , chunk_length_s=args.chunk_length_s , stride_length_s=args.stride_length_s )
snake_case : str = prediction["text"]
snake_case : Tuple = normalize_text(batch["sentence"] )
return batch
# run inference on all examples
snake_case : Dict = dataset.map(__lowerCamelCase , remove_columns=dataset.column_names )
# compute and log_results
# do not change function below
log_results(__lowerCamelCase , __lowerCamelCase )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
"""--model_id""", type=str, required=True, help="""Model identifier. Should be loadable with 🤗 Transformers"""
)
parser.add_argument(
"""--dataset""",
type=str,
required=True,
help="""Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets""",
)
parser.add_argument(
"""--config""", type=str, required=True, help="""Config of the dataset. *E.g.* `'en'` for Common Voice"""
)
parser.add_argument("""--split""", type=str, required=True, help="""Split of the dataset. *E.g.* `'test'`""")
parser.add_argument(
"""--chunk_length_s""", type=float, default=None, help="""Chunk length in seconds. Defaults to 5 seconds."""
)
parser.add_argument(
"""--stride_length_s""", type=float, default=None, help="""Stride of the audio chunks. Defaults to 1 second."""
)
parser.add_argument(
"""--log_outputs""", action="""store_true""", help="""If defined, write outputs to log file for analysis."""
)
parser.add_argument(
"""--device""",
type=int,
default=None,
help="""The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.""",
)
__lowerCamelCase = parser.parse_args()
main(args)
| 59
| 1
|
import argparse
from collections import defaultdict
import yaml
__lowerCamelCase = """docs/source/en/_toctree.yml"""
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
snake_case : Any = defaultdict(__lowerCamelCase )
for doc in model_doc:
counts[doc["local"]] += 1
snake_case : Optional[Any] = [key for key, value in counts.items() if value > 1]
snake_case : int = []
for duplicate_key in duplicates:
snake_case : Union[str, Any] = list({doc["title"] for doc in model_doc if doc["local"] == duplicate_key} )
if len(__lowerCamelCase ) > 1:
raise ValueError(
f"""{duplicate_key} is present several times in the documentation table of content at """
"`docs/source/en/_toctree.yml` with different *Title* values. Choose one of those and remove the "
"others." )
# Only add this once
new_doc.append({"local": duplicate_key, "title": titles[0]} )
# Add none duplicate-keys
new_doc.extend([doc for doc in model_doc if counts[doc["local"]] == 1] )
# Sort
return sorted(__lowerCamelCase , key=lambda __lowerCamelCase : s["title"].lower() )
def UpperCamelCase ( __lowerCamelCase : List[Any]=False ):
with open(__lowerCamelCase , encoding="utf-8" ) as f:
snake_case : Any = yaml.safe_load(f.read() )
# Get to the API doc
snake_case : Any = 0
while content[api_idx]["title"] != "API":
api_idx += 1
snake_case : str = content[api_idx]["sections"]
# Then to the model doc
snake_case : Dict = 0
while api_doc[model_idx]["title"] != "Models":
model_idx += 1
snake_case : Dict = api_doc[model_idx]["sections"]
snake_case : Tuple = [(idx, section) for idx, section in enumerate(__lowerCamelCase ) if "sections" in section]
snake_case : Tuple = False
for idx, modality_doc in modalities_docs:
snake_case : Dict = modality_doc["sections"]
snake_case : int = clean_model_doc_toc(__lowerCamelCase )
if old_modality_doc != new_modality_doc:
snake_case : int = True
if overwrite:
snake_case : Tuple = new_modality_doc
if diff:
if overwrite:
snake_case : int = model_doc
snake_case : Union[str, Any] = api_doc
with open(__lowerCamelCase , "w" , encoding="utf-8" ) as f:
f.write(yaml.dump(__lowerCamelCase , allow_unicode=__lowerCamelCase ) )
else:
raise ValueError(
"The model doc part of the table of content is not properly sorted, run `make style` to fix this." )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument("""--fix_and_overwrite""", action="""store_true""", help="""Whether to fix inconsistencies.""")
__lowerCamelCase = parser.parse_args()
check_model_doc(args.fix_and_overwrite)
| 59
|
from typing import Optional, Tuple, Union
import flax
import flax.linen as nn
import jax
import jax.numpy as jnp
from flax.core.frozen_dict import FrozenDict
from ..configuration_utils import ConfigMixin, flax_register_to_config
from ..utils import BaseOutput
from .embeddings_flax import FlaxTimestepEmbedding, FlaxTimesteps
from .modeling_flax_utils import FlaxModelMixin
from .unet_ad_blocks_flax import (
FlaxCrossAttnDownBlockaD,
FlaxCrossAttnUpBlockaD,
FlaxDownBlockaD,
FlaxUNetMidBlockaDCrossAttn,
FlaxUpBlockaD,
)
@flax.struct.dataclass
class UpperCAmelCase ( A_ ):
A__ : jnp.ndarray
@flax_register_to_config
class UpperCAmelCase ( nn.Module ,A_ ,A_ ):
A__ : int = 32
A__ : int = 4
A__ : int = 4
A__ : Tuple[str] = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
A__ : Tuple[str] = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
A__ : Union[bool, Tuple[bool]] = False
A__ : Tuple[int] = (3_20, 6_40, 12_80, 12_80)
A__ : int = 2
A__ : Union[int, Tuple[int]] = 8
A__ : Optional[Union[int, Tuple[int]]] = None
A__ : int = 12_80
A__ : float = 0.0
A__ : bool = False
A__ : jnp.dtype = jnp.floataa
A__ : bool = True
A__ : int = 0
A__ : bool = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : jax.random.KeyArray ) -> FrozenDict:
'''simple docstring'''
snake_case : Dict = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case : Any = jnp.zeros(snake_case__ , dtype=jnp.floataa )
snake_case : List[str] = jnp.ones((1,) , dtype=jnp.intaa )
snake_case : str = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case , snake_case : Optional[int] = jax.random.split(snake_case__ )
snake_case : Union[str, Any] = {"params": params_rng, "dropout": dropout_rng}
return self.init(snake_case__ , snake_case__ , snake_case__ , snake_case__ )["params"]
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : str = self.block_out_channels
snake_case : Optional[Any] = block_out_channels[0] * 4
if self.num_attention_heads is not None:
raise ValueError(
"At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19." )
# If `num_attention_heads` is not defined (which is the case for most models)
# it will default to `attention_head_dim`. This looks weird upon first reading it and it is.
# The reason for this behavior is to correct for incorrectly named variables that were introduced
# when this library was created. The incorrect naming was only discovered much later in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131
# Changing `attention_head_dim` to `num_attention_heads` for 40,000+ configurations is too backwards breaking
# which is why we correct for the naming here.
snake_case : Tuple = self.num_attention_heads or self.attention_head_dim
# input
snake_case : Tuple = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case : Union[str, Any] = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case : Dict = FlaxTimestepEmbedding(snake_case__ , dtype=self.dtype )
snake_case : List[str] = self.only_cross_attention
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (only_cross_attention,) * len(self.down_block_types )
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case : List[Any] = []
snake_case : Optional[int] = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case : List[Any] = output_channel
snake_case : Dict = block_out_channels[i]
snake_case : Optional[Any] = i == len(snake_case__ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case : List[Any] = FlaxCrossAttnDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , num_attention_heads=num_attention_heads[i] , add_downsample=not is_final_block , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Union[str, Any] = FlaxDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , add_downsample=not is_final_block , dtype=self.dtype , )
down_blocks.append(snake_case__ )
snake_case : Dict = down_blocks
# mid
snake_case : Optional[int] = FlaxUNetMidBlockaDCrossAttn(
in_channels=block_out_channels[-1] , dropout=self.dropout , num_attention_heads=num_attention_heads[-1] , use_linear_projection=self.use_linear_projection , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
# up
snake_case : Optional[Any] = []
snake_case : Optional[int] = list(reversed(snake_case__ ) )
snake_case : Dict = list(reversed(snake_case__ ) )
snake_case : Tuple = list(reversed(snake_case__ ) )
snake_case : Optional[Any] = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case : Optional[int] = output_channel
snake_case : List[Any] = reversed_block_out_channels[i]
snake_case : Union[str, Any] = reversed_block_out_channels[min(i + 1 , len(snake_case__ ) - 1 )]
snake_case : int = i == len(snake_case__ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case : Any = FlaxCrossAttnUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , num_attention_heads=reversed_num_attention_heads[i] , add_upsample=not is_final_block , dropout=self.dropout , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Optional[int] = FlaxUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , add_upsample=not is_final_block , dropout=self.dropout , dtype=self.dtype , )
up_blocks.append(snake_case__ )
snake_case : Optional[int] = output_channel
snake_case : Tuple = up_blocks
# out
snake_case : Optional[int] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case : List[str] = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__(self : Dict , snake_case__ : Dict , snake_case__ : Dict , snake_case__ : Optional[int] , snake_case__ : Tuple=None , snake_case__ : Union[str, Any]=None , snake_case__ : bool = True , snake_case__ : bool = False , ) -> Union[FlaxUNetaDConditionOutput, Tuple]:
'''simple docstring'''
if not isinstance(snake_case__ , jnp.ndarray ):
snake_case : List[Any] = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(snake_case__ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case : Any = timesteps.astype(dtype=jnp.floataa )
snake_case : int = jnp.expand_dims(snake_case__ , 0 )
snake_case : str = self.time_proj(snake_case__ )
snake_case : str = self.time_embedding(snake_case__ )
# 2. pre-process
snake_case : int = jnp.transpose(snake_case__ , (0, 2, 3, 1) )
snake_case : List[Any] = self.conv_in(snake_case__ )
# 3. down
snake_case : Optional[int] = (sample,)
for down_block in self.down_blocks:
if isinstance(snake_case__ , snake_case__ ):
snake_case , snake_case : List[Any] = down_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
else:
snake_case , snake_case : str = down_block(snake_case__ , snake_case__ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case : Tuple = ()
for down_block_res_sample, down_block_additional_residual in zip(
snake_case__ , snake_case__ ):
down_block_res_sample += down_block_additional_residual
new_down_block_res_samples += (down_block_res_sample,)
snake_case : Optional[int] = new_down_block_res_samples
# 4. mid
snake_case : Optional[int] = self.mid_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
if mid_block_additional_residual is not None:
sample += mid_block_additional_residual
# 5. up
for up_block in self.up_blocks:
snake_case : int = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case : Optional[Any] = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(snake_case__ , snake_case__ ):
snake_case : Optional[Any] = up_block(
snake_case__ , temb=snake_case__ , encoder_hidden_states=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train , )
else:
snake_case : Dict = up_block(snake_case__ , temb=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train )
# 6. post-process
snake_case : List[str] = self.conv_norm_out(snake_case__ )
snake_case : Any = nn.silu(snake_case__ )
snake_case : Optional[int] = self.conv_out(snake_case__ )
snake_case : Union[str, Any] = jnp.transpose(snake_case__ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=snake_case__ )
| 59
| 1
|
def UpperCamelCase ( __lowerCamelCase : int = 1000 ):
snake_case : List[Any] = 2**power
snake_case : Any = 0
while n:
snake_case , snake_case : Tuple = r + n % 10, n // 10
return r
if __name__ == "__main__":
print(solution(int(str(input()).strip())))
| 59
|
__lowerCamelCase = {
"joule": 1.0,
"kilojoule": 10_00,
"megajoule": 1_00_00_00,
"gigajoule": 10_00_00_00_00,
"wattsecond": 1.0,
"watthour": 36_00,
"kilowatthour": 3_60_00_00,
"newtonmeter": 1.0,
"calorie_nutr": 41_86.8,
"kilocalorie_nutr": 4_18_68_00.00,
"electronvolt": 1.602_176_634e-19,
"britishthermalunit_it": 10_55.0_55_85,
"footpound": 1.35_5818,
}
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : float ):
if to_type not in ENERGY_CONVERSION or from_type not in ENERGY_CONVERSION:
snake_case : List[Any] = (
f"""Incorrect 'from_type' or 'to_type' value: {from_type!r}, {to_type!r}\n"""
f"""Valid values are: {', '.join(__lowerCamelCase )}"""
)
raise ValueError(__lowerCamelCase )
return value * ENERGY_CONVERSION[from_type] / ENERGY_CONVERSION[to_type]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
import unittest
from transformers import EsmConfig, is_torch_available
from transformers.testing_utils import TestCasePlus, require_torch, slow, torch_device
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, ids_tensor, random_attention_mask
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from transformers import EsmForMaskedLM, EsmForSequenceClassification, EsmForTokenClassification, EsmModel
from transformers.models.esm.modeling_esm import (
ESM_PRETRAINED_MODEL_ARCHIVE_LIST,
EsmEmbeddings,
create_position_ids_from_input_ids,
)
class UpperCAmelCase :
def __init__(self : Tuple , snake_case__ : Optional[int] , snake_case__ : List[str]=13 , snake_case__ : List[Any]=7 , snake_case__ : Tuple=False , snake_case__ : Dict=True , snake_case__ : Optional[int]=False , snake_case__ : Dict=True , snake_case__ : Tuple=33 , snake_case__ : str=32 , snake_case__ : Tuple=5 , snake_case__ : int=4 , snake_case__ : str=37 , snake_case__ : Optional[Any]="gelu" , snake_case__ : Optional[int]=0.1 , snake_case__ : List[Any]=0.1 , snake_case__ : str=5_12 , snake_case__ : List[Any]=16 , snake_case__ : Union[str, Any]=2 , snake_case__ : List[str]=0.02 , snake_case__ : str=3 , snake_case__ : Optional[Any]=4 , snake_case__ : str=None , ) -> Tuple:
'''simple docstring'''
snake_case : int = parent
snake_case : Dict = batch_size
snake_case : List[Any] = seq_length
snake_case : List[str] = is_training
snake_case : Optional[Any] = use_input_mask
snake_case : Optional[Any] = use_token_type_ids
snake_case : Optional[Any] = use_labels
snake_case : Union[str, Any] = vocab_size
snake_case : Tuple = hidden_size
snake_case : List[Any] = num_hidden_layers
snake_case : int = num_attention_heads
snake_case : Any = intermediate_size
snake_case : Tuple = hidden_act
snake_case : Dict = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : str = type_vocab_size
snake_case : Optional[Any] = type_sequence_label_size
snake_case : Any = initializer_range
snake_case : Union[str, Any] = num_labels
snake_case : Tuple = num_choices
snake_case : str = scope
def _SCREAMING_SNAKE_CASE (self : int ) -> int:
'''simple docstring'''
snake_case : Any = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case : Union[str, Any] = None
if self.use_input_mask:
snake_case : List[Any] = random_attention_mask([self.batch_size, self.seq_length] )
snake_case : int = None
snake_case : Any = None
snake_case : Tuple = None
if self.use_labels:
snake_case : Dict = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : Optional[int] = ids_tensor([self.batch_size, self.seq_length] , self.num_labels )
snake_case : Optional[Any] = ids_tensor([self.batch_size] , self.num_choices )
snake_case : List[Any] = self.get_config()
return config, input_ids, input_mask, sequence_labels, token_labels, choice_labels
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
return EsmConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , pad_token_id=1 , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , initializer_range=self.initializer_range , )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[Any] , snake_case__ : int , snake_case__ : Union[str, Any] , snake_case__ : Optional[Any] , snake_case__ : List[str] , snake_case__ : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Tuple = EsmModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[str] = model(snake_case__ , attention_mask=snake_case__ )
snake_case : List[str] = model(snake_case__ )
snake_case : List[Any] = model(snake_case__ )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, self.seq_length, self.hidden_size) )
self.parent.assertEqual(result.pooler_output.shape , (self.batch_size, self.hidden_size) )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Dict , snake_case__ : List[str] , snake_case__ : str , snake_case__ : Optional[Any] , snake_case__ : List[Any] , snake_case__ : List[Any] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = EsmForMaskedLM(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : int = model(snake_case__ , attention_mask=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.vocab_size) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : str , snake_case__ : Union[str, Any] , snake_case__ : str , snake_case__ : Optional[int] , snake_case__ : Optional[Any] , snake_case__ : List[str] ) -> Optional[Any]:
'''simple docstring'''
snake_case : Any = self.num_labels
snake_case : int = EsmForTokenClassification(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Dict = model(snake_case__ , attention_mask=snake_case__ , labels=snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.seq_length, self.num_labels) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Tuple = self.prepare_config_and_inputs()
(
(
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) , (
snake_case
) ,
) : Tuple = config_and_inputs
snake_case : Optional[Any] = {"input_ids": input_ids, "attention_mask": input_mask}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : List[Any] = False
A__ : int = (
(
EsmForMaskedLM,
EsmModel,
EsmForSequenceClassification,
EsmForTokenClassification,
)
if is_torch_available()
else ()
)
A__ : Any = ()
A__ : Optional[Any] = (
{
"feature-extraction": EsmModel,
"fill-mask": EsmForMaskedLM,
"text-classification": EsmForSequenceClassification,
"token-classification": EsmForTokenClassification,
"zero-shot": EsmForSequenceClassification,
}
if is_torch_available()
else {}
)
A__ : List[Any] = True
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = EsmModelTester(self )
snake_case : Tuple = ConfigTester(self , config_class=snake_case__ , hidden_size=37 )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
self.config_tester.run_common_tests()
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[Any]:
'''simple docstring'''
snake_case : Dict = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = self.model_tester.prepare_config_and_inputs()
for type in ["absolute", "relative_key", "relative_key_query"]:
snake_case : Optional[int] = type
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> List[str]:
'''simple docstring'''
snake_case : Tuple = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*snake_case__ )
@slow
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
for model_name in ESM_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case : int = EsmModel.from_pretrained(snake_case__ )
self.assertIsNotNone(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> List[Any]:
'''simple docstring'''
snake_case : Dict = self.model_tester.prepare_config_and_inputs()[0]
snake_case : int = EsmEmbeddings(config=snake_case__ )
snake_case : List[Any] = torch.as_tensor([[12, 31, 13, model.padding_idx]] )
snake_case : List[Any] = torch.as_tensor(
[
[
0 + model.padding_idx + 1,
1 + model.padding_idx + 1,
2 + model.padding_idx + 1,
model.padding_idx,
]
] )
snake_case : Optional[int] = create_position_ids_from_input_ids(snake_case__ , model.padding_idx )
self.assertEqual(position_ids.shape , expected_positions.shape )
self.assertTrue(torch.all(torch.eq(snake_case__ , snake_case__ ) ) )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> List[Any]:
'''simple docstring'''
snake_case : List[Any] = self.model_tester.prepare_config_and_inputs()[0]
snake_case : Any = EsmEmbeddings(config=snake_case__ )
snake_case : Optional[int] = torch.empty(2 , 4 , 30 )
snake_case : Optional[Any] = [
0 + embeddings.padding_idx + 1,
1 + embeddings.padding_idx + 1,
2 + embeddings.padding_idx + 1,
3 + embeddings.padding_idx + 1,
]
snake_case : Dict = torch.as_tensor([expected_single_positions, expected_single_positions] )
snake_case : int = embeddings.create_position_ids_from_inputs_embeds(snake_case__ )
self.assertEqual(position_ids.shape , expected_positions.shape )
self.assertTrue(torch.all(torch.eq(snake_case__ , snake_case__ ) ) )
@unittest.skip("Esm does not support embedding resizing" )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Tuple:
'''simple docstring'''
pass
@unittest.skip("Esm does not support embedding resizing" )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Union[str, Any]:
'''simple docstring'''
pass
@unittest.skip("Will be fixed soon by reducing the size of the model used for common tests." )
def _SCREAMING_SNAKE_CASE (self : int ) -> List[Any]:
'''simple docstring'''
pass
@require_torch
class UpperCAmelCase ( A_ ):
@slow
def _SCREAMING_SNAKE_CASE (self : str ) -> str:
'''simple docstring'''
with torch.no_grad():
snake_case : Tuple = EsmForMaskedLM.from_pretrained("facebook/esm2_t6_8M_UR50D" )
model.eval()
snake_case : str = torch.tensor([[0, 1, 2, 3, 4, 5]] )
snake_case : Optional[Any] = model(snake_case__ )[0]
snake_case : List[str] = 33
snake_case : Tuple = torch.Size((1, 6, vocab_size) )
self.assertEqual(output.shape , snake_case__ )
snake_case : str = torch.tensor(
[[[8.9215, -10.5898, -6.4671], [-6.3967, -13.9114, -1.1212], [-7.7812, -13.9516, -3.7406]]] )
self.assertTrue(torch.allclose(output[:, :3, :3] , snake_case__ , atol=1e-4 ) )
@slow
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Tuple:
'''simple docstring'''
with torch.no_grad():
snake_case : Dict = EsmModel.from_pretrained("facebook/esm2_t6_8M_UR50D" )
model.eval()
snake_case : str = torch.tensor([[0, 6, 4, 13, 5, 4, 16, 12, 11, 7, 2]] )
snake_case : int = model(snake_case__ )[0]
# compare the actual values for a slice.
snake_case : Optional[Any] = torch.tensor(
[[[0.1444, 0.5413, 0.3248], [0.3034, 0.0053, 0.3108], [0.3228, -0.2499, 0.3415]]] )
self.assertTrue(torch.allclose(output[:, :3, :3] , snake_case__ , atol=1e-4 ) )
| 59
|
import logging
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
import datasets
import numpy as np
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PreTrainedTokenizer,
TFAutoModelForSequenceClassification,
TFTrainer,
TFTrainingArguments,
)
from transformers.utils import logging as hf_logging
hf_logging.set_verbosity_info()
hf_logging.enable_default_handler()
hf_logging.enable_explicit_format()
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : PreTrainedTokenizer , __lowerCamelCase : int , __lowerCamelCase : Optional[int] = None , ):
snake_case : int = {}
if train_file is not None:
snake_case : List[Any] = [train_file]
if eval_file is not None:
snake_case : Optional[int] = [eval_file]
if test_file is not None:
snake_case : Any = [test_file]
snake_case : int = datasets.load_dataset("csv" , data_files=__lowerCamelCase )
snake_case : str = list(ds[list(files.keys() )[0]].features.keys() )
snake_case : int = features_name.pop(__lowerCamelCase )
snake_case : str = list(set(ds[list(files.keys() )[0]][label_name] ) )
snake_case : str = {label: i for i, label in enumerate(__lowerCamelCase )}
snake_case : List[Any] = tokenizer.model_input_names
snake_case : List[Any] = {}
if len(__lowerCamelCase ) == 1:
for k in files.keys():
snake_case : Tuple = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
example[features_name[0]] , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" ) , batched=__lowerCamelCase , )
elif len(__lowerCamelCase ) == 2:
for k in files.keys():
snake_case : List[Any] = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
(example[features_name[0]], example[features_name[1]]) , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" , ) , batched=__lowerCamelCase , )
def gen_train():
for ex in transformed_ds[datasets.Split.TRAIN]:
snake_case : Dict = {k: v for k, v in ex.items() if k in input_names}
snake_case : Union[str, Any] = labelaid[ex[label_name]]
yield (d, label)
def gen_val():
for ex in transformed_ds[datasets.Split.VALIDATION]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : Any = labelaid[ex[label_name]]
yield (d, label)
def gen_test():
for ex in transformed_ds[datasets.Split.TEST]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : List[str] = labelaid[ex[label_name]]
yield (d, label)
snake_case : int = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TRAIN in transformed_ds
else None
)
if train_ds is not None:
snake_case : Optional[Any] = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) )
snake_case : Tuple = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.VALIDATION in transformed_ds
else None
)
if val_ds is not None:
snake_case : List[str] = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) )
snake_case : Optional[int] = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TEST in transformed_ds
else None
)
if test_ds is not None:
snake_case : str = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) )
return train_ds, val_ds, test_ds, labelaid
__lowerCamelCase = logging.getLogger(__name__)
@dataclass
class UpperCAmelCase :
A__ : int = field(metadata={"help": "Which column contains the label"} )
A__ : str = field(default=A_ ,metadata={"help": "The path of the training file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the development file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the test file"} )
A__ : int = field(
default=1_28 ,metadata={
"help": (
"The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
} ,)
A__ : bool = field(
default=A_ ,metadata={"help": "Overwrite the cached training and evaluation sets"} )
@dataclass
class UpperCAmelCase :
A__ : str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained config name or path if not the same as model_name"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"} )
A__ : bool = field(default=A_ ,metadata={"help": "Set this flag to use fast tokenization."} )
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
# or just modify its tokenizer_config.json.
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} ,)
def UpperCamelCase ( ):
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
snake_case : List[Any] = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) )
snake_case , snake_case , snake_case : int = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir )
and os.listdir(training_args.output_dir )
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
f"""Output directory ({training_args.output_dir}) already exists and is not empty. Use"""
" --overwrite_output_dir to overcome." )
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(
f"""n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, """
f"""16-bits training: {training_args.fpaa}""" )
logger.info(f"""Training/evaluation parameters {training_args}""" )
# Load pretrained model and tokenizer
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
snake_case : Tuple = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path , cache_dir=model_args.cache_dir , )
snake_case , snake_case , snake_case , snake_case : Tuple = get_tfds(
train_file=data_args.train_file , eval_file=data_args.dev_file , test_file=data_args.test_file , tokenizer=__lowerCamelCase , label_column_id=data_args.label_column_id , max_seq_length=data_args.max_seq_length , )
snake_case : Optional[Any] = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path , num_labels=len(__lowerCamelCase ) , labelaid=__lowerCamelCase , idalabel={id: label for label, id in labelaid.items()} , finetuning_task="text-classification" , cache_dir=model_args.cache_dir , )
with training_args.strategy.scope():
snake_case : int = TFAutoModelForSequenceClassification.from_pretrained(
model_args.model_name_or_path , from_pt=bool(".bin" in model_args.model_name_or_path ) , config=__lowerCamelCase , cache_dir=model_args.cache_dir , )
def compute_metrics(__lowerCamelCase : EvalPrediction ) -> Dict:
snake_case : Optional[int] = np.argmax(p.predictions , axis=1 )
return {"acc": (preds == p.label_ids).mean()}
# Initialize our Trainer
snake_case : int = TFTrainer(
model=__lowerCamelCase , args=__lowerCamelCase , train_dataset=__lowerCamelCase , eval_dataset=__lowerCamelCase , compute_metrics=__lowerCamelCase , )
# Training
if training_args.do_train:
trainer.train()
trainer.save_model()
tokenizer.save_pretrained(training_args.output_dir )
# Evaluation
snake_case : int = {}
if training_args.do_eval:
logger.info("*** Evaluate ***" )
snake_case : Any = trainer.evaluate()
snake_case : List[Any] = os.path.join(training_args.output_dir , "eval_results.txt" )
with open(__lowerCamelCase , "w" ) as writer:
logger.info("***** Eval results *****" )
for key, value in result.items():
logger.info(f""" {key} = {value}""" )
writer.write(f"""{key} = {value}\n""" )
results.update(__lowerCamelCase )
return results
if __name__ == "__main__":
main()
| 59
| 1
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
__lowerCamelCase = {
"""configuration_instructblip""": [
"""INSTRUCTBLIP_PRETRAINED_CONFIG_ARCHIVE_MAP""",
"""InstructBlipConfig""",
"""InstructBlipQFormerConfig""",
"""InstructBlipVisionConfig""",
],
"""processing_instructblip""": ["""InstructBlipProcessor"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""INSTRUCTBLIP_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""InstructBlipQFormerModel""",
"""InstructBlipPreTrainedModel""",
"""InstructBlipForConditionalGeneration""",
"""InstructBlipVisionModel""",
]
if TYPE_CHECKING:
from .configuration_instructblip import (
INSTRUCTBLIP_PRETRAINED_CONFIG_ARCHIVE_MAP,
InstructBlipConfig,
InstructBlipQFormerConfig,
InstructBlipVisionConfig,
)
from .processing_instructblip import InstructBlipProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_instructblip import (
INSTRUCTBLIP_PRETRAINED_MODEL_ARCHIVE_LIST,
InstructBlipForConditionalGeneration,
InstructBlipPreTrainedModel,
InstructBlipQFormerModel,
InstructBlipVisionModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
|
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Any ) -> List[str]:
'''simple docstring'''
snake_case : int = tempfile.mkdtemp()
# fmt: off
snake_case : Optional[int] = ["[UNK]", "[CLS]", "[SEP]", "[PAD]", "[MASK]", "want", "##want", "##ed", "wa", "un", "runn", "##ing", ",", "low", "lowest"]
# fmt: on
snake_case : List[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] ) )
snake_case : int = {
"do_resize": True,
"size": {"height": 18, "width": 18},
"do_normalize": True,
"image_mean": [0.5, 0.5, 0.5],
"image_std": [0.5, 0.5, 0.5],
}
snake_case : Optional[Any] = os.path.join(self.tmpdirname , snake_case__ )
with open(self.image_processor_file , "w" , encoding="utf-8" ) as fp:
json.dump(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : str ) -> Optional[int]:
'''simple docstring'''
return BertTokenizer.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : List[str] ) -> int:
'''simple docstring'''
return ViTImageProcessor.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
shutil.rmtree(self.tmpdirname )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : List[Any] = [np.random.randint(2_55 , size=(3, 30, 4_00) , dtype=np.uinta )]
snake_case : Optional[int] = [Image.fromarray(np.moveaxis(snake_case__ , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.get_tokenizer()
snake_case : Optional[Any] = self.get_image_processor()
snake_case : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
processor.save_pretrained(self.tmpdirname )
snake_case : Any = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
snake_case : Optional[int] = self.get_tokenizer(bos_token="(BOS)" , eos_token="(EOS)" )
snake_case : Tuple = self.get_image_processor(do_normalize=snake_case__ , padding_value=1.0 )
snake_case : List[str] = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token="(BOS)" , eos_token="(EOS)" , do_normalize=snake_case__ , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Optional[int] = self.get_tokenizer()
snake_case : List[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = self.prepare_image_inputs()
snake_case : str = image_processor(snake_case__ , return_tensors="np" )
snake_case : Any = processor(images=snake_case__ , return_tensors="np" )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.get_image_processor()
snake_case : int = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Tuple = "lower newer"
snake_case : Tuple = processor(text=snake_case__ )
snake_case : Union[str, Any] = tokenizer(snake_case__ )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[int]:
'''simple docstring'''
snake_case : List[Any] = self.get_image_processor()
snake_case : Dict = self.get_tokenizer()
snake_case : Dict = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : int = "lower newer"
snake_case : Dict = self.prepare_image_inputs()
snake_case : Union[str, Any] = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , ["input_ids", "token_type_ids", "attention_mask", "pixel_values"] )
# test if it raises when no input is passed
with self.assertRaises(snake_case__ ):
processor()
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : Tuple = self.get_image_processor()
snake_case : Optional[Any] = self.get_tokenizer()
snake_case : Tuple = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : List[str] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
snake_case : List[Any] = processor.batch_decode(snake_case__ )
snake_case : Union[str, Any] = tokenizer.batch_decode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Union[str, Any] = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = "lower newer"
snake_case : List[Any] = self.prepare_image_inputs()
snake_case : Tuple = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 59
| 1
|
from __future__ import annotations
import random
import unittest
from transformers import TransfoXLConfig, 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 (
TF_TRANSFO_XL_PRETRAINED_MODEL_ARCHIVE_LIST,
TFTransfoXLForSequenceClassification,
TFTransfoXLLMHeadModel,
TFTransfoXLModel,
)
class UpperCAmelCase :
def __init__(self : List[str] , snake_case__ : str , ) -> int:
'''simple docstring'''
snake_case : Optional[int] = parent
snake_case : List[str] = 13
snake_case : Tuple = 7
snake_case : str = 30
snake_case : str = self.seq_length + self.mem_len
snake_case : List[Any] = 15
snake_case : List[Any] = True
snake_case : int = True
snake_case : Dict = 99
snake_case : Optional[Any] = [10, 50, 80]
snake_case : Any = 32
snake_case : List[Any] = 32
snake_case : Tuple = 4
snake_case : Union[str, Any] = 8
snake_case : Optional[int] = 1_28
snake_case : List[str] = 2
snake_case : Optional[int] = 2
snake_case : Optional[Any] = None
snake_case : Optional[Any] = 1
snake_case : str = 0
snake_case : Any = 3
snake_case : Union[str, Any] = self.vocab_size - 1
snake_case : str = 0.01
def _SCREAMING_SNAKE_CASE (self : Any ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case : Union[str, Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case : Tuple = None
if self.use_labels:
snake_case : Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case : str = TransfoXLConfig(
vocab_size=self.vocab_size , mem_len=self.mem_len , clamp_len=self.clamp_len , cutoffs=self.cutoffs , d_model=self.hidden_size , d_embed=self.d_embed , n_head=self.num_attention_heads , d_head=self.d_head , d_inner=self.d_inner , div_val=self.div_val , n_layer=self.num_hidden_layers , eos_token_id=self.eos_token_id , pad_token_id=self.vocab_size - 1 , init_range=self.init_range , num_labels=self.num_labels , )
return (config, input_ids_a, input_ids_a, lm_labels)
def _SCREAMING_SNAKE_CASE (self : int ) -> int:
'''simple docstring'''
random.seed(self.seed )
tf.random.set_seed(self.seed )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Any , snake_case__ : List[str] , snake_case__ : Tuple , snake_case__ : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Optional[int] = TFTransfoXLModel(snake_case__ )
snake_case , snake_case : List[Any] = model(snake_case__ ).to_tuple()
snake_case : Optional[Any] = {"input_ids": input_ids_a, "mems": mems_a}
snake_case , snake_case : str = model(snake_case__ ).to_tuple()
self.parent.assertEqual(hidden_states_a.shape , (self.batch_size, self.seq_length, self.hidden_size) )
self.parent.assertEqual(hidden_states_a.shape , (self.batch_size, self.seq_length, self.hidden_size) )
self.parent.assertListEqual(
[mem.shape for mem in mems_a] , [(self.mem_len, self.batch_size, self.hidden_size)] * self.num_hidden_layers , )
self.parent.assertListEqual(
[mem.shape for mem in mems_a] , [(self.mem_len, self.batch_size, self.hidden_size)] * self.num_hidden_layers , )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Optional[int] , snake_case__ : List[Any] , snake_case__ : int , snake_case__ : Any ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[Any] = TFTransfoXLLMHeadModel(snake_case__ )
snake_case , snake_case : Dict = model(snake_case__ ).to_tuple()
snake_case : Any = {"input_ids": input_ids_a, "labels": lm_labels}
snake_case , snake_case : str = model(snake_case__ ).to_tuple()
snake_case , snake_case : List[Any] = model([input_ids_a, mems_a] ).to_tuple()
snake_case : List[Any] = {"input_ids": input_ids_a, "mems": mems_a, "labels": lm_labels}
snake_case , snake_case : Optional[int] = model(snake_case__ ).to_tuple()
self.parent.assertEqual(lm_logits_a.shape , (self.batch_size, self.seq_length, self.vocab_size) )
self.parent.assertListEqual(
[mem.shape for mem in mems_a] , [(self.mem_len, self.batch_size, self.hidden_size)] * self.num_hidden_layers , )
self.parent.assertEqual(lm_logits_a.shape , (self.batch_size, self.seq_length, self.vocab_size) )
self.parent.assertListEqual(
[mem.shape for mem in mems_a] , [(self.mem_len, self.batch_size, self.hidden_size)] * self.num_hidden_layers , )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Any , snake_case__ : Tuple , snake_case__ : Optional[Any] , snake_case__ : str ) -> List[Any]:
'''simple docstring'''
snake_case : int = TFTransfoXLForSequenceClassification(snake_case__ )
snake_case : Union[str, Any] = model(snake_case__ )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.num_labels) )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Dict:
'''simple docstring'''
snake_case : Tuple = self.prepare_config_and_inputs()
((snake_case) , (snake_case) , (snake_case) , (snake_case)) : str = config_and_inputs
snake_case : List[Any] = {"input_ids": input_ids_a}
return config, inputs_dict
@require_tf
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : Dict = (
(TFTransfoXLModel, TFTransfoXLLMHeadModel, TFTransfoXLForSequenceClassification) if is_tf_available() else ()
)
A__ : List[Any] = () if is_tf_available() else ()
A__ : Optional[Any] = (
{
"feature-extraction": TFTransfoXLModel,
"text-classification": TFTransfoXLForSequenceClassification,
"text-generation": TFTransfoXLLMHeadModel,
"zero-shot": TFTransfoXLForSequenceClassification,
}
if is_tf_available()
else {}
)
# TODO: add this test when TFTransfoXLLMHead has a linear output layer implemented
A__ : Optional[Any] = False
A__ : List[str] = False
A__ : Any = False
A__ : List[str] = False
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : int , snake_case__ : List[Any] , snake_case__ : Any , snake_case__ : List[str] , snake_case__ : Tuple ) -> Any:
'''simple docstring'''
if pipeline_test_casse_name == "TextGenerationPipelineTests":
# Get `ValueError: AttributeError: 'NoneType' object has no attribute 'new_ones'` or `AssertionError`.
# `TransfoXLConfig` was never used in pipeline tests: cannot create a simple
# tokenizer.
return True
return False
def _SCREAMING_SNAKE_CASE (self : Any ) -> int:
'''simple docstring'''
snake_case : Dict = TFTransfoXLModelTester(self )
snake_case : str = ConfigTester(self , config_class=snake_case__ , d_embed=37 )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
self.config_tester.run_common_tests()
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
self.model_tester.set_seed()
snake_case : List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_transfo_xl_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Dict:
'''simple docstring'''
self.model_tester.set_seed()
snake_case : Tuple = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_transfo_xl_lm_head(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_transfo_xl_for_sequence_classification(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : List[Any] = [TFTransfoXLForSequenceClassification]
for model_class in self.all_model_classes:
snake_case : str = model_class(snake_case__ )
assert isinstance(model.get_input_embeddings() , tf.keras.layers.Layer )
if model_class in list_other_models_with_output_ebd:
snake_case : Dict = model.get_output_embeddings()
assert isinstance(snake_case__ , tf.keras.layers.Layer )
snake_case : Optional[int] = model.get_bias()
assert name is None
else:
snake_case : Union[str, Any] = model.get_output_embeddings()
assert x is None
snake_case : Optional[int] = model.get_bias()
assert name is None
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Dict:
'''simple docstring'''
pass
@slow
def _SCREAMING_SNAKE_CASE (self : int ) -> Union[str, Any]:
'''simple docstring'''
for model_name in TF_TRANSFO_XL_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case : Any = TFTransfoXLModel.from_pretrained(snake_case__ )
self.assertIsNotNone(snake_case__ )
@unittest.skip(reason="This model doesn't play well with fit() due to not returning a single loss." )
def _SCREAMING_SNAKE_CASE (self : str ) -> Any:
'''simple docstring'''
pass
@require_tf
class UpperCAmelCase ( unittest.TestCase ):
@unittest.skip("Skip test until #12651 is resolved." )
@slow
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Tuple:
'''simple docstring'''
snake_case : Any = TFTransfoXLLMHeadModel.from_pretrained("transfo-xl-wt103" )
# fmt: off
snake_case : List[str] = tf.convert_to_tensor([[33,12_97,2,1,10_09,4,11_09,1_17_39,47_62,3_58,5,25,2_45,22,17_06,17,2_00_98,5,32_15,21,37,11_10,3,13,10_41,4,24,6_03,4_90,2,7_14_77,2_00_98,10_44_47,2,2_09_61,1,26_04,4,1,3_29,3,62_24,8_31,1_60_02,2,8,6_03,7_89_67,2_95_46,23,8_03,20,25,4_16,5,8,2_32,4,2_77,6,18_55,46_01,3,2_95_46,54,8,36_09,5,5_72_11,49,4,1,2_77,18,8,17_55,1_56_91,3,3_41,25,4_16,6_93,4_25_73,71,17,4_01,94,31,1_79_19,2,2_95_46,78_73,18,1,4_35,23,1_10_11,7_55,5,51_67,3,79_83,98,84,2,2_95_46,32_67,8,36_09,4,1,48_65,10_75,2,60_87,71,6,3_46,8,58_54,3,2_95_46,8_24,14_00,18_68,2,19,1_60,2,3_11,8,54_96,2,2_09_20,17,25,1_50_97,3,24,24,0]] , dtype=tf.intaa ) # noqa: E231
# fmt: on
# In 1991 , the remains of Russian Tsar Nicholas II and his family
# ( except for Alexei and Maria ) are discovered .
# The voice of Nicholas's young son , Tsarevich Alexei Nikolaevich , narrates the
# remainder of the story . 1883 Western Siberia ,
# a young Grigori Rasputin is asked by his father and a group of men to perform magic .
# Rasputin has a vision and denounces one of the men as a horse thief . Although his
# father initially slaps him for making such an accusation , Rasputin watches as the
# man is chased outside and beaten . Twenty years later , Rasputin sees a vision of
# the Virgin Mary , prompting him to become a priest . Rasputin quickly becomes famous ,
# with people , even a bishop , begging for his blessing . <eod> </s> <eos>
# fmt: off
snake_case : int = [33,12_97,2,1,10_09,4,11_09,1_17_39,47_62,3_58,5,25,2_45,22,17_06,17,2_00_98,5,32_15,21,37,11_10,3,13,10_41,4,24,6_03,4_90,2,7_14_77,2_00_98,10_44_47,2,2_09_61,1,26_04,4,1,3_29,3,62_24,8_31,1_60_02,2,8,6_03,7_89_67,2_95_46,23,8_03,20,25,4_16,5,8,2_32,4,2_77,6,18_55,46_01,3,2_95_46,54,8,36_09,5,5_72_11,49,4,1,2_77,18,8,17_55,1_56_91,3,3_41,25,4_16,6_93,4_25_73,71,17,4_01,94,31,1_79_19,2,2_95_46,78_73,18,1,4_35,23,1_10_11,7_55,5,51_67,3,79_83,98,84,2,2_95_46,32_67,8,36_09,4,1,48_65,10_75,2,60_87,71,6,3_46,8,58_54,3,2_95_46,8_24,14_00,18_68,2,19,1_60,2,3_11,8,54_96,2,2_09_20,17,25,1_50_97,3,24,24,0,33,1,18_57,2,1,10_09,4,11_09,1_17_39,47_62,3_58,5,25,2_45,28,11_10,3,13,10_41,4,24,6_03,4_90,2,7_14_77,2_00_98,10_44_47,2,2_09_61,1,26_04,4,1,3_29,3,0] # noqa: E231
# fmt: on
# In 1991, the remains of Russian Tsar Nicholas II and his family (
# except for Alexei and Maria ) are discovered. The voice of young son,
# Tsarevich Alexei Nikolaevich, narrates the remainder of the story.
# 1883 Western Siberia, a young Grigori Rasputin is asked by his father
# and a group of men to perform magic. Rasputin has a vision and
# denounces one of the men as a horse thief. Although his father initially
# slaps him for making such an accusation, Rasputin watches as the man
# is chased outside and beaten. Twenty years later, Rasputin sees a vision
# of the Virgin Mary, prompting him to become a priest.
# Rasputin quickly becomes famous, with people, even a bishop, begging for
# his blessing. <unk> <unk> <eos> In the 1990s, the remains of Russian Tsar
# Nicholas II and his family were discovered. The voice of <unk> young son,
# Tsarevich Alexei Nikolaevich, narrates the remainder of the story.<eos>
snake_case : Optional[int] = model.generate(snake_case__ , max_length=2_00 , do_sample=snake_case__ )
self.assertListEqual(output_ids[0].numpy().tolist() , snake_case__ )
| 59
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__lowerCamelCase = {
"""configuration_biogpt""": ["""BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP""", """BioGptConfig"""],
"""tokenization_biogpt""": ["""BioGptTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""BioGptForCausalLM""",
"""BioGptForTokenClassification""",
"""BioGptForSequenceClassification""",
"""BioGptModel""",
"""BioGptPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_biogpt import BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, BioGptConfig
from .tokenization_biogpt import BioGptTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_biogpt import (
BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
BioGptForCausalLM,
BioGptForSequenceClassification,
BioGptForTokenClassification,
BioGptModel,
BioGptPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
| 1
|
import argparse
import torch
# Step 1. clone https://github.com/microsoft/unilm
# Step 2. git checkout to https://github.com/microsoft/unilm/commit/b94ec76c36f02fb2b0bf0dcb0b8554a2185173cd
# Step 3. cd unilm
# Step 4. ln -s $(realpath wavlm/modules.py) ./ # create simlink
# import classes
from unilm.wavlm.WavLM import WavLM as WavLMOrig
from unilm.wavlm.WavLM import WavLMConfig as WavLMConfigOrig
from transformers import WavLMConfig, WavLMModel, logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""post_extract_proj""": """feature_projection.projection""",
"""encoder.pos_conv.0""": """encoder.pos_conv_embed.conv""",
"""self_attn.k_proj""": """encoder.layers.*.attention.k_proj""",
"""self_attn.v_proj""": """encoder.layers.*.attention.v_proj""",
"""self_attn.q_proj""": """encoder.layers.*.attention.q_proj""",
"""self_attn.out_proj""": """encoder.layers.*.attention.out_proj""",
"""self_attn.grep_linear""": """encoder.layers.*.attention.gru_rel_pos_linear""",
"""self_attn.relative_attention_bias""": """encoder.layers.*.attention.rel_attn_embed""",
"""self_attn.grep_a""": """encoder.layers.*.attention.gru_rel_pos_const""",
"""self_attn_layer_norm""": """encoder.layers.*.layer_norm""",
"""fc1""": """encoder.layers.*.feed_forward.intermediate_dense""",
"""fc2""": """encoder.layers.*.feed_forward.output_dense""",
"""final_layer_norm""": """encoder.layers.*.final_layer_norm""",
"""encoder.layer_norm""": """encoder.layer_norm""",
"""w2v_model.layer_norm""": """feature_projection.layer_norm""",
"""quantizer.weight_proj""": """quantizer.weight_proj""",
"""quantizer.vars""": """quantizer.codevectors""",
"""project_q""": """project_q""",
"""final_proj""": """project_hid""",
"""w2v_encoder.proj""": """ctc_proj""",
"""mask_emb""": """masked_spec_embed""",
}
__lowerCamelCase = [
"""ctc_proj""",
"""quantizer.weight_proj""",
"""quantizer.codevectors""",
"""project_q""",
"""project_hid""",
]
def UpperCamelCase ( __lowerCamelCase : List[Any] , __lowerCamelCase : Any , __lowerCamelCase : List[str] , __lowerCamelCase : Tuple , __lowerCamelCase : int ):
for attribute in key.split("." ):
snake_case : int = getattr(__lowerCamelCase , __lowerCamelCase )
if weight_type is not None:
snake_case : Union[str, Any] = getattr(__lowerCamelCase , __lowerCamelCase ).shape
else:
snake_case : Optional[int] = hf_pointer.shape
assert hf_shape == value.shape, (
f"""Shape of hf {key + '.' + weight_type if weight_type is not None else ''} is {hf_shape}, but should be"""
f""" {value.shape} for {full_name}"""
)
if weight_type == "weight":
snake_case : Optional[Any] = value
elif weight_type == "weight_g":
snake_case : str = value
elif weight_type == "weight_v":
snake_case : Dict = value
elif weight_type == "bias":
snake_case : Dict = value
else:
snake_case : Tuple = value
logger.info(f"""{key + '.' + weight_type if weight_type is not None else ''} was initialized from {full_name}.""" )
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Any ):
snake_case : List[Any] = []
snake_case : List[Any] = fairseq_model.state_dict()
snake_case : str = hf_model.feature_extractor
for name, value in fairseq_dict.items():
snake_case : Tuple = False
if "conv_layers" in name:
load_conv_layer(
__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , hf_model.config.feat_extract_norm == "group" , )
snake_case : Optional[Any] = True
else:
for key, mapped_key in MAPPING.items():
if key in name or key.split("w2v_model." )[-1] == name.split("." )[0]:
snake_case : Dict = True
if "*" in mapped_key:
snake_case : Union[str, Any] = name.split(__lowerCamelCase )[0].split("." )[-2]
snake_case : Dict = mapped_key.replace("*" , __lowerCamelCase )
if "weight_g" in name:
snake_case : int = "weight_g"
elif "weight_v" in name:
snake_case : Dict = "weight_v"
elif "bias" in name and "relative_attention_bias" not in name:
snake_case : Dict = "bias"
elif "weight" in name:
# TODO: don't match quantizer.weight_proj
snake_case : List[Any] = "weight"
else:
snake_case : List[str] = None
set_recursively(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
continue
if not is_used:
unused_weights.append(__lowerCamelCase )
logger.warning(f"""Unused weights: {unused_weights}""" )
def UpperCamelCase ( __lowerCamelCase : Any , __lowerCamelCase : Tuple , __lowerCamelCase : Optional[int] , __lowerCamelCase : Any , __lowerCamelCase : Any ):
snake_case : Dict = full_name.split("conv_layers." )[-1]
snake_case : List[Any] = name.split("." )
snake_case : Tuple = int(items[0] )
snake_case : List[Any] = int(items[1] )
if type_id == 0:
if "bias" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].conv.bias.data.shape, (
f"""{full_name} has size {value.shape}, but"""
f""" {feature_extractor.conv_layers[layer_id].conv.bias.data.shape} was found."""
)
snake_case : Optional[Any] = value
logger.info(f"""Feat extract conv layer {layer_id} was initialized from {full_name}.""" )
elif "weight" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].conv.weight.data.shape, (
f"""{full_name} has size {value.shape}, but"""
f""" {feature_extractor.conv_layers[layer_id].conv.weight.data.shape} was found."""
)
snake_case : List[str] = value
logger.info(f"""Feat extract conv layer {layer_id} was initialized from {full_name}.""" )
elif (type_id == 2 and not use_group_norm) or (type_id == 2 and layer_id == 0 and use_group_norm):
if "bias" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.bias.data.shape, (
f"""{full_name} has size {value.shape}, but {feature_extractor[layer_id].layer_norm.bias.data.shape} was"""
" found."
)
snake_case : str = value
logger.info(f"""Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.""" )
elif "weight" in name:
assert value.shape == feature_extractor.conv_layers[layer_id].layer_norm.weight.data.shape, (
f"""{full_name} has size {value.shape}, but"""
f""" {feature_extractor[layer_id].layer_norm.weight.data.shape} was found."""
)
snake_case : List[str] = value
logger.info(f"""Feat extract layer norm weight of layer {layer_id} was initialized from {full_name}.""" )
else:
unused_weights.append(__lowerCamelCase )
@torch.no_grad()
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : List[str] , __lowerCamelCase : Union[str, Any]=None ):
# load the pre-trained checkpoints
snake_case : Optional[int] = torch.load(__lowerCamelCase )
snake_case : List[Any] = WavLMConfigOrig(checkpoint["cfg"] )
snake_case : Optional[int] = WavLMOrig(__lowerCamelCase )
model.load_state_dict(checkpoint["model"] )
model.eval()
if config_path is not None:
snake_case : str = WavLMConfig.from_pretrained(__lowerCamelCase )
else:
snake_case : Optional[int] = WavLMConfig()
snake_case : Any = WavLMModel(__lowerCamelCase )
recursively_load_weights(__lowerCamelCase , __lowerCamelCase )
hf_wavlm.save_pretrained(__lowerCamelCase )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument("""--pytorch_dump_folder_path""", default=None, type=str, help="""Path to the output PyTorch model.""")
parser.add_argument("""--checkpoint_path""", default=None, type=str, help="""Path to fairseq checkpoint""")
parser.add_argument("""--config_path""", default=None, type=str, help="""Path to hf config.json of model to convert""")
__lowerCamelCase = parser.parse_args()
convert_wavlm_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path)
| 59
|
import collections
import inspect
import unittest
from typing import Dict, List, Tuple
from transformers import MaskFormerSwinConfig
from transformers.testing_utils import require_torch, require_torch_multi_gpu, torch_device
from transformers.utils import is_torch_available
from ...test_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import MaskFormerSwinBackbone
from transformers.models.maskformer import MaskFormerSwinModel
class UpperCAmelCase :
def __init__(self : Dict , snake_case__ : Dict , snake_case__ : Any=13 , snake_case__ : Any=32 , snake_case__ : Optional[Any]=2 , snake_case__ : Union[str, Any]=3 , snake_case__ : List[Any]=16 , snake_case__ : int=[1, 2, 1] , snake_case__ : Dict=[2, 2, 4] , snake_case__ : Dict=2 , snake_case__ : Tuple=2.0 , snake_case__ : Optional[int]=True , snake_case__ : Union[str, Any]=0.0 , snake_case__ : Any=0.0 , snake_case__ : Union[str, Any]=0.1 , snake_case__ : int="gelu" , snake_case__ : Optional[int]=False , snake_case__ : List[Any]=True , snake_case__ : List[str]=0.02 , snake_case__ : int=1e-5 , snake_case__ : List[str]=True , snake_case__ : Union[str, Any]=None , snake_case__ : List[Any]=True , snake_case__ : Optional[Any]=10 , snake_case__ : Optional[Any]=8 , snake_case__ : Any=["stage1", "stage2", "stage3"] , snake_case__ : Tuple=[1, 2, 3] , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = parent
snake_case : Optional[int] = batch_size
snake_case : Union[str, Any] = image_size
snake_case : Dict = patch_size
snake_case : Optional[Any] = num_channels
snake_case : Union[str, Any] = embed_dim
snake_case : int = depths
snake_case : List[str] = num_heads
snake_case : Union[str, Any] = window_size
snake_case : Union[str, Any] = mlp_ratio
snake_case : List[Any] = qkv_bias
snake_case : List[Any] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Union[str, Any] = drop_path_rate
snake_case : int = hidden_act
snake_case : Optional[int] = use_absolute_embeddings
snake_case : int = patch_norm
snake_case : Union[str, Any] = layer_norm_eps
snake_case : Any = initializer_range
snake_case : Optional[Any] = is_training
snake_case : Tuple = scope
snake_case : Optional[int] = use_labels
snake_case : Optional[Any] = type_sequence_label_size
snake_case : Union[str, Any] = encoder_stride
snake_case : Any = out_features
snake_case : Tuple = out_indices
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case : Optional[Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case : int = None
if self.use_labels:
snake_case : Tuple = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : Dict = self.get_config()
return config, pixel_values, labels
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> int:
'''simple docstring'''
return MaskFormerSwinConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , depths=self.depths , num_heads=self.num_heads , window_size=self.window_size , mlp_ratio=self.mlp_ratio , qkv_bias=self.qkv_bias , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , drop_path_rate=self.drop_path_rate , hidden_act=self.hidden_act , use_absolute_embeddings=self.use_absolute_embeddings , path_norm=self.patch_norm , layer_norm_eps=self.layer_norm_eps , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , out_features=self.out_features , out_indices=self.out_indices , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : List[Any] , snake_case__ : List[str] , snake_case__ : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
snake_case : Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case : int = int(config.embed_dim * 2 ** (len(config.depths ) - 1) )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, expected_seq_len, expected_dim) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = MaskFormerSwinBackbone(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [13, 16, 16, 16] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , [16, 32, 64] )
# verify ValueError
with self.parent.assertRaises(snake_case__ ):
snake_case : Tuple = ["stem"]
snake_case : List[Any] = MaskFormerSwinBackbone(config=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self.prepare_config_and_inputs()
snake_case , snake_case , snake_case : List[Any] = config_and_inputs
snake_case : int = {"pixel_values": pixel_values}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : List[str] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
A__ : str = {"feature-extraction": MaskFormerSwinModel} if is_torch_available() else {}
A__ : Optional[Any] = False
A__ : List[Any] = False
A__ : List[str] = False
A__ : List[str] = False
A__ : Union[str, Any] = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = MaskFormerSwinModelTester(self )
snake_case : Optional[int] = ConfigTester(self , config_class=snake_case__ , embed_dim=37 )
@require_torch_multi_gpu
@unittest.skip(
reason=(
"`MaskFormerSwinModel` outputs `hidden_states_spatial_dimensions` which doesn't work well with"
" `nn.DataParallel`"
) )
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : str ) -> List[str]:
'''simple docstring'''
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
return
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Dict:
'''simple docstring'''
snake_case : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*snake_case__ )
@unittest.skip("Swin does not use inputs_embeds" )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
pass
@unittest.skip("Swin does not support feedforward chunking" )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Dict:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : int = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case : List[Any] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , nn.Linear ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case , snake_case : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : str = model_class(snake_case__ )
snake_case : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case : Optional[Any] = [*signature.parameters.keys()]
snake_case : Tuple = ["pixel_values"]
self.assertListEqual(arg_names[:1] , snake_case__ )
@unittest.skip(reason="MaskFormerSwin is only used as backbone and doesn't support output_attentions" )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="MaskFormerSwin is only used as an internal backbone" )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
with torch.no_grad():
snake_case : Any = model(**self._prepare_for_class(snake_case__ , snake_case__ ) )
snake_case : int = outputs.hidden_states
snake_case : Union[str, Any] = getattr(
self.model_tester , "expected_num_hidden_layers" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case__ ) , snake_case__ )
# Swin has a different seq_length
snake_case : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : Tuple = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : int = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
for model_class in self.all_model_classes:
snake_case : int = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Dict = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case , snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Any = 3
snake_case : List[str] = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
snake_case : Tuple = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : str = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case : str = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes:
snake_case : str = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Optional[Any] = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
@unittest.skip(reason="MaskFormerSwin doesn't have pretrained checkpoints" )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : int ) -> str:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case__ : Union[str, Any] ):
snake_case : Any = 0
return t
def check_equivalence(snake_case__ : Union[str, Any] , snake_case__ : int , snake_case__ : List[str] , snake_case__ : Optional[int]={} ):
with torch.no_grad():
snake_case : Optional[Any] = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ )
snake_case : Tuple = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ ).to_tuple()
def recursive_check(snake_case__ : List[str] , snake_case__ : Optional[Any] ):
if isinstance(snake_case__ , (List, Tuple) ):
for tuple_iterable_value, dict_iterable_value in zip(snake_case__ , snake_case__ ):
recursive_check(snake_case__ , snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values() , dict_object.values() ):
recursive_check(snake_case__ , snake_case__ )
elif tuple_object is None:
return
else:
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(snake_case__ ) , set_nan_tensor_to_zero(snake_case__ ) , atol=1e-5 ) , msg=(
"Tuple and dict output are not equal. Difference:"
f""" {torch.max(torch.abs(tuple_object - dict_object ) )}. Tuple has `nan`:"""
f""" {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}. Dict has"""
f""" `nan`: {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}."""
) , )
recursive_check(snake_case__ , snake_case__ )
for model_class in self.all_model_classes:
snake_case : Optional[int] = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Union[str, Any] = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : Optional[Any] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Dict = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : List[Any] = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
snake_case : Any = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : List[str] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
@require_torch
class UpperCAmelCase ( unittest.TestCase ,A_ ):
A__ : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
A__ : int = MaskFormerSwinConfig
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModelTester(self )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case : Dict = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Optional[int] = inputs_dict["pixel_values"].shape[0]
for backbone_class in self.all_model_classes:
snake_case : Optional[int] = backbone_class(snake_case__ )
backbone.to(snake_case__ )
backbone.eval()
snake_case : Union[str, Any] = backbone(**snake_case__ )
# Test default outputs and verify feature maps
self.assertIsInstance(outputs.feature_maps , snake_case__ )
self.assertTrue(len(outputs.feature_maps ) == len(backbone.channels ) )
for feature_map, n_channels in zip(outputs.feature_maps , backbone.channels ):
self.assertTrue(feature_map.shape[:2] , (batch_size, n_channels) )
self.assertIsNone(outputs.hidden_states )
self.assertIsNone(outputs.attentions )
# Test output_hidden_states=True
snake_case : Optional[int] = backbone(**snake_case__ , output_hidden_states=snake_case__ )
self.assertIsNotNone(outputs.hidden_states )
self.assertTrue(len(outputs.hidden_states ) , len(backbone.stage_names ) )
# We skip the stem layer
for hidden_states, n_channels in zip(outputs.hidden_states[1:] , backbone.channels ):
for hidden_state in hidden_states:
# Hidden states are in the format (batch_size, (height * width), n_channels)
snake_case , snake_case , snake_case : Dict = hidden_state.shape
self.assertTrue((h_batch_size, h_n_channels) , (batch_size, n_channels) )
# Test output_attentions=True
if self.has_attentions:
snake_case : Optional[Any] = backbone(**snake_case__ , output_attentions=snake_case__ )
self.assertIsNotNone(outputs.attentions )
| 59
| 1
|
import multiprocessing
from typing import TYPE_CHECKING, Optional, Union
from .. import Dataset, Features, config
from ..formatting import query_table
from ..packaged_modules.sql.sql import Sql
from ..utils import logging
from .abc import AbstractDatasetInputStream
if TYPE_CHECKING:
import sqlitea
import sqlalchemy
class UpperCAmelCase ( A_ ):
def __init__(self : Tuple , snake_case__ : Union[str, "sqlalchemy.sql.Selectable"] , snake_case__ : Union[str, "sqlalchemy.engine.Connection", "sqlalchemy.engine.Engine", "sqlite3.Connection"] , snake_case__ : Optional[Features] = None , snake_case__ : str = None , snake_case__ : bool = False , **snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
super().__init__(features=snake_case__ , cache_dir=snake_case__ , keep_in_memory=snake_case__ , **snake_case__ )
snake_case : Tuple = Sql(
cache_dir=snake_case__ , features=snake_case__ , sql=snake_case__ , con=snake_case__ , **snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = None
snake_case : int = None
snake_case : List[str] = None
snake_case : int = None
self.builder.download_and_prepare(
download_config=snake_case__ , download_mode=snake_case__ , verification_mode=snake_case__ , base_path=snake_case__ , )
# Build dataset for splits
snake_case : Optional[int] = self.builder.as_dataset(
split="train" , verification_mode=snake_case__ , in_memory=self.keep_in_memory )
return dataset
class UpperCAmelCase :
def __init__(self : Union[str, Any] , snake_case__ : Dataset , snake_case__ : str , snake_case__ : Union[str, "sqlalchemy.engine.Connection", "sqlalchemy.engine.Engine", "sqlite3.Connection"] , snake_case__ : Optional[int] = None , snake_case__ : Optional[int] = None , **snake_case__ : Dict , ) -> Optional[Any]:
'''simple docstring'''
if num_proc is not None and num_proc <= 0:
raise ValueError(f"""num_proc {num_proc} must be an integer > 0.""" )
snake_case : Any = dataset
snake_case : Dict = name
snake_case : Optional[Any] = con
snake_case : Any = batch_size if batch_size else config.DEFAULT_MAX_BATCH_SIZE
snake_case : Optional[int] = num_proc
snake_case : str = to_sql_kwargs
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> int:
'''simple docstring'''
snake_case : int = self.to_sql_kwargs.pop("sql" , snake_case__ )
snake_case : Tuple = self.to_sql_kwargs.pop("con" , snake_case__ )
snake_case : Tuple = self.to_sql_kwargs.pop("index" , snake_case__ )
snake_case : Union[str, Any] = self._write(index=snake_case__ , **self.to_sql_kwargs )
return written
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case , snake_case : Tuple = args
snake_case : Optional[int] = {**to_sql_kwargs, "if_exists": "append"} if offset > 0 else to_sql_kwargs
snake_case : List[Any] = query_table(
table=self.dataset.data , key=slice(snake_case__ , offset + self.batch_size ) , indices=self.dataset._indices , )
snake_case : List[Any] = batch.to_pandas()
snake_case : Optional[Any] = df.to_sql(self.name , self.con , index=snake_case__ , **snake_case__ )
return num_rows or len(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : str , **snake_case__ : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : List[str] = 0
if self.num_proc is None or self.num_proc == 1:
for offset in logging.tqdm(
range(0 , len(self.dataset ) , self.batch_size ) , unit="ba" , disable=not logging.is_progress_bar_enabled() , desc="Creating SQL from Arrow format" , ):
written += self._batch_sql((offset, index, to_sql_kwargs) )
else:
snake_case , snake_case : List[str] = len(self.dataset ), self.batch_size
with multiprocessing.Pool(self.num_proc ) as pool:
for num_rows in logging.tqdm(
pool.imap(
self._batch_sql , [(offset, index, to_sql_kwargs) for offset in range(0 , snake_case__ , snake_case__ )] , ) , total=(num_rows // batch_size) + 1 if num_rows % batch_size else num_rows // batch_size , unit="ba" , disable=not logging.is_progress_bar_enabled() , desc="Creating SQL from Arrow format" , ):
written += num_rows
return written
| 59
|
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : List[str] = []
snake_case : Optional[int] = []
snake_case : Any = []
for rt in rc.restypes:
snake_case : List[Any] = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
snake_case : str = {name: i for i, name in enumerate(__lowerCamelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
snake_case : Optional[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : List[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : int = torch.tensor(
__lowerCamelCase , dtype=torch.floataa , device=protein["aatype"].device , )
snake_case : int = protein["aatype"].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
snake_case : List[Any] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : str = restype_atomaa_mask[protein_aatype]
snake_case : str = residx_atomaa_mask
snake_case : Any = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
snake_case : List[str] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : List[Any] = residx_atomaa_to_atomaa.long()
# create the corresponding mask
snake_case : Union[str, Any] = torch.zeros([21, 37] , dtype=torch.floataa , device=protein["aatype"].device )
for restype, restype_letter in enumerate(rc.restypes ):
snake_case : Optional[int] = rc.restype_atoa[restype_letter]
snake_case : Any = rc.residue_atoms[restype_name]
for atom_name in atom_names:
snake_case : List[Any] = rc.atom_order[atom_name]
snake_case : Optional[Any] = 1
snake_case : List[Any] = restype_atomaa_mask[protein_aatype]
snake_case : int = residx_atomaa_mask
return protein
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : Dict = tree_map(lambda __lowerCamelCase : torch.tensor(__lowerCamelCase , device=batch["aatype"].device ) , __lowerCamelCase , np.ndarray )
snake_case : List[str] = tensor_tree_map(lambda __lowerCamelCase : np.array(__lowerCamelCase ) , make_atomaa_masks(__lowerCamelCase ) )
return out
| 59
| 1
|
import operator as op
__lowerCamelCase = """scaler.pt"""
__lowerCamelCase = """pytorch_model"""
__lowerCamelCase = """random_states"""
__lowerCamelCase = """optimizer"""
__lowerCamelCase = """scheduler"""
__lowerCamelCase = """pytorch_model.bin"""
__lowerCamelCase = """pytorch_model.bin.index.json"""
__lowerCamelCase = """model.safetensors"""
__lowerCamelCase = """model.safetensors.index.json"""
__lowerCamelCase = """1.10.2"""
__lowerCamelCase = """py38"""
__lowerCamelCase = """4.17.0"""
__lowerCamelCase = ["""ml.p3.16xlarge""", """ml.p3dn.24xlarge""", """ml.p4dn.24xlarge"""]
__lowerCamelCase = ["""FULL_SHARD""", """SHARD_GRAD_OP""", """NO_SHARD""", """HYBRID_SHARD""", """HYBRID_SHARD_ZERO2"""]
__lowerCamelCase = ["""TRANSFORMER_BASED_WRAP""", """SIZE_BASED_WRAP""", """NO_WRAP"""]
__lowerCamelCase = ["""BACKWARD_PRE""", """BACKWARD_POST""", """NO_PREFETCH"""]
__lowerCamelCase = ["""FULL_STATE_DICT""", """LOCAL_STATE_DICT""", """SHARDED_STATE_DICT"""]
__lowerCamelCase = """2.0.1"""
__lowerCamelCase = ["""pdsh""", """standard""", """openmpi""", """mvapich"""]
__lowerCamelCase = ["""default""", """reduce-overhead""", """max-autotune"""]
__lowerCamelCase = {""">""": op.gt, """>=""": op.ge, """==""": op.eq, """!=""": op.ne, """<=""": op.le, """<""": op.lt}
# These are the args for `torch.distributed.launch` for pytorch < 1.9
__lowerCamelCase = [
"""nnodes""",
"""nproc_per_node""",
"""rdzv_backend""",
"""rdzv_endpoint""",
"""rdzv_id""",
"""rdzv_conf""",
"""standalone""",
"""max_restarts""",
"""monitor_interval""",
"""start_method""",
"""role""",
"""module""",
"""m""",
"""no_python""",
"""run_path""",
"""log_dir""",
"""r""",
"""redirects""",
"""t""",
"""tee""",
"""node_rank""",
"""master_addr""",
"""master_port""",
]
__lowerCamelCase = ["""DEEPSPEED""", """MULTI_GPU""", """FSDP""", """MEGATRON_LM"""]
__lowerCamelCase = ["""DEEPSPEED""", """MULTI_XPU""", """FSDP"""]
| 59
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from .tokenization_lxmert import LxmertTokenizer
__lowerCamelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {
"""unc-nlp/lxmert-base-uncased""": """https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""unc-nlp/lxmert-base-uncased""": (
"""https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/tokenizer.json"""
),
},
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": 5_12,
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": {"""do_lower_case""": True},
}
class UpperCAmelCase ( A_ ):
A__ : Any = VOCAB_FILES_NAMES
A__ : Optional[int] = PRETRAINED_VOCAB_FILES_MAP
A__ : Tuple = PRETRAINED_INIT_CONFIGURATION
A__ : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : List[Any] = LxmertTokenizer
def __init__(self : Dict , snake_case__ : Tuple=None , snake_case__ : Optional[Any]=None , snake_case__ : Optional[Any]=True , snake_case__ : Tuple="[UNK]" , snake_case__ : Optional[Any]="[SEP]" , snake_case__ : Optional[Any]="[PAD]" , snake_case__ : List[Any]="[CLS]" , snake_case__ : Tuple="[MASK]" , snake_case__ : Dict=True , snake_case__ : Union[str, Any]=None , **snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
snake_case : int = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , snake_case__ ) != do_lower_case
or normalizer_state.get("strip_accents" , snake_case__ ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , snake_case__ ) != tokenize_chinese_chars
):
snake_case : Union[str, Any] = getattr(snake_case__ , normalizer_state.pop("type" ) )
snake_case : str = do_lower_case
snake_case : List[Any] = strip_accents
snake_case : Optional[int] = tokenize_chinese_chars
snake_case : int = normalizer_class(**snake_case__ )
snake_case : Optional[Any] = do_lower_case
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Optional[Any] , snake_case__ : Dict=None ) -> Any:
'''simple docstring'''
snake_case : Optional[int] = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Optional[Any] = [self.sep_token_id]
snake_case : 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 ) * [0] + len(token_ids_a + sep ) * [1]
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
snake_case : List[Any] = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 59
| 1
|
from collections import OrderedDict
from ...utils import logging
from .auto_factory import _BaseAutoModelClass, _LazyAutoMapping, auto_class_update
from .configuration_auto import CONFIG_MAPPING_NAMES
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = OrderedDict(
[
# Base model mapping
("""albert""", """FlaxAlbertModel"""),
("""bart""", """FlaxBartModel"""),
("""beit""", """FlaxBeitModel"""),
("""bert""", """FlaxBertModel"""),
("""big_bird""", """FlaxBigBirdModel"""),
("""blenderbot""", """FlaxBlenderbotModel"""),
("""blenderbot-small""", """FlaxBlenderbotSmallModel"""),
("""clip""", """FlaxCLIPModel"""),
("""distilbert""", """FlaxDistilBertModel"""),
("""electra""", """FlaxElectraModel"""),
("""gpt-sw3""", """FlaxGPT2Model"""),
("""gpt2""", """FlaxGPT2Model"""),
("""gpt_neo""", """FlaxGPTNeoModel"""),
("""gptj""", """FlaxGPTJModel"""),
("""longt5""", """FlaxLongT5Model"""),
("""marian""", """FlaxMarianModel"""),
("""mbart""", """FlaxMBartModel"""),
("""mt5""", """FlaxMT5Model"""),
("""opt""", """FlaxOPTModel"""),
("""pegasus""", """FlaxPegasusModel"""),
("""regnet""", """FlaxRegNetModel"""),
("""resnet""", """FlaxResNetModel"""),
("""roberta""", """FlaxRobertaModel"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormModel"""),
("""roformer""", """FlaxRoFormerModel"""),
("""t5""", """FlaxT5Model"""),
("""vision-text-dual-encoder""", """FlaxVisionTextDualEncoderModel"""),
("""vit""", """FlaxViTModel"""),
("""wav2vec2""", """FlaxWav2Vec2Model"""),
("""whisper""", """FlaxWhisperModel"""),
("""xglm""", """FlaxXGLMModel"""),
("""xlm-roberta""", """FlaxXLMRobertaModel"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for pre-training mapping
("""albert""", """FlaxAlbertForPreTraining"""),
("""bart""", """FlaxBartForConditionalGeneration"""),
("""bert""", """FlaxBertForPreTraining"""),
("""big_bird""", """FlaxBigBirdForPreTraining"""),
("""electra""", """FlaxElectraForPreTraining"""),
("""longt5""", """FlaxLongT5ForConditionalGeneration"""),
("""mbart""", """FlaxMBartForConditionalGeneration"""),
("""mt5""", """FlaxMT5ForConditionalGeneration"""),
("""roberta""", """FlaxRobertaForMaskedLM"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForMaskedLM"""),
("""roformer""", """FlaxRoFormerForMaskedLM"""),
("""t5""", """FlaxT5ForConditionalGeneration"""),
("""wav2vec2""", """FlaxWav2Vec2ForPreTraining"""),
("""whisper""", """FlaxWhisperForConditionalGeneration"""),
("""xlm-roberta""", """FlaxXLMRobertaForMaskedLM"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Masked LM mapping
("""albert""", """FlaxAlbertForMaskedLM"""),
("""bart""", """FlaxBartForConditionalGeneration"""),
("""bert""", """FlaxBertForMaskedLM"""),
("""big_bird""", """FlaxBigBirdForMaskedLM"""),
("""distilbert""", """FlaxDistilBertForMaskedLM"""),
("""electra""", """FlaxElectraForMaskedLM"""),
("""mbart""", """FlaxMBartForConditionalGeneration"""),
("""roberta""", """FlaxRobertaForMaskedLM"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForMaskedLM"""),
("""roformer""", """FlaxRoFormerForMaskedLM"""),
("""xlm-roberta""", """FlaxXLMRobertaForMaskedLM"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Seq2Seq Causal LM mapping
("""bart""", """FlaxBartForConditionalGeneration"""),
("""blenderbot""", """FlaxBlenderbotForConditionalGeneration"""),
("""blenderbot-small""", """FlaxBlenderbotSmallForConditionalGeneration"""),
("""encoder-decoder""", """FlaxEncoderDecoderModel"""),
("""longt5""", """FlaxLongT5ForConditionalGeneration"""),
("""marian""", """FlaxMarianMTModel"""),
("""mbart""", """FlaxMBartForConditionalGeneration"""),
("""mt5""", """FlaxMT5ForConditionalGeneration"""),
("""pegasus""", """FlaxPegasusForConditionalGeneration"""),
("""t5""", """FlaxT5ForConditionalGeneration"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Image-classsification
("""beit""", """FlaxBeitForImageClassification"""),
("""regnet""", """FlaxRegNetForImageClassification"""),
("""resnet""", """FlaxResNetForImageClassification"""),
("""vit""", """FlaxViTForImageClassification"""),
]
)
__lowerCamelCase = OrderedDict(
[
("""vision-encoder-decoder""", """FlaxVisionEncoderDecoderModel"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Causal LM mapping
("""bart""", """FlaxBartForCausalLM"""),
("""bert""", """FlaxBertForCausalLM"""),
("""big_bird""", """FlaxBigBirdForCausalLM"""),
("""electra""", """FlaxElectraForCausalLM"""),
("""gpt-sw3""", """FlaxGPT2LMHeadModel"""),
("""gpt2""", """FlaxGPT2LMHeadModel"""),
("""gpt_neo""", """FlaxGPTNeoForCausalLM"""),
("""gptj""", """FlaxGPTJForCausalLM"""),
("""opt""", """FlaxOPTForCausalLM"""),
("""roberta""", """FlaxRobertaForCausalLM"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForCausalLM"""),
("""xglm""", """FlaxXGLMForCausalLM"""),
("""xlm-roberta""", """FlaxXLMRobertaForCausalLM"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Sequence Classification mapping
("""albert""", """FlaxAlbertForSequenceClassification"""),
("""bart""", """FlaxBartForSequenceClassification"""),
("""bert""", """FlaxBertForSequenceClassification"""),
("""big_bird""", """FlaxBigBirdForSequenceClassification"""),
("""distilbert""", """FlaxDistilBertForSequenceClassification"""),
("""electra""", """FlaxElectraForSequenceClassification"""),
("""mbart""", """FlaxMBartForSequenceClassification"""),
("""roberta""", """FlaxRobertaForSequenceClassification"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForSequenceClassification"""),
("""roformer""", """FlaxRoFormerForSequenceClassification"""),
("""xlm-roberta""", """FlaxXLMRobertaForSequenceClassification"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Question Answering mapping
("""albert""", """FlaxAlbertForQuestionAnswering"""),
("""bart""", """FlaxBartForQuestionAnswering"""),
("""bert""", """FlaxBertForQuestionAnswering"""),
("""big_bird""", """FlaxBigBirdForQuestionAnswering"""),
("""distilbert""", """FlaxDistilBertForQuestionAnswering"""),
("""electra""", """FlaxElectraForQuestionAnswering"""),
("""mbart""", """FlaxMBartForQuestionAnswering"""),
("""roberta""", """FlaxRobertaForQuestionAnswering"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForQuestionAnswering"""),
("""roformer""", """FlaxRoFormerForQuestionAnswering"""),
("""xlm-roberta""", """FlaxXLMRobertaForQuestionAnswering"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Token Classification mapping
("""albert""", """FlaxAlbertForTokenClassification"""),
("""bert""", """FlaxBertForTokenClassification"""),
("""big_bird""", """FlaxBigBirdForTokenClassification"""),
("""distilbert""", """FlaxDistilBertForTokenClassification"""),
("""electra""", """FlaxElectraForTokenClassification"""),
("""roberta""", """FlaxRobertaForTokenClassification"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForTokenClassification"""),
("""roformer""", """FlaxRoFormerForTokenClassification"""),
("""xlm-roberta""", """FlaxXLMRobertaForTokenClassification"""),
]
)
__lowerCamelCase = OrderedDict(
[
# Model for Multiple Choice mapping
("""albert""", """FlaxAlbertForMultipleChoice"""),
("""bert""", """FlaxBertForMultipleChoice"""),
("""big_bird""", """FlaxBigBirdForMultipleChoice"""),
("""distilbert""", """FlaxDistilBertForMultipleChoice"""),
("""electra""", """FlaxElectraForMultipleChoice"""),
("""roberta""", """FlaxRobertaForMultipleChoice"""),
("""roberta-prelayernorm""", """FlaxRobertaPreLayerNormForMultipleChoice"""),
("""roformer""", """FlaxRoFormerForMultipleChoice"""),
("""xlm-roberta""", """FlaxXLMRobertaForMultipleChoice"""),
]
)
__lowerCamelCase = OrderedDict(
[
("""bert""", """FlaxBertForNextSentencePrediction"""),
]
)
__lowerCamelCase = OrderedDict(
[
("""speech-encoder-decoder""", """FlaxSpeechEncoderDecoderModel"""),
("""whisper""", """FlaxWhisperForConditionalGeneration"""),
]
)
__lowerCamelCase = OrderedDict(
[
("""whisper""", """FlaxWhisperForAudioClassification"""),
]
)
__lowerCamelCase = _LazyAutoMapping(CONFIG_MAPPING_NAMES, FLAX_MODEL_MAPPING_NAMES)
__lowerCamelCase = _LazyAutoMapping(CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_PRETRAINING_MAPPING_NAMES)
__lowerCamelCase = _LazyAutoMapping(CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_MASKED_LM_MAPPING_NAMES)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_VISION_2_SEQ_MAPPING_NAMES)
__lowerCamelCase = _LazyAutoMapping(CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_CAUSAL_LM_MAPPING_NAMES)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES
)
__lowerCamelCase = _LazyAutoMapping(
CONFIG_MAPPING_NAMES, FLAX_MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES
)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Any = FLAX_MODEL_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModel)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Any = FLAX_MODEL_FOR_PRETRAINING_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModelForPreTraining, head_doc="""pretraining""")
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Optional[Any] = FLAX_MODEL_FOR_CAUSAL_LM_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModelForCausalLM, head_doc="""causal language modeling""")
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : int = FLAX_MODEL_FOR_MASKED_LM_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModelForMaskedLM, head_doc="""masked language modeling""")
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : str = FLAX_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING
__lowerCamelCase = auto_class_update(
FlaxAutoModelForSeqaSeqLM, head_doc="""sequence-to-sequence language modeling""", checkpoint_for_example="""t5-base"""
)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Dict = FLAX_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
__lowerCamelCase = auto_class_update(
FlaxAutoModelForSequenceClassification, head_doc="""sequence classification"""
)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Tuple = FLAX_MODEL_FOR_QUESTION_ANSWERING_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModelForQuestionAnswering, head_doc="""question answering""")
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : str = FLAX_MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING
__lowerCamelCase = auto_class_update(
FlaxAutoModelForTokenClassification, head_doc="""token classification"""
)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : int = FLAX_MODEL_FOR_MULTIPLE_CHOICE_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModelForMultipleChoice, head_doc="""multiple choice""")
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Optional[Any] = FLAX_MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING
__lowerCamelCase = auto_class_update(
FlaxAutoModelForNextSentencePrediction, head_doc="""next sentence prediction"""
)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Tuple = FLAX_MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING
__lowerCamelCase = auto_class_update(
FlaxAutoModelForImageClassification, head_doc="""image classification"""
)
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : int = FLAX_MODEL_FOR_VISION_2_SEQ_MAPPING
__lowerCamelCase = auto_class_update(FlaxAutoModelForVisionaSeq, head_doc="""vision-to-text modeling""")
class UpperCAmelCase ( _BaseAutoModelClass ):
A__ : Tuple = FLAX_MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING
__lowerCamelCase = auto_class_update(
FlaxAutoModelForSpeechSeqaSeq, head_doc="""sequence-to-sequence speech-to-text modeling"""
)
| 59
|
import torch
from diffusers import DDIMParallelScheduler
from .test_schedulers import SchedulerCommonTest
class UpperCAmelCase ( A_ ):
A__ : Dict = (DDIMParallelScheduler,)
A__ : Tuple = (("eta", 0.0), ("num_inference_steps", 50))
def _SCREAMING_SNAKE_CASE (self : Tuple , **snake_case__ : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case : Any = {
"num_train_timesteps": 10_00,
"beta_start": 0.0001,
"beta_end": 0.02,
"beta_schedule": "linear",
"clip_sample": True,
}
config.update(**snake_case__ )
return config
def _SCREAMING_SNAKE_CASE (self : Dict , **snake_case__ : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = self.scheduler_classes[0]
snake_case : Any = self.get_scheduler_config(**snake_case__ )
snake_case : Any = scheduler_class(**snake_case__ )
snake_case , snake_case : Union[str, Any] = 10, 0.0
snake_case : List[Any] = self.dummy_model()
snake_case : Any = self.dummy_sample_deter
scheduler.set_timesteps(snake_case__ )
for t in scheduler.timesteps:
snake_case : Optional[int] = model(snake_case__ , snake_case__ )
snake_case : List[str] = scheduler.step(snake_case__ , snake_case__ , snake_case__ , snake_case__ ).prev_sample
return sample
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> str:
'''simple docstring'''
for timesteps in [1_00, 5_00, 10_00]:
self.check_over_configs(num_train_timesteps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=snake_case__ )
snake_case : Optional[int] = self.scheduler_classes[0]
snake_case : Optional[int] = self.get_scheduler_config(steps_offset=1 )
snake_case : Union[str, Any] = scheduler_class(**snake_case__ )
scheduler.set_timesteps(5 )
assert torch.equal(scheduler.timesteps , torch.LongTensor([8_01, 6_01, 4_01, 2_01, 1] ) )
def _SCREAMING_SNAKE_CASE (self : int ) -> Tuple:
'''simple docstring'''
for beta_start, beta_end in zip([0.0001, 0.001, 0.01, 0.1] , [0.002, 0.02, 0.2, 2] ):
self.check_over_configs(beta_start=snake_case__ , beta_end=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
for timestep_spacing in ["trailing", "leading"]:
self.check_over_configs(timestep_spacing=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[Any]:
'''simple docstring'''
for rescale_betas_zero_snr in [True, False]:
self.check_over_configs(rescale_betas_zero_snr=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
self.check_over_configs(thresholding=snake_case__ )
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(
thresholding=snake_case__ , prediction_type=snake_case__ , sample_max_value=snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
for t in [1, 10, 49]:
self.check_over_forward(time_step=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Any:
'''simple docstring'''
for t, num_inference_steps in zip([1, 10, 50] , [10, 50, 5_00] ):
self.check_over_forward(time_step=snake_case__ , num_inference_steps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Optional[Any]:
'''simple docstring'''
for t, eta in zip([1, 10, 49] , [0.0, 0.5, 1.0] ):
self.check_over_forward(time_step=snake_case__ , eta=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Dict = self.scheduler_classes[0]
snake_case : Tuple = self.get_scheduler_config()
snake_case : Dict = scheduler_class(**snake_case__ )
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_20 , 4_00 ) - 0.14771 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_80 , 9_60 ) - 0.32460 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_87 , 4_86 ) - 0.00979 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_99 , 9_98 ) - 0.02 ) ) < 1e-5
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
snake_case : Union[str, Any] = self.scheduler_classes[0]
snake_case : List[Any] = self.get_scheduler_config()
snake_case : int = scheduler_class(**snake_case__ )
snake_case , snake_case : Any = 10, 0.0
scheduler.set_timesteps(snake_case__ )
snake_case : Optional[Any] = self.dummy_model()
snake_case : str = self.dummy_sample_deter
snake_case : Dict = self.dummy_sample_deter + 0.1
snake_case : Dict = self.dummy_sample_deter - 0.1
snake_case : Optional[Any] = samplea.shape[0]
snake_case : str = torch.stack([samplea, samplea, samplea] , dim=0 )
snake_case : Tuple = torch.arange(snake_case__ )[0:3, None].repeat(1 , snake_case__ )
snake_case : Tuple = model(samples.flatten(0 , 1 ) , timesteps.flatten(0 , 1 ) )
snake_case : List[str] = scheduler.batch_step_no_noise(snake_case__ , timesteps.flatten(0 , 1 ) , samples.flatten(0 , 1 ) , snake_case__ )
snake_case : Dict = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 1147.7904 ) < 1e-2
assert abs(result_mean.item() - 0.4982 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[Any] = self.full_loop()
snake_case : Optional[Any] = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 172.0067 ) < 1e-2
assert abs(result_mean.item() - 0.223967 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : str ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(prediction_type="v_prediction" )
snake_case : int = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[int] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 52.5302 ) < 1e-2
assert abs(result_mean.item() - 0.0684 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : str = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.8295 ) < 1e-2
assert abs(result_mean.item() - 0.1951 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : Tuple = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.0784 ) < 1e-2
assert abs(result_mean.item() - 0.1941 ) < 1e-3
| 59
| 1
|
def UpperCamelCase ( __lowerCamelCase : int ):
if p < 2:
raise ValueError("p should not be less than 2!" )
elif p == 2:
return True
snake_case : Dict = 4
snake_case : str = (1 << p) - 1
for _ in range(p - 2 ):
snake_case : Dict = ((s * s) - 2) % m
return s == 0
if __name__ == "__main__":
print(lucas_lehmer_test(7))
print(lucas_lehmer_test(11))
| 59
|
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )]
snake_case : int = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1 or len(__lowerCamelCase ) <= key:
return input_string
for position, character in enumerate(__lowerCamelCase ):
snake_case : Any = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append(__lowerCamelCase )
snake_case : List[str] = ["".join(__lowerCamelCase ) for row in temp_grid]
snake_case : Tuple = "".join(__lowerCamelCase )
return output_string
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : Dict = []
snake_case : Union[str, Any] = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1:
return input_string
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )] # generates template
for position in range(len(__lowerCamelCase ) ):
snake_case : List[str] = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append("*" )
snake_case : Tuple = 0
for row in temp_grid: # fills in the characters
snake_case : Union[str, Any] = input_string[counter : counter + len(__lowerCamelCase )]
grid.append(list(__lowerCamelCase ) )
counter += len(__lowerCamelCase )
snake_case : str = "" # reads as zigzag
for position in range(len(__lowerCamelCase ) ):
snake_case : Optional[int] = position % (lowest * 2) # puts it in bounds
snake_case : Tuple = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
output_string += grid[num][0]
grid[num].pop(0 )
return output_string
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : Tuple = {}
for key_guess in range(1 , len(__lowerCamelCase ) ): # tries every key
snake_case : Any = decrypt(__lowerCamelCase , __lowerCamelCase )
return results
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 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 UpperCAmelCase ( unittest.TestCase ):
@property
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Union[str, Any]:
'''simple docstring'''
torch.manual_seed(0 )
snake_case : Dict = 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 : Union[str, Any] ) -> int:
'''simple docstring'''
torch.manual_seed(0 )
snake_case : 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 : Tuple ) -> List[Any]:
'''simple docstring'''
torch.manual_seed(0 )
snake_case : int = 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=10_00 , )
return CLIPTextModel(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Dict:
'''simple docstring'''
snake_case : List[Any] = self.dummy_uncond_unet
snake_case : List[Any] = DDIMScheduler()
snake_case : Optional[Any] = self.dummy_vq_model
snake_case : str = LDMPipeline(unet=snake_case__ , vqvae=snake_case__ , scheduler=snake_case__ )
ldm.to(snake_case__ )
ldm.set_progress_bar_config(disable=snake_case__ )
snake_case : Tuple = torch.manual_seed(0 )
snake_case : Any = ldm(generator=snake_case__ , num_inference_steps=2 , output_type="numpy" ).images
snake_case : List[Any] = torch.manual_seed(0 )
snake_case : Tuple = ldm(generator=snake_case__ , num_inference_steps=2 , output_type="numpy" , return_dict=snake_case__ )[0]
snake_case : Tuple = image[0, -3:, -3:, -1]
snake_case : Optional[int] = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 64, 64, 3)
snake_case : int = np.array([0.8512, 0.818, 0.6411, 0.6808, 0.4465, 0.5618, 0.46, 0.6231, 0.5172] )
snake_case : Any = 1e-2 if torch_device != "mps" else 3e-2
assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < tolerance
@slow
@require_torch
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Any ) -> str:
'''simple docstring'''
snake_case : int = LDMPipeline.from_pretrained("CompVis/ldm-celebahq-256" )
ldm.to(snake_case__ )
ldm.set_progress_bar_config(disable=snake_case__ )
snake_case : List[str] = torch.manual_seed(0 )
snake_case : Any = ldm(generator=snake_case__ , num_inference_steps=5 , output_type="numpy" ).images
snake_case : Any = image[0, -3:, -3:, -1]
assert image.shape == (1, 2_56, 2_56, 3)
snake_case : List[str] = np.array([0.4399, 0.44975, 0.46825, 0.474, 0.4359, 0.4581, 0.45095, 0.4341, 0.4447] )
snake_case : Optional[int] = 1e-2 if torch_device != "mps" else 3e-2
assert np.abs(image_slice.flatten() - expected_slice ).max() < tolerance
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
if is_sentencepiece_available():
from ..ta.tokenization_ta import TaTokenizer
else:
from ...utils.dummy_sentencepiece_objects import TaTokenizer
__lowerCamelCase = TaTokenizer
if is_tokenizers_available():
from ..ta.tokenization_ta_fast import TaTokenizerFast
else:
from ...utils.dummy_tokenizers_objects import TaTokenizerFast
__lowerCamelCase = TaTokenizerFast
__lowerCamelCase = {"""configuration_mt5""": ["""MT5Config""", """MT5OnnxConfig"""]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""MT5EncoderModel""",
"""MT5ForConditionalGeneration""",
"""MT5ForQuestionAnswering""",
"""MT5Model""",
"""MT5PreTrainedModel""",
"""MT5Stack""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""TFMT5EncoderModel""", """TFMT5ForConditionalGeneration""", """TFMT5Model"""]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""FlaxMT5EncoderModel""", """FlaxMT5ForConditionalGeneration""", """FlaxMT5Model"""]
if TYPE_CHECKING:
from .configuration_mta import MTaConfig, MTaOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mta import (
MTaEncoderModel,
MTaForConditionalGeneration,
MTaForQuestionAnswering,
MTaModel,
MTaPreTrainedModel,
MTaStack,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel
else:
import sys
__lowerCamelCase = _LazyModule(
__name__,
globals()["""__file__"""],
_import_structure,
extra_objects={"""MT5Tokenizer""": MTaTokenizer, """MT5TokenizerFast""": MTaTokenizerFast},
module_spec=__spec__,
)
| 59
| 1
|
import json
import os
from typing import Dict, List, Optional, Tuple
import regex as re
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""vocab_file""": """vocab.json""",
"""merges_file""": """merges.txt""",
"""tokenizer_config_file""": """tokenizer_config.json""",
}
__lowerCamelCase = {
"""vocab_file""": {
"""facebook/blenderbot_small-90M""": """https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/vocab.json"""
},
"""merges_file""": {
"""facebook/blenderbot_small-90M""": """https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/merges.txt"""
},
"""tokenizer_config_file""": {
"""facebook/blenderbot_small-90M""": (
"""https://huggingface.co/facebook/blenderbot_small-90M/resolve/main/tokenizer_config.json"""
)
},
}
__lowerCamelCase = {"""facebook/blenderbot_small-90M""": 5_12}
def UpperCamelCase ( __lowerCamelCase : Any ):
snake_case : Optional[Any] = set()
snake_case : Tuple = word[0]
for char in word[1:]:
pairs.add((prev_char, char) )
snake_case : Tuple = char
snake_case : int = set(__lowerCamelCase )
return pairs
class UpperCAmelCase ( A_ ):
A__ : str = VOCAB_FILES_NAMES
A__ : Optional[int] = PRETRAINED_VOCAB_FILES_MAP
A__ : Any = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : Optional[int] = ["input_ids", "attention_mask"]
def __init__(self : Union[str, Any] , snake_case__ : int , snake_case__ : Optional[Any] , snake_case__ : Dict="__start__" , snake_case__ : Any="__end__" , snake_case__ : Tuple="__unk__" , snake_case__ : Dict="__null__" , **snake_case__ : List[Any] , ) -> List[str]:
'''simple docstring'''
super().__init__(unk_token=snake_case__ , bos_token=snake_case__ , eos_token=snake_case__ , pad_token=snake_case__ , **snake_case__ )
with open(snake_case__ , encoding="utf-8" ) as vocab_handle:
snake_case : Optional[int] = json.load(snake_case__ )
snake_case : Any = {v: k for k, v in self.encoder.items()}
with open(snake_case__ , encoding="utf-8" ) as merges_handle:
snake_case : List[Any] = merges_handle.read().split("\n" )[1:-1]
snake_case : int = [tuple(merge.split() ) for merge in merges]
snake_case : Any = dict(zip(snake_case__ , range(len(snake_case__ ) ) ) )
snake_case : Optional[Any] = {}
@property
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
return len(self.encoder )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Dict:
'''simple docstring'''
return dict(self.encoder , **self.added_tokens_encoder )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : str ) -> str:
'''simple docstring'''
if token in self.cache:
return self.cache[token]
snake_case : Union[str, Any] = re.sub("([.,!?()])" , R" \1" , snake_case__ )
snake_case : int = re.sub("(')" , R" \1 " , snake_case__ )
snake_case : str = re.sub(R"\s{2,}" , " " , snake_case__ )
if "\n" in token:
snake_case : List[Any] = token.replace("\n" , " __newln__" )
snake_case : Optional[Any] = token.split(" " )
snake_case : str = []
for token in tokens:
if not len(snake_case__ ):
continue
snake_case : Any = token.lower()
snake_case : Any = tuple(snake_case__ )
snake_case : str = tuple(list(word[:-1] ) + [word[-1] + "</w>"] )
snake_case : Optional[int] = get_pairs(snake_case__ )
if not pairs:
words.append(snake_case__ )
continue
while True:
snake_case : Optional[int] = min(snake_case__ , key=lambda snake_case__ : self.bpe_ranks.get(snake_case__ , float("inf" ) ) )
if bigram not in self.bpe_ranks:
break
snake_case , snake_case : Tuple = bigram
snake_case : Tuple = []
snake_case : str = 0
while i < len(snake_case__ ):
try:
snake_case : str = word.index(snake_case__ , snake_case__ )
new_word.extend(word[i:j] )
snake_case : Dict = j
except ValueError:
new_word.extend(word[i:] )
break
if word[i] == first and i < len(snake_case__ ) - 1 and word[i + 1] == second:
new_word.append(first + second )
i += 2
else:
new_word.append(word[i] )
i += 1
snake_case : Optional[Any] = tuple(snake_case__ )
snake_case : List[Any] = new_word
if len(snake_case__ ) == 1:
break
else:
snake_case : Any = get_pairs(snake_case__ )
snake_case : List[str] = "@@ ".join(snake_case__ )
snake_case : Union[str, Any] = word[:-4]
snake_case : int = word
words.append(snake_case__ )
return " ".join(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Tuple = []
snake_case : Tuple = re.findall(R"\S+\n?" , snake_case__ )
for token in words:
split_tokens.extend(list(self.bpe(snake_case__ ).split(" " ) ) )
return split_tokens
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> int:
'''simple docstring'''
snake_case : Tuple = token.lower()
return self.encoder.get(snake_case__ , self.encoder.get(self.unk_token ) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : int ) -> str:
'''simple docstring'''
return self.decoder.get(snake_case__ , self.unk_token )
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : List[str] ) -> str:
'''simple docstring'''
snake_case : str = " ".join(snake_case__ ).replace("@@ " , "" ).strip()
return out_string
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
if not os.path.isdir(snake_case__ ):
logger.error(f"""Vocabulary path ({save_directory}) should be a directory""" )
return
snake_case : int = os.path.join(
snake_case__ , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
snake_case : List[Any] = os.path.join(
snake_case__ , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["merges_file"] )
with open(snake_case__ , "w" , encoding="utf-8" ) as f:
f.write(json.dumps(self.encoder , indent=2 , sort_keys=snake_case__ , ensure_ascii=snake_case__ ) + "\n" )
snake_case : Tuple = 0
with open(snake_case__ , "w" , encoding="utf-8" ) as writer:
writer.write("#version: 0.2\n" )
for bpe_tokens, token_index in sorted(self.bpe_ranks.items() , key=lambda snake_case__ : kv[1] ):
if index != token_index:
logger.warning(
f"""Saving vocabulary to {merge_file}: BPE merge indices are not consecutive."""
" Please check that the tokenizer is not corrupted!" )
snake_case : List[Any] = token_index
writer.write(" ".join(snake_case__ ) + "\n" )
index += 1
return vocab_file, merge_file
| 59
|
import os
import shutil
from pathlib import Path
from typing import Optional, Union
import numpy as np
from huggingface_hub import hf_hub_download
from ..utils import ONNX_EXTERNAL_WEIGHTS_NAME, ONNX_WEIGHTS_NAME, is_onnx_available, logging
if is_onnx_available():
import onnxruntime as ort
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""tensor(bool)""": np.bool_,
"""tensor(int8)""": np.inta,
"""tensor(uint8)""": np.uinta,
"""tensor(int16)""": np.intaa,
"""tensor(uint16)""": np.uintaa,
"""tensor(int32)""": np.intaa,
"""tensor(uint32)""": np.uintaa,
"""tensor(int64)""": np.intaa,
"""tensor(uint64)""": np.uintaa,
"""tensor(float16)""": np.floataa,
"""tensor(float)""": np.floataa,
"""tensor(double)""": np.floataa,
}
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=None , **snake_case__ : Optional[Any] ) -> List[str]:
'''simple docstring'''
logger.info("`diffusers.OnnxRuntimeModel` is experimental and might change in the future." )
snake_case : Optional[Any] = model
snake_case : Dict = kwargs.get("model_save_dir" , snake_case__ )
snake_case : int = kwargs.get("latest_model_name" , snake_case__ )
def __call__(self : Tuple , **snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Union[str, Any] = {k: np.array(snake_case__ ) for k, v in kwargs.items()}
return self.model.run(snake_case__ , snake_case__ )
@staticmethod
def _SCREAMING_SNAKE_CASE (snake_case__ : Union[str, Path] , snake_case__ : Optional[int]=None , snake_case__ : Optional[int]=None ) -> Any:
'''simple docstring'''
if provider is None:
logger.info("No onnxruntime provider specified, using CPUExecutionProvider" )
snake_case : Optional[int] = "CPUExecutionProvider"
return ort.InferenceSession(snake_case__ , providers=[provider] , sess_options=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Union[str, Path] , snake_case__ : Optional[str] = None , **snake_case__ : Any ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = file_name if file_name is not None else ONNX_WEIGHTS_NAME
snake_case : Any = self.model_save_dir.joinpath(self.latest_model_name )
snake_case : str = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
snake_case : List[str] = self.model_save_dir.joinpath(snake_case__ )
if src_path.exists():
snake_case : Tuple = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, os.PathLike] , **snake_case__ : Optional[int] , ) -> str:
'''simple docstring'''
if os.path.isfile(snake_case__ ):
logger.error(f"""Provided path ({save_directory}) should be a directory, not a file""" )
return
os.makedirs(snake_case__ , exist_ok=snake_case__ )
# saving model weights/files
self._save_pretrained(snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Tuple , snake_case__ : Union[str, Path] , snake_case__ : Optional[Union[bool, str, None]] = None , snake_case__ : Optional[Union[str, None]] = None , snake_case__ : bool = False , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional["ort.SessionOptions"] = None , **snake_case__ : Tuple , ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(snake_case__ ):
snake_case : Any = OnnxRuntimeModel.load_model(
os.path.join(snake_case__ , snake_case__ ) , provider=snake_case__ , sess_options=snake_case__ )
snake_case : Union[str, Any] = Path(snake_case__ )
# load model from hub
else:
# download model
snake_case : Dict = hf_hub_download(
repo_id=snake_case__ , filename=snake_case__ , use_auth_token=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , )
snake_case : List[Any] = Path(snake_case__ ).parent
snake_case : Union[str, Any] = Path(snake_case__ ).name
snake_case : Dict = OnnxRuntimeModel.load_model(snake_case__ , provider=snake_case__ , sess_options=snake_case__ )
return cls(model=snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : Union[str, Path] , snake_case__ : bool = True , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , **snake_case__ : Dict , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = None
if len(str(snake_case__ ).split("@" ) ) == 2:
snake_case , snake_case : int = model_id.split("@" )
return cls._from_pretrained(
model_id=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , use_auth_token=snake_case__ , **snake_case__ , )
| 59
| 1
|
import os
from datetime import datetime as dt
from github import Github
__lowerCamelCase = [
"""good first issue""",
"""good second issue""",
"""good difficult issue""",
"""enhancement""",
"""new pipeline/model""",
"""new scheduler""",
"""wip""",
]
def UpperCamelCase ( ):
snake_case : int = Github(os.environ["GITHUB_TOKEN"] )
snake_case : Union[str, Any] = g.get_repo("huggingface/diffusers" )
snake_case : Optional[Any] = repo.get_issues(state="open" )
for issue in open_issues:
snake_case : Optional[int] = sorted(issue.get_comments() , key=lambda __lowerCamelCase : i.created_at , reverse=__lowerCamelCase )
snake_case : List[Any] = comments[0] if len(__lowerCamelCase ) > 0 else None
if (
last_comment is not None
and last_comment.user.login == "github-actions[bot]"
and (dt.utcnow() - issue.updated_at).days > 7
and (dt.utcnow() - issue.created_at).days >= 30
and not any(label.name.lower() in LABELS_TO_EXEMPT for label in issue.get_labels() )
):
# Closes the issue after 7 days of inactivity since the Stalebot notification.
issue.edit(state="closed" )
elif (
"stale" in issue.get_labels()
and last_comment is not None
and last_comment.user.login != "github-actions[bot]"
):
# Opens the issue if someone other than Stalebot commented.
issue.edit(state="open" )
issue.remove_from_labels("stale" )
elif (
(dt.utcnow() - issue.updated_at).days > 23
and (dt.utcnow() - issue.created_at).days >= 30
and not any(label.name.lower() in LABELS_TO_EXEMPT for label in issue.get_labels() )
):
# Post a Stalebot notification after 23 days of inactivity.
issue.create_comment(
"This issue has been automatically marked as stale because it has not had "
"recent activity. If you think this still needs to be addressed "
"please comment on this thread.\n\nPlease note that issues that do not follow the "
"[contributing guidelines](https://github.com/huggingface/diffusers/blob/main/CONTRIBUTING.md) "
"are likely to be ignored." )
issue.add_to_labels("stale" )
if __name__ == "__main__":
main()
| 59
|
import argparse
import json
from dataclasses import dataclass, field
from functools import partial
from pathlib import Path
from typing import Callable, Dict, List, Tuple
import timm
import torch
import torch.nn as nn
from classy_vision.models.regnet import RegNet, RegNetParams, RegNetYaagf, RegNetYaagf, RegNetYaaagf
from huggingface_hub import cached_download, hf_hub_url
from torch import Tensor
from vissl.models.model_helpers import get_trunk_forward_outputs
from transformers import AutoImageProcessor, RegNetConfig, RegNetForImageClassification, RegNetModel
from transformers.utils import logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger()
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : List[nn.Module] = field(default_factory=A_ )
A__ : list = field(default_factory=A_ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : Tensor , snake_case__ : Tensor ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = len(list(m.modules() ) ) == 1 or isinstance(snake_case__ , nn.Convad ) or isinstance(snake_case__ , nn.BatchNormad )
if has_not_submodules:
self.traced.append(snake_case__ )
def __call__(self : List[Any] , snake_case__ : Tensor ) -> List[Any]:
'''simple docstring'''
for m in self.module.modules():
self.handles.append(m.register_forward_hook(self._forward_hook ) )
self.module(snake_case__ )
[x.remove() for x in self.handles]
return self
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[int]:
'''simple docstring'''
return list(filter(lambda snake_case__ : len(list(x.state_dict().keys() ) ) > 0 , self.traced ) )
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : nn.Module
A__ : int = 1
A__ : List = field(default_factory=A_ )
A__ : List = field(default_factory=A_ )
A__ : bool = True
def __call__(self : List[Any] , snake_case__ : Tensor ) -> Any:
'''simple docstring'''
snake_case : str = Tracker(self.dest )(snake_case__ ).parametrized
snake_case : Optional[int] = Tracker(self.src )(snake_case__ ).parametrized
snake_case : List[str] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.src_skip , snake_case__ ) )
snake_case : Optional[Any] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.dest_skip , snake_case__ ) )
if len(snake_case__ ) != len(snake_case__ ) and self.raise_if_mismatch:
raise Exception(
f"""Numbers of operations are different. Source module has {len(snake_case__ )} operations while"""
f""" destination module has {len(snake_case__ )}.""" )
for dest_m, src_m in zip(snake_case__ , snake_case__ ):
dest_m.load_state_dict(src_m.state_dict() )
if self.verbose == 1:
print(f"""Transfered from={src_m} to={dest_m}""" )
class UpperCAmelCase ( nn.Module ):
def __init__(self : Tuple , snake_case__ : nn.Module ) -> Optional[Any]:
'''simple docstring'''
super().__init__()
snake_case : List[Tuple[str, nn.Module]] = []
# - get the stem
feature_blocks.append(("conv1", model.stem) )
# - get all the feature blocks
for k, v in model.trunk_output.named_children():
assert k.startswith("block" ), f"""Unexpected layer name {k}"""
snake_case : Union[str, Any] = len(snake_case__ ) + 1
feature_blocks.append((f"""res{block_index}""", v) )
snake_case : Optional[Any] = nn.ModuleDict(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Tensor ) -> Dict:
'''simple docstring'''
return get_trunk_forward_outputs(
snake_case__ , out_feat_keys=snake_case__ , feature_blocks=self._feature_blocks , )
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> str:
'''simple docstring'''
snake_case : List[Any] = x.split("-" )
return x_split[0] + x_split[1] + "_" + "".join(x_split[2:] )
def __getitem__(self : Optional[int] , snake_case__ : str ) -> Callable[[], Tuple[nn.Module, Dict]]:
'''simple docstring'''
if x not in self:
snake_case : Dict = self.convert_name_to_timm(snake_case__ )
snake_case : Union[str, Any] = partial(lambda: (timm.create_model(snake_case__ , pretrained=snake_case__ ).eval(), None) )
else:
snake_case : List[str] = super().__getitem__(snake_case__ )
return val
class UpperCAmelCase ( A_ ):
def __getitem__(self : Dict , snake_case__ : str ) -> Callable[[], nn.Module]:
'''simple docstring'''
if "seer" in x and "in1k" not in x:
snake_case : str = RegNetModel
else:
snake_case : Optional[Any] = RegNetForImageClassification
return val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Tuple[str, str]] ):
for from_key, to_key in keys:
snake_case : str = from_state_dict[from_key].clone()
print(f"""Copied key={from_key} to={to_key}""" )
return to_state_dict
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : RegNetConfig , __lowerCamelCase : Path , __lowerCamelCase : bool = True , ):
print(f"""Converting {name}...""" )
with torch.no_grad():
snake_case , snake_case : int = from_model_func()
snake_case : str = our_model_func(__lowerCamelCase ).eval()
snake_case : int = ModuleTransfer(src=__lowerCamelCase , dest=__lowerCamelCase , raise_if_mismatch=__lowerCamelCase )
snake_case : Dict = torch.randn((1, 3, 224, 224) )
module_transfer(__lowerCamelCase )
if from_state_dict is not None:
snake_case : str = []
# for seer - in1k finetuned we have to manually copy the head
if "seer" in name and "in1k" in name:
snake_case : Tuple = [("0.clf.0.weight", "classifier.1.weight"), ("0.clf.0.bias", "classifier.1.bias")]
snake_case : Optional[Any] = manually_copy_vissl_head(__lowerCamelCase , our_model.state_dict() , __lowerCamelCase )
our_model.load_state_dict(__lowerCamelCase )
snake_case : Any = our_model(__lowerCamelCase , output_hidden_states=__lowerCamelCase )
snake_case : Union[str, Any] = (
our_outputs.logits if isinstance(__lowerCamelCase , __lowerCamelCase ) else our_outputs.last_hidden_state
)
snake_case : Union[str, Any] = from_model(__lowerCamelCase )
snake_case : Dict = from_output[-1] if type(__lowerCamelCase ) is list else from_output
# now since I don't want to use any config files, vissl seer model doesn't actually have an head, so let's just check the last hidden state
if "seer" in name and "in1k" in name:
snake_case : Any = our_outputs.hidden_states[-1]
assert torch.allclose(__lowerCamelCase , __lowerCamelCase ), "The model logits don't match the original one."
if push_to_hub:
our_model.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add model" , use_temp_dir=__lowerCamelCase , )
snake_case : List[str] = 224 if "seer" not in name else 384
# we can use the convnext one
snake_case : int = AutoImageProcessor.from_pretrained("facebook/convnext-base-224-22k-1k" , size=__lowerCamelCase )
image_processor.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add image processor" , use_temp_dir=__lowerCamelCase , )
print(f"""Pushed {name}""" )
def UpperCamelCase ( __lowerCamelCase : Path , __lowerCamelCase : str = None , __lowerCamelCase : bool = True ):
snake_case : Union[str, Any] = "imagenet-1k-id2label.json"
snake_case : List[str] = 1000
snake_case : List[str] = (1, num_labels)
snake_case : Any = "huggingface/label-files"
snake_case : List[str] = num_labels
snake_case : Optional[Any] = json.load(open(cached_download(hf_hub_url(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) ) , "r" ) )
snake_case : List[Any] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : str = idalabel
snake_case : List[Any] = {v: k for k, v in idalabel.items()}
snake_case : Dict = partial(__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase )
snake_case : Optional[Any] = {
"regnet-x-002": ImageNetPreTrainedConfig(
depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 , layer_type="x" ),
"regnet-x-004": ImageNetPreTrainedConfig(
depths=[1, 2, 7, 12] , hidden_sizes=[32, 64, 160, 384] , groups_width=16 , layer_type="x" ),
"regnet-x-006": ImageNetPreTrainedConfig(
depths=[1, 3, 5, 7] , hidden_sizes=[48, 96, 240, 528] , groups_width=24 , layer_type="x" ),
"regnet-x-008": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 5] , hidden_sizes=[64, 128, 288, 672] , groups_width=16 , layer_type="x" ),
"regnet-x-016": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 2] , hidden_sizes=[72, 168, 408, 912] , groups_width=24 , layer_type="x" ),
"regnet-x-032": ImageNetPreTrainedConfig(
depths=[2, 6, 15, 2] , hidden_sizes=[96, 192, 432, 1008] , groups_width=48 , layer_type="x" ),
"regnet-x-040": ImageNetPreTrainedConfig(
depths=[2, 5, 14, 2] , hidden_sizes=[80, 240, 560, 1360] , groups_width=40 , layer_type="x" ),
"regnet-x-064": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 392, 784, 1624] , groups_width=56 , layer_type="x" ),
"regnet-x-080": ImageNetPreTrainedConfig(
depths=[2, 5, 15, 1] , hidden_sizes=[80, 240, 720, 1920] , groups_width=120 , layer_type="x" ),
"regnet-x-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 , layer_type="x" ),
"regnet-x-160": ImageNetPreTrainedConfig(
depths=[2, 6, 13, 1] , hidden_sizes=[256, 512, 896, 2048] , groups_width=128 , layer_type="x" ),
"regnet-x-320": ImageNetPreTrainedConfig(
depths=[2, 7, 13, 1] , hidden_sizes=[336, 672, 1344, 2520] , groups_width=168 , layer_type="x" ),
# y variant
"regnet-y-002": ImageNetPreTrainedConfig(depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 ),
"regnet-y-004": ImageNetPreTrainedConfig(
depths=[1, 3, 6, 6] , hidden_sizes=[48, 104, 208, 440] , groups_width=8 ),
"regnet-y-006": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 4] , hidden_sizes=[48, 112, 256, 608] , groups_width=16 ),
"regnet-y-008": ImageNetPreTrainedConfig(
depths=[1, 3, 8, 2] , hidden_sizes=[64, 128, 320, 768] , groups_width=16 ),
"regnet-y-016": ImageNetPreTrainedConfig(
depths=[2, 6, 17, 2] , hidden_sizes=[48, 120, 336, 888] , groups_width=24 ),
"regnet-y-032": ImageNetPreTrainedConfig(
depths=[2, 5, 13, 1] , hidden_sizes=[72, 216, 576, 1512] , groups_width=24 ),
"regnet-y-040": ImageNetPreTrainedConfig(
depths=[2, 6, 12, 2] , hidden_sizes=[128, 192, 512, 1088] , groups_width=64 ),
"regnet-y-064": ImageNetPreTrainedConfig(
depths=[2, 7, 14, 2] , hidden_sizes=[144, 288, 576, 1296] , groups_width=72 ),
"regnet-y-080": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 448, 896, 2016] , groups_width=56 ),
"regnet-y-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 ),
"regnet-y-160": ImageNetPreTrainedConfig(
depths=[2, 4, 11, 1] , hidden_sizes=[224, 448, 1232, 3024] , groups_width=112 ),
"regnet-y-320": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
# models created by SEER -> https://arxiv.org/abs/2202.08360
"regnet-y-320-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer": RegNetConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer": RegNetConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
# finetuned on imagenet
"regnet-y-320-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer-in1k": ImageNetPreTrainedConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
}
snake_case : Union[str, Any] = NameToOurModelFuncMap()
snake_case : str = NameToFromModelFuncMap()
# add seer weights logic
def load_using_classy_vision(__lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] ) -> Tuple[nn.Module, Dict]:
snake_case : List[Any] = torch.hub.load_state_dict_from_url(__lowerCamelCase , model_dir=str(__lowerCamelCase ) , map_location="cpu" )
snake_case : Dict = model_func()
# check if we have a head, if yes add it
snake_case : str = files["classy_state_dict"]["base_model"]["model"]
snake_case : Dict = model_state_dict["trunk"]
model.load_state_dict(__lowerCamelCase )
return model.eval(), model_state_dict["heads"]
# pretrained
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet32d/seer_regnet32gf_model_iteration244000.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Optional[int] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet64/seer_regnet64gf_model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : List[str] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/swav_ig1b_regnet128Gf_cnstant_bs32_node16_sinkhorn10_proto16k_syncBN64_warmup8k/model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet10B/model_iteration124500_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
# IN1K finetuned
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet32_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet64_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : str = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet128_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Dict = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_10b_finetuned_in1k_model_phase28_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
if model_name:
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , names_to_config[model_name] , __lowerCamelCase , __lowerCamelCase , )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , )
return config, expected_shape
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help=(
"""The name of the model you wish to convert, it must be one of the supported regnet* architecture,"""
""" currently: regnetx-*, regnety-*. If `None`, all of them will the converted."""
),
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default=None,
type=Path,
required=True,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument(
"""--push_to_hub""",
default=True,
type=bool,
required=False,
help="""If True, push model and image processor to the hub.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 59
| 1
|
import enum
import shutil
import sys
__lowerCamelCase, __lowerCamelCase = shutil.get_terminal_size()
__lowerCamelCase = {"""UP""": """A""", """DOWN""": """B""", """RIGHT""": """C""", """LEFT""": """D"""}
class UpperCAmelCase ( enum.Enum ):
A__ : Dict = 0
A__ : str = 1
def UpperCamelCase ( __lowerCamelCase : List[str] , __lowerCamelCase : Tuple="" ):
sys.stdout.write(str(__lowerCamelCase ) + end )
sys.stdout.flush()
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple="" ):
forceWrite(f"""\u001b[{color}m{content}\u001b[0m""" , __lowerCamelCase )
def UpperCamelCase ( ):
forceWrite("\r" )
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : str ):
forceWrite(f"""\033[{num_lines}{CURSOR_TO_CHAR[direction.upper()]}""" )
def UpperCamelCase ( ):
forceWrite(" " * TERMINAL_WIDTH )
reset_cursor()
def UpperCamelCase ( ):
reset_cursor()
forceWrite("-" * TERMINAL_WIDTH )
| 59
|
import warnings
from typing import Dict
import numpy as np
from ..utils import ExplicitEnum, add_end_docstrings, is_tf_available, is_torch_available
from .base import PIPELINE_INIT_ARGS, GenericTensor, Pipeline
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
return 1.0 / (1.0 + np.exp(-_outputs ))
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Tuple = np.max(_outputs , axis=-1 , keepdims=__lowerCamelCase )
snake_case : int = np.exp(_outputs - maxes )
return shifted_exp / shifted_exp.sum(axis=-1 , keepdims=__lowerCamelCase )
class UpperCAmelCase ( A_ ):
A__ : Any = "sigmoid"
A__ : str = "softmax"
A__ : int = "none"
@add_end_docstrings(
A_ ,r"\n return_all_scores (`bool`, *optional*, defaults to `False`):\n Whether to return all prediction scores or just the one of the predicted class.\n function_to_apply (`str`, *optional*, defaults to `\"default\"`):\n The function to apply to the model outputs in order to retrieve the scores. Accepts four different values:\n\n - `\"default\"`: if the model has a single label, will apply the sigmoid function on the output. If the model\n has several labels, will apply the softmax function on the output.\n - `\"sigmoid\"`: Applies the sigmoid function on the output.\n - `\"softmax\"`: Applies the softmax function on the output.\n - `\"none\"`: Does not apply any function on the output.\n " ,)
class UpperCAmelCase ( A_ ):
A__ : int = False
A__ : Union[str, Any] = ClassificationFunction.NONE
def __init__(self : List[str] , **snake_case__ : int ) -> str:
'''simple docstring'''
super().__init__(**snake_case__ )
self.check_model_type(
TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if self.framework == "tf"
else MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str]=None , snake_case__ : Optional[Any]=None , snake_case__ : Union[str, Any]="" , **snake_case__ : List[str] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = tokenizer_kwargs
snake_case : List[Any] = {}
if hasattr(self.model.config , "return_all_scores" ) and return_all_scores is None:
snake_case : Optional[int] = self.model.config.return_all_scores
if isinstance(snake_case__ , snake_case__ ) or top_k is None:
snake_case : List[Any] = top_k
snake_case : str = False
elif return_all_scores is not None:
warnings.warn(
"`return_all_scores` is now deprecated, if want a similar functionality use `top_k=None` instead of"
" `return_all_scores=True` or `top_k=1` instead of `return_all_scores=False`." , snake_case__ , )
if return_all_scores:
snake_case : List[str] = None
else:
snake_case : Optional[int] = 1
if isinstance(snake_case__ , snake_case__ ):
snake_case : Dict = ClassificationFunction[function_to_apply.upper()]
if function_to_apply is not None:
snake_case : Optional[int] = function_to_apply
return preprocess_params, {}, postprocess_params
def __call__(self : Dict , *snake_case__ : List[str] , **snake_case__ : int ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = super().__call__(*snake_case__ , **snake_case__ )
# TODO try and retrieve it in a nicer way from _sanitize_parameters.
snake_case : Tuple = "top_k" not in kwargs
if isinstance(args[0] , snake_case__ ) and _legacy:
# This pipeline is odd, and return a list when single item is run
return [result]
else:
return result
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Tuple , **snake_case__ : Union[str, Any] ) -> Dict[str, GenericTensor]:
'''simple docstring'''
snake_case : int = self.framework
if isinstance(snake_case__ , snake_case__ ):
return self.tokenizer(**snake_case__ , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ) and len(snake_case__ ) == 1 and isinstance(inputs[0] , snake_case__ ) and len(inputs[0] ) == 2:
# It used to be valid to use a list of list of list for text pairs, keeping this path for BC
return self.tokenizer(
text=inputs[0][0] , text_pair=inputs[0][1] , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
# This is likely an invalid usage of the pipeline attempting to pass text pairs.
raise ValueError(
"The pipeline received invalid inputs, if you are trying to send text pairs, you can try to send a"
" dictionary `{\"text\": \"My text\", \"text_pair\": \"My pair\"}` in order to send a text pair." )
return self.tokenizer(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Union[str, Any] ) -> int:
'''simple docstring'''
return self.model(**snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Optional[Any] , snake_case__ : List[str]=None , snake_case__ : Dict=1 , snake_case__ : Tuple=True ) -> str:
'''simple docstring'''
if function_to_apply is None:
if self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels == 1:
snake_case : Tuple = ClassificationFunction.SIGMOID
elif self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels > 1:
snake_case : Tuple = ClassificationFunction.SOFTMAX
elif hasattr(self.model.config , "function_to_apply" ) and function_to_apply is None:
snake_case : Tuple = self.model.config.function_to_apply
else:
snake_case : int = ClassificationFunction.NONE
snake_case : Any = model_outputs["logits"][0]
snake_case : List[str] = outputs.numpy()
if function_to_apply == ClassificationFunction.SIGMOID:
snake_case : Optional[Any] = sigmoid(snake_case__ )
elif function_to_apply == ClassificationFunction.SOFTMAX:
snake_case : Union[str, Any] = softmax(snake_case__ )
elif function_to_apply == ClassificationFunction.NONE:
snake_case : Optional[Any] = outputs
else:
raise ValueError(f"""Unrecognized `function_to_apply` argument: {function_to_apply}""" )
if top_k == 1 and _legacy:
return {"label": self.model.config.idalabel[scores.argmax().item()], "score": scores.max().item()}
snake_case : Optional[int] = [
{"label": self.model.config.idalabel[i], "score": score.item()} for i, score in enumerate(snake_case__ )
]
if not _legacy:
dict_scores.sort(key=lambda snake_case__ : x["score"] , reverse=snake_case__ )
if top_k is not None:
snake_case : Optional[int] = dict_scores[:top_k]
return dict_scores
| 59
| 1
|
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )]
snake_case : int = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1 or len(__lowerCamelCase ) <= key:
return input_string
for position, character in enumerate(__lowerCamelCase ):
snake_case : Any = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append(__lowerCamelCase )
snake_case : List[str] = ["".join(__lowerCamelCase ) for row in temp_grid]
snake_case : Tuple = "".join(__lowerCamelCase )
return output_string
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : Dict = []
snake_case : Union[str, Any] = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1:
return input_string
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )] # generates template
for position in range(len(__lowerCamelCase ) ):
snake_case : List[str] = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append("*" )
snake_case : Tuple = 0
for row in temp_grid: # fills in the characters
snake_case : Union[str, Any] = input_string[counter : counter + len(__lowerCamelCase )]
grid.append(list(__lowerCamelCase ) )
counter += len(__lowerCamelCase )
snake_case : str = "" # reads as zigzag
for position in range(len(__lowerCamelCase ) ):
snake_case : Optional[int] = position % (lowest * 2) # puts it in bounds
snake_case : Tuple = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
output_string += grid[num][0]
grid[num].pop(0 )
return output_string
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : Tuple = {}
for key_guess in range(1 , len(__lowerCamelCase ) ): # tries every key
snake_case : Any = decrypt(__lowerCamelCase , __lowerCamelCase )
return results
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
|
from __future__ import annotations
__lowerCamelCase = list[list[int]]
# assigning initial values to the grid
__lowerCamelCase = [
[3, 0, 6, 5, 0, 8, 4, 0, 0],
[5, 2, 0, 0, 0, 0, 0, 0, 0],
[0, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
# a grid with no solution
__lowerCamelCase = [
[5, 0, 6, 5, 0, 8, 4, 0, 3],
[5, 2, 0, 0, 0, 0, 0, 0, 2],
[1, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
def UpperCamelCase ( __lowerCamelCase : Matrix , __lowerCamelCase : int , __lowerCamelCase : int , __lowerCamelCase : int ):
for i in range(9 ):
if grid[row][i] == n or grid[i][column] == n:
return False
for i in range(3 ):
for j in range(3 ):
if grid[(row - row % 3) + i][(column - column % 3) + j] == n:
return False
return True
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for i in range(9 ):
for j in range(9 ):
if grid[i][j] == 0:
return i, j
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
if location := find_empty_location(__lowerCamelCase ):
snake_case , snake_case : Union[str, Any] = location
else:
# If the location is ``None``, then the grid is solved.
return grid
for digit in range(1 , 10 ):
if is_safe(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase ):
snake_case : List[Any] = digit
if sudoku(__lowerCamelCase ) is not None:
return grid
snake_case : Union[str, Any] = 0
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for row in grid:
for cell in row:
print(__lowerCamelCase , end=" " )
print()
if __name__ == "__main__":
# make a copy of grid so that you can compare with the unmodified grid
for example_grid in (initial_grid, no_solution):
print("""\nExample grid:\n""" + """=""" * 20)
print_solution(example_grid)
print("""\nExample grid solution:""")
__lowerCamelCase = sudoku(example_grid)
if solution is not None:
print_solution(solution)
else:
print("""Cannot find a solution.""")
| 59
| 1
|
import re
from filelock import FileLock
try:
import nltk
__lowerCamelCase = True
except (ImportError, ModuleNotFoundError):
__lowerCamelCase = False
if NLTK_AVAILABLE:
with FileLock(""".lock""") as lock:
nltk.download("""punkt""", quiet=True)
def UpperCamelCase ( __lowerCamelCase : str ):
re.sub("<n>" , "" , __lowerCamelCase ) # remove pegasus newline char
assert NLTK_AVAILABLE, "nltk must be installed to separate newlines between sentences. (pip install nltk)"
return "\n".join(nltk.sent_tokenize(__lowerCamelCase ) )
| 59
|
import logging
import numpy as np
import pytest
from scipy.linalg import eigh
logging.basicConfig(level=logging.INFO, format="""%(message)s""")
def UpperCamelCase ( __lowerCamelCase : np.ndarray ):
return input_array.reshape((input_array.size, 1) )
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Any = np.nan
for i in range(__lowerCamelCase ):
snake_case : List[str] = features[:, labels == i]
snake_case : Dict = data.mean(1 )
# Centralize the data of class i
snake_case : Optional[Any] = data - column_reshape(__lowerCamelCase )
if i > 0:
# If covariance_sum is not None
covariance_sum += np.dot(__lowerCamelCase , centered_data.T )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Optional[Any] = features.mean(1 )
snake_case : Tuple = np.nan
for i in range(__lowerCamelCase ):
snake_case : Tuple = features[:, labels == i]
snake_case : Tuple = data.shape[1]
snake_case : List[str] = data.mean(1 )
if i > 0:
# If covariance_sum is not None
covariance_sum += device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[int] = device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
# Check if the features have been loaded
if features.any():
snake_case : Tuple = features.mean(1 )
# Center the dataset
snake_case : List[str] = features - np.reshape(__lowerCamelCase , (data_mean.size, 1) )
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T ) / features.shape[1]
snake_case , snake_case : Dict = np.linalg.eigh(__lowerCamelCase )
# Take all the columns in the reverse order (-1), and then takes only the first
snake_case : Optional[Any] = eigenvectors[:, ::-1][:, 0:dimensions]
# Project the database on the new space
snake_case : Union[str, Any] = np.dot(filtered_eigenvectors.T , __lowerCamelCase )
logging.info("Principal Component Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int , __lowerCamelCase : int ):
assert classes > dimensions
# Check if features have been already loaded
if features.any:
snake_case , snake_case : str = eigh(
covariance_between_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , covariance_within_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , )
snake_case : str = eigenvectors[:, ::-1][:, :dimensions]
snake_case , snake_case , snake_case : int = np.linalg.svd(__lowerCamelCase )
snake_case : List[Any] = svd_matrix[:, 0:dimensions]
snake_case : Optional[Any] = np.dot(filtered_svd_matrix.T , __lowerCamelCase )
logging.info("Linear Discriminant Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( ):
# Create dummy dataset with 2 classes and 3 features
snake_case : str = np.array([[1, 2, 3, 4, 5], [2, 3, 4, 5, 6], [3, 4, 5, 6, 7]] )
snake_case : Union[str, Any] = np.array([0, 0, 0, 1, 1] )
snake_case : List[Any] = 2
snake_case : Any = 2
# Assert that the function raises an AssertionError if dimensions > classes
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : str = linear_discriminant_analysis(
__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
if isinstance(__lowerCamelCase , np.ndarray ):
raise AssertionError(
"Did not raise AssertionError for dimensions > classes" )
assert error_info.type is AssertionError
def UpperCamelCase ( ):
snake_case : List[str] = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]] )
snake_case : List[str] = 2
snake_case : int = np.array([[6.9282_0323, 8.6602_5404, 10.3923_0485], [3.0, 3.0, 3.0]] )
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : Union[str, Any] = principal_component_analysis(__lowerCamelCase , __lowerCamelCase )
if not np.allclose(__lowerCamelCase , __lowerCamelCase ):
raise AssertionError
assert error_info.type is AssertionError
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
import pickle
import numpy as np
from matplotlib import pyplot as plt
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Union[str, Any] , snake_case__ : Dict , snake_case__ : Optional[int] , snake_case__ : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : int=0.2 , snake_case__ : Any=0.2 ) -> Optional[Any]:
'''simple docstring'''
snake_case : Optional[int] = bp_numa
snake_case : Any = bp_numa
snake_case : Union[str, Any] = bp_numa
snake_case : Optional[Any] = conva_get[:2]
snake_case : Optional[int] = conva_get[2]
snake_case : List[Any] = size_pa
snake_case : List[Any] = rate_w
snake_case : Optional[Any] = rate_t
snake_case : Union[str, Any] = [
np.mat(-1 * np.random.rand(self.conva[0] , self.conva[0] ) + 0.5 )
for i in range(self.conva[1] )
]
snake_case : Tuple = np.mat(-1 * np.random.rand(self.num_bpa , self.num_bpa ) + 0.5 )
snake_case : Any = np.mat(-1 * np.random.rand(self.num_bpa , self.num_bpa ) + 0.5 )
snake_case : Dict = -2 * np.random.rand(self.conva[1] ) + 1
snake_case : List[str] = -2 * np.random.rand(self.num_bpa ) + 1
snake_case : int = -2 * np.random.rand(self.num_bpa ) + 1
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Any ) -> Tuple:
'''simple docstring'''
snake_case : Optional[int] = {
"num_bp1": self.num_bpa,
"num_bp2": self.num_bpa,
"num_bp3": self.num_bpa,
"conv1": self.conva,
"step_conv1": self.step_conva,
"size_pooling1": self.size_poolinga,
"rate_weight": self.rate_weight,
"rate_thre": self.rate_thre,
"w_conv1": self.w_conva,
"wkj": self.wkj,
"vji": self.vji,
"thre_conv1": self.thre_conva,
"thre_bp2": self.thre_bpa,
"thre_bp3": self.thre_bpa,
}
with open(snake_case__ , "wb" ) as f:
pickle.dump(snake_case__ , snake_case__ )
print(f"""Model saved： {save_path}""" )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Dict , snake_case__ : str ) -> Union[str, Any]:
'''simple docstring'''
with open(snake_case__ , "rb" ) as f:
snake_case : int = pickle.load(snake_case__ ) # noqa: S301
snake_case : Union[str, Any] = model_dic.get("conv1" )
conv_get.append(model_dic.get("step_conv1" ) )
snake_case : Any = model_dic.get("size_pooling1" )
snake_case : str = model_dic.get("num_bp1" )
snake_case : List[str] = model_dic.get("num_bp2" )
snake_case : Union[str, Any] = model_dic.get("num_bp3" )
snake_case : Optional[Any] = model_dic.get("rate_weight" )
snake_case : Union[str, Any] = model_dic.get("rate_thre" )
# create model instance
snake_case : str = CNN(snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# modify model parameter
snake_case : Optional[int] = model_dic.get("w_conv1" )
snake_case : int = model_dic.get("wkj" )
snake_case : Dict = model_dic.get("vji" )
snake_case : Dict = model_dic.get("thre_conv1" )
snake_case : Tuple = model_dic.get("thre_bp2" )
snake_case : Any = model_dic.get("thre_bp3" )
return conv_ins
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : int ) -> List[str]:
'''simple docstring'''
return 1 / (1 + np.exp(-1 * x ))
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[int] ) -> List[str]:
'''simple docstring'''
return round(snake_case__ , 3 )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : List[str] , snake_case__ : Tuple , snake_case__ : List[Any] , snake_case__ : int , snake_case__ : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : str = convs[0]
snake_case : Optional[Any] = convs[1]
snake_case : str = np.shape(snake_case__ )[0]
# get the data slice of original image data, data_focus
snake_case : Any = []
for i_focus in range(0 , size_data - size_conv + 1 , snake_case__ ):
for j_focus in range(0 , size_data - size_conv + 1 , snake_case__ ):
snake_case : Any = data[
i_focus : i_focus + size_conv, j_focus : j_focus + size_conv
]
data_focus.append(snake_case__ )
# calculate the feature map of every single kernel, and saved as list of matrix
snake_case : int = []
snake_case : Dict = int((size_data - size_conv) / conv_step + 1 )
for i_map in range(snake_case__ ):
snake_case : Union[str, Any] = []
for i_focus in range(len(snake_case__ ) ):
snake_case : Union[str, Any] = (
np.sum(np.multiply(data_focus[i_focus] , w_convs[i_map] ) )
- thre_convs[i_map]
)
featuremap.append(self.sig(snake_case__ ) )
snake_case : str = np.asmatrix(snake_case__ ).reshape(
snake_case__ , snake_case__ )
data_featuremap.append(snake_case__ )
# expanding the data slice to One dimenssion
snake_case : str = []
for each_focus in data_focus:
focusa_list.extend(self.Expand_Mat(snake_case__ ) )
snake_case : Dict = np.asarray(snake_case__ )
return focus_list, data_featuremap
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Dict , snake_case__ : List[str] , snake_case__ : List[str]="average_pool" ) -> List[str]:
'''simple docstring'''
snake_case : str = len(featuremaps[0] )
snake_case : Optional[int] = int(size_map / size_pooling )
snake_case : Union[str, Any] = []
for i_map in range(len(snake_case__ ) ):
snake_case : Any = featuremaps[i_map]
snake_case : Dict = []
for i_focus in range(0 , snake_case__ , snake_case__ ):
for j_focus in range(0 , snake_case__ , snake_case__ ):
snake_case : Optional[int] = feature_map[
i_focus : i_focus + size_pooling,
j_focus : j_focus + size_pooling,
]
if pooling_type == "average_pool":
# average pooling
map_pooled.append(np.average(snake_case__ ) )
elif pooling_type == "max_pooling":
# max pooling
map_pooled.append(np.max(snake_case__ ) )
snake_case : List[str] = np.asmatrix(snake_case__ ).reshape(snake_case__ , snake_case__ )
featuremap_pooled.append(snake_case__ )
return featuremap_pooled
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : Optional[int] = []
for i in range(len(snake_case__ ) ):
snake_case : Any = np.shape(data[i] )
snake_case : List[str] = data[i].reshape(1 , shapes[0] * shapes[1] )
snake_case : Optional[int] = data_listed.getA().tolist()[0]
data_expanded.extend(snake_case__ )
snake_case : Union[str, Any] = np.asarray(snake_case__ )
return data_expanded
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str] ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = np.asarray(snake_case__ )
snake_case : Tuple = np.shape(snake_case__ )
snake_case : Union[str, Any] = data_mat.reshape(1 , shapes[0] * shapes[1] )
return data_expanded
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] , snake_case__ : Optional[int] ) -> Tuple:
'''simple docstring'''
snake_case : Optional[int] = []
snake_case : Tuple = 0
for i_map in range(snake_case__ ):
snake_case : Tuple = np.ones((size_map, size_map) )
for i in range(0 , snake_case__ , snake_case__ ):
for j in range(0 , snake_case__ , snake_case__ ):
snake_case : Dict = pd_pool[
i_pool
]
snake_case : int = i_pool + 1
snake_case : Tuple = np.multiply(
snake_case__ , np.multiply(out_map[i_map] , (1 - out_map[i_map]) ) )
pd_all.append(snake_case__ )
return pd_all
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Optional[int] , snake_case__ : Optional[int] , snake_case__ : List[Any] , snake_case__ : Any , snake_case__ : Optional[int] , snake_case__ : Any=bool ) -> str:
'''simple docstring'''
print("----------------------Start Training-------------------------" )
print((" - - Shape: Train_Data ", np.shape(snake_case__ )) )
print((" - - Shape: Teach_Data ", np.shape(snake_case__ )) )
snake_case : Optional[Any] = 0
snake_case : str = []
snake_case : Optional[int] = 1_00_00
while rp < n_repeat and mse >= error_accuracy:
snake_case : Any = 0
print(f"""-------------Learning Time {rp}--------------""" )
for p in range(len(snake_case__ ) ):
# print('------------Learning Image: %d--------------'%p)
snake_case : List[Any] = np.asmatrix(datas_train[p] )
snake_case : Dict = np.asarray(datas_teach[p] )
snake_case , snake_case : str = self.convolute(
snake_case__ , self.conva , self.w_conva , self.thre_conva , conv_step=self.step_conva , )
snake_case : List[str] = self.pooling(snake_case__ , self.size_poolinga )
snake_case : Tuple = np.shape(snake_case__ )
snake_case : int = self._expand(snake_case__ )
snake_case : Tuple = data_bp_input
snake_case : str = np.dot(snake_case__ , self.vji.T ) - self.thre_bpa
snake_case : int = self.sig(snake_case__ )
snake_case : Tuple = np.dot(snake_case__ , self.wkj.T ) - self.thre_bpa
snake_case : Any = self.sig(snake_case__ )
# --------------Model Leaning ------------------------
# calculate error and gradient---------------
snake_case : Any = np.multiply(
(data_teach - bp_outa) , np.multiply(snake_case__ , (1 - bp_outa) ) )
snake_case : Union[str, Any] = np.multiply(
np.dot(snake_case__ , self.wkj ) , np.multiply(snake_case__ , (1 - bp_outa) ) )
snake_case : List[str] = np.dot(snake_case__ , self.vji )
snake_case : Tuple = pd_i_all / (self.size_poolinga * self.size_poolinga)
snake_case : str = pd_conva_pooled.T.getA().tolist()
snake_case : str = self._calculate_gradient_from_pool(
snake_case__ , snake_case__ , shape_featuremapa[0] , shape_featuremapa[1] , self.size_poolinga , )
# weight and threshold learning process---------
# convolution layer
for k_conv in range(self.conva[1] ):
snake_case : Union[str, Any] = self._expand_mat(pd_conva_all[k_conv] )
snake_case : Tuple = self.rate_weight * np.dot(snake_case__ , snake_case__ )
snake_case : int = self.w_conva[k_conv] + delta_w.reshape(
(self.conva[0], self.conva[0]) )
snake_case : Optional[Any] = (
self.thre_conva[k_conv]
- np.sum(pd_conva_all[k_conv] ) * self.rate_thre
)
# all connected layer
snake_case : Any = self.wkj + pd_k_all.T * bp_outa * self.rate_weight
snake_case : str = self.vji + pd_j_all.T * bp_outa * self.rate_weight
snake_case : int = self.thre_bpa - pd_k_all * self.rate_thre
snake_case : Optional[int] = self.thre_bpa - pd_j_all * self.rate_thre
# calculate the sum error of all single image
snake_case : Optional[int] = np.sum(abs(data_teach - bp_outa ) )
error_count += errors
# print(' ----Teach ',data_teach)
# print(' ----BP_output ',bp_out3)
snake_case : Tuple = rp + 1
snake_case : List[str] = error_count / patterns
all_mse.append(snake_case__ )
def draw_error():
snake_case : Optional[int] = [error_accuracy for i in range(int(n_repeat * 1.2 ) )]
plt.plot(snake_case__ , "+-" )
plt.plot(snake_case__ , "r--" )
plt.xlabel("Learning Times" )
plt.ylabel("All_mse" )
plt.grid(snake_case__ , alpha=0.5 )
plt.show()
print("------------------Training Complished---------------------" )
print((" - - Training epoch: ", rp, f""" - - Mse: {mse:.6f}""") )
if draw_e:
draw_error()
return mse
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Dict = []
print("-------------------Start Testing-------------------------" )
print((" - - Shape: Test_Data ", np.shape(snake_case__ )) )
for p in range(len(snake_case__ ) ):
snake_case : str = np.asmatrix(datas_test[p] )
snake_case , snake_case : Optional[Any] = self.convolute(
snake_case__ , self.conva , self.w_conva , self.thre_conva , conv_step=self.step_conva , )
snake_case : Tuple = self.pooling(snake_case__ , self.size_poolinga )
snake_case : Any = self._expand(snake_case__ )
snake_case : Dict = data_bp_input
snake_case : Union[str, Any] = bp_outa * self.vji.T - self.thre_bpa
snake_case : List[str] = self.sig(snake_case__ )
snake_case : Optional[Any] = bp_outa * self.wkj.T - self.thre_bpa
snake_case : Union[str, Any] = self.sig(snake_case__ )
produce_out.extend(bp_outa.getA().tolist() )
snake_case : List[Any] = [list(map(self.do_round , snake_case__ ) ) for each in produce_out]
return np.asarray(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Dict ) -> Dict:
'''simple docstring'''
snake_case : List[Any] = np.asmatrix(snake_case__ )
snake_case , snake_case : Tuple = self.convolute(
snake_case__ , self.conva , self.w_conva , self.thre_conva , conv_step=self.step_conva , )
snake_case : str = self.pooling(snake_case__ , self.size_poolinga )
return data_conveda, data_pooleda
if __name__ == "__main__":
pass
| 59
|
import pytest
from datasets import inspect_metric, list_metrics, load_metric
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : Optional[int] ):
monkeypatch.setattr("datasets.utils.deprecation_utils._emitted_deprecation_warnings" , set() )
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : str ):
class UpperCAmelCase :
def __init__(self : Optional[int] , snake_case__ : str ) -> Any:
'''simple docstring'''
snake_case : List[str] = metric_id
class UpperCAmelCase :
A__ : List[str] = [MetricMock(A_ ) for metric_id in ["accuracy", "mse", "precision", "codeparrot/apps_metric"]]
def _SCREAMING_SNAKE_CASE (self : int ) -> List[str]:
'''simple docstring'''
return self._metrics
monkeypatch.setattr("datasets.inspect.huggingface_hub" , HfhMock() )
@pytest.mark.parametrize(
"func, args" , [(load_metric, ("metrics/mse",)), (list_metrics, ()), (inspect_metric, ("metrics/mse", "tmp_path"))] )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Any , __lowerCamelCase : Any , __lowerCamelCase : int , __lowerCamelCase : Any ):
if "tmp_path" in args:
snake_case : str = tuple(arg if arg != "tmp_path" else tmp_path for arg in args )
with pytest.warns(__lowerCamelCase , match="https://huggingface.co/docs/evaluate" ):
func(*__lowerCamelCase )
| 59
| 1
|
__lowerCamelCase = """0.18.2"""
from .configuration_utils import ConfigMixin
from .utils import (
OptionalDependencyNotAvailable,
is_flax_available,
is_inflect_available,
is_invisible_watermark_available,
is_k_diffusion_available,
is_k_diffusion_version,
is_librosa_available,
is_note_seq_available,
is_onnx_available,
is_scipy_available,
is_torch_available,
is_torchsde_available,
is_transformers_available,
is_transformers_version,
is_unidecode_available,
logging,
)
try:
if not is_onnx_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_onnx_objects import * # noqa F403
else:
from .pipelines import OnnxRuntimeModel
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_pt_objects import * # noqa F403
else:
from .models import (
AutoencoderKL,
ControlNetModel,
ModelMixin,
PriorTransformer,
TaFilmDecoder,
TransformeraDModel,
UNetaDModel,
UNetaDConditionModel,
UNetaDModel,
UNetaDConditionModel,
VQModel,
)
from .optimization import (
get_constant_schedule,
get_constant_schedule_with_warmup,
get_cosine_schedule_with_warmup,
get_cosine_with_hard_restarts_schedule_with_warmup,
get_linear_schedule_with_warmup,
get_polynomial_decay_schedule_with_warmup,
get_scheduler,
)
from .pipelines import (
AudioPipelineOutput,
ConsistencyModelPipeline,
DanceDiffusionPipeline,
DDIMPipeline,
DDPMPipeline,
DiffusionPipeline,
DiTPipeline,
ImagePipelineOutput,
KarrasVePipeline,
LDMPipeline,
LDMSuperResolutionPipeline,
PNDMPipeline,
RePaintPipeline,
ScoreSdeVePipeline,
)
from .schedulers import (
CMStochasticIterativeScheduler,
DDIMInverseScheduler,
DDIMParallelScheduler,
DDIMScheduler,
DDPMParallelScheduler,
DDPMScheduler,
DEISMultistepScheduler,
DPMSolverMultistepInverseScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
EulerAncestralDiscreteScheduler,
EulerDiscreteScheduler,
HeunDiscreteScheduler,
IPNDMScheduler,
KarrasVeScheduler,
KDPMaAncestralDiscreteScheduler,
KDPMaDiscreteScheduler,
PNDMScheduler,
RePaintScheduler,
SchedulerMixin,
ScoreSdeVeScheduler,
UnCLIPScheduler,
UniPCMultistepScheduler,
VQDiffusionScheduler,
)
from .training_utils import EMAModel
try:
if not (is_torch_available() and is_scipy_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_scipy_objects import * # noqa F403
else:
from .schedulers import LMSDiscreteScheduler
try:
if not (is_torch_available() and is_torchsde_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_torchsde_objects import * # noqa F403
else:
from .schedulers import DPMSolverSDEScheduler
try:
if not (is_torch_available() and is_transformers_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_transformers_objects import * # noqa F403
else:
from .pipelines import (
AltDiffusionImgaImgPipeline,
AltDiffusionPipeline,
AudioLDMPipeline,
CycleDiffusionPipeline,
IFImgaImgPipeline,
IFImgaImgSuperResolutionPipeline,
IFInpaintingPipeline,
IFInpaintingSuperResolutionPipeline,
IFPipeline,
IFSuperResolutionPipeline,
ImageTextPipelineOutput,
KandinskyImgaImgPipeline,
KandinskyInpaintPipeline,
KandinskyPipeline,
KandinskyPriorPipeline,
KandinskyVaaControlnetImgaImgPipeline,
KandinskyVaaControlnetPipeline,
KandinskyVaaImgaImgPipeline,
KandinskyVaaInpaintPipeline,
KandinskyVaaPipeline,
KandinskyVaaPriorEmbaEmbPipeline,
KandinskyVaaPriorPipeline,
LDMTextToImagePipeline,
PaintByExamplePipeline,
SemanticStableDiffusionPipeline,
ShapEImgaImgPipeline,
ShapEPipeline,
StableDiffusionAttendAndExcitePipeline,
StableDiffusionControlNetImgaImgPipeline,
StableDiffusionControlNetInpaintPipeline,
StableDiffusionControlNetPipeline,
StableDiffusionDepthaImgPipeline,
StableDiffusionDiffEditPipeline,
StableDiffusionImageVariationPipeline,
StableDiffusionImgaImgPipeline,
StableDiffusionInpaintPipeline,
StableDiffusionInpaintPipelineLegacy,
StableDiffusionInstructPixaPixPipeline,
StableDiffusionLatentUpscalePipeline,
StableDiffusionLDMaDPipeline,
StableDiffusionModelEditingPipeline,
StableDiffusionPanoramaPipeline,
StableDiffusionParadigmsPipeline,
StableDiffusionPipeline,
StableDiffusionPipelineSafe,
StableDiffusionPixaPixZeroPipeline,
StableDiffusionSAGPipeline,
StableDiffusionUpscalePipeline,
StableUnCLIPImgaImgPipeline,
StableUnCLIPPipeline,
TextToVideoSDPipeline,
TextToVideoZeroPipeline,
UnCLIPImageVariationPipeline,
UnCLIPPipeline,
UniDiffuserModel,
UniDiffuserPipeline,
UniDiffuserTextDecoder,
VersatileDiffusionDualGuidedPipeline,
VersatileDiffusionImageVariationPipeline,
VersatileDiffusionPipeline,
VersatileDiffusionTextToImagePipeline,
VideoToVideoSDPipeline,
VQDiffusionPipeline,
)
try:
if not (is_torch_available() and is_transformers_available() and is_invisible_watermark_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_transformers_and_invisible_watermark_objects import * # noqa F403
else:
from .pipelines import StableDiffusionXLImgaImgPipeline, StableDiffusionXLPipeline
try:
if not (is_torch_available() and is_transformers_available() and is_k_diffusion_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_transformers_and_k_diffusion_objects import * # noqa F403
else:
from .pipelines import StableDiffusionKDiffusionPipeline
try:
if not (is_torch_available() and is_transformers_available() and is_onnx_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_transformers_and_onnx_objects import * # noqa F403
else:
from .pipelines import (
OnnxStableDiffusionImgaImgPipeline,
OnnxStableDiffusionInpaintPipeline,
OnnxStableDiffusionInpaintPipelineLegacy,
OnnxStableDiffusionPipeline,
OnnxStableDiffusionUpscalePipeline,
StableDiffusionOnnxPipeline,
)
try:
if not (is_torch_available() and is_librosa_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_torch_and_librosa_objects import * # noqa F403
else:
from .pipelines import AudioDiffusionPipeline, Mel
try:
if not (is_transformers_available() and is_torch_available() and is_note_seq_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_transformers_and_torch_and_note_seq_objects import * # noqa F403
else:
from .pipelines import SpectrogramDiffusionPipeline
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_flax_objects import * # noqa F403
else:
from .models.controlnet_flax import FlaxControlNetModel
from .models.modeling_flax_utils import FlaxModelMixin
from .models.unet_ad_condition_flax import FlaxUNetaDConditionModel
from .models.vae_flax import FlaxAutoencoderKL
from .pipelines import FlaxDiffusionPipeline
from .schedulers import (
FlaxDDIMScheduler,
FlaxDDPMScheduler,
FlaxDPMSolverMultistepScheduler,
FlaxKarrasVeScheduler,
FlaxLMSDiscreteScheduler,
FlaxPNDMScheduler,
FlaxSchedulerMixin,
FlaxScoreSdeVeScheduler,
)
try:
if not (is_flax_available() and is_transformers_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_flax_and_transformers_objects import * # noqa F403
else:
from .pipelines import (
FlaxStableDiffusionControlNetPipeline,
FlaxStableDiffusionImgaImgPipeline,
FlaxStableDiffusionInpaintPipeline,
FlaxStableDiffusionPipeline,
)
try:
if not (is_note_seq_available()):
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from .utils.dummy_note_seq_objects import * # noqa F403
else:
from .pipelines import MidiProcessor
| 59
|
import argparse
import dataclasses
import json
import logging
import os
import shutil
from typing import List, Optional
import datasets
from accelerate import Accelerator
from datasets import load_dataset
from finetuning import finetune
from tqdm.auto import tqdm
import transformers
from transformers import AutoConfig, set_seed
from transformers.trainer_utils import IntervalStrategy
__lowerCamelCase = logging.getLogger(__name__)
__lowerCamelCase = """pytorch_model.bin"""
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co."} ,)
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the training data."} )
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the data to predict on."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "A csv or a json file containing the validation data."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "The name of the task to train on."} ,)
A__ : Optional[List[str]] = dataclasses.field(
default=A_ ,metadata={"help": "The list of labels for the task."} )
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "The output directory where the model predictions and checkpoints will be written."} )
A__ : Optional[str] = dataclasses.field(
default="accuracy" ,metadata={"help": "The evaluation metric used for the task."} )
A__ : Optional[str] = dataclasses.field(
default="no" ,metadata={
"help": "The evaluation strategy to adopt during training. Possible values are: [\"no\", \"step\", \"epoch]"
} ,)
A__ : Optional[int] = dataclasses.field(
default=10 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={
"help": "How much the specified evaluation metric must improve to satisfy early stopping conditions."
} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the confidence score."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the validation performance."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to fine-tune on labeled data after pseudo training."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={"help": "Confidence threshold for pseudo-labeled data filtering."} ,)
A__ : Optional[int] = dataclasses.field(
default=1_00 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[int] = dataclasses.field(
default=A_ ,metadata={"help": "Random seed for initialization."} ,)
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Tuple , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Tuple , __lowerCamelCase : Dict , __lowerCamelCase : Optional[int] ):
snake_case : Tuple = datasets.concatenate_datasets([infer_input, infer_output] , axis=1 )
if args.do_filter_by_confidence:
snake_case : Optional[int] = dataset.filter(lambda __lowerCamelCase : example["probability"] > args.confidence_threshold )
if args.do_filter_by_val_performance:
assert eval_result >= 0.0 and eval_result <= 1.0
snake_case : int = int(eval_result * len(__lowerCamelCase ) )
print(__lowerCamelCase )
snake_case : List[str] = dataset.sort("probability" , reverse=__lowerCamelCase )
snake_case : Tuple = dataset.select(range(__lowerCamelCase ) )
snake_case : List[Any] = dataset.remove_columns(["label", "probability"] )
snake_case : Any = dataset.rename_column("prediction" , "label" )
snake_case : str = dataset.map(lambda __lowerCamelCase : {"label": idalabel[example["label"]]} )
snake_case : List[str] = dataset.shuffle(seed=args.seed )
snake_case : int = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.data_file_extension == "csv":
dataset.to_csv(__lowerCamelCase , index=__lowerCamelCase )
else:
dataset.to_json(__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : str , __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple , **__lowerCamelCase : List[Any] ):
snake_case : int = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(accelerator.state )
# Setup logging, we only want one process per machine to log things on the
# screen. accelerator.is_local_main_process is only True for one process per
# machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
snake_case : Dict = STModelArguments(model_name_or_path=__lowerCamelCase )
snake_case : Tuple = STDataArguments(train_file=__lowerCamelCase , infer_file=__lowerCamelCase )
snake_case : str = STTrainingArguments(output_dir=__lowerCamelCase )
snake_case : int = argparse.Namespace()
for arg_class in (model_args, data_args, training_args):
for key, value in vars(__lowerCamelCase ).items():
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
for key, value in kwargs.items():
if hasattr(__lowerCamelCase , __lowerCamelCase ):
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
# Sanity checks
snake_case : List[str] = {}
snake_case : Optional[int] = None
# You need to provide the training data and the data to predict on
assert args.train_file is not None
assert args.infer_file is not None
snake_case : str = args.train_file
snake_case : Tuple = args.infer_file
if args.evaluation_strategy != IntervalStrategy.NO.value:
assert args.eval_file is not None
snake_case : Tuple = args.eval_file
for key in data_files:
snake_case : List[Any] = data_files[key].split("." )[-1]
assert extension in ["csv", "json"], f"""`{key}_file` should be a csv or a json file."""
if args.data_file_extension is None:
snake_case : Union[str, Any] = extension
else:
assert extension == args.data_file_extension, f"""`{key}_file` should be a {args.data_file_extension} file`."""
assert (
args.eval_metric in datasets.list_metrics()
), f"""{args.eval_metric} not in the list of supported metrics {datasets.list_metrics()}."""
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed )
logger.info("Creating the initial data directory for self-training..." )
snake_case : List[Any] = f"""{args.output_dir}/self-train_iter-{{}}""".format
snake_case : Optional[int] = data_dir_format(0 )
if accelerator.is_main_process:
if args.output_dir is not None:
os.makedirs(args.output_dir , exist_ok=__lowerCamelCase )
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : Dict = None
snake_case : Union[str, Any] = None
snake_case : Tuple = 0
snake_case : List[Any] = False
# Show the progress bar
snake_case : List[Any] = tqdm(range(args.max_selftrain_iterations ) , disable=not accelerator.is_local_main_process )
# Self-train
for iteration in range(0 , int(args.max_selftrain_iterations ) ):
snake_case : str = data_dir_format(__lowerCamelCase )
assert os.path.exists(__lowerCamelCase )
# Stage 1: initial fine-tuning for iteration = 0 or pseudo-training for
# iteration > 0
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-1" )
snake_case : Optional[Any] = {
"accelerator": accelerator,
"model_name_or_path": args.model_name_or_path,
"cache_dir": args.cache_dir,
"do_train": True,
"train_file": data_files["train"] if iteration == 0 else data_files["train_pseudo"],
"do_eval": True if args.eval_file is not None else False,
"eval_file": data_files["eval"],
"do_predict": True,
"infer_file": data_files["infer"],
"task_name": args.task_name,
"label_list": args.label_list,
"output_dir": current_output_dir,
"eval_metric": args.eval_metric,
"evaluation_strategy": args.evaluation_strategy,
"early_stopping_patience": args.early_stopping_patience,
"early_stopping_threshold": args.early_stopping_threshold,
"seed": args.seed,
}
# Add additional training arguments
for key, value in kwargs.items():
if key not in arguments_dict and not hasattr(__lowerCamelCase , __lowerCamelCase ):
arguments_dict.update({key: value} )
snake_case : int = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 1." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 1 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 1." , __lowerCamelCase )
if iteration > 0 and args.finetune_on_labeled_data:
# Stage 2 (optional): fine-tuning on the original labeled data
snake_case : str = os.path.join(__lowerCamelCase , "best-checkpoint" )
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-2" )
# Update arguments_dict
snake_case : List[str] = model_path
snake_case : Optional[Any] = data_files["train"]
snake_case : Optional[Any] = current_output_dir
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 2." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 2 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 2." , __lowerCamelCase )
snake_case : int = iteration
snake_case : Tuple = data_dir_format(iteration + 1 )
snake_case : Tuple = AutoConfig.from_pretrained(os.path.join(__lowerCamelCase , "best-checkpoint" ) )
snake_case : Optional[int] = config.idalabel
snake_case : List[Any] = os.path.join(__lowerCamelCase , "eval_results_best-checkpoint.json" )
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "test_results_best-checkpoint.json" )
assert os.path.exists(__lowerCamelCase )
with open(__lowerCamelCase , "r" ) as f:
snake_case : Dict = float(json.load(__lowerCamelCase )[args.eval_metric] )
snake_case : Optional[int] = os.path.join(__lowerCamelCase , "infer_output_best-checkpoint.csv" )
assert os.path.exists(__lowerCamelCase )
# Loading the dataset from local csv or json files.
snake_case : Optional[Any] = load_dataset(args.data_file_extension , data_files={"data": data_files["infer"]} )["data"]
snake_case : Dict = load_dataset("csv" , data_files={"data": infer_output_file} )["data"]
if accelerator.is_main_process:
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) )
if os.path.exists(__lowerCamelCase ):
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""test_results_iter-{iteration}.json""" ) )
create_pseudo_labeled_data(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : str = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.evaluation_strategy != IntervalStrategy.NO.value:
snake_case : List[Any] = eval_result
if best_iteration is None:
snake_case : List[Any] = new_iteration
snake_case : int = new_eval_result
else:
if new_eval_result - best_eval_result > args.early_stopping_threshold:
snake_case : int = new_iteration
snake_case : Union[str, Any] = new_eval_result
snake_case : str = 0
else:
if new_eval_result == best_eval_result:
snake_case : Any = new_iteration
snake_case : Union[str, Any] = new_eval_result
early_stopping_patience_counter += 1
if early_stopping_patience_counter >= args.early_stopping_patience:
snake_case : Tuple = True
progress_bar.update(1 )
if should_training_stop:
break
if best_iteration is not None:
# Save the best iteration
logger.info("Best iteration: %d" , __lowerCamelCase )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
else:
# Assume that the last iteration is the best
logger.info("Best iteration: %d" , args.max_selftrain_iterations - 1 )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{args.max_selftrain_iterations - 1}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
| 59
| 1
|
from collections import OrderedDict
from typing import Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""microsoft/swin-tiny-patch4-window7-224""": (
"""https://huggingface.co/microsoft/swin-tiny-patch4-window7-224/resolve/main/config.json"""
),
# See all Swin models at https://huggingface.co/models?filter=swin
}
class UpperCAmelCase ( A_ ,A_ ):
A__ : Optional[Any] = "swin"
A__ : str = {
"num_attention_heads": "num_heads",
"num_hidden_layers": "num_layers",
}
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=2_24 , snake_case__ : List[str]=4 , snake_case__ : Union[str, Any]=3 , snake_case__ : Tuple=96 , snake_case__ : List[Any]=[2, 2, 6, 2] , snake_case__ : Dict=[3, 6, 12, 24] , snake_case__ : Optional[Any]=7 , snake_case__ : int=4.0 , snake_case__ : Tuple=True , snake_case__ : List[str]=0.0 , snake_case__ : Optional[int]=0.0 , snake_case__ : Tuple=0.1 , snake_case__ : Dict="gelu" , snake_case__ : int=False , snake_case__ : Dict=0.02 , snake_case__ : List[str]=1e-5 , snake_case__ : Union[str, Any]=32 , snake_case__ : Any=None , snake_case__ : Tuple=None , **snake_case__ : List[str] , ) -> Tuple:
'''simple docstring'''
super().__init__(**snake_case__ )
snake_case : Union[str, Any] = image_size
snake_case : List[Any] = patch_size
snake_case : str = num_channels
snake_case : List[Any] = embed_dim
snake_case : Dict = depths
snake_case : Optional[Any] = len(snake_case__ )
snake_case : Optional[Any] = num_heads
snake_case : Any = window_size
snake_case : str = mlp_ratio
snake_case : Optional[int] = qkv_bias
snake_case : Union[str, Any] = hidden_dropout_prob
snake_case : Optional[Any] = attention_probs_dropout_prob
snake_case : Optional[int] = drop_path_rate
snake_case : Optional[Any] = hidden_act
snake_case : Optional[int] = use_absolute_embeddings
snake_case : List[Any] = layer_norm_eps
snake_case : Any = initializer_range
snake_case : Optional[Any] = encoder_stride
# we set the hidden_size attribute in order to make Swin work with VisionEncoderDecoderModel
# this indicates the channel dimension after the last stage of the model
snake_case : Any = int(embed_dim * 2 ** (len(snake_case__ ) - 1) )
snake_case : Union[str, Any] = ["stem"] + [f"""stage{idx}""" for idx in range(1 , len(snake_case__ ) + 1 )]
snake_case , snake_case : Optional[Any] = get_aligned_output_features_output_indices(
out_features=snake_case__ , out_indices=snake_case__ , stage_names=self.stage_names )
class UpperCAmelCase ( A_ ):
A__ : Dict = version.parse("1.11" )
@property
def _SCREAMING_SNAKE_CASE (self : Any ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
return OrderedDict(
[
("pixel_values", {0: "batch", 1: "num_channels", 2: "height", 3: "width"}),
] )
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> float:
'''simple docstring'''
return 1e-4
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__lowerCamelCase = {"""configuration_xglm""": ["""XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP""", """XGLMConfig"""]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizer"""]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizerFast"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""XGLMForCausalLM""",
"""XGLMModel""",
"""XGLMPreTrainedModel""",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""FlaxXGLMForCausalLM""",
"""FlaxXGLMModel""",
"""FlaxXGLMPreTrainedModel""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""TFXGLMForCausalLM""",
"""TFXGLMModel""",
"""TFXGLMPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm import XGLMTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm_fast import XGLMTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_xglm import (
TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST,
TFXGLMForCausalLM,
TFXGLMModel,
TFXGLMPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure)
| 59
| 1
|
from typing import Optional, Union
import torch
from torch import nn
from ...configuration_utils import ConfigMixin, register_to_config
from ...models.modeling_utils import ModelMixin
class UpperCAmelCase ( A_ ,A_ ):
@register_to_config
def __init__(self : int , snake_case__ : int = 7_68 , ) -> Dict:
'''simple docstring'''
super().__init__()
snake_case : int = nn.Parameter(torch.zeros(1 , snake_case__ ) )
snake_case : Any = nn.Parameter(torch.ones(1 , snake_case__ ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Union[str, torch.device]] = None , snake_case__ : Optional[torch.dtype] = None , ) -> int:
'''simple docstring'''
snake_case : Dict = nn.Parameter(self.mean.to(snake_case__ ).to(snake_case__ ) )
snake_case : List[str] = nn.Parameter(self.std.to(snake_case__ ).to(snake_case__ ) )
return self
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : str = (embeds - self.mean) * 1.0 / self.std
return embeds
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : int ) -> Dict:
'''simple docstring'''
snake_case : List[str] = (embeds * self.std) + self.mean
return embeds
| 59
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
# See all MEGATRON_BERT models at https://huggingface.co/models?filter=bert
}
class UpperCAmelCase ( A_ ):
A__ : List[str] = "megatron-bert"
def __init__(self : Optional[int] , snake_case__ : List[str]=2_90_56 , snake_case__ : List[Any]=10_24 , snake_case__ : str=24 , snake_case__ : Tuple=16 , snake_case__ : Union[str, Any]=40_96 , snake_case__ : str="gelu" , snake_case__ : str=0.1 , snake_case__ : Optional[int]=0.1 , snake_case__ : Tuple=5_12 , snake_case__ : Union[str, Any]=2 , snake_case__ : Dict=0.02 , snake_case__ : List[Any]=1e-12 , snake_case__ : int=0 , snake_case__ : Tuple="absolute" , snake_case__ : Any=True , **snake_case__ : Union[str, Any] , ) -> Optional[Any]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , **snake_case__ )
snake_case : Tuple = vocab_size
snake_case : str = hidden_size
snake_case : str = num_hidden_layers
snake_case : str = num_attention_heads
snake_case : Optional[int] = hidden_act
snake_case : int = intermediate_size
snake_case : List[str] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : List[str] = type_vocab_size
snake_case : List[str] = initializer_range
snake_case : Tuple = layer_norm_eps
snake_case : int = position_embedding_type
snake_case : str = use_cache
| 59
| 1
|
import itertools
import json
import os
import unittest
from transformers import AddedToken, LongformerTokenizer, LongformerTokenizerFast
from transformers.models.longformer.tokenization_longformer import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers, slow
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class UpperCAmelCase ( A_ ,unittest.TestCase ):
A__ : Optional[Any] = LongformerTokenizer
A__ : str = True
A__ : Any = LongformerTokenizerFast
A__ : List[str] = True
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> str:
'''simple docstring'''
super().setUp()
# Adapted from Sennrich et al. 2015 and https://github.com/rsennrich/subword-nmt
snake_case : Union[str, Any] = [
"l",
"o",
"w",
"e",
"r",
"s",
"t",
"i",
"d",
"n",
"\u0120",
"\u0120l",
"\u0120n",
"\u0120lo",
"\u0120low",
"er",
"\u0120lowest",
"\u0120newer",
"\u0120wider",
"<unk>",
]
snake_case : List[str] = dict(zip(snake_case__ , range(len(snake_case__ ) ) ) )
snake_case : List[Any] = ["#version: 0.2", "\u0120 l", "\u0120l o", "\u0120lo w", "e r", ""]
snake_case : Any = {"unk_token": "<unk>"}
snake_case : Optional[int] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["vocab_file"] )
snake_case : 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(snake_case__ ) + "\n" )
with open(self.merges_file , "w" , encoding="utf-8" ) as fp:
fp.write("\n".join(snake_case__ ) )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , **snake_case__ : Optional[Any] ) -> Union[str, Any]:
'''simple docstring'''
kwargs.update(self.special_tokens_map )
return self.tokenizer_class.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str , **snake_case__ : Union[str, Any] ) -> str:
'''simple docstring'''
kwargs.update(self.special_tokens_map )
return self.rust_tokenizer_class.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : str ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Tuple = "lower newer"
snake_case : str = "lower newer"
return input_text, output_text
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case : int = self.tokenizer_class(self.vocab_file , self.merges_file , **self.special_tokens_map )
snake_case : Tuple = "lower newer"
snake_case : List[str] = ["l", "o", "w", "er", "\u0120", "n", "e", "w", "er"]
snake_case : Optional[Any] = tokenizer.tokenize(snake_case__ ) # , add_prefix_space=True)
self.assertListEqual(snake_case__ , snake_case__ )
snake_case : int = tokens + [tokenizer.unk_token]
snake_case : int = [0, 1, 2, 15, 10, 9, 3, 2, 15, 19]
self.assertListEqual(tokenizer.convert_tokens_to_ids(snake_case__ ) , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Any:
'''simple docstring'''
snake_case : str = self.get_tokenizer()
self.assertListEqual(tokenizer.encode("Hello world!" , add_special_tokens=snake_case__ ) , [0, 3_14_14, 2_32, 3_28, 2] )
self.assertListEqual(
tokenizer.encode("Hello world! cécé herlolip 418" , add_special_tokens=snake_case__ ) , [0, 3_14_14, 2_32, 3_28, 7_40, 11_40, 1_26_95, 69, 4_60_78, 15_88, 2] , )
@slow
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = self.tokenizer_class.from_pretrained("allenai/longformer-base-4096" )
snake_case : Any = tokenizer.encode("sequence builders" , add_special_tokens=snake_case__ )
snake_case : Optional[int] = tokenizer.encode("multi-sequence build" , add_special_tokens=snake_case__ )
snake_case : int = tokenizer.encode(
"sequence builders" , add_special_tokens=snake_case__ , add_prefix_space=snake_case__ )
snake_case : int = tokenizer.encode(
"sequence builders" , "multi-sequence build" , add_special_tokens=snake_case__ , add_prefix_space=snake_case__ )
snake_case : int = tokenizer.build_inputs_with_special_tokens(snake_case__ )
snake_case : int = tokenizer.build_inputs_with_special_tokens(snake_case__ , snake_case__ )
assert encoded_sentence == encoded_text_from_decode
assert encoded_pair == encoded_pair_from_decode
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Dict = self.get_tokenizer()
snake_case : str = "Encode this sequence."
snake_case : Optional[Any] = tokenizer.byte_encoder[" ".encode("utf-8" )[0]]
# Testing encoder arguments
snake_case : Optional[Any] = tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ , add_prefix_space=snake_case__ )
snake_case : int = tokenizer.convert_ids_to_tokens(encoded[0] )[0]
self.assertNotEqual(snake_case__ , snake_case__ )
snake_case : int = tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ , add_prefix_space=snake_case__ )
snake_case : List[Any] = tokenizer.convert_ids_to_tokens(encoded[0] )[0]
self.assertEqual(snake_case__ , snake_case__ )
tokenizer.add_special_tokens({"bos_token": "<s>"} )
snake_case : int = tokenizer.encode(snake_case__ , add_special_tokens=snake_case__ )
snake_case : Dict = tokenizer.convert_ids_to_tokens(encoded[1] )[0]
self.assertNotEqual(snake_case__ , snake_case__ )
# Testing spaces after special tokens
snake_case : Union[str, Any] = "<mask>"
tokenizer.add_special_tokens(
{"mask_token": AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ )} ) # mask token has a left space
snake_case : List[str] = tokenizer.convert_tokens_to_ids(snake_case__ )
snake_case : Tuple = "Encode <mask> sequence"
snake_case : Optional[Any] = "Encode <mask>sequence"
snake_case : Dict = tokenizer.encode(snake_case__ )
snake_case : Optional[Any] = encoded.index(snake_case__ )
snake_case : int = tokenizer.convert_ids_to_tokens(encoded[mask_loc + 1] )[0]
self.assertEqual(snake_case__ , snake_case__ )
snake_case : Dict = tokenizer.encode(snake_case__ )
snake_case : Union[str, Any] = encoded.index(snake_case__ )
snake_case : Union[str, Any] = tokenizer.convert_ids_to_tokens(encoded[mask_loc + 1] )[0]
self.assertNotEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> List[Any]:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(f"""{tokenizer.__class__.__name__} ({pretrained_name})""" ):
snake_case : Optional[int] = self.rust_tokenizer_class.from_pretrained(snake_case__ , **snake_case__ )
snake_case : List[Any] = self.tokenizer_class.from_pretrained(snake_case__ , **snake_case__ )
snake_case : Union[str, Any] = "A, <mask> AllenNLP sentence."
snake_case : Tuple = tokenizer_r.encode_plus(snake_case__ , add_special_tokens=snake_case__ , return_token_type_ids=snake_case__ )
snake_case : Optional[int] = tokenizer_p.encode_plus(snake_case__ , add_special_tokens=snake_case__ , return_token_type_ids=snake_case__ )
# token_type_ids should put 0 everywhere
self.assertEqual(sum(tokens_r["token_type_ids"] ) , sum(tokens_p["token_type_ids"] ) )
# attention_mask should put 1 everywhere, so sum over length should be 1
self.assertEqual(
sum(tokens_r["attention_mask"] ) / len(tokens_r["attention_mask"] ) , sum(tokens_p["attention_mask"] ) / len(tokens_p["attention_mask"] ) , )
snake_case : List[Any] = tokenizer_r.convert_ids_to_tokens(tokens_r["input_ids"] )
snake_case : int = tokenizer_p.convert_ids_to_tokens(tokens_p["input_ids"] )
# Rust correctly handles the space before the mask while python doesnt
self.assertSequenceEqual(tokens_p["input_ids"] , [0, 2_50, 6, 5_02_64, 38_23, 4_87, 2_19_92, 36_45, 4, 2] )
self.assertSequenceEqual(tokens_r["input_ids"] , [0, 2_50, 6, 5_02_64, 38_23, 4_87, 2_19_92, 36_45, 4, 2] )
self.assertSequenceEqual(
snake_case__ , ["<s>", "A", ",", "<mask>", "ĠAllen", "N", "LP", "Ġsentence", ".", "</s>"] )
self.assertSequenceEqual(
snake_case__ , ["<s>", "A", ",", "<mask>", "ĠAllen", "N", "LP", "Ġsentence", ".", "</s>"] )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
for trim_offsets, add_prefix_space in itertools.product([True, False] , repeat=2 ):
snake_case : Any = self.rust_tokenizer_class.from_pretrained(
self.tmpdirname , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : Dict = json.loads(tokenizer_r.backend_tokenizer.pre_tokenizer.__getstate__() )
snake_case : str = json.loads(tokenizer_r.backend_tokenizer.post_processor.__getstate__() )
self.assertEqual(pre_tokenizer_state["add_prefix_space"] , snake_case__ )
self.assertEqual(post_processor_state["add_prefix_space"] , snake_case__ )
self.assertEqual(post_processor_state["trim_offsets"] , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> int:
'''simple docstring'''
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(f"""{tokenizer.__class__.__name__} ({pretrained_name})""" ):
snake_case : Union[str, Any] = "hello" # `hello` is a token in the vocabulary of `pretrained_name`
snake_case : List[Any] = f"""{text_of_1_token} {text_of_1_token}"""
snake_case : int = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : str = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (0, len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (len(snake_case__ ) + 1, len(snake_case__ ) + 1 + len(snake_case__ )) , )
snake_case : Any = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : int = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (0, len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (len(snake_case__ ) + 1, len(snake_case__ ) + 1 + len(snake_case__ )) , )
snake_case : int = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : Union[str, Any] = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (0, len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (len(snake_case__ ), len(snake_case__ ) + 1 + len(snake_case__ )) , )
snake_case : Union[str, Any] = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : List[Any] = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (0, len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (len(snake_case__ ), len(snake_case__ ) + 1 + len(snake_case__ )) , )
snake_case : Any = f""" {text}"""
# tokenizer_r = self.rust_tokenizer_class.from_pretrained(
# pretrained_name, use_fast=True, add_prefix_space=True, trim_offsets=True
# )
# encoding = tokenizer_r(text, return_offsets_mapping=True, add_special_tokens=False)
# self.assertEqual(encoding.offset_mapping[0], (1, 1 + len(text_of_1_token)))
# self.assertEqual(
# encoding.offset_mapping[1],
# (1 + len(text_of_1_token) + 1, 1 + len(text_of_1_token) + 1 + len(text_of_1_token)),
# )
snake_case : List[str] = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : Optional[int] = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (1, 1 + len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (1 + len(snake_case__ ) + 1, 1 + len(snake_case__ ) + 1 + len(snake_case__ )) , )
snake_case : Optional[Any] = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : List[Any] = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (0, 1 + len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (1 + len(snake_case__ ), 1 + len(snake_case__ ) + 1 + len(snake_case__ )) , )
snake_case : Optional[int] = self.rust_tokenizer_class.from_pretrained(
snake_case__ , use_fast=snake_case__ , add_prefix_space=snake_case__ , trim_offsets=snake_case__ )
snake_case : Optional[Any] = tokenizer_r(snake_case__ , return_offsets_mapping=snake_case__ , add_special_tokens=snake_case__ )
self.assertEqual(encoding.offset_mapping[0] , (0, 1 + len(snake_case__ )) )
self.assertEqual(
encoding.offset_mapping[1] , (1 + len(snake_case__ ), 1 + len(snake_case__ ) + 1 + len(snake_case__ )) , )
| 59
|
import gc
import unittest
from parameterized import parameterized
from diffusers import FlaxUNetaDConditionModel
from diffusers.utils import is_flax_available
from diffusers.utils.testing_utils import load_hf_numpy, require_flax, slow
if is_flax_available():
import jax
import jax.numpy as jnp
@slow
@require_flax
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, Any] , snake_case__ : List[str] ) -> List[str]:
'''simple docstring'''
return f"""gaussian_noise_s={seed}_shape={'_'.join([str(snake_case__ ) for s in shape] )}.npy"""
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> int:
'''simple docstring'''
super().tearDown()
gc.collect()
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Optional[Any]=0 , snake_case__ : Any=(4, 4, 64, 64) , snake_case__ : List[Any]=False ) -> int:
'''simple docstring'''
snake_case : Optional[Any] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Optional[int] = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return image
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Tuple=False , snake_case__ : List[Any]="CompVis/stable-diffusion-v1-4" ) -> List[Any]:
'''simple docstring'''
snake_case : List[str] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : str = "bf16" if fpaa else None
snake_case , snake_case : Optional[int] = FlaxUNetaDConditionModel.from_pretrained(
snake_case__ , subfolder="unet" , dtype=snake_case__ , revision=snake_case__ )
return model, params
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any]=0 , snake_case__ : Union[str, Any]=(4, 77, 7_68) , snake_case__ : Dict=False ) -> List[str]:
'''simple docstring'''
snake_case : Any = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Any = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return hidden_states
@parameterized.expand(
[
# fmt: off
[83, 4, [-0.2323, -0.1304, 0.0813, -0.3093, -0.0919, -0.1571, -0.1125, -0.5806]],
[17, 0.55, [-0.0831, -0.2443, 0.0901, -0.0919, 0.3396, 0.0103, -0.3743, 0.0701]],
[8, 0.89, [-0.4863, 0.0859, 0.0875, -0.1658, 0.9199, -0.0114, 0.4839, 0.4639]],
[3, 10_00, [-0.5649, 0.2402, -0.5518, 0.1248, 1.1328, -0.2443, -0.0325, -1.0078]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : int , snake_case__ : Optional[int] , snake_case__ : Dict ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : List[str] = self.get_unet_model(model_id="CompVis/stable-diffusion-v1-4" , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_latents(snake_case__ , fpaa=snake_case__ )
snake_case : List[str] = self.get_encoder_hidden_states(snake_case__ , fpaa=snake_case__ )
snake_case : Dict = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : Optional[Any] = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Optional[int] = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, in the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[83, 4, [0.1514, 0.0807, 0.1624, 0.1016, -0.1896, 0.0263, 0.0677, 0.2310]],
[17, 0.55, [0.1164, -0.0216, 0.0170, 0.1589, -0.3120, 0.1005, -0.0581, -0.1458]],
[8, 0.89, [-0.1758, -0.0169, 0.1004, -0.1411, 0.1312, 0.1103, -0.1996, 0.2139]],
[3, 10_00, [0.1214, 0.0352, -0.0731, -0.1562, -0.0994, -0.0906, -0.2340, -0.0539]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[int] , snake_case__ : str , snake_case__ : Tuple ) -> str:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.get_unet_model(model_id="stabilityai/stable-diffusion-2" , fpaa=snake_case__ )
snake_case : List[str] = self.get_latents(snake_case__ , shape=(4, 4, 96, 96) , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_encoder_hidden_states(snake_case__ , shape=(4, 77, 10_24) , fpaa=snake_case__ )
snake_case : Optional[int] = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : int = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Dict = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, on the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
| 59
| 1
|
import argparse
from collections import defaultdict
import yaml
__lowerCamelCase = """docs/source/en/_toctree.yml"""
def UpperCamelCase ( __lowerCamelCase : Optional[Any] ):
snake_case : Any = defaultdict(__lowerCamelCase )
snake_case : List[Any] = []
snake_case : List[str] = []
for doc in doc_list:
if "local" in doc:
counts[doc["local"]] += 1
if doc["title"].lower() == "overview":
overview_doc.append({"local": doc["local"], "title": doc["title"]} )
else:
new_doc_list.append(__lowerCamelCase )
snake_case : List[str] = new_doc_list
snake_case : Optional[int] = [key for key, value in counts.items() if value > 1]
snake_case : Any = []
for duplicate_key in duplicates:
snake_case : List[str] = list({doc["title"] for doc in doc_list if doc["local"] == duplicate_key} )
if len(__lowerCamelCase ) > 1:
raise ValueError(
f"""{duplicate_key} is present several times in the documentation table of content at """
"`docs/source/en/_toctree.yml` with different *Title* values. Choose one of those and remove the "
"others." )
# Only add this once
new_doc.append({"local": duplicate_key, "title": titles[0]} )
# Add none duplicate-keys
new_doc.extend([doc for doc in doc_list if "local" not in counts or counts[doc["local"]] == 1] )
snake_case : Union[str, Any] = sorted(__lowerCamelCase , key=lambda __lowerCamelCase : s["title"].lower() )
# "overview" gets special treatment and is always first
if len(__lowerCamelCase ) > 1:
raise ValueError("{doc_list} has two 'overview' docs which is not allowed." )
overview_doc.extend(__lowerCamelCase )
# Sort
return overview_doc
def UpperCamelCase ( __lowerCamelCase : Any=False ):
with open(__lowerCamelCase , encoding="utf-8" ) as f:
snake_case : Optional[Any] = yaml.safe_load(f.read() )
# Get to the API doc
snake_case : List[str] = 0
while content[api_idx]["title"] != "API":
api_idx += 1
snake_case : str = content[api_idx]["sections"]
# Then to the model doc
snake_case : Optional[int] = 0
while api_doc[scheduler_idx]["title"] != "Schedulers":
scheduler_idx += 1
snake_case : Dict = api_doc[scheduler_idx]["sections"]
snake_case : Any = clean_doc_toc(__lowerCamelCase )
snake_case : str = False
if new_scheduler_doc != scheduler_doc:
snake_case : Dict = True
if overwrite:
snake_case : int = new_scheduler_doc
if diff:
if overwrite:
snake_case : Dict = api_doc
with open(__lowerCamelCase , "w" , encoding="utf-8" ) as f:
f.write(yaml.dump(__lowerCamelCase , allow_unicode=__lowerCamelCase ) )
else:
raise ValueError(
"The model doc part of the table of content is not properly sorted, run `make style` to fix this." )
def UpperCamelCase ( __lowerCamelCase : int=False ):
with open(__lowerCamelCase , encoding="utf-8" ) as f:
snake_case : str = yaml.safe_load(f.read() )
# Get to the API doc
snake_case : str = 0
while content[api_idx]["title"] != "API":
api_idx += 1
snake_case : List[Any] = content[api_idx]["sections"]
# Then to the model doc
snake_case : Tuple = 0
while api_doc[pipeline_idx]["title"] != "Pipelines":
pipeline_idx += 1
snake_case : int = False
snake_case : Optional[int] = api_doc[pipeline_idx]["sections"]
snake_case : Any = []
# sort sub pipeline docs
for pipeline_doc in pipeline_docs:
if "section" in pipeline_doc:
snake_case : str = pipeline_doc["section"]
snake_case : Optional[int] = clean_doc_toc(__lowerCamelCase )
if overwrite:
snake_case : Any = new_sub_pipeline_doc
new_pipeline_docs.append(__lowerCamelCase )
# sort overall pipeline doc
snake_case : Tuple = clean_doc_toc(__lowerCamelCase )
if new_pipeline_docs != pipeline_docs:
snake_case : Optional[int] = True
if overwrite:
snake_case : Optional[int] = new_pipeline_docs
if diff:
if overwrite:
snake_case : int = api_doc
with open(__lowerCamelCase , "w" , encoding="utf-8" ) as f:
f.write(yaml.dump(__lowerCamelCase , allow_unicode=__lowerCamelCase ) )
else:
raise ValueError(
"The model doc part of the table of content is not properly sorted, run `make style` to fix this." )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument("""--fix_and_overwrite""", action="""store_true""", help="""Whether to fix inconsistencies.""")
__lowerCamelCase = parser.parse_args()
check_scheduler_doc(args.fix_and_overwrite)
check_pipeline_doc(args.fix_and_overwrite)
| 59
|
import argparse
import re
from typing import Dict
import torch
from datasets import Audio, Dataset, load_dataset, load_metric
from transformers import AutoFeatureExtractor, pipeline
def UpperCamelCase ( __lowerCamelCase : Dataset , __lowerCamelCase : Dict[str, str] ):
snake_case : int = args.log_outputs
snake_case : Dict = "_".join(args.dataset.split("/" ) + [args.config, args.split] )
# load metric
snake_case : List[str] = load_metric("wer" )
snake_case : Tuple = load_metric("cer" )
# compute metrics
snake_case : List[Any] = wer.compute(references=result["target"] , predictions=result["prediction"] )
snake_case : int = cer.compute(references=result["target"] , predictions=result["prediction"] )
# print & log results
snake_case : int = f"""WER: {wer_result}\nCER: {cer_result}"""
print(__lowerCamelCase )
with open(f"""{dataset_id}_eval_results.txt""" , "w" ) as f:
f.write(__lowerCamelCase )
# log all results in text file. Possibly interesting for analysis
if log_outputs is not None:
snake_case : int = f"""log_{dataset_id}_predictions.txt"""
snake_case : List[Any] = f"""log_{dataset_id}_targets.txt"""
with open(__lowerCamelCase , "w" ) as p, open(__lowerCamelCase , "w" ) as t:
# mapping function to write output
def write_to_file(__lowerCamelCase : str , __lowerCamelCase : Optional[int] ):
p.write(f"""{i}""" + "\n" )
p.write(batch["prediction"] + "\n" )
t.write(f"""{i}""" + "\n" )
t.write(batch["target"] + "\n" )
result.map(__lowerCamelCase , with_indices=__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : List[Any] = "[,?.!\-\;\:\"“%‘”�—’…–]" # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
snake_case : List[Any] = re.sub(__lowerCamelCase , "" , text.lower() )
# In addition, we can normalize the target text, e.g. removing new lines characters etc...
# note that order is important here!
snake_case : Optional[Any] = ["\n\n", "\n", " ", " "]
for t in token_sequences_to_ignore:
snake_case : Dict = " ".join(text.split(__lowerCamelCase ) )
return text
def UpperCamelCase ( __lowerCamelCase : int ):
# load dataset
snake_case : str = load_dataset(args.dataset , args.config , split=args.split , use_auth_token=__lowerCamelCase )
# for testing: only process the first two examples as a test
# dataset = dataset.select(range(10))
# load processor
snake_case : List[Any] = AutoFeatureExtractor.from_pretrained(args.model_id )
snake_case : Union[str, Any] = feature_extractor.sampling_rate
# resample audio
snake_case : Union[str, Any] = dataset.cast_column("audio" , Audio(sampling_rate=__lowerCamelCase ) )
# load eval pipeline
if args.device is None:
snake_case : List[str] = 0 if torch.cuda.is_available() else -1
snake_case : str = pipeline("automatic-speech-recognition" , model=args.model_id , device=args.device )
# map function to decode audio
def map_to_pred(__lowerCamelCase : int ):
snake_case : Dict = asr(
batch["audio"]["array"] , chunk_length_s=args.chunk_length_s , stride_length_s=args.stride_length_s )
snake_case : str = prediction["text"]
snake_case : Tuple = normalize_text(batch["sentence"] )
return batch
# run inference on all examples
snake_case : Dict = dataset.map(__lowerCamelCase , remove_columns=dataset.column_names )
# compute and log_results
# do not change function below
log_results(__lowerCamelCase , __lowerCamelCase )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
"""--model_id""", type=str, required=True, help="""Model identifier. Should be loadable with 🤗 Transformers"""
)
parser.add_argument(
"""--dataset""",
type=str,
required=True,
help="""Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets""",
)
parser.add_argument(
"""--config""", type=str, required=True, help="""Config of the dataset. *E.g.* `'en'` for Common Voice"""
)
parser.add_argument("""--split""", type=str, required=True, help="""Split of the dataset. *E.g.* `'test'`""")
parser.add_argument(
"""--chunk_length_s""", type=float, default=None, help="""Chunk length in seconds. Defaults to 5 seconds."""
)
parser.add_argument(
"""--stride_length_s""", type=float, default=None, help="""Stride of the audio chunks. Defaults to 1 second."""
)
parser.add_argument(
"""--log_outputs""", action="""store_true""", help="""If defined, write outputs to log file for analysis."""
)
parser.add_argument(
"""--device""",
type=int,
default=None,
help="""The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.""",
)
__lowerCamelCase = parser.parse_args()
main(args)
| 59
| 1
|
import os
import numpy
import onnx
def UpperCamelCase ( __lowerCamelCase : List[str] , __lowerCamelCase : str ):
snake_case : Union[str, Any] = a.name
snake_case : Optional[Any] = b.name
snake_case : Any = ""
snake_case : str = ""
snake_case : str = a == b
snake_case : Optional[int] = name_a
snake_case : int = name_b
return res
def UpperCamelCase ( __lowerCamelCase : Tuple , __lowerCamelCase : Optional[Any] , __lowerCamelCase : List[str] ):
for i, input_name in enumerate(node_proto.input ):
if input_name == name:
node_proto.input.insert(__lowerCamelCase , __lowerCamelCase )
node_proto.input.pop(i + 1 )
if node_proto.op_type == "If":
_graph_replace_input_with(node_proto.attribute[0].g , __lowerCamelCase , __lowerCamelCase )
_graph_replace_input_with(node_proto.attribute[1].g , __lowerCamelCase , __lowerCamelCase )
if node_proto.op_type == "Loop":
_graph_replace_input_with(node_proto.attribute[0].g , __lowerCamelCase , __lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : List[str] , __lowerCamelCase : Optional[int] , __lowerCamelCase : List[Any] ):
for n in graph_proto.node:
_node_replace_input_with(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : List[Any] ):
snake_case : Optional[int] = list(model.graph.initializer )
snake_case : Optional[Any] = list(model_without_ext.graph.initializer )
for i, ref_i in ind_to_replace:
assert inits_with_data[i].name == inits[i].name
assert inits_with_data[ref_i].name == inits[ref_i].name
assert i > ref_i
snake_case : Union[str, Any] = inits[i].name
snake_case : Any = inits[ref_i].name
model_without_ext.graph.initializer.remove(inits[i] )
# for n in model.graph.node:
_graph_replace_input_with(model_without_ext.graph , __lowerCamelCase , __lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
snake_case : Union[str, Any] = os.path.dirname(__lowerCamelCase )
snake_case : Union[str, Any] = os.path.basename(__lowerCamelCase )
snake_case : Union[str, Any] = onnx.load(os.path.join(__lowerCamelCase , __lowerCamelCase ) )
snake_case : Optional[Any] = list(model.graph.initializer )
snake_case : Any = set()
snake_case : Optional[Any] = {}
snake_case : Tuple = []
snake_case : str = 0
for i in range(len(__lowerCamelCase ) ):
if i in dup_set:
continue
for j in range(i + 1 , len(__lowerCamelCase ) ):
if j in dup_set:
continue
if _is_equal_tensor_proto(inits[i] , inits[j] ):
dup_set.add(__lowerCamelCase )
dup_set.add(__lowerCamelCase )
snake_case : List[Any] = inits[j].data_type
snake_case : str = numpy.prod(inits[j].dims )
if dtype == 1:
mem_size *= 4
elif dtype == 6:
mem_size *= 4
elif dtype == 7 or dtype == 11:
mem_size *= 8
else:
print("unexpected data type: " , __lowerCamelCase )
total_reduced_size += mem_size
snake_case : Optional[int] = inits[i].name
snake_case : Optional[int] = inits[j].name
if name_i in dup_map:
dup_map[name_i].append(__lowerCamelCase )
else:
snake_case : Any = [name_j]
ind_to_replace.append((j, i) )
print("total reduced size: " , total_reduced_size / 1024 / 1024 / 1024 , "GB" )
snake_case : Tuple = sorted(__lowerCamelCase )
_remove_dup_initializers_from_model(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
snake_case : int = "optimized_" + model_file_name
snake_case : str = os.path.join(__lowerCamelCase , __lowerCamelCase )
onnx.save(__lowerCamelCase , __lowerCamelCase )
return new_model
| 59
|
from typing import Optional, Tuple, Union
import flax
import flax.linen as nn
import jax
import jax.numpy as jnp
from flax.core.frozen_dict import FrozenDict
from ..configuration_utils import ConfigMixin, flax_register_to_config
from ..utils import BaseOutput
from .embeddings_flax import FlaxTimestepEmbedding, FlaxTimesteps
from .modeling_flax_utils import FlaxModelMixin
from .unet_ad_blocks_flax import (
FlaxCrossAttnDownBlockaD,
FlaxCrossAttnUpBlockaD,
FlaxDownBlockaD,
FlaxUNetMidBlockaDCrossAttn,
FlaxUpBlockaD,
)
@flax.struct.dataclass
class UpperCAmelCase ( A_ ):
A__ : jnp.ndarray
@flax_register_to_config
class UpperCAmelCase ( nn.Module ,A_ ,A_ ):
A__ : int = 32
A__ : int = 4
A__ : int = 4
A__ : Tuple[str] = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
A__ : Tuple[str] = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
A__ : Union[bool, Tuple[bool]] = False
A__ : Tuple[int] = (3_20, 6_40, 12_80, 12_80)
A__ : int = 2
A__ : Union[int, Tuple[int]] = 8
A__ : Optional[Union[int, Tuple[int]]] = None
A__ : int = 12_80
A__ : float = 0.0
A__ : bool = False
A__ : jnp.dtype = jnp.floataa
A__ : bool = True
A__ : int = 0
A__ : bool = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : jax.random.KeyArray ) -> FrozenDict:
'''simple docstring'''
snake_case : Dict = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case : Any = jnp.zeros(snake_case__ , dtype=jnp.floataa )
snake_case : List[str] = jnp.ones((1,) , dtype=jnp.intaa )
snake_case : str = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case , snake_case : Optional[int] = jax.random.split(snake_case__ )
snake_case : Union[str, Any] = {"params": params_rng, "dropout": dropout_rng}
return self.init(snake_case__ , snake_case__ , snake_case__ , snake_case__ )["params"]
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : str = self.block_out_channels
snake_case : Optional[Any] = block_out_channels[0] * 4
if self.num_attention_heads is not None:
raise ValueError(
"At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19." )
# If `num_attention_heads` is not defined (which is the case for most models)
# it will default to `attention_head_dim`. This looks weird upon first reading it and it is.
# The reason for this behavior is to correct for incorrectly named variables that were introduced
# when this library was created. The incorrect naming was only discovered much later in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131
# Changing `attention_head_dim` to `num_attention_heads` for 40,000+ configurations is too backwards breaking
# which is why we correct for the naming here.
snake_case : Tuple = self.num_attention_heads or self.attention_head_dim
# input
snake_case : Tuple = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case : Union[str, Any] = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case : Dict = FlaxTimestepEmbedding(snake_case__ , dtype=self.dtype )
snake_case : List[str] = self.only_cross_attention
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (only_cross_attention,) * len(self.down_block_types )
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case : List[Any] = []
snake_case : Optional[int] = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case : List[Any] = output_channel
snake_case : Dict = block_out_channels[i]
snake_case : Optional[Any] = i == len(snake_case__ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case : List[Any] = FlaxCrossAttnDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , num_attention_heads=num_attention_heads[i] , add_downsample=not is_final_block , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Union[str, Any] = FlaxDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , add_downsample=not is_final_block , dtype=self.dtype , )
down_blocks.append(snake_case__ )
snake_case : Dict = down_blocks
# mid
snake_case : Optional[int] = FlaxUNetMidBlockaDCrossAttn(
in_channels=block_out_channels[-1] , dropout=self.dropout , num_attention_heads=num_attention_heads[-1] , use_linear_projection=self.use_linear_projection , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
# up
snake_case : Optional[Any] = []
snake_case : Optional[int] = list(reversed(snake_case__ ) )
snake_case : Dict = list(reversed(snake_case__ ) )
snake_case : Tuple = list(reversed(snake_case__ ) )
snake_case : Optional[Any] = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case : Optional[int] = output_channel
snake_case : List[Any] = reversed_block_out_channels[i]
snake_case : Union[str, Any] = reversed_block_out_channels[min(i + 1 , len(snake_case__ ) - 1 )]
snake_case : int = i == len(snake_case__ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case : Any = FlaxCrossAttnUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , num_attention_heads=reversed_num_attention_heads[i] , add_upsample=not is_final_block , dropout=self.dropout , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Optional[int] = FlaxUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , add_upsample=not is_final_block , dropout=self.dropout , dtype=self.dtype , )
up_blocks.append(snake_case__ )
snake_case : Optional[int] = output_channel
snake_case : Tuple = up_blocks
# out
snake_case : Optional[int] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case : List[str] = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__(self : Dict , snake_case__ : Dict , snake_case__ : Dict , snake_case__ : Optional[int] , snake_case__ : Tuple=None , snake_case__ : Union[str, Any]=None , snake_case__ : bool = True , snake_case__ : bool = False , ) -> Union[FlaxUNetaDConditionOutput, Tuple]:
'''simple docstring'''
if not isinstance(snake_case__ , jnp.ndarray ):
snake_case : List[Any] = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(snake_case__ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case : Any = timesteps.astype(dtype=jnp.floataa )
snake_case : int = jnp.expand_dims(snake_case__ , 0 )
snake_case : str = self.time_proj(snake_case__ )
snake_case : str = self.time_embedding(snake_case__ )
# 2. pre-process
snake_case : int = jnp.transpose(snake_case__ , (0, 2, 3, 1) )
snake_case : List[Any] = self.conv_in(snake_case__ )
# 3. down
snake_case : Optional[int] = (sample,)
for down_block in self.down_blocks:
if isinstance(snake_case__ , snake_case__ ):
snake_case , snake_case : List[Any] = down_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
else:
snake_case , snake_case : str = down_block(snake_case__ , snake_case__ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case : Tuple = ()
for down_block_res_sample, down_block_additional_residual in zip(
snake_case__ , snake_case__ ):
down_block_res_sample += down_block_additional_residual
new_down_block_res_samples += (down_block_res_sample,)
snake_case : Optional[int] = new_down_block_res_samples
# 4. mid
snake_case : Optional[int] = self.mid_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
if mid_block_additional_residual is not None:
sample += mid_block_additional_residual
# 5. up
for up_block in self.up_blocks:
snake_case : int = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case : Optional[Any] = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(snake_case__ , snake_case__ ):
snake_case : Optional[Any] = up_block(
snake_case__ , temb=snake_case__ , encoder_hidden_states=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train , )
else:
snake_case : Dict = up_block(snake_case__ , temb=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train )
# 6. post-process
snake_case : List[str] = self.conv_norm_out(snake_case__ )
snake_case : Any = nn.silu(snake_case__ )
snake_case : Optional[int] = self.conv_out(snake_case__ )
snake_case : Union[str, Any] = jnp.transpose(snake_case__ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=snake_case__ )
| 59
| 1
|
import copy
from typing import TYPE_CHECKING, Any, Mapping, Optional, OrderedDict
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
from ..auto.configuration_auto import AutoConfig
if TYPE_CHECKING:
from ... import PreTrainedTokenizerBase, TensorType
__lowerCamelCase = logging.get_logger(__name__)
class UpperCAmelCase ( A_ ):
A__ : int = "vision-encoder-decoder"
A__ : Tuple = True
def __init__(self : Optional[Any] , **snake_case__ : Optional[int] ) -> List[Any]:
'''simple docstring'''
super().__init__(**snake_case__ )
if "encoder" not in kwargs or "decoder" not in kwargs:
raise ValueError(
f"""A configuraton of type {self.model_type} cannot be instantiated because """
f"""not both `encoder` and `decoder` sub-configurations are passed, but only {kwargs}""" )
snake_case : int = kwargs.pop("encoder" )
snake_case : Tuple = encoder_config.pop("model_type" )
snake_case : int = kwargs.pop("decoder" )
snake_case : Optional[Any] = decoder_config.pop("model_type" )
snake_case : Tuple = AutoConfig.for_model(snake_case__ , **snake_case__ )
snake_case : List[Any] = AutoConfig.for_model(snake_case__ , **snake_case__ )
snake_case : List[Any] = True
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : PretrainedConfig , snake_case__ : PretrainedConfig , **snake_case__ : int ) -> PretrainedConfig:
'''simple docstring'''
logger.info("Setting `config.is_decoder=True` and `config.add_cross_attention=True` for decoder_config" )
snake_case : List[str] = True
snake_case : Optional[Any] = True
return cls(encoder=encoder_config.to_dict() , decoder=decoder_config.to_dict() , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
snake_case : Any = copy.deepcopy(self.__dict__ )
snake_case : List[Any] = self.encoder.to_dict()
snake_case : Tuple = self.decoder.to_dict()
snake_case : Union[str, Any] = self.__class__.model_type
return output
class UpperCAmelCase ( A_ ):
A__ : str = version.parse("1.11" )
@property
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
return OrderedDict(
[
("pixel_values", {0: "batch", 1: "num_channels", 2: "height", 3: "width"}),
] )
@property
def _SCREAMING_SNAKE_CASE (self : Any ) -> float:
'''simple docstring'''
return 1e-4
@property
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
return OrderedDict({"last_hidden_state": {0: "batch", 1: "encoder_sequence"}} )
class UpperCAmelCase ( A_ ):
@property
def _SCREAMING_SNAKE_CASE (self : Any ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
snake_case : Dict = OrderedDict()
snake_case : Optional[Any] = {0: "batch", 1: "past_decoder_sequence + sequence"}
snake_case : Tuple = {0: "batch", 1: "past_decoder_sequence + sequence"}
snake_case : List[Any] = {0: "batch", 1: "encoder_sequence"}
return common_inputs
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : "PreTrainedTokenizerBase" , snake_case__ : int = -1 , snake_case__ : int = -1 , snake_case__ : bool = False , snake_case__ : Optional["TensorType"] = None , ) -> Mapping[str, Any]:
'''simple docstring'''
import torch
snake_case : str = OrderedDict()
snake_case : Dict = super().generate_dummy_inputs(
snake_case__ , batch_size=snake_case__ , seq_length=snake_case__ , is_pair=snake_case__ , framework=snake_case__ )
snake_case , snake_case : Dict = dummy_input["input_ids"].shape
snake_case : Dict = (batch, encoder_sequence, self._config.encoder_hidden_size)
snake_case : List[str] = dummy_input.pop("input_ids" )
snake_case : Optional[int] = dummy_input.pop("attention_mask" )
snake_case : Union[str, Any] = torch.zeros(snake_case__ )
return common_inputs
class UpperCAmelCase ( A_ ):
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> None:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : PretrainedConfig ) -> OnnxConfig:
'''simple docstring'''
return VisionEncoderDecoderEncoderOnnxConfig(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : PretrainedConfig , snake_case__ : PretrainedConfig , snake_case__ : str = "default" ) -> OnnxConfig:
'''simple docstring'''
snake_case : Union[str, Any] = encoder_config.hidden_size
return VisionEncoderDecoderDecoderOnnxConfig(snake_case__ , snake_case__ )
| 59
|
__lowerCamelCase = {
"joule": 1.0,
"kilojoule": 10_00,
"megajoule": 1_00_00_00,
"gigajoule": 10_00_00_00_00,
"wattsecond": 1.0,
"watthour": 36_00,
"kilowatthour": 3_60_00_00,
"newtonmeter": 1.0,
"calorie_nutr": 41_86.8,
"kilocalorie_nutr": 4_18_68_00.00,
"electronvolt": 1.602_176_634e-19,
"britishthermalunit_it": 10_55.0_55_85,
"footpound": 1.35_5818,
}
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : float ):
if to_type not in ENERGY_CONVERSION or from_type not in ENERGY_CONVERSION:
snake_case : List[Any] = (
f"""Incorrect 'from_type' or 'to_type' value: {from_type!r}, {to_type!r}\n"""
f"""Valid values are: {', '.join(__lowerCamelCase )}"""
)
raise ValueError(__lowerCamelCase )
return value * ENERGY_CONVERSION[from_type] / ENERGY_CONVERSION[to_type]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
from diffusers.utils.testing_utils import require_onnxruntime
@require_onnxruntime
class UpperCAmelCase :
pass
| 59
|
import logging
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
import datasets
import numpy as np
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PreTrainedTokenizer,
TFAutoModelForSequenceClassification,
TFTrainer,
TFTrainingArguments,
)
from transformers.utils import logging as hf_logging
hf_logging.set_verbosity_info()
hf_logging.enable_default_handler()
hf_logging.enable_explicit_format()
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : PreTrainedTokenizer , __lowerCamelCase : int , __lowerCamelCase : Optional[int] = None , ):
snake_case : int = {}
if train_file is not None:
snake_case : List[Any] = [train_file]
if eval_file is not None:
snake_case : Optional[int] = [eval_file]
if test_file is not None:
snake_case : Any = [test_file]
snake_case : int = datasets.load_dataset("csv" , data_files=__lowerCamelCase )
snake_case : str = list(ds[list(files.keys() )[0]].features.keys() )
snake_case : int = features_name.pop(__lowerCamelCase )
snake_case : str = list(set(ds[list(files.keys() )[0]][label_name] ) )
snake_case : str = {label: i for i, label in enumerate(__lowerCamelCase )}
snake_case : List[Any] = tokenizer.model_input_names
snake_case : List[Any] = {}
if len(__lowerCamelCase ) == 1:
for k in files.keys():
snake_case : Tuple = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
example[features_name[0]] , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" ) , batched=__lowerCamelCase , )
elif len(__lowerCamelCase ) == 2:
for k in files.keys():
snake_case : List[Any] = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
(example[features_name[0]], example[features_name[1]]) , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" , ) , batched=__lowerCamelCase , )
def gen_train():
for ex in transformed_ds[datasets.Split.TRAIN]:
snake_case : Dict = {k: v for k, v in ex.items() if k in input_names}
snake_case : Union[str, Any] = labelaid[ex[label_name]]
yield (d, label)
def gen_val():
for ex in transformed_ds[datasets.Split.VALIDATION]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : Any = labelaid[ex[label_name]]
yield (d, label)
def gen_test():
for ex in transformed_ds[datasets.Split.TEST]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : List[str] = labelaid[ex[label_name]]
yield (d, label)
snake_case : int = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TRAIN in transformed_ds
else None
)
if train_ds is not None:
snake_case : Optional[Any] = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) )
snake_case : Tuple = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.VALIDATION in transformed_ds
else None
)
if val_ds is not None:
snake_case : List[str] = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) )
snake_case : Optional[int] = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TEST in transformed_ds
else None
)
if test_ds is not None:
snake_case : str = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) )
return train_ds, val_ds, test_ds, labelaid
__lowerCamelCase = logging.getLogger(__name__)
@dataclass
class UpperCAmelCase :
A__ : int = field(metadata={"help": "Which column contains the label"} )
A__ : str = field(default=A_ ,metadata={"help": "The path of the training file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the development file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the test file"} )
A__ : int = field(
default=1_28 ,metadata={
"help": (
"The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
} ,)
A__ : bool = field(
default=A_ ,metadata={"help": "Overwrite the cached training and evaluation sets"} )
@dataclass
class UpperCAmelCase :
A__ : str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained config name or path if not the same as model_name"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"} )
A__ : bool = field(default=A_ ,metadata={"help": "Set this flag to use fast tokenization."} )
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
# or just modify its tokenizer_config.json.
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} ,)
def UpperCamelCase ( ):
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
snake_case : List[Any] = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) )
snake_case , snake_case , snake_case : int = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir )
and os.listdir(training_args.output_dir )
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
f"""Output directory ({training_args.output_dir}) already exists and is not empty. Use"""
" --overwrite_output_dir to overcome." )
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(
f"""n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, """
f"""16-bits training: {training_args.fpaa}""" )
logger.info(f"""Training/evaluation parameters {training_args}""" )
# Load pretrained model and tokenizer
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
snake_case : Tuple = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path , cache_dir=model_args.cache_dir , )
snake_case , snake_case , snake_case , snake_case : Tuple = get_tfds(
train_file=data_args.train_file , eval_file=data_args.dev_file , test_file=data_args.test_file , tokenizer=__lowerCamelCase , label_column_id=data_args.label_column_id , max_seq_length=data_args.max_seq_length , )
snake_case : Optional[Any] = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path , num_labels=len(__lowerCamelCase ) , labelaid=__lowerCamelCase , idalabel={id: label for label, id in labelaid.items()} , finetuning_task="text-classification" , cache_dir=model_args.cache_dir , )
with training_args.strategy.scope():
snake_case : int = TFAutoModelForSequenceClassification.from_pretrained(
model_args.model_name_or_path , from_pt=bool(".bin" in model_args.model_name_or_path ) , config=__lowerCamelCase , cache_dir=model_args.cache_dir , )
def compute_metrics(__lowerCamelCase : EvalPrediction ) -> Dict:
snake_case : Optional[int] = np.argmax(p.predictions , axis=1 )
return {"acc": (preds == p.label_ids).mean()}
# Initialize our Trainer
snake_case : int = TFTrainer(
model=__lowerCamelCase , args=__lowerCamelCase , train_dataset=__lowerCamelCase , eval_dataset=__lowerCamelCase , compute_metrics=__lowerCamelCase , )
# Training
if training_args.do_train:
trainer.train()
trainer.save_model()
tokenizer.save_pretrained(training_args.output_dir )
# Evaluation
snake_case : int = {}
if training_args.do_eval:
logger.info("*** Evaluate ***" )
snake_case : Any = trainer.evaluate()
snake_case : List[Any] = os.path.join(training_args.output_dir , "eval_results.txt" )
with open(__lowerCamelCase , "w" ) as writer:
logger.info("***** Eval results *****" )
for key, value in result.items():
logger.info(f""" {key} = {value}""" )
writer.write(f"""{key} = {value}\n""" )
results.update(__lowerCamelCase )
return results
if __name__ == "__main__":
main()
| 59
| 1
|
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Optional[Any] = (1 + 24 * n) ** 0.5
return ((1 + root) / 6) % 1 == 0
def UpperCamelCase ( __lowerCamelCase : int = 5000 ):
snake_case : Optional[Any] = [(i * (3 * i - 1)) // 2 for i in range(1 , __lowerCamelCase )]
for i, pentagonal_i in enumerate(__lowerCamelCase ):
for j in range(__lowerCamelCase , len(__lowerCamelCase ) ):
snake_case : List[Any] = pentagonal_nums[j]
snake_case : Union[str, Any] = pentagonal_i + pentagonal_j
snake_case : List[Any] = pentagonal_j - pentagonal_i
if is_pentagonal(__lowerCamelCase ) and is_pentagonal(__lowerCamelCase ):
return b
return -1
if __name__ == "__main__":
print(F'{solution() = }')
| 59
|
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Any ) -> List[str]:
'''simple docstring'''
snake_case : int = tempfile.mkdtemp()
# fmt: off
snake_case : Optional[int] = ["[UNK]", "[CLS]", "[SEP]", "[PAD]", "[MASK]", "want", "##want", "##ed", "wa", "un", "runn", "##ing", ",", "low", "lowest"]
# fmt: on
snake_case : List[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] ) )
snake_case : int = {
"do_resize": True,
"size": {"height": 18, "width": 18},
"do_normalize": True,
"image_mean": [0.5, 0.5, 0.5],
"image_std": [0.5, 0.5, 0.5],
}
snake_case : Optional[Any] = os.path.join(self.tmpdirname , snake_case__ )
with open(self.image_processor_file , "w" , encoding="utf-8" ) as fp:
json.dump(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : str ) -> Optional[int]:
'''simple docstring'''
return BertTokenizer.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : List[str] ) -> int:
'''simple docstring'''
return ViTImageProcessor.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
shutil.rmtree(self.tmpdirname )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : List[Any] = [np.random.randint(2_55 , size=(3, 30, 4_00) , dtype=np.uinta )]
snake_case : Optional[int] = [Image.fromarray(np.moveaxis(snake_case__ , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.get_tokenizer()
snake_case : Optional[Any] = self.get_image_processor()
snake_case : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
processor.save_pretrained(self.tmpdirname )
snake_case : Any = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
snake_case : Optional[int] = self.get_tokenizer(bos_token="(BOS)" , eos_token="(EOS)" )
snake_case : Tuple = self.get_image_processor(do_normalize=snake_case__ , padding_value=1.0 )
snake_case : List[str] = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token="(BOS)" , eos_token="(EOS)" , do_normalize=snake_case__ , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Optional[int] = self.get_tokenizer()
snake_case : List[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = self.prepare_image_inputs()
snake_case : str = image_processor(snake_case__ , return_tensors="np" )
snake_case : Any = processor(images=snake_case__ , return_tensors="np" )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.get_image_processor()
snake_case : int = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Tuple = "lower newer"
snake_case : Tuple = processor(text=snake_case__ )
snake_case : Union[str, Any] = tokenizer(snake_case__ )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[int]:
'''simple docstring'''
snake_case : List[Any] = self.get_image_processor()
snake_case : Dict = self.get_tokenizer()
snake_case : Dict = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : int = "lower newer"
snake_case : Dict = self.prepare_image_inputs()
snake_case : Union[str, Any] = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , ["input_ids", "token_type_ids", "attention_mask", "pixel_values"] )
# test if it raises when no input is passed
with self.assertRaises(snake_case__ ):
processor()
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : Tuple = self.get_image_processor()
snake_case : Optional[Any] = self.get_tokenizer()
snake_case : Tuple = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : List[str] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
snake_case : List[Any] = processor.batch_decode(snake_case__ )
snake_case : Union[str, Any] = tokenizer.batch_decode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Union[str, Any] = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = "lower newer"
snake_case : List[Any] = self.prepare_image_inputs()
snake_case : Tuple = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 59
| 1
|
# Copyright 2023 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
__lowerCamelCase = {
"""configuration_mgp_str""": ["""MGP_STR_PRETRAINED_CONFIG_ARCHIVE_MAP""", """MgpstrConfig"""],
"""processing_mgp_str""": ["""MgpstrProcessor"""],
"""tokenization_mgp_str""": ["""MgpstrTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""MGP_STR_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""MgpstrModel""",
"""MgpstrPreTrainedModel""",
"""MgpstrForSceneTextRecognition""",
]
if TYPE_CHECKING:
from .configuration_mgp_str import MGP_STR_PRETRAINED_CONFIG_ARCHIVE_MAP, MgpstrConfig
from .processing_mgp_str import MgpstrProcessor
from .tokenization_mgp_str import MgpstrTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mgp_str import (
MGP_STR_PRETRAINED_MODEL_ARCHIVE_LIST,
MgpstrForSceneTextRecognition,
MgpstrModel,
MgpstrPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__lowerCamelCase = {
"""configuration_biogpt""": ["""BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP""", """BioGptConfig"""],
"""tokenization_biogpt""": ["""BioGptTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""BioGptForCausalLM""",
"""BioGptForTokenClassification""",
"""BioGptForSequenceClassification""",
"""BioGptModel""",
"""BioGptPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_biogpt import BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, BioGptConfig
from .tokenization_biogpt import BioGptTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_biogpt import (
BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
BioGptForCausalLM,
BioGptForSequenceClassification,
BioGptForTokenClassification,
BioGptModel,
BioGptPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
| 1
|
import os
from pathlib import Path
def UpperCamelCase ( ):
from torch.utils.cpp_extension import load
snake_case : str = Path(__lowerCamelCase ).resolve().parent.parent.parent / "kernels" / "deformable_detr"
snake_case : int = [
root / filename
for filename in [
"vision.cpp",
os.path.join("cpu" , "ms_deform_attn_cpu.cpp" ),
os.path.join("cuda" , "ms_deform_attn_cuda.cu" ),
]
]
load(
"MultiScaleDeformableAttention" , __lowerCamelCase , with_cuda=__lowerCamelCase , extra_include_paths=[str(__lowerCamelCase )] , extra_cflags=["-DWITH_CUDA=1"] , extra_cuda_cflags=[
"-DCUDA_HAS_FP16=1",
"-D__CUDA_NO_HALF_OPERATORS__",
"-D__CUDA_NO_HALF_CONVERSIONS__",
"-D__CUDA_NO_HALF2_OPERATORS__",
] , )
import MultiScaleDeformableAttention as MSDA
return MSDA
| 59
|
import collections
import inspect
import unittest
from typing import Dict, List, Tuple
from transformers import MaskFormerSwinConfig
from transformers.testing_utils import require_torch, require_torch_multi_gpu, torch_device
from transformers.utils import is_torch_available
from ...test_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import MaskFormerSwinBackbone
from transformers.models.maskformer import MaskFormerSwinModel
class UpperCAmelCase :
def __init__(self : Dict , snake_case__ : Dict , snake_case__ : Any=13 , snake_case__ : Any=32 , snake_case__ : Optional[Any]=2 , snake_case__ : Union[str, Any]=3 , snake_case__ : List[Any]=16 , snake_case__ : int=[1, 2, 1] , snake_case__ : Dict=[2, 2, 4] , snake_case__ : Dict=2 , snake_case__ : Tuple=2.0 , snake_case__ : Optional[int]=True , snake_case__ : Union[str, Any]=0.0 , snake_case__ : Any=0.0 , snake_case__ : Union[str, Any]=0.1 , snake_case__ : int="gelu" , snake_case__ : Optional[int]=False , snake_case__ : List[Any]=True , snake_case__ : List[str]=0.02 , snake_case__ : int=1e-5 , snake_case__ : List[str]=True , snake_case__ : Union[str, Any]=None , snake_case__ : List[Any]=True , snake_case__ : Optional[Any]=10 , snake_case__ : Optional[Any]=8 , snake_case__ : Any=["stage1", "stage2", "stage3"] , snake_case__ : Tuple=[1, 2, 3] , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = parent
snake_case : Optional[int] = batch_size
snake_case : Union[str, Any] = image_size
snake_case : Dict = patch_size
snake_case : Optional[Any] = num_channels
snake_case : Union[str, Any] = embed_dim
snake_case : int = depths
snake_case : List[str] = num_heads
snake_case : Union[str, Any] = window_size
snake_case : Union[str, Any] = mlp_ratio
snake_case : List[Any] = qkv_bias
snake_case : List[Any] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Union[str, Any] = drop_path_rate
snake_case : int = hidden_act
snake_case : Optional[int] = use_absolute_embeddings
snake_case : int = patch_norm
snake_case : Union[str, Any] = layer_norm_eps
snake_case : Any = initializer_range
snake_case : Optional[Any] = is_training
snake_case : Tuple = scope
snake_case : Optional[int] = use_labels
snake_case : Optional[Any] = type_sequence_label_size
snake_case : Union[str, Any] = encoder_stride
snake_case : Any = out_features
snake_case : Tuple = out_indices
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case : Optional[Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case : int = None
if self.use_labels:
snake_case : Tuple = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : Dict = self.get_config()
return config, pixel_values, labels
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> int:
'''simple docstring'''
return MaskFormerSwinConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , depths=self.depths , num_heads=self.num_heads , window_size=self.window_size , mlp_ratio=self.mlp_ratio , qkv_bias=self.qkv_bias , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , drop_path_rate=self.drop_path_rate , hidden_act=self.hidden_act , use_absolute_embeddings=self.use_absolute_embeddings , path_norm=self.patch_norm , layer_norm_eps=self.layer_norm_eps , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , out_features=self.out_features , out_indices=self.out_indices , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : List[Any] , snake_case__ : List[str] , snake_case__ : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
snake_case : Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case : int = int(config.embed_dim * 2 ** (len(config.depths ) - 1) )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, expected_seq_len, expected_dim) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = MaskFormerSwinBackbone(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [13, 16, 16, 16] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , [16, 32, 64] )
# verify ValueError
with self.parent.assertRaises(snake_case__ ):
snake_case : Tuple = ["stem"]
snake_case : List[Any] = MaskFormerSwinBackbone(config=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self.prepare_config_and_inputs()
snake_case , snake_case , snake_case : List[Any] = config_and_inputs
snake_case : int = {"pixel_values": pixel_values}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : List[str] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
A__ : str = {"feature-extraction": MaskFormerSwinModel} if is_torch_available() else {}
A__ : Optional[Any] = False
A__ : List[Any] = False
A__ : List[str] = False
A__ : List[str] = False
A__ : Union[str, Any] = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = MaskFormerSwinModelTester(self )
snake_case : Optional[int] = ConfigTester(self , config_class=snake_case__ , embed_dim=37 )
@require_torch_multi_gpu
@unittest.skip(
reason=(
"`MaskFormerSwinModel` outputs `hidden_states_spatial_dimensions` which doesn't work well with"
" `nn.DataParallel`"
) )
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : str ) -> List[str]:
'''simple docstring'''
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
return
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Dict:
'''simple docstring'''
snake_case : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*snake_case__ )
@unittest.skip("Swin does not use inputs_embeds" )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
pass
@unittest.skip("Swin does not support feedforward chunking" )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Dict:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : int = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case : List[Any] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , nn.Linear ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case , snake_case : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : str = model_class(snake_case__ )
snake_case : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case : Optional[Any] = [*signature.parameters.keys()]
snake_case : Tuple = ["pixel_values"]
self.assertListEqual(arg_names[:1] , snake_case__ )
@unittest.skip(reason="MaskFormerSwin is only used as backbone and doesn't support output_attentions" )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="MaskFormerSwin is only used as an internal backbone" )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
with torch.no_grad():
snake_case : Any = model(**self._prepare_for_class(snake_case__ , snake_case__ ) )
snake_case : int = outputs.hidden_states
snake_case : Union[str, Any] = getattr(
self.model_tester , "expected_num_hidden_layers" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case__ ) , snake_case__ )
# Swin has a different seq_length
snake_case : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : Tuple = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : int = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
for model_class in self.all_model_classes:
snake_case : int = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Dict = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case , snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Any = 3
snake_case : List[str] = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
snake_case : Tuple = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : str = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case : str = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes:
snake_case : str = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Optional[Any] = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
@unittest.skip(reason="MaskFormerSwin doesn't have pretrained checkpoints" )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : int ) -> str:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case__ : Union[str, Any] ):
snake_case : Any = 0
return t
def check_equivalence(snake_case__ : Union[str, Any] , snake_case__ : int , snake_case__ : List[str] , snake_case__ : Optional[int]={} ):
with torch.no_grad():
snake_case : Optional[Any] = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ )
snake_case : Tuple = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ ).to_tuple()
def recursive_check(snake_case__ : List[str] , snake_case__ : Optional[Any] ):
if isinstance(snake_case__ , (List, Tuple) ):
for tuple_iterable_value, dict_iterable_value in zip(snake_case__ , snake_case__ ):
recursive_check(snake_case__ , snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values() , dict_object.values() ):
recursive_check(snake_case__ , snake_case__ )
elif tuple_object is None:
return
else:
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(snake_case__ ) , set_nan_tensor_to_zero(snake_case__ ) , atol=1e-5 ) , msg=(
"Tuple and dict output are not equal. Difference:"
f""" {torch.max(torch.abs(tuple_object - dict_object ) )}. Tuple has `nan`:"""
f""" {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}. Dict has"""
f""" `nan`: {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}."""
) , )
recursive_check(snake_case__ , snake_case__ )
for model_class in self.all_model_classes:
snake_case : Optional[int] = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Union[str, Any] = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : Optional[Any] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Dict = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : List[Any] = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
snake_case : Any = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : List[str] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
@require_torch
class UpperCAmelCase ( unittest.TestCase ,A_ ):
A__ : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
A__ : int = MaskFormerSwinConfig
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModelTester(self )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case : Dict = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Optional[int] = inputs_dict["pixel_values"].shape[0]
for backbone_class in self.all_model_classes:
snake_case : Optional[int] = backbone_class(snake_case__ )
backbone.to(snake_case__ )
backbone.eval()
snake_case : Union[str, Any] = backbone(**snake_case__ )
# Test default outputs and verify feature maps
self.assertIsInstance(outputs.feature_maps , snake_case__ )
self.assertTrue(len(outputs.feature_maps ) == len(backbone.channels ) )
for feature_map, n_channels in zip(outputs.feature_maps , backbone.channels ):
self.assertTrue(feature_map.shape[:2] , (batch_size, n_channels) )
self.assertIsNone(outputs.hidden_states )
self.assertIsNone(outputs.attentions )
# Test output_hidden_states=True
snake_case : Optional[int] = backbone(**snake_case__ , output_hidden_states=snake_case__ )
self.assertIsNotNone(outputs.hidden_states )
self.assertTrue(len(outputs.hidden_states ) , len(backbone.stage_names ) )
# We skip the stem layer
for hidden_states, n_channels in zip(outputs.hidden_states[1:] , backbone.channels ):
for hidden_state in hidden_states:
# Hidden states are in the format (batch_size, (height * width), n_channels)
snake_case , snake_case , snake_case : Dict = hidden_state.shape
self.assertTrue((h_batch_size, h_n_channels) , (batch_size, n_channels) )
# Test output_attentions=True
if self.has_attentions:
snake_case : Optional[Any] = backbone(**snake_case__ , output_attentions=snake_case__ )
self.assertIsNotNone(outputs.attentions )
| 59
| 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.
from typing import TYPE_CHECKING
import torch
from ..models.auto import AutoModelForVisualQuestionAnswering, AutoProcessor
from ..utils import requires_backends
from .base import PipelineTool
if TYPE_CHECKING:
from PIL import Image
class UpperCAmelCase ( A_ ):
A__ : List[Any] = "dandelin/vilt-b32-finetuned-vqa"
A__ : List[str] = (
"This is a tool that answers a question about an image. It takes an input named `image` which should be the "
"image containing the information, as well as a `question` which should be the question in English. It "
"returns a text that is the answer to the question."
)
A__ : str = "image_qa"
A__ : List[str] = AutoProcessor
A__ : Dict = AutoModelForVisualQuestionAnswering
A__ : List[str] = ["image", "text"]
A__ : Union[str, Any] = ["text"]
def __init__(self : int , *snake_case__ : int , **snake_case__ : str ) -> List[Any]:
'''simple docstring'''
requires_backends(self , ["vision"] )
super().__init__(*snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : "Image" , snake_case__ : str ) -> Tuple:
'''simple docstring'''
return self.pre_processor(snake_case__ , snake_case__ , return_tensors="pt" )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : int ) -> Union[str, Any]:
'''simple docstring'''
with torch.no_grad():
return self.model(**snake_case__ ).logits
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Optional[int] = outputs.argmax(-1 ).item()
return self.model.config.idalabel[idx]
| 59
|
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : List[str] = []
snake_case : Optional[int] = []
snake_case : Any = []
for rt in rc.restypes:
snake_case : List[Any] = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
snake_case : str = {name: i for i, name in enumerate(__lowerCamelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
snake_case : Optional[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : List[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : int = torch.tensor(
__lowerCamelCase , dtype=torch.floataa , device=protein["aatype"].device , )
snake_case : int = protein["aatype"].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
snake_case : List[Any] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : str = restype_atomaa_mask[protein_aatype]
snake_case : str = residx_atomaa_mask
snake_case : Any = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
snake_case : List[str] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : List[Any] = residx_atomaa_to_atomaa.long()
# create the corresponding mask
snake_case : Union[str, Any] = torch.zeros([21, 37] , dtype=torch.floataa , device=protein["aatype"].device )
for restype, restype_letter in enumerate(rc.restypes ):
snake_case : Optional[int] = rc.restype_atoa[restype_letter]
snake_case : Any = rc.residue_atoms[restype_name]
for atom_name in atom_names:
snake_case : List[Any] = rc.atom_order[atom_name]
snake_case : Optional[Any] = 1
snake_case : List[Any] = restype_atomaa_mask[protein_aatype]
snake_case : int = residx_atomaa_mask
return protein
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : Dict = tree_map(lambda __lowerCamelCase : torch.tensor(__lowerCamelCase , device=batch["aatype"].device ) , __lowerCamelCase , np.ndarray )
snake_case : List[str] = tensor_tree_map(lambda __lowerCamelCase : np.array(__lowerCamelCase ) , make_atomaa_masks(__lowerCamelCase ) )
return out
| 59
| 1
|
from collections import OrderedDict
from typing import List, Mapping
from packaging import version
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""google/efficientnet-b7""": """https://huggingface.co/google/efficientnet-b7/resolve/main/config.json""",
}
class UpperCAmelCase ( A_ ):
A__ : Any = "efficientnet"
def __init__(self : Optional[Any] , snake_case__ : int = 3 , snake_case__ : int = 6_00 , snake_case__ : float = 2.0 , snake_case__ : float = 3.1 , snake_case__ : int = 8 , snake_case__ : List[int] = [3, 3, 5, 3, 5, 5, 3] , snake_case__ : List[int] = [32, 16, 24, 40, 80, 1_12, 1_92] , snake_case__ : List[int] = [16, 24, 40, 80, 1_12, 1_92, 3_20] , snake_case__ : List[int] = [] , snake_case__ : List[int] = [1, 2, 2, 2, 1, 2, 1] , snake_case__ : List[int] = [1, 2, 2, 3, 3, 4, 1] , snake_case__ : List[int] = [1, 6, 6, 6, 6, 6, 6] , snake_case__ : float = 0.25 , snake_case__ : str = "swish" , snake_case__ : int = 25_60 , snake_case__ : str = "mean" , snake_case__ : float = 0.02 , snake_case__ : float = 0.001 , snake_case__ : float = 0.99 , snake_case__ : float = 0.5 , snake_case__ : float = 0.2 , **snake_case__ : int , ) -> Union[str, Any]:
'''simple docstring'''
super().__init__(**snake_case__ )
snake_case : List[str] = num_channels
snake_case : Optional[Any] = image_size
snake_case : Any = width_coefficient
snake_case : Any = depth_coefficient
snake_case : Optional[int] = depth_divisor
snake_case : Tuple = kernel_sizes
snake_case : Dict = in_channels
snake_case : Tuple = out_channels
snake_case : Any = depthwise_padding
snake_case : Optional[Any] = strides
snake_case : Optional[int] = num_block_repeats
snake_case : Optional[Any] = expand_ratios
snake_case : List[Any] = squeeze_expansion_ratio
snake_case : Any = hidden_act
snake_case : Dict = hidden_dim
snake_case : Optional[Any] = pooling_type
snake_case : Optional[int] = initializer_range
snake_case : Union[str, Any] = batch_norm_eps
snake_case : Dict = batch_norm_momentum
snake_case : Optional[Any] = dropout_rate
snake_case : List[str] = drop_connect_rate
snake_case : int = sum(snake_case__ ) * 4
class UpperCAmelCase ( A_ ):
A__ : List[str] = version.parse("1.11" )
@property
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Mapping[str, Mapping[int, str]]:
'''simple docstring'''
return OrderedDict(
[
("pixel_values", {0: "batch", 1: "num_channels", 2: "height", 3: "width"}),
] )
@property
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> float:
'''simple docstring'''
return 1e-5
| 59
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from .tokenization_lxmert import LxmertTokenizer
__lowerCamelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {
"""unc-nlp/lxmert-base-uncased""": """https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""unc-nlp/lxmert-base-uncased""": (
"""https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/tokenizer.json"""
),
},
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": 5_12,
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": {"""do_lower_case""": True},
}
class UpperCAmelCase ( A_ ):
A__ : Any = VOCAB_FILES_NAMES
A__ : Optional[int] = PRETRAINED_VOCAB_FILES_MAP
A__ : Tuple = PRETRAINED_INIT_CONFIGURATION
A__ : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : List[Any] = LxmertTokenizer
def __init__(self : Dict , snake_case__ : Tuple=None , snake_case__ : Optional[Any]=None , snake_case__ : Optional[Any]=True , snake_case__ : Tuple="[UNK]" , snake_case__ : Optional[Any]="[SEP]" , snake_case__ : Optional[Any]="[PAD]" , snake_case__ : List[Any]="[CLS]" , snake_case__ : Tuple="[MASK]" , snake_case__ : Dict=True , snake_case__ : Union[str, Any]=None , **snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
snake_case : int = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , snake_case__ ) != do_lower_case
or normalizer_state.get("strip_accents" , snake_case__ ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , snake_case__ ) != tokenize_chinese_chars
):
snake_case : Union[str, Any] = getattr(snake_case__ , normalizer_state.pop("type" ) )
snake_case : str = do_lower_case
snake_case : List[Any] = strip_accents
snake_case : Optional[int] = tokenize_chinese_chars
snake_case : int = normalizer_class(**snake_case__ )
snake_case : Optional[Any] = do_lower_case
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Optional[Any] , snake_case__ : Dict=None ) -> Any:
'''simple docstring'''
snake_case : Optional[int] = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Optional[Any] = [self.sep_token_id]
snake_case : 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 ) * [0] + len(token_ids_a + sep ) * [1]
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
snake_case : List[Any] = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 59
| 1
|
import os
def UpperCamelCase ( __lowerCamelCase : Optional[int] ):
snake_case : Union[str, Any] = len(grid[0] )
snake_case : int = len(__lowerCamelCase )
snake_case : List[Any] = 0
snake_case : Dict = 0
snake_case : int = 0
# Check vertically, horizontally, diagonally at the same time (only works
# for nxn grid)
for i in range(__lowerCamelCase ):
for j in range(n_rows - 3 ):
snake_case : int = grid[j][i] * grid[j + 1][i] * grid[j + 2][i] * grid[j + 3][i]
snake_case : Dict = grid[i][j] * grid[i][j + 1] * grid[i][j + 2] * grid[i][j + 3]
# Left-to-right diagonal (\) product
if i < n_columns - 3:
snake_case : List[str] = (
grid[i][j]
* grid[i + 1][j + 1]
* grid[i + 2][j + 2]
* grid[i + 3][j + 3]
)
# Right-to-left diagonal(/) product
if i > 2:
snake_case : Tuple = (
grid[i][j]
* grid[i - 1][j + 1]
* grid[i - 2][j + 2]
* grid[i - 3][j + 3]
)
snake_case : Any = max(
__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
if max_product > largest:
snake_case : List[str] = max_product
return largest
def UpperCamelCase ( ):
snake_case : Optional[int] = []
with open(os.path.dirname(__lowerCamelCase ) + "/grid.txt" ) as file:
for line in file:
grid.append(line.strip("\n" ).split(" " ) )
snake_case : Optional[int] = [[int(__lowerCamelCase ) for i in grid[j]] for j in range(len(__lowerCamelCase ) )]
return largest_product(__lowerCamelCase )
if __name__ == "__main__":
print(solution())
| 59
|
import torch
from diffusers import DDIMParallelScheduler
from .test_schedulers import SchedulerCommonTest
class UpperCAmelCase ( A_ ):
A__ : Dict = (DDIMParallelScheduler,)
A__ : Tuple = (("eta", 0.0), ("num_inference_steps", 50))
def _SCREAMING_SNAKE_CASE (self : Tuple , **snake_case__ : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case : Any = {
"num_train_timesteps": 10_00,
"beta_start": 0.0001,
"beta_end": 0.02,
"beta_schedule": "linear",
"clip_sample": True,
}
config.update(**snake_case__ )
return config
def _SCREAMING_SNAKE_CASE (self : Dict , **snake_case__ : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = self.scheduler_classes[0]
snake_case : Any = self.get_scheduler_config(**snake_case__ )
snake_case : Any = scheduler_class(**snake_case__ )
snake_case , snake_case : Union[str, Any] = 10, 0.0
snake_case : List[Any] = self.dummy_model()
snake_case : Any = self.dummy_sample_deter
scheduler.set_timesteps(snake_case__ )
for t in scheduler.timesteps:
snake_case : Optional[int] = model(snake_case__ , snake_case__ )
snake_case : List[str] = scheduler.step(snake_case__ , snake_case__ , snake_case__ , snake_case__ ).prev_sample
return sample
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> str:
'''simple docstring'''
for timesteps in [1_00, 5_00, 10_00]:
self.check_over_configs(num_train_timesteps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=snake_case__ )
snake_case : Optional[int] = self.scheduler_classes[0]
snake_case : Optional[int] = self.get_scheduler_config(steps_offset=1 )
snake_case : Union[str, Any] = scheduler_class(**snake_case__ )
scheduler.set_timesteps(5 )
assert torch.equal(scheduler.timesteps , torch.LongTensor([8_01, 6_01, 4_01, 2_01, 1] ) )
def _SCREAMING_SNAKE_CASE (self : int ) -> Tuple:
'''simple docstring'''
for beta_start, beta_end in zip([0.0001, 0.001, 0.01, 0.1] , [0.002, 0.02, 0.2, 2] ):
self.check_over_configs(beta_start=snake_case__ , beta_end=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
for timestep_spacing in ["trailing", "leading"]:
self.check_over_configs(timestep_spacing=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[Any]:
'''simple docstring'''
for rescale_betas_zero_snr in [True, False]:
self.check_over_configs(rescale_betas_zero_snr=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
self.check_over_configs(thresholding=snake_case__ )
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(
thresholding=snake_case__ , prediction_type=snake_case__ , sample_max_value=snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
for t in [1, 10, 49]:
self.check_over_forward(time_step=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Any:
'''simple docstring'''
for t, num_inference_steps in zip([1, 10, 50] , [10, 50, 5_00] ):
self.check_over_forward(time_step=snake_case__ , num_inference_steps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Optional[Any]:
'''simple docstring'''
for t, eta in zip([1, 10, 49] , [0.0, 0.5, 1.0] ):
self.check_over_forward(time_step=snake_case__ , eta=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Dict = self.scheduler_classes[0]
snake_case : Tuple = self.get_scheduler_config()
snake_case : Dict = scheduler_class(**snake_case__ )
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_20 , 4_00 ) - 0.14771 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_80 , 9_60 ) - 0.32460 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_87 , 4_86 ) - 0.00979 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_99 , 9_98 ) - 0.02 ) ) < 1e-5
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
snake_case : Union[str, Any] = self.scheduler_classes[0]
snake_case : List[Any] = self.get_scheduler_config()
snake_case : int = scheduler_class(**snake_case__ )
snake_case , snake_case : Any = 10, 0.0
scheduler.set_timesteps(snake_case__ )
snake_case : Optional[Any] = self.dummy_model()
snake_case : str = self.dummy_sample_deter
snake_case : Dict = self.dummy_sample_deter + 0.1
snake_case : Dict = self.dummy_sample_deter - 0.1
snake_case : Optional[Any] = samplea.shape[0]
snake_case : str = torch.stack([samplea, samplea, samplea] , dim=0 )
snake_case : Tuple = torch.arange(snake_case__ )[0:3, None].repeat(1 , snake_case__ )
snake_case : Tuple = model(samples.flatten(0 , 1 ) , timesteps.flatten(0 , 1 ) )
snake_case : List[str] = scheduler.batch_step_no_noise(snake_case__ , timesteps.flatten(0 , 1 ) , samples.flatten(0 , 1 ) , snake_case__ )
snake_case : Dict = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 1147.7904 ) < 1e-2
assert abs(result_mean.item() - 0.4982 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[Any] = self.full_loop()
snake_case : Optional[Any] = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 172.0067 ) < 1e-2
assert abs(result_mean.item() - 0.223967 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : str ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(prediction_type="v_prediction" )
snake_case : int = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[int] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 52.5302 ) < 1e-2
assert abs(result_mean.item() - 0.0684 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : str = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.8295 ) < 1e-2
assert abs(result_mean.item() - 0.1951 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : Tuple = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.0784 ) < 1e-2
assert abs(result_mean.item() - 0.1941 ) < 1e-3
| 59
| 1
|
import os
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, BatchEncoding, PreTrainedTokenizer
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = """▁"""
__lowerCamelCase = {"""vocab_file""": """sentencepiece.bpe.model"""}
__lowerCamelCase = {
"""vocab_file""": {
"""facebook/mbart-large-50-one-to-many-mmt""": (
"""https://huggingface.co/facebook/mbart-large-50-one-to-many-mmt/resolve/main/sentencepiece.bpe.model"""
),
}
}
__lowerCamelCase = {
"""facebook/mbart-large-50-one-to-many-mmt""": 10_24,
}
# fmt: off
__lowerCamelCase = ["""ar_AR""", """cs_CZ""", """de_DE""", """en_XX""", """es_XX""", """et_EE""", """fi_FI""", """fr_XX""", """gu_IN""", """hi_IN""", """it_IT""", """ja_XX""", """kk_KZ""", """ko_KR""", """lt_LT""", """lv_LV""", """my_MM""", """ne_NP""", """nl_XX""", """ro_RO""", """ru_RU""", """si_LK""", """tr_TR""", """vi_VN""", """zh_CN""", """af_ZA""", """az_AZ""", """bn_IN""", """fa_IR""", """he_IL""", """hr_HR""", """id_ID""", """ka_GE""", """km_KH""", """mk_MK""", """ml_IN""", """mn_MN""", """mr_IN""", """pl_PL""", """ps_AF""", """pt_XX""", """sv_SE""", """sw_KE""", """ta_IN""", """te_IN""", """th_TH""", """tl_XX""", """uk_UA""", """ur_PK""", """xh_ZA""", """gl_ES""", """sl_SI"""]
class UpperCAmelCase ( A_ ):
A__ : Optional[Any] = VOCAB_FILES_NAMES
A__ : Union[str, Any] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : str = PRETRAINED_VOCAB_FILES_MAP
A__ : str = ["input_ids", "attention_mask"]
A__ : List[int] = []
A__ : List[int] = []
def __init__(self : List[Any] , snake_case__ : Tuple , snake_case__ : Dict=None , snake_case__ : Any=None , snake_case__ : int="</s>" , snake_case__ : List[Any]="</s>" , snake_case__ : Optional[int]="<s>" , snake_case__ : Tuple="<unk>" , snake_case__ : List[str]="<pad>" , snake_case__ : Dict="<mask>" , snake_case__ : Optional[Dict[str, Any]] = None , **snake_case__ : Tuple , ) -> None:
'''simple docstring'''
snake_case : int = AddedToken(snake_case__ , lstrip=snake_case__ , rstrip=snake_case__ ) if isinstance(snake_case__ , snake_case__ ) else mask_token
snake_case : Tuple = {} if sp_model_kwargs is None else sp_model_kwargs
snake_case : Union[str, Any] = kwargs.get("additional_special_tokens" , [] )
kwargs["additional_special_tokens"] += [
code for code in FAIRSEQ_LANGUAGE_CODES if code not in kwargs["additional_special_tokens"]
]
super().__init__(
src_lang=snake_case__ , tgt_lang=snake_case__ , eos_token=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , cls_token=snake_case__ , pad_token=snake_case__ , mask_token=snake_case__ , sp_model_kwargs=self.sp_model_kwargs , **snake_case__ , )
snake_case : Tuple = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(str(snake_case__ ) )
snake_case : Tuple = vocab_file
# Original fairseq vocab and spm vocab must be "aligned":
# Vocab | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9
# -------- | ------- | ------- | ------ | ------- | --- | --- | --- | ----- | ----- | ----
# fairseq | '<s>' | '<pad>' | '</s>' | '<unk>' | ',' | '.' | '▁' | 's' | '▁de' | '-'
# spm | '<unk>' | '<s>' | '</s>' | ',' | '.' | '▁' | 's' | '▁de' | '-' | '▁a'
# Mimic fairseq token-to-id alignment for the first 4 token
snake_case : Optional[int] = {"<s>": 0, "<pad>": 1, "</s>": 2, "<unk>": 3}
# The first "real" token "," has position 4 in the original fairseq vocab and position 3 in the spm vocab
snake_case : List[str] = 1
snake_case : int = len(self.sp_model )
snake_case : Any = {
code: self.sp_model_size + i + self.fairseq_offset for i, code in enumerate(snake_case__ )
}
snake_case : Union[str, Any] = {v: k for k, v in self.lang_code_to_id.items()}
snake_case : List[Any] = len(self.sp_model ) + len(self.lang_code_to_id ) + self.fairseq_offset
self.fairseq_tokens_to_ids.update(self.lang_code_to_id )
snake_case : Any = {v: k for k, v in self.fairseq_tokens_to_ids.items()}
snake_case : str = src_lang if src_lang is not None else "en_XX"
snake_case : Dict = self.lang_code_to_id[self._src_lang]
snake_case : int = tgt_lang
self.set_src_lang_special_tokens(self._src_lang )
@property
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> int:
'''simple docstring'''
return len(self.sp_model ) + len(self.lang_code_to_id ) + self.fairseq_offset + 1 # Plus 1 for the mask token
@property
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
return self._src_lang
@src_lang.setter
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : str ) -> None:
'''simple docstring'''
snake_case : Dict = new_src_lang
self.set_src_lang_special_tokens(self._src_lang )
def __getstate__(self : str ) -> Dict:
'''simple docstring'''
snake_case : Optional[int] = self.__dict__.copy()
snake_case : List[str] = None
return state
def __setstate__(self : str , snake_case__ : Dict ) -> None:
'''simple docstring'''
snake_case : List[Any] = d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs" ):
snake_case : Dict = {}
snake_case : int = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Dict:
'''simple docstring'''
snake_case : Dict = {self.convert_ids_to_tokens(snake_case__ ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : str ) -> List[str]:
'''simple docstring'''
return self.sp_model.encode(snake_case__ , out_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : str ) -> int:
'''simple docstring'''
if token in self.fairseq_tokens_to_ids:
return self.fairseq_tokens_to_ids[token]
snake_case : List[Any] = self.sp_model.PieceToId(snake_case__ )
# Need to return unknown token if the SP model returned 0
return spm_id + self.fairseq_offset if spm_id else self.unk_token_id
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : int ) -> str:
'''simple docstring'''
if index in self.fairseq_ids_to_tokens:
return self.fairseq_ids_to_tokens[index]
return self.sp_model.IdToPiece(index - self.fairseq_offset )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : List[Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : int = []
snake_case : Union[str, Any] = ""
snake_case : Tuple = False
for token in tokens:
# make sure that special tokens are not decoded using sentencepiece model
if token in self.all_special_tokens:
if not prev_is_special:
out_string += " "
out_string += self.sp_model.decode(snake_case__ ) + token
snake_case : List[Any] = True
snake_case : List[Any] = []
else:
current_sub_tokens.append(snake_case__ )
snake_case : Optional[Any] = False
out_string += self.sp_model.decode(snake_case__ )
return out_string.strip()
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
if not os.path.isdir(snake_case__ ):
logger.error(f"""Vocabulary path ({save_directory}) should be a directory""" )
return
snake_case : Any = os.path.join(
snake_case__ , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"] )
if os.path.abspath(self.vocab_file ) != os.path.abspath(snake_case__ ) and os.path.isfile(self.vocab_file ):
copyfile(self.vocab_file , snake_case__ )
elif not os.path.isfile(self.vocab_file ):
with open(snake_case__ , "wb" ) as fi:
snake_case : str = self.sp_model.serialized_model_proto()
fi.write(snake_case__ )
return (out_vocab_file,)
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None , snake_case__ : bool = False ) -> List[int]:
'''simple docstring'''
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=snake_case__ , token_ids_a=snake_case__ , already_has_special_tokens=snake_case__ )
snake_case : Union[str, Any] = [1] * len(self.prefix_tokens )
snake_case : int = [1] * len(self.suffix_tokens )
if token_ids_a is None:
return prefix_ones + ([0] * len(snake_case__ )) + suffix_ones
return prefix_ones + ([0] * len(snake_case__ )) + ([0] * len(snake_case__ )) + suffix_ones
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
if token_ids_a is None:
return self.prefix_tokens + token_ids_a + self.suffix_tokens
# We don't expect to process pairs, but leave the pair logic for API consistency
return self.prefix_tokens + token_ids_a + token_ids_a + self.suffix_tokens
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Dict , snake_case__ : str , snake_case__ : Optional[str] , snake_case__ : Optional[str] , **snake_case__ : Optional[int] ) -> Optional[int]:
'''simple docstring'''
if src_lang is None or tgt_lang is None:
raise ValueError("Translation requires a `src_lang` and a `tgt_lang` for this model" )
snake_case : str = src_lang
snake_case : List[str] = self(snake_case__ , add_special_tokens=snake_case__ , return_tensors=snake_case__ , **snake_case__ )
snake_case : Optional[int] = self.convert_tokens_to_ids(snake_case__ )
snake_case : Optional[int] = tgt_lang_id
return inputs
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : List[str] , snake_case__ : str = "en_XX" , snake_case__ : Optional[List[str]] = None , snake_case__ : str = "ro_RO" , **snake_case__ : Union[str, Any] , ) -> BatchEncoding:
'''simple docstring'''
snake_case : str = src_lang
snake_case : int = tgt_lang
return super().prepare_seqaseq_batch(snake_case__ , snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
return self.set_src_lang_special_tokens(self.src_lang )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> str:
'''simple docstring'''
return self.set_tgt_lang_special_tokens(self.tgt_lang )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : str ) -> None:
'''simple docstring'''
snake_case : List[str] = self.lang_code_to_id[src_lang]
snake_case : List[str] = [self.cur_lang_code_id]
snake_case : Any = [self.eos_token_id]
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : str ) -> None:
'''simple docstring'''
snake_case : Tuple = self.lang_code_to_id[tgt_lang]
snake_case : Union[str, Any] = [self.cur_lang_code_id]
snake_case : Tuple = [self.eos_token_id]
| 59
|
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )]
snake_case : int = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1 or len(__lowerCamelCase ) <= key:
return input_string
for position, character in enumerate(__lowerCamelCase ):
snake_case : Any = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append(__lowerCamelCase )
snake_case : List[str] = ["".join(__lowerCamelCase ) for row in temp_grid]
snake_case : Tuple = "".join(__lowerCamelCase )
return output_string
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : Dict = []
snake_case : Union[str, Any] = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1:
return input_string
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )] # generates template
for position in range(len(__lowerCamelCase ) ):
snake_case : List[str] = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append("*" )
snake_case : Tuple = 0
for row in temp_grid: # fills in the characters
snake_case : Union[str, Any] = input_string[counter : counter + len(__lowerCamelCase )]
grid.append(list(__lowerCamelCase ) )
counter += len(__lowerCamelCase )
snake_case : str = "" # reads as zigzag
for position in range(len(__lowerCamelCase ) ):
snake_case : Optional[int] = position % (lowest * 2) # puts it in bounds
snake_case : Tuple = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
output_string += grid[num][0]
grid[num].pop(0 )
return output_string
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : Tuple = {}
for key_guess in range(1 , len(__lowerCamelCase ) ): # tries every key
snake_case : Any = decrypt(__lowerCamelCase , __lowerCamelCase )
return results
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
import re
import string
from collections import Counter
import sacrebleu
import sacremoses
from packaging import version
import datasets
__lowerCamelCase = """
@inproceedings{xu-etal-2016-optimizing,
title = {Optimizing Statistical Machine Translation for Text Simplification},
authors={Xu, Wei and Napoles, Courtney and Pavlick, Ellie and Chen, Quanze and Callison-Burch, Chris},
journal = {Transactions of the Association for Computational Linguistics},
volume = {4},
year={2016},
url = {https://www.aclweb.org/anthology/Q16-1029},
pages = {401--415
},
@inproceedings{post-2018-call,
title = \"A Call for Clarity in Reporting {BLEU} Scores\",
author = \"Post, Matt\",
booktitle = \"Proceedings of the Third Conference on Machine Translation: Research Papers\",
month = oct,
year = \"2018\",
address = \"Belgium, Brussels\",
publisher = \"Association for Computational Linguistics\",
url = \"https://www.aclweb.org/anthology/W18-6319\",
pages = \"186--191\",
}
"""
__lowerCamelCase = """\
WIKI_SPLIT is the combination of three metrics SARI, EXACT and SACREBLEU
It can be used to evaluate the quality of machine-generated texts.
"""
__lowerCamelCase = """
Calculates sari score (between 0 and 100) given a list of source and predicted
sentences, and a list of lists of reference sentences. It also computes the BLEU score as well as the exact match score.
Args:
sources: list of source sentences where each sentence should be a string.
predictions: list of predicted sentences where each sentence should be a string.
references: list of lists of reference sentences where each sentence should be a string.
Returns:
sari: sari score
sacrebleu: sacrebleu score
exact: exact score
Examples:
>>> sources=[\"About 95 species are currently accepted .\"]
>>> predictions=[\"About 95 you now get in .\"]
>>> references=[[\"About 95 species are currently known .\"]]
>>> wiki_split = datasets.load_metric(\"wiki_split\")
>>> results = wiki_split.compute(sources=sources, predictions=predictions, references=references)
>>> print(results)
{'sari': 21.805555555555557, 'sacrebleu': 14.535768424205482, 'exact': 0.0}
"""
def UpperCamelCase ( __lowerCamelCase : int ):
def remove_articles(__lowerCamelCase : Dict ):
snake_case : Dict = re.compile(r"\b(a|an|the)\b" , re.UNICODE )
return re.sub(__lowerCamelCase , " " , __lowerCamelCase )
def white_space_fix(__lowerCamelCase : Dict ):
return " ".join(text.split() )
def remove_punc(__lowerCamelCase : Optional[int] ):
snake_case : List[Any] = set(string.punctuation )
return "".join(ch for ch in text if ch not in exclude )
def lower(__lowerCamelCase : Optional[Any] ):
return text.lower()
return white_space_fix(remove_articles(remove_punc(lower(__lowerCamelCase ) ) ) )
def UpperCamelCase ( __lowerCamelCase : Any , __lowerCamelCase : Dict ):
return int(normalize_answer(__lowerCamelCase ) == normalize_answer(__lowerCamelCase ) )
def UpperCamelCase ( __lowerCamelCase : List[Any] , __lowerCamelCase : List[str] ):
snake_case : Optional[int] = [any(compute_exact(__lowerCamelCase , __lowerCamelCase ) for ref in refs ) for pred, refs in zip(__lowerCamelCase , __lowerCamelCase )]
return (sum(__lowerCamelCase ) / len(__lowerCamelCase )) * 100
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : List[str] , __lowerCamelCase : Optional[Any] , __lowerCamelCase : List[Any] ):
snake_case : Any = [rgram for rgrams in rgramslist for rgram in rgrams]
snake_case : Optional[int] = Counter(__lowerCamelCase )
snake_case : Union[str, Any] = Counter(__lowerCamelCase )
snake_case : List[Any] = Counter()
for sgram, scount in sgramcounter.items():
snake_case : Optional[int] = scount * numref
snake_case : Dict = Counter(__lowerCamelCase )
snake_case : Optional[int] = Counter()
for cgram, ccount in cgramcounter.items():
snake_case : int = ccount * numref
# KEEP
snake_case : List[Any] = sgramcounter_rep & cgramcounter_rep
snake_case : Optional[Any] = keepgramcounter_rep & rgramcounter
snake_case : List[Any] = sgramcounter_rep & rgramcounter
snake_case : int = 0
snake_case : List[Any] = 0
for keepgram in keepgramcountergood_rep:
keeptmpscorea += keepgramcountergood_rep[keepgram] / keepgramcounter_rep[keepgram]
# Fix an alleged bug [2] in the keep score computation.
# keeptmpscore2 += keepgramcountergood_rep[keepgram] / keepgramcounterall_rep[keepgram]
keeptmpscorea += keepgramcountergood_rep[keepgram]
# Define 0/0=1 instead of 0 to give higher scores for predictions that match
# a target exactly.
snake_case : Optional[int] = 1
snake_case : Optional[Any] = 1
if len(__lowerCamelCase ) > 0:
snake_case : Tuple = keeptmpscorea / len(__lowerCamelCase )
if len(__lowerCamelCase ) > 0:
# Fix an alleged bug [2] in the keep score computation.
# keepscore_recall = keeptmpscore2 / len(keepgramcounterall_rep)
snake_case : List[Any] = keeptmpscorea / sum(keepgramcounterall_rep.values() )
snake_case : int = 0
if keepscore_precision > 0 or keepscore_recall > 0:
snake_case : List[Any] = 2 * keepscore_precision * keepscore_recall / (keepscore_precision + keepscore_recall)
# DELETION
snake_case : Optional[Any] = sgramcounter_rep - cgramcounter_rep
snake_case : Optional[int] = delgramcounter_rep - rgramcounter
snake_case : Optional[int] = sgramcounter_rep - rgramcounter
snake_case : int = 0
snake_case : Tuple = 0
for delgram in delgramcountergood_rep:
deltmpscorea += delgramcountergood_rep[delgram] / delgramcounter_rep[delgram]
deltmpscorea += delgramcountergood_rep[delgram] / delgramcounterall_rep[delgram]
# Define 0/0=1 instead of 0 to give higher scores for predictions that match
# a target exactly.
snake_case : Any = 1
if len(__lowerCamelCase ) > 0:
snake_case : Tuple = deltmpscorea / len(__lowerCamelCase )
# ADDITION
snake_case : Any = set(__lowerCamelCase ) - set(__lowerCamelCase )
snake_case : str = set(__lowerCamelCase ) & set(__lowerCamelCase )
snake_case : str = set(__lowerCamelCase ) - set(__lowerCamelCase )
snake_case : int = 0
for addgram in addgramcountergood:
addtmpscore += 1
# Define 0/0=1 instead of 0 to give higher scores for predictions that match
# a target exactly.
snake_case : List[str] = 1
snake_case : List[Any] = 1
if len(__lowerCamelCase ) > 0:
snake_case : Union[str, Any] = addtmpscore / len(__lowerCamelCase )
if len(__lowerCamelCase ) > 0:
snake_case : Tuple = addtmpscore / len(__lowerCamelCase )
snake_case : Optional[Any] = 0
if addscore_precision > 0 or addscore_recall > 0:
snake_case : str = 2 * addscore_precision * addscore_recall / (addscore_precision + addscore_recall)
return (keepscore, delscore_precision, addscore)
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Any , __lowerCamelCase : Optional[Any] ):
snake_case : List[str] = len(__lowerCamelCase )
snake_case : Optional[int] = ssent.split(" " )
snake_case : Optional[int] = csent.split(" " )
snake_case : Any = []
snake_case : Dict = []
snake_case : Union[str, Any] = []
snake_case : List[str] = []
snake_case : Any = []
snake_case : List[Any] = []
snake_case : Tuple = []
snake_case : str = []
snake_case : int = []
snake_case : str = []
for rsent in rsents:
snake_case : Tuple = rsent.split(" " )
snake_case : Dict = []
snake_case : int = []
snake_case : Optional[Any] = []
ragramslist.append(__lowerCamelCase )
for i in range(0 , len(__lowerCamelCase ) - 1 ):
if i < len(__lowerCamelCase ) - 1:
snake_case : Any = ragrams[i] + " " + ragrams[i + 1]
ragrams.append(__lowerCamelCase )
if i < len(__lowerCamelCase ) - 2:
snake_case : str = ragrams[i] + " " + ragrams[i + 1] + " " + ragrams[i + 2]
ragrams.append(__lowerCamelCase )
if i < len(__lowerCamelCase ) - 3:
snake_case : List[Any] = ragrams[i] + " " + ragrams[i + 1] + " " + ragrams[i + 2] + " " + ragrams[i + 3]
ragrams.append(__lowerCamelCase )
ragramslist.append(__lowerCamelCase )
ragramslist.append(__lowerCamelCase )
ragramslist.append(__lowerCamelCase )
for i in range(0 , len(__lowerCamelCase ) - 1 ):
if i < len(__lowerCamelCase ) - 1:
snake_case : List[str] = sagrams[i] + " " + sagrams[i + 1]
sagrams.append(__lowerCamelCase )
if i < len(__lowerCamelCase ) - 2:
snake_case : Dict = sagrams[i] + " " + sagrams[i + 1] + " " + sagrams[i + 2]
sagrams.append(__lowerCamelCase )
if i < len(__lowerCamelCase ) - 3:
snake_case : Any = sagrams[i] + " " + sagrams[i + 1] + " " + sagrams[i + 2] + " " + sagrams[i + 3]
sagrams.append(__lowerCamelCase )
for i in range(0 , len(__lowerCamelCase ) - 1 ):
if i < len(__lowerCamelCase ) - 1:
snake_case : Tuple = cagrams[i] + " " + cagrams[i + 1]
cagrams.append(__lowerCamelCase )
if i < len(__lowerCamelCase ) - 2:
snake_case : Optional[Any] = cagrams[i] + " " + cagrams[i + 1] + " " + cagrams[i + 2]
cagrams.append(__lowerCamelCase )
if i < len(__lowerCamelCase ) - 3:
snake_case : Any = cagrams[i] + " " + cagrams[i + 1] + " " + cagrams[i + 2] + " " + cagrams[i + 3]
cagrams.append(__lowerCamelCase )
((snake_case) , (snake_case) , (snake_case)) : int = SARIngram(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
((snake_case) , (snake_case) , (snake_case)) : Optional[int] = SARIngram(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
((snake_case) , (snake_case) , (snake_case)) : int = SARIngram(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
((snake_case) , (snake_case) , (snake_case)) : Tuple = SARIngram(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
snake_case : List[Any] = sum([keepascore, keepascore, keepascore, keepascore] ) / 4
snake_case : Union[str, Any] = sum([delascore, delascore, delascore, delascore] ) / 4
snake_case : Tuple = sum([addascore, addascore, addascore, addascore] ) / 4
snake_case : List[Any] = (avgkeepscore + avgdelscore + avgaddscore) / 3
return finalscore
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : bool = True , __lowerCamelCase : str = "13a" , __lowerCamelCase : bool = True ):
# Normalization is requried for the ASSET dataset (one of the primary
# datasets in sentence simplification) to allow using space
# to split the sentence. Even though Wiki-Auto and TURK datasets,
# do not require normalization, we do it for consistency.
# Code adapted from the EASSE library [1] written by the authors of the ASSET dataset.
# [1] https://github.com/feralvam/easse/blob/580bba7e1378fc8289c663f864e0487188fe8067/easse/utils/preprocessing.py#L7
if lowercase:
snake_case : str = sentence.lower()
if tokenizer in ["13a", "intl"]:
if version.parse(sacrebleu.__version__ ).major >= 2:
snake_case : Dict = sacrebleu.metrics.bleu._get_tokenizer(__lowerCamelCase )()(__lowerCamelCase )
else:
snake_case : List[Any] = sacrebleu.TOKENIZERS[tokenizer]()(__lowerCamelCase )
elif tokenizer == "moses":
snake_case : List[Any] = sacremoses.MosesTokenizer().tokenize(__lowerCamelCase , return_str=__lowerCamelCase , escape=__lowerCamelCase )
elif tokenizer == "penn":
snake_case : Union[str, Any] = sacremoses.MosesTokenizer().penn_tokenize(__lowerCamelCase , return_str=__lowerCamelCase )
else:
snake_case : List[str] = sentence
if not return_str:
snake_case : Any = normalized_sent.split()
return normalized_sent
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Optional[Any] , __lowerCamelCase : Tuple ):
if not (len(__lowerCamelCase ) == len(__lowerCamelCase ) == len(__lowerCamelCase )):
raise ValueError("Sources length must match predictions and references lengths." )
snake_case : int = 0
for src, pred, refs in zip(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ):
sari_score += SARIsent(normalize(__lowerCamelCase ) , normalize(__lowerCamelCase ) , [normalize(__lowerCamelCase ) for sent in refs] )
snake_case : Optional[int] = sari_score / len(__lowerCamelCase )
return 100 * sari_score
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Optional[int] , __lowerCamelCase : List[Any]="exp" , __lowerCamelCase : List[str]=None , __lowerCamelCase : Optional[Any]=False , __lowerCamelCase : Union[str, Any]=False , __lowerCamelCase : Optional[int]=False , ):
snake_case : Dict = len(references[0] )
if any(len(__lowerCamelCase ) != references_per_prediction for refs in references ):
raise ValueError("Sacrebleu requires the same number of references for each prediction" )
snake_case : Union[str, Any] = [[refs[i] for refs in references] for i in range(__lowerCamelCase )]
snake_case : Optional[int] = sacrebleu.corpus_bleu(
__lowerCamelCase , __lowerCamelCase , smooth_method=__lowerCamelCase , smooth_value=__lowerCamelCase , force=__lowerCamelCase , lowercase=__lowerCamelCase , use_effective_order=__lowerCamelCase , )
return output.score
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION ,_KWARGS_DESCRIPTION )
class UpperCAmelCase ( datasets.Metric ):
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Optional[int]:
'''simple docstring'''
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
"predictions": datasets.Value("string" , id="sequence" ),
"references": datasets.Sequence(datasets.Value("string" , id="sequence" ) , id="references" ),
} ) , codebase_urls=[
"https://github.com/huggingface/transformers/blob/master/src/transformers/data/metrics/squad_metrics.py",
"https://github.com/cocoxu/simplification/blob/master/SARI.py",
"https://github.com/tensorflow/tensor2tensor/blob/master/tensor2tensor/utils/sari_hook.py",
"https://github.com/mjpost/sacreBLEU",
] , reference_urls=[
"https://www.aclweb.org/anthology/Q16-1029.pdf",
"https://github.com/mjpost/sacreBLEU",
"https://en.wikipedia.org/wiki/BLEU",
"https://towardsdatascience.com/evaluating-text-output-in-nlp-bleu-at-your-own-risk-e8609665a213",
] , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : List[Any] , snake_case__ : Optional[int] , snake_case__ : List[Any] ) -> int:
'''simple docstring'''
snake_case : str = {}
result.update({"sari": compute_sari(sources=snake_case__ , predictions=snake_case__ , references=snake_case__ )} )
result.update({"sacrebleu": compute_sacrebleu(predictions=snake_case__ , references=snake_case__ )} )
result.update({"exact": compute_em(predictions=snake_case__ , references=snake_case__ )} )
return result
| 59
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
if is_sentencepiece_available():
from ..ta.tokenization_ta import TaTokenizer
else:
from ...utils.dummy_sentencepiece_objects import TaTokenizer
__lowerCamelCase = TaTokenizer
if is_tokenizers_available():
from ..ta.tokenization_ta_fast import TaTokenizerFast
else:
from ...utils.dummy_tokenizers_objects import TaTokenizerFast
__lowerCamelCase = TaTokenizerFast
__lowerCamelCase = {"""configuration_mt5""": ["""MT5Config""", """MT5OnnxConfig"""]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""MT5EncoderModel""",
"""MT5ForConditionalGeneration""",
"""MT5ForQuestionAnswering""",
"""MT5Model""",
"""MT5PreTrainedModel""",
"""MT5Stack""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""TFMT5EncoderModel""", """TFMT5ForConditionalGeneration""", """TFMT5Model"""]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""FlaxMT5EncoderModel""", """FlaxMT5ForConditionalGeneration""", """FlaxMT5Model"""]
if TYPE_CHECKING:
from .configuration_mta import MTaConfig, MTaOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mta import (
MTaEncoderModel,
MTaForConditionalGeneration,
MTaForQuestionAnswering,
MTaModel,
MTaPreTrainedModel,
MTaStack,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel
else:
import sys
__lowerCamelCase = _LazyModule(
__name__,
globals()["""__file__"""],
_import_structure,
extra_objects={"""MT5Tokenizer""": MTaTokenizer, """MT5TokenizerFast""": MTaTokenizerFast},
module_spec=__spec__,
)
| 59
| 1
|
from __future__ import annotations
import csv
import requests
from bsa import BeautifulSoup
def UpperCamelCase ( __lowerCamelCase : str = "" ):
snake_case : Dict = url or "https://www.imdb.com/chart/top/?ref_=nv_mv_250"
snake_case : Union[str, Any] = BeautifulSoup(requests.get(__lowerCamelCase ).text , "html.parser" )
snake_case : str = soup.find_all("td" , attrs="titleColumn" )
snake_case : Union[str, Any] = soup.find_all("td" , class_="ratingColumn imdbRating" )
return {
title.a.text: float(rating.strong.text )
for title, rating in zip(__lowerCamelCase , __lowerCamelCase )
}
def UpperCamelCase ( __lowerCamelCase : str = "IMDb_Top_250_Movies.csv" ):
snake_case : List[str] = get_imdb_top_aaa_movies()
with open(__lowerCamelCase , "w" , newline="" ) as out_file:
snake_case : Union[str, Any] = csv.writer(__lowerCamelCase )
writer.writerow(["Movie title", "IMDb rating"] )
for title, rating in movies.items():
writer.writerow([title, rating] )
if __name__ == "__main__":
write_movies()
| 59
|
import os
import shutil
from pathlib import Path
from typing import Optional, Union
import numpy as np
from huggingface_hub import hf_hub_download
from ..utils import ONNX_EXTERNAL_WEIGHTS_NAME, ONNX_WEIGHTS_NAME, is_onnx_available, logging
if is_onnx_available():
import onnxruntime as ort
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""tensor(bool)""": np.bool_,
"""tensor(int8)""": np.inta,
"""tensor(uint8)""": np.uinta,
"""tensor(int16)""": np.intaa,
"""tensor(uint16)""": np.uintaa,
"""tensor(int32)""": np.intaa,
"""tensor(uint32)""": np.uintaa,
"""tensor(int64)""": np.intaa,
"""tensor(uint64)""": np.uintaa,
"""tensor(float16)""": np.floataa,
"""tensor(float)""": np.floataa,
"""tensor(double)""": np.floataa,
}
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=None , **snake_case__ : Optional[Any] ) -> List[str]:
'''simple docstring'''
logger.info("`diffusers.OnnxRuntimeModel` is experimental and might change in the future." )
snake_case : Optional[Any] = model
snake_case : Dict = kwargs.get("model_save_dir" , snake_case__ )
snake_case : int = kwargs.get("latest_model_name" , snake_case__ )
def __call__(self : Tuple , **snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Union[str, Any] = {k: np.array(snake_case__ ) for k, v in kwargs.items()}
return self.model.run(snake_case__ , snake_case__ )
@staticmethod
def _SCREAMING_SNAKE_CASE (snake_case__ : Union[str, Path] , snake_case__ : Optional[int]=None , snake_case__ : Optional[int]=None ) -> Any:
'''simple docstring'''
if provider is None:
logger.info("No onnxruntime provider specified, using CPUExecutionProvider" )
snake_case : Optional[int] = "CPUExecutionProvider"
return ort.InferenceSession(snake_case__ , providers=[provider] , sess_options=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Union[str, Path] , snake_case__ : Optional[str] = None , **snake_case__ : Any ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = file_name if file_name is not None else ONNX_WEIGHTS_NAME
snake_case : Any = self.model_save_dir.joinpath(self.latest_model_name )
snake_case : str = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
snake_case : List[str] = self.model_save_dir.joinpath(snake_case__ )
if src_path.exists():
snake_case : Tuple = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, os.PathLike] , **snake_case__ : Optional[int] , ) -> str:
'''simple docstring'''
if os.path.isfile(snake_case__ ):
logger.error(f"""Provided path ({save_directory}) should be a directory, not a file""" )
return
os.makedirs(snake_case__ , exist_ok=snake_case__ )
# saving model weights/files
self._save_pretrained(snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Tuple , snake_case__ : Union[str, Path] , snake_case__ : Optional[Union[bool, str, None]] = None , snake_case__ : Optional[Union[str, None]] = None , snake_case__ : bool = False , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional["ort.SessionOptions"] = None , **snake_case__ : Tuple , ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(snake_case__ ):
snake_case : Any = OnnxRuntimeModel.load_model(
os.path.join(snake_case__ , snake_case__ ) , provider=snake_case__ , sess_options=snake_case__ )
snake_case : Union[str, Any] = Path(snake_case__ )
# load model from hub
else:
# download model
snake_case : Dict = hf_hub_download(
repo_id=snake_case__ , filename=snake_case__ , use_auth_token=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , )
snake_case : List[Any] = Path(snake_case__ ).parent
snake_case : Union[str, Any] = Path(snake_case__ ).name
snake_case : Dict = OnnxRuntimeModel.load_model(snake_case__ , provider=snake_case__ , sess_options=snake_case__ )
return cls(model=snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : Union[str, Path] , snake_case__ : bool = True , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , **snake_case__ : Dict , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = None
if len(str(snake_case__ ).split("@" ) ) == 2:
snake_case , snake_case : int = model_id.split("@" )
return cls._from_pretrained(
model_id=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , use_auth_token=snake_case__ , **snake_case__ , )
| 59
| 1
|
import math
from numpy import inf
from scipy.integrate import quad
def UpperCamelCase ( __lowerCamelCase : float ):
if num <= 0:
raise ValueError("math domain error" )
return quad(__lowerCamelCase , 0 , __lowerCamelCase , args=(__lowerCamelCase) )[0]
def UpperCamelCase ( __lowerCamelCase : float , __lowerCamelCase : float ):
return math.pow(__lowerCamelCase , z - 1 ) * math.exp(-x )
if __name__ == "__main__":
from doctest import testmod
testmod()
| 59
|
import argparse
import json
from dataclasses import dataclass, field
from functools import partial
from pathlib import Path
from typing import Callable, Dict, List, Tuple
import timm
import torch
import torch.nn as nn
from classy_vision.models.regnet import RegNet, RegNetParams, RegNetYaagf, RegNetYaagf, RegNetYaaagf
from huggingface_hub import cached_download, hf_hub_url
from torch import Tensor
from vissl.models.model_helpers import get_trunk_forward_outputs
from transformers import AutoImageProcessor, RegNetConfig, RegNetForImageClassification, RegNetModel
from transformers.utils import logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger()
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : List[nn.Module] = field(default_factory=A_ )
A__ : list = field(default_factory=A_ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : Tensor , snake_case__ : Tensor ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = len(list(m.modules() ) ) == 1 or isinstance(snake_case__ , nn.Convad ) or isinstance(snake_case__ , nn.BatchNormad )
if has_not_submodules:
self.traced.append(snake_case__ )
def __call__(self : List[Any] , snake_case__ : Tensor ) -> List[Any]:
'''simple docstring'''
for m in self.module.modules():
self.handles.append(m.register_forward_hook(self._forward_hook ) )
self.module(snake_case__ )
[x.remove() for x in self.handles]
return self
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[int]:
'''simple docstring'''
return list(filter(lambda snake_case__ : len(list(x.state_dict().keys() ) ) > 0 , self.traced ) )
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : nn.Module
A__ : int = 1
A__ : List = field(default_factory=A_ )
A__ : List = field(default_factory=A_ )
A__ : bool = True
def __call__(self : List[Any] , snake_case__ : Tensor ) -> Any:
'''simple docstring'''
snake_case : str = Tracker(self.dest )(snake_case__ ).parametrized
snake_case : Optional[int] = Tracker(self.src )(snake_case__ ).parametrized
snake_case : List[str] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.src_skip , snake_case__ ) )
snake_case : Optional[Any] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.dest_skip , snake_case__ ) )
if len(snake_case__ ) != len(snake_case__ ) and self.raise_if_mismatch:
raise Exception(
f"""Numbers of operations are different. Source module has {len(snake_case__ )} operations while"""
f""" destination module has {len(snake_case__ )}.""" )
for dest_m, src_m in zip(snake_case__ , snake_case__ ):
dest_m.load_state_dict(src_m.state_dict() )
if self.verbose == 1:
print(f"""Transfered from={src_m} to={dest_m}""" )
class UpperCAmelCase ( nn.Module ):
def __init__(self : Tuple , snake_case__ : nn.Module ) -> Optional[Any]:
'''simple docstring'''
super().__init__()
snake_case : List[Tuple[str, nn.Module]] = []
# - get the stem
feature_blocks.append(("conv1", model.stem) )
# - get all the feature blocks
for k, v in model.trunk_output.named_children():
assert k.startswith("block" ), f"""Unexpected layer name {k}"""
snake_case : Union[str, Any] = len(snake_case__ ) + 1
feature_blocks.append((f"""res{block_index}""", v) )
snake_case : Optional[Any] = nn.ModuleDict(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Tensor ) -> Dict:
'''simple docstring'''
return get_trunk_forward_outputs(
snake_case__ , out_feat_keys=snake_case__ , feature_blocks=self._feature_blocks , )
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> str:
'''simple docstring'''
snake_case : List[Any] = x.split("-" )
return x_split[0] + x_split[1] + "_" + "".join(x_split[2:] )
def __getitem__(self : Optional[int] , snake_case__ : str ) -> Callable[[], Tuple[nn.Module, Dict]]:
'''simple docstring'''
if x not in self:
snake_case : Dict = self.convert_name_to_timm(snake_case__ )
snake_case : Union[str, Any] = partial(lambda: (timm.create_model(snake_case__ , pretrained=snake_case__ ).eval(), None) )
else:
snake_case : List[str] = super().__getitem__(snake_case__ )
return val
class UpperCAmelCase ( A_ ):
def __getitem__(self : Dict , snake_case__ : str ) -> Callable[[], nn.Module]:
'''simple docstring'''
if "seer" in x and "in1k" not in x:
snake_case : str = RegNetModel
else:
snake_case : Optional[Any] = RegNetForImageClassification
return val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Tuple[str, str]] ):
for from_key, to_key in keys:
snake_case : str = from_state_dict[from_key].clone()
print(f"""Copied key={from_key} to={to_key}""" )
return to_state_dict
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : RegNetConfig , __lowerCamelCase : Path , __lowerCamelCase : bool = True , ):
print(f"""Converting {name}...""" )
with torch.no_grad():
snake_case , snake_case : int = from_model_func()
snake_case : str = our_model_func(__lowerCamelCase ).eval()
snake_case : int = ModuleTransfer(src=__lowerCamelCase , dest=__lowerCamelCase , raise_if_mismatch=__lowerCamelCase )
snake_case : Dict = torch.randn((1, 3, 224, 224) )
module_transfer(__lowerCamelCase )
if from_state_dict is not None:
snake_case : str = []
# for seer - in1k finetuned we have to manually copy the head
if "seer" in name and "in1k" in name:
snake_case : Tuple = [("0.clf.0.weight", "classifier.1.weight"), ("0.clf.0.bias", "classifier.1.bias")]
snake_case : Optional[Any] = manually_copy_vissl_head(__lowerCamelCase , our_model.state_dict() , __lowerCamelCase )
our_model.load_state_dict(__lowerCamelCase )
snake_case : Any = our_model(__lowerCamelCase , output_hidden_states=__lowerCamelCase )
snake_case : Union[str, Any] = (
our_outputs.logits if isinstance(__lowerCamelCase , __lowerCamelCase ) else our_outputs.last_hidden_state
)
snake_case : Union[str, Any] = from_model(__lowerCamelCase )
snake_case : Dict = from_output[-1] if type(__lowerCamelCase ) is list else from_output
# now since I don't want to use any config files, vissl seer model doesn't actually have an head, so let's just check the last hidden state
if "seer" in name and "in1k" in name:
snake_case : Any = our_outputs.hidden_states[-1]
assert torch.allclose(__lowerCamelCase , __lowerCamelCase ), "The model logits don't match the original one."
if push_to_hub:
our_model.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add model" , use_temp_dir=__lowerCamelCase , )
snake_case : List[str] = 224 if "seer" not in name else 384
# we can use the convnext one
snake_case : int = AutoImageProcessor.from_pretrained("facebook/convnext-base-224-22k-1k" , size=__lowerCamelCase )
image_processor.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add image processor" , use_temp_dir=__lowerCamelCase , )
print(f"""Pushed {name}""" )
def UpperCamelCase ( __lowerCamelCase : Path , __lowerCamelCase : str = None , __lowerCamelCase : bool = True ):
snake_case : Union[str, Any] = "imagenet-1k-id2label.json"
snake_case : List[str] = 1000
snake_case : List[str] = (1, num_labels)
snake_case : Any = "huggingface/label-files"
snake_case : List[str] = num_labels
snake_case : Optional[Any] = json.load(open(cached_download(hf_hub_url(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) ) , "r" ) )
snake_case : List[Any] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : str = idalabel
snake_case : List[Any] = {v: k for k, v in idalabel.items()}
snake_case : Dict = partial(__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase )
snake_case : Optional[Any] = {
"regnet-x-002": ImageNetPreTrainedConfig(
depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 , layer_type="x" ),
"regnet-x-004": ImageNetPreTrainedConfig(
depths=[1, 2, 7, 12] , hidden_sizes=[32, 64, 160, 384] , groups_width=16 , layer_type="x" ),
"regnet-x-006": ImageNetPreTrainedConfig(
depths=[1, 3, 5, 7] , hidden_sizes=[48, 96, 240, 528] , groups_width=24 , layer_type="x" ),
"regnet-x-008": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 5] , hidden_sizes=[64, 128, 288, 672] , groups_width=16 , layer_type="x" ),
"regnet-x-016": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 2] , hidden_sizes=[72, 168, 408, 912] , groups_width=24 , layer_type="x" ),
"regnet-x-032": ImageNetPreTrainedConfig(
depths=[2, 6, 15, 2] , hidden_sizes=[96, 192, 432, 1008] , groups_width=48 , layer_type="x" ),
"regnet-x-040": ImageNetPreTrainedConfig(
depths=[2, 5, 14, 2] , hidden_sizes=[80, 240, 560, 1360] , groups_width=40 , layer_type="x" ),
"regnet-x-064": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 392, 784, 1624] , groups_width=56 , layer_type="x" ),
"regnet-x-080": ImageNetPreTrainedConfig(
depths=[2, 5, 15, 1] , hidden_sizes=[80, 240, 720, 1920] , groups_width=120 , layer_type="x" ),
"regnet-x-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 , layer_type="x" ),
"regnet-x-160": ImageNetPreTrainedConfig(
depths=[2, 6, 13, 1] , hidden_sizes=[256, 512, 896, 2048] , groups_width=128 , layer_type="x" ),
"regnet-x-320": ImageNetPreTrainedConfig(
depths=[2, 7, 13, 1] , hidden_sizes=[336, 672, 1344, 2520] , groups_width=168 , layer_type="x" ),
# y variant
"regnet-y-002": ImageNetPreTrainedConfig(depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 ),
"regnet-y-004": ImageNetPreTrainedConfig(
depths=[1, 3, 6, 6] , hidden_sizes=[48, 104, 208, 440] , groups_width=8 ),
"regnet-y-006": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 4] , hidden_sizes=[48, 112, 256, 608] , groups_width=16 ),
"regnet-y-008": ImageNetPreTrainedConfig(
depths=[1, 3, 8, 2] , hidden_sizes=[64, 128, 320, 768] , groups_width=16 ),
"regnet-y-016": ImageNetPreTrainedConfig(
depths=[2, 6, 17, 2] , hidden_sizes=[48, 120, 336, 888] , groups_width=24 ),
"regnet-y-032": ImageNetPreTrainedConfig(
depths=[2, 5, 13, 1] , hidden_sizes=[72, 216, 576, 1512] , groups_width=24 ),
"regnet-y-040": ImageNetPreTrainedConfig(
depths=[2, 6, 12, 2] , hidden_sizes=[128, 192, 512, 1088] , groups_width=64 ),
"regnet-y-064": ImageNetPreTrainedConfig(
depths=[2, 7, 14, 2] , hidden_sizes=[144, 288, 576, 1296] , groups_width=72 ),
"regnet-y-080": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 448, 896, 2016] , groups_width=56 ),
"regnet-y-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 ),
"regnet-y-160": ImageNetPreTrainedConfig(
depths=[2, 4, 11, 1] , hidden_sizes=[224, 448, 1232, 3024] , groups_width=112 ),
"regnet-y-320": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
# models created by SEER -> https://arxiv.org/abs/2202.08360
"regnet-y-320-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer": RegNetConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer": RegNetConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
# finetuned on imagenet
"regnet-y-320-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer-in1k": ImageNetPreTrainedConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
}
snake_case : Union[str, Any] = NameToOurModelFuncMap()
snake_case : str = NameToFromModelFuncMap()
# add seer weights logic
def load_using_classy_vision(__lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] ) -> Tuple[nn.Module, Dict]:
snake_case : List[Any] = torch.hub.load_state_dict_from_url(__lowerCamelCase , model_dir=str(__lowerCamelCase ) , map_location="cpu" )
snake_case : Dict = model_func()
# check if we have a head, if yes add it
snake_case : str = files["classy_state_dict"]["base_model"]["model"]
snake_case : Dict = model_state_dict["trunk"]
model.load_state_dict(__lowerCamelCase )
return model.eval(), model_state_dict["heads"]
# pretrained
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet32d/seer_regnet32gf_model_iteration244000.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Optional[int] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet64/seer_regnet64gf_model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : List[str] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/swav_ig1b_regnet128Gf_cnstant_bs32_node16_sinkhorn10_proto16k_syncBN64_warmup8k/model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet10B/model_iteration124500_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
# IN1K finetuned
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet32_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet64_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : str = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet128_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Dict = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_10b_finetuned_in1k_model_phase28_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
if model_name:
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , names_to_config[model_name] , __lowerCamelCase , __lowerCamelCase , )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , )
return config, expected_shape
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help=(
"""The name of the model you wish to convert, it must be one of the supported regnet* architecture,"""
""" currently: regnetx-*, regnety-*. If `None`, all of them will the converted."""
),
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default=None,
type=Path,
required=True,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument(
"""--push_to_hub""",
default=True,
type=bool,
required=False,
help="""If True, push model and image processor to the hub.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 59
| 1
|
from ...processing_utils import ProcessorMixin
from ...tokenization_utils_base import BatchEncoding
class UpperCAmelCase ( A_ ):
A__ : Optional[Any] = ["image_processor", "tokenizer"]
A__ : Tuple = "AutoImageProcessor"
A__ : Optional[int] = "AutoTokenizer"
def __init__(self : str , snake_case__ : List[str] , snake_case__ : int ) -> Tuple:
'''simple docstring'''
super().__init__(snake_case__ , snake_case__ )
snake_case : int = self.image_processor
def __call__(self : Optional[Any] , snake_case__ : Dict=None , snake_case__ : Tuple=None , snake_case__ : Dict=None , **snake_case__ : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
if text is None and images is None:
raise ValueError("You have to specify either text or images. Both cannot be none." )
if text is not None:
snake_case : Tuple = self.tokenizer(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
if images is not None:
snake_case : List[str] = self.image_processor(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
if text is not None and images is not None:
snake_case : Tuple = image_features.pixel_values
return encoding
elif text is not None:
return encoding
else:
return BatchEncoding(data=dict(**snake_case__ ) , tensor_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict , *snake_case__ : Any , **snake_case__ : str ) -> Tuple:
'''simple docstring'''
return self.tokenizer.batch_decode(*snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Dict , *snake_case__ : Union[str, Any] , **snake_case__ : int ) -> Tuple:
'''simple docstring'''
return self.tokenizer.decode(*snake_case__ , **snake_case__ )
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
return ["input_ids", "attention_mask", "pixel_values"]
| 59
|
import warnings
from typing import Dict
import numpy as np
from ..utils import ExplicitEnum, add_end_docstrings, is_tf_available, is_torch_available
from .base import PIPELINE_INIT_ARGS, GenericTensor, Pipeline
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
return 1.0 / (1.0 + np.exp(-_outputs ))
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Tuple = np.max(_outputs , axis=-1 , keepdims=__lowerCamelCase )
snake_case : int = np.exp(_outputs - maxes )
return shifted_exp / shifted_exp.sum(axis=-1 , keepdims=__lowerCamelCase )
class UpperCAmelCase ( A_ ):
A__ : Any = "sigmoid"
A__ : str = "softmax"
A__ : int = "none"
@add_end_docstrings(
A_ ,r"\n return_all_scores (`bool`, *optional*, defaults to `False`):\n Whether to return all prediction scores or just the one of the predicted class.\n function_to_apply (`str`, *optional*, defaults to `\"default\"`):\n The function to apply to the model outputs in order to retrieve the scores. Accepts four different values:\n\n - `\"default\"`: if the model has a single label, will apply the sigmoid function on the output. If the model\n has several labels, will apply the softmax function on the output.\n - `\"sigmoid\"`: Applies the sigmoid function on the output.\n - `\"softmax\"`: Applies the softmax function on the output.\n - `\"none\"`: Does not apply any function on the output.\n " ,)
class UpperCAmelCase ( A_ ):
A__ : int = False
A__ : Union[str, Any] = ClassificationFunction.NONE
def __init__(self : List[str] , **snake_case__ : int ) -> str:
'''simple docstring'''
super().__init__(**snake_case__ )
self.check_model_type(
TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if self.framework == "tf"
else MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str]=None , snake_case__ : Optional[Any]=None , snake_case__ : Union[str, Any]="" , **snake_case__ : List[str] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = tokenizer_kwargs
snake_case : List[Any] = {}
if hasattr(self.model.config , "return_all_scores" ) and return_all_scores is None:
snake_case : Optional[int] = self.model.config.return_all_scores
if isinstance(snake_case__ , snake_case__ ) or top_k is None:
snake_case : List[Any] = top_k
snake_case : str = False
elif return_all_scores is not None:
warnings.warn(
"`return_all_scores` is now deprecated, if want a similar functionality use `top_k=None` instead of"
" `return_all_scores=True` or `top_k=1` instead of `return_all_scores=False`." , snake_case__ , )
if return_all_scores:
snake_case : List[str] = None
else:
snake_case : Optional[int] = 1
if isinstance(snake_case__ , snake_case__ ):
snake_case : Dict = ClassificationFunction[function_to_apply.upper()]
if function_to_apply is not None:
snake_case : Optional[int] = function_to_apply
return preprocess_params, {}, postprocess_params
def __call__(self : Dict , *snake_case__ : List[str] , **snake_case__ : int ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = super().__call__(*snake_case__ , **snake_case__ )
# TODO try and retrieve it in a nicer way from _sanitize_parameters.
snake_case : Tuple = "top_k" not in kwargs
if isinstance(args[0] , snake_case__ ) and _legacy:
# This pipeline is odd, and return a list when single item is run
return [result]
else:
return result
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Tuple , **snake_case__ : Union[str, Any] ) -> Dict[str, GenericTensor]:
'''simple docstring'''
snake_case : int = self.framework
if isinstance(snake_case__ , snake_case__ ):
return self.tokenizer(**snake_case__ , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ) and len(snake_case__ ) == 1 and isinstance(inputs[0] , snake_case__ ) and len(inputs[0] ) == 2:
# It used to be valid to use a list of list of list for text pairs, keeping this path for BC
return self.tokenizer(
text=inputs[0][0] , text_pair=inputs[0][1] , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
# This is likely an invalid usage of the pipeline attempting to pass text pairs.
raise ValueError(
"The pipeline received invalid inputs, if you are trying to send text pairs, you can try to send a"
" dictionary `{\"text\": \"My text\", \"text_pair\": \"My pair\"}` in order to send a text pair." )
return self.tokenizer(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Union[str, Any] ) -> int:
'''simple docstring'''
return self.model(**snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Optional[Any] , snake_case__ : List[str]=None , snake_case__ : Dict=1 , snake_case__ : Tuple=True ) -> str:
'''simple docstring'''
if function_to_apply is None:
if self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels == 1:
snake_case : Tuple = ClassificationFunction.SIGMOID
elif self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels > 1:
snake_case : Tuple = ClassificationFunction.SOFTMAX
elif hasattr(self.model.config , "function_to_apply" ) and function_to_apply is None:
snake_case : Tuple = self.model.config.function_to_apply
else:
snake_case : int = ClassificationFunction.NONE
snake_case : Any = model_outputs["logits"][0]
snake_case : List[str] = outputs.numpy()
if function_to_apply == ClassificationFunction.SIGMOID:
snake_case : Optional[Any] = sigmoid(snake_case__ )
elif function_to_apply == ClassificationFunction.SOFTMAX:
snake_case : Union[str, Any] = softmax(snake_case__ )
elif function_to_apply == ClassificationFunction.NONE:
snake_case : Optional[Any] = outputs
else:
raise ValueError(f"""Unrecognized `function_to_apply` argument: {function_to_apply}""" )
if top_k == 1 and _legacy:
return {"label": self.model.config.idalabel[scores.argmax().item()], "score": scores.max().item()}
snake_case : Optional[int] = [
{"label": self.model.config.idalabel[i], "score": score.item()} for i, score in enumerate(snake_case__ )
]
if not _legacy:
dict_scores.sort(key=lambda snake_case__ : x["score"] , reverse=snake_case__ )
if top_k is not None:
snake_case : Optional[int] = dict_scores[:top_k]
return dict_scores
| 59
| 1
|
import contextlib
import csv
import json
import os
import sqlitea
import tarfile
import textwrap
import zipfile
import pyarrow as pa
import pyarrow.parquet as pq
import pytest
import datasets
import datasets.config
@pytest.fixture(scope="session" )
def UpperCamelCase ( ):
snake_case : int = 10
snake_case : Any = datasets.Features(
{
"tokens": datasets.Sequence(datasets.Value("string" ) ),
"labels": datasets.Sequence(datasets.ClassLabel(names=["negative", "positive"] ) ),
"answers": datasets.Sequence(
{
"text": datasets.Value("string" ),
"answer_start": datasets.Value("int32" ),
} ),
"id": datasets.Value("int64" ),
} )
snake_case : Optional[int] = datasets.Dataset.from_dict(
{
"tokens": [["foo"] * 5] * n,
"labels": [[1] * 5] * n,
"answers": [{"answer_start": [97], "text": ["1976"]}] * 10,
"id": list(range(__lowerCamelCase ) ),
} , features=__lowerCamelCase , )
return dataset
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : List[str] ):
snake_case : List[Any] = str(tmp_path_factory.mktemp("data" ) / "file.arrow" )
dataset.map(cache_file_name=__lowerCamelCase )
return filename
# FILE_CONTENT + files
__lowerCamelCase = """\
Text data.
Second line of data."""
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Any ):
snake_case : Any = tmp_path_factory.mktemp("data" ) / "file.txt"
snake_case : Union[str, Any] = FILE_CONTENT
with open(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase )
return filename
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict ):
import bza
snake_case : Tuple = tmp_path_factory.mktemp("data" ) / "file.txt.bz2"
snake_case : List[Any] = bytes(__lowerCamelCase , "utf-8" )
with bza.open(__lowerCamelCase , "wb" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Tuple ):
import gzip
snake_case : Union[str, Any] = str(tmp_path_factory.mktemp("data" ) / "file.txt.gz" )
snake_case : Any = bytes(__lowerCamelCase , "utf-8" )
with gzip.open(__lowerCamelCase , "wb" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] ):
if datasets.config.LZ4_AVAILABLE:
import lza.frame
snake_case : List[Any] = tmp_path_factory.mktemp("data" ) / "file.txt.lz4"
snake_case : Optional[int] = bytes(__lowerCamelCase , "utf-8" )
with lza.frame.open(__lowerCamelCase , "wb" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[str] ):
if datasets.config.PY7ZR_AVAILABLE:
import pyazr
snake_case : Dict = tmp_path_factory.mktemp("data" ) / "file.txt.7z"
with pyazr.SevenZipFile(__lowerCamelCase , "w" ) as archive:
archive.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Union[str, Any] ):
import tarfile
snake_case : Optional[int] = tmp_path_factory.mktemp("data" ) / "file.txt.tar"
with tarfile.TarFile(__lowerCamelCase , "w" ) as f:
f.add(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Any ):
import lzma
snake_case : Optional[int] = tmp_path_factory.mktemp("data" ) / "file.txt.xz"
snake_case : int = bytes(__lowerCamelCase , "utf-8" )
with lzma.open(__lowerCamelCase , "wb" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : List[Any] , __lowerCamelCase : Any ):
import zipfile
snake_case : List[Any] = tmp_path_factory.mktemp("data" ) / "file.txt.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] ):
if datasets.config.ZSTANDARD_AVAILABLE:
import zstandard as zstd
snake_case : Optional[int] = tmp_path_factory.mktemp("data" ) / "file.txt.zst"
snake_case : Dict = bytes(__lowerCamelCase , "utf-8" )
with zstd.open(__lowerCamelCase , "wb" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Tuple ):
snake_case : Optional[Any] = tmp_path_factory.mktemp("data" ) / "file.xml"
snake_case : str = textwrap.dedent(
"\\n <?xml version=\"1.0\" encoding=\"UTF-8\" ?>\n <tmx version=\"1.4\">\n <header segtype=\"sentence\" srclang=\"ca\" />\n <body>\n <tu>\n <tuv xml:lang=\"ca\"><seg>Contingut 1</seg></tuv>\n <tuv xml:lang=\"en\"><seg>Content 1</seg></tuv>\n </tu>\n <tu>\n <tuv xml:lang=\"ca\"><seg>Contingut 2</seg></tuv>\n <tuv xml:lang=\"en\"><seg>Content 2</seg></tuv>\n </tu>\n <tu>\n <tuv xml:lang=\"ca\"><seg>Contingut 3</seg></tuv>\n <tuv xml:lang=\"en\"><seg>Content 3</seg></tuv>\n </tu>\n <tu>\n <tuv xml:lang=\"ca\"><seg>Contingut 4</seg></tuv>\n <tuv xml:lang=\"en\"><seg>Content 4</seg></tuv>\n </tu>\n <tu>\n <tuv xml:lang=\"ca\"><seg>Contingut 5</seg></tuv>\n <tuv xml:lang=\"en\"><seg>Content 5</seg></tuv>\n </tu>\n </body>\n </tmx>" )
with open(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase )
return filename
__lowerCamelCase = [
{"""col_1""": """0""", """col_2""": 0, """col_3""": 0.0},
{"""col_1""": """1""", """col_2""": 1, """col_3""": 1.0},
{"""col_1""": """2""", """col_2""": 2, """col_3""": 2.0},
{"""col_1""": """3""", """col_2""": 3, """col_3""": 3.0},
]
__lowerCamelCase = [
{"""col_1""": """4""", """col_2""": 4, """col_3""": 4.0},
{"""col_1""": """5""", """col_2""": 5, """col_3""": 5.0},
]
__lowerCamelCase = {
"""col_1""": ["""0""", """1""", """2""", """3"""],
"""col_2""": [0, 1, 2, 3],
"""col_3""": [0.0, 1.0, 2.0, 3.0],
}
__lowerCamelCase = [
{"""col_3""": 0.0, """col_1""": """0""", """col_2""": 0},
{"""col_3""": 1.0, """col_1""": """1""", """col_2""": 1},
]
__lowerCamelCase = [
{"""col_1""": """s0""", """col_2""": 0, """col_3""": 0.0},
{"""col_1""": """s1""", """col_2""": 1, """col_3""": 1.0},
{"""col_1""": """s2""", """col_2""": 2, """col_3""": 2.0},
{"""col_1""": """s3""", """col_2""": 3, """col_3""": 3.0},
]
@pytest.fixture(scope="session" )
def UpperCamelCase ( ):
return DATA_DICT_OF_LISTS
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] ):
snake_case : str = datasets.Dataset.from_dict(__lowerCamelCase )
snake_case : Optional[int] = str(tmp_path_factory.mktemp("data" ) / "dataset.arrow" )
dataset.map(cache_file_name=__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Tuple ):
snake_case : int = str(tmp_path_factory.mktemp("data" ) / "dataset.sqlite" )
with contextlib.closing(sqlitea.connect(__lowerCamelCase ) ) as con:
snake_case : Optional[int] = con.cursor()
cur.execute("CREATE TABLE dataset(col_1 text, col_2 int, col_3 real)" )
for item in DATA:
cur.execute("INSERT INTO dataset(col_1, col_2, col_3) VALUES (?, ?, ?)" , tuple(item.values() ) )
con.commit()
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : List[str] ):
snake_case : Union[str, Any] = str(tmp_path_factory.mktemp("data" ) / "dataset.csv" )
with open(__lowerCamelCase , "w" , newline="" ) as f:
snake_case : str = csv.DictWriter(__lowerCamelCase , fieldnames=["col_1", "col_2", "col_3"] )
writer.writeheader()
for item in DATA:
writer.writerow(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict ):
snake_case : Any = str(tmp_path_factory.mktemp("data" ) / "dataset2.csv" )
with open(__lowerCamelCase , "w" , newline="" ) as f:
snake_case : Optional[int] = csv.DictWriter(__lowerCamelCase , fieldnames=["col_1", "col_2", "col_3"] )
writer.writeheader()
for item in DATA:
writer.writerow(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : List[str] ):
import bza
snake_case : Optional[Any] = tmp_path_factory.mktemp("data" ) / "dataset.csv.bz2"
with open(__lowerCamelCase , "rb" ) as f:
snake_case : str = f.read()
# data = bytes(FILE_CONTENT, "utf-8")
with bza.open(__lowerCamelCase , "wb" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Tuple , __lowerCamelCase : Dict ):
snake_case : Union[str, Any] = tmp_path_factory.mktemp("data" ) / "dataset.csv.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : int , __lowerCamelCase : int ):
snake_case : str = tmp_path_factory.mktemp("data" ) / "dataset.csv.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename(csv_path.replace(".csv" , ".CSV" ) ) )
f.write(__lowerCamelCase , arcname=os.path.basename(csva_path.replace(".csv" , ".CSV" ) ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Any , __lowerCamelCase : Any ):
snake_case : str = tmp_path_factory.mktemp("data" ) / "dataset_with_dir.csv.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.join("main_dir" , os.path.basename(__lowerCamelCase ) ) )
f.write(__lowerCamelCase , arcname=os.path.join("main_dir" , os.path.basename(__lowerCamelCase ) ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Optional[int] ):
snake_case : Tuple = str(tmp_path_factory.mktemp("data" ) / "dataset.parquet" )
snake_case : Tuple = pa.schema(
{
"col_1": pa.string(),
"col_2": pa.intaa(),
"col_3": pa.floataa(),
} )
with open(__lowerCamelCase , "wb" ) as f:
snake_case : Tuple = pq.ParquetWriter(__lowerCamelCase , schema=__lowerCamelCase )
snake_case : Dict = pa.Table.from_pydict({k: [DATA[i][k] for i in range(len(__lowerCamelCase ) )] for k in DATA[0]} , schema=__lowerCamelCase )
writer.write_table(__lowerCamelCase )
writer.close()
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : List[str] = str(tmp_path_factory.mktemp("data" ) / "dataset.json" )
snake_case : Union[str, Any] = {"data": DATA}
with open(__lowerCamelCase , "w" ) as f:
json.dump(__lowerCamelCase , __lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict ):
snake_case : Optional[int] = str(tmp_path_factory.mktemp("data" ) / "dataset.json" )
snake_case : Tuple = {"data": DATA_DICT_OF_LISTS}
with open(__lowerCamelCase , "w" ) as f:
json.dump(__lowerCamelCase , __lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : List[str] = str(tmp_path_factory.mktemp("data" ) / "dataset.jsonl" )
with open(__lowerCamelCase , "w" ) as f:
for item in DATA:
f.write(json.dumps(__lowerCamelCase ) + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Any ):
snake_case : List[str] = str(tmp_path_factory.mktemp("data" ) / "dataset2.jsonl" )
with open(__lowerCamelCase , "w" ) as f:
for item in DATA:
f.write(json.dumps(__lowerCamelCase ) + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Any ):
snake_case : Tuple = str(tmp_path_factory.mktemp("data" ) / "dataset_312.jsonl" )
with open(__lowerCamelCase , "w" ) as f:
for item in DATA_312:
f.write(json.dumps(__lowerCamelCase ) + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Any ):
snake_case : str = str(tmp_path_factory.mktemp("data" ) / "dataset-str.jsonl" )
with open(__lowerCamelCase , "w" ) as f:
for item in DATA_STR:
f.write(json.dumps(__lowerCamelCase ) + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : List[str] ):
import gzip
snake_case : Tuple = str(tmp_path_factory.mktemp("data" ) / "dataset.txt.gz" )
with open(__lowerCamelCase , "rb" ) as orig_file:
with gzip.open(__lowerCamelCase , "wb" ) as zipped_file:
zipped_file.writelines(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[str] ):
import gzip
snake_case : int = str(tmp_path_factory.mktemp("data" ) / "dataset.jsonl.gz" )
with open(__lowerCamelCase , "rb" ) as orig_file:
with gzip.open(__lowerCamelCase , "wb" ) as zipped_file:
zipped_file.writelines(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Dict ):
snake_case : Optional[int] = tmp_path_factory.mktemp("data" ) / "dataset.jsonl.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple , __lowerCamelCase : Dict , __lowerCamelCase : List[Any] ):
snake_case : Optional[Any] = tmp_path_factory.mktemp("data" ) / "dataset_nested.jsonl.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.join("nested" , os.path.basename(__lowerCamelCase ) ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple , __lowerCamelCase : int ):
snake_case : Tuple = tmp_path_factory.mktemp("data" ) / "dataset_with_dir.jsonl.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.join("main_dir" , os.path.basename(__lowerCamelCase ) ) )
f.write(__lowerCamelCase , arcname=os.path.join("main_dir" , os.path.basename(__lowerCamelCase ) ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Optional[Any] ):
snake_case : Optional[int] = tmp_path_factory.mktemp("data" ) / "dataset.jsonl.tar"
with tarfile.TarFile(__lowerCamelCase , "w" ) as f:
f.add(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
f.add(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : List[Any] , __lowerCamelCase : Optional[int] , __lowerCamelCase : Optional[Any] , __lowerCamelCase : int ):
snake_case : str = tmp_path_factory.mktemp("data" ) / "dataset_nested.jsonl.tar"
with tarfile.TarFile(__lowerCamelCase , "w" ) as f:
f.add(__lowerCamelCase , arcname=os.path.join("nested" , os.path.basename(__lowerCamelCase ) ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Optional[Any] ):
snake_case : Dict = ["0", "1", "2", "3"]
snake_case : List[Any] = str(tmp_path_factory.mktemp("data" ) / "dataset.txt" )
with open(__lowerCamelCase , "w" ) as f:
for item in data:
f.write(item + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : List[Any] = ["0", "1", "2", "3"]
snake_case : int = str(tmp_path_factory.mktemp("data" ) / "dataset2.txt" )
with open(__lowerCamelCase , "w" ) as f:
for item in data:
f.write(item + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict ):
snake_case : Union[str, Any] = ["0", "1", "2", "3"]
snake_case : Any = tmp_path_factory.mktemp("data" ) / "dataset.abc"
with open(__lowerCamelCase , "w" ) as f:
for item in data:
f.write(item + "\n" )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : Dict , __lowerCamelCase : Union[str, Any] ):
snake_case : Optional[int] = tmp_path_factory.mktemp("data" ) / "dataset.text.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : List[Any] , __lowerCamelCase : Optional[Any] ):
snake_case : List[Any] = tmp_path_factory.mktemp("data" ) / "dataset_with_dir.text.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.join("main_dir" , os.path.basename(__lowerCamelCase ) ) )
f.write(__lowerCamelCase , arcname=os.path.join("main_dir" , os.path.basename(__lowerCamelCase ) ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : List[str] , __lowerCamelCase : Optional[Any] , __lowerCamelCase : Union[str, Any] ):
snake_case : Any = tmp_path_factory.mktemp("data" ) / "dataset.ext.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename("unsupported.ext" ) )
f.write(__lowerCamelCase , arcname=os.path.basename("unsupported_2.ext" ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Optional[Any] ):
snake_case : List[Any] = "\n".join(["First", "Second\u2029with Unicode new line", "Third"] )
snake_case : str = str(tmp_path_factory.mktemp("data" ) / "dataset_with_unicode_new_lines.txt" )
with open(__lowerCamelCase , "w" , encoding="utf-8" ) as f:
f.write(__lowerCamelCase )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( ):
return os.path.join("tests" , "features" , "data" , "test_image_rgb.jpg" )
@pytest.fixture(scope="session" )
def UpperCamelCase ( ):
return os.path.join("tests" , "features" , "data" , "test_audio_44100.wav" )
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : Tuple , __lowerCamelCase : List[str] ):
snake_case : str = tmp_path_factory.mktemp("data" ) / "dataset.img.zip"
with zipfile.ZipFile(__lowerCamelCase , "w" ) as f:
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ) )
f.write(__lowerCamelCase , arcname=os.path.basename(__lowerCamelCase ).replace(".jpg" , "2.jpg" ) )
return path
@pytest.fixture(scope="session" )
def UpperCamelCase ( __lowerCamelCase : List[str] ):
snake_case : Optional[Any] = tmp_path_factory.mktemp("data_dir" )
(data_dir / "subdir").mkdir()
with open(data_dir / "subdir" / "train.txt" , "w" ) as f:
f.write("foo\n" * 10 )
with open(data_dir / "subdir" / "test.txt" , "w" ) as f:
f.write("bar\n" * 10 )
# hidden file
with open(data_dir / "subdir" / ".test.txt" , "w" ) as f:
f.write("bar\n" * 10 )
# hidden directory
(data_dir / ".subdir").mkdir()
with open(data_dir / ".subdir" / "train.txt" , "w" ) as f:
f.write("foo\n" * 10 )
with open(data_dir / ".subdir" / "test.txt" , "w" ) as f:
f.write("bar\n" * 10 )
return data_dir
| 59
|
from __future__ import annotations
__lowerCamelCase = list[list[int]]
# assigning initial values to the grid
__lowerCamelCase = [
[3, 0, 6, 5, 0, 8, 4, 0, 0],
[5, 2, 0, 0, 0, 0, 0, 0, 0],
[0, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
# a grid with no solution
__lowerCamelCase = [
[5, 0, 6, 5, 0, 8, 4, 0, 3],
[5, 2, 0, 0, 0, 0, 0, 0, 2],
[1, 8, 7, 0, 0, 0, 0, 3, 1],
[0, 0, 3, 0, 1, 0, 0, 8, 0],
[9, 0, 0, 8, 6, 3, 0, 0, 5],
[0, 5, 0, 0, 9, 0, 6, 0, 0],
[1, 3, 0, 0, 0, 0, 2, 5, 0],
[0, 0, 0, 0, 0, 0, 0, 7, 4],
[0, 0, 5, 2, 0, 6, 3, 0, 0],
]
def UpperCamelCase ( __lowerCamelCase : Matrix , __lowerCamelCase : int , __lowerCamelCase : int , __lowerCamelCase : int ):
for i in range(9 ):
if grid[row][i] == n or grid[i][column] == n:
return False
for i in range(3 ):
for j in range(3 ):
if grid[(row - row % 3) + i][(column - column % 3) + j] == n:
return False
return True
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for i in range(9 ):
for j in range(9 ):
if grid[i][j] == 0:
return i, j
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
if location := find_empty_location(__lowerCamelCase ):
snake_case , snake_case : Union[str, Any] = location
else:
# If the location is ``None``, then the grid is solved.
return grid
for digit in range(1 , 10 ):
if is_safe(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase ):
snake_case : List[Any] = digit
if sudoku(__lowerCamelCase ) is not None:
return grid
snake_case : Union[str, Any] = 0
return None
def UpperCamelCase ( __lowerCamelCase : Matrix ):
for row in grid:
for cell in row:
print(__lowerCamelCase , end=" " )
print()
if __name__ == "__main__":
# make a copy of grid so that you can compare with the unmodified grid
for example_grid in (initial_grid, no_solution):
print("""\nExample grid:\n""" + """=""" * 20)
print_solution(example_grid)
print("""\nExample grid solution:""")
__lowerCamelCase = sudoku(example_grid)
if solution is not None:
print_solution(solution)
else:
print("""Cannot find a solution.""")
| 59
| 1
|
import argparse
import shutil
from pathlib import Path
from tqdm import tqdm
from transformers import AutoTokenizer
def UpperCamelCase ( __lowerCamelCase : Dict , __lowerCamelCase : List[Any] , __lowerCamelCase : Optional[Any] , __lowerCamelCase : List[Any]=1024 ):
snake_case , snake_case : Tuple = [], []
snake_case : str = list(zip(__lowerCamelCase , __lowerCamelCase ) )
snake_case , snake_case : List[str] = sorted_examples[0]
def is_too_big(__lowerCamelCase : int ):
return tok(__lowerCamelCase , return_tensors="pt" ).input_ids.shape[1] > max_tokens
for src, tgt in tqdm(sorted_examples[1:] ):
snake_case : Optional[Any] = new_src + " " + src
snake_case : Dict = new_tgt + " " + tgt
if is_too_big(__lowerCamelCase ) or is_too_big(__lowerCamelCase ): # cant fit, finalize example
finished_src.append(__lowerCamelCase )
finished_tgt.append(__lowerCamelCase )
snake_case , snake_case : Optional[Any] = src, tgt
else: # can fit, keep adding
snake_case , snake_case : Optional[Any] = cand_src, cand_tgt
# cleanup
if new_src:
assert new_tgt
finished_src.append(__lowerCamelCase )
finished_tgt.append(__lowerCamelCase )
return finished_src, finished_tgt
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Path , __lowerCamelCase : Optional[int] , __lowerCamelCase : Union[str, Any] ):
snake_case : Union[str, Any] = Path(__lowerCamelCase )
save_path.mkdir(exist_ok=__lowerCamelCase )
for split in ["train"]:
snake_case , snake_case : str = data_dir / f"""{split}.source""", data_dir / f"""{split}.target"""
snake_case : List[str] = [x.rstrip() for x in Path(__lowerCamelCase ).open().readlines()]
snake_case : Tuple = [x.rstrip() for x in Path(__lowerCamelCase ).open().readlines()]
snake_case , snake_case : Optional[int] = pack_examples(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
print(f"""packed {split} split from {len(__lowerCamelCase )} examples -> {len(__lowerCamelCase )}.""" )
Path(save_path / f"""{split}.source""" ).open("w" ).write("\n".join(__lowerCamelCase ) )
Path(save_path / f"""{split}.target""" ).open("w" ).write("\n".join(__lowerCamelCase ) )
for split in ["val", "test"]:
snake_case , snake_case : Optional[int] = data_dir / f"""{split}.source""", data_dir / f"""{split}.target"""
shutil.copyfile(__lowerCamelCase , save_path / f"""{split}.source""" )
shutil.copyfile(__lowerCamelCase , save_path / f"""{split}.target""" )
def UpperCamelCase ( ):
snake_case : Tuple = argparse.ArgumentParser()
parser.add_argument("--tok_name" , type=__lowerCamelCase , help="like facebook/bart-large-cnn,t5-base, etc." )
parser.add_argument("--max_seq_len" , type=__lowerCamelCase , default=128 )
parser.add_argument("--data_dir" , type=__lowerCamelCase )
parser.add_argument("--save_path" , type=__lowerCamelCase )
snake_case : Dict = parser.parse_args()
snake_case : Tuple = AutoTokenizer.from_pretrained(args.tok_name )
return pack_data_dir(__lowerCamelCase , Path(args.data_dir ) , args.max_seq_len , args.save_path )
if __name__ == "__main__":
packer_cli()
| 59
|
import logging
import numpy as np
import pytest
from scipy.linalg import eigh
logging.basicConfig(level=logging.INFO, format="""%(message)s""")
def UpperCamelCase ( __lowerCamelCase : np.ndarray ):
return input_array.reshape((input_array.size, 1) )
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Any = np.nan
for i in range(__lowerCamelCase ):
snake_case : List[str] = features[:, labels == i]
snake_case : Dict = data.mean(1 )
# Centralize the data of class i
snake_case : Optional[Any] = data - column_reshape(__lowerCamelCase )
if i > 0:
# If covariance_sum is not None
covariance_sum += np.dot(__lowerCamelCase , centered_data.T )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
snake_case : Optional[Any] = features.mean(1 )
snake_case : Tuple = np.nan
for i in range(__lowerCamelCase ):
snake_case : Tuple = features[:, labels == i]
snake_case : Tuple = data.shape[1]
snake_case : List[str] = data.mean(1 )
if i > 0:
# If covariance_sum is not None
covariance_sum += device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
else:
# If covariance_sum is np.nan (i.e. first loop)
snake_case : Optional[int] = device_data * np.dot(
column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase ) , (column_reshape(__lowerCamelCase ) - column_reshape(__lowerCamelCase )).T , )
return covariance_sum / features.shape[1]
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : int ):
# Check if the features have been loaded
if features.any():
snake_case : Tuple = features.mean(1 )
# Center the dataset
snake_case : List[str] = features - np.reshape(__lowerCamelCase , (data_mean.size, 1) )
snake_case : Optional[Any] = np.dot(__lowerCamelCase , centered_data.T ) / features.shape[1]
snake_case , snake_case : Dict = np.linalg.eigh(__lowerCamelCase )
# Take all the columns in the reverse order (-1), and then takes only the first
snake_case : Optional[Any] = eigenvectors[:, ::-1][:, 0:dimensions]
# Project the database on the new space
snake_case : Union[str, Any] = np.dot(filtered_eigenvectors.T , __lowerCamelCase )
logging.info("Principal Component Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( __lowerCamelCase : np.ndarray , __lowerCamelCase : np.ndarray , __lowerCamelCase : int , __lowerCamelCase : int ):
assert classes > dimensions
# Check if features have been already loaded
if features.any:
snake_case , snake_case : str = eigh(
covariance_between_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , covariance_within_classes(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) , )
snake_case : str = eigenvectors[:, ::-1][:, :dimensions]
snake_case , snake_case , snake_case : int = np.linalg.svd(__lowerCamelCase )
snake_case : List[Any] = svd_matrix[:, 0:dimensions]
snake_case : Optional[Any] = np.dot(filtered_svd_matrix.T , __lowerCamelCase )
logging.info("Linear Discriminant Analysis computed" )
return projected_data
else:
logging.basicConfig(level=logging.ERROR , format="%(message)s" , force=__lowerCamelCase )
logging.error("Dataset empty" )
raise AssertionError
def UpperCamelCase ( ):
# Create dummy dataset with 2 classes and 3 features
snake_case : str = np.array([[1, 2, 3, 4, 5], [2, 3, 4, 5, 6], [3, 4, 5, 6, 7]] )
snake_case : Union[str, Any] = np.array([0, 0, 0, 1, 1] )
snake_case : List[Any] = 2
snake_case : Any = 2
# Assert that the function raises an AssertionError if dimensions > classes
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : str = linear_discriminant_analysis(
__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
if isinstance(__lowerCamelCase , np.ndarray ):
raise AssertionError(
"Did not raise AssertionError for dimensions > classes" )
assert error_info.type is AssertionError
def UpperCamelCase ( ):
snake_case : List[str] = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]] )
snake_case : List[str] = 2
snake_case : int = np.array([[6.9282_0323, 8.6602_5404, 10.3923_0485], [3.0, 3.0, 3.0]] )
with pytest.raises(__lowerCamelCase ) as error_info:
snake_case : Union[str, Any] = principal_component_analysis(__lowerCamelCase , __lowerCamelCase )
if not np.allclose(__lowerCamelCase , __lowerCamelCase ):
raise AssertionError
assert error_info.type is AssertionError
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 1
|
def _a ( a :int = 1_000 ) -> int:
a = 2**power
a = str(a )
a = list(a )
a = 0
for i in list_num:
sum_of_num += int(a )
return sum_of_num
if __name__ == "__main__":
UpperCAmelCase__ = int(input("Enter the power of 2: ").strip())
print("2 ^ ", power, " = ", 2**power)
UpperCAmelCase__ = solution(power)
print("Sum of the digits is: ", result)
| 0
|
import pytest
from datasets import inspect_metric, list_metrics, load_metric
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : Optional[int] ):
monkeypatch.setattr("datasets.utils.deprecation_utils._emitted_deprecation_warnings" , set() )
@pytest.fixture
def UpperCamelCase ( __lowerCamelCase : str ):
class UpperCAmelCase :
def __init__(self : Optional[int] , snake_case__ : str ) -> Any:
'''simple docstring'''
snake_case : List[str] = metric_id
class UpperCAmelCase :
A__ : List[str] = [MetricMock(A_ ) for metric_id in ["accuracy", "mse", "precision", "codeparrot/apps_metric"]]
def _SCREAMING_SNAKE_CASE (self : int ) -> List[str]:
'''simple docstring'''
return self._metrics
monkeypatch.setattr("datasets.inspect.huggingface_hub" , HfhMock() )
@pytest.mark.parametrize(
"func, args" , [(load_metric, ("metrics/mse",)), (list_metrics, ()), (inspect_metric, ("metrics/mse", "tmp_path"))] )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : Any , __lowerCamelCase : Any , __lowerCamelCase : int , __lowerCamelCase : Any ):
if "tmp_path" in args:
snake_case : str = tuple(arg if arg != "tmp_path" else tmp_path for arg in args )
with pytest.warns(__lowerCamelCase , match="https://huggingface.co/docs/evaluate" ):
func(*__lowerCamelCase )
| 59
| 0
|
'''simple docstring'''
from __future__ import annotations
from collections import deque
class __A :
def __init__(self : List[Any] , __a : list[str] ):
UpperCAmelCase_ = []
self.adlist.append(
{"value": "", "next_states": [], "fail_state": 0, "output": []} )
for keyword in keywords:
self.add_keyword(__a )
self.set_fail_transitions()
def _lowercase (self : Union[str, Any] , __a : int , __a : str ):
for state in self.adlist[current_state]["next_states"]:
if char == self.adlist[state]["value"]:
return state
return None
def _lowercase (self : Optional[Any] , __a : str ):
UpperCAmelCase_ = 0
for character in keyword:
UpperCAmelCase_ = self.find_next_state(__a , __a )
if next_state is None:
self.adlist.append(
{
"value": character,
"next_states": [],
"fail_state": 0,
"output": [],
} )
self.adlist[current_state]["next_states"].append(len(self.adlist ) - 1 )
UpperCAmelCase_ = len(self.adlist ) - 1
else:
UpperCAmelCase_ = next_state
self.adlist[current_state]["output"].append(__a )
def _lowercase (self : int ):
UpperCAmelCase_ = deque()
for node in self.adlist[0]["next_states"]:
q.append(__a )
UpperCAmelCase_ = 0
while q:
UpperCAmelCase_ = q.popleft()
for child in self.adlist[r]["next_states"]:
q.append(__a )
UpperCAmelCase_ = self.adlist[r]["fail_state"]
while (
self.find_next_state(__a , self.adlist[child]["value"] ) is None
and state != 0
):
UpperCAmelCase_ = self.adlist[state]["fail_state"]
UpperCAmelCase_ = self.find_next_state(
__a , self.adlist[child]["value"] )
if self.adlist[child]["fail_state"] is None:
UpperCAmelCase_ = 0
UpperCAmelCase_ = (
self.adlist[child]["output"]
+ self.adlist[self.adlist[child]["fail_state"]]["output"]
)
def _lowercase (self : Dict , __a : str ):
UpperCAmelCase_ = {} # returns a dict with keywords and list of its occurrences
UpperCAmelCase_ = 0
for i in range(len(__a ) ):
while (
self.find_next_state(__a , string[i] ) is None
and current_state != 0
):
UpperCAmelCase_ = self.adlist[current_state]["fail_state"]
UpperCAmelCase_ = self.find_next_state(__a , string[i] )
if next_state is None:
UpperCAmelCase_ = 0
else:
UpperCAmelCase_ = next_state
for key in self.adlist[current_state]["output"]:
if key not in result:
UpperCAmelCase_ = []
result[key].append(i - len(__a ) + 1 )
return result
if __name__ == "__main__":
import doctest
doctest.testmod()
| 1
|
import argparse
import dataclasses
import json
import logging
import os
import shutil
from typing import List, Optional
import datasets
from accelerate import Accelerator
from datasets import load_dataset
from finetuning import finetune
from tqdm.auto import tqdm
import transformers
from transformers import AutoConfig, set_seed
from transformers.trainer_utils import IntervalStrategy
__lowerCamelCase = logging.getLogger(__name__)
__lowerCamelCase = """pytorch_model.bin"""
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co."} ,)
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the training data."} )
A__ : str = dataclasses.field(metadata={"help": "A csv or a json file containing the data to predict on."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "A csv or a json file containing the validation data."} )
A__ : Optional[str] = dataclasses.field(
default=A_ ,metadata={"help": "The name of the task to train on."} ,)
A__ : Optional[List[str]] = dataclasses.field(
default=A_ ,metadata={"help": "The list of labels for the task."} )
@dataclasses.dataclass
class UpperCAmelCase :
A__ : str = dataclasses.field(
metadata={"help": "The output directory where the model predictions and checkpoints will be written."} )
A__ : Optional[str] = dataclasses.field(
default="accuracy" ,metadata={"help": "The evaluation metric used for the task."} )
A__ : Optional[str] = dataclasses.field(
default="no" ,metadata={
"help": "The evaluation strategy to adopt during training. Possible values are: [\"no\", \"step\", \"epoch]"
} ,)
A__ : Optional[int] = dataclasses.field(
default=10 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={
"help": "How much the specified evaluation metric must improve to satisfy early stopping conditions."
} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the confidence score."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to filter the pseudo-labeled data based on the validation performance."} ,)
A__ : Optional[bool] = dataclasses.field(
default=A_ ,metadata={"help": "Whether to fine-tune on labeled data after pseudo training."} ,)
A__ : Optional[float] = dataclasses.field(
default=0.0 ,metadata={"help": "Confidence threshold for pseudo-labeled data filtering."} ,)
A__ : Optional[int] = dataclasses.field(
default=1_00 ,metadata={"help": "Number of evaluation calls with no improvement after which training will be stopped."} ,)
A__ : Optional[int] = dataclasses.field(
default=A_ ,metadata={"help": "Random seed for initialization."} ,)
def UpperCamelCase ( __lowerCamelCase : int , __lowerCamelCase : Tuple , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Tuple , __lowerCamelCase : Dict , __lowerCamelCase : Optional[int] ):
snake_case : Tuple = datasets.concatenate_datasets([infer_input, infer_output] , axis=1 )
if args.do_filter_by_confidence:
snake_case : Optional[int] = dataset.filter(lambda __lowerCamelCase : example["probability"] > args.confidence_threshold )
if args.do_filter_by_val_performance:
assert eval_result >= 0.0 and eval_result <= 1.0
snake_case : int = int(eval_result * len(__lowerCamelCase ) )
print(__lowerCamelCase )
snake_case : List[str] = dataset.sort("probability" , reverse=__lowerCamelCase )
snake_case : Tuple = dataset.select(range(__lowerCamelCase ) )
snake_case : List[Any] = dataset.remove_columns(["label", "probability"] )
snake_case : Any = dataset.rename_column("prediction" , "label" )
snake_case : str = dataset.map(lambda __lowerCamelCase : {"label": idalabel[example["label"]]} )
snake_case : List[str] = dataset.shuffle(seed=args.seed )
snake_case : int = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.data_file_extension == "csv":
dataset.to_csv(__lowerCamelCase , index=__lowerCamelCase )
else:
dataset.to_json(__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : Optional[int] , __lowerCamelCase : str , __lowerCamelCase : Optional[int] , __lowerCamelCase : Tuple , **__lowerCamelCase : List[Any] ):
snake_case : int = Accelerator()
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(accelerator.state )
# Setup logging, we only want one process per machine to log things on the
# screen. accelerator.is_local_main_process is only True for one process per
# machine.
logger.setLevel(logging.INFO if accelerator.is_local_main_process else logging.ERROR )
if accelerator.is_local_main_process:
datasets.utils.logging.set_verbosity_warning()
transformers.utils.logging.set_verbosity_info()
else:
datasets.utils.logging.set_verbosity_error()
transformers.utils.logging.set_verbosity_error()
snake_case : Dict = STModelArguments(model_name_or_path=__lowerCamelCase )
snake_case : Tuple = STDataArguments(train_file=__lowerCamelCase , infer_file=__lowerCamelCase )
snake_case : str = STTrainingArguments(output_dir=__lowerCamelCase )
snake_case : int = argparse.Namespace()
for arg_class in (model_args, data_args, training_args):
for key, value in vars(__lowerCamelCase ).items():
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
for key, value in kwargs.items():
if hasattr(__lowerCamelCase , __lowerCamelCase ):
setattr(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
# Sanity checks
snake_case : List[str] = {}
snake_case : Optional[int] = None
# You need to provide the training data and the data to predict on
assert args.train_file is not None
assert args.infer_file is not None
snake_case : str = args.train_file
snake_case : Tuple = args.infer_file
if args.evaluation_strategy != IntervalStrategy.NO.value:
assert args.eval_file is not None
snake_case : Tuple = args.eval_file
for key in data_files:
snake_case : List[Any] = data_files[key].split("." )[-1]
assert extension in ["csv", "json"], f"""`{key}_file` should be a csv or a json file."""
if args.data_file_extension is None:
snake_case : Union[str, Any] = extension
else:
assert extension == args.data_file_extension, f"""`{key}_file` should be a {args.data_file_extension} file`."""
assert (
args.eval_metric in datasets.list_metrics()
), f"""{args.eval_metric} not in the list of supported metrics {datasets.list_metrics()}."""
# If passed along, set the training seed now.
if args.seed is not None:
set_seed(args.seed )
logger.info("Creating the initial data directory for self-training..." )
snake_case : List[Any] = f"""{args.output_dir}/self-train_iter-{{}}""".format
snake_case : Optional[int] = data_dir_format(0 )
if accelerator.is_main_process:
if args.output_dir is not None:
os.makedirs(args.output_dir , exist_ok=__lowerCamelCase )
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : Dict = None
snake_case : Union[str, Any] = None
snake_case : Tuple = 0
snake_case : List[Any] = False
# Show the progress bar
snake_case : List[Any] = tqdm(range(args.max_selftrain_iterations ) , disable=not accelerator.is_local_main_process )
# Self-train
for iteration in range(0 , int(args.max_selftrain_iterations ) ):
snake_case : str = data_dir_format(__lowerCamelCase )
assert os.path.exists(__lowerCamelCase )
# Stage 1: initial fine-tuning for iteration = 0 or pseudo-training for
# iteration > 0
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-1" )
snake_case : Optional[Any] = {
"accelerator": accelerator,
"model_name_or_path": args.model_name_or_path,
"cache_dir": args.cache_dir,
"do_train": True,
"train_file": data_files["train"] if iteration == 0 else data_files["train_pseudo"],
"do_eval": True if args.eval_file is not None else False,
"eval_file": data_files["eval"],
"do_predict": True,
"infer_file": data_files["infer"],
"task_name": args.task_name,
"label_list": args.label_list,
"output_dir": current_output_dir,
"eval_metric": args.eval_metric,
"evaluation_strategy": args.evaluation_strategy,
"early_stopping_patience": args.early_stopping_patience,
"early_stopping_threshold": args.early_stopping_threshold,
"seed": args.seed,
}
# Add additional training arguments
for key, value in kwargs.items():
if key not in arguments_dict and not hasattr(__lowerCamelCase , __lowerCamelCase ):
arguments_dict.update({key: value} )
snake_case : int = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 1." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 1 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 1." , __lowerCamelCase )
if iteration > 0 and args.finetune_on_labeled_data:
# Stage 2 (optional): fine-tuning on the original labeled data
snake_case : str = os.path.join(__lowerCamelCase , "best-checkpoint" )
snake_case : Dict = os.path.join(__lowerCamelCase , "stage-2" )
# Update arguments_dict
snake_case : List[str] = model_path
snake_case : Optional[Any] = data_files["train"]
snake_case : Optional[Any] = current_output_dir
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "best-checkpoint" , __lowerCamelCase )
if os.path.exists(__lowerCamelCase ):
logger.info(
"Found existing model checkpoint at %s. Skipping self-training: iteration: %d, stage: 2." , __lowerCamelCase , __lowerCamelCase , )
else:
logger.info("***** Running self-training: iteration: %d, stage: 2 *****" , __lowerCamelCase )
finetune(**__lowerCamelCase )
accelerator.wait_for_everyone()
assert os.path.exists(__lowerCamelCase )
logger.info("Self-training job completed: iteration: %d, stage: 2." , __lowerCamelCase )
snake_case : int = iteration
snake_case : Tuple = data_dir_format(iteration + 1 )
snake_case : Tuple = AutoConfig.from_pretrained(os.path.join(__lowerCamelCase , "best-checkpoint" ) )
snake_case : Optional[int] = config.idalabel
snake_case : List[Any] = os.path.join(__lowerCamelCase , "eval_results_best-checkpoint.json" )
snake_case : Union[str, Any] = os.path.join(__lowerCamelCase , "test_results_best-checkpoint.json" )
assert os.path.exists(__lowerCamelCase )
with open(__lowerCamelCase , "r" ) as f:
snake_case : Dict = float(json.load(__lowerCamelCase )[args.eval_metric] )
snake_case : Optional[int] = os.path.join(__lowerCamelCase , "infer_output_best-checkpoint.csv" )
assert os.path.exists(__lowerCamelCase )
# Loading the dataset from local csv or json files.
snake_case : Optional[Any] = load_dataset(args.data_file_extension , data_files={"data": data_files["infer"]} )["data"]
snake_case : Dict = load_dataset("csv" , data_files={"data": infer_output_file} )["data"]
if accelerator.is_main_process:
os.makedirs(__lowerCamelCase , exist_ok=__lowerCamelCase )
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) )
if os.path.exists(__lowerCamelCase ):
shutil.copy(__lowerCamelCase , os.path.join(__lowerCamelCase , f"""test_results_iter-{iteration}.json""" ) )
create_pseudo_labeled_data(__lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase )
accelerator.wait_for_everyone()
snake_case : str = os.path.join(__lowerCamelCase , f"""train_pseudo.{args.data_file_extension}""" )
if args.evaluation_strategy != IntervalStrategy.NO.value:
snake_case : List[Any] = eval_result
if best_iteration is None:
snake_case : List[Any] = new_iteration
snake_case : int = new_eval_result
else:
if new_eval_result - best_eval_result > args.early_stopping_threshold:
snake_case : int = new_iteration
snake_case : Union[str, Any] = new_eval_result
snake_case : str = 0
else:
if new_eval_result == best_eval_result:
snake_case : Any = new_iteration
snake_case : Union[str, Any] = new_eval_result
early_stopping_patience_counter += 1
if early_stopping_patience_counter >= args.early_stopping_patience:
snake_case : Tuple = True
progress_bar.update(1 )
if should_training_stop:
break
if best_iteration is not None:
# Save the best iteration
logger.info("Best iteration: %d" , __lowerCamelCase )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{iteration}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
else:
# Assume that the last iteration is the best
logger.info("Best iteration: %d" , args.max_selftrain_iterations - 1 )
logger.info("Best evaluation result: %s = %f" , args.eval_metric , __lowerCamelCase )
accelerator.wait_for_everyone()
if accelerator.is_main_process:
shutil.copy(
os.path.join(__lowerCamelCase , f"""eval_results_iter-{args.max_selftrain_iterations - 1}.json""" ) , os.path.join(__lowerCamelCase , "eval_results_best-iteration.json" ) , )
| 59
| 0
|
'''simple docstring'''
import logging
import math
from functools import partial
from typing import Any, Callable, Dict, Iterable, List, Optional, Sequence, Tuple, Union
import torch
from .tensor_utils import tensor_tree_map, tree_map
def _SCREAMING_SNAKE_CASE (A ) -> List[Tuple[int, ...]]:
"""simple docstring"""
lowercase__ = []
if isinstance(A , A ):
for v in tree.values():
shapes.extend(_fetch_dims(A ) )
elif isinstance(A , (list, tuple) ):
for t in tree:
shapes.extend(_fetch_dims(A ) )
elif isinstance(A , torch.Tensor ):
shapes.append(tree.shape )
else:
raise ValueError('''Not supported''' )
return shapes
@torch.jit.ignore
def _SCREAMING_SNAKE_CASE (A , A ) -> Tuple[int, ...]:
"""simple docstring"""
lowercase__ = []
for d in reversed(A ):
idx.append(flat_idx % d )
lowercase__ = flat_idx // d
return tuple(reversed(A ) )
@torch.jit.ignore
def _SCREAMING_SNAKE_CASE (A , A , A , A = None , A = None , ) -> List[Tuple[slice, ...]]:
"""simple docstring"""
def reduce_edge_list(A ) -> None:
lowercase__ = True
for i in range(len(A ) ):
lowercase__ = -1 * (i + 1)
l[reversed_idx] &= tally
lowercase__ = l[reversed_idx]
if start_edges is None:
lowercase__ = [s == 0 for s in start]
reduce_edge_list(A )
if end_edges is None:
lowercase__ = [e == (d - 1) for e, d in zip(A , A )]
reduce_edge_list(A )
# Base cases. Either start/end are empty and we're done, or the final,
# one-dimensional tensor can be simply sliced
if len(A ) == 0:
return [()]
elif len(A ) == 1:
return [(slice(start[0] , end[0] + 1 ),)]
lowercase__ = []
lowercase__ = []
# Dimensions common to start and end can be selected directly
for s, e in zip(A , A ):
if s == e:
path_list.append(slice(A , s + 1 ) )
else:
break
lowercase__ = tuple(A )
lowercase__ = len(A )
# start == end, and we're done
if divergence_idx == len(A ):
return [path]
def upper() -> Tuple[Tuple[slice, ...], ...]:
assert start_edges is not None
assert end_edges is not None
lowercase__ = start[divergence_idx]
return tuple(
path + (slice(A , sdi + 1 ),) + s
for s in _get_minimal_slice_set(
start[divergence_idx + 1 :] , [d - 1 for d in dims[divergence_idx + 1 :]] , dims[divergence_idx + 1 :] , start_edges=start_edges[divergence_idx + 1 :] , end_edges=[True for _ in end_edges[divergence_idx + 1 :]] , ) )
def lower() -> Tuple[Tuple[slice, ...], ...]:
assert start_edges is not None
assert end_edges is not None
lowercase__ = end[divergence_idx]
return tuple(
path + (slice(A , edi + 1 ),) + s
for s in _get_minimal_slice_set(
[0 for _ in start[divergence_idx + 1 :]] , end[divergence_idx + 1 :] , dims[divergence_idx + 1 :] , start_edges=[True for _ in start_edges[divergence_idx + 1 :]] , end_edges=end_edges[divergence_idx + 1 :] , ) )
# If both start and end are at the edges of the subtree rooted at
# divergence_idx, we can just select the whole subtree at once
if start_edges[divergence_idx] and end_edges[divergence_idx]:
slices.append(path + (slice(start[divergence_idx] , end[divergence_idx] + 1 ),) )
# If just start is at the edge, we can grab almost all of the subtree,
# treating only the ragged bottom edge as an edge case
elif start_edges[divergence_idx]:
slices.append(path + (slice(start[divergence_idx] , end[divergence_idx] ),) )
slices.extend(lower() )
# Analogous to the previous case, but the top is ragged this time
elif end_edges[divergence_idx]:
slices.extend(upper() )
slices.append(path + (slice(start[divergence_idx] + 1 , end[divergence_idx] + 1 ),) )
# If both sides of the range are ragged, we need to handle both sides
# separately. If there's contiguous meat in between them, we can index it
# in one big chunk
else:
slices.extend(upper() )
lowercase__ = end[divergence_idx] - start[divergence_idx]
if middle_ground > 1:
slices.append(path + (slice(start[divergence_idx] + 1 , end[divergence_idx] ),) )
slices.extend(lower() )
return slices
@torch.jit.ignore
def _SCREAMING_SNAKE_CASE (A , A , A , A ) -> torch.Tensor:
"""simple docstring"""
lowercase__ = t.shape[:no_batch_dims]
lowercase__ = list(_flat_idx_to_idx(A , A ) )
# _get_minimal_slice_set is inclusive
lowercase__ = list(_flat_idx_to_idx(flat_end - 1 , A ) )
# Get an ordered list of slices to perform
lowercase__ = _get_minimal_slice_set(
A , A , A , )
lowercase__ = [t[s] for s in slices]
return torch.cat([s.view((-1,) + t.shape[no_batch_dims:] ) for s in sliced_tensors] )
def _SCREAMING_SNAKE_CASE (A , A , A , A , A = False , A = None , A = False , ) -> Any:
"""simple docstring"""
if not (len(A ) > 0):
raise ValueError('''Must provide at least one input''' )
lowercase__ = [shape[:no_batch_dims] for shape in _fetch_dims(A )]
lowercase__ = tuple([max(A ) for s in zip(*A )] )
def _prep_inputs(A ) -> torch.Tensor:
if not low_mem:
if not sum(t.shape[:no_batch_dims] ) == no_batch_dims:
lowercase__ = t.expand(orig_batch_dims + t.shape[no_batch_dims:] )
lowercase__ = t.reshape(-1 , *t.shape[no_batch_dims:] )
else:
lowercase__ = t.expand(orig_batch_dims + t.shape[no_batch_dims:] )
return t
lowercase__ = tensor_tree_map(_prep_inputs , A )
lowercase__ = None
if _out is not None:
lowercase__ = tensor_tree_map(lambda A : t.view([-1] + list(t.shape[no_batch_dims:] ) ) , _out )
lowercase__ = 1
for d in orig_batch_dims:
flat_batch_dim *= d
lowercase__ = flat_batch_dim // chunk_size + (flat_batch_dim % chunk_size != 0)
def _select_chunk(A ) -> torch.Tensor:
return t[i : i + chunk_size] if t.shape[0] != 1 else t
lowercase__ = 0
lowercase__ = prepped_outputs
for _ in range(A ):
# Chunk the input
if not low_mem:
lowercase__ = _select_chunk
else:
lowercase__ = partial(
_chunk_slice , flat_start=A , flat_end=min(A , i + chunk_size ) , no_batch_dims=len(A ) , )
lowercase__ = tensor_tree_map(A , A )
# Run the layer on the chunk
lowercase__ = layer(**A )
# Allocate space for the output
if out is None:
lowercase__ = tensor_tree_map(lambda A : t.new_zeros((flat_batch_dim,) + t.shape[1:] ) , A )
# Put the chunk in its pre-allocated space
if isinstance(A , A ):
def assign(A , A ) -> None:
for k, v in da.items():
if isinstance(A , A ):
assign(A , da[k] )
else:
if _add_into_out:
v[i : i + chunk_size] += da[k]
else:
lowercase__ = da[k]
assign(A , A )
elif isinstance(A , A ):
for xa, xa in zip(A , A ):
if _add_into_out:
xa[i : i + chunk_size] += xa
else:
lowercase__ = xa
elif isinstance(A , torch.Tensor ):
if _add_into_out:
out[i : i + chunk_size] += output_chunk
else:
lowercase__ = output_chunk
else:
raise ValueError('''Not supported''' )
i += chunk_size
lowercase__ = tensor_tree_map(lambda A : t.view(orig_batch_dims + t.shape[1:] ) , A )
return out
class __lowerCAmelCase :
'''simple docstring'''
def __init__(self : Tuple , UpperCamelCase : int = 512 , ):
'''simple docstring'''
lowercase__ = max_chunk_size
lowercase__ = None
lowercase__ = None
def UpperCamelCase__ (self : Optional[Any] , UpperCamelCase : Callable , UpperCamelCase : tuple , UpperCamelCase : int ):
'''simple docstring'''
logging.info('''Tuning chunk size...''' )
if min_chunk_size >= self.max_chunk_size:
return min_chunk_size
lowercase__ = [2**l for l in range(int(math.log(self.max_chunk_size , 2 ) ) + 1 )]
lowercase__ = [c for c in candidates if c > min_chunk_size]
lowercase__ = [min_chunk_size] + candidates
candidates[-1] += 4
def test_chunk_size(UpperCamelCase : int ) -> bool:
try:
with torch.no_grad():
fn(*UpperCamelCase , chunk_size=UpperCamelCase )
return True
except RuntimeError:
return False
lowercase__ = 0
lowercase__ = len(UpperCamelCase ) - 1
while i > min_viable_chunk_size_index:
lowercase__ = test_chunk_size(candidates[i] )
if not viable:
lowercase__ = (min_viable_chunk_size_index + i) // 2
else:
lowercase__ = i
lowercase__ = (i + len(UpperCamelCase ) - 1) // 2
return candidates[min_viable_chunk_size_index]
def UpperCamelCase__ (self : int , UpperCamelCase : Iterable , UpperCamelCase : Iterable ):
'''simple docstring'''
lowercase__ = True
for aa, aa in zip(UpperCamelCase , UpperCamelCase ):
assert type(UpperCamelCase ) == type(UpperCamelCase )
if isinstance(UpperCamelCase , (list, tuple) ):
consistent &= self._compare_arg_caches(UpperCamelCase , UpperCamelCase )
elif isinstance(UpperCamelCase , UpperCamelCase ):
lowercase__ = [v for _, v in sorted(aa.items() , key=lambda UpperCamelCase : x[0] )]
lowercase__ = [v for _, v in sorted(aa.items() , key=lambda UpperCamelCase : x[0] )]
consistent &= self._compare_arg_caches(UpperCamelCase , UpperCamelCase )
else:
consistent &= aa == aa
return consistent
def UpperCamelCase__ (self : int , UpperCamelCase : Callable , UpperCamelCase : tuple , UpperCamelCase : int , ):
'''simple docstring'''
lowercase__ = True
lowercase__ = tree_map(lambda UpperCamelCase : a.shape if isinstance(UpperCamelCase , torch.Tensor ) else a , UpperCamelCase , UpperCamelCase )
if self.cached_arg_data is not None:
# If args have changed shape/value, we need to re-tune
assert len(self.cached_arg_data ) == len(UpperCamelCase )
lowercase__ = self._compare_arg_caches(self.cached_arg_data , UpperCamelCase )
else:
# Otherwise, we can reuse the precomputed value
lowercase__ = False
if not consistent:
lowercase__ = self._determine_favorable_chunk_size(
UpperCamelCase , UpperCamelCase , UpperCamelCase , )
lowercase__ = arg_data
assert self.cached_chunk_size is not None
return self.cached_chunk_size
| 2
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__lowerCamelCase = {"""configuration_xglm""": ["""XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP""", """XGLMConfig"""]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizer"""]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""XGLMTokenizerFast"""]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""XGLMForCausalLM""",
"""XGLMModel""",
"""XGLMPreTrainedModel""",
]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""FlaxXGLMForCausalLM""",
"""FlaxXGLMModel""",
"""FlaxXGLMPreTrainedModel""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""TFXGLMForCausalLM""",
"""TFXGLMModel""",
"""TFXGLMPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_xglm import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP, XGLMConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm import XGLMTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_xglm_fast import XGLMTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_xglm import XGLM_PRETRAINED_MODEL_ARCHIVE_LIST, XGLMForCausalLM, XGLMModel, XGLMPreTrainedModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_xglm import FlaxXGLMForCausalLM, FlaxXGLMModel, FlaxXGLMPreTrainedModel
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_xglm import (
TF_XGLM_PRETRAINED_MODEL_ARCHIVE_LIST,
TFXGLMForCausalLM,
TFXGLMModel,
TFXGLMPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure)
| 59
| 0
|
'''simple docstring'''
from typing import Optional, Tuple, Union
import torch
from einops import rearrange, reduce
from diffusers import DDIMScheduler, DDPMScheduler, DiffusionPipeline, ImagePipelineOutput, UNetaDConditionModel
from diffusers.schedulers.scheduling_ddim import DDIMSchedulerOutput
from diffusers.schedulers.scheduling_ddpm import DDPMSchedulerOutput
lowercase : int = 8
def lowerCAmelCase_ ( snake_case__ , snake_case__=BITS ):
'''simple docstring'''
A : Optional[Any] = x.device
A : Optional[Any] = (x * 255).int().clamp(0 , 255 )
A : Union[str, Any] = 2 ** torch.arange(bits - 1 , -1 , -1 , device=snake_case__ )
A : int = rearrange(snake_case__ , '''d -> d 1 1''' )
A : Optional[Any] = rearrange(snake_case__ , '''b c h w -> b c 1 h w''' )
A : Union[str, Any] = ((x & mask) != 0).float()
A : int = rearrange(snake_case__ , '''b c d h w -> b (c d) h w''' )
A : Any = bits * 2 - 1
return bits
def lowerCAmelCase_ ( snake_case__ , snake_case__=BITS ):
'''simple docstring'''
A : Any = x.device
A : List[str] = (x > 0).int()
A : str = 2 ** torch.arange(bits - 1 , -1 , -1 , device=snake_case__ , dtype=torch.intaa )
A : Union[str, Any] = rearrange(snake_case__ , '''d -> d 1 1''' )
A : Tuple = rearrange(snake_case__ , '''b (c d) h w -> b c d h w''' , d=8 )
A : int = reduce(x * mask , '''b c d h w -> b c h w''' , '''sum''' )
return (dec / 255).clamp(0.0 , 1.0 )
def lowerCAmelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__ = 0.0 , snake_case__ = True , snake_case__=None , snake_case__ = True , ):
'''simple docstring'''
if self.num_inference_steps is None:
raise ValueError(
'''Number of inference steps is \'None\', you need to run \'set_timesteps\' after creating the scheduler''' )
# See formulas (12) and (16) of DDIM paper https://arxiv.org/pdf/2010.02502.pdf
# Ideally, read DDIM paper in-detail understanding
# Notation (<variable name> -> <name in paper>
# - pred_noise_t -> e_theta(x_t, t)
# - pred_original_sample -> f_theta(x_t, t) or x_0
# - std_dev_t -> sigma_t
# - eta -> η
# - pred_sample_direction -> "direction pointing to x_t"
# - pred_prev_sample -> "x_t-1"
# 1. get previous step value (=t-1)
A : Union[str, Any] = timestep - self.config.num_train_timesteps // self.num_inference_steps
# 2. compute alphas, betas
A : str = self.alphas_cumprod[timestep]
A : Any = self.alphas_cumprod[prev_timestep] if prev_timestep >= 0 else self.final_alpha_cumprod
A : int = 1 - alpha_prod_t
# 3. compute predicted original sample from predicted noise also called
# "predicted x_0" of formula (12) from https://arxiv.org/pdf/2010.02502.pdf
A : Union[str, Any] = (sample - beta_prod_t ** 0.5 * model_output) / alpha_prod_t ** 0.5
# 4. Clip "predicted x_0"
A : Optional[Any] = self.bit_scale
if self.config.clip_sample:
A : Union[str, Any] = torch.clamp(snake_case__ , -scale , snake_case__ )
# 5. compute variance: "sigma_t(η)" -> see formula (16)
# σ_t = sqrt((1 − α_t−1)/(1 − α_t)) * sqrt(1 − α_t/α_t−1)
A : Optional[Any] = self._get_variance(snake_case__ , snake_case__ )
A : Optional[int] = eta * variance ** 0.5
if use_clipped_model_output:
# the model_output is always re-derived from the clipped x_0 in Glide
A : Union[str, Any] = (sample - alpha_prod_t ** 0.5 * pred_original_sample) / beta_prod_t ** 0.5
# 6. compute "direction pointing to x_t" of formula (12) from https://arxiv.org/pdf/2010.02502.pdf
A : List[Any] = (1 - alpha_prod_t_prev - std_dev_t**2) ** 0.5 * model_output
# 7. compute x_t without "random noise" of formula (12) from https://arxiv.org/pdf/2010.02502.pdf
A : Optional[Any] = alpha_prod_t_prev ** 0.5 * pred_original_sample + pred_sample_direction
if eta > 0:
# randn_like does not support generator https://github.com/pytorch/pytorch/issues/27072
A : Dict = model_output.device if torch.is_tensor(snake_case__ ) else '''cpu'''
A : Optional[Any] = torch.randn(model_output.shape , dtype=model_output.dtype , generator=snake_case__ ).to(snake_case__ )
A : int = self._get_variance(snake_case__ , snake_case__ ) ** 0.5 * eta * noise
A : List[str] = prev_sample + variance
if not return_dict:
return (prev_sample,)
return DDIMSchedulerOutput(prev_sample=snake_case__ , pred_original_sample=snake_case__ )
def lowerCAmelCase_ ( self , snake_case__ , snake_case__ , snake_case__ , snake_case__="epsilon" , snake_case__=None , snake_case__ = True , ):
'''simple docstring'''
A : Dict = timestep
if model_output.shape[1] == sample.shape[1] * 2 and self.variance_type in ["learned", "learned_range"]:
A, A : int = torch.split(snake_case__ , sample.shape[1] , dim=1 )
else:
A : Any = None
# 1. compute alphas, betas
A : int = self.alphas_cumprod[t]
A : Any = self.alphas_cumprod[t - 1] if t > 0 else self.one
A : Union[str, Any] = 1 - alpha_prod_t
A : int = 1 - alpha_prod_t_prev
# 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 prediction_type == "epsilon":
A : Union[str, Any] = (sample - beta_prod_t ** 0.5 * model_output) / alpha_prod_t ** 0.5
elif prediction_type == "sample":
A : str = model_output
else:
raise ValueError(F'Unsupported prediction_type {prediction_type}.' )
# 3. Clip "predicted x_0"
A : Optional[int] = self.bit_scale
if self.config.clip_sample:
A : Any = torch.clamp(snake_case__ , -scale , snake_case__ )
# 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
A : Optional[Any] = (alpha_prod_t_prev ** 0.5 * self.betas[t]) / beta_prod_t
A : List[str] = self.alphas[t] ** 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
A : Union[str, Any] = pred_original_sample_coeff * pred_original_sample + current_sample_coeff * sample
# 6. Add noise
A : int = 0
if t > 0:
A : Tuple = torch.randn(
model_output.size() , dtype=model_output.dtype , layout=model_output.layout , generator=snake_case__ ).to(model_output.device )
A : str = (self._get_variance(snake_case__ , predicted_variance=snake_case__ ) ** 0.5) * noise
A : Dict = pred_prev_sample + variance
if not return_dict:
return (pred_prev_sample,)
return DDPMSchedulerOutput(prev_sample=snake_case__ , pred_original_sample=snake_case__ )
class A ( __snake_case ):
def __init__( self , SCREAMING_SNAKE_CASE , SCREAMING_SNAKE_CASE , SCREAMING_SNAKE_CASE = 1.0 , ) -> Dict:
"""simple docstring"""
super().__init__()
A : Optional[int] = bit_scale
A : str = (
ddim_bit_scheduler_step if isinstance(SCREAMING_SNAKE_CASE , SCREAMING_SNAKE_CASE ) else ddpm_bit_scheduler_step
)
self.register_modules(unet=SCREAMING_SNAKE_CASE , scheduler=SCREAMING_SNAKE_CASE )
@torch.no_grad()
def __call__( self , SCREAMING_SNAKE_CASE = 256 , SCREAMING_SNAKE_CASE = 256 , SCREAMING_SNAKE_CASE = 50 , SCREAMING_SNAKE_CASE = None , SCREAMING_SNAKE_CASE = 1 , SCREAMING_SNAKE_CASE = "pil" , SCREAMING_SNAKE_CASE = True , **SCREAMING_SNAKE_CASE , ) -> Union[Tuple, ImagePipelineOutput]:
"""simple docstring"""
A : List[Any] = torch.randn(
(batch_size, self.unet.config.in_channels, height, width) , generator=SCREAMING_SNAKE_CASE , )
A : Tuple = decimal_to_bits(SCREAMING_SNAKE_CASE ) * self.bit_scale
A : str = latents.to(self.device )
self.scheduler.set_timesteps(SCREAMING_SNAKE_CASE )
for t in self.progress_bar(self.scheduler.timesteps ):
# predict the noise residual
A : Any = self.unet(SCREAMING_SNAKE_CASE , SCREAMING_SNAKE_CASE ).sample
# compute the previous noisy sample x_t -> x_t-1
A : Union[str, Any] = self.scheduler.step(SCREAMING_SNAKE_CASE , SCREAMING_SNAKE_CASE , SCREAMING_SNAKE_CASE ).prev_sample
A : List[Any] = bits_to_decimal(SCREAMING_SNAKE_CASE )
if output_type == "pil":
A : Any = self.numpy_to_pil(SCREAMING_SNAKE_CASE )
if not return_dict:
return (image,)
return ImagePipelineOutput(images=SCREAMING_SNAKE_CASE )
| 3
|
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
# See all MEGATRON_BERT models at https://huggingface.co/models?filter=bert
}
class UpperCAmelCase ( A_ ):
A__ : List[str] = "megatron-bert"
def __init__(self : Optional[int] , snake_case__ : List[str]=2_90_56 , snake_case__ : List[Any]=10_24 , snake_case__ : str=24 , snake_case__ : Tuple=16 , snake_case__ : Union[str, Any]=40_96 , snake_case__ : str="gelu" , snake_case__ : str=0.1 , snake_case__ : Optional[int]=0.1 , snake_case__ : Tuple=5_12 , snake_case__ : Union[str, Any]=2 , snake_case__ : Dict=0.02 , snake_case__ : List[Any]=1e-12 , snake_case__ : int=0 , snake_case__ : Tuple="absolute" , snake_case__ : Any=True , **snake_case__ : Union[str, Any] , ) -> Optional[Any]:
'''simple docstring'''
super().__init__(pad_token_id=snake_case__ , **snake_case__ )
snake_case : Tuple = vocab_size
snake_case : str = hidden_size
snake_case : str = num_hidden_layers
snake_case : str = num_attention_heads
snake_case : Optional[int] = hidden_act
snake_case : int = intermediate_size
snake_case : List[str] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Dict = max_position_embeddings
snake_case : List[str] = type_vocab_size
snake_case : List[str] = initializer_range
snake_case : Tuple = layer_norm_eps
snake_case : int = position_embedding_type
snake_case : str = use_cache
| 59
| 0
|
'''simple docstring'''
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 =logging.getLogger(__name__)
class UpperCAmelCase_ ( __lowercase ):
def __UpperCAmelCase ( self : str , UpperCAmelCase__ : Optional[int] , UpperCAmelCase__ : Optional[Any] , UpperCAmelCase__ : List[str]=None , UpperCAmelCase__ : Optional[Any]=None ) -> str:
lowerCAmelCase = self.layer[current_layer](UpperCAmelCase__ , UpperCAmelCase__ , head_mask[current_layer] )
lowerCAmelCase = 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.''' , __lowercase , )
class UpperCAmelCase_ ( __lowercase ):
def __init__( self : int , UpperCAmelCase__ : Tuple ) -> Dict:
super().__init__(UpperCAmelCase__ )
lowerCAmelCase = BertEncoderWithPabee(UpperCAmelCase__ )
self.init_weights()
lowerCAmelCase = 0
lowerCAmelCase = 0
lowerCAmelCase = 0
lowerCAmelCase = 0
def __UpperCAmelCase ( self : Tuple , UpperCAmelCase__ : Union[str, Any] ) -> Union[str, Any]:
lowerCAmelCase = threshold
def __UpperCAmelCase ( self : Union[str, Any] , UpperCAmelCase__ : Dict ) -> Dict:
lowerCAmelCase = patience
def __UpperCAmelCase ( self : Union[str, Any] ) -> List[str]:
lowerCAmelCase = 0
lowerCAmelCase = 0
def __UpperCAmelCase ( self : int ) -> str:
lowerCAmelCase = self.inference_layers_num / self.inference_instances_num
lowerCAmelCase = (
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(UpperCAmelCase__ )
@add_start_docstrings_to_model_forward(UpperCAmelCase__ )
def __UpperCAmelCase ( self : Tuple , UpperCAmelCase__ : List[Any]=None , UpperCAmelCase__ : Dict=None , UpperCAmelCase__ : Any=None , UpperCAmelCase__ : List[str]=None , UpperCAmelCase__ : Any=None , UpperCAmelCase__ : Any=None , UpperCAmelCase__ : List[Any]=None , UpperCAmelCase__ : Optional[int]=None , UpperCAmelCase__ : Any=None , UpperCAmelCase__ : Union[str, Any]=None , UpperCAmelCase__ : Dict=False , ) -> List[Any]:
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:
lowerCAmelCase = input_ids.size()
elif inputs_embeds is not None:
lowerCAmelCase = inputs_embeds.size()[:-1]
else:
raise ValueError('You have to specify either input_ids or inputs_embeds' )
lowerCAmelCase = input_ids.device if input_ids is not None else inputs_embeds.device
if attention_mask is None:
lowerCAmelCase = torch.ones(UpperCAmelCase__ , device=UpperCAmelCase__ )
if token_type_ids is None:
lowerCAmelCase = torch.zeros(UpperCAmelCase__ , dtype=torch.long , device=UpperCAmelCase__ )
# 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.
lowerCAmelCase = self.get_extended_attention_mask(UpperCAmelCase__ , UpperCAmelCase__ , UpperCAmelCase__ )
# 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:
lowerCAmelCase , lowerCAmelCase , lowerCAmelCase = encoder_hidden_states.size()
lowerCAmelCase = (encoder_batch_size, encoder_sequence_length)
if encoder_attention_mask is None:
lowerCAmelCase = torch.ones(UpperCAmelCase__ , device=UpperCAmelCase__ )
lowerCAmelCase = self.invert_attention_mask(UpperCAmelCase__ )
else:
lowerCAmelCase = 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]
lowerCAmelCase = self.get_head_mask(UpperCAmelCase__ , self.config.num_hidden_layers )
lowerCAmelCase = self.embeddings(
input_ids=UpperCAmelCase__ , position_ids=UpperCAmelCase__ , token_type_ids=UpperCAmelCase__ , inputs_embeds=UpperCAmelCase__ )
lowerCAmelCase = embedding_output
if self.training:
lowerCAmelCase = []
for i in range(self.config.num_hidden_layers ):
lowerCAmelCase = self.encoder.adaptive_forward(
UpperCAmelCase__ , current_layer=UpperCAmelCase__ , attention_mask=UpperCAmelCase__ , head_mask=UpperCAmelCase__ )
lowerCAmelCase = self.pooler(UpperCAmelCase__ )
lowerCAmelCase = output_layers[i](output_dropout(UpperCAmelCase__ ) )
res.append(UpperCAmelCase__ )
elif self.patience == 0: # Use all layers for inference
lowerCAmelCase = self.encoder(
UpperCAmelCase__ , attention_mask=UpperCAmelCase__ , head_mask=UpperCAmelCase__ , encoder_hidden_states=UpperCAmelCase__ , encoder_attention_mask=UpperCAmelCase__ , )
lowerCAmelCase = self.pooler(encoder_outputs[0] )
lowerCAmelCase = [output_layers[self.config.num_hidden_layers - 1](UpperCAmelCase__ )]
else:
lowerCAmelCase = 0
lowerCAmelCase = None
lowerCAmelCase = 0
for i in range(self.config.num_hidden_layers ):
calculated_layer_num += 1
lowerCAmelCase = self.encoder.adaptive_forward(
UpperCAmelCase__ , current_layer=UpperCAmelCase__ , attention_mask=UpperCAmelCase__ , head_mask=UpperCAmelCase__ )
lowerCAmelCase = self.pooler(UpperCAmelCase__ )
lowerCAmelCase = output_layers[i](UpperCAmelCase__ )
if regression:
lowerCAmelCase = logits.detach()
if patient_result is not None:
lowerCAmelCase = patient_result.detach()
if (patient_result is not None) and torch.abs(patient_result - labels ) < self.regression_threshold:
patient_counter += 1
else:
lowerCAmelCase = 0
else:
lowerCAmelCase = logits.detach().argmax(dim=1 )
if patient_result is not None:
lowerCAmelCase = patient_result.detach().argmax(dim=1 )
if (patient_result is not None) and torch.all(labels.eq(UpperCAmelCase__ ) ):
patient_counter += 1
else:
lowerCAmelCase = 0
lowerCAmelCase = logits
if patient_counter == self.patience:
break
lowerCAmelCase = [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
the pooled output) e.g. for GLUE tasks. ''' , __lowercase , )
class UpperCAmelCase_ ( __lowercase ):
def __init__( self : List[Any] , UpperCAmelCase__ : List[Any] ) -> List[str]:
super().__init__(UpperCAmelCase__ )
lowerCAmelCase = config.num_labels
lowerCAmelCase = BertModelWithPabee(UpperCAmelCase__ )
lowerCAmelCase = nn.Dropout(config.hidden_dropout_prob )
lowerCAmelCase = 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(UpperCAmelCase__ )
def __UpperCAmelCase ( self : List[str] , UpperCAmelCase__ : List[Any]=None , UpperCAmelCase__ : List[str]=None , UpperCAmelCase__ : Optional[int]=None , UpperCAmelCase__ : Tuple=None , UpperCAmelCase__ : List[str]=None , UpperCAmelCase__ : int=None , UpperCAmelCase__ : Any=None , ) -> Any:
lowerCAmelCase = self.bert(
input_ids=UpperCAmelCase__ , attention_mask=UpperCAmelCase__ , token_type_ids=UpperCAmelCase__ , position_ids=UpperCAmelCase__ , head_mask=UpperCAmelCase__ , inputs_embeds=UpperCAmelCase__ , output_dropout=self.dropout , output_layers=self.classifiers , regression=self.num_labels == 1 , )
lowerCAmelCase = (logits[-1],)
if labels is not None:
lowerCAmelCase = None
lowerCAmelCase = 0
for ix, logits_item in enumerate(UpperCAmelCase__ ):
if self.num_labels == 1:
# We are doing regression
lowerCAmelCase = MSELoss()
lowerCAmelCase = loss_fct(logits_item.view(-1 ) , labels.view(-1 ) )
else:
lowerCAmelCase = CrossEntropyLoss()
lowerCAmelCase = loss_fct(logits_item.view(-1 , self.num_labels ) , labels.view(-1 ) )
if total_loss is None:
lowerCAmelCase = loss
else:
total_loss += loss * (ix + 1)
total_weights += ix + 1
lowerCAmelCase = (total_loss / total_weights,) + outputs
return outputs
| 4
|
import gc
import unittest
from parameterized import parameterized
from diffusers import FlaxUNetaDConditionModel
from diffusers.utils import is_flax_available
from diffusers.utils.testing_utils import load_hf_numpy, require_flax, slow
if is_flax_available():
import jax
import jax.numpy as jnp
@slow
@require_flax
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, Any] , snake_case__ : List[str] ) -> List[str]:
'''simple docstring'''
return f"""gaussian_noise_s={seed}_shape={'_'.join([str(snake_case__ ) for s in shape] )}.npy"""
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> int:
'''simple docstring'''
super().tearDown()
gc.collect()
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Optional[Any]=0 , snake_case__ : Any=(4, 4, 64, 64) , snake_case__ : List[Any]=False ) -> int:
'''simple docstring'''
snake_case : Optional[Any] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Optional[int] = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return image
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Tuple=False , snake_case__ : List[Any]="CompVis/stable-diffusion-v1-4" ) -> List[Any]:
'''simple docstring'''
snake_case : List[str] = jnp.bfloataa if fpaa else jnp.floataa
snake_case : str = "bf16" if fpaa else None
snake_case , snake_case : Optional[int] = FlaxUNetaDConditionModel.from_pretrained(
snake_case__ , subfolder="unet" , dtype=snake_case__ , revision=snake_case__ )
return model, params
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any]=0 , snake_case__ : Union[str, Any]=(4, 77, 7_68) , snake_case__ : Dict=False ) -> List[str]:
'''simple docstring'''
snake_case : Any = jnp.bfloataa if fpaa else jnp.floataa
snake_case : Any = jnp.array(load_hf_numpy(self.get_file_format(snake_case__ , snake_case__ ) ) , dtype=snake_case__ )
return hidden_states
@parameterized.expand(
[
# fmt: off
[83, 4, [-0.2323, -0.1304, 0.0813, -0.3093, -0.0919, -0.1571, -0.1125, -0.5806]],
[17, 0.55, [-0.0831, -0.2443, 0.0901, -0.0919, 0.3396, 0.0103, -0.3743, 0.0701]],
[8, 0.89, [-0.4863, 0.0859, 0.0875, -0.1658, 0.9199, -0.0114, 0.4839, 0.4639]],
[3, 10_00, [-0.5649, 0.2402, -0.5518, 0.1248, 1.1328, -0.2443, -0.0325, -1.0078]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : int , snake_case__ : Optional[int] , snake_case__ : Dict ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : List[str] = self.get_unet_model(model_id="CompVis/stable-diffusion-v1-4" , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_latents(snake_case__ , fpaa=snake_case__ )
snake_case : List[str] = self.get_encoder_hidden_states(snake_case__ , fpaa=snake_case__ )
snake_case : Dict = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : Optional[Any] = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Optional[int] = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, in the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
@parameterized.expand(
[
# fmt: off
[83, 4, [0.1514, 0.0807, 0.1624, 0.1016, -0.1896, 0.0263, 0.0677, 0.2310]],
[17, 0.55, [0.1164, -0.0216, 0.0170, 0.1589, -0.3120, 0.1005, -0.0581, -0.1458]],
[8, 0.89, [-0.1758, -0.0169, 0.1004, -0.1411, 0.1312, 0.1103, -0.1996, 0.2139]],
[3, 10_00, [0.1214, 0.0352, -0.0731, -0.1562, -0.0994, -0.0906, -0.2340, -0.0539]],
# fmt: on
] )
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : Optional[int] , snake_case__ : str , snake_case__ : Tuple ) -> str:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.get_unet_model(model_id="stabilityai/stable-diffusion-2" , fpaa=snake_case__ )
snake_case : List[str] = self.get_latents(snake_case__ , shape=(4, 4, 96, 96) , fpaa=snake_case__ )
snake_case : Union[str, Any] = self.get_encoder_hidden_states(snake_case__ , shape=(4, 77, 10_24) , fpaa=snake_case__ )
snake_case : Optional[int] = model.apply(
{"params": params} , snake_case__ , jnp.array(snake_case__ , dtype=jnp.intaa ) , encoder_hidden_states=snake_case__ , ).sample
assert sample.shape == latents.shape
snake_case : int = jnp.asarray(jax.device_get((sample[-1, -2:, -2:, :2].flatten()) ) , dtype=jnp.floataa )
snake_case : Dict = jnp.array(snake_case__ , dtype=jnp.floataa )
# Found torch (float16) and flax (bfloat16) outputs to be within this tolerance, on the same hardware
assert jnp.allclose(snake_case__ , snake_case__ , atol=1e-2 )
| 59
| 0
|
# DISCLAIMER: This file is strongly influenced by https://github.com/yang-song/score_sde_pytorch
import math
from typing import Union
import torch
from ..configuration_utils import ConfigMixin, register_to_config
from ..utils import randn_tensor
from .scheduling_utils import SchedulerMixin
class lowerCamelCase__ ( lowerCAmelCase , lowerCAmelCase):
SCREAMING_SNAKE_CASE__ = 1
@register_to_config
def __init__(self , UpperCAmelCase=2_0_0_0 , UpperCAmelCase=0.1 , UpperCAmelCase=2_0 , UpperCAmelCase=1e-3 ) -> List[str]:
_lowercase =None
_lowercase =None
_lowercase =None
def __A (self , UpperCAmelCase , UpperCAmelCase = None ) -> str:
_lowercase =torch.linspace(1 , self.config.sampling_eps , UpperCAmelCase , device=UpperCAmelCase )
def __A (self , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase , UpperCAmelCase=None ) -> Optional[int]:
if self.timesteps is None:
raise ValueError(
'''`self.timesteps` is not set, you need to run \'set_timesteps\' after creating the scheduler''' )
# TODO(Patrick) better comments + non-PyTorch
# postprocess model score
_lowercase =(
-0.25 * t**2 * (self.config.beta_max - self.config.beta_min) - 0.5 * t * self.config.beta_min
)
_lowercase =torch.sqrt(1.0 - torch.exp(2.0 * log_mean_coeff ) )
_lowercase =std.flatten()
while len(std.shape ) < len(score.shape ):
_lowercase =std.unsqueeze(-1 )
_lowercase =-score / std
# compute
_lowercase =-1.0 / len(self.timesteps )
_lowercase =self.config.beta_min + t * (self.config.beta_max - self.config.beta_min)
_lowercase =beta_t.flatten()
while len(beta_t.shape ) < len(x.shape ):
_lowercase =beta_t.unsqueeze(-1 )
_lowercase =-0.5 * beta_t * x
_lowercase =torch.sqrt(UpperCAmelCase )
_lowercase =drift - diffusion**2 * score
_lowercase =x + drift * dt
# add noise
_lowercase =randn_tensor(x.shape , layout=x.layout , generator=UpperCAmelCase , device=x.device , dtype=x.dtype )
_lowercase =x_mean + diffusion * math.sqrt(-dt ) * noise
return x, x_mean
def __len__(self ) -> str:
return self.config.num_train_timesteps
| 5
|
import argparse
import re
from typing import Dict
import torch
from datasets import Audio, Dataset, load_dataset, load_metric
from transformers import AutoFeatureExtractor, pipeline
def UpperCamelCase ( __lowerCamelCase : Dataset , __lowerCamelCase : Dict[str, str] ):
snake_case : int = args.log_outputs
snake_case : Dict = "_".join(args.dataset.split("/" ) + [args.config, args.split] )
# load metric
snake_case : List[str] = load_metric("wer" )
snake_case : Tuple = load_metric("cer" )
# compute metrics
snake_case : List[Any] = wer.compute(references=result["target"] , predictions=result["prediction"] )
snake_case : int = cer.compute(references=result["target"] , predictions=result["prediction"] )
# print & log results
snake_case : int = f"""WER: {wer_result}\nCER: {cer_result}"""
print(__lowerCamelCase )
with open(f"""{dataset_id}_eval_results.txt""" , "w" ) as f:
f.write(__lowerCamelCase )
# log all results in text file. Possibly interesting for analysis
if log_outputs is not None:
snake_case : int = f"""log_{dataset_id}_predictions.txt"""
snake_case : List[Any] = f"""log_{dataset_id}_targets.txt"""
with open(__lowerCamelCase , "w" ) as p, open(__lowerCamelCase , "w" ) as t:
# mapping function to write output
def write_to_file(__lowerCamelCase : str , __lowerCamelCase : Optional[int] ):
p.write(f"""{i}""" + "\n" )
p.write(batch["prediction"] + "\n" )
t.write(f"""{i}""" + "\n" )
t.write(batch["target"] + "\n" )
result.map(__lowerCamelCase , with_indices=__lowerCamelCase )
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : List[Any] = "[,?.!\-\;\:\"“%‘”�—’…–]" # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
snake_case : List[Any] = re.sub(__lowerCamelCase , "" , text.lower() )
# In addition, we can normalize the target text, e.g. removing new lines characters etc...
# note that order is important here!
snake_case : Optional[Any] = ["\n\n", "\n", " ", " "]
for t in token_sequences_to_ignore:
snake_case : Dict = " ".join(text.split(__lowerCamelCase ) )
return text
def UpperCamelCase ( __lowerCamelCase : int ):
# load dataset
snake_case : str = load_dataset(args.dataset , args.config , split=args.split , use_auth_token=__lowerCamelCase )
# for testing: only process the first two examples as a test
# dataset = dataset.select(range(10))
# load processor
snake_case : List[Any] = AutoFeatureExtractor.from_pretrained(args.model_id )
snake_case : Union[str, Any] = feature_extractor.sampling_rate
# resample audio
snake_case : Union[str, Any] = dataset.cast_column("audio" , Audio(sampling_rate=__lowerCamelCase ) )
# load eval pipeline
if args.device is None:
snake_case : List[str] = 0 if torch.cuda.is_available() else -1
snake_case : str = pipeline("automatic-speech-recognition" , model=args.model_id , device=args.device )
# map function to decode audio
def map_to_pred(__lowerCamelCase : int ):
snake_case : Dict = asr(
batch["audio"]["array"] , chunk_length_s=args.chunk_length_s , stride_length_s=args.stride_length_s )
snake_case : str = prediction["text"]
snake_case : Tuple = normalize_text(batch["sentence"] )
return batch
# run inference on all examples
snake_case : Dict = dataset.map(__lowerCamelCase , remove_columns=dataset.column_names )
# compute and log_results
# do not change function below
log_results(__lowerCamelCase , __lowerCamelCase )
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
parser.add_argument(
"""--model_id""", type=str, required=True, help="""Model identifier. Should be loadable with 🤗 Transformers"""
)
parser.add_argument(
"""--dataset""",
type=str,
required=True,
help="""Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets""",
)
parser.add_argument(
"""--config""", type=str, required=True, help="""Config of the dataset. *E.g.* `'en'` for Common Voice"""
)
parser.add_argument("""--split""", type=str, required=True, help="""Split of the dataset. *E.g.* `'test'`""")
parser.add_argument(
"""--chunk_length_s""", type=float, default=None, help="""Chunk length in seconds. Defaults to 5 seconds."""
)
parser.add_argument(
"""--stride_length_s""", type=float, default=None, help="""Stride of the audio chunks. Defaults to 1 second."""
)
parser.add_argument(
"""--log_outputs""", action="""store_true""", help="""If defined, write outputs to log file for analysis."""
)
parser.add_argument(
"""--device""",
type=int,
default=None,
help="""The device to run the pipeline on. -1 for CPU (default), 0 for the first GPU and so on.""",
)
__lowerCamelCase = parser.parse_args()
main(args)
| 59
| 0
|
import random
import unittest
import torch
from diffusers import IFImgaImgSuperResolutionPipeline
from diffusers.utils import floats_tensor
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.testing_utils import skip_mps, torch_device
from ..pipeline_params import TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS, TEXT_GUIDED_IMAGE_VARIATION_PARAMS
from ..test_pipelines_common import PipelineTesterMixin
from . import IFPipelineTesterMixin
@skip_mps
class __A( a , a , unittest.TestCase ):
snake_case_ = IFImgaImgSuperResolutionPipeline
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {'''width''', '''height'''}
snake_case_ = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS.union({'''original_image'''} )
snake_case_ = PipelineTesterMixin.required_optional_params - {'''latents'''}
def SCREAMING_SNAKE_CASE_ ( self ) -> int:
'''simple docstring'''
return self._get_superresolution_dummy_components()
def SCREAMING_SNAKE_CASE_ ( self , _snake_case , _snake_case=0 ) -> int:
'''simple docstring'''
if str(_snake_case ).startswith('''mps''' ):
__a = torch.manual_seed(_snake_case )
else:
__a = torch.Generator(device=_snake_case ).manual_seed(_snake_case )
__a = floats_tensor((1, 3, 32, 32) , rng=random.Random(_snake_case ) ).to(_snake_case )
__a = floats_tensor((1, 3, 16, 16) , rng=random.Random(_snake_case ) ).to(_snake_case )
__a = {
'''prompt''': '''A painting of a squirrel eating a burger''',
'''image''': image,
'''original_image''': original_image,
'''generator''': generator,
'''num_inference_steps''': 2,
'''output_type''': '''numpy''',
}
return inputs
@unittest.skipIf(
torch_device != '''cuda''' or not is_xformers_available() , reason='''XFormers attention is only available with CUDA and `xformers` installed''' , )
def SCREAMING_SNAKE_CASE_ ( self ) -> List[Any]:
'''simple docstring'''
self._test_xformers_attention_forwardGenerator_pass(expected_max_diff=1E-3 )
def SCREAMING_SNAKE_CASE_ ( self ) -> Any:
'''simple docstring'''
self._test_save_load_optional_components()
@unittest.skipIf(torch_device != '''cuda''' , reason='''float16 requires CUDA''' )
def SCREAMING_SNAKE_CASE_ ( self ) -> Any:
'''simple docstring'''
super().test_save_load_floataa(expected_max_diff=1E-1 )
def SCREAMING_SNAKE_CASE_ ( self ) -> List[Any]:
'''simple docstring'''
self._test_attention_slicing_forward_pass(expected_max_diff=1E-2 )
def SCREAMING_SNAKE_CASE_ ( self ) -> Union[str, Any]:
'''simple docstring'''
self._test_save_load_local()
def SCREAMING_SNAKE_CASE_ ( self ) -> Optional[int]:
'''simple docstring'''
self._test_inference_batch_single_identical(
expected_max_diff=1E-2 , )
| 6
|
from typing import Optional, Tuple, Union
import flax
import flax.linen as nn
import jax
import jax.numpy as jnp
from flax.core.frozen_dict import FrozenDict
from ..configuration_utils import ConfigMixin, flax_register_to_config
from ..utils import BaseOutput
from .embeddings_flax import FlaxTimestepEmbedding, FlaxTimesteps
from .modeling_flax_utils import FlaxModelMixin
from .unet_ad_blocks_flax import (
FlaxCrossAttnDownBlockaD,
FlaxCrossAttnUpBlockaD,
FlaxDownBlockaD,
FlaxUNetMidBlockaDCrossAttn,
FlaxUpBlockaD,
)
@flax.struct.dataclass
class UpperCAmelCase ( A_ ):
A__ : jnp.ndarray
@flax_register_to_config
class UpperCAmelCase ( nn.Module ,A_ ,A_ ):
A__ : int = 32
A__ : int = 4
A__ : int = 4
A__ : Tuple[str] = (
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"CrossAttnDownBlock2D",
"DownBlock2D",
)
A__ : Tuple[str] = ("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D")
A__ : Union[bool, Tuple[bool]] = False
A__ : Tuple[int] = (3_20, 6_40, 12_80, 12_80)
A__ : int = 2
A__ : Union[int, Tuple[int]] = 8
A__ : Optional[Union[int, Tuple[int]]] = None
A__ : int = 12_80
A__ : float = 0.0
A__ : bool = False
A__ : jnp.dtype = jnp.floataa
A__ : bool = True
A__ : int = 0
A__ : bool = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : jax.random.KeyArray ) -> FrozenDict:
'''simple docstring'''
snake_case : Dict = (1, self.in_channels, self.sample_size, self.sample_size)
snake_case : Any = jnp.zeros(snake_case__ , dtype=jnp.floataa )
snake_case : List[str] = jnp.ones((1,) , dtype=jnp.intaa )
snake_case : str = jnp.zeros((1, 1, self.cross_attention_dim) , dtype=jnp.floataa )
snake_case , snake_case : Optional[int] = jax.random.split(snake_case__ )
snake_case : Union[str, Any] = {"params": params_rng, "dropout": dropout_rng}
return self.init(snake_case__ , snake_case__ , snake_case__ , snake_case__ )["params"]
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : str = self.block_out_channels
snake_case : Optional[Any] = block_out_channels[0] * 4
if self.num_attention_heads is not None:
raise ValueError(
"At the moment it is not possible to define the number of attention heads via `num_attention_heads` because of a naming issue as described in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131. Passing `num_attention_heads` will only be supported in diffusers v0.19." )
# If `num_attention_heads` is not defined (which is the case for most models)
# it will default to `attention_head_dim`. This looks weird upon first reading it and it is.
# The reason for this behavior is to correct for incorrectly named variables that were introduced
# when this library was created. The incorrect naming was only discovered much later in https://github.com/huggingface/diffusers/issues/2011#issuecomment-1547958131
# Changing `attention_head_dim` to `num_attention_heads` for 40,000+ configurations is too backwards breaking
# which is why we correct for the naming here.
snake_case : Tuple = self.num_attention_heads or self.attention_head_dim
# input
snake_case : Tuple = nn.Conv(
block_out_channels[0] , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
# time
snake_case : Union[str, Any] = FlaxTimesteps(
block_out_channels[0] , flip_sin_to_cos=self.flip_sin_to_cos , freq_shift=self.config.freq_shift )
snake_case : Dict = FlaxTimestepEmbedding(snake_case__ , dtype=self.dtype )
snake_case : List[str] = self.only_cross_attention
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (only_cross_attention,) * len(self.down_block_types )
if isinstance(snake_case__ , snake_case__ ):
snake_case : List[Any] = (num_attention_heads,) * len(self.down_block_types )
# down
snake_case : List[Any] = []
snake_case : Optional[int] = block_out_channels[0]
for i, down_block_type in enumerate(self.down_block_types ):
snake_case : List[Any] = output_channel
snake_case : Dict = block_out_channels[i]
snake_case : Optional[Any] = i == len(snake_case__ ) - 1
if down_block_type == "CrossAttnDownBlock2D":
snake_case : List[Any] = FlaxCrossAttnDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , num_attention_heads=num_attention_heads[i] , add_downsample=not is_final_block , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Union[str, Any] = FlaxDownBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , dropout=self.dropout , num_layers=self.layers_per_block , add_downsample=not is_final_block , dtype=self.dtype , )
down_blocks.append(snake_case__ )
snake_case : Dict = down_blocks
# mid
snake_case : Optional[int] = FlaxUNetMidBlockaDCrossAttn(
in_channels=block_out_channels[-1] , dropout=self.dropout , num_attention_heads=num_attention_heads[-1] , use_linear_projection=self.use_linear_projection , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
# up
snake_case : Optional[Any] = []
snake_case : Optional[int] = list(reversed(snake_case__ ) )
snake_case : Dict = list(reversed(snake_case__ ) )
snake_case : Tuple = list(reversed(snake_case__ ) )
snake_case : Optional[Any] = reversed_block_out_channels[0]
for i, up_block_type in enumerate(self.up_block_types ):
snake_case : Optional[int] = output_channel
snake_case : List[Any] = reversed_block_out_channels[i]
snake_case : Union[str, Any] = reversed_block_out_channels[min(i + 1 , len(snake_case__ ) - 1 )]
snake_case : int = i == len(snake_case__ ) - 1
if up_block_type == "CrossAttnUpBlock2D":
snake_case : Any = FlaxCrossAttnUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , num_attention_heads=reversed_num_attention_heads[i] , add_upsample=not is_final_block , dropout=self.dropout , use_linear_projection=self.use_linear_projection , only_cross_attention=only_cross_attention[i] , use_memory_efficient_attention=self.use_memory_efficient_attention , dtype=self.dtype , )
else:
snake_case : Optional[int] = FlaxUpBlockaD(
in_channels=snake_case__ , out_channels=snake_case__ , prev_output_channel=snake_case__ , num_layers=self.layers_per_block + 1 , add_upsample=not is_final_block , dropout=self.dropout , dtype=self.dtype , )
up_blocks.append(snake_case__ )
snake_case : Optional[int] = output_channel
snake_case : Tuple = up_blocks
# out
snake_case : Optional[int] = nn.GroupNorm(num_groups=32 , epsilon=1e-5 )
snake_case : List[str] = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__(self : Dict , snake_case__ : Dict , snake_case__ : Dict , snake_case__ : Optional[int] , snake_case__ : Tuple=None , snake_case__ : Union[str, Any]=None , snake_case__ : bool = True , snake_case__ : bool = False , ) -> Union[FlaxUNetaDConditionOutput, Tuple]:
'''simple docstring'''
if not isinstance(snake_case__ , jnp.ndarray ):
snake_case : List[Any] = jnp.array([timesteps] , dtype=jnp.intaa )
elif isinstance(snake_case__ , jnp.ndarray ) and len(timesteps.shape ) == 0:
snake_case : Any = timesteps.astype(dtype=jnp.floataa )
snake_case : int = jnp.expand_dims(snake_case__ , 0 )
snake_case : str = self.time_proj(snake_case__ )
snake_case : str = self.time_embedding(snake_case__ )
# 2. pre-process
snake_case : int = jnp.transpose(snake_case__ , (0, 2, 3, 1) )
snake_case : List[Any] = self.conv_in(snake_case__ )
# 3. down
snake_case : Optional[int] = (sample,)
for down_block in self.down_blocks:
if isinstance(snake_case__ , snake_case__ ):
snake_case , snake_case : List[Any] = down_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
else:
snake_case , snake_case : str = down_block(snake_case__ , snake_case__ , deterministic=not train )
down_block_res_samples += res_samples
if down_block_additional_residuals is not None:
snake_case : Tuple = ()
for down_block_res_sample, down_block_additional_residual in zip(
snake_case__ , snake_case__ ):
down_block_res_sample += down_block_additional_residual
new_down_block_res_samples += (down_block_res_sample,)
snake_case : Optional[int] = new_down_block_res_samples
# 4. mid
snake_case : Optional[int] = self.mid_block(snake_case__ , snake_case__ , snake_case__ , deterministic=not train )
if mid_block_additional_residual is not None:
sample += mid_block_additional_residual
# 5. up
for up_block in self.up_blocks:
snake_case : int = down_block_res_samples[-(self.layers_per_block + 1) :]
snake_case : Optional[Any] = down_block_res_samples[: -(self.layers_per_block + 1)]
if isinstance(snake_case__ , snake_case__ ):
snake_case : Optional[Any] = up_block(
snake_case__ , temb=snake_case__ , encoder_hidden_states=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train , )
else:
snake_case : Dict = up_block(snake_case__ , temb=snake_case__ , res_hidden_states_tuple=snake_case__ , deterministic=not train )
# 6. post-process
snake_case : List[str] = self.conv_norm_out(snake_case__ )
snake_case : Any = nn.silu(snake_case__ )
snake_case : Optional[int] = self.conv_out(snake_case__ )
snake_case : Union[str, Any] = jnp.transpose(snake_case__ , (0, 3, 1, 2) )
if not return_dict:
return (sample,)
return FlaxUNetaDConditionOutput(sample=snake_case__ )
| 59
| 0
|
import inspect
import unittest
from transformers import RegNetConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from transformers.utils import cached_property, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor
if is_flax_available():
import jax
import jax.numpy as jnp
from transformers.models.regnet.modeling_flax_regnet import FlaxRegNetForImageClassification, FlaxRegNetModel
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class A ( unittest.TestCase ):
"""simple docstring"""
def __init__( self : Union[str, Any],lowercase_ : str,lowercase_ : Optional[int]=3,lowercase_ : Optional[Any]=3_2,lowercase_ : str=3,lowercase_ : List[str]=1_0,lowercase_ : List[Any]=[1_0, 2_0, 3_0, 4_0],lowercase_ : Dict=[1, 1, 2, 1],lowercase_ : List[str]=True,lowercase_ : Tuple=True,lowercase_ : Optional[Any]="relu",lowercase_ : Tuple=3,lowercase_ : Any=None,)-> Tuple:
'''simple docstring'''
A__ = parent
A__ = batch_size
A__ = image_size
A__ = num_channels
A__ = embeddings_size
A__ = hidden_sizes
A__ = depths
A__ = is_training
A__ = use_labels
A__ = hidden_act
A__ = num_labels
A__ = scope
A__ = len(lowercase_ )
def snake_case__ ( self : Any )-> List[Any]:
'''simple docstring'''
A__ = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
A__ = self.get_config()
return config, pixel_values
def snake_case__ ( self : Any )-> Dict:
'''simple docstring'''
return RegNetConfig(
num_channels=self.num_channels,embeddings_size=self.embeddings_size,hidden_sizes=self.hidden_sizes,depths=self.depths,hidden_act=self.hidden_act,num_labels=self.num_labels,image_size=self.image_size,)
def snake_case__ ( self : List[Any],lowercase_ : Optional[Any],lowercase_ : List[str] )-> List[str]:
'''simple docstring'''
A__ = FlaxRegNetModel(config=lowercase_ )
A__ = model(lowercase_ )
# Output shape (b, c, h, w)
self.parent.assertEqual(
result.last_hidden_state.shape,(self.batch_size, self.hidden_sizes[-1], self.image_size // 3_2, self.image_size // 3_2),)
def snake_case__ ( self : Any,lowercase_ : int,lowercase_ : List[str] )-> Optional[Any]:
'''simple docstring'''
A__ = self.num_labels
A__ = FlaxRegNetForImageClassification(config=lowercase_ )
A__ = model(lowercase_ )
self.parent.assertEqual(result.logits.shape,(self.batch_size, self.num_labels) )
def snake_case__ ( self : Tuple )-> str:
'''simple docstring'''
A__ = self.prepare_config_and_inputs()
A__ , A__ = config_and_inputs
A__ = {'pixel_values': pixel_values}
return config, inputs_dict
@require_flax
class A ( _UpperCAmelCase , unittest.TestCase ):
"""simple docstring"""
lowerCamelCase = (FlaxRegNetModel, FlaxRegNetForImageClassification) if is_flax_available() else ()
lowerCamelCase = False
lowerCamelCase = False
lowerCamelCase = False
def snake_case__ ( self : Tuple )-> None:
'''simple docstring'''
A__ = FlaxRegNetModelTester(self )
A__ = ConfigTester(self,config_class=lowercase_,has_text_modality=lowercase_ )
def snake_case__ ( self : Union[str, Any] )-> List[Any]:
'''simple docstring'''
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def snake_case__ ( self : List[str] )-> int:
'''simple docstring'''
return
def snake_case__ ( self : str )-> Union[str, Any]:
'''simple docstring'''
A__ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*lowercase_ )
def snake_case__ ( self : str )-> Tuple:
'''simple docstring'''
A__ = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*lowercase_ )
@unittest.skip(reason='RegNet does not use inputs_embeds' )
def snake_case__ ( self : int )-> Optional[Any]:
'''simple docstring'''
pass
@unittest.skip(reason='RegNet does not support input and output embeddings' )
def snake_case__ ( self : Optional[Any] )-> str:
'''simple docstring'''
pass
def snake_case__ ( self : int )-> Dict:
'''simple docstring'''
A__ , A__ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
A__ = model_class(lowercase_ )
A__ = inspect.signature(model.__call__ )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
A__ = [*signature.parameters.keys()]
A__ = ['pixel_values']
self.assertListEqual(arg_names[:1],lowercase_ )
def snake_case__ ( self : Optional[Any] )-> Optional[Any]:
'''simple docstring'''
def check_hidden_states_output(lowercase_ : List[str],lowercase_ : List[Any],lowercase_ : Optional[int] ):
A__ = model_class(lowercase_ )
A__ = model(**self._prepare_for_class(lowercase_,lowercase_ ) )
A__ = outputs.encoder_hidden_states if config.is_encoder_decoder else outputs.hidden_states
A__ = self.model_tester.num_stages
self.assertEqual(len(lowercase_ ),expected_num_stages + 1 )
A__ , A__ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
A__ = True
check_hidden_states_output(lowercase_,lowercase_,lowercase_ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
A__ = True
check_hidden_states_output(lowercase_,lowercase_,lowercase_ )
def snake_case__ ( self : Tuple )-> Tuple:
'''simple docstring'''
A__ , A__ = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
with self.subTest(model_class.__name__ ):
A__ = self._prepare_for_class(lowercase_,lowercase_ )
A__ = model_class(lowercase_ )
@jax.jit
def model_jitted(lowercase_ : int,**lowercase_ : Dict ):
return model(pixel_values=lowercase_,**lowercase_ )
with self.subTest('JIT Enabled' ):
A__ = model_jitted(**lowercase_ ).to_tuple()
with self.subTest('JIT Disabled' ):
with jax.disable_jit():
A__ = model_jitted(**lowercase_ ).to_tuple()
self.assertEqual(len(lowercase_ ),len(lowercase_ ) )
for jitted_output, output in zip(lowercase_,lowercase_ ):
self.assertEqual(jitted_output.shape,output.shape )
def _snake_case( ) -> Dict:
'''simple docstring'''
A__ = Image.open('./tests/fixtures/tests_samples/COCO/000000039769.png' )
return image
@require_flax
class A ( unittest.TestCase ):
"""simple docstring"""
@cached_property
def snake_case__ ( self : Dict )-> Optional[int]:
'''simple docstring'''
return AutoImageProcessor.from_pretrained('facebook/regnet-y-040' ) if is_vision_available() else None
@slow
def snake_case__ ( self : Union[str, Any] )-> str:
'''simple docstring'''
A__ = FlaxRegNetForImageClassification.from_pretrained('facebook/regnet-y-040' )
A__ = self.default_image_processor
A__ = prepare_img()
A__ = image_processor(images=lowercase_,return_tensors='np' )
A__ = model(**lowercase_ )
# verify the logits
A__ = (1, 1_0_0_0)
self.assertEqual(outputs.logits.shape,lowercase_ )
A__ = jnp.array([-0.4_180, -1.5_051, -3.4_836] )
self.assertTrue(jnp.allclose(outputs.logits[0, :3],lowercase_,atol=1E-4 ) )
| 7
|
__lowerCamelCase = {
"joule": 1.0,
"kilojoule": 10_00,
"megajoule": 1_00_00_00,
"gigajoule": 10_00_00_00_00,
"wattsecond": 1.0,
"watthour": 36_00,
"kilowatthour": 3_60_00_00,
"newtonmeter": 1.0,
"calorie_nutr": 41_86.8,
"kilocalorie_nutr": 4_18_68_00.00,
"electronvolt": 1.602_176_634e-19,
"britishthermalunit_it": 10_55.0_55_85,
"footpound": 1.35_5818,
}
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : float ):
if to_type not in ENERGY_CONVERSION or from_type not in ENERGY_CONVERSION:
snake_case : List[Any] = (
f"""Incorrect 'from_type' or 'to_type' value: {from_type!r}, {to_type!r}\n"""
f"""Valid values are: {', '.join(__lowerCamelCase )}"""
)
raise ValueError(__lowerCamelCase )
return value * ENERGY_CONVERSION[from_type] / ENERGY_CONVERSION[to_type]
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 0
|
from __future__ import annotations
from typing import Dict
from ...configuration_utils import PretrainedConfig
lowerCAmelCase_ = {
'''susnato/ernie-m-base_pytorch''': '''https://huggingface.co/susnato/ernie-m-base_pytorch/blob/main/config.json''',
'''susnato/ernie-m-large_pytorch''': '''https://huggingface.co/susnato/ernie-m-large_pytorch/blob/main/config.json''',
}
class snake_case_ ( __A ):
'''simple docstring'''
SCREAMING_SNAKE_CASE : Dict = "ernie_m"
SCREAMING_SNAKE_CASE : Dict[str, str] = {"dropout": "classifier_dropout", "num_classes": "num_labels"}
def __init__( self : List[str] , _UpperCamelCase : int = 2_5_0_0_0_2 , _UpperCamelCase : int = 7_6_8 , _UpperCamelCase : int = 1_2 , _UpperCamelCase : int = 1_2 , _UpperCamelCase : int = 3_0_7_2 , _UpperCamelCase : str = "gelu" , _UpperCamelCase : float = 0.1 , _UpperCamelCase : float = 0.1 , _UpperCamelCase : int = 5_1_4 , _UpperCamelCase : float = 0.02 , _UpperCamelCase : int = 1 , _UpperCamelCase : float = 1e-05 , _UpperCamelCase : List[str]=None , _UpperCamelCase : Union[str, Any]=False , _UpperCamelCase : Optional[int]=0.0 , **_UpperCamelCase : Optional[Any] , ) ->Union[str, Any]:
super().__init__(pad_token_id=_UpperCamelCase , **_UpperCamelCase )
snake_case_ = vocab_size
snake_case_ = hidden_size
snake_case_ = num_hidden_layers
snake_case_ = num_attention_heads
snake_case_ = intermediate_size
snake_case_ = hidden_act
snake_case_ = hidden_dropout_prob
snake_case_ = attention_probs_dropout_prob
snake_case_ = max_position_embeddings
snake_case_ = initializer_range
snake_case_ = layer_norm_eps
snake_case_ = classifier_dropout
snake_case_ = is_decoder
snake_case_ = act_dropout
| 8
|
import logging
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
import datasets
import numpy as np
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PreTrainedTokenizer,
TFAutoModelForSequenceClassification,
TFTrainer,
TFTrainingArguments,
)
from transformers.utils import logging as hf_logging
hf_logging.set_verbosity_info()
hf_logging.enable_default_handler()
hf_logging.enable_explicit_format()
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : str , __lowerCamelCase : PreTrainedTokenizer , __lowerCamelCase : int , __lowerCamelCase : Optional[int] = None , ):
snake_case : int = {}
if train_file is not None:
snake_case : List[Any] = [train_file]
if eval_file is not None:
snake_case : Optional[int] = [eval_file]
if test_file is not None:
snake_case : Any = [test_file]
snake_case : int = datasets.load_dataset("csv" , data_files=__lowerCamelCase )
snake_case : str = list(ds[list(files.keys() )[0]].features.keys() )
snake_case : int = features_name.pop(__lowerCamelCase )
snake_case : str = list(set(ds[list(files.keys() )[0]][label_name] ) )
snake_case : str = {label: i for i, label in enumerate(__lowerCamelCase )}
snake_case : List[Any] = tokenizer.model_input_names
snake_case : List[Any] = {}
if len(__lowerCamelCase ) == 1:
for k in files.keys():
snake_case : Tuple = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
example[features_name[0]] , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" ) , batched=__lowerCamelCase , )
elif len(__lowerCamelCase ) == 2:
for k in files.keys():
snake_case : List[Any] = ds[k].map(
lambda __lowerCamelCase : tokenizer.batch_encode_plus(
(example[features_name[0]], example[features_name[1]]) , truncation=__lowerCamelCase , max_length=__lowerCamelCase , padding="max_length" , ) , batched=__lowerCamelCase , )
def gen_train():
for ex in transformed_ds[datasets.Split.TRAIN]:
snake_case : Dict = {k: v for k, v in ex.items() if k in input_names}
snake_case : Union[str, Any] = labelaid[ex[label_name]]
yield (d, label)
def gen_val():
for ex in transformed_ds[datasets.Split.VALIDATION]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : Any = labelaid[ex[label_name]]
yield (d, label)
def gen_test():
for ex in transformed_ds[datasets.Split.TEST]:
snake_case : str = {k: v for k, v in ex.items() if k in input_names}
snake_case : List[str] = labelaid[ex[label_name]]
yield (d, label)
snake_case : int = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TRAIN in transformed_ds
else None
)
if train_ds is not None:
snake_case : Optional[Any] = train_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TRAIN] ) ) )
snake_case : Tuple = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.VALIDATION in transformed_ds
else None
)
if val_ds is not None:
snake_case : List[str] = val_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.VALIDATION] ) ) )
snake_case : Optional[int] = (
tf.data.Dataset.from_generator(
__lowerCamelCase , ({k: tf.intaa for k in input_names}, tf.intaa) , ({k: tf.TensorShape([None] ) for k in input_names}, tf.TensorShape([] )) , )
if datasets.Split.TEST in transformed_ds
else None
)
if test_ds is not None:
snake_case : str = test_ds.apply(tf.data.experimental.assert_cardinality(len(ds[datasets.Split.TEST] ) ) )
return train_ds, val_ds, test_ds, labelaid
__lowerCamelCase = logging.getLogger(__name__)
@dataclass
class UpperCAmelCase :
A__ : int = field(metadata={"help": "Which column contains the label"} )
A__ : str = field(default=A_ ,metadata={"help": "The path of the training file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the development file"} )
A__ : Optional[str] = field(default=A_ ,metadata={"help": "The path of the test file"} )
A__ : int = field(
default=1_28 ,metadata={
"help": (
"The maximum total input sequence length after tokenization. Sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
} ,)
A__ : bool = field(
default=A_ ,metadata={"help": "Overwrite the cached training and evaluation sets"} )
@dataclass
class UpperCAmelCase :
A__ : str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained config name or path if not the same as model_name"} )
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"} )
A__ : bool = field(default=A_ ,metadata={"help": "Set this flag to use fast tokenization."} )
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
# or just modify its tokenizer_config.json.
A__ : Optional[str] = field(
default=A_ ,metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"} ,)
def UpperCamelCase ( ):
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
snake_case : List[Any] = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments) )
snake_case , snake_case , snake_case : int = parser.parse_args_into_dataclasses()
if (
os.path.exists(training_args.output_dir )
and os.listdir(training_args.output_dir )
and training_args.do_train
and not training_args.overwrite_output_dir
):
raise ValueError(
f"""Output directory ({training_args.output_dir}) already exists and is not empty. Use"""
" --overwrite_output_dir to overcome." )
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s" , datefmt="%m/%d/%Y %H:%M:%S" , level=logging.INFO , )
logger.info(
f"""n_replicas: {training_args.n_replicas}, distributed training: {bool(training_args.n_replicas > 1 )}, """
f"""16-bits training: {training_args.fpaa}""" )
logger.info(f"""Training/evaluation parameters {training_args}""" )
# Load pretrained model and tokenizer
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
snake_case : Tuple = AutoTokenizer.from_pretrained(
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path , cache_dir=model_args.cache_dir , )
snake_case , snake_case , snake_case , snake_case : Tuple = get_tfds(
train_file=data_args.train_file , eval_file=data_args.dev_file , test_file=data_args.test_file , tokenizer=__lowerCamelCase , label_column_id=data_args.label_column_id , max_seq_length=data_args.max_seq_length , )
snake_case : Optional[Any] = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path , num_labels=len(__lowerCamelCase ) , labelaid=__lowerCamelCase , idalabel={id: label for label, id in labelaid.items()} , finetuning_task="text-classification" , cache_dir=model_args.cache_dir , )
with training_args.strategy.scope():
snake_case : int = TFAutoModelForSequenceClassification.from_pretrained(
model_args.model_name_or_path , from_pt=bool(".bin" in model_args.model_name_or_path ) , config=__lowerCamelCase , cache_dir=model_args.cache_dir , )
def compute_metrics(__lowerCamelCase : EvalPrediction ) -> Dict:
snake_case : Optional[int] = np.argmax(p.predictions , axis=1 )
return {"acc": (preds == p.label_ids).mean()}
# Initialize our Trainer
snake_case : int = TFTrainer(
model=__lowerCamelCase , args=__lowerCamelCase , train_dataset=__lowerCamelCase , eval_dataset=__lowerCamelCase , compute_metrics=__lowerCamelCase , )
# Training
if training_args.do_train:
trainer.train()
trainer.save_model()
tokenizer.save_pretrained(training_args.output_dir )
# Evaluation
snake_case : int = {}
if training_args.do_eval:
logger.info("*** Evaluate ***" )
snake_case : Any = trainer.evaluate()
snake_case : List[Any] = os.path.join(training_args.output_dir , "eval_results.txt" )
with open(__lowerCamelCase , "w" ) as writer:
logger.info("***** Eval results *****" )
for key, value in result.items():
logger.info(f""" {key} = {value}""" )
writer.write(f"""{key} = {value}\n""" )
results.update(__lowerCamelCase )
return results
if __name__ == "__main__":
main()
| 59
| 0
|
from __future__ import annotations
def _UpperCamelCase ( lowercase__ ): # This function is recursive
__SCREAMING_SNAKE_CASE : Optional[int] = len(lowercase__ )
# If the array contains only one element, we return it (it's the stop condition of
# recursion)
if array_length <= 1:
return array
# Else
__SCREAMING_SNAKE_CASE : Dict = array[0]
__SCREAMING_SNAKE_CASE : Tuple = False
__SCREAMING_SNAKE_CASE : List[str] = 1
__SCREAMING_SNAKE_CASE : list[int] = []
while not is_found and i < array_length:
if array[i] < pivot:
__SCREAMING_SNAKE_CASE : List[Any] = True
__SCREAMING_SNAKE_CASE : Dict = [element for element in array[i:] if element >= array[i]]
__SCREAMING_SNAKE_CASE : Optional[Any] = longest_subsequence(lowercase__ )
if len(lowercase__ ) > len(lowercase__ ):
__SCREAMING_SNAKE_CASE : Optional[int] = temp_array
else:
i += 1
__SCREAMING_SNAKE_CASE : Dict = [element for element in array[1:] if element >= pivot]
__SCREAMING_SNAKE_CASE : Tuple = [pivot, *longest_subsequence(lowercase__ )]
if len(lowercase__ ) > len(lowercase__ ):
return temp_array
else:
return longest_subseq
if __name__ == "__main__":
import doctest
doctest.testmod()
| 9
|
import json
import os
import shutil
import tempfile
import unittest
import numpy as np
from transformers import BertTokenizerFast
from transformers.models.bert.tokenization_bert import VOCAB_FILES_NAMES, BertTokenizer
from transformers.testing_utils import require_tokenizers, require_vision
from transformers.utils import IMAGE_PROCESSOR_NAME, is_vision_available
if is_vision_available():
from PIL import Image
from transformers import VisionTextDualEncoderProcessor, ViTImageProcessor
@require_tokenizers
@require_vision
class UpperCAmelCase ( unittest.TestCase ):
def _SCREAMING_SNAKE_CASE (self : Any ) -> List[str]:
'''simple docstring'''
snake_case : int = tempfile.mkdtemp()
# fmt: off
snake_case : Optional[int] = ["[UNK]", "[CLS]", "[SEP]", "[PAD]", "[MASK]", "want", "##want", "##ed", "wa", "un", "runn", "##ing", ",", "low", "lowest"]
# fmt: on
snake_case : List[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] ) )
snake_case : int = {
"do_resize": True,
"size": {"height": 18, "width": 18},
"do_normalize": True,
"image_mean": [0.5, 0.5, 0.5],
"image_std": [0.5, 0.5, 0.5],
}
snake_case : Optional[Any] = os.path.join(self.tmpdirname , snake_case__ )
with open(self.image_processor_file , "w" , encoding="utf-8" ) as fp:
json.dump(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : str ) -> Optional[int]:
'''simple docstring'''
return BertTokenizer.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , **snake_case__ : List[str] ) -> int:
'''simple docstring'''
return ViTImageProcessor.from_pretrained(self.tmpdirname , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
shutil.rmtree(self.tmpdirname )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : List[Any] = [np.random.randint(2_55 , size=(3, 30, 4_00) , dtype=np.uinta )]
snake_case : Optional[int] = [Image.fromarray(np.moveaxis(snake_case__ , 0 , -1 ) ) for x in image_inputs]
return image_inputs
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.get_tokenizer()
snake_case : Optional[Any] = self.get_image_processor()
snake_case : Optional[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
processor.save_pretrained(self.tmpdirname )
snake_case : Any = VisionTextDualEncoderProcessor.from_pretrained(self.tmpdirname )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : str = VisionTextDualEncoderProcessor(
tokenizer=self.get_tokenizer() , image_processor=self.get_image_processor() )
processor.save_pretrained(self.tmpdirname )
snake_case : Optional[int] = self.get_tokenizer(bos_token="(BOS)" , eos_token="(EOS)" )
snake_case : Tuple = self.get_image_processor(do_normalize=snake_case__ , padding_value=1.0 )
snake_case : List[str] = VisionTextDualEncoderProcessor.from_pretrained(
self.tmpdirname , bos_token="(BOS)" , eos_token="(EOS)" , do_normalize=snake_case__ , padding_value=1.0 )
self.assertEqual(processor.tokenizer.get_vocab() , tokenizer_add_kwargs.get_vocab() )
self.assertIsInstance(processor.tokenizer , (BertTokenizer, BertTokenizerFast) )
self.assertEqual(processor.image_processor.to_json_string() , image_processor_add_kwargs.to_json_string() )
self.assertIsInstance(processor.image_processor , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> int:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Optional[int] = self.get_tokenizer()
snake_case : List[Any] = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = self.prepare_image_inputs()
snake_case : str = image_processor(snake_case__ , return_tensors="np" )
snake_case : Any = processor(images=snake_case__ , return_tensors="np" )
for key in input_feat_extract.keys():
self.assertAlmostEqual(input_feat_extract[key].sum() , input_processor[key].sum() , delta=1e-2 )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.get_image_processor()
snake_case : int = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Tuple = "lower newer"
snake_case : Tuple = processor(text=snake_case__ )
snake_case : Union[str, Any] = tokenizer(snake_case__ )
for key in encoded_tok.keys():
self.assertListEqual(encoded_tok[key] , encoded_processor[key] )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Optional[int]:
'''simple docstring'''
snake_case : List[Any] = self.get_image_processor()
snake_case : Dict = self.get_tokenizer()
snake_case : Dict = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : int = "lower newer"
snake_case : Dict = self.prepare_image_inputs()
snake_case : Union[str, Any] = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , ["input_ids", "token_type_ids", "attention_mask", "pixel_values"] )
# test if it raises when no input is passed
with self.assertRaises(snake_case__ ):
processor()
def _SCREAMING_SNAKE_CASE (self : str ) -> Tuple:
'''simple docstring'''
snake_case : Tuple = self.get_image_processor()
snake_case : Optional[Any] = self.get_tokenizer()
snake_case : Tuple = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : List[str] = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9]]
snake_case : List[Any] = processor.batch_decode(snake_case__ )
snake_case : Union[str, Any] = tokenizer.batch_decode(snake_case__ )
self.assertListEqual(snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = self.get_image_processor()
snake_case : Union[str, Any] = self.get_tokenizer()
snake_case : Any = VisionTextDualEncoderProcessor(tokenizer=snake_case__ , image_processor=snake_case__ )
snake_case : Optional[Any] = "lower newer"
snake_case : List[Any] = self.prepare_image_inputs()
snake_case : Tuple = processor(text=snake_case__ , images=snake_case__ )
self.assertListEqual(list(inputs.keys() ) , processor.model_input_names )
| 59
| 0
|
import os
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
__A = logging.get_logger(__name__)
__A = "▁"
__A = {"vocab_file": "sentencepiece.bpe.model"}
__A = {
"vocab_file": {
"facebook/xglm-564M": "https://huggingface.co/facebook/xglm-564M/resolve/main/sentencepiece.bpe.model",
}
}
__A = {
"facebook/xglm-564M": 2048,
}
class _SCREAMING_SNAKE_CASE ( __SCREAMING_SNAKE_CASE ):
'''simple docstring'''
lowercase_ = VOCAB_FILES_NAMES
lowercase_ = PRETRAINED_VOCAB_FILES_MAP
lowercase_ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
lowercase_ = ["input_ids", "attention_mask"]
def __init__(self : List[Any] , UpperCAmelCase_ : List[Any] , UpperCAmelCase_ : Any="<s>" , UpperCAmelCase_ : Optional[int]="</s>" , UpperCAmelCase_ : List[Any]="</s>" , UpperCAmelCase_ : Union[str, Any]="<s>" , UpperCAmelCase_ : Any="<unk>" , UpperCAmelCase_ : str="<pad>" , UpperCAmelCase_ : Optional[Dict[str, Any]] = None , **UpperCAmelCase_ : str , ) ->None:
'''simple docstring'''
lowerCamelCase__: Dict ={} if sp_model_kwargs is None else sp_model_kwargs
# Compatibility with the original tokenizer
lowerCamelCase__: Tuple =7
lowerCamelCase__: int =[F"""<madeupword{i}>""" for i in range(self.num_madeup_words)]
lowerCamelCase__: str =kwargs.get("additional_special_tokens" , [])
kwargs["additional_special_tokens"] += [
word for word in madeup_words if word not in kwargs["additional_special_tokens"]
]
super().__init__(
bos_token=UpperCAmelCase_ , eos_token=UpperCAmelCase_ , unk_token=UpperCAmelCase_ , sep_token=UpperCAmelCase_ , cls_token=UpperCAmelCase_ , pad_token=UpperCAmelCase_ , sp_model_kwargs=self.sp_model_kwargs , **UpperCAmelCase_ , )
lowerCamelCase__: Any =spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.Load(str(UpperCAmelCase_))
lowerCamelCase__: List[str] =vocab_file
# Original fairseq vocab and spm vocab must be "aligned":
# Vocab | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9
# -------- | ------- | ------- | ------ | ------- | --- | --- | --- | ----- | ----- | ----
# fairseq | '<s>' | '<pad>' | '</s>' | '<unk>' | ',' | '.' | '▁' | 's' | '▁de' | '-'
# spm | '<unk>' | '<s>' | '</s>' | ',' | '.' | '▁' | 's' | '▁de' | '-' | '▁a'
# The first "real" token "," has position 4 in the original fairseq vocab and position 3 in the spm vocab
lowerCamelCase__: Union[str, Any] =1
# Mimic fairseq token-to-id alignment for the first 4 token
lowerCamelCase__: Union[str, Any] ={"<s>": 0, "<pad>": 1, "</s>": 2, "<unk>": 3}
lowerCamelCase__: Any =len(self.sp_model)
lowerCamelCase__: Tuple ={F"""<madeupword{i}>""": sp_size + i + self.fairseq_offset for i in range(self.num_madeup_words)}
self.fairseq_tokens_to_ids.update(UpperCAmelCase_)
lowerCamelCase__: int ={v: k for k, v in self.fairseq_tokens_to_ids.items()}
def __getstate__(self : Dict) ->List[str]:
'''simple docstring'''
lowerCamelCase__: List[Any] =self.__dict__.copy()
lowerCamelCase__: str =None
lowerCamelCase__: Any =self.sp_model.serialized_model_proto()
return state
def __setstate__(self : Any , UpperCAmelCase_ : Tuple) ->Optional[int]:
'''simple docstring'''
lowerCamelCase__: Optional[Any] =d
# for backward compatibility
if not hasattr(self , "sp_model_kwargs"):
lowerCamelCase__: int ={}
lowerCamelCase__: str =spm.SentencePieceProcessor(**self.sp_model_kwargs)
self.sp_model.LoadFromSerializedProto(self.sp_model_proto)
def SCREAMING_SNAKE_CASE_ (self : List[str] , UpperCAmelCase_ : List[int] , UpperCAmelCase_ : Optional[List[int]] = None) ->List[int]:
'''simple docstring'''
if token_ids_a is None:
return [self.sep_token_id] + token_ids_a
lowerCamelCase__: Any =[self.sep_token_id]
return sep + token_ids_a + sep + sep + token_ids_a
def SCREAMING_SNAKE_CASE_ (self : int , UpperCAmelCase_ : List[int] , UpperCAmelCase_ : Optional[List[int]] = None , UpperCAmelCase_ : bool = False) ->List[int]:
'''simple docstring'''
if already_has_special_tokens:
return super().get_special_tokens_mask(
token_ids_a=UpperCAmelCase_ , token_ids_a=UpperCAmelCase_ , already_has_special_tokens=UpperCAmelCase_)
if token_ids_a is None:
return [1] + ([0] * len(UpperCAmelCase_))
return [1] + ([0] * len(UpperCAmelCase_)) + [1, 1] + ([0] * len(UpperCAmelCase_))
def SCREAMING_SNAKE_CASE_ (self : Dict , UpperCAmelCase_ : List[int] , UpperCAmelCase_ : Optional[List[int]] = None) ->List[int]:
'''simple docstring'''
lowerCamelCase__: List[str] =[self.sep_token_id]
if token_ids_a is None:
return len(sep + token_ids_a) * [0]
return len(sep + token_ids_a + sep + sep + token_ids_a) * [0]
@property
def SCREAMING_SNAKE_CASE_ (self : List[Any]) ->Optional[int]:
'''simple docstring'''
return len(self.sp_model) + self.fairseq_offset + self.num_madeup_words
def SCREAMING_SNAKE_CASE_ (self : int) ->str:
'''simple docstring'''
lowerCamelCase__: str ={self.convert_ids_to_tokens(UpperCAmelCase_): i for i in range(self.vocab_size)}
vocab.update(self.added_tokens_encoder)
return vocab
def SCREAMING_SNAKE_CASE_ (self : Any , UpperCAmelCase_ : str) ->List[str]:
'''simple docstring'''
return self.sp_model.encode(UpperCAmelCase_ , out_type=UpperCAmelCase_)
def SCREAMING_SNAKE_CASE_ (self : List[str] , UpperCAmelCase_ : Union[str, Any]) ->Dict:
'''simple docstring'''
if token in self.fairseq_tokens_to_ids:
return self.fairseq_tokens_to_ids[token]
lowerCamelCase__: Optional[int] =self.sp_model.PieceToId(UpperCAmelCase_)
# Need to return unknown token if the SP model returned 0
return spm_id + self.fairseq_offset if spm_id else self.unk_token_id
def SCREAMING_SNAKE_CASE_ (self : List[str] , UpperCAmelCase_ : List[str]) ->Optional[Any]:
'''simple docstring'''
if index in self.fairseq_ids_to_tokens:
return self.fairseq_ids_to_tokens[index]
return self.sp_model.IdToPiece(index - self.fairseq_offset)
def SCREAMING_SNAKE_CASE_ (self : Any , UpperCAmelCase_ : Dict) ->Dict:
'''simple docstring'''
lowerCamelCase__: str ="".join(UpperCAmelCase_).replace(UpperCAmelCase_ , " ").strip()
return out_string
def SCREAMING_SNAKE_CASE_ (self : Dict , UpperCAmelCase_ : str , UpperCAmelCase_ : Optional[str] = None) ->Tuple[str]:
'''simple docstring'''
if not os.path.isdir(UpperCAmelCase_):
logger.error(F"""Vocabulary path ({save_directory}) should be a directory""")
return
lowerCamelCase__: List[str] =os.path.join(
UpperCAmelCase_ , (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"])
if os.path.abspath(self.vocab_file) != os.path.abspath(UpperCAmelCase_) and os.path.isfile(self.vocab_file):
copyfile(self.vocab_file , UpperCAmelCase_)
elif not os.path.isfile(self.vocab_file):
with open(UpperCAmelCase_ , "wb") as fi:
lowerCamelCase__: Optional[int] =self.sp_model.serialized_model_proto()
fi.write(UpperCAmelCase_)
return (out_vocab_file,)
| 10
|
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__lowerCamelCase = {
"""configuration_biogpt""": ["""BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP""", """BioGptConfig"""],
"""tokenization_biogpt""": ["""BioGptTokenizer"""],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST""",
"""BioGptForCausalLM""",
"""BioGptForTokenClassification""",
"""BioGptForSequenceClassification""",
"""BioGptModel""",
"""BioGptPreTrainedModel""",
]
if TYPE_CHECKING:
from .configuration_biogpt import BIOGPT_PRETRAINED_CONFIG_ARCHIVE_MAP, BioGptConfig
from .tokenization_biogpt import BioGptTokenizer
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_biogpt import (
BIOGPT_PRETRAINED_MODEL_ARCHIVE_LIST,
BioGptForCausalLM,
BioGptForSequenceClassification,
BioGptForTokenClassification,
BioGptModel,
BioGptPreTrainedModel,
)
else:
import sys
__lowerCamelCase = _LazyModule(__name__, globals()["""__file__"""], _import_structure, module_spec=__spec__)
| 59
| 0
|
from __future__ import annotations
def _UpperCAmelCase (UpperCamelCase__ : list , UpperCamelCase__ : int , UpperCamelCase__ : int , UpperCamelCase__ : int ):
_A : str = []
_A , _A : Dict = input_list[low:mid], input_list[mid : high + 1]
while left and right:
result.append((left if left[0] <= right[0] else right).pop(0 ) )
_A : int = result + left + right
return input_list
def _UpperCAmelCase (UpperCamelCase__ : list ):
if len(UpperCamelCase__ ) <= 1:
return input_list
_A : Optional[Any] = list(UpperCamelCase__ )
# iteration for two-way merging
_A : Dict = 2
while p <= len(UpperCamelCase__ ):
# getting low, high and middle value for merge-sort of single list
for i in range(0 , len(UpperCamelCase__ ) , UpperCamelCase__ ):
_A : List[str] = i
_A : Dict = i + p - 1
_A : Optional[Any] = (low + high + 1) // 2
_A : Any = merge(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
# final merge of last two parts
if p * 2 >= len(UpperCamelCase__ ):
_A : int = i
_A : Any = merge(UpperCamelCase__ , 0 , UpperCamelCase__ , len(UpperCamelCase__ ) - 1 )
break
p *= 2
return input_list
if __name__ == "__main__":
lowerCAmelCase__ = input('Enter numbers separated by a comma:\n').strip()
if user_input == "":
lowerCAmelCase__ = []
else:
lowerCAmelCase__ = [int(item.strip()) for item in user_input.split(',')]
print(iter_merge_sort(unsorted))
| 11
|
import collections
import inspect
import unittest
from typing import Dict, List, Tuple
from transformers import MaskFormerSwinConfig
from transformers.testing_utils import require_torch, require_torch_multi_gpu, torch_device
from transformers.utils import is_torch_available
from ...test_backbone_common import BackboneTesterMixin
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import MaskFormerSwinBackbone
from transformers.models.maskformer import MaskFormerSwinModel
class UpperCAmelCase :
def __init__(self : Dict , snake_case__ : Dict , snake_case__ : Any=13 , snake_case__ : Any=32 , snake_case__ : Optional[Any]=2 , snake_case__ : Union[str, Any]=3 , snake_case__ : List[Any]=16 , snake_case__ : int=[1, 2, 1] , snake_case__ : Dict=[2, 2, 4] , snake_case__ : Dict=2 , snake_case__ : Tuple=2.0 , snake_case__ : Optional[int]=True , snake_case__ : Union[str, Any]=0.0 , snake_case__ : Any=0.0 , snake_case__ : Union[str, Any]=0.1 , snake_case__ : int="gelu" , snake_case__ : Optional[int]=False , snake_case__ : List[Any]=True , snake_case__ : List[str]=0.02 , snake_case__ : int=1e-5 , snake_case__ : List[str]=True , snake_case__ : Union[str, Any]=None , snake_case__ : List[Any]=True , snake_case__ : Optional[Any]=10 , snake_case__ : Optional[Any]=8 , snake_case__ : Any=["stage1", "stage2", "stage3"] , snake_case__ : Tuple=[1, 2, 3] , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Any = parent
snake_case : Optional[int] = batch_size
snake_case : Union[str, Any] = image_size
snake_case : Dict = patch_size
snake_case : Optional[Any] = num_channels
snake_case : Union[str, Any] = embed_dim
snake_case : int = depths
snake_case : List[str] = num_heads
snake_case : Union[str, Any] = window_size
snake_case : Union[str, Any] = mlp_ratio
snake_case : List[Any] = qkv_bias
snake_case : List[Any] = hidden_dropout_prob
snake_case : Union[str, Any] = attention_probs_dropout_prob
snake_case : Union[str, Any] = drop_path_rate
snake_case : int = hidden_act
snake_case : Optional[int] = use_absolute_embeddings
snake_case : int = patch_norm
snake_case : Union[str, Any] = layer_norm_eps
snake_case : Any = initializer_range
snake_case : Optional[Any] = is_training
snake_case : Tuple = scope
snake_case : Optional[int] = use_labels
snake_case : Optional[Any] = type_sequence_label_size
snake_case : Union[str, Any] = encoder_stride
snake_case : Any = out_features
snake_case : Tuple = out_indices
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case : Optional[Any] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case : int = None
if self.use_labels:
snake_case : Tuple = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case : Dict = self.get_config()
return config, pixel_values, labels
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> int:
'''simple docstring'''
return MaskFormerSwinConfig(
image_size=self.image_size , patch_size=self.patch_size , num_channels=self.num_channels , embed_dim=self.embed_dim , depths=self.depths , num_heads=self.num_heads , window_size=self.window_size , mlp_ratio=self.mlp_ratio , qkv_bias=self.qkv_bias , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , drop_path_rate=self.drop_path_rate , hidden_act=self.hidden_act , use_absolute_embeddings=self.use_absolute_embeddings , path_norm=self.patch_norm , layer_norm_eps=self.layer_norm_eps , initializer_range=self.initializer_range , encoder_stride=self.encoder_stride , out_features=self.out_features , out_indices=self.out_indices , )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : List[Any] , snake_case__ : List[str] , snake_case__ : Tuple ) -> Optional[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModel(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
snake_case : Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case : int = int(config.embed_dim * 2 ** (len(config.depths ) - 1) )
self.parent.assertEqual(result.last_hidden_state.shape , (self.batch_size, expected_seq_len, expected_dim) )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Union[str, Any] , snake_case__ : List[Any] , snake_case__ : Union[str, Any] ) -> str:
'''simple docstring'''
snake_case : Optional[int] = MaskFormerSwinBackbone(config=snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : List[Any] = model(snake_case__ )
# verify feature maps
self.parent.assertEqual(len(result.feature_maps ) , len(config.out_features ) )
self.parent.assertListEqual(list(result.feature_maps[0].shape ) , [13, 16, 16, 16] )
# verify channels
self.parent.assertEqual(len(model.channels ) , len(config.out_features ) )
self.parent.assertListEqual(model.channels , [16, 32, 64] )
# verify ValueError
with self.parent.assertRaises(snake_case__ ):
snake_case : Tuple = ["stem"]
snake_case : List[Any] = MaskFormerSwinBackbone(config=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
snake_case : Union[str, Any] = self.prepare_config_and_inputs()
snake_case , snake_case , snake_case : List[Any] = config_and_inputs
snake_case : int = {"pixel_values": pixel_values}
return config, inputs_dict
@require_torch
class UpperCAmelCase ( A_ ,A_ ,unittest.TestCase ):
A__ : List[str] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
A__ : str = {"feature-extraction": MaskFormerSwinModel} if is_torch_available() else {}
A__ : Optional[Any] = False
A__ : List[Any] = False
A__ : List[str] = False
A__ : List[str] = False
A__ : Union[str, Any] = False
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[str]:
'''simple docstring'''
snake_case : str = MaskFormerSwinModelTester(self )
snake_case : Optional[int] = ConfigTester(self , config_class=snake_case__ , embed_dim=37 )
@require_torch_multi_gpu
@unittest.skip(
reason=(
"`MaskFormerSwinModel` outputs `hidden_states_spatial_dimensions` which doesn't work well with"
" `nn.DataParallel`"
) )
def _SCREAMING_SNAKE_CASE (self : str ) -> Optional[Any]:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : str ) -> List[str]:
'''simple docstring'''
self.create_and_test_config_common_properties()
self.config_tester.create_and_test_config_to_json_string()
self.config_tester.create_and_test_config_to_json_file()
self.config_tester.create_and_test_config_from_and_save_pretrained()
self.config_tester.create_and_test_config_with_num_labels()
self.config_tester.check_config_can_be_init_without_params()
self.config_tester.check_config_arguments_init()
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> List[Any]:
'''simple docstring'''
return
def _SCREAMING_SNAKE_CASE (self : Dict ) -> str:
'''simple docstring'''
snake_case : str = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Dict:
'''simple docstring'''
snake_case : Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_backbone(*snake_case__ )
@unittest.skip("Swin does not use inputs_embeds" )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
pass
@unittest.skip("Swin does not support feedforward chunking" )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Dict:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
snake_case , snake_case : Any = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : int = model_class(snake_case__ )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case : List[Any] = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case__ , nn.Linear ) )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Dict:
'''simple docstring'''
snake_case , snake_case : str = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case : str = model_class(snake_case__ )
snake_case : Optional[int] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case : Optional[Any] = [*signature.parameters.keys()]
snake_case : Tuple = ["pixel_values"]
self.assertListEqual(arg_names[:1] , snake_case__ )
@unittest.skip(reason="MaskFormerSwin is only used as backbone and doesn't support output_attentions" )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="MaskFormerSwin is only used as an internal backbone" )
def _SCREAMING_SNAKE_CASE (self : Tuple ) -> Any:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : List[Any] , snake_case__ : str , snake_case__ : List[Any] , snake_case__ : Tuple ) -> Optional[int]:
'''simple docstring'''
snake_case : Tuple = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
with torch.no_grad():
snake_case : Any = model(**self._prepare_for_class(snake_case__ , snake_case__ ) )
snake_case : int = outputs.hidden_states
snake_case : Union[str, Any] = getattr(
self.model_tester , "expected_num_hidden_layers" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case__ ) , snake_case__ )
# Swin has a different seq_length
snake_case : Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : Tuple = (image_size[1] // patch_size[1]) * (image_size[0] // patch_size[0])
self.assertListEqual(
list(hidden_states[0].shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def _SCREAMING_SNAKE_CASE (self : Dict ) -> Union[str, Any]:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : int = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
for model_class in self.all_model_classes:
snake_case : int = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Dict = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int ) -> Any:
'''simple docstring'''
snake_case , snake_case : Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Any = 3
snake_case : List[str] = (
self.model_tester.image_size
if isinstance(self.model_tester.image_size , collections.abc.Iterable )
else (self.model_tester.image_size, self.model_tester.image_size)
)
snake_case : Tuple = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case : str = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case : str = image_size[1] + patch_size[1] - (image_size[1] % patch_size[1])
for model_class in self.all_model_classes:
snake_case : str = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
# check that output_hidden_states also work using config
del inputs_dict["output_hidden_states"]
snake_case : Optional[Any] = True
self.check_hidden_states_output(snake_case__ , snake_case__ , snake_case__ , (padded_height, padded_width) )
@unittest.skip(reason="MaskFormerSwin doesn't have pretrained checkpoints" )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> str:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
pass
@unittest.skip(reason="This will be fixed once MaskFormerSwin is replaced by native Swin" )
def _SCREAMING_SNAKE_CASE (self : int ) -> str:
'''simple docstring'''
pass
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
snake_case , snake_case : List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case__ : Union[str, Any] ):
snake_case : Any = 0
return t
def check_equivalence(snake_case__ : Union[str, Any] , snake_case__ : int , snake_case__ : List[str] , snake_case__ : Optional[int]={} ):
with torch.no_grad():
snake_case : Optional[Any] = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ )
snake_case : Tuple = model(**snake_case__ , return_dict=snake_case__ , **snake_case__ ).to_tuple()
def recursive_check(snake_case__ : List[str] , snake_case__ : Optional[Any] ):
if isinstance(snake_case__ , (List, Tuple) ):
for tuple_iterable_value, dict_iterable_value in zip(snake_case__ , snake_case__ ):
recursive_check(snake_case__ , snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
for tuple_iterable_value, dict_iterable_value in zip(
tuple_object.values() , dict_object.values() ):
recursive_check(snake_case__ , snake_case__ )
elif tuple_object is None:
return
else:
self.assertTrue(
torch.allclose(
set_nan_tensor_to_zero(snake_case__ ) , set_nan_tensor_to_zero(snake_case__ ) , atol=1e-5 ) , msg=(
"Tuple and dict output are not equal. Difference:"
f""" {torch.max(torch.abs(tuple_object - dict_object ) )}. Tuple has `nan`:"""
f""" {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}. Dict has"""
f""" `nan`: {torch.isnan(snake_case__ ).any()} and `inf`: {torch.isinf(snake_case__ )}."""
) , )
recursive_check(snake_case__ , snake_case__ )
for model_class in self.all_model_classes:
snake_case : Optional[int] = model_class(snake_case__ )
model.to(snake_case__ )
model.eval()
snake_case : Union[str, Any] = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Tuple = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : Optional[Any] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ )
snake_case : Dict = self._prepare_for_class(snake_case__ , snake_case__ )
snake_case : List[Any] = self._prepare_for_class(snake_case__ , snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
snake_case : Any = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
snake_case : List[str] = self._prepare_for_class(snake_case__ , snake_case__ , return_labels=snake_case__ )
check_equivalence(snake_case__ , snake_case__ , snake_case__ , {"output_hidden_states": True} )
@require_torch
class UpperCAmelCase ( unittest.TestCase ,A_ ):
A__ : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
A__ : int = MaskFormerSwinConfig
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> Any:
'''simple docstring'''
snake_case : Union[str, Any] = MaskFormerSwinModelTester(self )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case , snake_case : Dict = self.model_tester.prepare_config_and_inputs_for_common()
snake_case : Optional[int] = inputs_dict["pixel_values"].shape[0]
for backbone_class in self.all_model_classes:
snake_case : Optional[int] = backbone_class(snake_case__ )
backbone.to(snake_case__ )
backbone.eval()
snake_case : Union[str, Any] = backbone(**snake_case__ )
# Test default outputs and verify feature maps
self.assertIsInstance(outputs.feature_maps , snake_case__ )
self.assertTrue(len(outputs.feature_maps ) == len(backbone.channels ) )
for feature_map, n_channels in zip(outputs.feature_maps , backbone.channels ):
self.assertTrue(feature_map.shape[:2] , (batch_size, n_channels) )
self.assertIsNone(outputs.hidden_states )
self.assertIsNone(outputs.attentions )
# Test output_hidden_states=True
snake_case : Optional[int] = backbone(**snake_case__ , output_hidden_states=snake_case__ )
self.assertIsNotNone(outputs.hidden_states )
self.assertTrue(len(outputs.hidden_states ) , len(backbone.stage_names ) )
# We skip the stem layer
for hidden_states, n_channels in zip(outputs.hidden_states[1:] , backbone.channels ):
for hidden_state in hidden_states:
# Hidden states are in the format (batch_size, (height * width), n_channels)
snake_case , snake_case , snake_case : Dict = hidden_state.shape
self.assertTrue((h_batch_size, h_n_channels) , (batch_size, n_channels) )
# Test output_attentions=True
if self.has_attentions:
snake_case : Optional[Any] = backbone(**snake_case__ , output_attentions=snake_case__ )
self.assertIsNotNone(outputs.attentions )
| 59
| 0
|
# 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.
from ..models.auto import AutoModelForSeqaSeqLM, AutoTokenizer
from .base import PipelineTool
class lowerCamelCase__( __lowerCamelCase):
UpperCAmelCase__ : Any = 'philschmid/bart-large-cnn-samsum'
UpperCAmelCase__ : str = (
'This is a tool that summarizes an English text. It takes an input `text` containing the text to summarize, '
'and returns a summary of the text.'
)
UpperCAmelCase__ : List[str] = 'summarizer'
UpperCAmelCase__ : Tuple = AutoTokenizer
UpperCAmelCase__ : Tuple = AutoModelForSeqaSeqLM
UpperCAmelCase__ : Any = ['text']
UpperCAmelCase__ : List[str] = ['text']
def lowerCAmelCase__ ( self: Union[str, Any] , UpperCamelCase_: Optional[int] ):
return self.pre_processor(UpperCamelCase_ , return_tensors="""pt""" , truncation=UpperCamelCase_ )
def lowerCAmelCase__ ( self: int , UpperCamelCase_: Union[str, Any] ):
return self.model.generate(**UpperCamelCase_ )[0]
def lowerCAmelCase__ ( self: Tuple , UpperCamelCase_: List[Any] ):
return self.pre_processor.decode(UpperCamelCase_ , skip_special_tokens=UpperCamelCase_ , clean_up_tokenization_spaces=UpperCamelCase_ )
| 12
|
from typing import Dict
import numpy as np
import torch
from . import residue_constants as rc
from .tensor_utils import tensor_tree_map, tree_map
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : List[str] = []
snake_case : Optional[int] = []
snake_case : Any = []
for rt in rc.restypes:
snake_case : List[Any] = rc.restype_name_to_atomaa_names[rc.restype_atoa[rt]]
restype_atomaa_to_atomaa_list.append([(rc.atom_order[name] if name else 0) for name in atom_names] )
snake_case : str = {name: i for i, name in enumerate(__lowerCamelCase )}
restype_atomaa_to_atomaa_list.append(
[(atom_name_to_idxaa[name] if name in atom_name_to_idxaa else 0) for name in rc.atom_types] )
restype_atomaa_mask_list.append([(1.0 if name else 0.0) for name in atom_names] )
# Add dummy mapping for restype 'UNK'
restype_atomaa_to_atomaa_list.append([0] * 14 )
restype_atomaa_to_atomaa_list.append([0] * 37 )
restype_atomaa_mask_list.append([0.0] * 14 )
snake_case : Optional[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : List[Any] = torch.tensor(
__lowerCamelCase , dtype=torch.intaa , device=protein["aatype"].device , )
snake_case : int = torch.tensor(
__lowerCamelCase , dtype=torch.floataa , device=protein["aatype"].device , )
snake_case : int = protein["aatype"].to(torch.long )
# create the mapping for (residx, atom14) --> atom37, i.e. an array
# with shape (num_res, 14) containing the atom37 indices for this protein
snake_case : List[Any] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : str = restype_atomaa_mask[protein_aatype]
snake_case : str = residx_atomaa_mask
snake_case : Any = residx_atomaa_to_atomaa.long()
# create the gather indices for mapping back
snake_case : List[str] = restype_atomaa_to_atomaa[protein_aatype]
snake_case : List[Any] = residx_atomaa_to_atomaa.long()
# create the corresponding mask
snake_case : Union[str, Any] = torch.zeros([21, 37] , dtype=torch.floataa , device=protein["aatype"].device )
for restype, restype_letter in enumerate(rc.restypes ):
snake_case : Optional[int] = rc.restype_atoa[restype_letter]
snake_case : Any = rc.residue_atoms[restype_name]
for atom_name in atom_names:
snake_case : List[Any] = rc.atom_order[atom_name]
snake_case : Optional[Any] = 1
snake_case : List[Any] = restype_atomaa_mask[protein_aatype]
snake_case : int = residx_atomaa_mask
return protein
def UpperCamelCase ( __lowerCamelCase : Dict[str, torch.Tensor] ):
snake_case : Dict = tree_map(lambda __lowerCamelCase : torch.tensor(__lowerCamelCase , device=batch["aatype"].device ) , __lowerCamelCase , np.ndarray )
snake_case : List[str] = tensor_tree_map(lambda __lowerCamelCase : np.array(__lowerCamelCase ) , make_atomaa_masks(__lowerCamelCase ) )
return out
| 59
| 0
|
def A_ ( _UpperCAmelCase ):
SCREAMING_SNAKE_CASE_: List[Any] = generate_pascal_triangle(_UpperCAmelCase )
for row_idx in range(_UpperCAmelCase ):
# Print left spaces
for _ in range(num_rows - row_idx - 1 ):
print(end=" " )
# Print row values
for col_idx in range(row_idx + 1 ):
if col_idx != row_idx:
print(triangle[row_idx][col_idx] , end=" " )
else:
print(triangle[row_idx][col_idx] , end="" )
print()
def A_ ( _UpperCAmelCase ):
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError("The input value of 'num_rows' should be 'int'" )
if num_rows == 0:
return []
elif num_rows < 0:
raise ValueError(
"The input value of 'num_rows' should be greater than or equal to 0" )
SCREAMING_SNAKE_CASE_: list[list[int]] = []
for current_row_idx in range(_UpperCAmelCase ):
SCREAMING_SNAKE_CASE_: List[Any] = populate_current_row(_UpperCAmelCase , _UpperCAmelCase )
triangle.append(_UpperCAmelCase )
return triangle
def A_ ( _UpperCAmelCase , _UpperCAmelCase ):
SCREAMING_SNAKE_CASE_: Dict = [-1] * (current_row_idx + 1)
# first and last elements of current row are equal to 1
SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_: Tuple = 1, 1
for current_col_idx in range(1 , _UpperCAmelCase ):
calculate_current_element(
_UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase )
return current_row
def A_ ( _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , _UpperCAmelCase , ):
SCREAMING_SNAKE_CASE_: str = triangle[current_row_idx - 1][current_col_idx - 1]
SCREAMING_SNAKE_CASE_: Optional[int] = triangle[current_row_idx - 1][current_col_idx]
SCREAMING_SNAKE_CASE_: str = above_to_left_elt + above_to_right_elt
def A_ ( _UpperCAmelCase ):
if not isinstance(_UpperCAmelCase , _UpperCAmelCase ):
raise TypeError("The input value of 'num_rows' should be 'int'" )
if num_rows == 0:
return []
elif num_rows < 0:
raise ValueError(
"The input value of 'num_rows' should be greater than or equal to 0" )
SCREAMING_SNAKE_CASE_: list[list[int]] = [[1]]
for row_index in range(1 , _UpperCAmelCase ):
SCREAMING_SNAKE_CASE_: Any = [0] + result[-1] + [0]
SCREAMING_SNAKE_CASE_: Tuple = row_index + 1
# Calculate the number of distinct elements in a row
SCREAMING_SNAKE_CASE_: Any = sum(divmod(_UpperCAmelCase , 2 ) )
SCREAMING_SNAKE_CASE_: Optional[int] = [
temp_row[i - 1] + temp_row[i] for i in range(1 , distinct_elements + 1 )
]
SCREAMING_SNAKE_CASE_: Tuple = row_first_half[: (row_index + 1) // 2]
row_second_half.reverse()
SCREAMING_SNAKE_CASE_: List[str] = row_first_half + row_second_half
result.append(_UpperCAmelCase )
return result
def A_ ( ):
from collections.abc import Callable
from timeit import timeit
def benchmark_a_function(_UpperCAmelCase , _UpperCAmelCase ) -> None:
SCREAMING_SNAKE_CASE_: int = f"{func.__name__}({value})"
SCREAMING_SNAKE_CASE_: List[str] = timeit(f"__main__.{call}" , setup="import __main__" )
# print(f"{call:38} = {func(value)} -- {timing:.4f} seconds")
print(f"{call:38} -- {timing:.4f} seconds" )
for value in range(15 ): # (1, 7, 14):
for func in (generate_pascal_triangle, generate_pascal_triangle_optimized):
benchmark_a_function(_UpperCAmelCase , _UpperCAmelCase )
print()
if __name__ == "__main__":
import doctest
doctest.testmod()
benchmark()
| 13
|
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from .tokenization_lxmert import LxmertTokenizer
__lowerCamelCase = {"""vocab_file""": """vocab.txt""", """tokenizer_file""": """tokenizer.json"""}
__lowerCamelCase = {
"""vocab_file""": {
"""unc-nlp/lxmert-base-uncased""": """https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/vocab.txt""",
},
"""tokenizer_file""": {
"""unc-nlp/lxmert-base-uncased""": (
"""https://huggingface.co/unc-nlp/lxmert-base-uncased/resolve/main/tokenizer.json"""
),
},
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": 5_12,
}
__lowerCamelCase = {
"""unc-nlp/lxmert-base-uncased""": {"""do_lower_case""": True},
}
class UpperCAmelCase ( A_ ):
A__ : Any = VOCAB_FILES_NAMES
A__ : Optional[int] = PRETRAINED_VOCAB_FILES_MAP
A__ : Tuple = PRETRAINED_INIT_CONFIGURATION
A__ : Tuple = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
A__ : List[Any] = LxmertTokenizer
def __init__(self : Dict , snake_case__ : Tuple=None , snake_case__ : Optional[Any]=None , snake_case__ : Optional[Any]=True , snake_case__ : Tuple="[UNK]" , snake_case__ : Optional[Any]="[SEP]" , snake_case__ : Optional[Any]="[PAD]" , snake_case__ : List[Any]="[CLS]" , snake_case__ : Tuple="[MASK]" , snake_case__ : Dict=True , snake_case__ : Union[str, Any]=None , **snake_case__ : Dict , ) -> Optional[int]:
'''simple docstring'''
super().__init__(
snake_case__ , tokenizer_file=snake_case__ , do_lower_case=snake_case__ , unk_token=snake_case__ , sep_token=snake_case__ , pad_token=snake_case__ , cls_token=snake_case__ , mask_token=snake_case__ , tokenize_chinese_chars=snake_case__ , strip_accents=snake_case__ , **snake_case__ , )
snake_case : int = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , snake_case__ ) != do_lower_case
or normalizer_state.get("strip_accents" , snake_case__ ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , snake_case__ ) != tokenize_chinese_chars
):
snake_case : Union[str, Any] = getattr(snake_case__ , normalizer_state.pop("type" ) )
snake_case : str = do_lower_case
snake_case : List[Any] = strip_accents
snake_case : Optional[int] = tokenize_chinese_chars
snake_case : int = normalizer_class(**snake_case__ )
snake_case : Optional[Any] = do_lower_case
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] , snake_case__ : Optional[Any] , snake_case__ : Dict=None ) -> Any:
'''simple docstring'''
snake_case : Optional[int] = [self.cls_token_id] + token_ids_a + [self.sep_token_id]
if token_ids_a:
output += token_ids_a + [self.sep_token_id]
return output
def _SCREAMING_SNAKE_CASE (self : Optional[int] , snake_case__ : List[int] , snake_case__ : Optional[List[int]] = None ) -> List[int]:
'''simple docstring'''
snake_case : Optional[Any] = [self.sep_token_id]
snake_case : 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 ) * [0] + len(token_ids_a + sep ) * [1]
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str , snake_case__ : Optional[str] = None ) -> Tuple[str]:
'''simple docstring'''
snake_case : List[Any] = self._tokenizer.model.save(snake_case__ , name=snake_case__ )
return tuple(snake_case__ )
| 59
| 0
|
import argparse
import torch
from safetensors.torch import load_file
from diffusers import StableDiffusionPipeline
def SCREAMING_SNAKE_CASE ( lowercase_ , lowercase_ , lowercase_ , lowercase_ , lowercase_ ) -> List[Any]:
"""simple docstring"""
A__ = StableDiffusionPipeline.from_pretrained(lowercase_ , torch_dtype=torch.floataa )
# load LoRA weight from .safetensors
A__ = load_file(lowercase_ )
A__ = []
# directly update weight in diffusers model
for key in state_dict:
# it is suggested to print out the key, it usually will be something like below
# "lora_te_text_model_encoder_layers_0_self_attn_k_proj.lora_down.weight"
# as we have set the alpha beforehand, so just skip
if ".alpha" in key or key in visited:
continue
if "text" in key:
A__ = key.split('''.''' )[0].split(LORA_PREFIX_TEXT_ENCODER + '''_''' )[-1].split('''_''' )
A__ = pipeline.text_encoder
else:
A__ = key.split('''.''' )[0].split(LORA_PREFIX_UNET + '''_''' )[-1].split('''_''' )
A__ = pipeline.unet
# find the target layer
A__ = layer_infos.pop(0 )
while len(lowercase_ ) > -1:
try:
A__ = curr_layer.__getattr__(lowercase_ )
if len(lowercase_ ) > 0:
A__ = layer_infos.pop(0 )
elif len(lowercase_ ) == 0:
break
except Exception:
if len(lowercase_ ) > 0:
temp_name += "_" + layer_infos.pop(0 )
else:
A__ = layer_infos.pop(0 )
A__ = []
if "lora_down" in key:
pair_keys.append(key.replace('''lora_down''' , '''lora_up''' ) )
pair_keys.append(lowercase_ )
else:
pair_keys.append(lowercase_ )
pair_keys.append(key.replace('''lora_up''' , '''lora_down''' ) )
# update weight
if len(state_dict[pair_keys[0]].shape ) == 4:
A__ = state_dict[pair_keys[0]].squeeze(3 ).squeeze(2 ).to(torch.floataa )
A__ = state_dict[pair_keys[1]].squeeze(3 ).squeeze(2 ).to(torch.floataa )
curr_layer.weight.data += alpha * torch.mm(lowercase_ , lowercase_ ).unsqueeze(2 ).unsqueeze(3 )
else:
A__ = state_dict[pair_keys[0]].to(torch.floataa )
A__ = state_dict[pair_keys[1]].to(torch.floataa )
curr_layer.weight.data += alpha * torch.mm(lowercase_ , lowercase_ )
# update visited list
for item in pair_keys:
visited.append(lowercase_ )
return pipeline
if __name__ == "__main__":
_lowerCamelCase : Tuple = argparse.ArgumentParser()
parser.add_argument(
"""--base_model_path""", default=None, type=str, required=True, help="""Path to the base model in diffusers format."""
)
parser.add_argument(
"""--checkpoint_path""", default=None, type=str, required=True, help="""Path to the checkpoint to convert."""
)
parser.add_argument("""--dump_path""", default=None, type=str, required=True, help="""Path to the output model.""")
parser.add_argument(
"""--lora_prefix_unet""", default="""lora_unet""", type=str, help="""The prefix of UNet weight in safetensors"""
)
parser.add_argument(
"""--lora_prefix_text_encoder""",
default="""lora_te""",
type=str,
help="""The prefix of text encoder weight in safetensors""",
)
parser.add_argument("""--alpha""", default=0.75, type=float, help="""The merging ratio in W = W0 + alpha * deltaW""")
parser.add_argument(
"""--to_safetensors""", action="""store_true""", help="""Whether to store pipeline in safetensors format or not."""
)
parser.add_argument("""--device""", type=str, help="""Device to use (e.g. cpu, cuda:0, cuda:1, etc.)""")
_lowerCamelCase : Tuple = parser.parse_args()
_lowerCamelCase : List[Any] = args.base_model_path
_lowerCamelCase : Optional[int] = args.checkpoint_path
_lowerCamelCase : Dict = args.dump_path
_lowerCamelCase : Optional[Any] = args.lora_prefix_unet
_lowerCamelCase : Optional[int] = args.lora_prefix_text_encoder
_lowerCamelCase : List[Any] = args.alpha
_lowerCamelCase : int = convert(base_model_path, checkpoint_path, lora_prefix_unet, lora_prefix_text_encoder, alpha)
_lowerCamelCase : Tuple = pipe.to(args.device)
pipe.save_pretrained(args.dump_path, safe_serialization=args.to_safetensors)
| 14
|
import torch
from diffusers import DDIMParallelScheduler
from .test_schedulers import SchedulerCommonTest
class UpperCAmelCase ( A_ ):
A__ : Dict = (DDIMParallelScheduler,)
A__ : Tuple = (("eta", 0.0), ("num_inference_steps", 50))
def _SCREAMING_SNAKE_CASE (self : Tuple , **snake_case__ : Optional[int] ) -> Optional[Any]:
'''simple docstring'''
snake_case : Any = {
"num_train_timesteps": 10_00,
"beta_start": 0.0001,
"beta_end": 0.02,
"beta_schedule": "linear",
"clip_sample": True,
}
config.update(**snake_case__ )
return config
def _SCREAMING_SNAKE_CASE (self : Dict , **snake_case__ : Optional[int] ) -> Any:
'''simple docstring'''
snake_case : List[Any] = self.scheduler_classes[0]
snake_case : Any = self.get_scheduler_config(**snake_case__ )
snake_case : Any = scheduler_class(**snake_case__ )
snake_case , snake_case : Union[str, Any] = 10, 0.0
snake_case : List[Any] = self.dummy_model()
snake_case : Any = self.dummy_sample_deter
scheduler.set_timesteps(snake_case__ )
for t in scheduler.timesteps:
snake_case : Optional[int] = model(snake_case__ , snake_case__ )
snake_case : List[str] = scheduler.step(snake_case__ , snake_case__ , snake_case__ , snake_case__ ).prev_sample
return sample
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> str:
'''simple docstring'''
for timesteps in [1_00, 5_00, 10_00]:
self.check_over_configs(num_train_timesteps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> int:
'''simple docstring'''
for steps_offset in [0, 1]:
self.check_over_configs(steps_offset=snake_case__ )
snake_case : Optional[int] = self.scheduler_classes[0]
snake_case : Optional[int] = self.get_scheduler_config(steps_offset=1 )
snake_case : Union[str, Any] = scheduler_class(**snake_case__ )
scheduler.set_timesteps(5 )
assert torch.equal(scheduler.timesteps , torch.LongTensor([8_01, 6_01, 4_01, 2_01, 1] ) )
def _SCREAMING_SNAKE_CASE (self : int ) -> Tuple:
'''simple docstring'''
for beta_start, beta_end in zip([0.0001, 0.001, 0.01, 0.1] , [0.002, 0.02, 0.2, 2] ):
self.check_over_configs(beta_start=snake_case__ , beta_end=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Union[str, Any] ) -> Union[str, Any]:
'''simple docstring'''
for schedule in ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : str ) -> Dict:
'''simple docstring'''
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> List[str]:
'''simple docstring'''
for clip_sample in [True, False]:
self.check_over_configs(clip_sample=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] ) -> List[Any]:
'''simple docstring'''
for timestep_spacing in ["trailing", "leading"]:
self.check_over_configs(timestep_spacing=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> List[Any]:
'''simple docstring'''
for rescale_betas_zero_snr in [True, False]:
self.check_over_configs(rescale_betas_zero_snr=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
self.check_over_configs(thresholding=snake_case__ )
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(
thresholding=snake_case__ , prediction_type=snake_case__ , sample_max_value=snake_case__ , )
def _SCREAMING_SNAKE_CASE (self : Any ) -> Any:
'''simple docstring'''
for t in [1, 10, 49]:
self.check_over_forward(time_step=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Any:
'''simple docstring'''
for t, num_inference_steps in zip([1, 10, 50] , [10, 50, 5_00] ):
self.check_over_forward(time_step=snake_case__ , num_inference_steps=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] ) -> Optional[Any]:
'''simple docstring'''
for t, eta in zip([1, 10, 49] , [0.0, 0.5, 1.0] ):
self.check_over_forward(time_step=snake_case__ , eta=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[int]:
'''simple docstring'''
snake_case : Dict = self.scheduler_classes[0]
snake_case : Tuple = self.get_scheduler_config()
snake_case : Dict = scheduler_class(**snake_case__ )
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_20 , 4_00 ) - 0.14771 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_80 , 9_60 ) - 0.32460 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(0 , 0 ) - 0.0 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(4_87 , 4_86 ) - 0.00979 ) ) < 1e-5
assert torch.sum(torch.abs(scheduler._get_variance(9_99 , 9_98 ) - 0.02 ) ) < 1e-5
def _SCREAMING_SNAKE_CASE (self : Optional[int] ) -> Dict:
'''simple docstring'''
snake_case : Union[str, Any] = self.scheduler_classes[0]
snake_case : List[Any] = self.get_scheduler_config()
snake_case : int = scheduler_class(**snake_case__ )
snake_case , snake_case : Any = 10, 0.0
scheduler.set_timesteps(snake_case__ )
snake_case : Optional[Any] = self.dummy_model()
snake_case : str = self.dummy_sample_deter
snake_case : Dict = self.dummy_sample_deter + 0.1
snake_case : Dict = self.dummy_sample_deter - 0.1
snake_case : Optional[Any] = samplea.shape[0]
snake_case : str = torch.stack([samplea, samplea, samplea] , dim=0 )
snake_case : Tuple = torch.arange(snake_case__ )[0:3, None].repeat(1 , snake_case__ )
snake_case : Tuple = model(samples.flatten(0 , 1 ) , timesteps.flatten(0 , 1 ) )
snake_case : List[str] = scheduler.batch_step_no_noise(snake_case__ , timesteps.flatten(0 , 1 ) , samples.flatten(0 , 1 ) , snake_case__ )
snake_case : Dict = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 1147.7904 ) < 1e-2
assert abs(result_mean.item() - 0.4982 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Optional[Any] ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[Any] = self.full_loop()
snake_case : Optional[Any] = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 172.0067 ) < 1e-2
assert abs(result_mean.item() - 0.223967 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : str ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(prediction_type="v_prediction" )
snake_case : int = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[int] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 52.5302 ) < 1e-2
assert abs(result_mean.item() - 0.0684 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : Any ) -> Optional[Any]:
'''simple docstring'''
snake_case : Dict = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : str = torch.sum(torch.abs(snake_case__ ) )
snake_case : Optional[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.8295 ) < 1e-2
assert abs(result_mean.item() - 0.1951 ) < 1e-3
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[Any]:
'''simple docstring'''
snake_case : int = self.full_loop(set_alpha_to_one=snake_case__ , beta_start=0.01 )
snake_case : Tuple = torch.sum(torch.abs(snake_case__ ) )
snake_case : List[Any] = torch.mean(torch.abs(snake_case__ ) )
assert abs(result_sum.item() - 149.0784 ) < 1e-2
assert abs(result_mean.item() - 0.1941 ) < 1e-3
| 59
| 0
|
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
SCREAMING_SNAKE_CASE :Optional[int] = '▁'
SCREAMING_SNAKE_CASE :List[Any] = get_tests_dir('fixtures/test_sentencepiece.model')
@require_sentencepiece
@require_tokenizers
class UpperCAmelCase ( __SCREAMING_SNAKE_CASE , unittest.TestCase ):
'''simple docstring'''
snake_case_ = BigBirdTokenizer
snake_case_ = BigBirdTokenizerFast
snake_case_ = True
snake_case_ = True
def UpperCamelCase_ ( self : Dict ):
super().setUp()
__A = self.tokenizer_class(A ,keep_accents=A )
tokenizer.save_pretrained(self.tmpdirname )
def UpperCamelCase_ ( self : Optional[Any] ):
__A = "<s>"
__A = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(A ) ,A )
self.assertEqual(self.get_tokenizer()._convert_id_to_token(A ) ,A )
def UpperCamelCase_ ( self : int ):
__A = list(self.get_tokenizer().get_vocab().keys() )
self.assertEqual(vocab_keys[0] ,"<unk>" )
self.assertEqual(vocab_keys[1] ,"<s>" )
self.assertEqual(vocab_keys[-1] ,"[MASK]" )
self.assertEqual(len(A ) ,10_04 )
def UpperCamelCase_ ( self : Dict ):
self.assertEqual(self.get_tokenizer().vocab_size ,10_00 )
def UpperCamelCase_ ( self : str ):
if not self.test_rust_tokenizer:
return
__A = self.get_tokenizer()
__A = self.get_rust_tokenizer()
__A = "I was born in 92000, and this is falsé."
__A = tokenizer.tokenize(A )
__A = rust_tokenizer.tokenize(A )
self.assertListEqual(A ,A )
__A = tokenizer.encode(A ,add_special_tokens=A )
__A = rust_tokenizer.encode(A ,add_special_tokens=A )
self.assertListEqual(A ,A )
__A = self.get_rust_tokenizer()
__A = tokenizer.encode(A )
__A = rust_tokenizer.encode(A )
self.assertListEqual(A ,A )
def UpperCamelCase_ ( self : Any ):
__A = BigBirdTokenizer(A ,keep_accents=A )
__A = tokenizer.tokenize("This is a test" )
self.assertListEqual(A ,["▁This", "▁is", "▁a", "▁t", "est"] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(A ) ,[2_85, 46, 10, 1_70, 3_82] ,)
__A = 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",
".",
] ,)
__A = tokenizer.convert_tokens_to_ids(A )
self.assertListEqual(
A ,[8, 21, 84, 55, 24, 19, 7, 0, 6_02, 3_47, 3_47, 3_47, 3, 12, 66, 46, 72, 80, 6, 0, 4] ,)
__A = 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 UpperCamelCase_ ( self : str ):
return BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
@slow
def UpperCamelCase_ ( self : List[str] ):
__A = "Hello World!"
__A = [65, 1_85_36, 22_60, 1_01, 66]
self.assertListEqual(A ,self.big_tokenizer.encode(A ) )
@slow
def UpperCamelCase_ ( self : Dict ):
__A = (
"This is a very long text with a lot of weird characters, such as: . , ~ ? ( ) \" [ ] ! : - . Also we will"
" add words that should not exsist and be tokenized to <unk>, such as saoneuhaoesuth"
)
# fmt: off
__A = [65, 8_71, 4_19, 3_58, 9_46, 9_91, 25_21, 4_52, 3_58, 13_57, 3_87, 77_51, 35_36, 1_12, 9_85, 4_56, 1_26, 8_65, 9_38, 54_00, 57_34, 4_58, 13_68, 4_67, 7_86, 24_62, 52_46, 11_59, 6_33, 8_65, 45_19, 4_57, 5_82, 8_52, 25_57, 4_27, 9_16, 5_08, 4_05, 3_43_24, 4_97, 3_91, 4_08, 1_13_42, 12_44, 3_85, 1_00, 9_38, 9_85, 4_56, 5_74, 3_62, 1_25_97, 32_00, 31_29, 11_72, 66] # noqa: E231
# fmt: on
self.assertListEqual(A ,self.big_tokenizer.encode(A ) )
@require_torch
@slow
def UpperCamelCase_ ( self : Any ):
import torch
from transformers import BigBirdConfig, BigBirdModel
# Build sequence
__A = list(self.big_tokenizer.get_vocab().keys() )[:10]
__A = " ".join(A )
__A = self.big_tokenizer.encode_plus(A ,return_tensors="pt" ,return_token_type_ids=A )
__A = self.big_tokenizer.batch_encode_plus(
[sequence + " " + sequence] ,return_tensors="pt" ,return_token_type_ids=A )
__A = BigBirdConfig(attention_type="original_full" )
__A = 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 UpperCamelCase_ ( self : List[str] ):
__A = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base" )
__A = tokenizer.decode(tokenizer("Paris is the [MASK]." ).input_ids )
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]" )
@slow
def UpperCamelCase_ ( self : str ):
# fmt: off
__A = {"input_ids": [[65, 3_92_86, 4_58, 3_63_35, 20_01, 4_56, 1_30_73, 1_32_66, 4_55, 1_13, 77_46, 17_41, 1_11_57, 3_91, 1_30_73, 1_32_66, 4_55, 1_13, 39_67, 3_54_12, 1_13, 49_36, 1_09, 38_70, 23_77, 1_13, 3_00_84, 4_57_20, 4_58, 1_34, 1_74_96, 1_12, 5_03, 1_16_72, 1_13, 1_18, 1_12, 56_65, 1_33_47, 3_86_87, 1_12, 14_96, 3_13_89, 1_12, 32_68, 4_72_64, 1_34, 9_62, 1_12, 1_63_77, 80_35, 2_31_30, 4_30, 1_21_69, 1_55_18, 2_85_92, 4_58, 1_46, 4_16_97, 1_09, 3_91, 1_21_69, 1_55_18, 1_66_89, 4_58, 1_46, 4_13_58, 1_09, 4_52, 7_26, 40_34, 1_11, 7_63, 3_54_12, 50_82, 3_88, 19_03, 1_11, 90_51, 3_91, 28_70, 4_89_18, 19_00, 11_23, 5_50, 9_98, 1_12, 95_86, 1_59_85, 4_55, 3_91, 4_10, 2_29_55, 3_76_36, 1_14, 66], [65, 4_48, 1_74_96, 4_19, 36_63, 3_85, 7_63, 1_13, 2_75_33, 28_70, 32_83, 1_30_43, 16_39, 2_47_13, 5_23, 6_56, 2_40_13, 1_85_50, 25_21, 5_17, 2_70_14, 2_12_44, 4_20, 12_12, 14_65, 3_91, 9_27, 48_33, 3_88, 5_78, 1_17_86, 1_14, 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, 4_84, 21_69, 76_87, 2_19_32, 1_81_46, 7_26, 3_63, 1_70_32, 33_91, 1_14, 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" ,)
| 15
|
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )]
snake_case : int = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1 or len(__lowerCamelCase ) <= key:
return input_string
for position, character in enumerate(__lowerCamelCase ):
snake_case : Any = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append(__lowerCamelCase )
snake_case : List[str] = ["".join(__lowerCamelCase ) for row in temp_grid]
snake_case : Tuple = "".join(__lowerCamelCase )
return output_string
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : int ):
snake_case : Dict = []
snake_case : Union[str, Any] = key - 1
if key <= 0:
raise ValueError("Height of grid can't be 0 or negative" )
if key == 1:
return input_string
snake_case : list[list[str]] = [[] for _ in range(__lowerCamelCase )] # generates template
for position in range(len(__lowerCamelCase ) ):
snake_case : List[str] = position % (lowest * 2) # puts it in bounds
snake_case : Optional[int] = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
temp_grid[num].append("*" )
snake_case : Tuple = 0
for row in temp_grid: # fills in the characters
snake_case : Union[str, Any] = input_string[counter : counter + len(__lowerCamelCase )]
grid.append(list(__lowerCamelCase ) )
counter += len(__lowerCamelCase )
snake_case : str = "" # reads as zigzag
for position in range(len(__lowerCamelCase ) ):
snake_case : Optional[int] = position % (lowest * 2) # puts it in bounds
snake_case : Tuple = min(__lowerCamelCase , lowest * 2 - num ) # creates zigzag pattern
output_string += grid[num][0]
grid[num].pop(0 )
return output_string
def UpperCamelCase ( __lowerCamelCase : str ):
snake_case : Tuple = {}
for key_guess in range(1 , len(__lowerCamelCase ) ): # tries every key
snake_case : Any = decrypt(__lowerCamelCase , __lowerCamelCase )
return results
if __name__ == "__main__":
import doctest
doctest.testmod()
| 59
| 0
|
"""simple docstring"""
from __future__ import annotations
from collections import namedtuple
def __UpperCAmelCase ( __lowerCamelCase , __lowerCamelCase , __lowerCamelCase ) -> tuple:
lowercase__ : Any = namedtuple('''result''' , '''name value''' )
if (voltage, current, power).count(0 ) != 1:
raise ValueError('''Only one argument must be 0''' )
elif power < 0:
raise ValueError(
'''Power cannot be negative in any electrical/electronics system''' )
elif voltage == 0:
return result('''voltage''' , power / current )
elif current == 0:
return result('''current''' , power / voltage )
elif power == 0:
return result('''power''' , float(round(abs(voltage * current ) , 2 ) ) )
else:
raise ValueError('''Exactly one argument must be 0''' )
if __name__ == "__main__":
import doctest
doctest.testmod()
| 16
|
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_sentencepiece_available,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
if is_sentencepiece_available():
from ..ta.tokenization_ta import TaTokenizer
else:
from ...utils.dummy_sentencepiece_objects import TaTokenizer
__lowerCamelCase = TaTokenizer
if is_tokenizers_available():
from ..ta.tokenization_ta_fast import TaTokenizerFast
else:
from ...utils.dummy_tokenizers_objects import TaTokenizerFast
__lowerCamelCase = TaTokenizerFast
__lowerCamelCase = {"""configuration_mt5""": ["""MT5Config""", """MT5OnnxConfig"""]}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = [
"""MT5EncoderModel""",
"""MT5ForConditionalGeneration""",
"""MT5ForQuestionAnswering""",
"""MT5Model""",
"""MT5PreTrainedModel""",
"""MT5Stack""",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""TFMT5EncoderModel""", """TFMT5ForConditionalGeneration""", """TFMT5Model"""]
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__lowerCamelCase = ["""FlaxMT5EncoderModel""", """FlaxMT5ForConditionalGeneration""", """FlaxMT5Model"""]
if TYPE_CHECKING:
from .configuration_mta import MTaConfig, MTaOnnxConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mta import (
MTaEncoderModel,
MTaForConditionalGeneration,
MTaForQuestionAnswering,
MTaModel,
MTaPreTrainedModel,
MTaStack,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mta import TFMTaEncoderModel, TFMTaForConditionalGeneration, TFMTaModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_mta import FlaxMTaEncoderModel, FlaxMTaForConditionalGeneration, FlaxMTaModel
else:
import sys
__lowerCamelCase = _LazyModule(
__name__,
globals()["""__file__"""],
_import_structure,
extra_objects={"""MT5Tokenizer""": MTaTokenizer, """MT5TokenizerFast""": MTaTokenizerFast},
module_spec=__spec__,
)
| 59
| 0
|
"""simple docstring"""
from math import pow, sqrt
def _A ( *UpperCamelCase_ : float) -> bool:
'''simple docstring'''
__lowercase = len(UpperCamelCase_) > 0 and all(value > 0.0 for value in values)
return result
def _A ( UpperCamelCase_ : float, UpperCamelCase_ : float) -> float | ValueError:
'''simple docstring'''
return (
round(sqrt(molar_mass_a / molar_mass_a), 6)
if validate(UpperCamelCase_, UpperCamelCase_)
else ValueError("Input Error: Molar mass values must greater than 0.")
)
def _A ( UpperCamelCase_ : float, UpperCamelCase_ : float, UpperCamelCase_ : float) -> float | ValueError:
'''simple docstring'''
return (
round(effusion_rate * sqrt(molar_mass_a / molar_mass_a), 6)
if validate(UpperCamelCase_, UpperCamelCase_, UpperCamelCase_)
else ValueError(
"Input Error: Molar mass and effusion rate values must greater than 0.")
)
def _A ( UpperCamelCase_ : float, UpperCamelCase_ : float, UpperCamelCase_ : float) -> float | ValueError:
'''simple docstring'''
return (
round(effusion_rate / sqrt(molar_mass_a / molar_mass_a), 6)
if validate(UpperCamelCase_, UpperCamelCase_, UpperCamelCase_)
else ValueError(
"Input Error: Molar mass and effusion rate values must greater than 0.")
)
def _A ( UpperCamelCase_ : float, UpperCamelCase_ : float, UpperCamelCase_ : float) -> float | ValueError:
'''simple docstring'''
return (
round(molar_mass / pow(effusion_rate_a / effusion_rate_a, 2), 6)
if validate(UpperCamelCase_, UpperCamelCase_, UpperCamelCase_)
else ValueError(
"Input Error: Molar mass and effusion rate values must greater than 0.")
)
def _A ( UpperCamelCase_ : float, UpperCamelCase_ : float, UpperCamelCase_ : float) -> float | ValueError:
'''simple docstring'''
return (
round(pow(effusion_rate_a / effusion_rate_a, 2) / molar_mass, 6)
if validate(UpperCamelCase_, UpperCamelCase_, UpperCamelCase_)
else ValueError(
"Input Error: Molar mass and effusion rate values must greater than 0.")
)
| 17
|
import os
import shutil
from pathlib import Path
from typing import Optional, Union
import numpy as np
from huggingface_hub import hf_hub_download
from ..utils import ONNX_EXTERNAL_WEIGHTS_NAME, ONNX_WEIGHTS_NAME, is_onnx_available, logging
if is_onnx_available():
import onnxruntime as ort
__lowerCamelCase = logging.get_logger(__name__)
__lowerCamelCase = {
"""tensor(bool)""": np.bool_,
"""tensor(int8)""": np.inta,
"""tensor(uint8)""": np.uinta,
"""tensor(int16)""": np.intaa,
"""tensor(uint16)""": np.uintaa,
"""tensor(int32)""": np.intaa,
"""tensor(uint32)""": np.uintaa,
"""tensor(int64)""": np.intaa,
"""tensor(uint64)""": np.uintaa,
"""tensor(float16)""": np.floataa,
"""tensor(float)""": np.floataa,
"""tensor(double)""": np.floataa,
}
class UpperCAmelCase :
def __init__(self : Optional[Any] , snake_case__ : Optional[Any]=None , **snake_case__ : Optional[Any] ) -> List[str]:
'''simple docstring'''
logger.info("`diffusers.OnnxRuntimeModel` is experimental and might change in the future." )
snake_case : Optional[Any] = model
snake_case : Dict = kwargs.get("model_save_dir" , snake_case__ )
snake_case : int = kwargs.get("latest_model_name" , snake_case__ )
def __call__(self : Tuple , **snake_case__ : str ) -> List[str]:
'''simple docstring'''
snake_case : Union[str, Any] = {k: np.array(snake_case__ ) for k, v in kwargs.items()}
return self.model.run(snake_case__ , snake_case__ )
@staticmethod
def _SCREAMING_SNAKE_CASE (snake_case__ : Union[str, Path] , snake_case__ : Optional[int]=None , snake_case__ : Optional[int]=None ) -> Any:
'''simple docstring'''
if provider is None:
logger.info("No onnxruntime provider specified, using CPUExecutionProvider" )
snake_case : Optional[int] = "CPUExecutionProvider"
return ort.InferenceSession(snake_case__ , providers=[provider] , sess_options=snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : Union[str, Path] , snake_case__ : Optional[str] = None , **snake_case__ : Any ) -> List[Any]:
'''simple docstring'''
snake_case : Tuple = file_name if file_name is not None else ONNX_WEIGHTS_NAME
snake_case : Any = self.model_save_dir.joinpath(self.latest_model_name )
snake_case : str = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
snake_case : List[str] = self.model_save_dir.joinpath(snake_case__ )
if src_path.exists():
snake_case : Tuple = Path(snake_case__ ).joinpath(snake_case__ )
try:
shutil.copyfile(snake_case__ , snake_case__ )
except shutil.SameFileError:
pass
def _SCREAMING_SNAKE_CASE (self : str , snake_case__ : Union[str, os.PathLike] , **snake_case__ : Optional[int] , ) -> str:
'''simple docstring'''
if os.path.isfile(snake_case__ ):
logger.error(f"""Provided path ({save_directory}) should be a directory, not a file""" )
return
os.makedirs(snake_case__ , exist_ok=snake_case__ )
# saving model weights/files
self._save_pretrained(snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Tuple , snake_case__ : Union[str, Path] , snake_case__ : Optional[Union[bool, str, None]] = None , snake_case__ : Optional[Union[str, None]] = None , snake_case__ : bool = False , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , snake_case__ : Optional["ort.SessionOptions"] = None , **snake_case__ : Tuple , ) -> Tuple:
'''simple docstring'''
snake_case : List[str] = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(snake_case__ ):
snake_case : Any = OnnxRuntimeModel.load_model(
os.path.join(snake_case__ , snake_case__ ) , provider=snake_case__ , sess_options=snake_case__ )
snake_case : Union[str, Any] = Path(snake_case__ )
# load model from hub
else:
# download model
snake_case : Dict = hf_hub_download(
repo_id=snake_case__ , filename=snake_case__ , use_auth_token=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , )
snake_case : List[Any] = Path(snake_case__ ).parent
snake_case : Union[str, Any] = Path(snake_case__ ).name
snake_case : Dict = OnnxRuntimeModel.load_model(snake_case__ , provider=snake_case__ , sess_options=snake_case__ )
return cls(model=snake_case__ , **snake_case__ )
@classmethod
def _SCREAMING_SNAKE_CASE (cls : Optional[Any] , snake_case__ : Union[str, Path] , snake_case__ : bool = True , snake_case__ : Optional[str] = None , snake_case__ : Optional[str] = None , **snake_case__ : Dict , ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = None
if len(str(snake_case__ ).split("@" ) ) == 2:
snake_case , snake_case : int = model_id.split("@" )
return cls._from_pretrained(
model_id=snake_case__ , revision=snake_case__ , cache_dir=snake_case__ , force_download=snake_case__ , use_auth_token=snake_case__ , **snake_case__ , )
| 59
| 0
|
class a__ :
def __init__( self : List[str],_A : list ):
"""simple docstring"""
SCREAMING_SNAKE_CASE_ : Union[str, Any] = set_counts
SCREAMING_SNAKE_CASE_ : List[Any] = max(_A )
SCREAMING_SNAKE_CASE_ : str = len(_A )
SCREAMING_SNAKE_CASE_ : List[str] = [1] * num_sets
SCREAMING_SNAKE_CASE_ : Union[str, Any] = list(range(_A ) )
def __UpperCamelCase ( self : Union[str, Any],_A : int,_A : int ):
"""simple docstring"""
SCREAMING_SNAKE_CASE_ : Optional[Any] = self.get_parent(_A )
SCREAMING_SNAKE_CASE_ : Dict = self.get_parent(_A )
if src_parent == dst_parent:
return False
if self.ranks[dst_parent] >= self.ranks[src_parent]:
self.set_counts[dst_parent] += self.set_counts[src_parent]
SCREAMING_SNAKE_CASE_ : List[str] = 0
SCREAMING_SNAKE_CASE_ : Union[str, Any] = dst_parent
if self.ranks[dst_parent] == self.ranks[src_parent]:
self.ranks[dst_parent] += 1
SCREAMING_SNAKE_CASE_ : str = self.set_counts[dst_parent]
else:
self.set_counts[src_parent] += self.set_counts[dst_parent]
SCREAMING_SNAKE_CASE_ : Tuple = 0
SCREAMING_SNAKE_CASE_ : str = src_parent
SCREAMING_SNAKE_CASE_ : Dict = self.set_counts[src_parent]
SCREAMING_SNAKE_CASE_ : Optional[Any] = max(self.max_set,_A )
return True
def __UpperCamelCase ( self : Any,_A : int ):
"""simple docstring"""
if self.parents[disj_set] == disj_set:
return disj_set
SCREAMING_SNAKE_CASE_ : str = self.get_parent(self.parents[disj_set] )
return self.parents[disj_set]
| 18
|
import argparse
import json
from dataclasses import dataclass, field
from functools import partial
from pathlib import Path
from typing import Callable, Dict, List, Tuple
import timm
import torch
import torch.nn as nn
from classy_vision.models.regnet import RegNet, RegNetParams, RegNetYaagf, RegNetYaagf, RegNetYaaagf
from huggingface_hub import cached_download, hf_hub_url
from torch import Tensor
from vissl.models.model_helpers import get_trunk_forward_outputs
from transformers import AutoImageProcessor, RegNetConfig, RegNetForImageClassification, RegNetModel
from transformers.utils import logging
logging.set_verbosity_info()
__lowerCamelCase = logging.get_logger()
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : List[nn.Module] = field(default_factory=A_ )
A__ : list = field(default_factory=A_ )
def _SCREAMING_SNAKE_CASE (self : Optional[Any] , snake_case__ : Optional[Any] , snake_case__ : Tensor , snake_case__ : Tensor ) -> Optional[Any]:
'''simple docstring'''
snake_case : List[str] = len(list(m.modules() ) ) == 1 or isinstance(snake_case__ , nn.Convad ) or isinstance(snake_case__ , nn.BatchNormad )
if has_not_submodules:
self.traced.append(snake_case__ )
def __call__(self : List[Any] , snake_case__ : Tensor ) -> List[Any]:
'''simple docstring'''
for m in self.module.modules():
self.handles.append(m.register_forward_hook(self._forward_hook ) )
self.module(snake_case__ )
[x.remove() for x in self.handles]
return self
@property
def _SCREAMING_SNAKE_CASE (self : int ) -> Optional[int]:
'''simple docstring'''
return list(filter(lambda snake_case__ : len(list(x.state_dict().keys() ) ) > 0 , self.traced ) )
@dataclass
class UpperCAmelCase :
A__ : nn.Module
A__ : nn.Module
A__ : int = 1
A__ : List = field(default_factory=A_ )
A__ : List = field(default_factory=A_ )
A__ : bool = True
def __call__(self : List[Any] , snake_case__ : Tensor ) -> Any:
'''simple docstring'''
snake_case : str = Tracker(self.dest )(snake_case__ ).parametrized
snake_case : Optional[int] = Tracker(self.src )(snake_case__ ).parametrized
snake_case : List[str] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.src_skip , snake_case__ ) )
snake_case : Optional[Any] = list(filter(lambda snake_case__ : type(snake_case__ ) not in self.dest_skip , snake_case__ ) )
if len(snake_case__ ) != len(snake_case__ ) and self.raise_if_mismatch:
raise Exception(
f"""Numbers of operations are different. Source module has {len(snake_case__ )} operations while"""
f""" destination module has {len(snake_case__ )}.""" )
for dest_m, src_m in zip(snake_case__ , snake_case__ ):
dest_m.load_state_dict(src_m.state_dict() )
if self.verbose == 1:
print(f"""Transfered from={src_m} to={dest_m}""" )
class UpperCAmelCase ( nn.Module ):
def __init__(self : Tuple , snake_case__ : nn.Module ) -> Optional[Any]:
'''simple docstring'''
super().__init__()
snake_case : List[Tuple[str, nn.Module]] = []
# - get the stem
feature_blocks.append(("conv1", model.stem) )
# - get all the feature blocks
for k, v in model.trunk_output.named_children():
assert k.startswith("block" ), f"""Unexpected layer name {k}"""
snake_case : Union[str, Any] = len(snake_case__ ) + 1
feature_blocks.append((f"""res{block_index}""", v) )
snake_case : Optional[Any] = nn.ModuleDict(snake_case__ )
def _SCREAMING_SNAKE_CASE (self : Tuple , snake_case__ : Tensor ) -> Dict:
'''simple docstring'''
return get_trunk_forward_outputs(
snake_case__ , out_feat_keys=snake_case__ , feature_blocks=self._feature_blocks , )
class UpperCAmelCase ( A_ ):
def _SCREAMING_SNAKE_CASE (self : Any , snake_case__ : str ) -> str:
'''simple docstring'''
snake_case : List[Any] = x.split("-" )
return x_split[0] + x_split[1] + "_" + "".join(x_split[2:] )
def __getitem__(self : Optional[int] , snake_case__ : str ) -> Callable[[], Tuple[nn.Module, Dict]]:
'''simple docstring'''
if x not in self:
snake_case : Dict = self.convert_name_to_timm(snake_case__ )
snake_case : Union[str, Any] = partial(lambda: (timm.create_model(snake_case__ , pretrained=snake_case__ ).eval(), None) )
else:
snake_case : List[str] = super().__getitem__(snake_case__ )
return val
class UpperCAmelCase ( A_ ):
def __getitem__(self : Dict , snake_case__ : str ) -> Callable[[], nn.Module]:
'''simple docstring'''
if "seer" in x and "in1k" not in x:
snake_case : str = RegNetModel
else:
snake_case : Optional[Any] = RegNetForImageClassification
return val
def UpperCamelCase ( __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : List[Tuple[str, str]] ):
for from_key, to_key in keys:
snake_case : str = from_state_dict[from_key].clone()
print(f"""Copied key={from_key} to={to_key}""" )
return to_state_dict
def UpperCamelCase ( __lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : Callable[[], nn.Module] , __lowerCamelCase : RegNetConfig , __lowerCamelCase : Path , __lowerCamelCase : bool = True , ):
print(f"""Converting {name}...""" )
with torch.no_grad():
snake_case , snake_case : int = from_model_func()
snake_case : str = our_model_func(__lowerCamelCase ).eval()
snake_case : int = ModuleTransfer(src=__lowerCamelCase , dest=__lowerCamelCase , raise_if_mismatch=__lowerCamelCase )
snake_case : Dict = torch.randn((1, 3, 224, 224) )
module_transfer(__lowerCamelCase )
if from_state_dict is not None:
snake_case : str = []
# for seer - in1k finetuned we have to manually copy the head
if "seer" in name and "in1k" in name:
snake_case : Tuple = [("0.clf.0.weight", "classifier.1.weight"), ("0.clf.0.bias", "classifier.1.bias")]
snake_case : Optional[Any] = manually_copy_vissl_head(__lowerCamelCase , our_model.state_dict() , __lowerCamelCase )
our_model.load_state_dict(__lowerCamelCase )
snake_case : Any = our_model(__lowerCamelCase , output_hidden_states=__lowerCamelCase )
snake_case : Union[str, Any] = (
our_outputs.logits if isinstance(__lowerCamelCase , __lowerCamelCase ) else our_outputs.last_hidden_state
)
snake_case : Union[str, Any] = from_model(__lowerCamelCase )
snake_case : Dict = from_output[-1] if type(__lowerCamelCase ) is list else from_output
# now since I don't want to use any config files, vissl seer model doesn't actually have an head, so let's just check the last hidden state
if "seer" in name and "in1k" in name:
snake_case : Any = our_outputs.hidden_states[-1]
assert torch.allclose(__lowerCamelCase , __lowerCamelCase ), "The model logits don't match the original one."
if push_to_hub:
our_model.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add model" , use_temp_dir=__lowerCamelCase , )
snake_case : List[str] = 224 if "seer" not in name else 384
# we can use the convnext one
snake_case : int = AutoImageProcessor.from_pretrained("facebook/convnext-base-224-22k-1k" , size=__lowerCamelCase )
image_processor.push_to_hub(
repo_path_or_name=save_directory / name , commit_message="Add image processor" , use_temp_dir=__lowerCamelCase , )
print(f"""Pushed {name}""" )
def UpperCamelCase ( __lowerCamelCase : Path , __lowerCamelCase : str = None , __lowerCamelCase : bool = True ):
snake_case : Union[str, Any] = "imagenet-1k-id2label.json"
snake_case : List[str] = 1000
snake_case : List[str] = (1, num_labels)
snake_case : Any = "huggingface/label-files"
snake_case : List[str] = num_labels
snake_case : Optional[Any] = json.load(open(cached_download(hf_hub_url(__lowerCamelCase , __lowerCamelCase , repo_type="dataset" ) ) , "r" ) )
snake_case : List[Any] = {int(__lowerCamelCase ): v for k, v in idalabel.items()}
snake_case : str = idalabel
snake_case : List[Any] = {v: k for k, v in idalabel.items()}
snake_case : Dict = partial(__lowerCamelCase , num_labels=__lowerCamelCase , idalabel=__lowerCamelCase , labelaid=__lowerCamelCase )
snake_case : Optional[Any] = {
"regnet-x-002": ImageNetPreTrainedConfig(
depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 , layer_type="x" ),
"regnet-x-004": ImageNetPreTrainedConfig(
depths=[1, 2, 7, 12] , hidden_sizes=[32, 64, 160, 384] , groups_width=16 , layer_type="x" ),
"regnet-x-006": ImageNetPreTrainedConfig(
depths=[1, 3, 5, 7] , hidden_sizes=[48, 96, 240, 528] , groups_width=24 , layer_type="x" ),
"regnet-x-008": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 5] , hidden_sizes=[64, 128, 288, 672] , groups_width=16 , layer_type="x" ),
"regnet-x-016": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 2] , hidden_sizes=[72, 168, 408, 912] , groups_width=24 , layer_type="x" ),
"regnet-x-032": ImageNetPreTrainedConfig(
depths=[2, 6, 15, 2] , hidden_sizes=[96, 192, 432, 1008] , groups_width=48 , layer_type="x" ),
"regnet-x-040": ImageNetPreTrainedConfig(
depths=[2, 5, 14, 2] , hidden_sizes=[80, 240, 560, 1360] , groups_width=40 , layer_type="x" ),
"regnet-x-064": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 392, 784, 1624] , groups_width=56 , layer_type="x" ),
"regnet-x-080": ImageNetPreTrainedConfig(
depths=[2, 5, 15, 1] , hidden_sizes=[80, 240, 720, 1920] , groups_width=120 , layer_type="x" ),
"regnet-x-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 , layer_type="x" ),
"regnet-x-160": ImageNetPreTrainedConfig(
depths=[2, 6, 13, 1] , hidden_sizes=[256, 512, 896, 2048] , groups_width=128 , layer_type="x" ),
"regnet-x-320": ImageNetPreTrainedConfig(
depths=[2, 7, 13, 1] , hidden_sizes=[336, 672, 1344, 2520] , groups_width=168 , layer_type="x" ),
# y variant
"regnet-y-002": ImageNetPreTrainedConfig(depths=[1, 1, 4, 7] , hidden_sizes=[24, 56, 152, 368] , groups_width=8 ),
"regnet-y-004": ImageNetPreTrainedConfig(
depths=[1, 3, 6, 6] , hidden_sizes=[48, 104, 208, 440] , groups_width=8 ),
"regnet-y-006": ImageNetPreTrainedConfig(
depths=[1, 3, 7, 4] , hidden_sizes=[48, 112, 256, 608] , groups_width=16 ),
"regnet-y-008": ImageNetPreTrainedConfig(
depths=[1, 3, 8, 2] , hidden_sizes=[64, 128, 320, 768] , groups_width=16 ),
"regnet-y-016": ImageNetPreTrainedConfig(
depths=[2, 6, 17, 2] , hidden_sizes=[48, 120, 336, 888] , groups_width=24 ),
"regnet-y-032": ImageNetPreTrainedConfig(
depths=[2, 5, 13, 1] , hidden_sizes=[72, 216, 576, 1512] , groups_width=24 ),
"regnet-y-040": ImageNetPreTrainedConfig(
depths=[2, 6, 12, 2] , hidden_sizes=[128, 192, 512, 1088] , groups_width=64 ),
"regnet-y-064": ImageNetPreTrainedConfig(
depths=[2, 7, 14, 2] , hidden_sizes=[144, 288, 576, 1296] , groups_width=72 ),
"regnet-y-080": ImageNetPreTrainedConfig(
depths=[2, 4, 10, 1] , hidden_sizes=[168, 448, 896, 2016] , groups_width=56 ),
"regnet-y-120": ImageNetPreTrainedConfig(
depths=[2, 5, 11, 1] , hidden_sizes=[224, 448, 896, 2240] , groups_width=112 ),
"regnet-y-160": ImageNetPreTrainedConfig(
depths=[2, 4, 11, 1] , hidden_sizes=[224, 448, 1232, 3024] , groups_width=112 ),
"regnet-y-320": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
# models created by SEER -> https://arxiv.org/abs/2202.08360
"regnet-y-320-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer": RegNetConfig(depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer": RegNetConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer": RegNetConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
# finetuned on imagenet
"regnet-y-320-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[232, 696, 1392, 3712] , groups_width=232 ),
"regnet-y-640-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 5, 12, 1] , hidden_sizes=[328, 984, 1968, 4920] , groups_width=328 ),
"regnet-y-1280-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[528, 1056, 2904, 7392] , groups_width=264 ),
"regnet-y-2560-seer-in1k": ImageNetPreTrainedConfig(
depths=[3, 7, 16, 1] , hidden_sizes=[640, 1696, 2544, 5088] , groups_width=640 ),
"regnet-y-10b-seer-in1k": ImageNetPreTrainedConfig(
depths=[2, 7, 17, 1] , hidden_sizes=[2020, 4040, 11110, 28280] , groups_width=1010 ),
}
snake_case : Union[str, Any] = NameToOurModelFuncMap()
snake_case : str = NameToFromModelFuncMap()
# add seer weights logic
def load_using_classy_vision(__lowerCamelCase : str , __lowerCamelCase : Callable[[], nn.Module] ) -> Tuple[nn.Module, Dict]:
snake_case : List[Any] = torch.hub.load_state_dict_from_url(__lowerCamelCase , model_dir=str(__lowerCamelCase ) , map_location="cpu" )
snake_case : Dict = model_func()
# check if we have a head, if yes add it
snake_case : str = files["classy_state_dict"]["base_model"]["model"]
snake_case : Dict = model_state_dict["trunk"]
model.load_state_dict(__lowerCamelCase )
return model.eval(), model_state_dict["heads"]
# pretrained
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet32d/seer_regnet32gf_model_iteration244000.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Optional[int] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet64/seer_regnet64gf_model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : List[str] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/swav_ig1b_regnet128Gf_cnstant_bs32_node16_sinkhorn10_proto16k_syncBN64_warmup8k/model_final_checkpoint_phase0.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_regnet10B/model_iteration124500_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
# IN1K finetuned
snake_case : List[Any] = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet32_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : Tuple = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet64_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaagf() ) , )
snake_case : str = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_regnet128_finetuned_in1k_model_final_checkpoint_phase78.torch" , lambda: FakeRegNetVisslWrapper(RegNetYaaagf() ) , )
snake_case : Dict = partial(
__lowerCamelCase , "https://dl.fbaipublicfiles.com/vissl/model_zoo/seer_finetuned/seer_10b_finetuned_in1k_model_phase28_conso.torch" , lambda: FakeRegNetVisslWrapper(
RegNet(RegNetParams(depth=27 , group_width=1010 , w_a=1744 , w_a=620.83 , w_m=2.52 ) ) ) , )
if model_name:
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , names_to_config[model_name] , __lowerCamelCase , __lowerCamelCase , )
else:
for model_name, config in names_to_config.items():
convert_weight_and_push(
__lowerCamelCase , names_to_from_model_map[model_name] , names_to_ours_model_map[model_name] , __lowerCamelCase , __lowerCamelCase , __lowerCamelCase , )
return config, expected_shape
if __name__ == "__main__":
__lowerCamelCase = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"""--model_name""",
default=None,
type=str,
help=(
"""The name of the model you wish to convert, it must be one of the supported regnet* architecture,"""
""" currently: regnetx-*, regnety-*. If `None`, all of them will the converted."""
),
)
parser.add_argument(
"""--pytorch_dump_folder_path""",
default=None,
type=Path,
required=True,
help="""Path to the output PyTorch model directory.""",
)
parser.add_argument(
"""--push_to_hub""",
default=True,
type=bool,
required=False,
help="""If True, push model and image processor to the hub.""",
)
__lowerCamelCase = parser.parse_args()
__lowerCamelCase = args.pytorch_dump_folder_path
pytorch_dump_folder_path.mkdir(exist_ok=True, parents=True)
convert_weights_and_push(pytorch_dump_folder_path, args.model_name, args.push_to_hub)
| 59
| 0
|
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_torch
if is_torch_available():
import torch
from transformers.generation import DisjunctiveConstraint
@require_torch
class _SCREAMING_SNAKE_CASE ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE_( self ) -> Optional[int]:
# For consistency across different places the DisjunctiveConstraint is called,
# dc.token_ids is a list of integers. It is also initialized only by integers.
lowerCamelCase_ = [[1, 2, 4], [1, 2, 3, 4]]
lowerCamelCase_ = DisjunctiveConstraint(lowercase )
self.assertTrue(isinstance(dc.token_ids , lowercase ) )
with self.assertRaises(lowercase ):
DisjunctiveConstraint(torch.LongTensor([[1, 2, 4], [1, 2, 3]] ) )
with self.assertRaises(lowercase ):
DisjunctiveConstraint([torch.LongTensor([1, 2, 4] ), torch.LongTensor([1, 2, 3, 4, 5] )] )
def SCREAMING_SNAKE_CASE_( self ) -> Union[str, Any]:
# We can't have constraints that are complete subsets of another. This leads to a preverse
# interpretation of "constraint fulfillment": does generating [1,2,3] fulfill the constraint?
# It would mean that it generated [1,2] which fulfills it, but it's in the middle of potentially
# fulfilling [1,2,3,4]. If we believe that [1,2,3] does fulfill the constraint, then the algorithm
# will necessarily never reach [1,2,3,4], giving users a false sense of control (better to just not allow it).
lowerCamelCase_ = [[1, 2], [1, 2, 3, 4]]
with self.assertRaises(lowercase ):
DisjunctiveConstraint(lowercase ) # fails here
def SCREAMING_SNAKE_CASE_( self ) -> Optional[Any]:
lowerCamelCase_ = [[1, 2, 3], [1, 2, 4]]
lowerCamelCase_ = DisjunctiveConstraint(lowercase )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(1 )
lowerCamelCase_ = stepped is True and completed is False and reset is False
self.assertTrue(lowercase )
self.assertTrue(not dc.completed )
self.assertTrue(dc.current_seq == [1] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(2 )
lowerCamelCase_ = stepped is True and completed is False and reset is False
self.assertTrue(lowercase )
self.assertTrue(not dc.completed )
self.assertTrue(dc.current_seq == [1, 2] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(3 )
lowerCamelCase_ = stepped is True and completed is True and reset is False
self.assertTrue(lowercase )
self.assertTrue(dc.completed ) # Completed!
self.assertTrue(dc.current_seq == [1, 2, 3] )
def SCREAMING_SNAKE_CASE_( self ) -> List[Any]:
lowerCamelCase_ = [[1, 2, 3], [1, 2, 4, 5], [1, 2, 5]]
lowerCamelCase_ = DisjunctiveConstraint(lowercase )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(1 )
self.assertTrue(not dc.completed )
self.assertTrue(dc.current_seq == [1] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(2 )
self.assertTrue(not dc.completed )
self.assertTrue(dc.current_seq == [1, 2] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(4 )
self.assertTrue(not dc.completed )
self.assertTrue(dc.current_seq == [1, 2, 4] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(5 )
self.assertTrue(dc.completed ) # Completed!
self.assertTrue(dc.current_seq == [1, 2, 4, 5] )
dc.reset()
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(1 )
self.assertTrue(not dc.completed )
self.assertTrue(dc.remaining() == 3 )
self.assertTrue(dc.current_seq == [1] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(2 )
self.assertTrue(not dc.completed )
self.assertTrue(dc.remaining() == 2 )
self.assertTrue(dc.current_seq == [1, 2] )
lowerCamelCase_ , lowerCamelCase_ , lowerCamelCase_ = dc.update(5 )
self.assertTrue(dc.completed ) # Completed!
self.assertTrue(dc.remaining() == 0 )
self.assertTrue(dc.current_seq == [1, 2, 5] )
| 19
|
import warnings
from typing import Dict
import numpy as np
from ..utils import ExplicitEnum, add_end_docstrings, is_tf_available, is_torch_available
from .base import PIPELINE_INIT_ARGS, GenericTensor, Pipeline
if is_tf_available():
from ..models.auto.modeling_tf_auto import TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if is_torch_available():
from ..models.auto.modeling_auto import MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
def UpperCamelCase ( __lowerCamelCase : List[Any] ):
return 1.0 / (1.0 + np.exp(-_outputs ))
def UpperCamelCase ( __lowerCamelCase : int ):
snake_case : Tuple = np.max(_outputs , axis=-1 , keepdims=__lowerCamelCase )
snake_case : int = np.exp(_outputs - maxes )
return shifted_exp / shifted_exp.sum(axis=-1 , keepdims=__lowerCamelCase )
class UpperCAmelCase ( A_ ):
A__ : Any = "sigmoid"
A__ : str = "softmax"
A__ : int = "none"
@add_end_docstrings(
A_ ,r"\n return_all_scores (`bool`, *optional*, defaults to `False`):\n Whether to return all prediction scores or just the one of the predicted class.\n function_to_apply (`str`, *optional*, defaults to `\"default\"`):\n The function to apply to the model outputs in order to retrieve the scores. Accepts four different values:\n\n - `\"default\"`: if the model has a single label, will apply the sigmoid function on the output. If the model\n has several labels, will apply the softmax function on the output.\n - `\"sigmoid\"`: Applies the sigmoid function on the output.\n - `\"softmax\"`: Applies the softmax function on the output.\n - `\"none\"`: Does not apply any function on the output.\n " ,)
class UpperCAmelCase ( A_ ):
A__ : int = False
A__ : Union[str, Any] = ClassificationFunction.NONE
def __init__(self : List[str] , **snake_case__ : int ) -> str:
'''simple docstring'''
super().__init__(**snake_case__ )
self.check_model_type(
TF_MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING
if self.framework == "tf"
else MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING )
def _SCREAMING_SNAKE_CASE (self : List[Any] , snake_case__ : List[str]=None , snake_case__ : Optional[Any]=None , snake_case__ : Union[str, Any]="" , **snake_case__ : List[str] ) -> Union[str, Any]:
'''simple docstring'''
snake_case : Dict = tokenizer_kwargs
snake_case : List[Any] = {}
if hasattr(self.model.config , "return_all_scores" ) and return_all_scores is None:
snake_case : Optional[int] = self.model.config.return_all_scores
if isinstance(snake_case__ , snake_case__ ) or top_k is None:
snake_case : List[Any] = top_k
snake_case : str = False
elif return_all_scores is not None:
warnings.warn(
"`return_all_scores` is now deprecated, if want a similar functionality use `top_k=None` instead of"
" `return_all_scores=True` or `top_k=1` instead of `return_all_scores=False`." , snake_case__ , )
if return_all_scores:
snake_case : List[str] = None
else:
snake_case : Optional[int] = 1
if isinstance(snake_case__ , snake_case__ ):
snake_case : Dict = ClassificationFunction[function_to_apply.upper()]
if function_to_apply is not None:
snake_case : Optional[int] = function_to_apply
return preprocess_params, {}, postprocess_params
def __call__(self : Dict , *snake_case__ : List[str] , **snake_case__ : int ) -> Optional[int]:
'''simple docstring'''
snake_case : Optional[int] = super().__call__(*snake_case__ , **snake_case__ )
# TODO try and retrieve it in a nicer way from _sanitize_parameters.
snake_case : Tuple = "top_k" not in kwargs
if isinstance(args[0] , snake_case__ ) and _legacy:
# This pipeline is odd, and return a list when single item is run
return [result]
else:
return result
def _SCREAMING_SNAKE_CASE (self : Dict , snake_case__ : Tuple , **snake_case__ : Union[str, Any] ) -> Dict[str, GenericTensor]:
'''simple docstring'''
snake_case : int = self.framework
if isinstance(snake_case__ , snake_case__ ):
return self.tokenizer(**snake_case__ , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ) and len(snake_case__ ) == 1 and isinstance(inputs[0] , snake_case__ ) and len(inputs[0] ) == 2:
# It used to be valid to use a list of list of list for text pairs, keeping this path for BC
return self.tokenizer(
text=inputs[0][0] , text_pair=inputs[0][1] , return_tensors=snake_case__ , **snake_case__ )
elif isinstance(snake_case__ , snake_case__ ):
# This is likely an invalid usage of the pipeline attempting to pass text pairs.
raise ValueError(
"The pipeline received invalid inputs, if you are trying to send text pairs, you can try to send a"
" dictionary `{\"text\": \"My text\", \"text_pair\": \"My pair\"}` in order to send a text pair." )
return self.tokenizer(snake_case__ , return_tensors=snake_case__ , **snake_case__ )
def _SCREAMING_SNAKE_CASE (self : int , snake_case__ : Union[str, Any] ) -> int:
'''simple docstring'''
return self.model(**snake_case__ )
def _SCREAMING_SNAKE_CASE (self : List[str] , snake_case__ : Optional[Any] , snake_case__ : List[str]=None , snake_case__ : Dict=1 , snake_case__ : Tuple=True ) -> str:
'''simple docstring'''
if function_to_apply is None:
if self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels == 1:
snake_case : Tuple = ClassificationFunction.SIGMOID
elif self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels > 1:
snake_case : Tuple = ClassificationFunction.SOFTMAX
elif hasattr(self.model.config , "function_to_apply" ) and function_to_apply is None:
snake_case : Tuple = self.model.config.function_to_apply
else:
snake_case : int = ClassificationFunction.NONE
snake_case : Any = model_outputs["logits"][0]
snake_case : List[str] = outputs.numpy()
if function_to_apply == ClassificationFunction.SIGMOID:
snake_case : Optional[Any] = sigmoid(snake_case__ )
elif function_to_apply == ClassificationFunction.SOFTMAX:
snake_case : Union[str, Any] = softmax(snake_case__ )
elif function_to_apply == ClassificationFunction.NONE:
snake_case : Optional[Any] = outputs
else:
raise ValueError(f"""Unrecognized `function_to_apply` argument: {function_to_apply}""" )
if top_k == 1 and _legacy:
return {"label": self.model.config.idalabel[scores.argmax().item()], "score": scores.max().item()}
snake_case : Optional[int] = [
{"label": self.model.config.idalabel[i], "score": score.item()} for i, score in enumerate(snake_case__ )
]
if not _legacy:
dict_scores.sort(key=lambda snake_case__ : x["score"] , reverse=snake_case__ )
if top_k is not None:
snake_case : Optional[int] = dict_scores[:top_k]
return dict_scores
| 59
| 0
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.